People v. Jackson

Full title: THE PEOPLE, Plaintiff and Respondent, v. KENNETH ALLEN JACKSON et al.…

Court: COURT OF APPEAL OF THE STATE OF CALIFORNIA FIFTH APPELLATE DISTRICT

Date published: Jul 17, 2018

F069966 (Cal. Ct. App. Jul. 17, 2018)

Opinion

F069966

07-17-2018

THE PEOPLE, Plaintiff and Respondent, v. KENNETH ALLEN JACKSON et al., Defendants and Appellants.

Stephen M. Hinkle, under appointment by the Court of Appeal, for Defendant Appellant Kenneth Allen Jackson. Audrey R. Chavez, under appointment by the Court of Appeal, for Defendant and Appellant Alice Waterman. Kamala D. Harris, Attorney General, Michael P. Farrell, Assistant Attorney General, Brian G. Smiley and Laura Wetzel Simpton, Deputy Attorneys General, for Plaintiff and Respondent.

LEVY, Acting P.J.

NOT TO BE PUBLISHED IN THE OFFICIAL REPORTS California Rules of Court, rule 8.1115(a), prohibits courts and parties from citing or relying on opinions not certified for publication or ordered published, except as specified by rule 8.1115(b). This opinion has not been certified for publication or ordered published for purposes of rule 8.1115. (Super. Ct. Nos. SCR013981A & SCR013981B)

OPINION

APPEAL from a judgment of the Superior Court of Madera County. Dale J. Blea, Judge. Stephen M. Hinkle, under appointment by the Court of Appeal, for Defendant Appellant Kenneth Allen Jackson. Audrey R. Chavez, under appointment by the Court of Appeal, for Defendant and Appellant Alice Waterman. Kamala D. Harris, Attorney General, Michael P. Farrell, Assistant Attorney General, Brian G. Smiley and Laura Wetzel Simpton, Deputy Attorneys General, for Plaintiff and Respondent.

-ooOoo-

INTRODUCTION

Appellants Kenneth Allen Jackson and Alice Waterman, husband and wife, were prosecuted together as arsonists for 31 charged fires. The fires occurred in May and June 2013, mostly in and around the Sierra foothills development of Yosemite Lakes Park (Yosemite Lakes), where appellants resided.

A jury found Jackson guilty of arson in 21 of the fires but not guilty in the other 10 fires.¹ (Pen. Code, § 451, subd. (c).)² He was also found guilty of conspiracy to commit arson (§ 182, subd. (a)(1); count 32); battery on a peace officer (§ 243, subd. (b); count 33), and resisting arrest (§ 148, subd. (a)(1); count 34). He received an aggregate prison sentence of 30 years eight months.

1 Jackson was found guilty of arson in counts 5, 6, 9, 10, 11, 13, 14, 15, 16, 19, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30 and 31. He was found not guilty of arson in counts 1, 2, 3, 4, 7, 8, 12, 17, 18, and 20.

2 All future statutory references are to the Penal Code unless otherwise noted.

Waterman was found guilty of arson in six of the fires (§ 451, subd. (c); counts 14, 15, 16, 18, 20 & 31) and guilty of conspiracy to commit arson (§ 182, subd. (a)(1); count 32). She received an aggregate prison sentence of 10 years eight months.

On appeal, appellants argue that the evidence was insufficient to support the arson and conspiracy convictions. They contend that the trial court erred in allowing the prosecution's experts to opine that arson caused some of these fires. They maintain that the trial court erred when it denied their motion for a mistrial based on a discovery violation. They claim the trial court erred in excluding certain third-party culpability evidence that suggested juveniles may have started some of the charged fires. Finally, they assert that the trial court erred when it allowed the prosecution to proceed against Jackson on six counts which were not charged in the complaint.

We find no error in the admission of the prosecution's expert testimony regarding causation. We find no abuse of discretion regarding the discovery violation and the exclusion of certain third-party culpability evidence. We find sufficient evidence supporting the conspiracy convictions (count 32) and a majority of the arson convictions. As to Jackson, we affirm the arson convictions in counts 9, 10, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, and 31. As to Waterman, we affirm the arson convictions in counts 15, 16, 20, and 31. However, based on insufficient evidence, we reverse eight of the arson convictions against Jackson (counts 5, 6, 11, 13, 14, 15, 16, and 19), and two of the arson convictions against Waterman (counts 14 and 18). The reversal of these convictions renders moot the final issue regarding the prosecution of Jackson for the six counts not pled.

The matter is remanded to the trial court for resentencing consistent with this opinion. We otherwise affirm the judgments.

BACKGROUND

This trial spanned 74 days and created a reporter's transcript exceeding 18,000 designated pages. The following is a relevant synopsis of the trial evidence.

I. An Overview Of The Charged Fires.

The 31 charged fires were known generally as the Gold Mine Fires of 2013. As explained below, these 31 fires occurred in an approximate 46-day span from May 11, 2013 through and including June 25, 2013. The last two fires occurred on the same day, with the last one occurring just before appellants were arrested.

Two groups of fires occurred. One group, consisting of 10 fires, was clustered around East Revis Circle in Yosemite Lakes next to and behind appellants' residence. Because of the terrain and location, whoever started these fires did so on foot. We refer to these as the Cluster Fires, which were charged in counts 4-6, 14-16, 18-20 and 31. These fires generally occurred between 4:00 p.m. and 9:00 p.m. Jackson was found guilty in seven of these counts (5, 6, 14, 15, 16, 19 and 31) and Waterman in six (14, 15, 16, 18, 20 and 31).

The second group, consisting of 21 fires, was scattered throughout and outside Yosemite Lakes. Each of these fires originated near a roadside. We refer to these as the Roadside Fires, which were charged in counts 1-3, 7-13, 17, and 21-30. A majority of these occurred on weekends between 10:00 a.m. and 3:00 p.m. Only Jackson was charged for these fires, and he was found guilty of setting 14 of them (counts 9-11, 13, and 21-30).

Personnel from the California Department of Forestry and Fire Prevention (CAL FIRE) typically prepared two reports for these fires; one report came from a responder, who suppressed the fire, and the second report came from an investigator, who analyzed the fire site. Both reports discussed potential causation. The CAL FIRE personnel generally relied upon a cause exclusion analysis (or negative corpus) to determine causation. This technique rules out possible causes.

The trial evidence generally established that the investigators received more training (FI-210) and were more qualified than some of the responders in determining fire causation. For some of these fires, an incendiary device was found and the investigator determined arson. For many of these fires, however, the investigator determined arson without evidence of an incendiary device but after ruling out all other possible causes. Finally, for a minority of these fires, the investigator could not exclude arson (sometimes along with one or more other potential causes).

II. An Overview Of The Roadside Fires.

Between May 11, 2013 and May 22, 2013, six fires started in and around Yosemite Lakes (three were Roadside Fires, counts 1-3, and three were Cluster Fires, counts 4-6). CAL FIRE officials believed that these six fires were set intentionally. The number of suspicious fires at that point was "unusual" during a fire season, even during the drought. Officials installed cameras intending to capture images of all vehicles entering and exiting Yosemite Lakes. By around June 3, 2013, four cameras were in place.

Yosemite Lakes is accessible via two thoroughfares, Yosemite Springs Parkway and Road 400. A camera designated as YSP East was placed at the southeast end of Yosemite Springs Parkway, near the intersection with Highway 41. A camera designated as YSP West was placed at the opposite end of Yosemite Springs Parkway at the intersection with Road 400.

A camera designated as 400 East was placed near the intersection of Road 400 and Road 415 at the northeast corner of Yosemite Lakes. A camera designated as 400 West was placed by Road 400 near Best Way some distance west of the intersection of Road 400 and Yosemite Springs Parkway.

As of June 8, 2013, CAL FIRE began collecting and reviewing camera data. Officials initially decided to analyze camera footage in a two-hour window, one hour before each reported fire and one hour after, but that proved too time consuming. As a result, starting June 18, 2013, CAL FIRE began reviewing camera footage 20 minutes before and 20 minutes after each reported fire.

Initially, CAL FIRE could only capture less than 60 percent of the license plates on vehicles. On June 13, 2013, officials modified the cameras. Thereafter, CAL FIRE had an approximate 95 percent capture rate. CAL FIRE compiled a database of about 7,576 vehicles captured on camera before and after each fire.

The initial five Roadside Fires (counts 1-3, 7 and 8) occurred between May 11 and June 1, 2013. Except for count 8, these fires burned before cameras were installed. Two cameras, YSP East and YSP West were in place when the fire in count 8 occurred, but the YSP West camera was not operating and it had no data. The one operating camera did not have any significant data, and it did not capture any license plates belonging to either Jackson or Waterman. From June 8 through June 25, 2013, the remaining 16 Roadside Fires occurred. Jackson's vehicles were observed all 16 times (usually on camera) typically in a 40-minute window either coming and/or going towards each fire site.³ The jury convicted him of setting 14 of these fires (counts 9-11, 13, and 21-30).

3 For the fire charged in count 29, Jackson's vehicles were never observed on any of the cameras. Instead, surveilling personnel heard Jackson's truck leave his residence, drive toward this fire's ultimate location, and return home all while remaining within audible distance.

For the final 16 Roadside Fires, CAL FIRE fire captain specialist Branden Smith calculated an ignition window for each fire. This was based on when each fire was reported, the time when Jackson's vehicle was observed, and the drive time to the fire site. To determine the drive time, Smith drove between the two applicable locations without stopping. The ignition window represented the time when Jackson would have passed the fire site and when that fire was reported. Smith's analysis assumed that an ignition source was placed, dropped or tossed at each fire's location. His analysis also assumed that Jackson drove at or below the posted speed limit and, when applicable, he did not turn off the road leading towards a fire site.

There were 110 vehicles that appeared in the applicable time window for two or more fires. Based on the license plates, Smith researched these registered owners. For a subset of these 110 vehicles, Smith analyzed when those vehicles would have passed by fire sites. Over time, however, he became less concerned with other vehicles. Other than Jackson's vehicles, no other vehicle could have been involved in all of the final 16 Roadside Fires and Jackson's vehicles could not be excluded for any of these fires. III. The Initial Five Roadside Fires Before Cameras (Counts 1-3, 7 and 8).

The initial five Roadside Fires (counts 1-3, 7 and 8) occurred between May 11 and June 1, 2013. Except for count 8, these fires burned before cameras were installed. Two cameras, YSP East and YSP West were in place when the fire in count 8 occurred, but the YSP West camera was not operating and it had no data. The one operating camera did not have any significant data, and it did not capture any license plates belonging to either Jackson or Waterman. Investigators determined that these five fires were intentionally set and/or arson could not be ruled out. The jury found Jackson not guilty of arson for these five fires.

A firework was recovered at the fire charged in count 2 (which was the first fire chronologically). For the fires charged in counts 1 and 8, cigarette butts were recovered that were the same color (orangish brown) as cigarettes later found in Jackson's possession. Based on burn markings, these recovered cigarette butts appeared to have been incendiary devices.

At trial, CAL FIRE battalion chief Bernie Quinn discussed a mark that he believed was visible on the cigarette butt recovered from the fire charged in count 1. He opined that a match head had caused the mark. He admitted, however, that he had been unable to duplicate those distinctive markings during experiments conducted with Jackson's brand of cigarettes. He also admitted that burning organic material could have caused the mark.

No other circumstantial evidence, other than a correlation to the remaining 16 Roadside Fires discussed below, linked Jackson to these five fires. As noted above, the jury found him not guilty of arson for these fires (counts 1-3, 7, and 8). IV. The Roadside Fires With Camera Footage (Counts 9-13, 17, 21-30).

Jackson owned a unique black 1998 Dodge four-wheel-drive pickup truck with raised suspension and distinctive chrome parts, including vertical chrome exhaust pipes.⁴ He also owned a 2003 Yamaha motorcycle.⁵ Waterman drove a Toyota Sequoia. At trial, Waterman denied driving any other vehicle.

4 Prior to trial, a Madera County Fire mechanic inspected Jackson's truck. It was in good working order and not emitting any sparks.

5 The motorcycle was only mentioned in connection with two fires, counts 10 and 11.

We summarize the relevant facts from each of the final 16 Roadside Fires. We have combined the discussion for those fires occurring on the same day.

A. The three Roadside Fires on June 8, 2013 (Counts 9, 10 and 11).

Three Roadside Fires occurred on June 8, 2013. The fire in count 9 was reported at about 12:25 p.m. and it occurred by Road 415. The fire in count 10 was reported around 2:57 p.m. and occurred by Lilley Mountain Drive. Finally, the fire in count 11 was reported at around 3:17 p.m. and it occurred by John Muir Drive, just minutes away from the fire in count 10. Jackson's vehicles were seen near all three fires.

For the fire in count 9, Jackson's truck was seen driving away from it about 10 minutes after it was reported. When spotted, his truck was a little over three minutes away from the fire site. After this fire was reported, Jackson's truck was also spotted twice more on camera moving towards this fire and then away from it again. Smith calculated a 13-minute ignition window for this fire.

For the fire in count 10, Jackson's motorcycle was spotted on camera at 2:36 p.m. traveling towards this fire site about 21 minutes before this fire was reported. At that time, he was about five minutes away from the fire site. Smith calculated a 15-minute ignition window for this fire.

Jackson's motorcycle was also seen on a private camera system belonging to a resident on John Muir Drive. The motorcycle was seen at about 2:52 p.m. traveling away from the fires in counts 10 and 11. Smith calculated a 25-minute ignition window for the fire charged in count 11.

CAL FIRE captain Tim McCann investigated these three fires and performed a cause exclusion analysis.⁶ He found cigarette butts at the fire site in count 9, but no such evidence at the other sites. He believed at least one cigarette butt had been an improvised incendiary device. After detailing how he examined each site, and after ruling out other potential causes, McCann opined at trial that arson was the cause of all three fires. However, he admitted it was possible that a tossed cigarette from a car had started the fire in count 9.

6 At the time of trial, McCann was a fire apparatus engineer.

The jury found Jackson guilty of arson for these three fires in counts 9, 10 and 11.

B. The two Roadside Fires on June 9, 2013 (Counts 12 and 13).

Two Roadside Fires occurred on June 9, 2013.⁷ The first fire (count 12) occurred in Oakhurst by Highway 41 and was reported at about 10:49 a.m. The other fire (count 13) occurred by Road 400 east of Yosemite Springs Parkway and was reported around 11:35 a.m. Jackson's truck was spotted on camera for both fires.

7 In addition, a Cluster Fire (count 14) also occurred on June 9, 2013. This Cluster Fire burned near appellants' residence and was reported at about 6:50 p.m.

For the fire charged in count 12, Jackson's truck appeared at 10:58 a.m. traveling on Yosemite Springs Parkway away from this fire. He was about 16 and a half minutes from the fire site. Smith calculated a seven-minute ignition window for this fire.

For the fire charged in count 13, Jackson's truck was seen at 11:50 a.m. traveling on Road 400. He was less than two minutes away from this fire. At trial, Smith opined that Jackson's vehicle would have been at this fire's origin around 11:49 a.m. Smith agreed, however, that this fire was reported about 14 minutes before that time.

1. The cause of the fire charged in count 12.

CAL FIRE battalion chief Mark Pimentel investigated the fire charged in count 12.⁸ First responders reported a burgundy colored sedan rapidly leaving the area. A cigarette butt was recovered at the scene and a cigarette lighter was found in the road. The butt was similar in color (orangish brown) to Jackson's brand. Pimentel determined this butt caused this fire. At trial, he said this was an incendiary device and he agreed it was consistent with previous incendiary devices found at earlier fires.

8 At the time, Pimentel was a fire captain specialist.

This cigarette butt subsequently became controversial in this trial. After Pimentel testified, CAL FIRE battalion chief Matthew Gilbert reexamined this evidence (away from the courthouse) and, for the first time, saw a match head. Over a month later, Gilbert testified about this match head and explained it may have shaken loose when the exhibit was transported.⁹ As we discuss later, the prosecution failed to disclose this new evidence to the defense, and the jury was instructed on the discovery violation.

9 The defense established that this exhibit had been stored in an unsealed box and 13 people had access to it. At trial, Pimentel admitted that the butt looked different than when he had first seized it at the fire site but he believed it was still the same item.

The jury found Jackson not guilty of arson for this fire charged in count 12.

2. The cause of the fire charged in count 13.

McCann investigated the fire charged in count 13. He determined a specific origin area, which included "burnt paper remnants." He could not exclude cigarettes and arson. A weathered cigarette butt was found at this site, but it was ruled out as causing the fire.

The jury found Jackson guilty of arson for this fire charged in count 13.

C. The Roadside Fire on June 12, 2013 (Count 17).

The fire charged in count 17 occurred on June 12, 2013, off Yosemite Springs Parkway near Highway 41.¹⁰ It was reported at about 2:51 p.m. Jackson's truck was spotted on camera at 2:36 p.m. moving away from this fire site. This fire site was approximately less than 30 seconds away from where Jackson's truck was seen on camera. Smith calculated a 15-minute ignition window.

10 In addition, two Cluster Fires (counts 18 and 19) also occurred on June 12, 2013. These Cluster Fires burned near appellants' residence and were reported at about 4:38 p.m. and 9:30 p.m., respectively.

CAL FIRE fire captain specialist William Cacho investigated this fire. He found a white colored cigarette, which he believed had caused this fire, but it was unknown whether it was accidental or not. Nothing indicated an incendiary device. A witness reported to Cacho that a white car had sped away from this fire when it was still very small.

The jury found Jackson not guilty for this fire charged in count 17.

D. The three Roadside Fires on June 16, 2013 (Counts 21, 22 and 23).

Three Roadside Fires occurred on June 16, 2013. The fire in count 21 occurred near Road 400 and was reported about 1:30 p.m. Jackson's truck was seen at 1:08 p.m. and 1:13 p.m., first traveling towards this fire site and then away. Smith calculated an ignition window of about 10 minutes.

The fire in count 22 occurred by John Muir Drive and it was reported at 3:35 p.m. Jackson's truck was seen on John Muir Drive at 3:25 p.m. traveling eastbound (away from this fire site). This footage was taken from a resident's camera about three minutes east of this fire. Smith calculated a 13-minute ignition window.

The fire in count 23 occurred along Highway 41 and was reported at about 8:16 p.m. Jackson's truck was spotted at about 8:00 p.m. traveling towards this fire on Yosemite Springs Parkway about four and a half minutes away. Smith calculated an approximate 11-minute ignition window. At trial, Smith agreed that Jackson could have turned off the highway onto another road before reaching this fire. Smith also agreed that this fire occurred on the east side of Highway 41 so that an incendiary device thrown from a moving vehicle would have come from the passenger side.

1. The cause of the fire charged in count 21.

McCann investigated the fire charged in count 21. He found the specific origin area and conducted a cause and exclusion analysis. He ruled out everything except arson. At trial, he agreed he could not rule out arson, in part, based on the recent fire activity in the immediate area. When asked whether he could have excluded arson if he had no information on the previous fires, he replied that there would have to be "a lot of difference there" and he would "have to really take a look at it."

The jury found Jackson guilty of this charge.

2. The cause of the fire charged in count 22.

After the fire charged in count 22 was suppressed, fire officials found part of a catalytic converter in the general area of origin. Photos were taken of it, along with the ground underneath it. McCann investigated this fire. Based on the catalytic converter, he ruled out arson. At trial, McCann explained that, around this time, the community was frustrated and he had felt "general pressure" to find a cause for the fires.

About two weeks later, and after appellants were arrested, McCann changed his opinion about this fire. After talking with colleagues and reexamining the catalytic converter, he determined that dirt, "other fibrous things," or "maybe grass" were lodged in the converter's honeycombs. Based on that and how the grass had burned around it, he concluded the catalytic converter could not have started this fire. He changed his report, ruled out the catalytic converter and he found arson.

The jury found Jackson guilty of this charge.

3. The cause of the fire charged in count 23.

McCann investigated the fire charged in count 23. A vehicle's electronic mirror was recovered in the specific origin area. McCann initially believed the mirror was responsible for this fire. After collecting the mirror, however, McCann reexamined it and he changed his opinion. He realized that the glass on the mirror, which was mostly missing, could not have reflected sunlight because it had been lying face down. The mirror also had no power source. He decided the mirror was not the cause. He listed both vehicles and arson as potential causes for this fire. At trial, McCann denied that somebody pressured him to change his opinion.

The jury found Jackson guilty of the fire charged in count 23.

E. The two Roadside Fires on June 20, 2013 (Counts 24 and 25).

Two Roadside Fires occurred on June 20, 2013. The fire charged in count 24 burned near Road 415 and was reported at 1:46 p.m. Just minutes later, the fire charged in count 25 occurred near Highway 41 just north of Road 416. This fire was reported at 1:58 p.m. One camera captured Jackson's truck around 1:32 p.m. traveling towards these fires. He was about 30 to 45 seconds away from the first fire (count 24). These two fires were separated by a drive time of six and a half to seven minutes. Smith calculated ignition windows of 13 minutes and 20 minutes, respectively.

In connection with these fires, a witness reported a white colored Subaru hatchback driving strangely down Road 400 headed towards Road 415. The Subaru was driving about 30 miles an hour and it would slow down to about 15 miles an hour near sites of old fires on Road 400 as if the driver was admiring them. Papers were seen flying out of this vehicle, and the driver ran the stop sign at Road 415. At some point, the Subaru stopped on Road 415, and the car sped away when the witness approached. The Subaru stopped again, and the witness spotted the fire near Road 415.¹¹

11 At trial, Smith confirmed that, when reviewing the camera footage for these fires, he had observed a white colored Subaru (and possibly also a silver colored Subaru). The white Subaru appeared on camera at 1:43 p.m. (11 minutes after Jackson's truck). This was the only time this particular Subaru appeared on camera for any of the Roadside Fires.

Cacho investigated both fires. He spoke with the witness about the Subaru. Cacho was unable to exclude either a vehicle or arson for the fire in count 24. For count 25, he could not exclude vehicles, a discarded cigarette, or arson.¹²

12 For the fire charged in count 25, Cacho found numerous cigarette butts, but they appeared weathered and sun damaged.

The jury found Jackson guilty of the fires charged in counts 24 and 25.

F. The four Roadside Fires on June 23, 2013 (Counts 26, 27, 28 and 29).

Four Roadside Fires occurred on June 23, 2013. First, the fire charged in count 26 occurred by Highway 41 north of Yosemite Lakes, and was reported at about 10:20 a.m. Second, the fire charged in count 27 occurred about five minutes away by Highway 41, and was reported at about 10:27 a.m. Third, the fire charged in count 28 occurred off Road 400 near Lilley Mountain Drive in Yosemite Lakes, and was reported about 3:31 p.m. Finally, the fire charged in count 29 occurred about a minute's drive from appellants' residence. This fire was reported around 7:25 p.m. Different cameras and eyewitnesses placed Jackson's truck at each of these fires.

1. The evidence for counts 26 and 27.

For the fires charged in counts 26 and 27, Jackson's truck was seen on camera at 9:49 a.m. on Yosemite Springs Parkway generally heading towards these fires. It was an approximate 13-minute and 18-minute drive, respectively, from this camera to these fires. Smith calculated ignition windows of 18 and 20 minutes, respectively.

In addition to the camera evidence, two CAL FIRE employees both saw Jackson's truck by the site of the fire charged in count 27. The employees also saw a white male near this truck flagging down and slowing traffic. In his opening brief, Jackson concedes that he was present at this fire after fire personnel responded to the scene.

Pimentel investigated the fire charged in count 26. Fire suppression had damaged the area and he could not find any ignition source. He did a cause exclusion analysis, and he could not rule out arson or a vehicle. McCann investigated the fire charged in count 27. After a cause exclusion analysis, he ruled out everything but arson.

At trial, Jackson called Martin Powell to testify. Powell owned a Napa store in Coarsegold, which was located on Highway 41 just south of where the fire occurred in count 26, and farther south from the fire in count 27. Powell had a receipt from his business that showed Jackson making a purchase on June 23, 2013, at about 10:24 a.m. Powell believed that Jackson had appeared at his store before 10:00 a.m. that morning. He said Jackson stayed for an unknown duration of time. Powell had a police scanner, and he recalled that Jackson and he heard a call on the scanner for a fire. Jackson and Powell had talked about the fire. Jackson left when other customers entered the store.

In rebuttal, Smith opined that it was not possible for Jackson to have arrived at the Napa store by 9:45 a.m. because he had been spotted on camera on Yosemite Springs Parkway at 9:49 a.m.

The jury found Jackson guilty of the fires charged in counts 26 and 27.

2. The evidence for count 28.

The fire charged in count 28 was reported at about 3:31 p.m., and Jackson's truck was spotted on camera on Yosemite Springs Parkway at 3:16 p.m. heading towards this fire site. It was spotted two minutes later traveling away from it. The travel time from this camera to this fire site was about 30 to 60 seconds. Smith calculated an ignition window of about 13 to 14 minutes.

In addition to the camera evidence, CAL FIRE personnel saw Jackson that day. At about 12:45 p.m., fire captain specialist David LaClair happened to pass Jackson's truck as LaClair was entering the Yosemite Lakes area. LaClair saw Jackson driving. At about 3:19 p.m., Quinn observed Jackson driving his truck on Yosemite Springs Parkway away from the fire charged in count 28. Jackson was again identified as the driver. As he drove, Jackson was seen "looking behind him the whole time" in the direction where this fire originated. No vehicles were behind Jackson when he was looking back. Finally, while responders were suppressing the fire charged in count 28, Jackson parked his truck nearby and told fire personnel that he had observed a suspicious car.

Pimentel investigated the fire charged in count 28. After conducting a cause exclusion analysis, he determined arson. At trial, Pimentel noted that the arsonist may have hiked down a small ravine, which dropped off away from the road. Pimentel, however, did not see any tire tracks in the dirt on the side of the road. He agreed it was possible a hiker caused this fire, but it was also possible that something had been thrown from a passing vehicle.

The jury found Jackson guilty of the fire charged in count 28.

3. The evidence regarding count 29.

For the fire charged in count 29, appellants' vehicles were not spotted on a camera but were heard by surveilling personnel.¹³ At about 7:19 p.m., officers heard Jackson's truck leave his residence. It drove toward this fire's ultimate location, and it returned at about 7:25 p.m. The truck remained within audible distance of fire personnel the entire time. Smith calculated an ignition window of 16 to 20 minutes.

13 Beginning June 11, 2013, CAL FIRE personnel began surveilling appellants' residence. From June 11 through June 25, 2013, appellants' residence was watched nearly every day, with only two days missed.

About two days after this fire, appellants were arrested. Waterman told an investigator that Jackson's truck never left their residence on the evening of this fire. She told LaClair that Jackson did not leave that night. She said an "imaginary" person must have been driving his truck.

McCann investigated the fire charged in count 29. As part of his investigation, he interviewed some juveniles, including M.S. and L.H. M.S. lived less than one block from this fire. McCann went through 13 causes and exclusions, and he was unable to exclude arson. He did not find any evidence suggesting that juveniles were responsible.

The jury found Jackson guilty of this charge.

G. The final Roadside Fire on June 25, 2013 (Count 30).

The final Roadside Fire occurred on June 25, 2013.¹⁴ This occurred by Road 400 south of Lilley Mountain Drive and it was reported at 3:28 p.m. Prior to this fire, fire officials had placed a GPS tracker on Jackson's truck. The GPS data showed Jackson's truck passing this fire's site at 3:05 p.m. and again at 3:08 p.m. Both times, his truck drove at about 37 miles per hour. The camera footage corroborated the GPS data regarding when Jackson's truck approached and left this fire site. Smith opined that this fire had either a 20- or 23-minute ignition window, depending on when Jackson placed an ignition source.

14 In addition, a Cluster Fire (count 31) also occurred on June 25, 2013, near appellants' residence. This fire was spotted around 6:40 p.m. by undercover personnel.

On the day of this fire, CAL FIRE officers, including Quinn, were following Jackson's truck. At about 3:03 p.m., Quinn was parked near Road 400 and he saw Jackson's vehicle pass. At about 3:14 p.m., another surveilling officer confirmed that a white male was driving Jackson's truck. At about 3:28 p.m., Quinn heard the dispatch for this fire. He turned and looked back. He could see a large column of smoke building. Quinn had passed that fire location two times while following Jackson's vehicle, but he had not previously seen any flames or smoke. Quinn had last passed that site sometime after 3:08 p.m.

Gilbert investigated this fire. He performed a cause exclusion analysis and he determined arson. He reached his determination after ruling out other potential causes. He also factored the past fires, along with the camera and GPS data.

The jury found Jackson guilty of this charge.

V. The Prosecution's Theory Of Causation Regarding All Roadside Fires.

After all of the investigators had testified at trial, Gilbert provided a global opinion regarding the Roadside Fires. He believed the camera data linked these fires together and indicated that Jackson had used a time delayed incendiary device to start all 21 of these fires. Gilbert noted that incendiary devices had been located at six of the Roadside Fires, including a firework at one fire and cigarettes at the others. He thought the cigarette butts linked the Roadside Fires together. He also mentioned the apparent match head fragment he saw connected with the cigarette recovered in count 12.

The jury learned that CAL FIRE experts had conducted experiments with Jackson's brand of cigarettes. They stuffed the cigarettes with match heads. These devices could be thrown or placed on the ground with the cigarette lit. When the cigarette burned down to the match head, it would flare.¹⁵ Gilbert opined that such a time delayed incendiary device could take less than five minutes, or up to 20 minutes, to start a wildland fire. Gilbert had never seen one take as long as 20 minutes but he had read about it.

15 An unmodified cigarette can start a fire when the air temperature is above 80 degrees and the relative humidity is between 18 and 22 percent.

At trial, Gilbert theorized that Jackson had made incendiary devices by placing matches or match heads inside cigarettes. He believed that most of these devices were either consumed in the fires or not detected. He noted that fire suppression could have damaged or destroyed this evidence.

VI. Jackson's Statements While In Custody.

Jackson was arrested on June 25, 2013 and he was interviewed later that night. He stated that Waterman usually drove the Toyota Sequoia. He generally denied starting any fires intentionally. He admitted carelessly throwing a finished cigarette out of his truck earlier that day when he was driving. He said it was possible his cigarette had caused a fire.

VII. Jackson's Ability And Possible Motive To Commit Arson.

When searched, Jackson's truck had two lighters and numerous packages of cigarettes and paper matches. Numerous cigarette butts were in the truck's bed. More paper matches and lighters were located in appellants' residence.¹⁶

16 In addition, a magnesium strike block (a tool used to start campfires) was discovered in appellants' kitchen. At trial, however, Gilbert testified that, in his opinion, the magnesium strike block had nothing to do with the charged fires.

In addition, Jackson's truck contained 12 small lead fishing weights in a package originally holding 40. No other fishing gear was located in the truck. At trial, LaClair noted that such weights could be added to a cigarette incendiary device to increase its weight. Gilbert, however, testified that he did not know whether fishing weights had anything to do with the charged fires because no fishing weights were ever found at the fire sites.

Jackson was employed part-time by a CAL FIRE contractor, involved in hauling equipment to and from fire sites. He also drove a big rig truck and worked with his father. He did not have a set schedule. A CAL FIRE incident action plan for one of the charged fires was found in Jackson's truck. This is a CAL FIRE document given to the people assigned to a fire. It gave the radio channels, the assignments for each day, and the plan to complete each assignment. In addition, a notebook listing radio frequencies, some of which firefighters used, and a radio scanner were found in appellants' living room. There was also a reference book used by firefighters and others working at fire scenes. A computer in appellants' residence displayed a picture of a wildfire.

VIII. Concerns About The Case For The Roadside Fires.

At trial, certain concerns were raised regarding the case against Jackson for the Roadside Fires. We summarize those concerns here.

A. Incomplete camera coverage.

When CAL FIRE installed the four surveillance cameras, fire personnel believed they would capture all traffic entering and leaving Yosemite Lakes. Late in its investigation, however, CAL FIRE realized that one entrance/exit point had not been monitored by a camera. This was the intersection of Road 400 and Lilley Mountain Drive, near the southwest portion of the area. CAL FIRE installed a camera at this location on June 24 or 25, 2013. This fifth camera was installed around the time the final two fires in the series, count 30 (a Roadside Fire) and count 31 (a Cluster Fire), occurred on June 25, 2013. Appellants were also arrested on that same day. Before this fifth camera was installed, it was possible for a person to drive into Yosemite Lakes and go nearly everywhere inside that area without being detected by a CAL FIRE camera.¹⁷

17 For the fires charged in counts 10, 11 and 22, CAL FIRE used private camera footage from a resident who lived on John Muir Drive. In addition, surveilling officers near appellants' residence heard Jackson's truck in relationship to the fire charged in count 29.

B. The numerous other vehicles.

As noted earlier, more than 7,000 vehicles were recorded by CAL FIRE's cameras during the time windows before and after each fire. Smith testified that there were 110 vehicles that appeared in the applicable time window for two or more fires. Based on the license plates, Smith researched these registered owners. For a subset of these 110 vehicles, Smith analyzed when those vehicles would have passed by fire sites. Over time, however, he became less concerned with other vehicles. Other than Jackson's vehicles, he determined that no other vehicle could have started all of the final 16 Roadside Fires and Jackson's vehicles could not be excluded for any of these fires. For the final 16 Roadside Fires from June 8 through June 25, 2013, Jackson's vehicles were observed all 16 times (usually on camera) typically in a 40-minute window either coming and/or going towards each fire site.¹⁸ The jury convicted him of setting 14 of these fires (counts 9-11, 13, and 21-30).

18 For the fire charged in count 29, Jackson's vehicles were never observed on any of the cameras. Instead, surveilling personnel heard Jackson's truck leave his residence, drive toward this fire's ultimate location, and return home all while remaining within audible distance.

At trial, Smith admitted that he did not always account for all vehicles. In some instances, he stopped reviewing footage after he saw Jackson's truck. In other instances, he did not watch for additional vehicles on the other cameras or log those other vehicles into the database.

C. The white or silver colored Subaru.

Apart from Jackson's vehicles, Smith specifically testified about one other suspicious vehicle, a light colored Subaru identified by its license plate number. The Subaru appeared on camera about six times for eight Roadside Fires (counts 10, 11, 13, 17, 21, 24, 25, and 30).¹⁹ The Subaru was generally spotted within the applicable time window before or after each fire, but count 21 was an exception; the Subaru was spotted more than an hour before that fire was reported.

19 The captures for the eight fires showed the Subaru's license plate. Smith said no when asked if he included other captures where license plates were not seen.

At trial, Smith explained that, during his investigation, he stopped looking for the Subaru because it had not appeared in all of the target areas and it had limited captures. He admitted, however, he had not reviewed the camera footage for the Subaru until the night before his trial testimony.

At trial, Smith initially opined that an occupant of the Subaru could have started seven of these Roadside Fires, and he explained how the Subaru could have arrived at these fires. He excluded the Subaru in count 30 because it was seen on camera about five minutes after that fire was reported. It was not possible for the Subaru to have been at this fire's location within the ignition window.

For the fire charged in count 21, in which the Subaru was captured more than an hour before that fire was reported, Smith said it was possible that it was involved in this fire, but he did not explain why. Later, Smith denied that the Subaru could have been involved in the fires charged in counts 24 and 25 with an assumed ignition window up to 20 minutes.

For some of the seven fires, Smith was questioned about whether he considered it probable that an occupant of the Subaru started them, but objections were sustained to all but one of these questions. The one question that received no objection was whether Smith considered it probable that the Subaru was connected with the fires charged in counts 24 and 25. He said no, but was not asked for a basis for this opinion. The jury found Jackson guilty in each of the fires (counts 10, 11 13, 21, 24, 25, and 30) in which the Subaru also appeared on camera.

D. Other uncharged roadside fires in the area.

Two uncharged fires occurred on June 17 and 18, 2013. Initially, CAL FIRE included these two additional fires as part of the Gold Mine series. These were known as the Harsh and Hensley fires. On about June 30, 2013, however, CAL FIRE determined that Jackson was not responsible because he had been working out of the area when these two fires had occurred.²⁰

20 In addition, Jackson worked in Bass Lake on June 10, 2013, from 8:00 a.m. until 2:00 or 3:00 p.m. The two Cluster Fires charged in counts 15 and 16 occurred that same night and were reported about 8:38 p.m. The jury also heard that Jackson's employer had called him for work on June 13, 2013, but Jackson had not responded to phone calls and he later told his employer he had been working in Bass Lake with his father. No fires occurred that day.

The Harsh fire occurred on June 17, 2013, beside Highway 41 to the south of Yosemite Lakes. CAL FIRE personnel included arson as a possible cause because there had been other fires in the area. Sparks or hot molten objects from vehicles were included as a possible cause because metal fragments were found. Smoking was also included as a possible cause based on recovered cigarette butts and atmospheric conditions.

The Hensley fire occurred on June 18, 2013, beside Road 400 near Hensley Lake. CAL FIRE personnel concluded that arson caused this fire. They found a firework and all other potential causes were excluded.

As stated during Waterman's closing arguments, the defense presented evidence of the Harsh and Hensley fires to show that CAL FIRE created an artificial fire pattern. Waterman's counsel argued that the prosecution chose the size of its map and which fires to include.

IX. The Cluster Fires.

Gilbert informed the jury that the Cluster Fires generally occurred near appellants' residence in a cluster pattern, all of these fires were away from the road, and they could not have been started from inside or beside a car. Gilbert believed that the Cluster Fires were all started "hot" or without an incendiary device. Gilbert opined that, based on the terrain elevation, these fires posed "minimal danger" to appellants' residence when they occurred.²¹

21 At trial, Waterman denied that her house was in danger from the Cluster Fire charged in count 18. When the prosecutor said, "No?" Waterman answered, "Fire travels uphill."

We summarize these fires, some of which occurred very close together in time and location. For convenience, we have combined the facts for those fires.

A. The Cluster Fire on May 18, 2013 (Count 4).

The fire in count 4 occurred on May 18, 2013, on Revis Road. Fire personnel were dispatched at about 7:30 p.m. This fire burned on private property belonging to Kevin Olsen. Appellants' residence was about a hundred feet away from this fire. Arrows and rocks with strike marks were found in this fire's origin area. Rock strikes could have caused a spark, which could have started a fire. Olsen testified at trial that he had last shot arrows behind his house at least a month before this fire. McCann investigated this fire. Following a cause exclusion analysis, he could not rule out arson or human activity.

The jury found both Jackson and Waterman not guilty in count 4.

B. The two Cluster Fires on May 22, 2013 (Counts 5 and 6).

The fires charged in counts 5 and 6 occurred on May 22, 2013. They were reported at about 8:30 p.m. and they occurred about 100 feet from appellants' residence. These fires started about 10 to 15 feet apart. No ignition source was located, but a weathered lighter was found at an intersection about 600 to 700 yards away.

Responders saw Jackson holding a garden hose about 50 feet away from flames; he said he was planning to extinguish the flames if they came closer to his residence. CAL FIRE received reports that children were heard in the area shortly before these fires started. It was also reported that children were known to frequent this area. About 90 minutes before these fires started, a juvenile, L.H., had started a fire with a lighter in a backyard about half a mile away.

At trial, CAL FIRE fire prevention specialist Darrin McCully opined that these fires had been intentionally set. He determined arson after eliminating other potential causes. On cross-examination, McCully acknowledged that a juvenile had started a fire with a lighter around the time these fires occurred. He told the jury, however, that he had found no evidence suggesting a juvenile was responsible for these fires.

At trial, Quinn opined that a juvenile had not set either of these fires or the one charged in count 4. The arsonist would have walked between or within people's yards, which usually did not have fences. There were at least two dogs in the area, which "aggressively" barked. Quinn did not believe a juvenile could have accessed this area without raising suspicion.

At trial, Waterman testified that she was in Fresno playing in an 8:15 p.m. soccer match when these fires occurred. She usually left no later than 6:15 p.m. to attend her games. Jackson had texted a picture of these fires to her. Several of Waterman's soccer teammates testified at trial, confirming that she had attended the soccer game on the night of these fires.

The jury found Jackson guilty for the fires charged in counts 5 and 6, but it found Waterman not guilty in both counts.

C. The Cluster Fire on June 9, 2013 (Count 14).

The fire charged in count 14 occurred on June 9, 2013, in very close proximity to both the fire charged in count 3 (a Roadside Fire) and the Cluster Fires that occurred in counts 5 and 6 (which had burned in May). Fire personnel responded around 6:50 p.m. Fire investigators did not find anything significant.

At trial, McCann opined that arson caused this fire after he ruled out all other possible causes. He agreed that the other fires in this area (the fires charged in counts 3, 5 and 6) played a role in determining arson.

The day after this fire, McCann spoke with Waterman at her residence. He noticed shrubbery next to their front walkway, which appeared burnt. Waterman told him that "they'd been gone that night" and she had noticed the shrubbery when they got home, but she had not reported it. McCann found her explanation odd. However, he did not examine the cause of this apparent shrubbery fire. At trial, Waterman disputed that the shrubbery had been burned, claiming it just had "blank spots" on it. She believed she had been home when the fire in count 14 occurred, either making or eating dinner.

The jury found both Jackson and Waterman guilty for the fire charged in count 14.

D. The two Cluster Fires on June 10, 2013 (Counts 15 and 16).

The fires charged in counts 15 and 16 both occurred on June 10, 2013, next to each other on Revis Circle East. The fire in count 15 was reported at about 8:38 p.m. Responders arrived and found the other fire also burning. These fires occurred near appellants' residence and they were about 50 feet apart. The larger fire (count 15) occurred "behind a residence" on a "little hill" and the smaller fire (count 16) occurred below it by Revis Circle East. At trial, the responding fire captain, Joseph Felix, who had undergone FI-210 training, opined that the larger fire (count 15) started first.

Pimentel investigated these fires. At trial, he reviewed People's exhibit 47, a photograph of an aerial view of the fire sites. He explained that the fire charged in count 15 (the Circle A fire) was much larger than the fire charged in count 16 (the Circle B fire). The Circle A fire occurred behind appellants' residence (and one other house) and the other fire occurred down the hill closer to the driveway that extended past appellants' residence. He opined that these fires were started separately. A cause exclusion analysis was performed, which ruled out everything but arson.

Shortly after these fires were suppressed, Waterman made statements to fire personnel. She had heard either a bicycle and/or a vehicle. She went outside to investigate and she had observed smoke coming from some grass about 25 feet from the neighbor's driveway. The smoke turned into active flames. When she first saw the fire, it was about the size of a roll of duct tape. She mentioned smelling a chemical odor. She tried to stomp out the fire, but it quickly expanded. She claimed that a vehicle had left the area at a high rate of speed.²²

22 Bystanders to this fire indicated that a teenager on a bicycle and a speeding white pickup had both been observed in the area. Fire personnel, however, did not follow up on this information.

On the evening of these fires, Waterman gave an interview with a television news reporter. She claimed, in part, that she had "just happened to walk up and spot the fire starting" and she had called 911. At trial, however, the television reporter claimed that Waterman had initiated the interview.

Waterman told the jury that she and Jackson were outside walking as they did regularly because of the past fires. She was near her neighbor's driveway when she smelled something funny and a fire suddenly started. She ran to her house and called 911. She saw the second fire start when she hung up with 911. She agreed that she saw one of these fires when it was the size of a baseball. She denied seeking out the television interview, claiming the reporter had first approached her. She could not recall hearing a bicycle before these fires.

At trial, McCann noted that when someone sees a fire beginning, as Waterman had reported, that person is usually at that fire's ignition. McCann told the jury that he had not believed Waterman's explanation of events.

The jury found both Jackson and Waterman guilty of the fires charged in counts 15 and 16.

E. The two Cluster Fires on June 12, 2013 (Counts 18 and 19).

The fires in counts 18 and 19 both occurred on June 12, 2013. The fire in count 18 occurred less than a two-minute walk to appellants' residence, and it was reported at about 4:38 p.m. The fire charged in count 19 was reported at about 9:30 p.m. These fires occurred about 200 yards apart. Using a cause exclusion analysis, Pimentel determined arson caused both fires.²³

23 After the fire in count 18 started, surveilling officers received a report of a suspicious green Nissan in the area. Units found and followed the Nissan for about two hours. The driver eventually went inside a casino, where officers continued to observe him. While watching that driver, the fire in count 19 occurred. As a result, fire officials excluded this driver as an arson suspect.

Waterman's Toyota Sequoia was spotted near her residence about 29 minutes before the first fire was reported. She later told an investigator that a white SUV was in their area just before the first fire occurred. A witness also told the same investigator that a white Chevy truck left the area at an accelerated speed. On a local camera, fire officials later saw a white Ford SUV enter the street at about 4:15 p.m. and leave a minute later.

At trial, Waterman testified that she had left for a soccer match in Fresno between 5:30 and 5:45 p.m. while fire personnel were battling the first fire. She also testified that she returned home around 10:45 p.m. and Jackson was asleep in bed. Several of Waterman's soccer teammates testified at trial, confirming that she had attended the soccer game on the night of this fire. The game had started at about 8:15 p.m. and the players typically arrived no later than 7:45 p.m.

At trial, Jackson's father testified that he and Jackson were in appellants' pool when the first fire started. He saw Waterman several times before the fire started, and she was never near its origin site. Waterman was preparing to leave for a soccer match and she was worried fire responders might block the driveway.

The jury found Jackson not guilty of the fire charged in count 18, but it found Waterman guilty. In contrast, the jury found Jackson guilty of the fire charged in count 19, but it found Waterman not guilty.

F. The Cluster Fire on June 14, 2013 (Count 20).

The fire charged in count 20 occurred on June 14, 2013, near appellants' residence. Fire personnel were dispatched at about 7:30 p.m. Cacho determined that arson caused this fire.

1. LaClair places Waterman at the origin of this fire.

At trial, an eyewitness, LaClair, identified Waterman near the origin of this fire. This testimony, however, raised credibility concerns.

On June 14, 2013, fire personnel were surveilling the area around appellants' residence. One of those officers, LaClair, was wearing a camouflaged outfit (a "ghillie" suit). LaClair saw a suspect walking in the brush. He described the suspect as a white person of unknown gender with black hair, wearing blue shorts, a white shirt, and black sunglasses. The suspect was about 180 pounds, and walked with a waddle, with arms held out. The suspect left his view and a fire started in that same area.

LaClair quickly went down a slope to the fire. As soon as he got to the fire, a female came behind him yelling about fire, and asking who he was and/or what he was doing there. This woman was wearing brown colored clothing and she appeared smaller in size than the suspect had been. After a heated exchange, LaClair left the area while this woman followed him a short distance. He did not initially believe that the suspect and this woman were the same, but, at some point, he realized that the woman was Waterman.

Two days after this encounter, LaClair again saw Waterman while he was conducting surveillance near her residence. Another officer took pictures of Waterman, and the photos were shown to the jury. Waterman was wearing sunglasses, a blue shirt, and blue shorts. She was walking in the area behind a nearby vacant house. She picked up a stick and began beating some tall grass with it. LaClair believed she was looking for concealed people.

Over time, LaClair determined that Waterman was the suspect. At trial, he asserted that the shorts, sunglasses and shoes in Waterman's pictures matched what the suspect had worn. He also believed that Waterman's posturing, her walk and her hair color were similar to the suspect's. At trial, he confirmed that the suspect had weighed 180 pounds, and, in court, he estimated Waterman to be somewhere between 110 and 120 pounds. He confirmed that he had not initially believed that Waterman and the suspect were the same person. However, he told the jury he was certain Waterman was that suspect.

2. Waterman's trial testimony regarding this fire.

At trial, Waterman denied starting this fire, and she denied that she had been the suspect LaClair had spotted. She did admit, however, that she had encountered LaClair and she had chased him away. She claimed that she (along with Jackson) had left their residence to investigate the sound of squealing tires. As she rounded a corner, she saw a man in camouflage. At that time, she did not know he was an undercover officer and she thought this person might have set the fires. They both cursed at each other and they had a heated exchange. She also saw another person in camouflage. She followed "them" because she thought these people might be arsonists. She agreed she might have weighed 140 pounds when this encounter occurred.

3. Waterman's statements to Cacho.

On June 14, 2013, Cacho spoke with Waterman about this fire. She explained that she had left her residence to investigate the sound of tires squealing.²⁴ She had heard crackling and walked up the driveway. She saw and approached a person in camouflage. She had asked that person if he was conducting surveillance in the area for the fires. At trial, Cacho agreed that Waterman's question was unusual.

24 Cacho received reports about multiple suspicious vehicles in the area of this fire. One was a white Jeep Liberty (which was ruled out because that was a CAL FIRE vehicle used in this surveillance). However, a pickup and a Volvo were also reported. Quinn told Cacho to ignore the reported vehicles. As such, Cacho did not follow up on any of the reported vehicles.

The jury found Jackson not guilty for the fire charged in count 20, but Waterman was found guilty in this count.

G. The final Cluster Fire on June 25, 2013 (Count 31).

1. Fire officials attempt to fool Waterman.

On June 16, 2013, Waterman was observed beating some grass with a stick in an area behind her residence. In the coming days, however, surveilling officers stopped seeing her in that area, which was unusual. Officials devised a ruse to put her at ease. On June 21, 2013, before the final Cluster Fire occurred, a marked CAL FIRE vehicle was driven near appellants' residence and two uniformed personnel removed yellow tape that had been put up. The personnel used loud radios and made a lot of noise in an effort to show that CAL FIRE was "clearing the area." Two days later, Waterman was again spotted outside her residence.

2. The final Cluster Fire.

The final Cluster Fire occurred on June 25, 2013, shortly before appellants were arrested that same day. This fire occurred behind appellants' residence, and it was spotted around 6:40 p.m. by undercover personnel. Gilbert investigated this fire. He eliminated all other causes and determined arson was the cause.

3. Surveilling officers spot Waterman in the area of this fire.

When this fire occurred, officers were in surveillance positions near appellants' residence. A smoke column was observed coming up from behind appellants' residence, and Waterman was seen walking slowly up the driveway away from her residence. She stopped and looked back towards her residence two or three times, but she kept walking. Rising smoke was visible in the direction she was looking. Around that time, surveilling officers could hear emergency vehicles approaching the area. When the sirens started to come up her street, Waterman stopped, turned, threw her hands into the air and began running towards her residence. She was heard yelling "fire" and expressing surprise that the arsonist started another one.

4. Responding firefighters see Jackson act strangely.

Two responding firefighters saw Jackson standing near his residence and near this fire. He was holding a hose that did not have water coming from it. Jackson made eye contact with one firefighter, who testified at trial that Jackson froze. Both firefighters believed that Jackson was faking efforts to suppress this fire.

Around that time, officers moved on Jackson to arrest him. A firefighter saw the officers tell Jackson to get down. At that point, Jackson was still holding his hose with no water coming out of it.²⁵

25 Testimony from one firefighter suggested that water may have been seen coming from this hose, but it is not clear when that happened.

At trial, Waterman claimed that she had given Jackson the hose after she had turned it on "full blast." She then claimed that she had "turned it down a little" to keep the water from spraying everywhere. She said the hose had a "fireman's nozzle" that allowed the operator to adjust the stream. She claimed that, after giving the hose to Jackson, she ran towards the fire and then ran to their driveway. She wanted to clear some tree limbs that a neighbor had placed so that fire personnel could reach the fire.

The jury convicted both Jackson and Waterman for the fire charged in count 31.

X. Waterman's Statements To LaClair.

On June 25, 2013, Waterman was detained and LaClair interviewed her. The interview was recorded, and it was played for the jury.

During the interview, Waterman repeatedly denied that she had started any fires. She denied being a smoker. She repeatedly denied being the person LaClair said he saw before the fire on June 14, 2013 (count 20). She claimed that she did not know how to start a fire except in the fireplace using paper, kindling, "and a clicky."

Despite these denials, however, Waterman made statements to LaClair which suggested she was the suspect he had spotted on June 14, 2013. She told LaClair that she has a distinctive walk and she is "bow-legged." She admitted following him. At one point, she said "[m]aybe" she was at the fire site on June 14, 2013, but said she could not keep track because she was "up here" many times. When pressed, she said she remembered this fire "[b]ecause I saw you and I chased your ass." LaClair noted that "you already admitted now that you were up there and you came back." She immediately answered, "Because I hear stuff." A short time later, she claimed she had gone to her house, heard squealing in the street, and she went to investigate. She told Jackson to call 911 because she heard "crackling." She then saw LaClair.

At trial, Waterman clarified that she had been patrolling the area in general before this fire, and she never intended to suggest she was at this fire right before it started. She admitted hitting the ground with a stick several days later. She admitted she had been looking for people while doing so but she denied trying to flush out people. She denied that people could have been hiding in the area where she was tapping.

XI. Gilbert's Testimony Regarding The Cluster Fires.

At trial, Gilbert opined that Waterman's statements and observations made by LaClair were significant in linking the Cluster Fires together. According to Gilbert, Waterman had observed "two or three" fires in their initial stages. Based on his analysis, Gilbert opined that two people were involved in setting the Roadside Fires and the Cluster Fires, which generally occurred during different times of the day. He believed that all 31 charged fires were caused by arson. He took into account all of the information from the investigation, including the GPS and camera data, information from the surveillance teams, and the overall pattern of the fires.

At trial, Gilbert admitted that he had opined at the preliminary hearing that Waterman had been solely responsible for the Cluster Fires. However, after learning that Waterman had presented a credible alibi defense for some of these charges, Gilbert now believed Jackson had started the other Cluster Fires.

XII. The Cluster Fires Stopped After Appellants Were Arrested.

Appellants were arrested on June 25, 2013. Three residents testified at trial regarding the occurrence of fires after appellants were arrested.²⁶ Patrick Denis, who lived in Coarsegold and was Jackson's part-time employer, gave his "general impression" that the fires stopped in the area around Yosemite Lakes. Before this testimony, however, Denis had testified, "I believe they stopped or, you know, I don't know exactly," and "I don't know for sure. I mean, you know, I mean, seemed like the fires stopped."

26 At trial, Quinn admitted that some arson and undetermined fires had occurred after appellants were arrested. These fires were not in the Yosemite Lakes area, but at least one arson fire had started in Coarsegold after appellants were arrested.

Jack Valpey, who had lived a quarter of a mile away from appellants, was asked what happened to the fires "in your area" after appellants were arrested. He said, "No more fires."

Gaile Jeanette Lippner, who lived across the street from appellants, recalled when appellants were arrested. When asked if anything significant happened regarding the fires after their arrest, she answered, "There weren't any."

XIII. Materials Located In Waterman's Toyota Sequoia.

After appellants were arrested, Waterman's Toyota Sequoia was searched. A new magnesium strike block was found in her SUV. In addition, a box of wooden matches was recovered, which was missing 65 of the 250 matches.

XIV. Evidence Of Resisting Arrest And Battery.

On June 25, 2013, multiple CAL FIRE officials approached Jackson outside appellants' residence while first responders were battling the last Cluster Fire. The men identified themselves as police officers. Jackson was ordered to drop his hose and lie on the ground. He turned to look at the officers, but failed to comply after the order was repeated. An officer drew his service weapon, and he tried to pull Jackson to the ground but Jackson pushed his hand away. Another officer tried to bear hug Jackson, and Jackson pushed back, which caused the officer to fall down. The officers tackled Jackson and cuffed him.

XV. Additional Defense Evidence.

Some of the defense evidence was included in the summary above. Below is a relevant summary of the additional defense evidence introduced at trial.

A. Expert testimony.

Steven Carman, an owner of a fire and explosion investigation company, testified as appellants' expert on fire investigation. In preparation for his trial testimony, Carman had read about 4,500 pages of CAL FIRE reports from this investigation, an analysis of vehicle traffic, and transcripts from the preliminary hearing. At trial, Carman raised concerns regarding an exclusion method to determine a fire's cause. He referenced the National Fire Protection Association (NFPA) and its publication, NFPA 921, as grounds that a lack of evidence could not be used to determine arson. He said fire investigators at one time erroneously believed this was a reliable indicator of arson. He asserted that a hypothesis cannot be tested adequately if no evidence supported it. Such a hypothesis, based on a negative corpus, was improper.

At trial, Carman believed that some of the CAL FIRE opinions of arson were cursory. In some of the CAL FIRE reports, Carman found references to leads which investigators had obtained but failed to pursue. In some instances, the reports showed that reporting parties and other witnesses were identified and their telephone numbers collected, or vehicles seen in the area were identified, but there was no indication of follow-up. Carman opined that if these reports reflected what actually happened, these investigations were inadequate; the dropped leads meant possible causation hypotheses were not developed and tested.

Carman believed the geospatial and temporal patterns were not sufficient grounds for CAL FIRE to determine arson. In his view, these patterns were investigative tools, showing the direction for further inquiry and justifying a focus on a particular suspect. However, arson required actual evidence that a fire was ignited in a particular way. Likewise, Carman believed that the surveillance camera data had value in pointing to Jackson as a suspect, but it did not establish arson.

Carman informed the jury that an investigator could infer arson from the existence of prior fires with undetermined causes or suspected as arson so long as these fires were very close together in space and time, i.e., a few feet and a few minutes. He did not think an inference of arson was supported from prior fires that were separated by days and miles. Instead, an investigator could suspect arson and investigate further.

Carman noted for the jury that the destruction of a fire's origin area by firefighting activities would weaken the investigators' ability to find or exclude possible causes. Small items of evidence, such as hot particles from vehicles' brake shoes, could be lost in this way. Related to this problem, Carman explained that, based on the documents he reviewed, CAL FIRE only sifted the soil for the last two fires in this series. This called into question the thoroughness of the other investigations at the specific origin areas. Sifting soil with a sifter is a technique used to find non-magnetic metal and non-metal particles that could be related to a fire, such as particles thrown from vehicles. According to Carman, this was particularly important with respect to the fires by roadsides as it affected the ability to exclude vehicles properly as a potential cause.

Carman believed some of the deficiencies he observed could be an indication that the CAL FIRE investigators were subject to confirmation bias or expectation bias. This involves cognitive tendencies to look for explanations that one expects, and to focus inquiries on confirming an expected explanation. The belief that prior arsons supported the conclusion that the next fire was arson was a possible example of these biases in action.

Giving his overall assessment of the CAL FIRE investigation as reflected in the reports he had reviewed, Carman testified that some of the investigations were inadequate. He could not say "the entire population" of investigations met the standard of care. He told the jury that, at one point, he had informed the defense that he believed the investigations as reflected in the reports were inadequate in 12 of the 31 cases. For those 12 cases, Carman had "real questions" regarding whether CAL FIRE could claim arson. He did not, however, specifically identify those 12 investigations to the jury.

B. DNA evidence.

The defense introduced DNA evidence. Five cigarette butts were tested for DNA. A full DNA profile was obtained from one butt, and a partial profile was obtained from another butt. Neither results matched Jackson's or Waterman's DNA. The other three butts had no DNA on them. CAL FIRE officials apparently failed to maintain a record of where exactly these butts were found. An expert testified that temperatures above 95 degrees Celsius (above about 200 degrees Fahrenheit) destroy DNA.

C. Carl Jackson's testimony.

Jackson's father, Carl Jackson, testified that he often spent time at appellants' residence and people were "always" walking through that area. On numerous occasions over the years, he had seen as many as four or five kids at a time in the nature area behind appellants' residence.

D. Waterman's trial testimony.

At trial, Waterman denied having an agreement with Jackson to start fires. She had no idea why fires occurred so close to their residence and she denied starting them. She denied ever seeing Jackson build incendiary devices and she had no indication he was possibly starting fires.

DISCUSSION

We address Jackson's and Waterman's respective appellate issues in the order that they generally arose during the trial proceedings.

I. The Trial Court Did Not Abuse Its Discretion Regarding The Admission Of Expert Testimony And Any Presumed Error Was Harmless.

Waterman argues that, based on CAL FIRE's cause exclusion analysis (or negative corpus), the trial court should have precluded the CAL FIRE experts from testifying about causation.²⁷ Prior to trial, the defense had moved to exclude any opinion testimony regarding fire causation that was based on a lack of evidence. The trial court reserved ruling on this issue and asked for more briefing, but noted that these concerns could be raised during trial cross-examination.²⁸ At trial, the defense expert, Carman, criticized the CAL FIRE investigation and cited NFPA 921 as grounds that a lack of evidence (negative corpus) could not be used to determine arson.²⁹

27 We note that the term "negative corpus" generally refers to a fire investigator's determination that a fire was incendiary based on the lack of evidence showing an accidental cause. (Schlesinger v. United States (E.D.N.Y. 2012) 898 F.Supp.2d 489, 491-492.)

28 After the prosecution rested at trial, the defense made a motion for judgment of acquittal pursuant to section 1118.1, asserting that negative corpus was insufficient to establish arson. After hearing argument from both sides, the trial court denied that motion. In raising the present claim on appeal, Waterman does not contend that the trial court erred in denying the motion for judgment of acquittal. As such, we do not address that posttrial motion.

29 On our own motion, we take judicial notice of NFPA 921 (Evid. Code, § 452, subd. (h)), but we do not take judicial notice of the truth of the matters stated therein.

A. Standard of review.

An expert's opinion may not be based on assumptions of fact without evidentiary support. (Sargon Enterprises, Inc. v. University of Southern California (2012) 55 Cal.4th 747, 770 (Sargon).) Likewise, such an opinion may not be based on speculative or conjectural factors. (Ibid.) "Thus, under Evidence Code sections 801, subdivision (b), and 802, the trial court acts as a gatekeeper to exclude expert opinion testimony that is (1) based on matter of a type on which an expert may not reasonably rely, (2) based on reasons unsupported by the material on which the expert relies, or (3) speculative. Other provisions of law, including decisional law, may also provide reasons for excluding expert opinion testimony." (Id. at pp. 771-772.)

An appellate court reviews for abuse of discretion a trial court's ruling excluding or admitting expert testimony. (Sargon, supra, 55 Cal.4th at p. 773.) "A ruling that constitutes an abuse of discretion has been described as one that is 'so irrational or arbitrary that no reasonable person could agree with it.' [Citation.]" (Ibid.)

B. Analysis.

Waterman contends that the prosecution's experts relied on faulty logic when opining that certain fires were caused by arson. She impliedly argues that the trial court abused its discretion in permitting this testimony. We find these arguments unpersuasive. As we explain below, nothing from the preliminary hearing, the text of NFPA 921, or relevant case law establishes an abuse of discretion. Moreover, even if we presume error occurred, we find no prejudice.

1. The trial court did not abuse its discretion.

Before ruling on the defense's motion to exclude the expert testimony, the trial court heard the prosecution's experts testify during the preliminary hearing. Nothing from the preliminary hearing suggested that the investigators made assumptions of fact without evidentiary support. To the contrary, the CAL FIRE investigators described their training and experience. The investigators generally explained how they performed a cause and origin investigation at a fire site. They discussed how they determined that some of the charged fires were caused by arson. They also explained how they determined that arson could not be excluded as a cause (sometimes with other potential causes) for the other charged fires.

During the preliminary hearing, some of the CAL FIRE investigators were asked about NFPA 921 and negative corpus. McCully said he was familiar with NFPA 921 and negative corpus. He agreed that a fire should not be deemed intentionally started just from ruling out accidental causes. Cacho agreed that he had heard of negative corpus, which he defined as "what's left behind" after all other possible causes are ruled out. Cacho was uncertain, however, whether or not NFPA 921 criticized negative corpus.

Gilbert was aware of NFPA 921 and he agreed that, in 2011, NFPA criticized negative corpus. He did not agree that he had relied on NFPA 921 in formulating his opinions, but it "assisted" during the process. Gilbert defined "negative corpus" as "a process of eliminating all potential causes of the fire with the exception of one cause for the fire and determining that to be the cause of the fire." He denied that the inability to exclude arson as a cause of fire is evidence that an arson was committed.

Nothing from the preliminary hearing reasonably suggested that the CAL FIRE experts based their opinions on speculative or improper matter. To the contrary, these witnesses explained how they analyzed the suspicious fires and how they reached their conclusions. They generally agreed that arson could not be determined merely from ruling out all other potential causes.

Further, Waterman does not cite, and we have not found, any authority holding that a reliance on negative corpus automatically requires the exclusion of expert testimony regarding the cause and origin of a fire. We have also not found, and Waterman does not cite, any authority holding that a failure to follow the methodology outlined in NFPA 921 automatically requires the exclusion of expert testimony at trial.

NFPA is an international nonprofit organization that is devoted to minimizing fire damage and risks. (Schlesinger v. United States, supra, 898 F.Supp.2d at p. 492.) NFPA publishes a Guide for Fire and Explosion Investigations, NFPA 921, which is widely regarded as the standard in the field of fire investigation. (See United Fire & Cas. Co. v. Whirlpool Corp. (11th Cir. 2013) 704 F.3d 1338, 1341; United States v. Hebshie (D.Mass. 2010) 754 F.Supp.2d 89, 109, fn. 39; Schlesinger v. United States, supra, 898 F.Supp.2d at p. 504.) In 2011, NFPA revised its publication and rejected negative corpus as a valid methodology to determine fire causation. (Somnis v. Country Mut. Ins. Co. (D.Minn. 2012) 840 F.Supp.2d 1166, 1172, fn. 2.)

NFPA 921 states that its purpose "is to establish guidelines and recommendations for the safe and systematic investigation or analysis of fire and explosion incidents." (NFPA 921 (2011) § 1.2.1.) It says that deviations from its procedures "are not necessarily wrong or inferior but need to be justified." (Id. at § 1.3.) It notes that every fire "is in some way unique and different from any other," and this document is not designed to encompass all necessary components of a complete investigation but the scientific method should be applied in every instance. (Id. at § 1.3.2.) It acknowledges that a fire investigator may be unable to apply its recommendations or techniques based on limitations due to time, resources, or existing policies. (Id. at § 1.3.4.)

A fair reading of NFPA 921 indicates that, while its recommendations carry great weight, any deviations from those recommendations do not, without more, render an expert's opinions invalid. To the contrary, NFPA 921 acknowledges that deviations can and do occur, but the investigator should be prepared to justify his or her conclusions.

Finally, although California case law is silent on this issue, some federal courts have examined NFPA 921 and its impact on expert testimony regarding fire causation.³⁰ These federal opinions generally support our conclusion that any deviation from NFPA 921 was not enough to require exclusion of the expert opinions at the trial in this matter.

30 "Although we are not bound by the decisions of the lower federal courts, 'they are persuasive and entitled to great weight.' [Citation.]" (People v. Camacho (2000) 23 Cal.4th 824, 843.)

Experts in arson cases have been allowed to testify even when their opinions were based on methodology other than that prescribed in NFPA 921. (See Manuel v. MDOW Ins. Co. (8th Cir. 2015) 791 F.3d 838, 845-846 [finding no clear error where investigator opined about cause of fire using negative corpus but based on his personal observations and 40 years of experience]; Russell v. Whirlpool Corp. (8th Cir. 2012) 702 F.3d 450, 455 [NFPA 921 is not the only reliable way to investigate a fire]; Schlesinger v. United States, supra, 898 F.Supp.2d at p. 504; Pekarek v. Sunbeam Products, Inc. (D.Kan. 2008) 672 F.Supp.2d 1161, 1175 [" . . . a failure to strictly adhere to NFPA 921 does not render an investigation per se unreliable."]; Thompson v. State Farm Fire & Cas. Co. (W.D.Tenn. 2008) 548 F.Supp.2d 588, 592 [following NFPA 921 indicates reliability but is not grounds for automatic disqualification].)³¹

31 For completeness, we note that some federal courts have excluded an expert from rendering opinions when the expert expressly relied on NFPA 921 but then deviated from its methodology. (See generally Presley v. Lakewood Eng'g & Mfg. Co. (8th Cir. 2009) 553 F.3d 638, 645; Fireman's Fund Insurance Company v. Canon U.S.A., Inc. (8th Cir. 2005) 394 F.3d 1054, 1057-1058.)

Moreover, some federal courts have permitted an expert to opine on causation even in the absence of an accidental explanation for a fire. (See Schlesinger v. United States, supra, 898 F.Supp.2d at pp. 504-505; Somnis v. Country Mut. Ins. Co., supra, 840 F.Supp.2d at p. 1171; Hickerson v. Dist. Pride Mobility Prods. Corp. (8th Cir. 2006) 470 F.3d 1252, 1257-1258 [holding, without any reference to NFPA 921, that expert could opine on cause of fire based on elimination of other possible causes and inferences therefrom].)

Based on this record, the trial court's ruling was neither irrational nor arbitrary. Nothing from the preliminary hearing reasonably suggested that the investigators based their opinions on improper or speculative matter. To the contrary, the investigators relied on their own observations and experience. Any concerns regarding CAL FIRE's investigation methodology were proper material for vigorous cross-examination. Accordingly, an abuse of discretion is not present and we reject this claim. In any event, any presumed error was also harmless.

2. Any presumed error was harmless.

Due process is not offended by the admission of relevant evidence unless it is so prejudicial as to render the criminal trial fundamentally unfair. (People v. Jablonski (2006) 37 Cal.4th 774, 805.) In the absence of fundamental unfairness, the harmless error test of People v. Watson (1956) 46 Cal.2d 818, 836 (Watson) is used to analyze an evidentiary error that involves state law. (People v. Partida (2005) 37 Cal.4th 428, 439.) The question is "whether it is reasonably probable the verdict would have been more favorable to the defendant absent the error." (Ibid.)

Here, within 46 days (May 11 through June 25, 2013), there were 31 suspicious fires reported in and around Yosemite Lakes. The number of suspicious fires was unusual, even during the fire season and in a drought. These suspicious fires occurred in either a pattern or in specific areas. The circumstantial evidence placed Jackson at a vast majority of the Roadside Fires within assumed ignition windows. All of the Cluster Fires occurred right next to appellants' residence and these fires stopped after appellants were arrested. Both Jackson and Waterman were at the scene of some of the Cluster Fires under circumstances that further suggested their guilt. Both Jackson and Waterman readily possessed the necessary tools to start these fires.³² All of this evidence strongly suggests the charged fires were started intentionally and appellants were responsible.

32 Although only Waterman raises this appellate issue, we also address the lack of prejudice against Jackson to render a more complete analysis.

The jury heard extensive testimony regarding how each investigator analyzed the charged fires and reached their respective conclusions. For a minority of fires, an incendiary device was recovered. For other fires, arson (sometimes with other potential causes) could not be excluded. For a majority of fires, the investigator determined arson in the absence of an incendiary device. After all investigators testified, Gilbert opined that all 31 charged fires were caused by arson. He took into account all of the information from the investigation, including the GPS and camera data, information from the surveillance teams, and the overall pattern of the fires.

We cannot say that admitting the investigators' opinions resulted in fundamental unfairness. The defense expert, Carman, raised substantial questions about the reliability of CAL FIRE's investigation, including its methodology, its failure to follow up on certain leads, and the strength of its conclusions. The defense was able to cross-examine the CAL FIRE witnesses regarding their opinions and how they investigated these fires. The defense was able to raise its concerns with the jury during closing arguments. The jury returned numerous not guilty verdicts, which suggests it scrutinized the evidence.

At no point did the fire investigators rule out arson for any of the 31 suspicious fires. Moreover, the jury saw all of the information from the investigation, including the GPS and camera data, information from the surveillance teams, and the overall pattern of the fires, which reasonably linked the charged fires together. The sheer number of suspicious fires, the short time period in which they occurred, their locations, and appellants' respective and repeated proximity to a vast majority of them overwhelmingly pointed to arson.

A reasonable probability does not exist that appellants would have received a better result had the trial court granted the defense motion and precluded the experts from opining that some of the fires were caused by arson without evidence of an incendiary device. The totality of the evidence was sufficient to establish arson as the cause of these 31 suspicious fires. Based on this record, any presumed error was harmless. Accordingly, prejudice is not present and this claim fails.

II. The Trial Court Did Not Abuse Its Discretion In Denying A Motion For Mistrial Following The Prosecution's Discovery Violation.

Jackson contends he did not receive a fair trial based on a discovery violation. He argues that the trial court abused its discretion in denying a motion for mistrial. Waterman joins this argument.

A. Background.

On March 19, 2014, the 46th day of trial, Gilbert was called back to the stand after the fire investigators had testified regarding all 31 charged fires. Gilbert testified that he believed all 21 Roadside Fires were started with cigarette incendiary devices, and the 10 Cluster Fires were started hot. His opinion was bolstered by his belated discovery of a purported match head fragment found in the cigarette butt recovered from the site of the fire charged in count 12.³³ The defense made an immediate motion for a mistrial based on the prosecution's failure to disclose this evidence. After the jury was dismissed, Gilbert explained that he had discovered this match head on the evening of February 13, 2014, after the 28th day of trial. He had advised the prosecutor that night.

33 During that 28th day of trial, Pimentel had testified about this same cigarette butt. Although Pimentel had previously opined that this butt was an incendiary device, his opinion had been based on its burn markings.

At the subsequent hearing, the prosecutor conceded that she had failed to comply with her discovery obligations. She explained that she had learned of this evidence on a Friday afternoon, and a holiday had occurred the following Monday. She got very busy and forgot about this evidence, which she admitted was not a legitimate excuse. She argued, however, that her omission was not willful.

After hearing from both sides, the trial court accepted the prosecutor's representations and found no willful discovery violation. The court denied the motion for a mistrial, noting there was time to cure prejudice. The court ruled that the defense could further examine any necessary witnesses regarding Gilbert's findings, and it could retain an expert to inspect the purported match head. The court indicated it would instruct the jury on the discovery violation, and it asked the defense if any additional remedies should be ordered. No other remedies were suggested.

The court instructed the jury on the discovery violation. The jury was told it could consider the effect, if any, of the late disclosure in evaluating the weight and significance of this evidence.

B. Standard of review.

A trial court should grant a mistrial only when the chances of either party receiving a fair trial have been irreparably damaged. (People v. Bolden (2002) 29 Cal.4th 515, 555.) A deferential abuse of discretion standard is used to review a trial court's ruling denying a mistrial. (Ibid.) "A ruling that constitutes an abuse of discretion has been described as one that is 'so irrational or arbitrary that no reasonable person could agree with it.' [Citation.]" (Sargon, supra, 55 Cal.4th at p. 773.)

C. Analysis.

Trial courts have broad discretion in determining the appropriate sanction for a discovery violation, and sanctions ranging from dismissal to the giving of special jury instructions may be required in order to ensure that the defendant receives a fair trial. (People v. Jenkins (2000) 22 Cal.4th 900, 951; see also People v. Verdugo (2010) 50 Cal.4th 263, 280 [citing section 1054.5, subdivision (b), and explaining options to remedy a violation of reciprocal discovery statute].) If a party willfully fails to provide discovery to obtain a tactical advantage, the court may exclude the relevant witness's testimony. (People v. Jackson (1993) 15 Cal.App.4th 1197, 1203.) However, absent a showing of significant prejudice and willful conduct, exclusion of testimony is not an appropriate sanction. (People v. Gonzales (1994) 22 Cal.App.4th 1744, 1758.)

Here, the trial court did not believe that prejudice from this discovery violation was incurable. We agree. The defense was permitted to recall witnesses and it was allowed to obtain an expert to view the purported match head. We do not find it persuasive that recalling witnesses was potentially so burdensome on the jury that appellants lost a chance for a fair trial.³⁴ The defense had ample grounds to attack the credibility of Gilbert's opinion, including the limited physical evidence supporting it, the timing of its disclosure during trial, and the fact Pimentel never testified about the purported match head. In addition, the court provided an appropriate jury instruction for this violation. We note that the jury found Jackson not guilty of arson in count 12 despite this fire having strong evidence of a time-delay incendiary device.

34 The trial continued another 28 court days after this violation came to light on March 19, 2014.

Further, prior to this discovery violation, the jury had heard about incendiary devices. Based on its burn markings, Pimentel had opined that this same cigarette butt was an incendiary device, and the jury had learned that incendiary devices had been recovered for the fires charged in counts 1, 8 and 9. These included cigarette butts that were the same color (orangish brown) as Jackson's brand.³⁵ Smith had also opined that an incendiary device had been used in counts 10 and 11.

35 Cigarette butts were also either observed or recovered from the fires charged in counts 13, 17 and 25, but they were not deemed incendiary devices. The butt involved in count 17 was white in color.

Regarding count 1, Quinn had opined that a match head had caused the distinctive burn markings on the recovered cigarette butt. On cross-examination, Quinn was asked about the butts recovered in counts 1 and 8. He was asked if someone could stuff match heads into a cigarette and how a match might ignite.

Regarding count 8, McCann confirmed that he had taken pictures of the burnt cigarette that he had collected. On cross-examination, he was asked if he thought "that someone stuffed something in the cigarette or anything like that?" McCann answered no.

Based on these examples, we disagree with Jackson's claim that the defense must have necessarily assumed all charged fires were started hot. We also reject Jackson's argument that the defense must have labored under a false theory of prosecution. This discovery violation did not hamper the defense to the extent Jackson now claims.

Finally, Jackson argues it is "highly unlikely" that the prosecutor "simply forgot" about this evidence. He contends the prosecutor's actions were willful. The trial court, however, was acquainted with the attorneys, understood the complexity of this trial, and was familiar with the evidence. The judge considered both sides' arguments and heard firsthand the prosecutor's explanations. We will not disturb the judge's exercise of discretion under these circumstances.

Based on this record, this discovery violation did not irreparably damage the defense's chance for a fair trial. Further, the trial court's ruling on the motion for mistrial was not so irrational or arbitrary that no reasonable person could agree with it. Accordingly, an abuse of discretion is not present and we reject this claim.

III. The Trial Court Did Not Abuse Its Discretion In Excluding Certain Third-Party Culpability Evidence And Any Presumed Error Is Harmless.

Both Jackson and Waterman contend that the trial court abused its discretion in excluding evidence and closing arguments that juveniles may have been responsible for some of the charged fires.

A. Background.

At trial, the jury learned that certain juveniles lived near appellants' residence, including L.H. and M.S. The jury learned that L.H. had lit a fire in his own backyard on May 22, 2013, about a half mile from appellants' residence. L.H.'s fire had started about 90 minutes before the Cluster Fires which were charged in counts 5 and 6. At least two other juveniles, including M.S., were present when L.H. started his uncharged fire. At trial, Quinn explained that he had investigated the area around L.H.'s home and he saw three separate burnt areas that were within 1,000 feet of each other. Quinn told the jury that he did not believe a juvenile had started the fires charged in counts 4, 5 or 6. He believed it would have been difficult for a stranger to access that remote area, particularly without alerting nearby dogs.

McCann investigated the fire charged in count 29. As part of his investigation, he interviewed some juveniles, including M.S. and L.H. M.S. lived less than one block from this fire. McCann did not find any evidence suggesting that juveniles were responsible.

At trial, the defense wanted to introduce additional evidence alleging L.H.'s involvement in other fires. Via an offer of proof, the defense alleged that L.H. had started three small brush fires in Oakhurst in June and August 2013. The defense alleged that L.H. had resided in Oakhurst when these three small fires had occurred, and they started near his residence. The prosecution opposed this evidence, in part, by arguing that L.H. had moved from Oakhurst when these fires started. The trial court determined that L.H.'s alleged involvement in the three uncharged Oakhurst fires was not sufficiently linked to the charged fires in this matter. The trial court prevented the admission of any evidence or argument regarding these three uncharged fires.

In addition, the defense wanted to argue to the jury that M.S. may have been responsible for some of the charged fires or he should have been investigated. The defense asserted that M.S. was friends with L.H. and M.S. lived only about a half mile from appellants' residence. Defense counsel noted that M.S. resembled the suspect that LaClair had observed. The defense claimed that M.S.'s resemblance showed flaws in the investigation. The trial court found an insufficient link between M.S. and the charged fires, and it precluded any argument involving M.S.

During closing arguments, Waterman's counsel emphasized that L.H. had used a lighter to start the uncharged fire near appellants' residence on May 22, 2013. It was noted that a lighter had been recovered near the fires charged in counts 5 and 6. No one mentioned M.S. during closing arguments.

B. Standard of review.

We review relevancy and Evidence Code section 352 rulings for abuse of discretion. (People v. Weaver (2001) 26 Cal.4th 876, 933.) A trial court abuses its discretion when a reasonable person could not agree with its ruling because it is so irrational or arbitrary. (Sargon, supra, 55 Cal.4th at p. 773.)

C. Analysis.

Appellants generally claim that, because one or more juveniles set fires in the same area and time frame as the fires charged in this matter, any inference of appellants' guilt was undermined. They argue that the third-party evidence was relevant because it impacted the prosecution's pattern theory, and it suggested that arson fires in the area did not stop after appellants were arrested. They assert that the circumstantial evidence linking the third parties to the charged fires was just as strong as the evidence linking appellants. As such, they contend that the trial court erred in precluding much of this evidence and limiting closing arguments, which they assert denied them their constitutional rights to counsel and to present a defense.

We find these claims unpersuasive. As we explain below, the trial court did not abuse its discretion and any presumed error was harmless.

1. The trial court did not abuse its discretion.

For an uncharged offense to be admitted to establish a third party's identity as the perpetrator of the charged crimes, the pattern and characteristics of the crimes must be so unusual and distinctive that it represents a signature. (People v. Elliott (2012) 53 Cal.4th 535, 581.) Evidence of a third party's culpability is admissible if it is capable of raising a reasonable doubt of the defendant's guilt. (People v. Prince (2007) 40 Cal.4th 1179, 1242.) However, a third party's mere motive or opportunity to commit the crime, without more, is not sufficient to raise reasonable doubt about a defendant's guilt. (Ibid.) Instead, direct or circumstantial evidence must link the third person to the actual perpetration of the crime. (Ibid.)

Here, other than L.H.'s arson fire on May 22, 2013, about a half mile from appellants' residence, no evidence linked any juvenile to the actual perpetration of the fires charged in this matter. Although juveniles may have had the motive and opportunity to start the charged fires, that was not a sufficient bases to admit additional third-party culpability evidence.

Based on this record, the third party evidence which the defense proffered, and which the trial court excluded, did not tend to raise a reasonable doubt about appellants' guilt. The trial court's rulings were not so irrational or arbitrary that no reasonable person could agree with it. As such, the trial court did not abuse its discretion and we reject this claim. In any event, we also find any presumed error harmless.

2. Any presumed error was harmless.

Appellants contend that the trial court's rulings were prejudicial, and denied them certain constitutional guarantees, such as the right to present a defense. They claim a review for prejudice should occur pursuant to the federal standard under Chapman v. California (1967) 386 U.S. 18. We disagree.

A trial court's discretionary ruling involving the ordinary rules of evidence does not normally implicate the federal Constitution. (People v. Cudjo (1993) 6 Cal.4th 585, 611; accord People v. Bradford (1997) 15 Cal.4th 1229, 1325 [disallowing defense evidence is not also the denial of the right to present a defense].) Under the Watson standard, we ask whether it is reasonably probable the verdict would have been more favorable to the defense had the trial court admitted the evidence regarding the culpability of the third parties. (People v. Bradford, supra, 15 Cal.4th at p. 1325.)

Here, this record does not establish prejudice. As Jackson concedes in his opening brief, the jury heard some third-party culpability evidence. The jury learned about L.H.'s arson fire near appellants' residence. Gilbert acknowledged that L.H.'s arson fire, along with about two other burned areas in that same area or backyard, could have been included in the map for these charged fires.

The jurors heard about two other uncharged fires, Harsh and Hensley. The jury learned that an unnamed arson fire had occurred in Coarsegold after appellants were arrested. Based on all of this evidence, the jury was made aware that another arsonist, and possibly more, had been active in the general time period and area as the charged fires.

Further, the jury heard details about M.S., including his proximity to some of these fires. Quinn acknowledged that M.S. had black curly hair, and he was tall and slender. The jury heard that, in June 2013, CAL FIRE investigators took a picture of L.H. and M.S. together (along with a third juvenile). That photograph, exhibit 222, was shown to the jury. The jury heard that juveniles, including these suspects, had been seen at various locations in Yosemite Lakes.

During trial, the defense generally cross-examined the witnesses about other cars, bicycles, and any juveniles that were spotted at or near the start of the charged fires. The jury heard cross-examination and closing argument attacking LaClair's identification of Waterman as the suspect with black hair who had started the fire charged in count 20.

In light of this record, a manifest miscarriage of justice did not occur. There is no reasonable probability the jury would have reached more favorable verdicts had it heard of L.H.'s alleged link to three uncharged fires in Oakhurst, and/or if the defense had been permitted to argue more about juveniles' potential liability for the charged fires. Any presumed error in limiting this evidence and argument was harmless. Accordingly, prejudice is not present and this claim fails. IV. Although The Evidence Is Sufficient To Affirm A Majority Of The Guilty Verdicts, We Reverse For Insufficient Evidence Eight Of Jackson's And Two Of Waterman's Arson Convictions.

The jury convicted Jackson of arson in 14 of the Roadside Fires (counts 9-11, 13, and 21-30). For the Cluster Fires, Jackson was found guilty in seven of those counts (5, 6, 14, 15, 16, 19 and 31) and Waterman in six (14, 15, 16, 18, 20 and 31). They were both found guilty of conspiracy (count 32). Jackson maintains there was insufficient evidence to prove any of the arson or conspiracy counts beyond a reasonable doubt. Waterman joins in this argument.

A. Standard of review.

For an appeal challenging the sufficiency of evidence, we review the entire record in the light most favorable to the judgment to determine whether a reasonable jury could have found the defendant guilty beyond a reasonable doubt based on "'"evidence that is reasonable, credible, and of solid value ...."'" (People v. Jones (2013) 57 Cal.4th 899, 960.) In doing this review, we are not required to ask whether we believe the trial evidence established guilt beyond a reasonable doubt. (People v. Johnson (1980) 26 Cal.3d 557, 576.) Rather, the issue is whether any rational jury could have found the essential elements of the crime beyond a reasonable doubt after viewing the evidence favorably for the prosecution. (Id. at p. 576.) We are to presume the existence of any fact the jury could have reasonably deduced from the evidence in support of the judgment. (Ibid.) It is not our role on appellate review to reweigh evidence or judge witness credibility. (People v. Lindberg (2008) 45 Cal.4th 1, 27.)

B. Analysis.

"A person is guilty of arson when he or she willfully and maliciously sets fire to or burns or causes to be burned ... any structure, forest land, or property." (§ 451.) Arson is a general intent crime. (People v. Atkins (2001) 25 Cal.4th 76, 84.) Arson's malice requirement ensures that the act is done with a design to do an intentional wrongful act. (Id. at p. 88.) This eliminates accidental or unintentional ignitions. Instead, an arson fire must have an incendiary origin. (Ibid.)

An arson case is often based on circumstantial evidence. (People v. Williams (1988) 197 Cal.App.3d 1320, 1326; People v. Green (1983) 146 Cal.App.3d 369, 377.) If circumstantial evidence could not be used in arson cases, then, in the absence of a guilty plea, arson could not be prosecuted because incendiary fires are rarely started in the presence of eyewitnesses. (People v. Andrews (1963) 222 Cal.App.2d 242, 244.) Although the prosecution has the burden to show that a fire was of criminal origin, the coincidence of a series of fires tends to rebut the possibility that a particular fire was caused by an accident. (Id. at p. 246.)

The standard of review for sufficiency of the evidence is the same in which a conviction is based primarily on circumstantial evidence. (People v. Clark (2016) 63 Cal.4th 522, 625.) "In a case built solely on circumstantial evidence, none of the individual pieces of evidence 'alone' is sufficient to convict. The sufficiency of the individual components, however, is not the test on appeal." (People v. Daya (1994) 29 Cal.App.4th 697, 708.) Rather, when reviewing the sufficiency of circumstantial evidence, we must consider such evidence cumulatively and determine whether a reasonable jury could have found the defendant guilty beyond a reasonable doubt. (Id. at p. 709.)

A jury has a duty to acquit a defendant if it finds two interpretations from the circumstantial evidence, one which suggests guilt and the other innocence. (People v. Abilez (2007) 41 Cal.4th 472, 504.) However, it is the jury and not the appellate court that must be convinced of the defendant's guilt beyond a reasonable doubt. If the circumstances reasonably justify the jury's findings, a reversal of the judgment is not warranted if the reviewing court determines that the circumstances might also reasonably be reconciled with a contrary finding. (Ibid.) Guilt beyond a reasonable doubt may be based on circumstantial evidence, which may be sufficient to connect a defendant with the crime. (Ibid.)

With these principles in mind, we analyze whether substantial evidence supports the guilty verdicts for the Roadside Fires, the Cluster Fires, and then conspiracy.

1. Substantial evidence regarding the Roadside Fires.

The sheer number of suspicious fires that occurred in a short time period, their locations, and Jackson's repeated proximity to a vast majority of them rebutted the possibility that a particular fire was started accidentally. (People v. Andrews, supra, 222 Cal.App.2d at p. 246 [although the prosecution has the burden to show that a fire was of criminal origin, the coincidence of a series of fires tends to rebut the possibility that a particular fire was caused by an accident].) Cigarette incendiary devices were located at four of the Roadside Fires (counts 1, 8, 9 and 12). Some of those recovered cigarette butts were similar in color to Jackson's brand.

When searched, Jackson's truck had two lighters and numerous packages of cigarettes and paper matches. Numerous cigarette butts were in the truck's bed. More paper matches and lighters were located in appellants' residence. A CAL FIRE incident action plan for one of the charged fires was found in Jackson's truck. This is a CAL FIRE document given to the people assigned to a fire. It gave the radio channels, the assignments for each day, and the plan to complete each assignment. In addition, a notebook listing radio frequencies, some of which firefighters used, and a radio scanner were found in appellants' living room. There was also a reference book used by firefighters and others working at fire scenes. A computer in appellants' residence displayed a picture of a wildfire.

The jury was informed where each of the Roadside Fires occurred, the nature and circumstances of each fire, how the responder and/or investigator analyzed each fire, and their respective conclusions. The jury was informed where and how Jackson's vehicles were spotted, the amount of driving time from the respective observation to each fire site, and the believed ignition windows. After the cameras were installed, Jackson's vehicles were observed in the proximity of all 16 of the final Roadside Fires. The unrefuted evidence overwhelmingly established that Jackson was driving his vehicles, he was in a position to start a vast majority of these fires, and he readily possessed the tools necessary to start these fires. The trial evidence suggested his motive to commit arson based on his apparent fascination with fire and/or fire suppression.

As the prosecutor noted during closing arguments, Jackson was the only person repeatedly observed in sufficient proximity to all of the Roadside Fires after camera footage was available. Jackson's proximity to each of these fires strongly suggests something more than randomness, bad luck or an innocent explanation. His repeated proximity to all of these Roadside Fires, many of which occurred on the same day but separated by many miles, overwhelmingly pointed to his guilt. Further, the jury was aware that 10 additional suspicious fires started near Jackson's residence. The jury was entitled to make reasonable inferences based on the circumstantial evidence. (People v. Livingston (2012) 53 Cal.4th 1145, 1166.) The reasonable inferences from this record do not point to Jackson's innocence. Instead, the only reasonable conclusion to be drawn from the circumstantial evidence is that Jackson intentionally started the Roadside Fires.

We disagree with Jackson's contention that the expert opinions against him were speculative. To the contrary, the CAL FIRE experts based their opinions on the facts of the case. Cigarette butts, which were described as incendiary devices, were recovered at several fire sites. A purported match head was observed in one cigarette butt, and a marking on another butt could have been caused from a match head. The camera and surveillance evidence, along with GPS data, established Jackson's movements and whereabouts relative to each fire site. Jackson admitted during his interrogation that he had carelessly thrown a cigarette from his truck. The CAL FIRE theories about the ignition windows and time delayed incendiary devices were based on assumptions derived from the evidence. (See Sargon, supra, 55 Cal.4th at p. 770 [an expert's opinion may not be based on assumptions of fact without evidentiary support].) The credibility of those opinions was an issue for the jury to resolve. (See People v. Lindberg, supra, 45 Cal.4th at p. 27 [it is not our role on appellate review to reweigh evidence or judge witness credibility].)

The jury heard Carman, the defense expert, attack the inadequacies of CAL FIRE's investigation, including its methodology. For instance, according to Carman, a vehicle repeatedly seen in the vicinity of suspicious fires might be a lead, but that does not determine ignition. It was the jury's role, however, to resolve the conflicting expert opinions and make credibility determinations. (§ 1127b.) The jury was properly instructed to weigh any conflicting expert opinions and examine the reasons given for each opinion, along with the facts or other matters on which the witness relied. The jury was invited to compare the experts' qualifications. Based on the verdicts rendered, it is clear that the jury generally credited the prosecution's experts over Carman's conflicting testimony, and we will not disturb those credibility findings on appeal. (See People v. Lindberg, supra, 45 Cal.4th at p. 27.)

We review the evidence for each of the 14 Roadside Fires in which the jury found Jackson guilty. When reviewing the sufficiency of circumstantial evidence, we must consider the evidence cumulatively. (People v. Daya, supra, 29 Cal.App.4th at pp. 708-709.) We must determine whether a reasonable jury could have found the defendant guilty beyond a reasonable doubt based on the entire record. (Id. at pp. 708-709.)

In addition to the evidence for each particular Roadside Fire, the jury was aware of Jackson's proximity to each of the final 16 Roadside Fires after cameras were installed. The jury heard about all of the camera and GPS data, the observations from the surveillance, and the overall pattern of the fires. The jury was aware of the suspicious nature of all 31 charged fires. In addition to Jackson's proximity to a vast majority of the Roadside Fires, the jury was aware of the Cluster Fires that occurred near his residence. All of this circumstantial evidence pointed to his guilt and did not reasonably indicate his innocence.

When reviewing the evidence for each particular Roadside Fire, along with the totality of the circumstantial evidence from this record, we find substantial evidence to affirm the jury's convictions for 12 of these fires (counts 9, 10, 21, 22, 23, 24, 25, 26, 27, 28, 29, and 30). A rational jury could have determined that these fires were caused by arson and Jackson was the perpetrator. However, we find insufficient evidence supporting the convictions for counts 11 and 13. The prosecution failed to establish that Jackson was at the site of these fires with sufficient time to have started them. Therefore, we reverse counts 11 and 13 for insufficient evidence.

We summarize the evidence for each of the Roadside Fires in which the jury rendered guilty verdicts.

a. Count 9.

The Roadside Fire charged in count 9 occurred on June 8, 2013. This fire was reported at about 12:25 p.m. and it occurred by Road 415. Jackson's truck was seen driving away from this fire about 10 minutes after it was reported. When spotted, his truck was a little over three minutes away from the fire site.³⁶ Smith calculated a 13-minute ignition window for this fire. McCann investigated this fire and performed a cause exclusion analysis. He determined arson caused this fire.

36 After this fire was reported, Jackson's truck was also spotted twice more on camera moving towards this fire and then away from it.

Based on this record, and the totality of the circumstantial evidence from this trial, the jury had substantial evidence to determine that Jackson intentionally started the Roadside Fire charged in count 9. Substantial evidence placed Jackson at this fire within the assumed ignition window. The only reasonable conclusion to be drawn from the evidence pointed to Jackson's guilt. A rational jury could have found the essential elements of the crime beyond a reasonable doubt after viewing the evidence favorably for the prosecution. (People v. Johnson, supra, 26 Cal.3d at p. 576.) This evidence was reasonable, credible and of solid value. Accordingly, we affirm the conviction in count 9.

b. Count 10.

The Roadside Fire charged in count 10 occurred on June 8, 2013. This fire was reported around 2:57 p.m. and occurred by Lilley Mountain Drive. Jackson's motorcycle was spotted on camera at 2:36 p.m., about 21 minutes before this fire was reported. At that time, he was about five minutes away from the fire site. Smith calculated a 15-minute ignition window for this fire. McCann investigated this fire and performed a cause exclusion analysis. He determined arson caused this fire.

Based on this record, and the totality of the circumstantial evidence from this trial, the jury had substantial evidence to determine that Jackson intentionally started the Roadside Fire charged in count 10. Substantial evidence placed Jackson at this fire within the assumed ignition window. The only reasonable conclusion to be drawn from the evidence pointed to Jackson's guilt. A rational jury could have found the essential elements of the crime beyond a reasonable doubt after viewing the evidence favorably for the prosecution. (People v. Johnson, supra, 26 Cal.3d at p. 576.) This evidence was reasonable, credible and of solid value. Accordingly, we affirm the conviction in count 10.

c. Count 11.

The Roadside Fire charged in count 11 occurred on June 8, 2013. This fire was reported at around 3:17 p.m. and it occurred by John Muir Drive, just minutes away from the fire in count 10. Jackson's motorcycle was seen on a private camera system belonging to a resident on John Muir Drive. The motorcycle was seen at about 2:52 p.m. traveling away from this fire. Smith calculated a 25-minute ignition window for the fire charged in count 11.

The prosecution's experts, however, made it clear that a time delayed incendiary device, such as the type the prosecution alleged Jackson to have used, might take up to 20 minutes, at the longest, to ignite a wildland fire. No evidence established or even suggested that such a device could take more than 20 minutes to start a fire. In fact, the testimony strongly suggested that even a 20-minute delay was rare.

Based on the totality of the circumstantial evidence, a rational jury could have determined that arson caused this fire. The prosecution, however, did not provide substantial evidence linking Jackson as the perpetrator of this fire. It is speculative to believe that Jackson started this fire with a time delayed incendiary device that took 25 minutes to ignite. A finding of fact must be an inference drawn from the evidence and not a speculation as to probabilities without evidence. (People v. Davis (2013) 57 Cal.4th 353, 360.) As such, based on this record, insufficient evidence links Jackson as the arsonist for this fire. Accordingly, the conviction in count 11 is reversed.

d. Count 13.

The Roadside Fire charged in count 13 occurred on June 9, 2013. This fire was reported around 11:35 p.m. and burned by Road 400 east of Yosemite Springs Parkway. Jackson's truck was spotted on camera at 11:50 a.m. traveling on Road 400. He was less than two minutes away from this fire. At trial, Smith opined that Jackson's vehicle would have been at this fire's origin around 11:49 a.m. Smith agreed, however, that this fire was reported about 14 minutes before that time.

The prosecution's evidence for this fire placed Jackson at this fire about 14 minutes after it was reported. The prosecution presented no evidence that placed Jackson at the site of this fire before it was reported.

Based on the totality of the circumstantial evidence, a rational jury could have determined that arson caused this fire. The prosecution, however, did not provide substantial evidence linking Jackson as the perpetrator of this fire. A finding of fact must be an inference drawn from the evidence and not a speculation as to probabilities without evidence. (People v. Davis, supra, 57 Cal.4th at p. 360.) As such, substantial evidence does not establish that Jackson was the perpetrator of the fire charged in count 13. We reverse this conviction for insufficient evidence.

e. Count 21.

The Roadside Fire charged in count 21 occurred on June 16, 2013. This fire was reported about 1:21 p.m. and occurred by Road 400. Jackson's truck was seen at 1:08 p.m. and 1:13 p.m., first traveling towards this fire site and then away. Smith calculated a 9- to 10-minute ignition window for this fire. McCann investigated this fire. After conducting a cause exclusion analysis, he ruled out everything except arson.

Based on this record, and the totality of the circumstantial evidence from this trial, the jury had substantial evidence to determine that Jackson intentionally started the Roadside Fire charged in count 21. Substantial evidence placed Jackson at this fire within the assumed ignition window. The only reasonable conclusion to be drawn from the evidence pointed to Jackson's guilt. A rational jury could have found the essential elements of the crime beyond a reasonable doubt after viewing the evidence favorably for the prosecution. (People v. Johnson, supra, 26 Cal.3d at p. 576.) This evidence was reasonable, credible and of solid value. Accordingly, we affirm the conviction in count 21.

f. Count 22.

The Roadside Fire charged in count 22 occurred on June 16, 2013. This fire was reported about 3:35 p.m. and occurred by John Muir Drive. Jackson's truck was seen on John Muir Drive at 3:25 p.m. traveling away from this fire site. Smith calculated a 13-minute ignition window for this fire. McCann investigated this fire. After determining that a recovered catalytic converter did not start this fire, McCann determined that this fire was caused by arson.

Based on this record, and the totality of the circumstantial evidence from this trial, the jury had substantial evidence to determine that Jackson intentionally started the Roadside Fire charged in count 22. Substantial evidence placed Jackson at this fire within the assumed ignition window. The only reasonable conclusion to be drawn from the evidence pointed to Jackson's guilt. A rational jury could have found the essential elements of the crime beyond a reasonable doubt after viewing the evidence favorably for the prosecution. (People v. Johnson, supra, 26 Cal.3d at p. 576.) This evidence was reasonable, credible and of solid value. Accordingly, we affirm the conviction in count 22.

g. Count 23.

The Roadside Fire charged in count 23 occurred on June 16, 2013. This fire was reported about 8:16 p.m. and burned along Highway 41. Jackson's truck was seen at about 8:00 p.m. traveling towards this fire on Yosemite Springs Parkway about four and a half minutes away. Smith calculated an 11-minute ignition window for this fire. McCann investigated this fire. After determining that a recovered electronic mirror could not have started this fire, McCann could not exclude arson or a vehicle as the cause for this fire.

Based on this record, and the totality of the circumstantial evidence from this trial, the jury had substantial evidence to determine that Jackson intentionally started the Roadside Fire charged in count 23. Substantial evidence placed Jackson at this fire within the assumed ignition window. The only reasonable conclusion to be drawn from the evidence pointed to Jackson's guilt. A rational jury could have found the essential elements of the crime beyond a reasonable doubt after viewing the evidence favorably for the prosecution. (People v. Johnson, supra, 26 Cal.3d at p. 576.) This evidence was reasonable, credible and of solid value. Accordingly, we affirm the conviction in count 23.

h. Count 24.

The Roadside Fire charged in count 24 occurred on June 20, 2013. This fire was reported at 1:46 p.m. and burned near Road 415. Jackson's truck was seen at about 1:32 p.m. traveling towards this fire. He was 30 to 45 seconds away from this fire's site. Smith calculated a 13-minute ignition window for this fire. Cacho investigated this fire. After performing a cause exclusion analysis, he was unable to exclude either a vehicle or arson as the cause of this fire.

Based on this record, and the totality of the circumstantial evidence from this trial, the jury had substantial evidence to determine that Jackson intentionally started the Roadside Fire charged in count 24. Substantial evidence placed Jackson at this fire within the assumed ignition window. The only reasonable conclusion to be drawn from the evidence pointed to Jackson's guilt. A rational jury could have found the essential elements of the crime beyond a reasonable doubt after viewing the evidence favorably for the prosecution. (People v. Johnson, supra, 26 Cal.3d at p. 576.) This evidence was reasonable, credible and of solid value. Accordingly, we affirm the conviction in count 24.

i. Count 25.

The Roadside Fire charged in count 25 occurred on June 20, 2013. This fire was reported at 1:58 p.m. and occurred near Highway 41 just north of Road 416. Jackson's truck was seen at about 1:32 p.m. traveling towards this fire. He was about seven minutes away from this fire's site. Smith calculated a 20-minute ignition window for this fire. Cacho investigated this fire. After performing a cause exclusion analysis, he was unable to exclude either a vehicle, a discarded cigarette or arson as the cause of this fire.

Based on this record, and the totality of the circumstantial evidence from this trial, the jury had substantial evidence to determine that Jackson intentionally started the Roadside Fire charged in count 25. Substantial evidence placed Jackson at this fire within the assumed ignition window. The only reasonable conclusion to be drawn from the evidence pointed to Jackson's guilt. The prosecution's experts made it clear that a time delayed incendiary device, such as the type the prosecution alleged Jackson to have used, might take up to 20 minutes to ignite a wildland fire. As such, a rational jury could have found the essential elements of the crime beyond a reasonable doubt after viewing the evidence favorably for the prosecution. (People v. Johnson, supra, 26 Cal.3d at p. 576.) This evidence was reasonable, credible and of solid value. Accordingly, we affirm the conviction in count 25.

j. Count 26.

The Roadside Fire charged in count 26 occurred on June 23, 2013. This fire was reported at 10:20 a.m. and burned by Highway 41 north of Yosemite Lakes. Jackson's truck was seen at about 9:49 a.m. traveling towards this fire on Yosemite Springs Parkway. He was about 13 minutes away from this fire's site. Smith calculated an 18-minute ignition window for this fire. Pimentel investigated this fire. He did a cause exclusion analysis and could not rule out arson or a vehicle.

In addition to the camera evidence, two CAL FIRE employees saw Jackson's truck by the site of the fire charged in count 27, which occurred about five minutes away from this fire. A white male was observed flagging down and slowing traffic.³⁷

37 In his opening brief, Jackson admits that he was at the site of the fire charged in count 27 after fire personnel responded.

Based on this record, and the totality of the circumstantial evidence from this trial, the jury had substantial evidence to determine that Jackson intentionally started the Roadside Fire charged in count 26. Substantial evidence placed Jackson at this fire within the assumed ignition window. The only reasonable conclusion to be drawn from the evidence pointed to Jackson's guilt. A rational jury could have found the essential elements of the crime beyond a reasonable doubt after viewing the evidence favorably for the prosecution. (People v. Johnson, supra, 26 Cal.3d at p. 576.) This evidence was reasonable, credible and of solid value. Accordingly, we affirm the conviction in count 26.

k. Count 27.

The Roadside Fire charged in count 27 occurred on June 23, 2013. This fire was reported at 10:27 a.m. and occurred by Highway 41. Jackson's truck was seen at about 9:49 a.m. traveling towards this fire on Yosemite Springs Parkway. He was about 18 minutes away from this fire's site. In addition to the camera evidence, two CAL FIRE employees saw Jackson's truck by the site of this fire. A white male was observed flagging down and slowing traffic.³⁸ Smith calculated a 20-minute ignition window for this fire. McCann investigated this fire. After a cause exclusion analysis, he ruled out everything but arson.

38 In his opening brief, Jackson admits that he was at this fire site after fire personnel responded.

Based on this record, and the totality of the circumstantial evidence from this trial, the jury had substantial evidence to determine that Jackson intentionally started the Roadside Fire charged in count 27. Substantial evidence placed Jackson at this fire within the assumed ignition window. The only reasonable conclusion to be drawn from the evidence pointed to Jackson's guilt. The prosecution's experts made it clear that a time delayed incendiary device, such as the type the prosecution alleged Jackson to have used, might take up to 20 minutes to ignite a wildland fire. As such, a rational jury could have found the essential elements of the crime beyond a reasonable doubt after viewing the evidence favorably for the prosecution. (People v. Johnson, supra, 26 Cal.3d at p. 576.) This evidence was reasonable, credible and of solid value. Accordingly, we affirm the conviction in count 27.

l. Count 28.

The Roadside Fire charged in count 28 occurred on June 23, 2013. This fire was reported at about 3:31 p.m. and occurred off Road 400 near Lilley Mountain Drive in Yosemite Lakes. Jackson's truck was seen on camera at about 3:16 p.m. traveling towards this fire site. He was spotted two minutes later on the same camera traveling away from this fire site. He was about one minute away from this fire's site when spotted both times. Smith calculated an ignition window of 13 to 14 minutes for this fire. Pimentel investigated this fire. After conducting a cause exclusion analysis, he determined that arson caused this fire.

In addition to the camera evidence, Quinn observed Jackson driving his truck on Yosemite Springs Parkway at about 3:19 p.m. Jackson was driving away from the site of this fire. Jackson was seen looking behind him the entire time but no vehicles were behind him. While responders were suppressing this fire, Jackson parked his truck nearby and told fire personnel about a suspicious car he had observed.

Based on this record, and the totality of the circumstantial evidence from this trial, the jury had substantial evidence to determine that Jackson intentionally started the Roadside Fire charged in count 28. Substantial evidence placed Jackson at this fire within the assumed ignition window. The only reasonable conclusion to be drawn from the evidence pointed to Jackson's guilt. A rational jury could have found the essential elements of the crime beyond a reasonable doubt after viewing the evidence favorably for the prosecution. (People v. Johnson, supra, 26 Cal.3d at p. 576.) This evidence was reasonable, credible and of solid value. Accordingly, we affirm the conviction in count 28.

m. Count 29.

The Roadside Fire charged in count 29 occurred on June 23, 2013. This fire was reported around 7:20 p.m. and occurred about a minute's drive from appellant's residence. Surveilling personnel heard Jackson's truck leave his residence at about 7:19 p.m. It drove towards this fire's ultimate location and returned at about 7:25 p.m. The truck remained within audible distance of fire personnel the entire time. Smith calculated an ignition window of 16 to 20 minutes for this fire. McCann investigated this fire. He went through 13 causes and exclusions, and he was unable to exclude arson as the cause of this fire.

About two days after this fire, appellants were arrested. Waterman told an investigator that Jackson's truck never left their residence on the evening of this fire. She told LaClair that Jackson did not leave that night. She said an "imaginary" person must have been driving his truck.

Based on this record, and the totality of the circumstantial evidence from this trial, the jury had substantial evidence to determine that Jackson intentionally started the Roadside Fire charged in count 29. Substantial evidence placed Jackson at this fire within the assumed ignition window. The only reasonable conclusion to be drawn from the evidence pointed to Jackson's guilt. A rational jury could have found the essential elements of the crime beyond a reasonable doubt after viewing the evidence favorably for the prosecution. (People v. Johnson, supra, 26 Cal.3d at p. 576.) This evidence was reasonable, credible and of solid value. Accordingly, we affirm the conviction in count 29.

n. Count 30.

The final Roadside Fire occurred on June 25, 2013. This occurred by Road 400 south of Lilley Mountain Drive and was reported at 3:28 p.m. Prior to this fire, fire officials had placed a GPS tracker on Jackson's truck. The GPS data showed Jackson's truck passing this fire's site at 3:05 p.m. and again at 3:08 p.m. Both times, his truck drove at about 37 miles per hour. The camera footage corroborated the GPS data regarding when Jackson's truck approached and left this fire site. Smith opined that this fire had either a 20- or 23-minute ignition window, depending on when Jackson placed an ignition source. Gilbert investigated this fire. He performed a cause exclusion analysis and determined that arson caused this fire.

On the day of this fire, CAL FIRE officers, including Quinn, were following Jackson's truck. At about 3:03 p.m., Quinn was parked near Road 400 and he saw Jackson's vehicle pass. At about 3:14 p.m., another surveilling officer confirmed that a white male was driving Jackson's truck. At about 3:28 p.m., Quinn heard the dispatch for this fire. He turned and looked back. He could see a large column of smoke building. Quinn had passed that fire location two times while following Jackson's vehicle, but he had not previously seen any flames or smoke. Quinn had last passed that site sometime after 3:08 p.m.

Based on this record, and the totality of the circumstantial evidence from this trial, the jury had substantial evidence to determine that Jackson intentionally started the Roadside Fire charged in count 30. Substantial evidence placed Jackson at this fire within the assumed ignition window. The only reasonable conclusion to be drawn from the evidence pointed to Jackson's guilt. The prosecution's experts made it clear that a time delayed incendiary device, such as the type the prosecution alleged Jackson to have used, might take up to 20 minutes to ignite a wildland fire. The jury could have concluded that Jackson threw an incendiary device the second time he passed this fire site, creating a 20-minute ignition window. We are to presume the existence of any fact the jury could have reasonably deduced from the evidence in support of the judgment. (People v. Johnson, supra, 26 Cal.3d at p. 576.) As such, a rational jury could have found the essential elements of the crime beyond a reasonable doubt after viewing the evidence favorably for the prosecution. (Ibid.) This evidence was reasonable, credible and of solid value. Accordingly, we affirm the conviction in count 30.

When the entire record is viewed in the light most favorable to the judgment, the jury had substantial evidence to convict Jackson for the Roadside Fires in counts 9, 10, 21, 22, 23, 24, 25, 26, 27, 28, 29, and 30. A rational jury could have found the essential elements of arson beyond a reasonable doubt.

Our conclusion is not altered by the problems surrounding the video evidence or the number of other potential suspects. It is the jury, and not this reviewing court, which must be convinced of Jackson's guilt beyond a reasonable doubt. (People v. Abilez, supra, 41 Cal.4th at p. 504.) The circumstances reasonably justify the jury's findings so that a reversal of the judgment is not warranted. (Ibid.) In fact, the circumstances are not reasonably reconciled with a contrary finding.

Further, the jury acquitted Jackson of numerous other Roadside Fires (counts 1-3, 7-8, 12 and 17). We are convinced that the jury would have acquitted Jackson of the remaining Roadside Fires had it felt reasonable doubt existed from (1) the numerous other potential suspects; (2) the limited physical evidence supporting the theory of a time delayed incendiary device for all Roadside Fires; (3) the assumption that Jackson drove past the applicable fire sites; (4) any gap in camera coverage because CAL FIRE's fifth camera was not installed until the end of its investigation; (5) how CAL FIRE conducted its investigation; and/or (6) how CAL FIRE determined causation. From the verdicts rendered, it is clear the jury carefully examined the record and weighed the evidence. Substantial evidence (other than for counts 11 and 13) supports the jury's verdicts.

We note that, under the doctrine of chances, similar results do not usually occur through abnormal causes. (People v. Robbins (1988) 45 Cal.3d 867, 879-880, superseded by statute on another ground as stated in People v. Jennings (1991) 53 Cal.3d 334, 387, fn. 13.) Instead, a recurrence of a similar result (i.e., an unlawful act) tends, with each instance, to negate accidental, inadvertent, good faith, or other innocent explanations. (People v. Robbins, supra, 45 Cal.3d at p. 880.) Repeated results increasingly tend to suggest criminal intent. (Ibid.) The doctrine of chances tells us it is extremely unlikely that, through either bad luck or coincidence, an innocent person would be associated with so many arson fires occurring so often in so many different areas. (People v. Erving (1998) 63 Cal.App.4th 652, 663.)

We acknowledge that courts do not use the doctrine of chances to analyze whether sufficient evidence supports a judgment. Instead, this doctrine is employed to justify the admission of evidence of uncharged acts to prove criminal intent and identity. (People v. Erving, supra, 63 Cal.App.4th at p. 662.) However, although the doctrine does not apply in this situation, its reasoning and logic are relevant and meaningful here.

When this record is reviewed in the light most favorable to the judgment, which it must be under these circumstances, the jury had substantial circumstantial evidence to determine that these 14 Roadside Fires were caused by arson and that Jackson was the perpetrator for all of the fires except counts 11 and 13. The evidence, and the logical inferences arising therefrom, is reasonable, credible, and of solid value. Accordingly, we affirm the convictions in counts 9-10 and 21-30. We reverse, however, the convictions in counts 11 and 13 for insufficient evidence. The convictions reversed for insufficient evidence may not be retried. (Lockhart v. Nelson (1988) 488 U.S. 33, 39.)

2. Substantial evidence regarding the Cluster Fires.

The fires around appellants' residence were suspicious and they stopped when appellants were arrested. Appellants, and very few other people, could access the areas where these fires occurred without alerting nearby dogs or neighbors. Appellants had numerous incendiary tools. Despite not being a smoker and claiming she could only start a fire in a fireplace with paper, kindling and a "clicky" thing, Waterman possessed numerous matches and had a magnesium strike block in her Sequoia. The totality of the circumstantial evidence overwhelmingly suggests that Jackson and/or Waterman were responsible for the Cluster Fires.

This logical inference, however, is not substantial evidence that either Jackson and/or Waterman started a particular Cluster Fire. Other people lived in the area where these fires occurred, and people did access this area. In some instances, drivers or bicyclists were seen or heard in close proximity to a fire. Under these circumstances, more evidence is necessary to link appellants to these suspicious fires beyond the general overwhelming inference of guilt.

As we explain below, as to Waterman, we find substantial evidence supporting the convictions in counts 15, 16, 20 and 31. As to Jackson, we find substantial evidence supporting the conviction only in count 31. The remaining guilty verdicts for the Cluster Fires have insufficient evidence and they are reversed on that basis. The convictions we reverse for Jackson are counts 5, 6, 14, 15, 16, and 19. The convictions we reverse for Waterman are counts 14 and 18.

a. Counts 5 and 6.

The fires charged in counts 5 and 6 occurred on May 22, 2013. These fires were reported at about 8:30 p.m. Fire responders arrived and saw Jackson standing about 50 feet from the fire holding a hose. Jackson said he was planning to use it if the fire got closer to his residence.

Although Jackson was home when these fires occurred, nothing else in this record establishes or suggests his guilt. Unlike the eyewitness testimony in count 31 (where Jackson was seen holding a hose that was not spraying water, he acted strangely and the responders believed he was involved in starting that fire), the evidence here does not suggest Jackson started these fires.

Other than a strong inference that appellants must have been involved in the Cluster Fires, nothing substantially links Jackson as an arsonist for these convictions. As such, insufficient evidence supports the convictions in counts 5 and 6 and they are reversed.

b. Count 14.

The fire charged in count 14 occurred on June 9, 2013. Fire personnel responded around 6:50 p.m. Fire investigators did not find anything significant in the origin area for this fire.

The day after this fire, McCann spoke with Waterman at her residence and he noticed burnt shrubs next to their front walkway. She told him they had been gone that night and she had noticed the shrubs upon returning, but she had not reported it. McCann found that odd. At trial, Waterman disputed that her shrubs appeared burnt and she believed she was home when this fire occurred.

Nothing in this record establishes or suggests that either Jackson or Waterman had any involvement in this fire. We reject respondent's claim that Waterman made inconsistent statements that amount to substantial evidence of her guilt. Other than a strong inference that appellants must have been involved in the Cluster Fires, nothing substantially links appellants as arsonists for this conviction. As such, insufficient evidence supports their convictions in count 14 and they are reversed.

c. Counts 15 and 16.

The fires charged in counts 15 and 16 both occurred on June 10, 2013, next to each other on Revis Circle East. They occurred near appellants' residence about 50 feet apart. The first fire was reported at about 8:38 p.m. Responders arrived and found the other fire also burning. At trial, Pimentel opined that these fires started separately. A cause exclusion analysis was performed, which ruled out everything but arson. Felix, the fire responder, opined that the larger fire (count 15) started first.

On the night these fires were suppressed, Waterman made statements to various people that suggested her guilt. McCann told the jury that he had not found her initial story credible. Waterman had told fire personnel that she had heard either a bicycle and/or a vehicle, which prompted her to go outside and investigate. She saw smoke turn into active flames near her neighbor's driveway. She smelled a chemical. She claimed the fire was about the size of a roll of duct tape before it grew larger.

Waterman gave a television interview that night and claimed she had "just happened to walk up and spot the fire starting" and she had called 911. At trial, the reporter indicated that Waterman had sought her out for the interview.

At trial, without mentioning either a bicycle or a vehicle, Waterman claimed she was walking outside with Jackson as they did regularly because of the past fires. She was near her neighbor's driveway when a fire started. She saw the second fire start when she hung up with 911. She told the jury that the television reporter had approached her for the interview.

Nothing substantially links Jackson as an arsonist for these fires other than the strong inference appellants were responsible for all of the Cluster Fires. In contrast, Waterman gave conflicting statements about why she was outside and the chronology of events. She admitted she was at the origin of one fire and her trial testimony strongly implied that the driveway fire (count 16) started first. Felix, however, said the other larger fire (count 15) started first.

Based on her inconsistent statements and testimony, a reasonable jury could have determined that Waterman was lying about which fire started first and how these fires started. A reasonable jury could have also determined that Waterman lied about the circumstances surrounding the television interview. Her efforts to appear on television suggested a motive to receive attention over these fires. These facts are reasonably deduced from the evidence in support of Waterman's guilty verdicts and we are to presume the existence of any fact the jury could have reasonably deduced from the evidence in support of the judgment. (People v. Johnson, supra, 26 Cal.3d at p. 576.)

Combined with the overwhelming inference that either Jackson and/or Waterman were responsible for all of the Cluster Fires, a reasonable jury could have determined that Waterman started these fires. This evidence was reasonable, credible and of solid value. Accordingly, we reverse these convictions against Jackson in counts 15 and 16 but affirm these convictions against Waterman.

d. Counts 18 and 19.

The fires in counts 18 and 19 both occurred on June 12, 2013. The fire in count 18 occurred very close to appellants' residence and it was reported at about 4:28 p.m. The fire charged in count 19 was reported at about 9:30 p.m. These fires occurred about 200 yards apart. Following a cause exclusion analysis, Pimentel determined arson for both fires.

At trial, Jackson's father testified that he and Jackson were in appellants' pool at their residence when the first fire started. Waterman was getting ready for her soccer match and she never went near the first fire. Some of Waterman's teammates confirmed that she attended a soccer match that night in Fresno. She testified that she returned home around 10:45 p.m. and Jackson was asleep in bed.

The jury found Jackson not guilty of the fire charged in count 18 but it found Waterman guilty for this count. In contrast, the jury found Jackson guilty of the fire charged in count 19, but it found Waterman not guilty for that count.

Although she was home at the time, nothing else suggests that Waterman had any involvement in starting the fire charged in count 18. In addition, although Waterman testified that Jackson was home asleep at about 10:45 p.m. that night, nothing else suggests that he had any involvement in starting the fire charged in count 19. Other than a strong inference that appellants must have been involved in the Cluster Fires, nothing substantially links appellants as the arsonists for these convictions. As such, insufficient evidence supports the convictions in counts 18 and 19 and they are reversed.

e. Count 20.

For the fire charged in count 20, Cacho determined arson and LaClair identified Waterman as the suspect who was seen entering the area just before this fire started. LaClair's testimony raised obvious credibility concerns because he initially described a 180 pound suspect wearing blue shorts, a white shirt, black hair and black sunglasses. He testified that Waterman was then spotted wearing brown clothing and she appeared smaller. However, LaClair explained to the jury why and how he subsequently believed that Waterman was that suspect.

During closing arguments, the prosecutor asked the jury to resolve this credibility dispute. The prosecutor showed the jury a photograph (exhibit 51) of Waterman wearing a baggy soccer uniform. The prosecutor contended that LaClair's assessment of Waterman was reasonable under the circumstances. She asked the jury to find it believable.

In rendering its verdict on this count, it is clear that the jury ultimately resolved this credibility dispute against Waterman. The jury had the luxury of seeing LaClair's demeanor while he explained himself. The jurors also observed Waterman first hand in the courtroom, they saw how she moved (which may have mirrored LaClair's descriptions of how the suspect moved with a particular gait), and the jurors were able to observe Waterman testify. The jurors were able to glean important details from their firsthand observations.

It is not our role in reviewing the sufficiency of the evidence to reweigh evidence or judge witness credibility. (People v. Lindberg, supra, 45 Cal.4th at p. 27.) Of course, inherently improbable testimony may be rejected on appeal. (People v. Mayberry (1975) 15 Cal.3d 143, 150; People v. Ontiveros (1975) 46 Cal.App.3d 110, 117.) To reject on appeal a witness who was believed by the trier of fact, either the witness's statements must be physically impossible that they are true or their falsity must be apparent without resorting to inferences or deductions. (People v. Mayberry, supra, 15 Cal.3d at p. 150.) The trier of fact has the exclusive role to determine witness credibility and the truth or falsity of the facts upon which a determination rests. (Ibid.) A judgment should not be reversed because of conflicts or even testimony that is subject to justifiable suspicion. (Ibid.)

Here, LaClair's testimony was not inherently improbable. He explained to the jury what happened and how he reached his ultimate conclusion that Waterman was the suspect seen near this fire. Two days after this encounter, LaClair saw Waterman while he was conducting surveillance near her residence. Another officer took pictures of Waterman, and the photos were shown to the jury. Waterman was wearing sunglasses, a blue shirt, and blue shorts. She was walking in the area behind a nearby vacant house. She picked up a stick and began beating some tall grass with it. LaClair believed she was looking for concealed people.

At trial, LaClair asserted that the shorts, sunglasses and shoes in Waterman's pictures matched what the suspect had worn. He also believed that Waterman's posturing, her walk and her hair color were similar to the suspect's. At trial, he confirmed that the suspect had weighed 180 pounds, and, in court, he estimated Waterman to be somewhere between 110 and 120 pounds. He confirmed that he had not initially believed that Waterman and the suspect were the same person. However, he told the jury he was certain Waterman was that suspect. LaClair was a percipient witness and his credibility was for the jury to resolve. (See People v. Lindberg, supra, 45 Cal.4th at p. 27 [it is not our role on appellate review to reweigh evidence or judge witness credibility].)

We note the jury had other evidence that tended to bolster LaClair's credibility in this regard. Waterman admitted that she had encountered LaClair and followed him when he walked away. She denied weighing 180 pounds then, but agreed she might have weighed 140 pounds when this encounter occurred. During her interview with LaClair, she made a statement that suggested she might have been in that area just before this fire.

Finally, Cacho spoke with Waterman at this fire scene. Waterman admitted that she had seen a person in camouflage. She had asked that person if he was conducting surveillance in the area for the fires. A reasonable juror could have viewed this statement as suggesting her guilt.

In doing our review, we are not required to ask whether we believe the trial evidence established guilt beyond a reasonable doubt. (People v. Johnson, supra, 26 Cal.3d at p. 576.) Rather, the issue is whether any rational jury could have found the essential elements of the crime beyond a reasonable doubt after viewing the evidence favorably for the prosecution. (Ibid.) Based on LaClair's percipient testimony, which the jury found credible, and the other circumstances of this fire, substantial evidence exists in this record to support the jury's conviction of Waterman in count 20. This evidence was reasonable, credible and of solid value. A rational jury could have found Waterman guilty of this crime beyond a reasonable doubt. Accordingly, we affirm the conviction in count 20.

f. Count 31.

Following Waterman's encounter with LaClair on June 14, 2013, the fires around appellants' residence halted for the next 11 days. Only after CAL FIRE officials made a "big scene" of removing yellow tape from the area on June 21, 2013, did the final Cluster Fire occur on June 25, 2013.

Gilbert responded to the fire charged in count 31 and he investigated it. He determined arson. When this fire occurred, surveilling officers saw a smoke column rising behind appellants' residence. Waterman was seen walking slowly away from her residence. She stopped and looked back towards her residence two or three times. She was looking towards the rising smoke. Around that time, surveilling officers could hear emergency vehicles approaching the area. When the sirens started to come up her street, Waterman stopped, turned, threw her hands into the air and began running towards her residence. She was heard yelling about fire and expressing surprise that the arsonist lit another one.

Responding firefighters saw Jackson standing near his residence and near this fire. He was holding a hose that did not have water coming from it. Jackson made eye contact with one firefighter and Jackson froze. It appeared that Jackson was faking efforts to suppress this fire.

Around that time, fire prevention officers moved in on Jackson to arrest him. A firefighter saw the officers tell Jackson to get down. At that point, Jackson was observed still holding his hose with no water coming out of it.

At trial, Waterman claimed that she had given Jackson the hose after she had turned it on "full blast." She then claimed that she had "turned it down a little" to keep the water from spraying everywhere. She said the hose had a "fireman's nozzle" that allowed the operator to adjust the stream. She claimed that, after giving the hose to Jackson, she ran towards the fire and then ran to their driveway. She wanted to clear some tree limbs that a neighbor had placed so that fire personnel could reach the fire.

We are to presume the existence of any fact the jury could have reasonably deduced from the evidence in support of the judgment. (People v. Johnson, supra, 26 Cal.3d at p. 576.) A reasonable jury could have determined that Jackson was faking efforts to suppress this fire. A reasonable jury could have determined that Waterman knew about this fire before sirens were heard approaching her street and she pretended to be surprised. A reasonable jury could have determined that both appellants acted suspiciously in a way that suggested their guilt.

Based on this record, a reasonable jury could have found both Jackson and Waterman guilty of arson beyond a reasonable doubt. This conclusion is based on appellants' respective proximity to this particular fire and their observed reactions to it, along with the overwhelming inference that they were responsible for all of the Cluster Fires. The evidence supporting these convictions was reasonable, credible and of solid value. Accordingly, we affirm their convictions.

g. Conclusion.

Based on the analysis above, we affirm the convictions against Waterman in counts 15, 16, 20 and 31. We affirm the conviction against Jackson in count 31. We reverse for insufficient evidence the convictions in counts 5, 6, 14, 15, 16 and 19 (Jackson) and in counts 14 and 18 (Waterman). The convictions reversed for insufficient evidence may not be retried. (Lockhart v. Nelson, supra, 488 U.S. at p. 39.)

3. Substantial evidence regarding conspiracy (count 32).

"A conviction for conspiracy requires proof of four elements: (1) an agreement between two or more people, (2) who have the specific intent to agree or conspire to commit an offense, (3) the specific intent to commit that offense, and (4) an overt act committed by one or more of the parties to the agreement for the purpose of carrying out the object of the conspiracy. [Citations.]" (People v. Vu (2006) 143 Cal.App.4th 1009, 1024.) Circumstantial evidence can be used to prove these elements, and the agreement proved can be a tacit one. (Id. at p. 1025.)

At trial, Waterman denied having any agreement with Jackson to start any of the charged fires. However, appellants were convicted of, and we affirm, multiple counts of arson, including one Cluster Fire in which they acted together, count 31.

In addition, Waterman lied to an investigator regarding Jackson's movements relative to the Roadside Fire charged in count 29. For this fire, surveilling personnel heard Jackson's truck leave his residence. It drove toward this fire's ultimate location, and it returned home at about 7:25 p.m. The truck remained within audible distance of fire personnel the entire time. About two days after this fire, appellants were arrested. Waterman told LaClair that Jackson's truck never left their residence on the evening of this fire and Jackson did not leave that night. She said an "imaginary" person must have been driving his truck.

Based on this record, a rational jury could have determined beyond a reasonable doubt that Jackson and Waterman conspired together to commit arson. This evidence was reasonable, credible and of solid value. Accordingly, we affirm these convictions. V. In Light Of The Reversals Already Ordered, We Need Not Address Jackson's Final Claim.

Jackson argues that he was not convicted properly on counts 5, 6, 14, 15, 16 and 19 because the complaint did not charge him with these counts, no evidence was presented against him on them at the preliminary hearing, and he was not bound over for trial on them. We need not, however, resolve these issues because we have reversed these counts for insufficient evidence.

DISPOSITION

We affirm the arson convictions against Jackson in counts 9, 10, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30 and 31. We affirm the arson convictions against Waterman in counts 15, 16, 20 and 31. We affirm the conspiracy convictions against both Jackson and Waterman in count 32. We reverse for insufficient evidence the arson convictions against Jackson in counts 5, 6, 11, 13, 14, 15, 16, and 19. We reverse for insufficient evidence the arson convictions against Waterman in counts 14 and 18. The matter is remanded to the trial court for resentencing consistent with this opinion. Upon resentencing, the trial court shall prepare amended abstracts of judgment and forward them to the appropriate authorities. We otherwise affirm the judgments.

/s/_________

LEVY, Acting P.J. I CONCUR: /s/_________
BLACK, J.^* SMITH, J. - Concurring and dissenting.

* Judge of the Fresno Superior Court assigned by the Chief Justice pursuant to article VI section 6 of the California Constitution.

In my view, the prosecution's evidence in this case was fundamentally deficient. Based on insufficiency of the evidence, I would reverse all the arson convictions against both defendants except for three against Jackson, and I would also reverse the convictions for conspiracy. For the reasons set forth in detail below, I respectfully dissent from the affirmances of Jackson's convictions on counts 9, 10, 21, 22, 23, 24, 25, 26, 27, 31, and 32, and from the affirmances of Waterman's convictions on counts 15, 16, 20, 31, and 32. I concur in the balance of the judgment, i.e., the affirmances of Jackson's convictions on counts 28, 29, 30, 33, and 34, the reversals of Jackson's convictions on counts 5, 6, 11, 13, 14, 15, 16, and 19, and the reversals of Waterman's convictions on counts 14 and 18.

No direct evidence tied Jackson or Waterman to any of the fires. For one group of fires, the prosecution's case for identifying the culprit consisted mainly of surveillance camera data indicating that in each instance, Jackson's pickup truck or motorcycle was one of many vehicles that passed by a camera miles from the fire during a 40-minute to two-hour window around the time the fire was reported, plus the fact that no other individual owned vehicles that were recorded during the window for all of those fires. For another group of fires, the main prosecution evidence of the identity of the perpetrators was the fact that this group was clustered around Jackson and Waterman's house, and they were home and sometimes present outdoors when firefighters arrived. The convictions depended on speculative inferences from the evidence, and thus were not proved beyond a reasonable doubt under the substantial evidence standard.

As I believe the majority opinion's account of the facts fails to convey adequately some of the weaknesses of the prosecution's case, I include my own account below. I. Information

The People filed a 34-count information against Jackson and Waterman on November 12, 2013. The charges are summarized in the following table. For convenience, the verdict for each count is also shown for each defendant, with G standing for guilty and NG for not guilty.

Cnt.	Incid-entNo.39	IncidentName	Address and Date	Pen. CodeSection	Defendant
1	4	Lilley 2	31391 Road 400 and BestWayMay 19, 2013	451, subd. (b)(arson ofinhabitedstructure)	Jackson (NG)
2	1	Springs 1	28401 Yosemite SpringsPkwy.May 11, 2013	451, subd. (c)(arson ofstructure orforest land)	Jackson (NG)
3	2	Ranger	43153 Ranger Circle Dr.May 12, 2013	451, subd. (c)	Jackson (NG)
4	3	Revis 1	30125 Revis Rd.May 18, 2013	451, subd. (c)	Jackson (NG)Waterman(NG)
5	5	Revis 2	43114 E. Revis CircleMay 22, 2013	451, subd. (c)	Jackson (G)Waterman(NG)
6	6	Smaller	43114 E. Revis CircleMay 22, 2013	451, subd. (c)	Jackson (G)Waterman(NG)
7	7	Brandy	Road 415 and Road 600May 24, 2013	451, subd. (c)	Jackson (NG)
8	8	Lakes 1	32510 Road 400June 1, 2013	451, subd. (c)	Jackson (NG)
9	9	Lakes 2	44410 Road 415June 8, 2013	451, subd. (c)	Jackson (G)
10	10	Mountain1	40705 Lilley Mountain Dr.June 8, 2013	451, subd. (c)	Jackson (G)
11	11	John Muir	42225 John Muir Dr.	451, subd. (c)	Jackson (G)

			June 8, 2013
12	12	Oakhurst[1]	38245 Road 425BJune 9, 2013	451, subd. (c)	Jackson (NG)
13	13	Springs 2	31877 Road 400June 9, 2013	451, subd. (c)	Jackson (G)
14	14	Parkway	43114 E. Revis CircleJune 9, 2013	451, subd. (c)	Jackson (G)Waterman (G)
15	15	Circle A	43117 E. Revis CircleJune 10, 2013	451, subd. (c)	Jackson (G)Waterman (G)
16	15	Circle B	43117 E. Revis CircleJune 10, 2013	451, subd. (c)	Jackson (G)Waterman (G)
17	16	Yosemite1	28111 Yosemite SpringsPkwy.June 12, 2013	451, subd. (c)	Jackson (NG)
18	17	Revis A	43117 E. Revis CircleJune 12, 2013	451, subd. (c)	Jackson (NG)Waterman (G)
19	17	Revis B	43114 E. Revis CircleJune 12, 2013	451, subd. (c)	Jackson (G)Waterman(NG)
20	19	Lakes 3	29971 Revis Rd.June 14, 2013	451, subd. (c)	Jackson (NG)Waterman (G)
21	20	Yosemite2	31965 Road 400June 16, 2013	451, subd. (c)	Jackson (G)
22	21	John	42058 John Muir Dr.June 16, 2013	451, subd. (c)	Jackson (G)
23	22	Mecca	30099 Hwy. 41June 16, 2013	451, subd. (c)	Jackson (G)
24	25	Lakes 4	33160 Sesame St.June 20, 2013	451, subd. (c)	Jackson (G)
25	26	416	31194 Hwy. 41	451, subd. (c)	Jackson (G)

			June 20, 2013
26	27	Highway	35788 Hwy. 41June 23, 2013	451, subd. (c)	Jackson (G)
27	28	Oakhurst2	38797 Hwy. 41June 23, 2013	451, subd. (c)	Jackson (G)
28	29	Lilley 1	31446 Yosemite SpringsPkwy.June 23, 2013	451, subd. (c)	Jackson (G)
29	30	RevisCourt	42917 Revis CourtJune 23, 2013	451, subd. (c)	Jackson (G)
30	31	Mountain[2]	29025 Road 400June 25, 2013	451, subd. (c)	Jackson (G)
31	32	Court	43114 E. Revis CircleJune 25, 2013	451, subd. (c)	Jackson (G)Waterman (G)
32			May 11, 2013 to June 25,2013	182, subd.(a)(1); 451,subd. (b) and(c)(conspiracy tocommitarson)	Jackson (G)Waterman (G)
33			May 2, 201340	243, subd. (b)(misdemeanorbattery on apeace officer	Jackson (G)
34			May 2, 2013	148, subd.(a)(1)(misdemeanorresistingarrest)	Jackson (G)

39 The fires charged in this case were given the collective name Gold Mine Fires by the California Department of Forestry and Fire Prevention (Cal Fire). Cal Fire also gave each fire a name, and numbered the fires with a series of incident numbers that differ from the numbers of the counts in the information.

40 The information stated the date for counts 33 and 34 incorrectly. In reality, these counts were based on Jackson's behavior when he was arrested on June 25, 2013, the date of the last charged fires.

The fires charged in the information formed two distinct groups. One group, consisting of 10 fires, was clustered within an area surrounding defendants' house at 43114 East Revis Circle in Yosemite Lakes Park. The size of this area is described variously in the record as a radius of 200 yards around defendants' house or an area a half-mile square.⁴¹ These were the fires charged in counts 4, 5, 6, 14, 15, 16, 18, 19, 20 and 31. Jackson was found guilty on seven of these counts (5, 6, 14, 15, 16, 19 and 31) and Waterman on six (14, 15, 16, 18, 20 and 31). According to a prosecution witness, these fires did not start near the roadside and a person intentionally starting them would have had to proceed to their points of origin on foot.

41 A circle with a radius of 200 yards has an area of about 125,000 square yards. A half-mile square (i.e., a square of which each side is a half-mile long, or 0.25 square miles) equals about 774,000 square yards. Half a square mile (0.5 square miles) is about 1.5 million square yards.

The second group, consisting of 21 fires, was scattered throughout and outside the Yosemite Lakes Park development. Each of these fires originated near a roadside. These were the fires charged in counts 1-3, 7-13, 17, and 21-30. Only Jackson was charged in the roadside fires, and he was found guilty of setting 14 of them (counts 9-11, 13, and 21-30). II. Trial Evidence

A. Evidence of the identity of Jackson as the person who set the roadside fires

1. Surveillance camera network

The first of the charged fires took place on May 11, 2013. By May 22, 2013, there had been six fires, and Cal Fire, which was investigating, suspected arson. Bernie Quinn, who was Cal Fire's battalion chief in charge of the Law Enforcement and Fire Prevention Bureau for Madera, Mariposa, and Merced Counties, asked Cal Fire to provide surveillance cameras for installation around the development to help with the investigation. On May 28, 2013, and June 4, 2013, work was carried out to install four cameras.

The installation of the cameras was part of an investigation that would later grow to include a live surveillance team and a staff to review the camera recordings. Quinn testified that, because of the fires, the situation in the Yosemite Lakes Park community had become tense. The development had a security force that had increased its hours and rented vehicles to try to find persons responsible for the fires. Citizen groups had begun patrolling. These civilian efforts had not resulted in the gathering of useful information, however. Investigators were "inundated" with civilian reports, and responding to these "was taking away from their duties to investigate the cause of these fires." Quinn thought the cameras and staff would allow the investigation to be more orderly and objective. Under examination by the prosecutor, he testified as follows:

"Q With regard to the situation within Yosemite Lakes Park, was there an emotional component to what you saw going on?

"A With the public?

"Q Yes.

"A Yes.

"Q And how would you describe that?

"A The public was upset and the public was upset with CAL FIRE. They wanted—and law enforcement. They wanted these people caught. They wanted these fires to stop. I mean, we had already lost a house and some vehicles [in the fire in count 1, on which there was no conviction]. People had suffered some damages and losses. They wanted it to end.

"Q And you mentioned vigilantes. Is there anything particularly concerning to you about the fact that you felt there was a rise in vigilantism within the community?

"A Yeah. I just felt it could jeopardize our chances of catching our arson suspects. It could interfere with our investigations.

"Q And given the emotional tenor and the resources that you had, did you feel it was best to focus your resources on more objective evidence rather than the subjective stuff you were getting from the residents?

"A Yes."

Quinn further testified that there was "a sense of pressure" on him from citizens "to find [out] what was happening," because it was his "responsibility to make it stop." Quinn answered yes when asked if the situation felt to him "like a public relations nightmare."

Roughly speaking, Yosemite Lakes Park can be thought of as forming a right triangle, with one boundary running north and south, another east and west at right angles to it, and the hypotenuse joining the ends of the other two sides in a line running from the southwest to the northeast. The development has two thoroughfares: Yosemite Springs Parkway and Road 400. Yosemite Springs Parkway bisects the development in a line running diagonally from northwest to southeast. It meets Highway 41 at the southeast vertex of the development, and forms a T-junction with Road 400 at its other end, near the midpoint of the triangle's southwest-northeast hypotenuse. Road 400 roughly follows the hypotenuse, terminating at Road 415 at the northeast end and leaving the development, heading toward Hensley Lake, at the southwest end.

The four cameras installed on May 28, 2013, and June 4, 2013, were placed so that vehicles entering and exiting the development by the main routes would pass them. The camera referred to in the record as YSP East was placed at the southeast end of Yosemite Springs Parkway, near the intersection with Highway 41. The YSP West camera was at the opposite end of Yosemite Springs Parkway, at the intersection with Road 400. A camera called 400 East was installed near the intersection of Road 400 and Road 415, at the northeast corner of the development. The fourth camera, 400 West, was located on Road 400 near Best Way, some distance west of the intersection of Road 400 and Yosemite Springs Parkway.

Quinn testified that the cameras installed on May 28, 2013, and June 4, 2013, would cover all traffic entering and leaving the development, but late in the investigation, Cal Fire realized that one entrance/exit point was not monitored by a camera. This was the intersection of Road 400 and Lilley Mountain Drive, near the southwest vertex of the development's triangular shape. A driver turning to the east off of Road 400, onto Lilley Mountain Drive, would be able to reach virtually any point inside the development without passing any of the Cal Fire cameras. A camera, which we will call camera 5, was installed at this intersection on June 24 or 25, 2013, but by this time the investigation was nearly over, for the final two fires in the series (counts 30 and 31) started on June 25, 2013, and defendants were arrested on that date. Thus, for virtually the entire period of the Gold Mine Fires, it was possible for a person to drive into the development, go to nearly all locations inside, and leave again without being detected by Cal Fire's cameras. (It also was always possible, of course, for a person already inside the development, such as a resident, to drive around within it without passing any of the entrance/exit points and thus to be undetected by the cameras.)

There was a sixth camera about which evidence was presented at trial. This was a home surveillance camera belonging to a homeowner who lived on John Muir Drive within the development, Gerald Arneson. Footage from this camera was discussed in connection with three of the roadside fires, those charged in counts 10, 11, and 22.

A further limitation of the camera surveillance system Cal Fire installed was that nine of the roadside fires, out of the 21 charged, took place outside the boundaries of the development and thus could have been started by someone from outside in a vehicle that need never have passed the Cal Fire cameras even if those cameras had monitored all the entrances. These were the nine fires charged in counts 7, 9, 12, 17, and 23 through 27. Jackson was found guilty of starting six of these (counts 9 and 23 to 27), almost half of the 14 roadside fires he was convicted of starting.

2. Surveillance camera data

Cal Fire's strategy for using the surveillance cameras was not to catch anyone in the act of starting fires, but to compile data on the vehicles that passed a camera within a time frame based on the time a fire was reported, and then to study the data for vehicles that appeared multiple times. Cal Fire Captain Branden Smith was assigned to carry out this analysis along with five or six other investigators. He testified that at the beginning, the investigators planned to examine the data on the vehicles that passed for a period of one hour before and after each fire, but "[t]here was a lot of information that we had to review and the staffing was limited," so they reduced the window to a period of 20 minutes before and after each fire. Smith testified that the 20-minute period was sufficient because he believed this would be enough time for a perpetrator driving directly from the most distant fire under investigation to reach a camera or vice versa, assuming the perpetrator's route took him or her past a camera.

Smith testified that from the start of the camera surveillance until June 12, 2013, the cameras were capturing less than 60 percent of the license plate numbers of the vehicles that passed. Adjustments were made and the capture rate rose to 94 or 95 percent from June 13, 2013, onwards.

In addition to considering all the vehicles that passed by a camera within the 20-minutes-before-and-after window, the investigators focused on the actual direct driving time from a fire to a camera. For each roadside fire that occurred after the cameras were installed, the investigators determined the driving time from the fire to a camera and compared this with the time when the fire was reported and the times when vehicles passed by and the direction in which they were travelling. For each of the roadside fires in connection with which a vehicle of Jackson's was seen, all these data points for Jackson's vehicles,⁴² along with a summary of the circumstances of each fire, are collected in the table attached to this opinion as appendix A. For example, the fire charged in count 9 was reported at 12:25 p.m. on June 8, 2013. Jackson's 1998 Dodge pickup truck was photographed by the 400 East camera driving away from the direction of the fire at 12:15 p.m., again driving toward the fire's location at 12:41 p.m., and a third time headed away from the fire at 12:47 p.m. The driving distance between the fire and the camera was three minutes and 19 seconds. (See appendix A, count 9.) (The table in appendix A also shows the opinions of prosecution witnesses on the causes of the fires, which will be discussed in section C below. In most instances, a Cal Fire responder and a Cal Fire investigator gave separate opinions about whether they believed a fire was caused by arson.)

42 Jackson owned a black 1998 Dodge four-wheel-drive pickup truck with raised suspension, distinctive chrome parts, including vertical chrome exhaust pipes, and license plate No. 9584B. He also owned a 2003 Yamaha motorcycle. In testimony about the detection of Jackson's vehicles by the surveillance cameras, the motorcycle is only mentioned in connection with two fires, counts 10 and 11. Usually the vehicle detected is referred to in testimony merely as a "vehicle," but it appears it was the Dodge truck that was seen in the time windows for the other fires.
According to the People's brief, Jackson's father, Carl Jackson, testified that only Jackson drove the pickup truck, but we found no such statement on the page the People cited or elsewhere in Carl Jackson's testimony. His testimony was only that he had seen Jackson "drive it lots." Jackson himself said in his statement to Cal Fire investigators that Waterman did not like driving the truck and did not drive it "that I'm aware of." Her Toyota Sequoia was "her main ride."

The investigators did not always apply their methodology in a way that would consistently have identified all the vehicles that passed by each camera within the 40-minute or two-hour window or within the actual drive time between the fire and the camera. For example, the fire charged in count 12, which occurred in Oakhurst, 13 miles outside the development, was estimated to have started at 10:35 a.m. and was reported at 10:49 a.m. on June 9, 2013. The driving time between the fire and the YSP East camera was 16 minutes and 30 seconds, and that camera recorded Jackson's truck driving toward the development (away from the fire) at 10:58 a.m. Smith admitted he stopped reviewing the camera footage after he saw Jackson's truck, instead of watching for the entire window, and also that he did not review the footage from any of the other cameras for that fire. Thus, he would not have been aware of any other vehicle entering or leaving the development and passing the YSP East camera during the window after the time Jackson's truck passed, and would not have known of any vehicle entering or exiting the development during the window via another route. (Of course, for this fire, like all those occurring outside the development, the cameras would not have recorded a perpetrator who did not come from or go to the development at all.)

The roads monitored by the cameras were well-traveled, and Smith compiled a database of 7,576 vehicle captures for all the time windows combined over the course of the investigation. The cameras recorded Jackson's truck or motorcycle on seven dates. On some of these dates, a vehicle of Jackson's appeared in the window for more than one fire; and in the windows for some fires, Jackson's vehicles appeared more than one time. Jackson's vehicles appeared a total of 21 times during the windows for the 16 fires on these seven dates. Fourteen of these were the 14 roadside fires Jackson was convicted of starting; he was acquitted on the other two (counts 12 and 17).

For each of those 16 fires, Smith calculated what he termed an ignition window. The variables in this calculation were the time when the fire was reported, the time when Jackson's vehicle was recorded by a camera, and the driving time between the camera and the fire location, as determined by Smith himself driving directly between the two locations without stopping. The ignition window was the length of time that would have passed between Jackson being at the fire location (had he driven as Smith did) and the report of the fire being received. For example, on June 20, 2013, Jackson's truck was recorded passing the 400 East camera, heading east on Road 415, at 1:32 p.m. The fire charged in count 24 was reported at 1:46 p.m. at a point along Road 415 that Jackson must have passed (unless he stopped, waited, and turned back the way he came outside the 40-minute window). The fire charged in count 25 was reported at 1:58 p.m. at a point on Road 416 that Jackson could have passed if he made a right turn at the intersection of Road 415 and Road 416. According to the driving times measured by Smith, the count 24 location was about 30 to 45 seconds from the 400 East camera, and the count 25 location was six and a half to seven minutes from the camera. Driving as Smith did, Jackson thus would have been at the count 24 location at 1:33 p.m. and the count 25 location at 1:39 p.m., resulting in an ignition window of 13 minutes for count 24 and 20 minutes for count 25. In other words, the count 24 fire started no more than 13 minutes after Jackson would have passed by and the count 25 fire started no more than 20 minutes after Jackson would have passed by had he continued driving as Smith assumed. A summary of all the ignition windows calculated by Smith is shown in appendix B to this opinion.

The point seems never to have been explicitly made during the trial, but it was possible for the ignition window to be a negative number, meaning that, under the driving-directly assumption, the fire was reported before Jackson's vehicle passed the location (i.e., a negative amount of time passed between his hypothesized presence at the fire location and the fire being reported). If the driving time from a camera to the fire was less than the time between the camera capture and the report of the fire, and the capture took place after the report, then Jackson's vehicle necessarily passed the fire location after the fire was reported, assuming it was driven directly between the two points, so that the occupant of the vehicle could not have started the fire unless he or she failed to proceed directly to the camera after starting it. (Similarly, if the capture preceded the fire report, but did so by an amount less than the driving time between the camera and the fire location, and the vehicle was heading toward the fire location, then the vehicle could not have reached the fire location in time to start the fire, unless it managed to cover the ground faster than Smith did when measuring.) This was the situation in one case, count 13. (See appx. B.) Jackson's vehicle passed the 400 West camera at 11:50 a.m., proceeding away from the fire location. The fire was reported at 11:35 a.m. The driving time from the fire to the camera was about one minute. If the vehicle went directly from the fire location to the camera, then it was at the fire location at 11:49 a.m., 14 minutes after the fire was reported. Smith testified to these facts, but was not directly asked by the prosecutor whether this meant the driver of the vehicle did not start the fire if he or she drove directly between the two points. Instead, the prosecutor asked him about whether Jackson could have started the fire, driven home from its location, and then gone out again later and passed the camera without this being shown on another camera. (He could.) Setting this example aside, the calculations showed that a vehicle belonging to Jackson could have been at the location of each of the roadside fires between seven minutes and 25 minutes prior to the fire being reported, if driven at the speed and along the route Smith assumed. (See appx. B.)

The significance of these calculations lay, of course, in their potential for placing Jackson at the site of each fire at a time when he could have started it, consistent with the time when it was reported. Except for count 13, when the data showed Jackson's truck would not have been at the fire location until 14 minutes after the fire was reported, Smith's calculations indicated that a Jackson vehicle could have been at the site of each of the roadside fires (for which there was camera data) on the early side—seven to 25 minutes before the fire was reported.

Cal Fire Chief Matthew Gilbert testified that this timing was consistent with Jackson starting each of these fires using a time-delayed incendiary device.⁴³ Gilbert theorized that Jackson used handmade devices consisting of cigarettes with matches or match heads inserted in or added to them. Such a device could be thrown or placed on the ground with the cigarette lit; it would then flare up, potentially causing a fire, when the cigarette burned down to the matches or match heads.⁴⁴ This could allow for a delay from less than five minutes to as many as 20 minutes between the placement of the device and the ignition of the fire.⁴⁵ (I will discuss below the physical evidence relevant to Gilbert's theory of ignition by incendiary devices.) Given the likelihood of some delay between the ignition of a fire and the same fire being reported, the camera data indicated that a person in a Jackson vehicle could have been at the location of each fire at a suitable time for starting it with such a device, according to Gilbert. In fact, Gilbert's testimony suggested that because of the timing, a person driving Jackson's vehicles could not have started the roadside fires unless he used time-delayed incendiary devices. When asked whether the camera data were inconsistent with the hypothesis of a "hot start" (a fire that started in the arsonist's presence, with no time delay) for any of the roadside fires, Gilbert replied that "the data on the videos indicates that a time delayed incendiary device was used."

43 Gilbert actually opined that all 21 of the roadside fires were started by Jackson using time-delayed incendiary devices, not just the 16 for which there was camera data. But for the five roadside fires for which there was no camera data—counts 1, 2, 3, 7, and 8—there was no evidence at all regarding when Jackson might or might not have been at the fires' locations, so the issue of timing had no role in supporting the hypotheses that Jackson started those fires and that an incendiary device was used to start them. The jury found Jackson not guilty on those counts.

44 An unmodified cigarette will not ignite a vegetation fire unless certain conditions exist. Among other things, the relative humidity must not exceed a figure between 18 and 22 percent, depending on other conditions, according to one Cal Fire employee who testified.

45 Gilbert testified that he personally had made 50 to 75 cigarette incendiary devices as experiments, but none of these had ever taken as long as 20 minutes to flare up. He had read of devices that took up to 20 minutes, however

The data were consistent with the occupants of many other vehicles being at the fire locations at times when they could have started the fires with or without time-delayed devices. As indicated above, more than 7,000 vehicles were recorded by Cal Fire's cameras during the time windows before and after each fire. Cars that passed the cameras within the windows were numerous on a fire-by-fire basis as well. For example the Arneson camera, which detected Jackson's truck for the fire in count 22, photographed 31 other cars passing along Arneson's residential street within the 40-minute time window Smith reviewed for that fire. Smith testified that he did not log any of these in his database. The camera that captured Jackson's truck while monitoring Yosemite Springs Parkway, a major thoroughfare, during the window for the fire in count 23, captured over 200 other cars in addition. Many of these other cars undoubtedly could have passed the fire locations at times when their occupants could have started the fires consistent with the times the fires were reported, if the vehicles were driven in the manner assumed by Smith.

Jackson's vehicles also were not the only ones that appeared within the window for multiple fires. Smith testified that there were 110 vehicles that appeared in the window for two or more fires. For a subset of this set of 110 repeating vehicles—those vehicles for which there were "generally maybe four to five captures"—Smith said he carried out a detailed analysis to determine when the vehicles would have passed by fire sites if they had travelled certain routes. Smith indicated that he was less likely to carry out this type of analysis if, as the investigation went on, the vehicle stopped making appearances on the video. Such vehicles "automatically kind of fell off" the field of his attention and "were no longer becoming a concern to me as I analyzed the data," he said. These remarks apparently reflected an assumption that the target of the investigation was one perpetrator who was active throughout the course of the series of fires, rather than two or more perpetrators who might have been active at different times. Smith testified that other than Jackson's vehicles, there was no vehicle in the database he compiled from the camera data that could have been in position to start all the roadside fires that took place after the cameras began operating.

Apart from Jackson's vehicles, Smith testified about only one of these detailed analyses. This involved a white or silver Subaru identified by its license plate number. The Subaru appeared in the before-and-after window a total of six or seven times, and these times were within the windows for eight fires (as some of the fires occurred close together in time).⁴⁶ First, it appeared on the 400 East camera at 2:59 p.m. on June 8, 2013. Smith said this appearance was relevant to the fires charged in counts 10 and 11, which were reported on the same day at 2:57 p.m. and either 3:17 p.m. or 3:25 p.m. respectively. (Jackson's motorcycle was captured by the Arneson camera at 2:52 p.m. or 2:55 p.m. that day.)

46 The captures for the eight fires were captures in which the Subaru's license plate could be read. Smith said no when asked the following question: "And are you including other captures where you didn't catch the license plate?" This question and answer appear to mean there might have been additional fires in connection with which the Subaru appeared.

Next, the Subaru was photographed by the 400 East camera on June 9, 2013, at 11:54 a.m. This was relevant to the fire charged in count 13, which was reported that day at 11:35 a.m. (Jackson's truck appeared on the 400 West camera that day at 11:50 a.m.)

On June 12, 2013, the Subaru was captured by the YSP West camera at 2:24 p.m. This was relevant to the fire charged in count 17, which was reported at 2:51 p.m. the same day. (Jackson's truck was photographed by the YSP East camera at 2:36 p.m. that day.)

The Subaru appeared next on the 400 East camera at 12:15 p.m. on June 16, 2013. In Smith's opinion, this capture was relevant to the fire charged in count 21, even though it happened more than an hour before the fire was reported at 1:21 p.m. (Jackson's truck passed the 400 East camera at 1:08 p.m. and the YSP West camera at 1:13 p.m. that day.)

Next, the Subaru was photographed on June 20, 2013, at 1:28 p.m., by the 400 East camera. This was relevant to the fires charged in counts 24 and 25, which were reported at 1:46 p.m. and 1:58 p.m. (Jackson's truck appeared on the 400 East camera at 1:32 p.m. that day.)

Finally, the Subaru appeared on the 400 East camera at 3:33 p.m. on June 25, 2013, and Smith said this capture was relevant to the fire charged in count 30, which was reported at 3:28 p.m. that day. (Jackson's truck was photographed by the 400 West camera at 2:59 p.m. and by camera 5 at 3:10 p.m. that day.) Smith excluded the Subaru as a factor in that fire, however, because it did not appear on any other camera, as it would necessarily have done if it had proceeded from the 400 East camera to the location of the fire. Jackson was found guilty of starting all the fires to which Smith considered the Subaru captures relevant.

Smith opined that an occupant of the Subaru could have started seven of these fires (all but the one in count 30); he described the routes the Subaru would have taken from the cameras to the sites of each fire. For some of the seven fires, Smith was questioned about whether he considered it probable that an occupant of the Subaru started them, but objections were sustained to all but one of these questions. The one question that got through was whether Smith considered it probable that the Subaru was connected with the fires charged in counts 24 and 25. He said no, but was not asked for a basis for this opinion.

3. Evidence other than the camera data

At the site of the fire charged in count 17, an eyewitness told Cal Fire personnel of a white car speeding away when the fire was only the size of a manhole cover. At the site of count 24, an eyewitness saw a white Subaru driving away when the fire was small. The Subaru repeatedly sped up and slowed down, and had paper flying out of its window. Other car sightings reported by witnesses at the scenes of some of the roadside fires included a black Jetta or Corolla driving away quickly (count 8); a dark sedan and a white Hummer with a Yosemite Lakes Park Public Works sign on the door (count 11); and a burgundy sedan with out-of-state plates leaving the scene rapidly (count 12).

Although the fire in count 27 took place in Oakhurst, a considerable distance from the development, Jackson's truck was seen parked at the scene 13 minutes after the fire was reported and Cal Fire personnel saw a man standing near the truck directing traffic.

The fire in count 28 happened within the development. Cal Fire personnel were conducting surveillance near Jackson and Waterman's house. They heard Jackson's truck start up and leave the house, saw it on the road at a point between the house and the fire location, and heard it return. One of the Cal Fire investigators saw Jackson driving a short time before the fire was reported; the place that turned out to be the fire location was behind him and he was looking at his side mirror. The truck returned home ten minutes before the fire was reported. At the scene of the fire, the truck again appeared and parked; Jackson got out and spoke to Cal Fire personnel, saying he thought he saw a suspicious car.

The fire in count 29 occurred a quarter to a half mile from Jackson and Waterman's house. Cal Fire surveillance personnel stationed near the house heard Jackson's truck leave the house, drive toward the fire location, and return home, remaining within audible distance all the while. The record is unclear on exactly when the fire was reported, but it was within about 15 minutes of the time the truck returned, possibly sooner. Later, Waterman falsely told a Cal Fire investigator that Jackson never left home that evening.

Finally, prior to the fire in count 30, Cal Fire investigators placed a GPS tracking device on Jackson's truck. The fire was within the development and was reported at 3:28 p.m. The tracking device showed the truck leaving Jackson and Waterman's house at 2:53 or 2:54 p.m. It drove around the development, passing the fire location at 3:05 p.m. at 37 miles per hour. About one mile beyond that point, the truck made a U-turn and passed the fire location again at 3:08 p.m. at a similar speed. The truck continued to drive within the development, returning home at 3:21 p.m. A Cal Fire officer saw a white man driving Jackson's truck inside the development at 3:14 p.m.

Jackson was a smoker. In the driver's door of the pickup truck was a box of King Mountain Premium Red Blend cigarettes. A carton and a half of the same brand was on the back seat. Four books of paper matches and two lighters were in the center console. Thirty-five cigarette butts were in the truck bed. Two more books of paper matches and two more lighters were in the nightstand in the bedroom of the house. More books of matches and another lighter were inside a lunch box. In the kitchen there also was a used magnesium strike block—a tool used for starting campfires—although the prosecution never suggested such a tool was used to start any of the roadside fires, and instead theorized that all were started with time-delayed incendiary devices.⁴⁷

47 Gilbert testified that, in his opinion, the magnesium strike block had nothing to do with the charged fires.

The truck contained a package that had held 40 small lead fishing weights. Twelve of the weights remained in the package. No other fishing gear was found in the truck. Cal Fire Captain David LaClair, one of the investigators, opined that the fishing weights could be incorporated into cigarette incendiary devices to add weight.⁴⁸ No cut or altered matches or cigarettes and no loose tobacco were found.

48 Gilbert testified that he did not know whether the fishing weights had anything to do with the charged fires, and confirmed that no fishing weights were found at any of the fire sites.

Jackson was employed part-time by a Cal Fire contractor, involved in hauling equipment to and from the sites of fires. In the living room were a radio scanner and a notebook listing radio frequencies, some of which were frequencies used by firefighters. There also was a reference book called an "incident response pocketbook," intended for use by firefighters and others working at fire scenes. Patrick Denis, Jackson's employer, testified that he and Jackson had received copies of this book at a training class conducted by Cal Fire. In the truck there was a Cal Fire document dated June 13, 2013, called an IAP or incident action plan for a fire designated "Revis." This was a briefing document, copies of which were given to each person assigned to work on the fire. (The fires named Revis A and Revis B, both assigned incident No. 17 by Cal Fire and charged as counts 18 and 19 in the information, took place on June 12, 2013, as indicated in the table above summarizing the information. These were among the fires clustered around Jackson and Waterman's house. Revis B occurred at Jackson and Waterman's address, 43114 Revis Circle East, and Revis A took place next door at 43117 Revis Circle East. Despite the date on the IAP, it appears to have been issued for the Revis fires charged in this case, as there was testimony from a Cal Fire witness that the IAP was for one of the charged fires.)

A computer in the house had a screen saver showing a picture of a forest fire. The Cal Fire employee who testified about finding this said he had seen the same screen saver on other computers before; he described it as "a popular picture."

A mechanic inspected Jackson's truck. He found it was in good working order and not emitting any sparks.

After he was arrested on June 25, 2013, Jackson was interviewed by a Cal Fire officer and made statements. The officer encouraged Jackson to admit he had lit many fires. The officer's interrogation techniques included falsely claiming Jackson's truck had been photographed by 13 cameras, that a GPS tracker had shown him driving to 13 fire sites, and that other officers had seen him throwing things out of the window of his truck. The officer said he knew Jackson lit all the fires and he knew Jackson was lying when he said he did not know how the fires started. Jackson only admitted carelessly throwing a cigarette out of his truck one time, after smoking. This seems to have been in reference to the last roadside fire, the one charged in count 25, which took place the day of the arrest.

B. Evidence of the identity of defendants as those who set the fires clustered near their house

The information charged Jackson and Waterman with setting 10 fires clustered in an area behind their house. These took place on May 18, 2013 (count 4), May 22, 2013 (counts 5 and 6), June 9 and 10, 2013 (counts 14, 15 and 16), June 12, 2013 (counts 18 and 19), June 14, 2013 (count 20), and finally June 25, 2013 (count 31), when defendants were arrested. Except for count 4, either Jackson, Waterman or both were found guilty of starting these fires. The facts on each fire, including expert opinions on causation, are summarized in the table in appendix C.

Cal Fire initiated undercover surveillance in the nature area behind defendants' house on June 11, 2013, after several fires occurred in that area—those charged in counts 4, 5, 6, 14, 15, and 16. This surveillance thus was in operation at the time of the cluster fires charged in counts 18, 19, 20, and 31. LaClair and two other officers positioned themselves in the nature area, wearing ghillie suits. These were camouflage outfits with artificial foliage attached, allowing the officers to blend in with the brush. The officers conducted surveillance daily from 3:00 p.m. to 9:00 p.m. Local firefighters and law enforcement were not informed of this surveillance operation (or the other aspects of Cal Fire's surveillance).

Despite the surveillance efforts, the facts developed by the investigation and presented at trial are remarkable for the paucity of material that incriminates Jackson or Waterman. Roughly speaking, the facts of most of the counts can be described as follows: The fires happened in the wilderness area behind defendants' house and a vacant neighboring house. One or both defendants were at home, and in some instances were seen outdoors around the time of the fires. A Cal Fire investigator concluded the fires were arson because they could find no innocent or accidental causes, although a Cal Fire responder often found there were innocent or accidental causes that could not be ruled out. And no other specific potential culprits were identified.

On May 22, 2013, the date of the fires in counts 5 and 6, responders arrived and found Jackson trying to extinguish the fire in count 5 with a hose. The fire in count 6 had started around the same time, about 10 feet away. Waterman was in Fresno playing in a soccer match. A Cal Fire witness suggested that neighboring dogs might have barked if the fires had been started by a stranger. A juvenile started a fire with a lighter about half a mile away, about two hours earlier. Jackson was convicted.

On the evening of June 9, 2013, the fire in count 14 was reported and extinguished in the wilderness area behind defendants' house. A Cal Fire investigator spoke to Waterman the next day. The investigator noticed what he described as burned shrubbery along the walkway leading to defendants' front door. A picture displayed to the jury showed a bush, a portion of which lacked foliage, but did not appear blackened. The investigator testified that Waterman said she noticed the spot and said no when he asked if she reported it. He did not investigate the purportedly burned bush and formed no opinion about how it got that way, but he thought it was odd. Waterman testified that she did not remember the discussion and the bush merely had spots where the foliage had died off; it was not burned. Waterman and Jackson were both convicted.

On June 10, 2013, the fires in counts 15 and 16 started around the same time, about 50 feet apart, in the wilderness area behind the house next to defendants' house. Waterman was standing nearby as firefighters worked on the fire. A Cal Fire responder testified that Waterman said she saw smoke in the grass and then flames where the first fire started. A Cal Fire investigator testified that Waterman reported that she first saw the fire when it was the size of a roll of duct tape. She said she heard a bicycle on the driveway of the neighboring house and also heard a car driving away. Other eyewitnesses confirmed a teenager on a bicycle and a speeding white pickup truck. The Cal Fire investigator opined that usually, if a person sees a fire when it is very small, he or she was present when it started. Waterman and Jackson were both convicted.

On June 12, 2013, in the nature area, about 200 yards apart, the fire in count 18 was reported at 4:28 p.m. and the fire in count 19 was reported about 9:40 p.m. A game camera set up by a Yosemite Lakes Park security guard showed Waterman's car coming home at 4:09 p.m. and then leaving the development some hours before the second fire was reported. She said she left for a soccer match in Fresno around 5:30 p.m. Jackson and his father were in the house when the first fire started. Waterman testified that she returned home around 10:45 p.m. and found Jackson in bed asleep. A camera showed two other cars entering and immediately leaving the dead-end street on which defendants' house was located, one around 4:15 p.m. and one around 4:30 p.m. Waterman was found guilty of starting the fire in count 18 and Jackson was found guilty of starting the fire in count 19.

An instance in which the evidence against Waterman might be viewed as going beyond her mere presence in or near her home when a fire started is count 20, in which Waterman was convicted and Jackson acquitted. As will be seen, however, the key incriminating testimony there involved what seems a near impossibility: Waterman's appearance first as a 180-pound person, who walked with a waddle and was allegedly seen where the fire started, and again minutes later in different clothes as the 110- to 140-pound soccer player she actually was, according to the account of the one witness who claimed he saw her.

LaClair testified that he was wearing his ghillie suit, conducting surveillance with two other officers in the nature area behind the vacant house adjacent to defendants' house around 7:00 p.m. on June 14, 2013, when he saw a person walking in the direction of the vacant house. The person was white, but LaClair could not determine his or her gender. LaClair estimated the person's weight as 180 pounds. He or she had black hair, and was wearing black sunglasses, blue shorts, and a white shirt. The person walked with a waddle, with arms held away from the body. LaClair transmitted this description to his fellow officers by text message, but did not think of using his phone to take a picture. The terrain was sloped, and the person walked into an area hidden from LaClair's view. After a few minutes, the person reappeared and walked away from the vacant house. LaClair then heard a crackling noise. He walked into the area where the person had been out of sight for a few minutes and discovered the fire. Instead of trying to track the person, he went toward the fire to see if he could find an incendiary device.

As LaClair was looking at the fire, a woman came up behind him, yelling, "Fire, fire!" The woman was smaller than the person LaClair had seen going into the fire area and was dressed all in brown, wearing what LaClair described as a completely different outfit from that of the person he saw before. LaClair believed she was a different person. Not knowing he was a Cal Fire officer, the woman demanded to know what he was doing there. He yelled at her to "get the fuck away" from him and then ran away, distressed that his cover had been blown. He called for and was picked up by a Cal Fire vehicle (a rented white Jeep Liberty SUV) and left the area.

Waterman testified that it was she who confronted LaClair near the fire location on the evening of June 14, 2013. She was in her house with Jackson, having dinner, when she heard tires squealing. She went out to investigate and saw a man in a ghillie suit, who she later learned was LaClair, standing next to the fire. She argued with him. He cursed at her, she chased him, and he fled.

Waterman testified that her weight at the time was no more than 140 pounds. LaClair estimated that Waterman weighed 110 to 120 pounds. (In her pre-arrest statement to LaClair, Waterman said she had a "[d]istinct walk" and was "bow-legged.")

At some point after this encounter, LaClair learned that the woman with whom he had argued was Waterman. He still did not believe, however, that Waterman was the 180-pound person who apparently started the fire.

On June 16, 2013, LaClair again saw Waterman while he was conducting surveillance. An officer accompanying him took pictures of her, which were shown to the jury. She was wearing sunglasses, a blue shirt, and blue shorts. At trial, LaClair asserted that the shorts and sunglasses in the June 16, 2013 pictures were the same as he had seen on the 180-pound person on June 14, who also was wearing similar shoes. He also thought the way she was holding her arms, her walk and her hair color in the pictures was similar to the 180-pound person. At the time the pictures were taken, however, he still did not connect Waterman with that 180-pound person.

According to LaClair, when he saw Waterman on June 16, 2013, she was walking in the area behind the vacant house when she picked up a stick and began beating some tall grass with it. LaClair testified that he supposed she must be looking for camouflaged people, since she had seen him in his ghillie suit on June 14, 2013. In her testimony, Waterman admitted she was hitting the ground with a stick and was looking for people, but denied she had LaClair or a man in a ghillie suit in mind, and said she was not actually trying to flush out concealed people by striking them with the stick.

It was on June 25, 2013, the day Jackson and Waterman were arrested, that LaClair finally came to believe that the person who appeared to set the June 14, 2013 fire, who did not look like Waterman and was wearing clothes different from those in which Waterman appeared shortly thereafter, must really have been her after all. He testified that, at the time of trial, there was no doubt in his mind about this.

While still outside her house on June 25, 2013, just before she was arrested, Waterman made statements to LaClair that were recorded and transcribed. She repeatedly denied that she had started any fires. She also repeatedly contradicted LaClair's assertions that she had been at the fire location on June 14, 2013, just before the fire started; she said that if LaClair saw someone before she came out shouting about the fire, it was not her. When LaClair continued to repeat that it had been her, Waterman said she might have been in the location at some earlier point checking for fires, but she found nothing and went back inside and ate dinner before coming out again. She did not go to the fire location, leave, and then come back right away shouting, as LaClair claimed. Below are the passages in which Waterman gave this account. (The June 14, 2013 fire, which was charged in count 20, is referred to here as fire No. 8.)

"[LaClair]: But then you were tryin' a tell me you weren't there.

"[Waterman]: Maybe I was.

"[LaClair]: I know you were there.

"[Waterman]: I don't—you know what? I come up here so many times, I don't keep count. Next time when—when you let me go tonight, I will start documenting when I came up and said, no fire."

"[LaClair]: But. . . you already admitted now that you were up there and you came back...

"[Waterman]: Because I hear stuff.

"[LaClair]: ...down. You're—you're—hear stuff.

"[Waterman]: I wasn't—I coulda been up here. I come up here all the time (unintelligible)."

"[LaClair]: [N]ow calm down. Now, fire number 8, you walked over to the playground [i.e., play equipment at the vacant house]. You heard something. You walked over there and then...

"[Waterman]: I wah...

"[LaClair]: ...that particular day, you walked back out without goin' around the house.

"[Waterman]: Because I thought, okay, there's nothing. I heard something. I checked. I go back in—I go back in to eat dinner.

"[LaClair]: And how far did you get when you realized there was a fire?

"[Waterman]: There was nothing. I went back to my house. We heard squealing in the cul-de-sac. We come—I opened the garage door, come runnin' out here. I get to where about my truck is and hit the driveway. I go up. I yell at (Ken) to call 911 because I hear the crackling. I come around here and obviously now I know there you were when I hit about that tree up there.

"[LaClair]: Yeah. Y...

"[Waterman]: And then you cursed at me and then I followed you. And then you told the per—your partner, 'We gotta get the fuck outta here right now.' And then you started running and I started running. 'Kay, I did not start any of these fires."

At trial, Waterman denied she had been in the fire area in the minutes before the fire started and said she did not start the fire. She acknowledged that, in the pre-arrest interview, she had said she might have been in that area before the fire. She again said she meant she had probably been patrolling around the area earlier, not just before the fire started.

June 25, 2013, was also the date of the last of the fires, that charged in count 31. One of the surveillance officers testified that, over the course of the surveillance period, he often saw Waterman walking from her house to the vacant house and back. At around 6:40 p.m. on June 25, 2013, the officer saw Waterman walking toward the vacant house, as he had seen her doing before. Behind her, from a location near her house, a column of smoke rose where the fire had started. The officer saw her twice turn and look in the direction of her house and then turn back and continue walking toward the vacant house. Then the officer heard sirens of fire trucks approaching. At about the same time, Waterman turned back and began running toward her house, while yelling, "Fire, fire! Can't believe the arsonist lit another fucking fire! This is bullshit!"

When responders arrived at the fire, they encountered Jackson near it holding a hose. At first, there was no water coming out of the hose. Then water did come out. A firefighter testified that he suspected Jackson had started the fire and was only pretending to be trying to put it out. Jackson and Waterman were both convicted of starting the fire.

Jackson and Waterman were arrested for the whole series of fires as the final fire was being fought. Officers planning to arrest them arrived at the same time as the firefighters.

Waterman's testimony at trial included a statement that she did not start any fires. She also testified that she did not agree with Jackson to start any fires. In response to questioning by the prosecutor, Waterman confirmed that her statement to LaClair on the night of her arrest included an assertion that she was not an arsonist, but she did not know about Jackson. Waterman told LaClair she prayed it was not Jackson and said she would divorce him if it was.

Waterman testified that she and Jackson were married in 2002, and had lived in their home since 2005. They declared bankruptcy in 2010, but did not lose the house and were not, at the time of the fires, having trouble paying their bills. But Patrick Denis, the Cal Fire contractor for whom Jackson worked hauling fire equipment, testified that Jackson said he had not made any mortgage payments for a couple of years. Waterman testified that she was employed part-time by the Central California Autism Center, caring for an autistic child. She had been doing this work for two and a half years. She testified that, in addition to his work for Cal Fire, Jackson was self-employed driving a semi truck and also did maintenance work with his father at camps. A search of Waterman's Toyota found an unopened magnesium strike block and a box of wooden matches. The box had contained 250 matches and 65 were missing.

The jury was shown a video recording of a brief interview Waterman gave to a television reporter on June 10, 2013. The fires charged in counts 15 and 16 were burning at the time of the interview, which was given on the scene. Waterman told the reporter she was angry and freaked out about the fires. She said there had been several near her house. The reporter said neighbors had suggested the fires were being set intentionally; she asked if Waterman had heard that or seen anyone suspicious. Waterman replied that she had only heard rumors.

The People's appellate brief suggests that all suspicious fire activity for 2013 in the relevant area stopped after defendants were arrested, but the record does not actually establish this. Testifying on this topic, Chief Quinn spoke about arson fires and fires with undetermined causes that took place in the Battalion 5 area. He was not asked what the boundaries of that area were and said only that "most" of the charged fires were within it. Two fires in Oakhurst, for instance—those charged in counts 12 and 27—were in the Battalion 4 area, about which no data were presented regarding fires after the arrest date. Quinn conceded that even within the Battalion 5 area, there were "some" arson fires and "some" fires with undetermined causes in 2013, after defendants' arrests. He said these were not in "the Yosemite Lakes Park area," but admitted at least one was in Coarsegold. Quinn did not testify about whether there were any suspicious fires after defendants' arrests outside the Battalion 5 area but still within the area encompassed by the series of charged fires, which ranged as far from the development as Oakhurst and Raymond. Quinn's testimony does, however, support the proposition that there were no more arson or undetermined-cause fires within the Yosemite Lakes Park development itself in 2013 after the arrests.

Two other witnesses gave testimony on the topic of whether the fires stopped after defendants were arrested. Patrick Denis, who lived in Coarsegold, was allowed to testify to his "general impression" that "[t]he fires stopped" in the area "[a]round Yosemite Lakes Park," but this testimony came after Denis had already twice said he did not know. (He said, "I believe they stopped or, you know, I don't know exactly," and "I don't know for sure. I mean, you know, I mean, seemed like the fires stopped.") Jack Valpey, who lived a quarter of a mile away from defendants, was asked what happened "with regard to the fires in your area" after defendants were arrested. He said, "No more fires." He was not asked what he regarded as being included in his area, but it seems clear from the context that he took the question to be referring to his own immediate neighborhood.

C. Evidence that arson caused the fires

At trial, virtually no direct evidence was presented that any of the charged fires were caused by arson.⁴⁹ The primary technique used by prosecution witnesses to show that the fires were caused by arson, and not some other cause, is referred to in the record as exclusion analysis or negative corpus analysis.⁵⁰ As will be seen, this technique relies on a lack of indicators of the presence of certain innocent causes from a standardized list (or affirmative indicators of the absence of those causes), plus an inference that the absence of another cause means there must have been arson. This approach was not used exclusively, however.

49 "A person is guilty of arson when he or she willfully and maliciously sets fire to or burns or causes to be burned or who aids, counsels, or procures the burning of, any structure, forest land, or property." (Pen. Code, § 451.) The prosecution must prove "a general intent to willfully commit the act of setting on fire under such circumstances that the direct, natural, and highly probable consequences would be the burning of the relevant structure or property." (In re V.V. (2011) 51 Cal.4th 1020, 1029.)

50 Gilbert answered yes when asked: "Isn't it true that negative corpus is the process of determining the ignition source of a fire by eliminating all possible ignition sources known or believed to have been present near an origin and then claiming such methodology is proof of an ignition source for which there is no supporting evidence of its existence?"

1. Exclusion analysis

While testifying about the first fire (chronologically) in the series (count 2), Cal Fire Apparatus Engineer Neil Tito explained that fire personnel use exclusion analysis to try to determine the causes of fires. They consider a set of possible causes, such as power lines, smoking, and vehicles, and eliminate these one after another if the circumstances indicate that they are impossible or unlikely. For example, Tito excluded smoking, even though he found cigarette butts at the scene of the fire, because the humidity was too high at the time for a cigarette to ignite the grass that fueled the fire. He eliminated lightning because the day was cloudless and there had been no thunderstorms. Sparks from trains were excluded because there was no railroad nearby. After going through an analysis like this for several other potential causes,⁵¹ Tito was left with people playing with fire and arson as the only unexcluded possibilities. A firework was found on the ground in the area where the fire started, and Tito concluded the firework started the fire, either intentionally or accidentally.

51 According to Cal Fire witnesses, 13 causes were considered in the standard Cal Fire exclusion analysis. These were (1) smoking; (2) lightning; (3) spontaneous combustion; (4) campfires; (5) debris burning; (6) vehicles; (7) equipment use; (8) railroads; (9) fireworks; (10) children playing with fire; (11) power lines; (12) glass refraction; and (13) arson.

Tito explained that after a cause analysis of this kind is done by a responder like himself, it is done a second time by an investigating fire prevention officer with more training. In the case of the fire in count 2, this second analysis was carried out by Cal Fire Captain Tim McCann. McCann agreed with the conclusion that the firework ignited the fire. In his view, however, arson was the only cause that could not be excluded, so he opined that an arsonist used the firework to start the fire.⁵²

52 It was never made clear in McCann's testimony what the basis was for his opinion that the fire was not started with the firework by accident. The only factor he mentioned as a basis for distinguishing an intentionally set fire from an accidentally set one where a cause like fireworks is involved was that when people set fires accidentally, they sometimes admit it. But the fact that no one came forward to confess was obviously an insufficient basis for making the distinction in this instance. There was no conviction on this count.

The results of the responders' and investigators' exclusion analyses for all the fires are shown in appendices A and C. For the 23 counts that led to convictions, the analysis results fall into several categories. In the largest category, including 14 of the counts, the investigator ruled out all causes except arson, but the responder believed arson plus one or more other causes could not be ruled out.⁵³ For four counts, the responder and investigator both testified that arson and one or more other causes could not be ruled out.⁵⁴ In three instances, the investigator testified that arson was the only cause not ruled out; no responder testified to a causation analysis in those instances.⁵⁵ Only in the remaining two counts did a responder and an investigator both testify that arson was the only unexcluded cause.⁵⁶ As shown in the appendices, sometimes the conclusion that a fire was arson was stated affirmatively, while other times the witness stated only that arson was the only cause not ruled out.

53 These were counts 9, 10, 11, 21, 22, 23, and 27 as shown in appendix A, and counts 5, 6, 15, 16, 18, 19, and 20 as shown in appendix C. On count 9, the investigator testified inconsistently that the fire was caused by arson and that he could not rule out smoking as a cause. (Appx. A, pp. 4-5.) On count 22, a responder and two investigators all agreed that the fire was caused by a piece of a catalytic converter falling from a vehicle; but after defendants' arrest, one of the investigators reconsidered, decided the catalytic converter fragment was not the cause, and determined the fire was arson. (Appx. A, pp. 9-10.)

54 Counts 13, 24, 25, and 26 as shown in appendix A. On count 24, however, the responder testified inconsistently, stating both that arson was the cause and that he had not ruled out vehicles as a cause. (Appx. A, p. 11.)

55 Counts 29 and 30 as shown in appendix A, and count 31 as shown in appendix C.

56 Count 28 as shown in appendix A and count 14 as shown in appendix C.

With a few exceptions, the Cal Fire witnesses were not asked to go through all of the 13 causes and explain why each was or was not excluded. Instead, they focused on a few possible causes. For the causes that were discussed, the reasoning in support of exclusion was sometimes weak, at least for purposes of a criminal trial. Testifying about count 28—one of the roadside fires—Cal Fire Apparatus Engineer Joseph Quistorff testified that he excluded vehicles as a possible cause although the fire occurred beside a heavily-travelled road; but he also answered yes when asked if "there was nothing particular about this fire that would allow you to exclude vehicles." Cal Fire Chief Mark Pimentel, testifying about the same fire, said he excluded vehicles as a cause because "[u]sually when we have a vehicle failure there is multiple fires going down the road. People usually don't realize that they are causing fires until a later period in time." Yet on cross-examination, he testified that a vehicle dragging an object and causing sparks might cause only one fire. "It's very possible," he said.

Cal Fire Chief Matthew Gilbert, testifying about the fire charged in count 31, said he excluded smoking as a possible cause because, at 40 percent, the relative humidity was too high for a discarded cigarette to have started the fire. The humidity was measured at the scene at 7:58 p.m., after the fire had been extinguished with water. The fire, however, was reported at 6:43 p.m., and Gilbert conceded that night falling and the application of water in the interim would have caused the humidity to increase.

In discussing three fires, those charged in counts 15, 16, and 18, Pimentel testified that the specific origin areas—the portions of the burned areas where each fire began—had been obliterated by firefighting activities before he could identify and examine them. He conceded that this could affect his ability to find any evidence of the cause of the fires that might have been present. In one instance—count 16—he added that it was dark by the time he investigated, so he could not see well. In spite of these problems, Pimentel concluded that the fires were arson by way of exclusion analysis. For instance, he excluded smoking as a cause because he found no cigarette butts in the origin area of the fire in count 18. He indicated that his confidence in his conclusions was not diminished by the destruction of the origin areas.

On counts 18 and 19 (two side-by-side fires near defendants' house, given the single Cal Fire incident No. 17), Pimentel testified that he excluded all of the 13 possible causes except arson. Asked why, he said only, "Being the nature, where it was located." When asked about some of the specific other possible causes (smoking, playing with fire), he indicated he relied on an absence of evidence at the site or in the area to support these. He added, however, that statements he received in witness interviews confirmed his conclusion. What he learned from the interviews was that two unknown vehicles, a white Chevrolet pickup and a white Ford SUV, were seen on defendants' street close to the time of the fires. The Chevrolet pickup was seen leaving the area quickly by a neighbor. Pimentel said this was suspicious and pointed to arson. Yet when the name of the registered owner of the pickup was found, Pimentel did not follow up, because this was "the only incident where that vehicle was found." Thus, the primary reason for believing the fire was arson was rejected in the end, but the suspicion that had been raised was transferred to defendants anyway. (At the end of his testimony, Pimentel did add that the position of the fire away from the road was another indicator of arson, because it suggested to him that someone took the trouble to walk into the brush to light it.)

2. Other considerations

The Cal Fire witnesses often did not rely solely on the exclusion of other causes to support their arson findings, but relied on other considerations as well. These additional considerations, however, often failed to supply the explanatory power the exclusion analysis lacked.

a. Prior fires

In several instances, witnesses testified that in reaching their arson findings, they relied on the fact that, by the time of the fire in question, Cal Fire believed that the previous fires in the series were arson. In connection with count 14, Tito said he concluded the fire was arson because there had been "a numerous amount of fires" in the area near defendants' house and they were all "like after lunch, during the hotter portions of the day." In reality, three of the cluster fires had taken place before the fire in count 14: counts 4, 5, and 6. The fire in count 4 was reported at 7:30 p.m. and those in counts 5 and 6 (which happened around the same time a few feet apart) around 8:30 p.m. McCann also said he relied on prior fires in finding count 14 to be arson.

On counts 15 and 16, Cal Fire Captain Joseph Felix believed the fires were arson because he had excluded the other causes on the list and also "due to the fact that the area has been plagued with several suspicious fires within the previous time or month."

Cal Fire Captain Tim McCann, discussing the fire charged in count 21, said he believed the fire was arson partly because of "similarities regarding recent fire activity in the immediate area." When asked on cross-examination whether his conclusion would have been different if he had not known of the earlier fires, he replied that he could not answer this hypothetical question.

On count 28, Quistorff said he found the fire was arson "because of the general fires in the area, same time of day, similar times of day." The fire in count 28 was reported at 3:31 p.m. and the prior fires in the series had been reported at many different times of day, from around 10:30 a.m. to around 9:30 p.m. Pimentel also testified that the prior fires were circumstances supporting his finding that the fire in count 28 was arson. Gilbert testified that he considered the prior fires in concluding that the fire in count 30 was arson.

b. Patterns

Two other methods of analyzing circumstantial evidence that the prosecution relied on to show causation were "geospatial analysis" and "time and date analysis," which involved plotting fires on a map and a time and date chart, and looking for patterns. Explaining the geospatial analysis he performed, Gilbert testified that the roadside fires formed a "linear pattern" or a set of linear patterns, in that they formed a series of points along roads. For instance, six of the charged fires took place along Road 400 (counts 1, 8, 13, 21, 28, and 30) and four were along Highway 41 (counts 23, 25, 26, and 27). A subset of the fires on Road 400 (counts 13, 21, and 28) also formed a "multiple reoccurring series," according to Gilbert, because each was not far from the intersection of Road 400 and Yosemite Springs Parkway. A common feature of all these fires was that they were close to the road, so that a person in a car could have started them without getting out and walking into the brush.

Gilbert further testified that the fires near defendants' house formed a "cluster of fires." All these were away from the road and could not have been started from inside or beside a car. Two fires on John Muir Drive (counts 11 and 22) and two in the general vicinity of the intersection of Yosemite Springs Parkway and Highway 41 (counts 2 and 17) were also "clusters."

Gilbert's time and date analysis indicated that a "majority" of the roadside fires were reported between 10:00 a.m. and 3:00 p.m. on weekends. The cluster fires "generally" occurred between 4:00 p.m. and 9:00 p.m.

To put Gilbert's analysis more simply: Some of the fires were along the area's roads and close to the roadside; some of them were close together, including those near defendants' house; the ones near defendants' house were not close to the road; and some of the fires took place within broadly defined time parameters.

Gilbert concluded from his analysis that there were two groups of linked fires: the roadside fires and the cluster fires. The links between the fires in each group tended to point to a single person as the cause and thus supported the view that the fires were caused by arson. Gilbert thought the two patterns suggested two arsonists.

c. Cigarette butts and incendiary devices

Gilbert also testified about cigarette butts. He described cigarette butts found at the sites of five of the roadside fires (counts 1, 8, 9, 12, and 17). He described the butts from four of the fires (counts 1, 8, 9, and 12) as being the same color (orangish brown) as Jackson's brand. (Of these fires, a conviction was obtained only on count 9.) Gilbert considered this to be a link connecting those four fires. The remaining butt, from the fire in count 17 (also no conviction) was white.

The prosecution presented a small quantity of physical evidence from the fire sites to support the proposition that the fires were started with time-delayed incendiary devices. Cal Fire witnesses mentioned cigarette butts found on the ground at the sites of seven of the roadside fires, pointing out in some instances that cigarette butts can be remnants of incendiary devices made from cigarettes: counts 1, 8, 9, 12, 13, 17, and 25. Gilbert, as we have just mentioned, testified about his examination of the butts found at five of these, which he said were the only ones he considered to be significant in the case. He referred to these when asked "how many incendiary devices" he found, but he only gave reasons for thinking one was an incendiary device and not just a cigarette butt.⁵⁷ This was the one found at the site of the fire in count 12, in which he said he found the remains of a match head. (There was no conviction on this count.) He was 90 to 95 percent sure it was the remains of a match head. The jury was shown a photograph of the cigarette butt and Gilbert pointed to the portion he thought showed the match head fragment. The photograph was taken when Gilbert found the fragment; there was no photograph from the time of his earlier examination. Gilbert testified that he first noticed the match head fragment in February 2013, while the trial was underway, and had missed it when he first examined the cigarette butt six months earlier. (This finding was the subject of the late-discovery violation; the prosecution failed to disclose the match head fragment to the defense until later in the trial.) Gilbert thought the cigarette butt had deteriorated while the box containing it had been moved around in the interim, exposing the match head fragment.⁵⁸

57 Gilbert also thought a firework found at the site of the fire charged in count 2 was an incendiary device. This opinion was based on McCann's report stating that the firework appeared to have been thrown into the brush, not ignited on the pavement. No conviction was obtained on that count.

58 During a hearing on chain of custody, Gilbert admitted that, between these two examinations, the cigarette butt was not kept in a sealed evidence box. During that time, 13 people had access to the evidence.

None of the other cigarette butts had match remnants in them. Two had what Gilbert described as a horseshoe-shaped burn pattern, but he did not explain what significance this pattern might have.

Three Cal Fire witnesses testified about experiments they performed with cigarette incendiary devices they had constructed themselves. Quinn testified about cigarette butts found at the sites of the fires in counts 1 and 8, on the sides of which burn marks appeared. (These possibly were the cigarette butts with a horseshoe-shaped burn mark about which Gilbert testified.) Quinn hypothesized that the butts were the remains of incendiary devices and the burn marks were caused by match heads inserted into the cigarettes. He prepared an incendiary device made from a cigarette of Jackson's brand and some match heads, and lit it to see if the same burn pattern would appear. The experiment failed and Quinn was unable to reproduce the burn pattern. He nevertheless was permitted to testify to his opinion that the burn patterns on the cigarette butts found at the fire sites were caused by match heads and the fires were started with incendiary devices. Counts 1 and 8 resulted in acquittals.

Gilbert testified about an attempt to make a cigarette incendiary device that would be consumed in a fire and leave no material behind. He participated in an experiment in which a device was made by inserting match heads into a cigarette from which the filter had been removed. When this device was ignited, it burned and left ash. The ash was not recognizable as a remnant of an incendiary device.

Finally, LaClair testified about an experiment with cigarette incendiary devices designed to test ignition times. He and other Cal Fire personnel made three trials of an experiment in which they threw a cigarette incendiary device from a moving car into some dry grass, and then waited to see how long it took for the device to flare up and set the grass on fire. The times recorded were four minutes 40 seconds, six minutes, and seven minutes. As indicated in appendix B, the ignition windows for roadside fires of which Jackson was convicted ranged from nine to 25 minutes.

Despite the meagerness of this evidence, the prosecution chose to found its entire case as to the roadside fires on a theory of time-delayed incendiary devices. As discussed above, Gilbert opined that they all started that way. As his reason for accepting this theory of causation, he stated that it conformed to the surveillance camera data on Jackson's—i.e., that assuming Jackson started the fires from his truck, he would have had to do it using time-delayed devices, given the hypothesized times when his truck would have passed the fire sites.

D. Additional defense evidence on the fires

1. Defense expert

Defendants' expert, Steven Carman, reviewed Cal Fire's documents and testified about deficiencies he perceived in the investigation and specifically the basis for the investigators' conclusions that the fires were caused by arson.

Regarding the negative corpus or exclusion method of determining the causes of fires, Carman testified that the topic had been a subject of controversy among fire investigation professionals. He did not opine that the method is inappropriate in general. It can be valid if used carefully as a means of testing a hypothesis regarding the cause of a fire; but it also can be abused. He suggested that exclusion analysis was a term describing the proper use of the technique and negative corpus its improper use.

Explaining the improper use of the method, Carman emphasized that a mere lack of evidence of causes other than arson was insufficient to show arson. He said fire investigators at one time erroneously believed this was a reliable indicator of arson. He gave a hypothetical example:

"As an example, we would go into a fire, let's say, and we'd look around for a cause of the fire in a house and we didn't see anything obvious. We didn't see an electrical failure in an item. We didn't see a pattern that might indicate that flammable liquids had spilled and ignited. We didn't see something else, anything that we could put our finger on.

"And so what we would do was then take the next step and say, well, it isn't A, B, C or D, so therefore it has to be what's left. And most often we called that arson. We would say, we haven't found an accidental cause which makes any sense, so it has to be arson. That's the only thing left.

"And that is a hypothesis based on the lack of information. There is no information in that case, that instance, that there is arson, but in your mind you have eliminated everything else, so therefore, that's what it has to be."

Over about the last six to ten years, however, this approach came to be viewed as wrong, Carman said. A publication called NFPA 921 (publication No. 921 published by the National Fire Protection Association, also known as Guide to Fire and Explosion Investigation, a standard work in the field, first published in 1992 and updated about every three years), took the position that the exclusion method cannot properly be used in this way: a mere lack of evidence of other causes does not alone suffice to show that a fire was caused by arson.⁵⁹ Carman stated, "[T]he idea is, if you have no evidence to form a hypothesis, then you can't adequately test that hypothesis. You are unable to test it and for that reason it's considered improper."

59 Gilbert also testified about NFPA 921 and its discussion of negative corpus. He stated that NFPA 921 is a "guide" for conducting fire investigations, but not "a policy or a hard line directive." He also testified that the 2014 edition of the volume included pertinent changes compared with the previous 2011 edition.
We have not found a copy of the relevant portions of NFPA 921 in the appellate record, but the People's appellate brief cites them and quotes them in part, relying on portions the People consider favorable. The relevant text is section 18.6.5 of the 2011 edition and section 19.6.5 of the 2014 edition. The 2011 edition rejects the negative corpus approach in unequivocal language. The 2014 edition continues to reject negative corpus, but also explains how a process of elimination can properly be used in this context. I reproduce both sections here for reference.
"Inappropriate Use of the Process of Elimination. The process of determining the ignition source for a fire, by eliminating all ignition sources found, known, or believed to have been present in the area of origin, and then claiming such methodology is proof of an ignition source for which there is no evidence of its existence, is referred to by some investigators as 'negative corpus.' Negative corpus has typically been used in classifying fires as incendiary, although the process has also been used to characterize fires classified as accidental. This process is not consistent with the Scientific Method, is inappropriate, and should not be used because it generates un-testable hypotheses, and may result in incorrect determinations of the ignition source and first fuel ignited. Any hypothesis formulated for the causal factors (e.g., first fuel, ignition source, and ignition sequence), must be based on facts. Those facts are derived from evidence, observations, calculations, experiments, and the laws of science. Speculative information cannot be included in the analysis." (NFPA 921 (2011 ed.), § 18.6.5.)
"Appropriate Use. The process of elimination is an integral part of the scientific method. Alternative hypotheses should be considered and challenged against the facts. Elimination of a testable hypothesis by disproving the hypothesis with reliable evidence is a fundamental part of the scientific method. However, the process of elimination can be used inappropriately. The process of determining the ignition source for a fire, by eliminating all ignition sources found, known, or believed to have been present in the area of origin, and then claiming such methodology is proof of an ignition source for which there is no supporting evidence of its existence, is referred to by some investigators as negative corpus. Negative corpus has typically been used in classifying fires as incendiary, although the process has also been used to characterize fires classified as accidental. This process is not consistent with the scientific method, is inappropriate, and should not be used because it generates untestable hypotheses, and may result in incorrect determinations of the ignition source and first fuel ignited. Any hypotheses formulated for the causal factors (e.g., first fuel, ignition source, and ignition sequence), must be based on the analysis of facts. Those facts are derived from evidence, observations, calculations, experiments, and the laws of science. Speculative information cannot be included in the analysis." (NFPA 921 (2014 ed.) § 19.6.5.)

Carman further testified that in a number of the Cal Fire reports on the fires in this case, the method appeared to have been used improperly:

"I had questions on several of them where I felt the exclusion analysis is essentially hypothesis testing and I got the distinct impression that the careful and serious challenge called for in hypothesis testing wasn't necessarily as serious a challenge as it might be. I thought it was somewhat cursory in a few cases the way it was described. Or in some cases. I wouldn't want to put a number on it. There were several instances where as I was reading the reports I did not get the feeling that or get the impression that the challenge that was described was thorough."

Carman gave his opinion on several other aspects of the investigations he considered to be substandard. All of these were related in one way or another to the process of making an indirect arson finding, either by eliminating non-arson causes or relying on other forms of circumstantial evidence.

In some of the Cal Fire reports, Carman found references to leads investigators had obtained but not followed up on. In some instances, the reports showed that reporting parties and other witnesses were identified and their telephone numbers collected, or vehicles seen in the area were identified, but there was no indication of follow-up. Carman opined that if these reports reflected what actually happened, these investigations were inadequate, because the dropped leads meant possible causation hypotheses were not being developed and tested.

On the geospatial and temporal patterns, Carman opined that Cal Fire should not have treated these as grounds for believing the fires were arson. In his view, these patterns could be valuable as investigative tools, pointing the direction for further inquiry and justifying a focus on a particular suspect; but actual evidence that a fire was ignited in a certain way was needed to support a finding of arson.

Carman's opinion about the value of the surveillance camera data in this case was similar. He thought it had value in pointing to Jackson as a suspect, but it was not the sort of evidence that could establish arson as the cause of any fire, in his view.

On the question of whether it was proper for an investigator to use the existence of prior fires with undetermined causes or fires suspected of being arson as a reason for believing the next fire is arson, Carman said this could be appropriate if the fires were very close together in space and time—a few feet and a few minutes. For fires separated by days and miles, he did not think the inference was supported.

The destruction of the origin area by firefighting activities would weaken investigators' ability to find or exclude possible causes. Small items of evidence, such as hot particles from vehicles' brake shoes, could be lost in this way. A related problem was that, according to the documents Carman reviewed, the soil was sifted in only two cases, the last two fires in the series. Sifting soil with a sifter is a technique used to find non-magnetic metal and non-metal particles that could be related to a fire, such as particles thrown from vehicles. This was particularly important with respect to the fires that were by the sides of roads, as it affects the investigator's ability to properly exclude vehicles as a potential cause.

Carman believed some of the deficiencies he observed could be an indication that the investigators were subject to confirmation bias or expectation bias: cognitive tendencies to look for explanations one expects, and to focus inquiries on confirming an expected explanation. The belief that prior arsons supported the view that the next fire was arson was a possible example of these biases in action.

Giving his overall assessment of the Cal Fire investigation as reflected in the reports he reviewed, Carman testified that the job done on some of the fires was inadequate: "Some of those were done much closer to the standard of care than others.... Across the board, I would not say that the entire population of cases met the—what I would consider meeting the standard of care, the base-line of where it should be." He testified that at one point he told defense counsel that he believed the investigations as reflected in the reports were inadequate in 12 of the 31 cases. He did not state which cases these were.

2. DNA evidence

The defense introduced DNA evidence. Five cigarette butts were tested for DNA. A full DNA profile was obtained from one butt and a partial profile from another. Neither matched Jackson's or Waterman's DNA. The other three butts had no DNA on them. Cal Fire apparently failed to maintain a record of where exactly the butts were found. An expert testified that temperatures above 95 degrees Celsius destroy DNA, so the butts that had DNA were probably found outside the specific origin areas of the fires where they were collected.

3. Uncharged fires

The defense presented evidence of two fires, named Harsh and Hensley by Cal Fire, which the investigators had excluded from the Gold Mine series. The Harsh fire took place beside Highway 41, to the south of Yosemite Lakes Park, and the Hensley fire beside Road 400. These fires were similar in some respects to the fires included in the series. For the Harsh fire, which took place on June 17, 2013, Cal Fire personnel included arson as a possible cause because there had been other fires in the area. Sparks or hot molten objects from vehicles were included as a possible cause because metal fragments were found, and smoking was included as a possible cause because cigarette butts were found and the atmospheric conditions were right. The specific origin areas were disturbed by firefighting activity. For the Hensley fire, which took place near Hensley Lake on June 18, 2013, Cal Fire personnel concluded the cause was arson, based on finding a firework and seeing no evidence of other causes. The defense presented this evidence to support the view that the selections of fires for inclusion and exclusion from the series created an artificial appearance of a coherent pattern in the roadside fires pointing to Jackson.

4. Carl Jackson

Jackson's father, Carl, testified that he often spent time at defendants' house and they "always had people walking through" the area. He had seen as many as four or five kids at a time in the nature area behind the house on numerous occasions over the years. This testimony contrasted with the testimony of a Cal Fire witness who excluded children playing with fire as causes of the fires near the house because he saw no children and the playground equipment at the vacant neighboring house appeared disused.

E. Evidence of resisting , battery , and conspiracy

Cal Fire Captain Smith testified that he, Cal Fire Chief Andy Andersen and other officers approached Jackson outside defendants' house on the night of the last charged fire and identified themselves as police officers. Jackson was ordered to drop his hose and lie on the ground. He turned to look at the officers but failed to comply after the order was repeated. Smith drew his gun. He tried to pull Jackson to the ground, but Jackson pushed his hand away. Andersen tried to bear hug Jackson and Jackson pushed back, causing Andersen to fall down. Finally, the officers tackled Jackson and handcuffed him.

No evidence was introduced that was specifically relevant to the elements of conspiracy. III. Verdicts and sentences

The jury returned verdicts and the court imposed sentences as detailed in the majority opinion. IV. Sufficiency of evidence

Jackson maintains there was insufficient evidence to prove any of the arson counts or the conspiracy counts beyond a reasonable doubt. Waterman joins in this argument. I agree with this contention for the conspiracy convictions and for all but three of the arson convictions (Jackson's convictions on counts 28, 29, and 30).

When considering a challenge to the sufficiency of the evidence to support a judgment, we review the record in the light most favorable to the judgment and decide whether it contains substantial evidence from which a reasonable finder of fact could make the necessary finding beyond a reasonable doubt. The evidence must be reasonable, credible and of solid value. We presume every inference in support of the judgment that the finder of fact could reasonably have made. We do not reweigh the evidence or reevaluate witness credibility. We cannot reverse the judgment merely because the evidence could be reconciled with a contrary finding. (People v. D'Arcy (2010) 48 Cal.4th 257, 293.)

Although this standard is deferential, it is not the case that "an appellate court should sustain a conviction supported by any evidence which taken in isolation might appear substantial, even if on the whole record no reasonable trier of fact would place credit in that evidence." (People v. Johnson (1980) 26 Cal.3d 557, 577 (Johnson).) Our Supreme Court has cautioned against misinterpreting the substantial evidence standard in this way:

"As Chief Justice Traynor explained, the 'seemingly sensible' substantial evidence rule may be distorted in this fashion, to take 'some strange twists.' 'Occasionally,' he observes, 'an appellate court affirms the trier of fact on isolated evidence torn from the context of the whole record. Such a court leaps from an acceptable premise, that a trier of fact could reasonably believe the isolated evidence, to the dubious conclusion that the trier of fact reasonably rejected everything that controverted the isolated evidence. Had the appellate court examined the whole record, it might have found that a reasonable trier of fact could not have made the finding in issue. One of the very purposes of review is to uncover just such irrational findings and thus preclude the risk of affirming a finding that should be disaffirmed as a matter of law.' (Traynor, The Riddle of Harmless Error (1969) p. 27.) (Fns. omitted.)" (Johnson, supra, 26 Cal.3d at pp. 577-578.)

In a case in which the judgment rests upon circumstantial evidence, the substantial evidence standard is the same as in other cases (People v. Rodriguez (1999) 20 Cal.4th 1, 11), but there is case law specifically addressing the application of this standard in situations where a judgment may rest on speculative inferences. In People v. Morris (1988) 46 Cal.3d 1, 19-20, overruled on other grounds by In re Sassounian (1995) 9 Cal.4th 535, 544-545, footnotes 5-6, the question was whether substantial evidence supported a finding that a murder was committed with the special circumstance of robbery. The victim was found unclothed, having been shot; he died shortly after the police found him. Three days later, the defendant was found in possession of a credit card the victim had once borrowed from a third party. There was no other evidence of robbery. Our Supreme Court held that the jury could not reasonably infer that the defendant had taken the credit card from the victim by force or fear or that the murder had occurred during the taking. "We may speculate about any number of scenarios that may have occurred on the morning in question. A reasonable inference, however, 'may not be based on suspicion alone, or on imagination, speculation, supposition, surmise, conjecture, or guess work. [¶] ... A finding of fact must be an inference drawn from evidence rather than ... a mere speculation as to probabilities without evidence.'" (People v. Morris, supra, at p. 21.)

A. Roadside fires (Jackson only)

1. Counts 9-11, 13, and 21-27

For 11 of the 14 roadside fires of which Jackson was convicted—all but counts 28 to 30—I conclude the evidence was insufficient to allow a reasonable factfinder to find Jackson guilty beyond a reasonable doubt.

I begin with the surveillance camera data, upon which the prosecution placed its primary reliance as means of establishing Jackson's identity as the setter of all these fires. At best, the camera data could show only that vehicles passing the cameras could have reached the fire sites at suitable times if they continued driving at the speed and along the route Smith assumed. Further, the cameras were capable of identifying only vehicles, not occupants. There was evidence that Jackson often drove his truck and Waterman did not drive it, but there was no evidence that Jackson was actually in the truck on any of the 11 occasions here at issue. The data merely potentially identified Jackson's vehicles as being among those potentially at the scenes of the fires at times roughly approximating the times the fires started. That's all. Thus, even if the camera surveillance network had been a perfect means of capturing all vehicles that could have driven past the fire sites at the relevant times, the data would not have been sufficient to identify Jackson as the perpetrator.

And the surveillance network was far from being a perfect means of capturing all vehicles that could have driven past the fire sites at the relevant times. The license-plate capture rate was only 60 percent for much of the investigation and was never better than 95 percent. One route into and out of the development lacked camera coverage throughout the investigation. Vehicles that never left the development could have gone to all the fire sites inside it without passing a camera, while vehicles that never entered the development could have gone to all the fire sites outside it without passing a camera.

Among the vehicles the surveillance network did photograph, Jackson's were very far from being the only ones that could have been at the right place at the right time for any fire. Many other cars could have done so, some for more than one fire.

Further, by Cal Fire's expert's own account, the data were incriminating as to Jackson's vehicles only under a very specific theory of arson, without which the data tended to exonerate Jackson by placing his vehicle at the scene too early: arson by time-delayed incendiary device. Yet for the fires Jackson was convicted of setting, there was perilously close to no evidence at all that such devices were involved. This evidence consisted of one cigarette butt with a burn mark, found at the site of the fire in count 9, which Cal Fire could not, despite its efforts, confirm was a trace of such a device. Cal Fire Chief Gilbert testified he was inclined to believe all the roadside fires were started by means of time-delayed incendiary devices because that theory would be consistent with the camera data, assuming Jackson started the fires—but this only shows that the incendiary-device theory rested on the view that Jackson was guilty instead of supporting it.

The materials found in the search of defendants' property added little. Everything found was innocent in itself and failed to form any link between Jackson and the start of any fire. The matches and cigarettes were normal possessions of a smoker and the magnesium strike block was conceded by the prosecution's expert to be unrelated to the fires. There was no reason to think possession of firefighting information was unusual for a person in Jackson's line of work. The radio scanner and list of fire channels also was consistent with Jackson's work, and not anomalous for anyone living in a fire-prone area during fire season.⁶⁰ The use of a popular screensaver showing a fire signified nothing. The asserted link between the fishing weights and the fires was speculative, depending on assumptions that incendiary devices started the fires and fishing weights were used to make the incendiary devices. (Remember that the incendiary-device theory really rested on the camera data, not on physical evidence of incendiary devices, let alone evidence of a fishing weight.)

60 For example, Martin Powell, a Coarsegold store owner who testified for Jackson about Jackson's whereabouts on the morning of the fire charged in count 27, said he had a scanner on in his store that morning and heard a call for a fire. --------

The statement Jackson gave after his arrest was not incriminating. The worst thing in it was his concession, under the pressure of his interrogator's false claims of evidence against him, that he carelessly threw a cigarette out of his truck window that day.

The People assert that no more fires occurred in the relevant area after defendants' arrest, but, as has been seen, the evidence did not establish this. It established that there were no more fires deemed suspicious after that time within the development itself, but the Gold Mine series was not confined to the development.

The "geospatial" and time-date "patterns" discussed by Gilbert would not have implicated Jackson in the roadside fires even if, as Gilbert opined, they were indicative of arson as the cause. (That is, whatever value this evidence had was relevant to the causes of fires, not to the identity of perpetrators.) Further, although Gilbert believed the "patterns" indicated that the roadside fires were started by one arsonist (and the fires near defendants' house by another), there was nothing in his testimony to support this opinion. He described several "geospatial" patterns: a linear pattern along Highway 41, another along Road 400, a cluster near the intersection of Road 400 and Yosemite Springs Parkway, and so on. He presented no reason independent of the camera data for thinking all the patterns were created by one arsonist.

All that really remains of the case against Jackson on these 11 roadside fires, after the People's contentions have been carefully analyzed, is the notion that it would be a great and improbable coincidence if the one owner whose vehicles appeared on camera within the time window for all of these fires were not the one person responsible for all of them. The People argue that in light of this notion, the "doctrine of chances" (People v. Erving (1998) 63 Cal.App.4th 652, 660, 661 (Erving)) allows the evidence to be regarded as providing proof beyond a reasonable doubt.

Before examining this notion, it is important to consider whether the coincidence would really be so great in this case. The camera data did not implicate Jackson's vehicles in all 21 roadside fires, as Gilbert's testimony implied. The cameras were not installed or not operating for the first five roadside fires (counts 1-3 and 7-8). On two of the roadside fires for which there was camera data (counts 12 and 17), Jackson was found not guilty. This makes 14—the 11 I would reverse plus the three (28-30) I would affirm based on other evidence—for which the camera data were relevant. One more (count 13) should be subtracted because the camera data show Jackson's truck would not have made it to the fire location until 14 minutes after the fire was reported, so the data do not actually provide any support at all for the conviction. That leaves 13 roadside convictions for which the People are relying on the implausibility of coincidence. Yet, another vehicle—the light-colored Subaru—was photographed within the window for eight fires, counting only those captures for which a license plate number was obtained. The People's case requires that this be accepted as innocent coincidence, although it involves only five fewer fires.

It also is necessary to take account of the fact that no evidence was presented regarding how much driving Jackson did in the vicinity outside the 40-minute or two-hour fire windows. A driver who spent more overall time on the roads in question would be more likely to appear innocently within the fire windows than one who spent less. If the Subaru driver did less driving overall than Jackson, the coincidence of the Subaru appearing in the windows for eight fires might have been greater and more implausible than that of Jackson's vehicles appearing in the windows for 13.

In fact, the jury was not informed in any manner about what the odds were of Jackson's vehicles appearing innocently at the camera locations within the fire windows for 13 fires. It was not provided with information from which it could rationally form even a rough supposition of the likelihood of this. In considering how great and implausible the coincidence would be, the jury could only speculate.

Returning to Erving, that case does not show the evidence in this case was sufficient to prove Jackson guilty beyond a reasonable doubt. The Court of Appeal relied on the "doctrine of chances" in Erving to show the admissibility of evidence, not its sufficiency. (Erving, supra, 63 Cal.App.4th at pp. 661-663.) There is no contention that the surveillance camera evidence was inadmissible in this case. Further, the unlikeliness of coincidence was far clearer in Erving:

"Appellant lived in four different, geographically distant neighborhoods. Arson fires regularly occurred either at her home, or within easy walking distance of it. The fires were set during her residency and stopped when she moved. In one neighborhood, appellant was seen attempting to set a fire. In another, she was the only person seen on the street while five arson fires burned a few blocks away. The doctrine of chances tells us it is extremely unlikely that, through bad luck or coincidence, an innocent person would live near so many arson fires, occurring so frequently, in so many different neighborhoods." (Erving, supra, 63 Cal.App.4th at p. 663.)

In this case, no one saw Jackson lighting or attempting to light any fires, fires did not follow him as he changed residences, and he was never the only person seen as five fires burned nearby.

Other cases cited by the People as part of their discussion of this notion—that the coincidence would be too great if Jackson were not responsible for the roadside fires—are similarly unhelpful. (People v. Andrews (1963) 222 Cal.App.2d 242, 243-246 [circumstantial evidence of incendiary cause of fire sufficient to establish corpus delicti]; People v. Edwards (2013) 57 Cal.4th 658, 710-712 [discussing admissibility of similar uncharged misconduct to show identity under Evid. Code, § 1101, subd. (b)].)

In summary, a considerable series of speculative inferences was necessary to arrive at the guilty verdicts on these 11 roadside fires. The drawing of these inferences amounted to "'speculation as to probabilities'" (Morris, supra, 46 Cal.3d at p. 21), and the jury could not "reasonably reject[] everything that controverted" them (Johnson, supra, 26 Cal.3d at p. 577). A reasonable factfinder could not deem the evidence proof beyond a reasonable doubt.

2. Counts 28-30

The case against Jackson on three of the roadside fires, those charged in counts 28, 29, and 30, was of a different character from the evidence on the remaining counts. It did not depend heavily on inferences from the camera data (although those data helped for counts 28 and 30), or on borderline irrelevancies like the screensaver or the magnesium strike block that no one thought was a fire cause. Instead, it depended primarily on witnesses' actual observations and (for count 30) on data from a GPS tracker.

On count 28, LaClair testified that he drove into Yosemite Lakes Park on June 23, 2013, to do undercover surveillance in the area behind defendants' house, arriving at around 12:45 p.m. As he drove into the development, he saw Jackson driving his truck along Yosemite Springs Parkway. He and his team got into position around 1:30 p.m. Jackson's truck had a distinctive sound and LaClair heard it arriving at and departing from defendants' house several times that day. One of the times he heard it leaving was 3:09 p.m. As the truck drove, LaClair could tell from the sound that it exited defendants' street, turned right on Revis Road and headed north.

Revis Road intersects with Yosemite Springs Parkway at a point north of defendants' street. The YSP West camera showed Jackson's truck proceeding north on Yosemite Springs Parkway at 3:16 p.m., going toward the location of the fire, which was one minute to 30 seconds beyond the camera. The same camera showed Jackson's truck proceeding south two minutes later, at 3:18 p.m. LaClair heard the truck return to the house at 3:21 p.m. The fire was reported at 3:31 p.m. LaClair could see the smoke from his surveillance position.

LaClair heard the truck leave the house again at 3:37 p.m. It was photographed by the YSP West camera, 30 seconds to one minute away from the fire location, at 3:41 p.m., going north, toward the fire. At 3:51 p.m., the same camera showed Johnson as a pedestrian.

Cal Fire Chief Chris Christopherson arrived at the fire site at 3:47 p.m. to take over as incident commander. After driving around to survey the scene, he saw a black truck parked near the intersection of Road 400 and Yosemite Springs Parkway, close to the fire. Christopherson also parked, and Jackson got out of the black truck and approached him. Jackson told Christopherson he had seen a white Jeep Liberty in the area of the fire and gave him a partial license plate number. A white Jeep Liberty had been in use by the Cal Fire surveillance officers. Jackson's truck passed the YSP West camera again going south at 3:54 p.m.

All this evidence is circumstantial, but it is far superior to the evidence for the counts discussed above. It is reasonable, credible, and of solid value. From it, the jury could reasonably infer, beyond a reasonable doubt, that Jackson left home at 3:09 p.m., proceeded toward the fire location and reached it between 3:16 and 3:18 p.m., left something there capable of setting the brush on fire, and returned home. After the smoke was visible at his home, he returned to the scene to see the results and to give diversionary information to Cal Fire.

The fire in count 29 took place later the same day as the fire in count 28, June 23, 2013. It was located a quarter to a half a mile, or less than one minute, away from defendants' house, to the north on Revis Road. There were no Cal Fire cameras between the house and the fire, and defendants' vehicles were not photographed during the 40-minute window. LaClair, still conducting surveillance near defendants' house, heard Jackson's truck start at 7:19 p.m. and again leave defendants' street and turn right, heading north on Revis Road. This time it remained within earshot until the sound began to be closer again. The truck returned to the house at 7:25 p.m. Waterman was outside the house walking a dog during this time. LaClair was notified of the fire at 7:25 or 7:28 p.m. Quinn said the fire was reported to him at 7:41 p.m. At trial, Waterman admitted she later told LaClair that Jackson never left home that night, and that if LaClair heard the truck being driven it must have been imaginary. This was also when Waterman said she prayed Jackson was not an arsonist.

This evidence also was sufficient to support the jury's inference, beyond a reasonable doubt, that Jackson drove to the location of the fire and started it as LaClair was listening. The jury could infer that Waterman denied he left the house in order to protect him.

On count 30, Cal Fire used a GPS tracker to trace Jackson's truck as it traveled from defendants' house to the fire location on the opposite side of Yosemite Lakes Park. The truck reached this location at 3:05 p.m. and passed it at 37 miles per hour. A short distance beyond that point, the truck turned around; it passed the fire location again going the other way at 3:08 p.m. at 36 to 39 miles per hour. At 3:14 p.m., Cal Fire surveillance officer Kirk Gramberg saw the truck on the road and saw a white male driving it. The tracker showed the truck arriving back at defendants' house at 3:21 p.m. The fire was reported at 3:28 p.m.

Again, the facts here do not require speculation about whether and when Jackson's truck was at the fire site or which member of his household was driving it. The evidence affirmatively places him at the scene at a time when he could have set a process to start the fire in motion. The GPS tracker data in this instance—like the witnesses' observations in counts 28 and 29—supported a reasonable inference that Jackson left home specifically to proceed to the fire location and that he returned home more or less directly after going to it. The fire was 26 feet from the road and the defense expert, Carman, testified that he did not believe a cigarette could be thrown that far from a car moving at 37 miles per hour; but the jury was not required to accept this. A reasonable factfinder could infer, under the reasonable doubt standard, that Jackson started the fire.

B. Cluster fires (Jackson and Waterman)

The prosecution did not present substantial evidence to convict Jackson or Waterman of any of the fires clustered in the nature area behind their house and the neighboring house. The prosecution's case again rested on speculative inferences, being founded mainly on the occurrence of multiple fires in this area.

The occurrence near a defendant's home of multiple fires, believed by investigators to be arson, is certainly relevant to the question of the defendant's guilt of arson. In this case, it is not enough to amount to proof beyond a reasonable doubt. The area in which the fires occurred was not exclusively at defendants' home, but was rather a zone as large as half a square mile mostly situated in a nature area bordering their property and the property of neighbors. Many people had access to this area, and in some instances, drivers or bicyclists were actually seen or heard coming and going close to the times of the fires. Under these conditions, other substantial evidence in addition to the location and multiplicity of the fires was necessary to link Jackson and Waterman to them. As we will explain, the other evidence that was introduced was not adequate.

1. Jackson

For Jackson, this other evidence is all but nonexistent. He was found guilty on counts 5, 6, 14, 15, 16, 19, and 31. Counts 5 and 6 happened a few feet away from each other around the same time. Cal Fire responders arrived and saw Jackson standing 50 feet from the fire and holding a hose, saying he was planning to use it if the fire got closer to his house. He texted a picture of the fires to Waterman.

The testimony on count 14 did not refer to Jackson at all. A Cal Fire investigator spoke to Waterman the next day. They talked about a bush the investigator claimed was, and Waterman claimed was not, burned. There was no evidence that the possibly burned bush had anything to do with any of the charged fires. The fires in counts 15 and 16 again happened near one another and close to the same time. Waterman indicated that Jackson was home at the time.

On count 19, the evidence was that Waterman was at a soccer game when the fire started and Jackson was home. This fire started while another fire, 200 yards away, was still in the process of being extinguished. It was located 30 feet from defendants' house. The People assert, as if it were a fact, that "[n]o one else besides Jackson could have started this fire, so close to Jackson's house, without alerting either Jackson or the other firefighters." The People do not, however, cite any evidence in the record in support of this assertion, and indeed it is unsupported. Jackson could have been asleep (as Waterman said he was when she got home an hour later), the firefighters were 200 yards away, and it was nighttime.

On count 31, Jackson was again found by firefighters standing near the fire with a hose. Witnesses said the hose first was not, then was, spraying water.

In summary, the evidence against Jackson on the cluster fires was that they happened within an area up to half a square mile adjacent to his house, he was at home during some of them, and he was standing outside with a hose during some of those. This is not evidence of solid value. The innocent items found in the house and in his truck did not significantly bolster the case.

2. Waterman

Except for count 20, the evidence against Waterman for the cluster fires was similar: It consisted mainly of the locations of the fires and her presence at her home. Waterman was convicted on counts 14, 15, 16, 18, 20, and 31. On count 14, as mentioned above, the additional evidence against Waterman consisted of a conversation she had with an investigator about a bush that was unrelated to any of the charged fires. On counts 15 and 16, she told an investigator she saw the fire as it began, either seeing smoke and then flame or just seeing a small area of flame. There was Cal Fire testimony that a person who sees a fire starting or when it is small was likely there when it started, which Waterman did not deny. (At one point, Gilbert testified that he understood Waterman to have said she saw two or three fires when they were small, but, as far as we can determine, this referred to her statement about counts 15 and 16 and to the fact that LaClair claimed he saw her at the ignition of the fire in count 20.) On count 18, Waterman's car arrived home about 20 minutes before the fire was reported. On count 31, one of the undercover surveillance officers said he saw Waterman outside as smoke began rising from the fire some distance away; Waterman looked over her shoulder toward the smoke but did not yell "Fire!" until a fire truck approached. This evidence might reasonably be viewed as giving mild support to the suspicion aroused by the locations of the fires, but that is all. The discovery of innocent property (matches, an unopened magnesium block, a police scanner, etc.) connected with fire in a general sense again did not add anything meaningful to the evidence.

Count 20 had evidence of a different nature, but it also was insufficient. This was the fire in connection with which LaClair insisted that, 11 days after he thought he saw two starkly dissimilar people near the fire—one before and one after it started—the truth entered his mind that they were really both Waterman and that she started the fire. LaClair's claim that the person who appeared to light the fire was Waterman was, of course, incriminating. But it also was inexplicable in light of the rest of his testimony about the incident. It was "evidence which taken in isolation might appear substantial," even though "on the whole record no reasonable trier of fact would place credit in" it. (Johnson, supra, 26 Cal.3d at p. 577.) It was not possible for Waterman to have both of these appearances, and LaClair offered no explanation for his alleged mistake in describing the first person he saw. Accepting his postarrest belief that they were the same person over his on-the-scene determination that they differed greatly would be the sort of "irrational finding[]" Chief Justice Traynor warned against as a potential pitfall in the application of the substantial evidence standard. (Id. at p. 578.) It is not evidence that is reasonable, credible and of solid value, and it does not suffice to support the conviction on count 20 alone or in combination with other evidence.

The People's appellate brief attempts to support the judgment against Waterman on the cluster fires by claiming she got caught up in making some inconsistent statements about them at various points. The claimed inconsistencies are trivial or not supported by the record at all.

On count 14, the People claim Waterman told McCann, the Cal Fire investigator, that she was not home when the fire took place, but told the television reporter she was home, and also said she was home at trial. In reality, McCann only said Waterman told him she had been "gone that night" at some unspecified time. He never claimed she said she was absent during the fire. The People say that at trial, when Waterman said she did not remember talking to McCann about the supposedly burned bush, she contradicted something she told McCann about it. But McCann only said Waterman said she "noticed that" when he pointed to the bush. There is no evidence she ever agreed with him that it was burned. Her testimony that she did not remember the discussion was not incriminating. In any event, the bush is a red herring. There is no reason to think it had anything to do with any issue in the case.

The People say Waterman claimed she saw both the fire in count 15 and the fire in count 16 start, but her account had them starting in the wrong order. The only relevant remark in the 43-page block of her testimony the People cite for this contention, however, simply says she saw one of the two fires start.

The People also claim there was an inconsistency in Waterman's description of what she heard when the fires in counts 15 and 16 were starting because she referred to hearing a bicycle and to hearing a car. What the record shows is this: Cal Fire Apparatus Engineer Neil Tito said Waterman told him she heard a vehicle. Engineer McCann said Waterman told him she heard a bicycle before she went out to look, and then heard a car afterward. At trial, she testified that she did not remember anything about a bicycle and did not mention a bicycle when she called 911. When shown a document (which does not appear to be in the appellate record), she conceded that she could have said what the document indicated, but did not remember. The foundation of the document was challenged; the challenge was not resolved; and the matter appears not to have been revisited.

On count 31, the People contend that Waterman's account at trial of her whereabouts when the fire started was inconsistent with a Cal Fire witness's description of what he saw. The Cal Fire witness said he saw Waterman walking outside, seeing the smoke, and shouting "Fire!" when a fire truck approached. Waterman testified that she was starting to eat dinner when she and Jackson heard about the fire on the police scanner. Then she walked outside, saw the smoke, and began running and yelling to Jackson.

These supposed inconsistencies lent no significant support to the case against Waterman. They do not change our view that the evidence was insufficient to allow a reasonable factfinder to find Waterman guilty beyond a reasonable doubt.

C. Conspiracy (Jackson and Waterman)

To prove a charge of conspiracy, the prosecution must show: "(1) an agreement between two or more people, (2) who have the specific intent to agree or conspire to commit an offense, (3) the specific intent to commit that offense, and (4) an overt act committed by one or more of the parties to the agreement for the purpose of carrying out the object of the conspiracy." (People v. Vu (2006) 143 Cal.App.4th 1009, 1024.) Circumstantial evidence can be used to prove these elements, and the agreement proved can be a tacit one. (Id. at p. 1025.)

In arguing that there was sufficient evidence to prove the conspiracy charges, the People rely on the fact that Jackson and Waterman were each convicted of setting many fires, including the four they were both convicted of setting. The People say it is "inconceivable that two people, living together as husband and wife in the same house, located in such close proximity to so many fires, would each be independently lighting multiple fires without ever mentioning it to the other." This is indeed unlikely, if not "inconceivable," but the picture changes drastically when only three counts, by only one defendant, were supported by substantial evidence. It is easily conceivable, and not especially unlikely, that Jackson could set the fires in counts 28, 29, and 30 while keeping this activity secret from Waterman. The People did not prove beyond a reasonable doubt that it did not happen that way.

The People refer again in this context to the matches, lighters, strike blocks, police scanner, list of scanner channels, screensaver, and Cal Fire reference materials. As we have said, these are all innocent in themselves. They are not circumstantial evidence of an agreement between Jackson and Waterman to commit arson.

The People refer to various actions taken by Jackson and Waterman at their home prior to the dates of the three arsons that were proved by substantial evidence. For instance, they say agreement is illustrated by a handwritten statement Jackson prepared after the fire in count 20. Jackson wrote that after a person in a ghillie suit cursed at Waterman, he (Jackson) yelled back and chased the person. Given that there was no conviction based on substantial evidence for any of the offenses predating count 28, however—and given that Waterman was not convicted of setting any fires on the basis of substantial evidence—actions like these do not prove an agreement to commit arson.

The People also argue that Jackson and Waterman "coordinated their movements" on June 23, 2013, the date of the fires charged in counts 28 and 29. But this is merely a tendentious description of the fact that Waterman was walking outside the house while Jackson was out driving. There is no evidence of coordination. Similarly, the People suggest an agreement is proved by the fact that, on June 25, 2013, the date of the fire charged in count 30, Waterman was outside and shouted "fire" when firefighters arrived—giving rise to a surveillance officer's impression that she looked at the smoke and then did nothing until the firefighters arrived—while Jackson was in the back yard "watching the fire." The suggestion here that Jackson and Waterman agreed to stall and wait while the fire burned, and thus must have agreed to start the fire, is speculation. In any event, the fire near their house on June 25, 2013, was count 31, of which there was insufficient evidence of the guilt of either of them.

Finally, the People rely on the fact that Waterman falsely told investigators Jackson did not leave the house on the evening of June 23, 2013, the time when the fire in count 29 took place. Even assuming she knew he did, this is not substantial evidence of conspiracy. It is at least as likely that she was simply covering for Jackson after the fact.

I find no evidence in the record that is reasonable, credible and of solid value indicating that Jackson and Waterman ever reached any express or tacit agreement to start any fire.

Regarding the remaining issues briefed by the defendants, I would not find reversible error. The trial court did not abuse its discretion in its handling of the prosecutor's discovery violation. Regarding the exclusion of third-party culpability evidence, and the admission of expert testimony relying on the exclusion method to show causation, I would not find reversible error as to counts 28, 29, 30, 33, and 34, and I would express no view on these points as to the remaining counts. The problem of counts not included in the information or supported at the preliminary hearing would be rendered moot by my views on the sufficiency of the evidence for those counts. V. Conclusion

This case is about taking the reasonable doubt standard seriously. The intuitively strongest aspect of the prosecution's case is that it seems unlikely that there would be a group of fires correlated with the camera data on Jackson's vehicles and a group clustered near defendants' house unless Jackson or Waterman or both lit at least some of them. The trouble with this intuition is that it does not go far enough in light of the requirements of a trial in our system of criminal justice. The intuition is not enough to support a reasonable factfinder in finding beyond a reasonable doubt that both defendants were guilty of starting all the fires they were convicted of starting. But if all the convictions have not been supported by proof beyond a reasonable doubt, then which have, and how do we identify these? There must be sufficient proof to support each conviction. The intuition that the scenario as a whole cannot be a coincidence provides the necessary proof neither for all the counts collectively nor for any specific count. That is why I would affirm the convictions only on those counts as to which the necessary individual proof was presented.

I am mindful of the community anxiety occasioned by the facts in this case. But this court is bound to uphold the principle that a criminal conviction must be supported by substantial evidence that can reasonably be found to prove guilt beyond a reasonable doubt.

/s/_________

SMITH, J. Appendix A: Roadside fires (Jackson only charged)

Cnt./ date Witness Truck or motorcycle at camera Fire report- ed Drive time: camera to or from fire Other Circum- stances Causes not ruled out Causes determined Verdict 2 5/11 Firework and piece of exhaust pipe found in general origin area (GOA). Civilian man, teenager and others found at scene trying to extinguish fire. Between a quarter and a half of an acre burned. Not guilty Tito (responder) Children playing with firework; Arson by firework Firework McCann (investigator) Arson by firework 141 3 5/12 Neighbor reported fire across street. Fifteen acres burned. No cigarette butts or other items found that could have caused fire. Not guilty Pimentel (investigator) Arson 1 5/19 Home and 91 acres burned. Cigarette butt found that could have been remains of incendiary device. Not guilty Stickles (responder) Smoking; vehicle; arson; other/ miscellaneous Pimentel (investigator) Arson by incendiary device 7 5/24 Fire started in brush pile on Not guilty 142 range land in Raymond. One acre burned. No cigarette remains or other items found that could have caused fire. Waag (responder) Burning debris; arson McCann (investigator) Arson; human- caused 8 6/1 Cigarette butt found that could have been remains of an incendiary device. Witness tried to extinguish fire with limeade. Black Jetta or Corolla seen driving quickly away from fire by Not guilty 143 reporting witness. McCann (investigator) Smoking; arson by cigarette Cigarette Quinn (investigator) Arson by cigarette incendiary device 9 6/8 Truck at 400 East: 12:15 p.m. (driving away); 12:41 p.m. (towards); 12:47 (away) 12:25 p.m. 3 min. 19 sec. Cigarette butt found that could have been remains of incendiary device. About one acre burned. Estimated ignition time 12:20 p.m. Guilty Quistorff (responder) Smoking; arson by cigarette or lighter McCann (investigator) Smoking; arson by cigarette incendiary device Arson (testimony incon- sistent) 144 10 6/8 Motorcycle at 400 West: 2:36 p.m. (toward); also at Arneson camera, see count 11 below 2:57 p.m. 5 min. Residents found trying to suppress fire with shovels. 0.4 acres burned. No cigarette butts found. Guilty Tito (responder) Vehicles, arson; playing with fire McCann (investigator) Arson, but not by cigarette or cigarette incendiary device 11 6/8 Motorcycle at Arneson: 2:52 or 2:55 p.m. (away); also at 400 West camera, see count 10 above 3:17 or 3:25 p.m. Insig- nificant (distance 200 to 300 ft.) First spotted by fire crew heading for fire in count 10. About 100 sq. ft. burned. No cigarette butts found. Dark sedan and white YLP Public Works Hummer seen Guilty 145 by responders. Hummer driver stopped and spoke to responder. Estimated time of ignition about a minute before crew spotted fire. Dowell (responder) Arson; playing with fire McCann (investigator) Arson 12 6/9 Truck at YSP East: 10:58 a.m. (away) 10:49 or 10:50 a.m. 16 and a half minutes Responders saw burgundy sedan with out-of-state plates leaving scene rapidly at about 11:09 a.m.; also saw man trying to extinguish fire with shovel. A quarter of an acre burned. Cigarette butt Not guilty 146 and lighter found. Tito (responder) Arson by lighter Pimentel (investigator) Arson by cigarette incendiary device 13 6/9 Truck at 400 West: 11:50 a.m. (away) 11:35 a.m. Less than 2 minutes Cigarette butt and burned pamphlets found. Nine acres burned. Guilty McCann (investigator) Arson; smoking 17 6/12 Truck at YSP East: 2:36 p.m. (away) 2:51 p.m. Less than 30 seconds Six acres burned. Witness saw white car speed away when fire was size of manhole cover. Witness spoke to responder and tried to extinguish fire Not guilty 147 with extinguisher. Cigarette butt found with no indication of match heads having been added. Cacho (investigator) Accident with cigarette; arson by cigarette Cigarette 21 6/16 Truck at 400 East: 1:08 p.m. (towards) Truck at YSP West: 1:13 p.m. (away) (Fire location was between these two cameras.) 1:21 p.m. 3-4 minutes One acre burned. Responders spoke to reporting party at scene. No cigarette butts found. Guilty Quistorff (responder) Arson, vehicle, miscellaneous McCann (investigator) Arson 22 Truck at 3:35 2 to 3 One acre Guilty 148 6/16 Arneson camera: 3:25 p.m. (away) p.m. minutes burned. Responders encountered passer-by attempting to extinguish fire with garden hose. Metal from catalytic converter found in SOA. Quistorff (responder) Vehicle (catalytic converter) Cacho (investigator) Vehicle (catalytic converter) McCann (investigator) Vehicle (catalytic converter) —before arrest. Arson— after arrest (McCann reexamined metal piece and re- considered condition of 149 the ground under it) 23 6/16 Truck at YSP East: 8:00 (toward) 8:16 p.m. 4 and a half minutes 0.2 acres burned. Electric car mirror found burning in SOA. Guilty Chase (responder) Electrical power (from the mirror); arson McCann (investigator) In his report: Electrical power (from the mirror); arson After re- considering: Arson 24 6/20 Truck at 400 East: 1:32 p.m. (toward) 1:46 p.m. Less than one minute Three-fourths of an acre burned. Nothing suspicious found. Witness saw white Subaru driving away from site when fire was small, Guilty 150 speeding up and slowing down, with paper flying out of window. Teenage boy on bike contacted in area during investigation. Quistorff (responder) Arson, vehicles Arson Cacho (investigator) Arson, vehicles 25 6/20 Truck at 400 East: 1:32 p.m. (toward) 1:58 p.m. Six and a half to seven minutes A quarter of an acre burned. Cigarette butts found. Guilty Cooper (responder) Smoking, vehicles, arson Cacho (investigator) Smoking, vehicles, arson 26 6/23 Truck at YSP East: 9:49 a.m. (towards). 10:19 a.m. 13 minutes 20-by-20 foot area burned. Passer-by extinguished Guilty 151 Outside 40- minute window. fire with a branch after reporting party reported, and left before responders arrived. Pieces of metal found, probably from vehicle. Crabill (responder) Vehicles, arson Pimentel (investigator) Vehicles, arson 27 6/23 Truck at YSP East: 9:49 a.m. (towards). Outside 40- minute window. 10:27 a.m. 18 minutes (13 to count 26 site plus five more) 15-by-15 foot area burned. Passersby tried to extinguish fire with sand. At 10:40 a.m., Cal Fire personnel saw Jackson's truck parked by the road near the fire and saw a Guilty 152 man, 5 feet 9 or 10 inches tall and round 200 pounds, standing by the truck directing traffic. Receipt indicated Jackson made purchase at store in Coarsegold at 10:24 a.m. Store owner thought Jackson arrived before 10:00 a.m. Russell (responder) Vehicle, arson McCann (investigator) Arson 28 6/23 Truck at YSP West: 3:16 p.m. (towards); 3:18 p.m. (away) 3:31 p.m. 30 seconds to one minute Nine acres burned. Nothing suspicious found. Someone was Guilty 153 seen trying to suppress the fire with a shovel. Cal Fire personnel heard Jackson's truck leaving home and returning, and saw it driving between his home and fire location. In once instance it left at 3:09 p.m., drove toward fire location, and returned at 3:21 p.m. YSP West camera showed truck coming and going a number of times. Jackson parked near fire while 154 responders were working, and told Cal Fire personnel he saw a white Jeep. Quinn saw Jackson driving at 3:19 and looking in side mirror when fire location was behind him. Silver or white Jeep shown on 400 East camera at 3:35 p.m. and YSP West camera at 3:39 p.m., heading toward fire, but after it was reported. Quistorff (responder) Arson Pimentel (investigator) Arson 155 29 6/23 Jackson's vehicles did not appear on cameras during 40- minute window. Around 7:20 p.m., 7:25 to 7:28 p.m., or 7:41 p.m. 0.2 acres burned. Fire located a quarter to half a mile, or less than one minute, from defendants' house. Cal Fire personnel surveilling defendants' house discovered and extinguished fire. Surveillance personnel heard Jackson's truck leave home at 7:19 p.m., drive toward fire location, and return home at 7:25 p.m., remaining within audible Guilty 156 distance in between. Waterman told a Cal Fire investigator Jackson never left home that evening. McCann (investigator) Arson 30 6/25 Truck at 400 West: 2:59 p.m. (toward) Truck at camera 5: 3:10 p.m. (toward) 3.28 p.m. 6 minutes (from 400 West) Nine acres burned. Nothing suspicious found. GPS tracker showed Jackson's truck leaving home at 2:53 or 2:54 p.m., driving around the development, passing the fire location at 3:05 p.m., making a U- turn one mile beyond that Guilty 157 location at 3:08 p.m., passing the fire location again, continuing to drive around the development, and returning home at 3:21 p.m. Cal Fire surveillance officer saw white male driving truck at 3:14 p.m. Gilbert (investigator) Arson 158 Appendix B: Ignition windows for roadside fires Count/ Verdict as to Jackson Incident number Incident name Ignition window 9 (G) 9 Lakes 2 13 min. 10 (G) 10 Mountain [1] 15min. 11 (G) 11 John Muir 25 min. 12 (NG) 12 Oakhurst [1] 7 min. 13 (G) 13 Springs 2 -14 min. 17 (NG) 16 Yosemite 1 15 min. 21 (G) 20 Yosemite 2 9-10 min. 22 (G) 21 John 13 min. 23 (G) 22 Mecca 11 min. 24 (G) 25 Lakes 4 13 min. 25 (G) 26 416 20 min. 26 (G) 27 Highway 18 min. 27 (G) 28 Oakhurst 2 20 min. 28 (G) 29 Lilley 1 13-14 min. 29 (G) 30 Revis Court 16-20 min. 30 (G) 31 Mountain [2] 20 min. or 23 min. 159 Appendix C: Fires near house Cnt./ date Witness Circumstances Causes not ruled out Causes found Verdict 4 5/18 Fire at about 7:30 p.m. on a Saturday burned one and half acres of Kevin Olson's property. Olson's property bordered on same 500-acre nature area as defendants' home. Arrows and rocks with strike marks found in origin area. Olson said he last shot arrows at least a month before. Jackson NG Waterman NG Russell (responder) Spark from homeowner's arrow hitting rock (arson ruled out) McCann (investigator) Arson 5 5/22 Fire reported about 8:30 p.m. on a Wednesday behind defendants' house, in nature area. Nothing suspicious found in origin area. 0.38 acres burned. Responders saw Jackson 50 feet from the fire holding a hose; he said he was Jackson G Waterman NG 160 Fairbanks (responder) planning to try to put it out if it came closer to the house. Waterman and a teammate testified that she was in Fresno playing in a soccer match at the time of the fire. Children lived nearby and frequented the area. Two hours before the fire, a juvenile lit a fire with a lighter in his back yard about half a mile away. Cal Fire witness saw dogs that barked at strangers. Arson; children playing with fire; smoking; other/ miscellaneous McCully (investigator) Arson 6 5/22 Fire started ten feet away from fire in count 5, around same time. 0.19 acres burned. Jackson G Waterman NG Fairbanks (responder) Arson; children playing with fire; smoking; other/ miscellaneous McCully (investigator) Arson 14 Fire reported Jackson G 161 6/9 around 6:50 p.m. on a Sunday in nature area behind defendants' house. Nothing suspicious found. Investigator spoke to Waterman next day. Saw what he described as burned shrubbery next to front walkway. Said Waterman told him she noticed but did not report it. Waterman testified bush was not burned and she did not remember discussing it with investigator. Bush had "blank spots" where it had "died off." Jury was shown photo of bush, part of which lacked leaves. Leafless spot did not appear blackened in photo. Investigator never formed opinion about whether bush was arson. Thought Waterman's nonreporting was "odd." Waterman G Tito (responder) Arson McCann (investigator) Arson 162 15 6/10 Fire burned about 2,000 square feet in nature area behind vacant house next door to defendants' house around 8:40 p.m. on a Monday. Vacant house had a playground. Waterman was present during investigation after fire was extinguished. She either said she saw smoke in the grass and then saw flame, or she saw a glow the size of a roll of duct tape which then spread. She told investigators she heard a bicycle on neighbor's driveway and heard a vehicle driving away quickly. She smelled a chemical smell or a smell like gasoline. Jackson was home. Bystanders confirmed teenager on bicycle and speeding white pickup. McCann believed a person who sees a fire when very small Jackson G Waterman G 163 usually was present when it started. Felix (responder) Arson; children playing with fire Pimentel (investigator) Arson 16 6/10 Fire located about 50 feet from fire in count 15 and reported at same time. Burned about 100 square feet. Jackson G Waterman G Felix (responder) Arson; children playing with fire Pimentel (investigator) Arson 18 6/12 Fire reported in nature area, about a one to two minute walk from defendants' house, at 4:28 p.m. on a Wednesday. 25 acres burned. Game camera showed Waterman's black Toyota Sequoia entering defendants' dead- end street and heading toward her house at 4:09 p.m. Waterman told investigator she saw a white SUV. Camera showed white Ford SUV Jackson NG Waterman G 164 entering street at 4:15 p.m. and leaving at 4:16 p.m. Neighbor reported white Chevy pickup. Camera showed white Chevy pickup entering street at 4:32 p.m. and leaving at 4:33 p.m. Waterman said she left for a soccer match in Fresno around 5:30 p.m., after the fire trucks came. Investigator did not see footage from the game camera showing her car leaving, but record does not indicate whether he watched footage for 5:30. Investigator confirmed that a white SUV and a white pickup had been seen in area over a couple of days. Jackson's father testified that he and Jackson were in pool at house when fire started. Lindo (responder) Smoking; playing with fire; arson Pimentel (investigator) Arson 165 19 6/12 Fire reported about 9:40 p.m. approximately 200 yards from fire in count 18. Less than 2,000 square feet burned. YSP East camera showed Waterman's Toyota leaving the development earlier in the evening. Waterman testified that she was at a soccer match in Fresno when the fire was reported, and Jackson was asleep in bed when she got home around 10:45 p.m. Jackson G Waterman NG Lindo (responder) Arson; playing with fire Pimentel (investigator) Arson 20 6/14 Fire reported at 7:30 p.m. in nature area behind vacant house adjacent to defendants' house. One and a half acres burned. LaClair, undercover, saw a person walking toward vacant house next door between 6:30 and 7:00 p.m. Person Jackson NG Waterman G 166 disappeared for a few minutes in sloped terrain, then reappeared and walked away. LaClair found fire in the hidden area a short time later. Then Waterman appeared and yelled "Fire!" She was wearing different clothes, and was a different size, than the person LaClair saw going into the fire area. LaClair and Waterman argued and he cursed at her. LaClair fled and was picked up by a Cal Fire vehicle. Later, he later decided she was the same person he had seen in the fire area. A crowd gathered during the fire. Multiple witnesses saw people entering a white SUV and speeding off. This was the Cal Fire vehicle. A Volvo and a pickup truck also were seen. Brechbuehl (responder) Arson; playing with fire Cacho (investigator) Arson 167 31 6/25 Fire reported around 6:40 p.m. near defendants' house, in the nature area. 0.3 acres burned. Undercover officer saw Waterman walking away from the house, turning to look back toward it as smoke rose from the area. When sirens were heard, officer saw Waterman yell "Fire!" and run back toward house. Responders found Jackson standing near fire with a hose. No water came from the hose for a moment; then water came out. Jackson G Waterman G Gilbert (responder) Arson