Current through Register Vol. 56, No. 21, November 4, 2024
GUIDELINES FOR A GROUND WATER MODELING EFFORT
1. The model of use must have a history that documents its ability to represent real world situations. In addition it should also be demonstrated that the model of choice has the ability for proposed management of ground water resources.2. The set of equations, that govern ground water flow and pollutant, and the derivations of these equations must be presented.3. The numerical methods used to solve the set of ground water flow and pollutant transport equations must be presented.4. The Boundary Conditions and Initial Conditions used in solving the ground water flow and pollutant transport equation sets should be presented both mathematically and in narrative form.5. A technical narrative describing the model to be used and a justification for the application of this to the specific problem should be presented. This should include whether the model is finite element, finite difference or some other scheme. The objective of the model should be stated up front.6. The unknown quantities that the model is solving for should be described and explained. In addition those parameters derived from the initial unknown quantities should also be described and explained.7. Appropriate analytical methods should be used to verify the validity of the numerical technique used to solve the flow equations in the model.8. A sensitivity study of the error tolerance used and modal spacing needs to be conducted. The results should be presented and explained.9. Perform mass balance calculations on selected elements in the model to verify physical validity.10. The model must be calibrated against field data. It is important to note that if there is insufficient field data available for calibration then the model will extrapolate values of unknown accuracies. This is particularly important since there is no one unique solution to a model and the most accurate solution (that closest to the real world situation) is a result of sufficient field data collection and model calibration with that data. It often takes more than 25 runs with the same data to properly calibrate a model to the real world situation. The level of field data considered to be sufficient should be agreed to before the modeling effort is initiated.11. Limits and confidence on model predictions should be established and stated in the beginning of the modeling report.12. All inputs and outputs to the computer program should be listed and explained in technical narrative. N.J. Admin. Code Tit. 7, ch. 26, subch. 2A, app A