Correctly predicting water levels is key to the success of created wetlands. The Pierce method is a commonly used technique for modeling and designing mitigation wetlands that assumes minimal groundwater interaction with the wetland. This technique for mitigation wetland design relies primarily on surface water inputs, assuming a relatively impermeable substrate (perched system), and level pool routing. The Pierce method was applied utilizing two different evapotranspiration estimation methods: Thornthwaite (IPM) and FAO-56 Penman-Monteith (IPM-FAO). A second process-based model, utilizing MODFLOW-2005, was constructed to better predict water levels in mitigation wetlands. Modeled processes included groundwater movement and vegetative resistance to flow, which can be a significant factor in wetland water levels. The two versions of the Pierce method were compared to the process-based wetland representation developed in MODFLOW-2005 using data from an existing mitigation wetland.

Output from these models were compared to observed data from an existing mitigation wetland near Manassas, VA, USA. Results indicate the use of Thornthwaite's method to estimate wetland evapotranspiration (ET) does not capture the timing or magnitude of wetland ET losses, leading to over-prediction of wetland water levels during the growing season. The Modflow-based approach resulted in more accurate hydroperiod predictions on a yearly basis than the Pierce Method. However, the Integrated Pierce method model, utilizing the FAO-56 Penman-Monteith method of estimating potential evapotranspiration instead of Thornthwaite's method most accurately predicted water levels during the growing season (March-October).