GIS-based, python modeling of the spatial and temporal distribution of water on the landscape for wetlands decision making
Wetlands provide many benefits for humans and the natural environment, but land use changes have reduced their number and areal extent. Interest has grown in examining surface water distribution both spatially and temporally, which help to determine those locations for which there is the greatest priority for wetland preservation or mitigation. This research first proposes a methodology to support that examination through the application of open channel hydraulics principles to flow over a landscape. The methodology, implemented through a Python script, automatically extracts landscape characteristics from a DEM and calculates hydraulic parameters. The parameters are used to determine water surface profiles using the Modified Euler's method. Multiple tests show that the script accurately produces profiles of flow between wetlands over a landscape. Such determinations are the first step in understanding where water will exist on the surface and where there may be infiltration to support wetland functions. Furthermore, a water balance methodology (where water will exist, how much will be there and for what period of time) is developed and demonstrated that focuses on small depressions, as locations where conservation efforts to create or regenerate wetlands may be achievable. Integral to this analysis is a detailed treatment of depressions in the landscape. Utilizing a digital elevation model, the methodology incorporates a cell-by-cell analysis to appropriately capture small-scale processes. Instead of treating these vital depressions as errors or being insignificant to the water balance calculations, they are retained. Flow direction is dynamically determined by the land surface and water characteristics. With potentially shallow flow in depressions, the use of Manning's equation incorporates stratified flow where differing values of Manning's n describe flow through and above vegetation. This real-time overland runoff model based on a short time step is implemented through a Python code using ArcGIS. Exercises on an artificial DEM with simulated precipitation demonstrate the ability of the model to accurately represent hydraulics principles. Simulations of two field sites over a period of a year, and incorporating precipitation, infiltration and evapotranspiration, demonstrate the ability to track water surface locations and extents with an accuracy necessary for decision making. Additionally, this research optimizes the Green Ampt infiltration model which allows for the calculation of infiltration rates with unsteady rainfall and then couples this Modified Green Ampt (MGA) model with a previously developed Dynamic Flow Direction (DFD) model to simulate overland flow. To test the accuracy of the improvements, results show shorter times to ponding, smaller total infiltration at the time of ponding and larger total infiltration with this Modified Green Ampt (MGA) model as compared with the results with a Traditional Green Ampt (TGA) model. Additionally, coupled with the DFD model, the MGA model takes surface water movement into consideration. The total water volume on the landscape with MGA is less than predicted by the TGA. Additionally, the inundation area is deeper than 0.05 m with MGA and is also smaller than the result with the TGA.
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 License.