MODELING INFILTRATION BASIN HYDROLOGY USING PARFLOW.CLM

We applied the integrated hydrologic model ParFlow-CLM using a variable-depth and terrain- following grid to simulate flow dynamics at two infiltration basins at Wakefield Park in Philadelphia, PA. The model is gridded at a 1.524-m horizontal resolution over an area of 39,130 m². Vertical resolution progressively increases from 0.1 m in the top soil layer to 68 m in the lowest bedrock layer to a total depth of 199 m. The model incorporates hydrologic properties for soil, saprolite, and fractured bedrock. Surface layer properties vary depending on whether the cells are impervious or pervious. Infiltration basin geometries are generated by modifying lidar-derived topography to create a topographic low at each basin location. Surface water is diverted from streets to the basins and allowed to flow through every model cell of each basin. CLM uses hourly NLDAS data to provide model precipitation and evapotranspiration (ET) forcing. This is an important feature that enables evaluation of spatially- and temporally-variable atmospheric and subsurface processes. For green infrastructure such as infiltration basins, these processes are often overlooked, studied independently from each other, or simplified in their implementation. Models are often simplified because during precipitation, infiltration rates are highly dependent on hysteresis of unsaturated zone hydraulic conductivity. At the land surface, cells become progressively drier at different rates due to ET and infiltration differing among land cover types. After long dry periods, the hydraulic gradient between top layers is upward due to ET fluxes at surface layers. ET is also weather-dependent, so it can be highly variable across days and seasons. For example in February, 2008, during daylight hours ET ranges from negative values when there is dew deposition, to 0.09 mm during a precipitation event, to between 0.05 mm and 1.23 mm when there is no precipitation. Daily rates differ dramatically for summer days characterized by full leaf-out, higher temperatures, and longer days. Therefore for each precipitation event, infiltration rates are affected by soil states caused by variable ET. Our aim of using this model is to quantify the effectiveness of infiltration basins across precipitation events and seasons and how they affect subsurface flow.