Hydrus-3D Estimates of Runoff from Pollinator Habitat at Ground-Mounted Solar Photovoltaic Sites

Ground-mounted photovoltaic (PV) sites are often treated as impervious surfaces. This ignores the
disconnected pervious soils beneath and between solar arrays, and leads to overestimation of runoff.
Our objective was to improve stormwater runoff estimates at ground-mounted PV facilities. Soil
moisture and precipitation measurements and Hydrus-3D hydrologic modeling were conducted beneath
arrays, under array drip edges and in the vegetated area between arrays at five solar PV sites located in
CO, GA, MN, NY and OR. Sites represent diverse climatic, topographic and soil conditions, with either
fixed or tracking solar arrays, and pollinator vegetation. A Hydrus-3D model for stormwater runoff was
developed based on measurements of precipitation, soil textures, soil depths, soil bulk densities, spacing
of solar arrays, type of ground cover, and slope steepness values. The model was capable of accounting
for the complex 3-dimensional nature of precipitation, drip edge redistribution of rainfall, infiltration,
runoff and evapotranspiration in the disconnected pervious areas below and between solar arrays.
RMSE values between measured daily soil moisture and predicted moisture averaged 0.028 across all
five sites, indicating the accuracy of Hydrus-3D across a range of ground solar PV sites with perennial
vegetation. Factors with a large impact on runoff included storm precipitation depth, saturated soil
hydraulic conductivity, soil bulk density and soil profile depth. Factors with a moderate impact on runoff
included type of vegetative cover, and spacing between panel arrays. Runoff across the five
experimental sites estimated using Hydrus-3D was, on average, 38% lower than runoff estimated using
the standard NRCS runoff curve number method. These improved runoff estimates using Hydrus-3D
could significantly reduce the need to install expensive stormwater mitigation practices at ground-
mounted solar PV sites.