Drywell Infiltration Performance: Tests, Monitoring, Simple, and Detailed Models

Abstract Drywells, perforated above the water table, are an attractive tool for both reducing the risk of floods, and increasing groundwater recharge in urban\suburban areas. Various simplifications of the relationship between the injection discharge (Q) and the water‐level rise in the drywell during water injection (H) are available. This work presents observations and models that improve our understanding of the drywell performance, namely the ratio Q/H for injections varying in time and dynamics. The drywell screen is at 22–27 m below surface in sandy porous medium, where the water table is at 40 m depth. The first set of observations were of six injection tests of constant Q, each lasting ∼30 min, performed on a daily basis. The Q/H ratio just before the end of each injection decreased each day. A simplified infiltration model assuming a radial sharp‐wetting‐front with an increase in the distance of the wetting front from the well fits the observations. A three‐dimensional variably saturated numerical flow model simulating the six injection events showed that the sharp wetting front at increasing radius is a reasonable simplification for this type of injection schedule. Monitoring of operational injection of rainwater harvested from an adjacent rooftop for a few months of a Mediterranean winter shows the opposite—a slight increase in the Q/H ratio as winter progresses. When the plume of relatively high pressure‐head reaches the water‐table, a continuous passage of higher hydraulic conductivity between the drywell and the aquifer is opened, and the Q/H ratio is expected to increase.