Stormwater infiltration facilities help reduce the hydrologic impacts of residential and commercial development. The design of these facilities is particularly challenging because of large uncertainties associated with predictions of both short-term and long-term infiltration rates.
Full-scale "flood tests" conducted at four infiltration facilities in western Washington suggest that lateral flow along the sides of the ponds may be significant. This is similar to "bank storage" that occurs in stream channels. More efficient designs may require a larger ratio of side area to bottom area and that maintenance activities should be considered for the sides as well as the bottom of the pond.
Saturated hydraulic conductivity values estimated from measuring air conductivity and from regression equations derived from grain size parameters were compared to full-scale infiltration rates for 15 sites in western Washington. The estimated values for saturated hydraulic conductivity were up to two orders-of-magnitude larger than the full-scale infiltration rates for some sites and were two orders-of-magnitude smaller at others. These results show that infiltration rates cannot be reliability estimated on the basis of soil properties alone; information related to the hydraulic gradient is also important.
Computer models were compared to identify the flow systems for which saturated models provide reasonable approximations. The difference between saturated and unsaturated flow models was lowest in highly permeable soils and increased as the hydraulic conductivity of the soil decreased. The simulations suggest that steady-state infiltration rates calculated with a saturated model will be 20 to 30 percent smaller than rates calculated with an unsaturated model for the range of hydraulic conductivities typically found beneath Western Washington infiltration ponds. A comparison of steady-state and transient simulations showed that the steadystate assumption may significantly underestimate infiltration rates.