The hydrologic performance of in situ bioinfiltration systems (bioretention systems with no fill media or underdrain) is quantified and soil classes are evaluated as proxies for design requirements. A one-dimensional (1D) Richard's equation model of a bioinfiltration system is used in combination with a dataset of soil hydraulic properties to conduct a Monte Carlo analysis of the effect of soil hydraulic properties; the results are summarized both by soil textural class and by hydrologic soil group (HSG), showing that textural class is generally a poor proxy for estimating the infiltration performance of in situ bioinfiltration cells (R 2 ¼ 0.40). Because infiltration measurements are required to estimate the HSG, they are a better proxy for bioinfiltration performance (R 2 ¼ 0.89). It is found that soil proxies do provide certain reliable guidelines: HSG-D soils always require engineered fill media with an underdrain; whereas underdrains are not necessary for sand, loamy sand, HSG-A, and HSG-B native soils. Minimum bounds on the design capture volume are generated for these soils which may be substantially larger than the surface storage volume.

No abstract

This article presents a method to estimate soil matrix hydraulic properties to be used for variably-saturated hydrologic analysis, such as rainfall/runoff and stormwater management analyses, among others. The methods are described and verified with the U.S. Department of Agriculture's Unsaturated Soil Database (UNSODA), and the resultant soil hydraulic properties are presented. For conservative estimates of soil infiltration, a filter is recommended to remove soils from the analysis with high-saturated hydraulic conductivity that are likely to be macropore-dominated, and it is confirmed that the removed soils contain substantially more macropore influence than the retained soils. It is shown with the UNSODA database that good estimations of the soil matrix hydraulic conductivity function may be obtained by using the modified van Genuchten-Mualem model normalized to the saturated hydraulic conductivity with an immutable model parameter h e ¼ −0.1 cm if data are not available to calibrate h e ; this allows for the hydraulic conductivity curve to be estimated without laboratory-measured unsaturated data and overcomes some of the limitations of the van Genuchten-Mualem model.

No abstract