Sensitivity of LISEM predicted catchment discharge to initial soil moisture content of soil profile

This study conducts a broad sensitivity analysis, taking into account the influence of initial soil moisture content in two soil layers, layer depths, event properties, and two infiltration models. A distributed hydrology and soil erosion model (LISEM) is used. Using the terrain data from the Catsop research catchment and two different rainfall events, the sensitivity of discharge is investigated for a range of pre-event soil moisture contents (0.1–0.40 cm3 cm−3) in two-layers for a two-layer Green–Ampt as well as Richards infiltration model. The sensitivity of the predicted discharge to the initial condition of soil moisture appears to depend highly on all factors: infiltration model, event properties, topsoil/subsoil depth configuration and the level of the initial condition itself. There are interaction effects between all the factors. However, the effect of the different infiltration models is most pronounced. The Green–Ampt model shows less sensitivity to moisture content variation of both top and subsoil. Top/subsoil depth configuration rarely influences the results of the Green–Ampt model. The Richards model shows a highly variable discharge – initial soil moisture relation with changing rainfall intensity and topsoil/subsoil depth configurations. Two methods of sensitivity analysis, relative sensitivity and One factor-At-a Time sensitivity, have been used. The two methods gave comparable results. Depending on the other parameter values, 1% changes in topsoil moisture content resulted into 0.8–1.81% and 0.03–3.5% changes in total discharge predicted by the Green–Ampt and Richards models, respectively.