Sequential Conceptual Simplification of the Effective Rainfall Component of a Rainfall-Streamflow Model for a Small Kenyan Catchment

Sequential Conceptual Simplification of the Effective Rainfall Component of a Rainfall-Streamflow Model for a Small Kenyan Catchment

A six-parameter soil-vegetation-atmosphere-transfer (SVAT) model, calibrated independently by other researchers using neutron probe soil moisture observations, is applied in series with an efficient three-parameter unit hydrograph (UH) in order to simulate continuous daily streamflow for a small catchment in Kenya. The resultant nine-parameter SVAT-UH model is applied to periods of record not used for its calibration to show that it is a reasonable characterisation of the catchment. An experiment was undertaken assuming that only rainfall, streamflow and air temperature (or potential evaporation) data were available for calibrating this structure of rainfall-streamflow model for the Kenyan catchment. Sequentially reducing the structural complexity of the SVAT module, and consequently the number of its parameters from six to three, while leaving the structure of the UH part unchanged, was not accompanied by a decrease in model calibration performance. This indicates the low level of process complexity in SVAT schemes that can be identified when only rainfall, streamflow and evaporation surrogate data are available for model calibration. However, the efficacy of the independently calibrated six-parameter SVAT module when included in a rainfall-streamflow model is demonstrated for the Kenyan catchment, indicating the value of additional soil moisture measurements for rainfall-streamflow modelling.