HRRTLE (High Resolution Runoff and Transmission Loss Estimator): a novel tool for mapping connectivity of runoff in ephemeral stream networks to aid the siting of water harvesting structures

Abstract. Water harvesting is predominantly carried out in arid and semi-arid regions. Site selection studies often rely on a methodology that calculates runoff using curve numbers to generate runoff maps. These maps, typically used as part of a multi-criteria selection process, identify areas conducive to the siting of water harvesting structures. However, traditional runoff maps do not account for transmission losses that occur along the surface flow path to the catchment outlet, and these losses can be significant in arid and semi-arid regions. Here we introduce a methodology that incorporates a curve number runoff method while also addressing transmission losses. Our approach, utilising three global datasets, was validated against observed runoff data from 28 catchments worldwide, and infers hydraulic characteristics of both overland and channel flow from curve number values. This involves leveraging the curve number dataset twice: initially for calculating runoff and subsequently for forecasting transmission losses. The outcomes include a runoff connectivity map, at a spatial resolution of 250 m × 250 m, presenting the runoff depth (in mm) for each pixel based on the direct runoff generated at that pixel and reaching the catchment outlet. This connectivity map aids planners in comprehending the dynamics of surface runoff towards a catchment outlet, assisting in identifying potential locations for future water harvesting structures. The process integrates 38 years of precipitation data, enabling predictions not only for average annual runoff but also for the return periods of various annual runoff volumes. Despite the simplicity of the model, a positive Nash-Sutcliffe efficiency value was observed in 11 out of the 28 catchments.