A complex network theory based approach to better understand the infiltration-excess runoff generation thresholds

Abstract

Understanding hillslope-scale surface runoff generation processes are challenging due to the non-linear interactions among the different hydro-meteorological variables. Often the complex connection between different hydrologic variables is little understood (i.e., the spatial–temporal connection among the same variable and between two different variables). Therefore, in this paper, we present a network theory-based approach to understand the complex connections of hydrological processes. The article mainly focuses on identifying runoff thresholds in grassed (GA) and agro-forested (AgF) hillslopes and on testing of suitability of network theory for explaining the hillslope hydrological processes. After analyzing 22 rainfall events of AgF and 37 events from GA hillslope, our findings highlight that the runoff generation is associated with the peak rainfall intensity, initial soil moisture conditions and rainfall duration. The runoff thresholds of GA hillslope are higher than the AgF hillslope due to its surface resistance and presence of extremely low soil hydraulic conductivity (<6 mm/h) at the upslope location. Further, from network analysis, we found that the network's topology was highly dependent on event characteristics and soil hydraulic conductivity. The strong network connectivity was observed between runoff and soil moisture nodes during high runoff generating events as the runoff and infiltration take place simultaneously. However, during the low runoff generating events, a weak connection was observed between runoff and soil moisture nodes as most rainwater infiltrates over the hillslope. The outcomes of this study indicate the usefulness of network theory in hillslope-scale runoff process understanding to identify the significant connections.