Modeling Nitrate Exporting Patterns during Storm Events for a Semi-arid Mountain Watershed

Abstract:

In steep semi-arid areas like coastal Southern California, most nitrate export occurs during storm events. Observed data from the Santa Barbara Long term ecological research site shows that concentration and concentration-discharge relationships vary both throughout storm events and seasonally. Nitrate sources can be inferred from these data by comparing the arriving time of the nitrate concentration peak and that of the stream flow peak. For example, for small storm events in September or October after the long sever summer drought, nitrate concentration peak several hours earlier than stream flow peak, and may indicates the nitrate flux is from near-stream riparian areas. In contrast, storm events in the winter or early spring after a series of precipitation events tend to produce the nitrate concentration peaks several hours later than the stream flow peak, and indicates the nitrate flux is from hydrologically connected upslopes area. To model these dynamics, hydrologic models must be able to capture temporal shifts in hydrologically connected source areas as well as seasonal shifts in nitrogen availability associated with ecosystem biogeochemical cycling. We assess the ability of Regional Hydro-Ecologic Simulation System (RHESSys), to reproduce the seasonal/multi-yearly nitrate’s export patterns and compare the success of different model implementations. Specifically we compare a new developed hourly time-step implementation versus a daily time-step model and a model implementation where lateral routing is based on the “fill and spill” hypothesis versus the more traditional approach based on subsurface transmissivity functions. We also compare behavior of different parameterizations of the 'fill and spill' approach, including comparing uniform and spatially heterogeneous storage capacities. Results highlight the importance of both the hourly time-step and the threshold based response of a “fill and spill” connectivity model.