Stochastic structures between quantity and quality responses of rainfall-runoff at an upland agricultural watershed

The causality between hydrologic parameters and water quality variations has been studied for a small upland agricultural watershed using a time-series analysis approach. The stage at the outlet of the watershed, which is a proxy for the outflow, and the average saturation deficit, which is a surrogate for the groundwater level, were used as hydrologic variables. The stages were measured at a triangular weir at the watershed outlet. The saturation deficits were obtained by means of a rainfall-runoff simulation model for agricultural upland watersheds (an extended version of TOPMODEL). Sequential water quality data were also collected at the watershed outlet. A systematic procedure for testing the causality between the hydrologic and water quality time-series is presented and used to investigate possible relationships between the two systems. The stochastic structure of a causality between hydrologic and water quality systems is shown. This stochastic structure is associated not only with the transport pathways (surface or subsurface) of each water quality parameter but also with the characteristics of the rainfall events.