As traditional hydrologic measurements, stream stage and precipitation gauge observations often have the benefit of long records. Even though they have good value for calibrating conceptual models of basin hydrologic units, they are not adequate for understanding basic (spatial) physical processes such as rainfall-runoff transformation. Stream discharge estimated from stage measurements represents the aggregate outflow from the catchment draining at the measurement location. Processes governing runoff nevertheless vary considerably across the terrain. Hillslope recharge-discharge patterns, surface water--groundwater interaction, and the emergence of riparian zones are examples of processes that vary across the basin. Their dynamics are not evident in stream gauge hydrographs. In this paper a methodology is introduced that allows the extraction of some spatial information from the analysis of precipitation and discharge time series. The marginal distribution of available soil water storage across the basin is evident in the subtle and nonlinear covariations between storm precipitation and storm runoff volumes in the time series. The approach is nonparametric in that it does not assume nor require a conceptual hydrologic model. The methodology is introduced and tested using both synthetic and observed gauge records. A scalar and nondimensional measure of hydrologic complexity is introduced to characterize the spatial variability of basin rainfall-runoff response behavior.