Observational analyses of hydrologic scaling: The roles of heterogeneity and non-local interaction as inferred from soil moisture and precipitation data.

摘要

The relationship between soil moisture and outflow (evapotranspiration, drainage, and runoff) is fundamental to many hydrologic applications (e.g. water balance monitoring and modeling). However, characterizing it at any level above a point scale becomes increasingly complex and uncertain due to various factors, including climate and surface heterogeneity, the presence of lateral transports, and the mismatches between the spatial resolutions of various measurements and models. Using sparsely and infrequently measured soil moisture data, and a technique using conditional averaging of precipitation to estimate outflow, we investigated the factors of heterogeneity and non-local interactions in scaling this relationship from points to larger areas.; Methods were developed to characterize and account for surface parameter heterogeneity by non-dimensionalizing the dependent variable (soil moisture) using observed distribution properties. This was successful in reducing variations in outflow response for similar forcing.; To investigate the presence of non-local interactions, locations are modeled as a set of independent columns and the moisture-outflow relationships are aggregated in such a way as to account for heterogeneity; this estimate is compared statistically to the largescale response estimated from aggregate data. Significant differences would suggest the system is not well represented by the independence assumption, i.e. local outflow is dependent on local moisture and also is independently influenced by large-scale moisture.; We applied these methods to data from three systems - a hillslope, an intermediate-sized watershed, and the state of Illinois, and found that heterogeneity coupled with nonlinearity of the governing processes significantly affected scaling in all three. After accounting for these, significant differences remained between the aggregated point-scale estimates and the large-scale response. This difference is attributable to non-local influences on the local systems; in each area studied, the effect is to decrease (increase) local outflow during large-scale dry (wet) anomalies; evidence in Illinois points to a possible pathway through wind speed. The apparent effect of the interactions is to prolong small spatial anomalies of moisture, and at the same time to decrease the temporal variability of the large-scale system. The results here suggest possible common dynamic and scaling effects in water balance at various scales.