Scale Dependence in Water Balance Sensitivity to Soil Moisture

Trustees Of Boston University, Boston

Investigators

Abstract

0074045 Salvucci Many questions in hydrology have as a common theme the inter-relation of water fluxes and storage at large scales. The issue arises for simple reasons: Transport processes may be reasonably well understood and predictable at small scales, but the scales at which transports become most relevant (either as resources/hazards or with respect to atmosphere interaction) are much larger. At large scales critical parameters (soil, vegetation, and topography) and forcings (radiation, precipitation, wind) are spatially heterogeneous, and transport mechanisms are not only vertical, but lateral (e.g., groundwater redistribution, boundary layer advection). Thus even the simplest question, i.e. "what controls the partitioning of precipitation into evapotranspiration, runoff, drainage and storage change", becomes difficult to answer at large scale. Here we propose a new and simple approach for using direct measurement to address the related question: how does the sensitivity of water balance to soil moisture change with scale? The novelty of the proposed approach is the use of monthly precipitation (P) measurements as a surrogate measure of total outflow from the root zone (evapotranspiration (ET), drainage (D), and runoff (R)). The basis of the method is that the changes in root zone moisture storage (S) over an interval, which make up the difference P-ET-D-R, are themselves uncorrelated with the mean storage over that interval. This lack of correlation is a direct result of the statistical stationary of root zone moisture. The utility of this property is that all of the covariance of ET, D and R with S can be estimated by measuring the covariance of P with S (or by calculating the precipitation rate conditionally averaged at given moisture levels, and taking derivatives). The stationarity condition thus imparts on conditionally averaged precipitation utility as a diagnostic measure of the sensitivity of combined evaporation and drainage/runoff fluxes to soil moisture. Because this simple methodology is based only on mass balance and conditional averaging, it can be applied at any area scale through simple aggregation of P and S data, and thus can be used to directly estimate both the form and the scale-dependence of moisture-water balance sensitivity. As such it can be used to quantify how variability in moisture due to lateral surface and groundwater redistribution processes and land cover/soil/meteorological forcing heterogeneity influence the large-scale sensitivity of water balance to soil moisture.