HR: 09:30h
AN: H11A-07    [PDF]
TI: Evaluating the Influence of Spatial Processes and Information on Local Moisture-Outflow Relations: Characterizing the Scale of Interaction Using Soil Moisture and Precipitation Observations
AU: * Saleem, J A
EM: jsaleem@bu.edu
AF: Boston University, Department of Geography 675 Commonwealth Ave., Boston, MA 02215 United States
AU: Salvucci, G D
EM: gdsalvuc@bu.edu
AF: Boston University, Dept. of Geography and Dept. of Earth Sciences 675 Commonwealth Ave., Boston, MA 02215
AB: The issue of scale in land surface hydrology is often important. Components of root zone outflow (evapotranspiration, drainage, and runoff processes) are dependent on soil moisture. In some cases, this dependence can be reasonably described at the point scale (e.g. the Darcy and Richards equations). However, at larger scales, these interrelationships become increasingly complex and uncertain. Small-scale processes are one of many factors that may influence large-scale behavior. In modeling or monitoring frameworks, the presence of subgrid heterogeneity in landcover, soil parameters or forcings, non-linear flux behavior, and lateral flows or processes all can have substantial impacts on large scale behavior. A successful hydrologic model must balance the importance of these different factors with the need for computational efficiency and tractability. For example, one can conceive of two possible models for soil moisture - outflow relationships at a given location. A simpler model is an "independent columns" approach; i.e. outflow can be completely represented as a function of the soil moisture at that location only. However, in many cases this is a not valid model; lateral interactions and flow among sites may exist and influence the components of outflow at measured point, and/or large-scale phenomena may affect outflow in a way that is not captured by a locally independent model. In such cases, some sort of spatial aspect must be incorporated and addressed if moisture relations are to be successfully described, predicted or aggregated. Here we use a non-linear water balance parameterization, fit to observational precipitation and soil moisture data. The parameters are optimized by exploiting the equilibrium tendency of soil moisture time series. Spatial information is introduced through two parameters that modify the drainage and evaporation functions, and the importance of these parameters to the performance of the model is assessed by statistically evaluating their effect on the stationarity of the modeled soil moisture trace through entropy estimates; thereby determining whether lateral flows or processes are contributing to the observed behavior. An example with data from the Illinois Climate Network is presented.
DE: 1866 Soil moisture
DE: 1899 General or miscellaneous
SC: Hydrology [H]
MN: 2003 Fall Meeting