Influence of soil textural properties on hydrologic fluxes in the Mississippi river basin.

Hydrological simulation of states and fluxes represents a viable approach for predicting the occurrence of extreme events and diagnosing their underlying causes. In recent decades, databases of soil textural properties have been published that provide a basis for estimating hydrologic model parameters. The aim of this study is to quantify the impact of soil textural assumptions on hydrologic states and fluxes: an often overlooked source of hydrologic modelling uncertainty. Mesoscale hydrologic model simulations driven by the digital general soil map of the USA (STATSGO2) soil database (1 : 250 000) were compared with those using the Food and Agriculture Organization (FAO)-based harmonized world soil database (1 : 5 000 000) over the Mississippi river basin. Annual and seasonal differences in modelled water fluxes and states were explored, in addition to major historical flood and drought events during 1988, 1993 and 2012. Model simulations using STATSGO2 consistently led to greater separation between fast and slow runoff responses (by approximately 24-39%) and, generally, more extreme responses to historic flood and drought events. Simulated soil moisture and runoff from the coarser harmonized world soil database produced greater persistence by roughly 0.10 autocorrelation points at 1-, 3- and 6-month lags, due to larger active water storage capacity. Overall, the choice of soil database altered the partitioning of precipitation between evapotranspiration and runoff, and affected the correlation structure between forcing and modelled fluxes by up to 0.2 points. As compared with other decisions needed to make hydrologic predictions, this analysis demonstrates that choice of soil textural properties for a single model can be an appreciable source of uncertainty and therefore warrants careful consideration. Copyright © 2015 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]