Modelling soil water dynamics and evapotranspiration for heterogeneous surfaces of the steppe and forest-steppe zones on a regional scale

The main aim of the present work is to develop a mesoscale dynamical-stochastical model describing the processes of heat and water exchange at heterogeneous surfaces during the warm period of a year and combining both determinism of physical processes and stochastic spatial structure of the main influencing factors. The analysis of mechanism of formation of soil water storage and evapotranspiration has shown that to attain the aim it is necessary to take into account the spatial variability of precipitation, soil moisture and leaf area index. A first version of a dynamical-stochastical model which takes into account the spatial variability of the aforementioned factors on a regional scale is presented. The model allows one to calculate temporal variation of statistical parameters of spatial distribution of soil water storage (namely, area-averaged value, standard deviation and even function of distribution) using the two first statistical moments of initial soil water storage, current precipitation and current leaf area index. A comparison of simulated and observed values of area averaged water storage in the top 1 m soil layer and their spatial variability for the Kiev and Kursk regions situated in the forest-steppe zone of the Former Soviet Union has demonstrated satisfactory agreement. Analysis of the results obtained shows that taking into account the spatial variability allows one to: (1) reduce the value of root-mean-square deviation between observed and calculated area-average soil water storage from 20 to 14 mm; (2) obtain the temporal variation of spatial variability of soil water storage; (3) draw a conclusion that area-averaged evapotranspiration can be overestimated by approximately 7% without accounting for spatial heterogeneity. In addition, it was found that current spatial variability of soil water storage is mainly affected by current variability of precipitation and leaf area index rather than the variability of initial soil water storage. The main contribution (up to 70-75%) is made by precipitation in the regions under study.