Download 2_87

УДК 504.064
ESTIMATING SOIL WATER RETENTION USING SOIL COMPONENT
ADDITIVITY MODEL AND PEDOTRANSFER FUNCTIONS
А. Zeiliguer, O. Ermolaeva, T. Kremleva
Moscow State University of Environmental Engineering, Moscow, Russia
Y. Pachepsky
USDA-ARS Hydrology Laboratory, Beltsville, U.S.A.
W.Rawls
USDA-ARS Hydrology and Remote Sensing Laboratory, Beltsville, U.S.A
А. Nemes
Research Institute for Soil Science and Agricultural Chemistry of the Hungarian Academy of
Sciences, Budapest, Hungary
H. Wösten
Alterra Green World Research, Wageningen, The Netherlands
Actual soil water retention (SWR) measurements are relatively time-consuming, and
become impractical when soil hydraulic estimates are needed for spatial hydrological modeling
of large areas. During last two decencies, some special models to estimate SWR from readily
available soil property data have been proposed to overcome this obstacle. From theoretical
point of view, these models can be divided into two groups. Models of first group are based on
empirical approach of transfer function use (called pedotransfer function) linking or water
content of specific water pressure head values, or parameter values of some equation used for
SWR fitting. Second group is composed of physically-based models that are use some
hypothesis to describe a distribution of soil component within a soil space. At this paper two
types of SWR estimating models representing both groups are compared by using data stored
into two large databases UNSODA and HYPRES.
Models of first group were represented by Models of second group were represented by additivity model based on the hypothesis that
soil water retention may be estimated as an additive function obtained by summing up a priory
known SWR of pore subspaces. These soil pore spaces are associated with soil textural and/or
structural fractional components, represented by particle and aggregate size distributions, as well
as organic matter. This hypothesis has been used in the additivity model developing with the
following assumptions.
(a)
The additivity is applicable to gravimetric water contents, soil water potentials
ranging from -1500kPa to 0 kPa.
(b)
A priory known water retention associated with textural fractions has been
measured in packed samples consisting exclusively of these fraction's particles.
(c)
Water retention of textural fractions contributes in soil water retention in
proportion to relative volumes of pore subspaces virtually associated (by additivity model) to
each fraction.
(d)
Water retention of the clay textural fraction depends on predominant clay
minerals.
(e)
Water retention of soil organic matter depends on organic matter.
Transfer functions and the additivity model and were tested with 550 soil samples from the
international database UNSODA and 2667 soil samples from the European database HYPRES
containing all textural soil classes after USDA soil texture classification.
The root mean square errors (RMSEs) of the volumetric water content estimates for
UNSODA vary from 0.021 m3m-3 for coarse sandy loam to 0.075 m3m-3 for sandy clay.
Obtained RMSEs are at the lower end of the RMSE range for regression-based water retention
estimates found in literature. Including retention estimates of organic matter significantly
improved RMSEs. The attained accuracy warrants testing the 'additivity' model with additional
soil data and improving this model to accommodate various types of soil structure