Use of spatial and temporal probability in the study of soil-water-landscape relationships.

摘要

Old feedlots have been found to enhance the production and movement of nitrate nitrogen. The main objectives of the first part of the research were: (1) to describe the spatial variability in nitrate and organic matter concentrations, with depth, in an old feedlot area; and (2) to explain, based on the probability of the presence of perched water, the relationship between the distribution of surface organic matter and subsurface nitrate concentrations. Geostatistical analysis delineated an area around an old silo, high in soil organic matter, which was not always well related to the soil nitrate distribution. It appeared that the nitrate was being attenuated by landscape-induced denitrification in some areas. Non-parametric geostatistics described the spatially probable distribution of water saturated soil in the landscape, which explained the nitrate attenuation. The "probability of presence of perched water" was related to elevation (r = -0.84) and to soil map unit (greater probability for Loring than Memphis map units).; The temporal and spatial context must be taken into account in the study of soil moisture in the landscape. The objective of the second part of the research was to use a stochastic approach to spatially and temporally analyze soil moisture change within an experimental field, in order to explain highly variable leachate volumes collected by 20 lysimeters for more than 10 years. The study was performed on a 0.36 ha field, predominantly Maury silt loam (Typic Paleudalf). Soil moisture was measured in two (Oct 2001--Jan 2002), and one (Feb--May 2002) day intervals. Soil physical properties (bulk density and texture) and primary landscape characteristics (plan and profile curvature, slope and elevation) were related to measured moisture contents. The stochastic approach illuminated areas with a higher probability of maintaining greater soil moisture for a longer time. These areas were not randomly allocated in the landscape, but were related to soil and landscape properties. The stochastic approach showed promise in developing an understanding of soil moisture distributions/change in space and time. The change in soil moisture was strongly correlated to lysimeter leachate volume (r = 0.69).