Effect of soil variability, within and between soil taxonomic units, on simulated nitrate leaching under arable farming, New Zealand

Abstract

A Monte Carlo approach was used to predict the effect of soil variability on nitrate leaching from 8 soil series, encompassing a wide range of drainage, texture, and age of soil development characteristics. A database of soil physical properties consisting of a minimum of 9 profiles per soil series was used to derive correlated probability distribution functions of key soil properties. The distribution functions were then used for random sampling to derive input soil-data for the GLEAMS (Groundwater Loading Effects of Agricultural Management Systems) simulation model. Variability in soil properties found within single soil taxonomic units (depth/stoniness phases of soil series) resulted in an appreciable range of predicted leaching of nitrate. Leaching from soils with greatest variability (generally shallow and stony phases) had an inter-quartile range of predicted leaching of up to 19 kg N/ha in 1992. Sensitivity analysis indicated that organic matter content, depth to the base of the upper 2 horizons, and available water storage were important drivers of variability within soil taxonomic units. Despite wide variation within soil taxonomic units, there were still clear differences between them. Effective soil depth accounted for most of this variance, which was attributed to differences in total profile available water storage. Soil drainage had some effect on risk of leaching. This effect would probably have been greater if water table effects had been accounted for. Soil series distinctions related to soil age had no significant effect of leaching risk. These results indicate that nitrogen leaching risk assessment using GLEAMS is dependent upon soil maps with accurate identification of soil depth/stoniness phases and organic matter content.