Geostatistical and GIS analyses on soil organic carbon concentrations in grassland of southeastern Ireland from two different periods

Abstract

A total of 191 soil samples, taken in 1964 from grassland of southeastern Ireland, and 220 samples, taken in 1996 from the same area, were examined for soil organic carbon (SOC). Temporal and spatial changes in SOC concentration after the ca. 30-year interval were evaluated using conventional statistics, geostatistics and geographic information system (GIS) analyses. A lognormal distribution feature was observed for both data sets, and the results of a t-test showed that the difference between them was not significant. Spatial outliers were detected with the index of local Moran's I in order to obtain a robust variogram. Kriging was used for spatial prediction for both data sets using the same grid system. The difference between the two maps was determined using map algebraic functions and showed considerable spatial change. The analysis has enabled the delineation of two discrete sectors in the study area, which have clearly behaved differently: mean increase of SOC concentration in the eastern coastal sector was 30.2% compared to a mean decrease of 16.0% in the inland sector. Changes in land use or cultivation did not adequately account for the spatial difference. The division between the sectors appeared to coincide approximately with and to reflect both hill land and geological patterns. A satisfactory explanation for the difference has not been identified. Thus, a combination of geostatistics and GIS map algebra provides a useful tool for the examination of spatio-temporal changes in the environmental sciences and may detect features that are not discernible when only conventional statistics are used.