Using neighbourhood statistics and GIS to quantify and visualize spatial variation in geochemical variables: An example using Ni concentrations in the topsoils of Northern Ireland

Abstract Spatial variation is a typical feature of geochemical variables, providing a challenge for sampling design and environmental monitoring. It is generally qualitatively but not quantitatively described using spatial distribution maps. In this study, the feasibility of quantifying spatial variation is investigated using neighbourhood statistics within a GIS environment, using, as an example, near-total Ni concentrations in the surface soils of Northern Ireland. A total of 6138 topsoil samples were collected at an average sampling density close to 1 sample per km 2 . At this sampling density it was possible to calculate neighbourhood statistics directly from the raw data. Neighbourhood statistics of local mean, local standard deviation and local coefficient of variation were calculated using window sizes of 3 km × 3 km, 6 km × 6 km, 9 km × 9 km, 12 km × 12 km, 24 km × 24 km and 48 km × 48 km and visualized using GIS mapping techniques. The results showed that the highest soil Ni concentrations were located in the northern part of Northern Ireland where basalt is the main rock type. Lowest soil Ni concentrations were found in the western region of the Province on schist and limestone geologies. The granite area in the south-eastern region of Northern Ireland also displayed low soil Ni values. In terms of assessing the degree of spatial variation, high local standard deviation values were found to be associated with high local mean values thereby limiting the usefulness of local standard deviation as an indicator of spatial variation. This effect did not occur when local coefficient of variation values were used in place of local standard deviation so the coefficient of variation values are recommended as a more appropriate indicator to quantify spatial variation. The strongest spatial variations were observed on the western edge of the basalt area along the boundary of the basalt–sandstone areas and the schist area. Within each rock type, spatial variations were relatively weak and this was most clearly demonstrated in the basalt area. As the window size used for calculation of neighbourhood statistics was increased, so too was the resulting smoothing effect which led to clearer patterns but with loss of detail in the spatial variation observed. Neighbourhood statistics, coupled to a GIS, were found to be an effective way of quantifying and visualizing spatial variation in environmental geochemistry.