SEGH Articles

Beyond mapping: new applications of GIS in environmental geochemistry and health

21 April 2011
GIS has been widely used to produce professional maps. However, the potential applications of GIS are far beyond mapping.

Dr. Chaosheng Zhang is head of GIS Centre, Ryan Institute at National University of Ireland, Galway. His research interest focuses on spatial analyses of environmental variables.

One of the most widely recognized functions of a geographic information system (GIS) is computerized mapping, which has helped to greatly improve the efficiency and quality of map production in environmental geochemistry and health. However, the potential applications of GIS are far beyond mapping. This short article explains some basic concepts of GIS spatial analyses, with an aim to promote their applications.

The spatial information contained in geochemical databases is presented in the form of their corresponding longitudes-latitudes or X-Y coordinates. A GIS is designed to handle such spatial information, and can easily produce maps in the format of symbols, contours and filled surfaces. Such mapping functions can aid "visual" interpretation for geochemical features. One of the examples is the soil geochemical atlas of Ireland. This map shows soil organic carbon distribution in Ireland, produced using a GIS and a spatial interpolation method of trans-Gaussian kriging based on a total of 1310 samples. Elevated SOC values are located in western Ireland where peat is widely distributed. The relatively low SOC values in south-eastern Ireland are related to the intensive agricultural activities.


New and advanced applications of GIS analyses in environmental geochemistry include: spatial outlier identification and hotspot analysis; spatial autocorrelation; spatial interpolation; and spatial variation. It should be noted that many spatial analysis techniques are still under development.

Spatial Outlier Identification and Hotspot Analysis: It is common that outliers exist in environmental geochemical databases. They originate from rare processes such as mineralization or human pollution, as well as possible errors during sampling and laboratory processes. They need to be identified and properly handled. They may contain useful information for identification of potentially contaminated land and clusters of diseases. The local index of spatial association (LISA) of local Moran's I has been found useful in spatial outlier identification and hotspot analysis.

Spatial Autocorrelation: Environmental geochemical data have the feature of spatial autocorrelation: Neighbouring data tend to have similar values. This feature is theorised by Tobler's "First Law of Geography": Everything is related to everything else, but near things are more related than distant things. Spatial autocorrelation can be quantified by Moran' I index and the variogram in geostatistics.

Spatial Interpolation: It is virtually impossible for us to have all the materials from an area of interest analysed. For example, in an urban soil geochemical survey, local government would not allow researchers to take all the soils to their labs for analyses. Meanwhile, the budget for laboratory analyses is also often limited. To get the full picture of a study area, spatial interpolation is needed to estimate the values at un-sampled locations based on analyses of sampled locations, and then spatial distribution maps can be produced. There are many spatial interpolation techniques available, and the most popularly used ones are IDW (inverse distance weighted) and kriging in geostatistics.

Spatial Variation: Spatial heterogeneity in environmental geochemistry exists at different spatial scales: ranging from global geochemical variation to variation within a small land. Even when the environmental factors are the same, e.g., same rock type, soil type, vegetation and land use, micro-scale spatial variation can still be observed. One of the solutions to spatial variation in environmental geochemical investigations is to use composite samples containing multiple subsamples from a small area. The description of spatial variation is traditionally based on "visualization" of maps. However, local or neighbourhood statistics can be used to quantify spatial variation, especially when the sampling density is high.

I hope the summarised information provided here helps to inspire creative thinking and interesting applications in environmental geochemistry and health.

Contact details: Dr. Chaosheng Zhang, School of Geography and Archaeology, National University of Ireland, Galway, Ireland.