Geostatistical Analysis of Contamination of Soils in an Urban Area in Ghana

Urbanization remains one of the unique predominant factors which is linked to the destruction of urban environment and its associated cases of soil contamination by heavy metals through the natural and anthropogenic activities. These activities are important sources of toxic heavy metals such as arsenic (As), cadmium ( Cd ), chromium ( Cr ), copper $(\mathrm{Cu})$, iron $(\mathrm{Fe})$ manganese $(\mathrm{Mn})$, and lead $(\mathrm{Pb})$, nickle $(\mathrm{Ni})$ and zinc $(\mathrm{Zn})$. Often, these heavy metals lead to increased levels in some areas due to the impact of atmospheric deposition caused by their proximity to industrial plants or the indiscriminately burning of substances. Information gathered on potentially hazardous levels of these heavy metals in soils lead to establish serious health and urban agriculture implications. However, characterisation of spatial variations of soil contamination by heavy metals in Ghana is limited. Kumasi is a Metropolitan city in Ghana, West Africa and is challenged with the current deterioration of soil quality due to rapid economic development and other activities such "Galamsey", illegal mining operations. The paper seeks to use multivariate geostatistical analysis to assess the spatial distribution of heavy metals in soils and the potential risk associated with ingestion of sources of soil contamination in the Metropolis. Geostatistical tools have the ability to detect changes in correlation structure and how a good knowledge of the study area can help to explain the different scales of variation detected. To achieve this task, point referenced data on heavy metals measured from topsoil samples in a previous study, were collected at various locations. Linear models of regionalisation and coregionalisation were fitted to all experimental semivariograms to describe the spatial dependence between the topsoil heavy metals at different spatial scales, which led to ordinary and cokriging at unsampled locations and production of risk maps of these heavy metals. Results obtained from both the linear regionalisation and coregionalization of the semivariogram models showed strong spatial dependence with range of correlations ranging up to 300 meters. The risk maps indicated strong spatial heterogeneity for almost all the heavy metals with highly risk of the soil contamination found close to areas with commercial and industrial activities. Hence, ongoing pollution interventions should be geared towards these areas for efficient management of soil contamination to avert further pollution in the Metropolis.