SPATIAL STATISTICS: A TOOL FOR SPATIAL ANALYSIS OF GEOCONTAMINANTS IN GROUNDWATER

Fluoride contamination in groundwater is growing concern worldwide, affecting more than 200 million people from 25 nations. Several natural and anthropogenic sources contribute to the geochemical cycling of fluorine in the environment. The geogenic fluoride mobilization in the environment is associated with many mineral deposits containing fluoride-bearing minerals, weathering, dissolution and other pedogenic processes such as volcanic eruptions. Fluoride-rich water sources and associated health effects are widely spread in many regions of Tanzania. Even though the fluoride problem is well-recognized in Tanzania, the lack of systematic assessment, databasing and mapping at regional level has made it very difficult to gain a broad overview of the precise locations and scale of the problem. This study aims at providing improved generic understanding of fluoride distribution and controls in high-fluoride aquifers within Great Rift Valley regions in northern Tanzania. Spatial statistics techniques were used to assess spatial dependence of fluoride concentrations measured in drinking water sources. Fluoride concentrations from 1,182 water samples indicate sufficient positive autocorrelation signifying that similar values are close to each (Global Moran’s I statistic=0.77, p-value=0.001).Similarly, local Moran’s I statistical tests indicated that fluoride concentrations does not occur at random instead it forms spatial patterns, some of which were statistical significant at 5% significance level. Spatial statistical modeling of causes indicated that fluoride occurrence and mobility in northern part of Tanzania is controlled by geology, landform evolution and human activities.