| dc.description.abstract | Estimating spatial variability of trace geothermal signatures is an important factor is evaluating the geothermal field potential. It is expected that a particular geothermal feature (hot spring, fumarole, geyser, etc.) exhibits unique 'marker' trace element/ and radiogenic signatures that may be used to prospect the field over large spatial dimension (for example mapping of passive sites) based on the results of a few measurements on the feature. Geostatistical methods provide means to study the heterogeneous nature of the 'marker' spatial distribution especially in an area of poor accessibility. The geothermal signatures in this study are trace elements in thermal water that uniquely characterize the Kerio Valley geothermal field, which lies in the mid Rift System of Kenya. The region has been associated with elevated background radiation to the extent that certain areas in have been characterized as high background radiation areas (HBRA) thereby providing a perfect setting to understand the relations between geothermal characterizing trace elements and naturally occurring radioactive material (NORM) signatures. X-Ray Fluorescence (XRF) analysed Sr, K, Rb, Br, Ca and Cr showed a strong positive correlation with water temperature and were considered for mapping their variability and prediction of unsampled areas. Variogram and Kriging analysis was performed using ArchMap 9.3. The results show that the spatial distribution of the sampling points is insufficient to map the whole area of interest. Various variogram models fit well for Sr, Rb and Cr. However, a fitting model for K, Br, and Ca could not be found. This is due to the unharmonized nature of the covariance between the sampled points. The kriging maps, which are a product of the variograms, capture the sense and importance of sampling design (DoE) in geostatistical modelling. This study therefore serves as a base for the design and systematic sampling approach for the sparse nature of the Kerio Valley goethermal signatures. Springs of elevated thermal gradient have been identified as the sampling points. Both water and soil (i.e. sediment) will be sample accordingly to 'markers' have shown a strong positive correlation with thermal gradient in water, and the soil has provided the same correlation pattern as the water. Therefore in cases of dried springs, soil samples can confidently provide a good model | en |
