Multivariate Chemometric Analysis Of Radon Concentration In Geothermal Fluids From Homa Mountain, High Background Radiation Area (Hbra

ABSTRACT The activity concentration of radon in geothermal waters from Homa mountain region of south-western Kenya has been measured. Homa mountain is one of the regions in Kenya with elevated background radiation. This region is covered with geothermal grass and has some cold and hot springs. The activity concentration of radon in spring waters, river waters and pond waters from (i) geothermally active HBRA (ii) non-geothermally active HBRA and (iii) non-geothermally active non-HBRA areas of this region was measured using Liquid Scintillation Counting (LSC) method. The average concentrations of radon in spring waters, river waters and pond waters were above USEPA contamination limit of 11 Bq/L. Multivariate chemometric techniques namely Principal Components Analysis (PCA), Hierarchical Cluster Analysis (HCA) and Soft Independent Modeling of Class Analogy (SIMCA) were applied on the activity concentration of radon in the geothermal water samples. Principal components analysis was used to deciphere patterns within the data set based on the activity concentrations of radon, pH and temperature. The score plots displayed distinct clusters in relation to the water source. Hierarchical cluster analysis was used to partition the radon concentrations into non-overlapping clusters; the dendrograms revealed that radon concentration in geothermal fluids form clusters which can be related to the water source as displayed by the PCA score plots. The formation of these clusters can be regarded to be controlled by geothermal activities in this region. Diagnostics of the geothermal potential in a typical high background radiation area based on the radon signatures was performed using SIMCA approach, in order to identify local models for geothermal waters from Homa mountain and to predict a probable class membership for new observations from the same region. Test samples were successfully classified into respective models after the Cooman’s plot with 64% efficiency. The modeling power of the activity concentration of radon in relation to pH and temperature in this study was more than 0.3. Therefore, these variables were important in constructing SIMCA models in this study. In order to increase this classification efficiency, other factors such as chemicals and isotopic components in geothermal waters need to be investigated and SIMCA models constructed on the basis of these chemicals and isotopic components and the activity concentration of radon in a geothermally active HBRA.