DRAI-AHP VULNERABILITY ASSESSMENT OF THE QUATERNARY AQUIFER IN THE KETA STRIP

ABSTRACT The Quaternary aquifer in the Keta basin is a source of fresh water to the over 130,000 inhabitants living along the Keta Strip. Groundwater quality is deteriorating due to pollution from mainly agriculture, sewage, septic tanks and solid waste disposal in the area. This research was undertaken to assess the intrinsic vulnerability to pollution of the aquifer. The DRASTIC methodology was applied to delineate zones vulnerable to pollution. Data was gathered on the seven DRASTIC parameters which include static water levels, soil sampling to determine soil media and aquifer media types, hydraulic conductivity, elevations and water quality. Using GIS methods the data was prepared, processed and analysed. Interpolation techniques were used to generate raster layers of the parameters. The seven raster maps were combined in a GIS environment to produce a composite vulnerability map. The map was reclassified into five classes namely, very low, low, moderate, high and very high pollution potential. The resulting map shows the relative intrinsic vulnerability of the unconfined aquifer to pollution. The map was tested by plotting on it April, 2004 nitrate concentration data. Statistical correlation between nitrate values and vulnerability class values extracted from the nitrate points yielded a low Spearman rho correlation coefficient of 36.3%. The model was calibrated using statistical correlations between each of the seven parameters and nitrate values. Soil media, topography and hydraulic conductivity parameters were found not significantly correlated to nitrate pollution and were eliminated from the model parameters leaving depth to water (D), recharge (R), aquifer media (A) and influence of the vadose media (I) to form the acronym DRAI. The weights of the parameters were revised based on the seven correlation coefficients. The rating values were also revised based on the mean nitrate values for each rating class of the original DRASTIC. The model accuracy was tested by applying correlation to the vulnerability classes and nitrate concentration data. A correlation coefficient of 62 % was obtained. The Analytic Hierarchy Process (AHP) was used to calibrate the model by determining a different set of parameter weights. The model accuracy improved to 66% after correlation between nitrate values and intrinsic vulnerability classes. Based on the correlation coefficient the AHP model was selected for the study area. The AHP optimized intrinsic vulnerability map shows that 10.6%, 26.8%, 30.1%,22.6% and 9.9% of the study area have very low, low, moderate, high and very high vulnerability to groundwater pollution respectively. Nitrate concentration values sampled in July, 2003 was used to evaluate validation of the AHP model. The nitrate values were plotted on the AHP vulnerability map and statistical correlation tests between nitrate values and vulnerability values yielded a correlation coefficient of 75%, confirming a close agreement between model values and actual nitrate pollution on the ground. Anloga and Woe area where agriculture is predominant have the lowest intrinsic vulnerability. Dzelukope and Keta which are heavily populated residential areas have very high groundwater pollution potential. The validation evaluation shows that the research results are realistic and representative of the actual groundwater pollution on the ground. Therefore, the DRAI model is applicable in the study area.