The Binomial Logistic and Multiple Linear Regression-aided Mapping of Aquifer Sulphate Levels in the Landheer Area, Dadaab Sub-County

Abstract

Abstract-The Landheere centre is a newly unveiled settlement in the Dadaab subcounty, which is located some 6 to 9 kilometers away from the Laghdera flow course in Dagahaley, a refugee camp neighboring the Dadaab Axis zone, categorized as the central Merti Aquifer unit. The study was aimed at availing potable water to the community via Hydrogeological and Geophysical investigations. It was then decided that an aspect of water quality, sulphate levels, be also factored as waters with anomalous levels beyond 200mg/liter have been known to bear toxicity associated with health hazards for the nomadic pastoralist communities living in Garissa county. To predict the sulphate levels, secondary data of the Merti Aquifer was used to generate a predictive model based on Multiple Linear regression and Logistic regression. The original Merti hydrochemistry data was analysed using the correlation plot function in R and it gave a matrix table output showing the strongest predictors of sulphates, one of which was found to be the electrical conductivity, or EC. The EC of the areas closest to Landheer are well known. Dagahaley is one such area and has EC levels way below 900 mg/Liter, at the very worst. Since this EC may easily be measured in the field, alongside proposed site’s coordinates and aquifer depths, the Models were used to predict sulphate levels expected in the borehole yet-to-be-drilled, and both algorithms determined that Landheer area is suitable for development of a well, to the extent that it is suitable in terms of water quality parameters. The MLR and Logistic Regression are thus useful Statistical Techniques ideal for the Water Resources Assessment for Development in the Merti Aquifer.