Spatial assessment of termites interaction with groundwater potential conditioning parameters in Keffi, Nigeria

© 2019 Elsevier B.V. Termite mounds are traditionally presumed to be good indicators of groundwater in places they inhabit but this hypothesis is yet to be scientifically substantiated. To confirm this assertion, it is expected that termite mounds would have strong correlations with groundwater conditioning parameters (GCPs). In this study, termite mounds distribution covering an area of about 156 km2 were mapped and their structural characteristics documented with the aim of examining their relationships with twelve (12) chosen GCPs. Other specific objectives were to identify specific mound types with affinity to groundwater and to produce a groundwater potential map of the study area. To achieve this, 12 GCPs including geology, drainage density, lineament density, lineament intersection density, land use/land cover, topographic wetness index (TWI), normalized difference vegetation index (NDVI), slope, elevation, plan curvature, static water level and groundwater level fluctuation were extracted from relevant sources. Frequency ratio (FR) and Spearman's rank correlation were used to find relationships and direction of such relationships. The result revealed a consistent agreement between FR and Spearman's rank correlation that tall (≥1.8 m) and Cathedral designed mounds are good indicators of groundwater. Further, the groundwater potential map produced from the Random Forest (RF) model via Correlation-based Feature Selection (CFS) using best-first algorithm depicted an erratic nature of groundwater distribution in the study area. This was then classified using natural break into very-high, high, moderate, low and very low potential classes and area under curve (AUC) of the receiver operating characteristics (ROC) showed an 86.5% validity of the model. About 75% of mapped termite mounds fell within the very-high to moderate potential classes thereby suggesting that although tall and cathedral mounds in particular showed good correlations with a number of GCPs, high mound density in a locality is also an indication of good groundwater potential