Micro-Catchment Typology for Integrated Catchment Management

Catchment management plans in data-scarce regions usually treat erosion, flooding, runoff and water harvesting separately, so interventions are prioritised hazard by hazard rather than by the combined condition of each land unit. This study developed a multivariate typology of 163 micro-catchments in the 1,641.7 km2 Idah Catchment, north-central Nigeria. Nine indicators derived from a 12.5 m digital elevation model and open global datasets were reduced with principal component analysis and grouped with k-means clustering. The first component (57.6% of variance) described a gradient from wooded plateau units to cultivated, flood-prone lowland units. Four management zones were identified: wooded plateau headwaters (50.1% of the area), steep, flashy plateau margins (21.7%), cultivated erosion–runoff hotspots (10.4%) and flood-prone floodplain (17.8%). The zones agreed with Ward clustering (adjusted Rand index 0.69), were reproduced by a random forest with 92% out-of-bag accuracy and were spatially coherent (65% of neighbouring units shared a zone, against 32% expected by chance). The hotspot and floodplain zones occupied 28.2% of the area but generated 54.9% of the soil loss and 58.3% of the 25-year runoff, contained 26 of the 28 significant soil-loss hotspots and had the highest groundwater recharge potential. Scenario analysis indicated that conservation agriculture with contour bunds on cropland in these two zones would reduce soil loss by 31.8%, whereas clearing 20% of headwater tree cover would increase it by 89.6%. Check dams outside the floodplain could retain up to 18.7% of the 25-year runoff. Annual rainfall declined significantly in the lowland zones. The typology concentrates management effort where hazards coincide and offers a transferable framework for integrated planning in ungauged tropical catchments.