Dynamics of surface runoff and soil loss from a toposequence under varied land use practices in Rwizi catchment, Lake Victoria Basin

Abstract

In this study, we quantified surface runoff and soil loss along a toposequence considering four land use systems and three topographic segment positions in Rwizi catchment, Lake Victoria Basin (LVB). The land use systems were grassland dominated by a local species known as “Omugugu” (<i>Cyperus</i> spp.), tree plantation of pine (<i>Pinus sabiniana</i>) and eucalyptus (<i>Eucalyptus globus</i>), mulched banana (<i>Musa Paradisiaca</i>), and unmulched banana, topographically located on the foot slope, midslope, and summit. Runoff and soil loss magnitude and trends under varied rainfall regimes were measured using closed runoff plots measuring 2 × 20 m (40 m²). A split plot experimental design was adopted for the study. In total, 36 runoff plots were installed considering the land use and topographic factors with three replications. Observed monthly rainfall values varied from 9.3 to 167 mm (75 ± 54 mm). Annually, the registered rainfall depth was 680 mm. The annual rainfall depth was below the long-term annual mean for the area. Nearly all rainfall events with a depth greater than 10 mm contributed relatively heavy runoff, while rainfall events less than 10 mm in general caused minor surface runoff at the sites. Annual runoff within the catchment at the studied sites varied from 42 to 411 m³ ha⁻¹ y⁻¹. The average observed cross-site annual runoff was 151 ± 95 m³ ha⁻¹ y⁻¹. Results show that annual runoff was in the order of tree plantation &gt; unmulched banana sites &gt; grassland = mulched banana. Soil losses were greatest (1.5 t ha⁻¹ y⁻¹) and lowest (0.8 t ha⁻¹ y⁻¹) on unmulched banana and mulched banana sites, respectively. Annual soil loss showed an increase with topographic slope segment position. Average annual soil loss on the upper slope position facets (1.7 t ha⁻¹ y⁻¹) were 2.4 folds higher than those observed on the footslope position facets (0.7 t ha⁻¹ y⁻¹), giving a percentage difference of 83%. We did not detect a statistical difference in runoff (<i>p</i> &gt; 0.05) due to either land use system or toposequence position. However, statistical differences were observed in soil loss (<i>p</i> &lt; 0.05) for land use system and toposequence position. While the observed annual soil loss magnitude is within the tolerable thresholds, a sustained maintenance of best land use practices is still critical in maintaining the delicate balance on the fragile hillslopes in the study area.