Poor agricultural practices among small-scale sub-Saharan African farmers can lead to soil erosion and reduce agricultural productivity. However, information on such practices is normally not well documented, making it challenging to design future mitigation strategies. We conducted a fine-scale agricultural survey on 200 farm households within the transboundary Sio Malaba Malakisi River Basin (SMMRB) between Kenya and Uganda to quantify the frequency and type of soil conservation practices (SWCPs) implemented. Information on farm sizes, ownership, crops grown, soil fertility, soil erosion, soil water conservation practices, and the decision-making processes was collected. Descriptive and chi-squared statistics were used to present trends in land use, decision-making processes and the extent of adoption of SWCPs, as well as to analyse the relationship between the SWCPs and the farmers’ perceptions on soil erosion. The region showed highly fragmented farms (mean area: 0.6 ha), primarily practising rain-fed subsistence farming. The principal decision-makers of each farm were mainly (63%) male. Various farmers (28%) lacked soil and water conservation practices (SWCPs). However, most farmers (35%) implemented one type of soil and water conservation practice, while 37% practised a combination of two to five soil and water conservation practices. Extensive soil and water conservation practices such as intercropping were widely practised as they were more affordable than intensive measures. Results on the farmers’ perceptions on soil erosion showed that most farmers in the SMMRB reported soil erosion (60%) and even more (92%) reported to have experienced a loss of soil fertility over the last 5 years. There was a significant positive correlation (X2 (2, n = 198) = 92.8, p = < 0.001) between the perception of soil erosion and the perception of the change in soil fertility, suggesting that reducing soil erosion could result in a reduction in the loss of soil fertility. Thus, there is still a need for strategies and measures to address the soil erosion risks currently faced by Sio Malaba Malakisi River Basin farmers. This study is a baseline study that shows the importance of farmers’ perceptions on the practice of soil and water conservation measures in the Sio Malaba Malakisi River Basin and therefore becomes an important avenue for improving the currently practised soil and water conservation measures as well as developing adoption programs as well as future studies that combine scientific and farmers’ perception/knowledge for sustainable agriculture. Further research into the efficiency of currently adopted SWCPs as well as the extent of the farmers’ knowledge and the accuracy of their perceptions is recommended.

Changes in land use/cover are among the most important anthropogenic transformation on the physical environment affecting proper functioning of the earth system. Hitherto, land characterization has often been studied using archived satellite data products to understandd trends in space and time. However, due to future uncertainties in land use change in developing countries and the associated impacts on the physical environment, there is need to model these changes at a local scale. A modelling framework to simulate empirically quantified relations between land use and its driving factors was used in the Sio-Malaba-Malakisi catchment between Kenya and Uganda. Changes for the catchment were simulated for a period of 30 years (2017 – 2047) using model parameters that define location characteristics, spatial policies, area restrictions, land use demand and conversion elasticity settings. Elevation, slope, population density, soil organic carbon, soil CEC and precipitation were potential factors selected to evaluate the suitability of devoting a grid cell to a land use type using a stepwise regression model. The scenarios evaluated include first growth, slow growth and an urbanization scenario. The high ROC value in all statistical tests (>0.72) indicated that the spatial distribution of some land use types in the basin could be explained by the selected driving variables. In a fast growth scenario (under policy restriction), areas under open soil and shrubland would be converted to cropland when demand for cash crop goes up in the region. Areas under open trees and marshland outside protected zones, would be converted to agricultural land while barren land with rock outcrops would remain largely unchanged over the period. In a slow growth scenario, expansion of the area under cropland would follow historical trend at 1.25% growth per annum. Marshland areas unsuitable for agricultural expansion are projected to remain the same. In an urbanization scenario, built-up areas would increase steadily at >1% per annum especially in areas earmarked for infrastructural development. In all the scenarios explored, topography, precipitation, soil characteristics and population density were identified as the key drivers of land use change. Results of this study would enhance the understanding of the complexities in projecting future land cover changes and provide baseline data for supporting ongoing soil and land management programs in a data scarce area.Key words: Land use change; CLUE-S model; Scenario analysis; Sio-Malaba-Malakisi catchment; Transboundary basin

Flooding associated with rising inland lake levels represents a distinct and understudied hazard compared to conventional river or coastal flooding. Unlike riverine floods, lake inundation develops gradually and persists over extended periods, yet integrated flood risk assessments specifically addressing this phenomenon remain scarce. This study uses Lake Baringo in Kenya’s Rift Valley as a case study, a lake that has experienced exceptional water level rises of nearly 10 m since 2010, to assess flood risk through the integration of scenario-based flood hazard modelling, spatial exposure assessment, and household vulnerability analysis. Three lake level scenarios were developed, spanning from maximum observed conditions to a worst-case threshold at the lake’s sill point. Results show that under the worst-case scenario, the lake area could expand by approximately 64%, potentially exposing nearly 18,000 people and 156 km of road infrastructure. A Flood Vulnerability Index (FVI) ranging from 0.58 to 0.73 indicated consistently high socioeconomic vulnerability across all shoreline communities, with vulnerability accounting for 42–53% of overall flood risk across scenarios. Flood risk was highest along the flat southern shoreline, where topography amplifies lateral inundation extent. The findings highlight the importance of integrating lake level dynamics into spatial planning and flood risk management, and the need for improved monitoring, livelihood diversification, and institutional coordination in semi-arid inland lake regions facing increasing hydro-climatic variability.