The uncertainty in climate change and high water demand pose pressure on the natural water resources supply. Not only does this require better understanding but also a call for immediate interventions, mitigation and adaptive measures. This study evaluates catchment water resources in the Luwombwa sub-catchment in Zambia through statistical analysis in the downscaling of past, present and future climatic variables from the CMIP6 climatic model. These variables are then integrated into WEAP - a semi-distributed hydrological and water evaluation model - to perform water demand and allocation scenario modelling. Mult-site calibration and validation were conducted on five selected micro-catchments within Luwombwa sub-catchment. The model performance was assessed usng the R², NSE and PBIAS as the objective functions. Satisfactory values of 92 % for R², 82 % for NSE and 6.9 % for PBIAS were achieved. This allowed for scenario modelling on water demand and allocation among competing users. Three future scenarios (2022–2050) were developed from the historical to baseline (1988–2022) and included state of water resources availability under climate change, expansion of irrigation area and impact of dam construction in the sub-catchment. The study reveals a decrease of 20 % in sub-catchment's water availability resulting from 9.3 % (equivalent to 4^oC) rise in maximum temperature and 4.5 % reduction in rainfall within the entire sub-catchment. This is especially under the persistence of SSP370 climate variability scenario projections downscaled from four GCM models by the year 2050. The study further revealed that the change point for anticipated future climate extremes is likely to occur between 2027 and 2030. The results are indicative of downward trends in streamflow under climate change and socioeconomic development leading to increase in water value and water scarcity. The insights from the study are critical to inform formulation of effective catchment water resources management strategies such as the development of management plans and adapation measures in the face of climate change and the needs for different stakeholders involvement.

This research implemented event-based hydrodynamic modelling techniques to characterise flash flooding in Nangoma catchment. The HEC-RAS 2D hydraulic model was calibrated and validated with acceptable overall statistical objective functions NSE of 84.6%-, R2 at 99% and PBIAS of -2.5%. Representative design hyetographs were derived from the hybrid clustering optimised for four clusters under the Inter-event Time Definition of three hours. The catchment flash flood threshold curve estimated the R2 of 0.96 confirming that the power-law model provided an excellent fit to the flood generating event population capturing 96% of the variance in average intensity as a function of duration across the full range of observed convective events especially under convective storms with 1.2 – 1.5 hours, triggered by corresponding rainfall amounts above 39 mm. To assess possible interventions, five scenarios were simulated for optimal NbS at the respective three monitoring sites (CP1-3) indicated that rainwater harvesting implemented on the main river channel, tributaries, and off-river storage improved water storage while reducing flooding risks by 51% on CP1, 30% on CP2, and 36% at the outlet on CP3. Therefore, such outputs from this integrated approach are key to providing valuable insights and actionable water-resilient nature-based solutions such as rainwater harvesting and enhanced catchment storage for communities facing recurring flash flood challenges.