Groundwater resources management has been a growing concern as the consequences of climate change and other anthropogenic forces have heavily impacted freshwater resources, especially in semi-arid regions. The adoptions of managed aquifer recharge (MAR) remains at its earliest stages of assessment in developing countries such as Lebanon. Regional conflicts and the current refugee status have amplified the need for accessible uncontaminated freshwater resources in the semi-arid Bekaa region. At a pilot scale, this thesis investigates the hydraulic properties of a well drilled in the Miocene alluvial consolidated deposits in the Litani Basin, Bekaa in Lebanon, which is devised to store surface runoff during the winter, to be utilized for agricultural irrigation purposes in the summer (Aquifer Storage and Recovery scheme). The small scale subsurface characterization of aquifer properties was done using grain size analysis of borehole cuttings and several pumping tests to assess transmissivity in the recharged aquifer at the borehole scale. In the latter case, water availability poses a challenge on sustainability of MAR. to answer this question, an integrated hydrological model was constructed and calibrated using MIKE SHE (DHI, 2017a) for the entire Upper Litani Basin catchment to simulate flow in steady state for years 1970 and 2010 based on existing water level data to evaluate the water balance in the area including lateral and return flow and recharge to groundwater) in comparison to pumping and River exchange. Transient flow was simulated to assess the degree of groundwater depletion, the availability of water resources and recharge during high flow periods, while accounting for the different hydrological components; namely climate, river, saturated and unsaturated zone, The model will serve as an integrated decision support tool to predict the change in groundwater levels in the future under climate change scenarios and to ensure proper sustainable water management. Additionally, the testing of the sensitivity of model output to model parameters will help refine model uncertainty and identifies the need of specific additional measurements to be implemented on ULB.