The Rambla de Algeciras lake in Murcia is a reservoir for drinking water and contributes to the reduction of flooding. With a semi-arid climate and a very friable nature of the geological formations at the lakeshore level, the emergence and development of bank gullies is favored and poses a problem of silting of the dam. A study was conducted on these lakeshores to estimate the sediment input from the bank gullies. In 2018, three gullies of different types were the subject of three UAV photography missions to model in high resolution their low topographic change, using the SfM-MVS photogrammetry method. The combination of two configurations of nadir and oblique photography allowed us to obtain a complete high-resolution modeling of complex bank gullies with overhangs, as it was the case in site 3. To study annual lakeshore variability and sediment dynamics we used LiDAR data from the PNOA project taken in 2009 and 2016. For a better error analysis of UAV photogrammetry data we compared spatially variable and uniform uncertainty models, while taking into account the different sources of error. For LiDAR data, on the other hand, we used a spatially uniform error model. Depending on the geomorphology of the gullies and the configuration of the data capture, we chose the most appropriate method to detect geomorphological changes on the surfaces of the bank gullies. At site 3 the gully topography is complex, so we performed a 3D distance calculation between point clouds using the M3C2 algorithm to estimate the sediment budget. On sites 1 and 2 we used the DoD technique to estimate the sediment budget as it was the case for the LiDAR data. The results of the LiDAR and UAV data reveal significant lakeshore erosion activity by bank gullies since the annual inflow from the banks is estimated at 39 T/ha/year.