RESERVOIR SEDIMENTATION ESTIMATES AND LANDSLIDE SUSCEPTIBILITY IN PUERTO RICO AFTER HURRICANE MARIA
This study focused on the drainage basins for the following drinking-water reservoirs: Dos Bocas, Carraízo, Lucchetti, Caonillas, and La Plata. Landslide volume was calculated using an area-scaling relationship developed for shallow failures globally. Results indicate that the sediment liberated during these events surpass the remaining capacity of each reservoir. However, not all sediment is delivered to the fluvial network and suspended sediment will not be captured in the reservoir. Therefore, it is essential to determine the efficiency of sediment trapping in each lake. We estimate that Dos Bocas was the most affected reservoir, potentially reaching a capacity filled value of 92% if the sediment delivery was just 1% efficient; this represents a calculated 28% increase from Hurricane María triggered sediment liberation. We estimate that La Plata was the least affected, having a capacity filled value of 31% if the sediment delivery was 1% efficient, marking an event increase of less than 4%. These estimates are sensitive to the area-volume scaling factor used, but are important to prioritize bathymetric surveys to assess the actual capacity of each reservoir.
Within the target drainage basins, landslide frequency was much more dense than in the island as a whole. Using a pre-existing landslide susceptibility map, more than 90% of the landslides found in the drainage basins fell in areas classified as high to very high risk. In the Dos Bocas catchment, 30% of the area categorized as high to very high risk suffered mass wasting, whereas in the La Plata basin, only 4% was affected. This is indicative that this type of map can be useful in predicting reservoir basins that are vulnerable to landslide sediment accumulation. The combined results show that the drainage basins in Puerto Rico were likely greatly affected by the hurricane and that a revised empirical susceptibility map using the inventory will aid in improving the forecast for future landsliding events in addition to reservoir infilling.