Modelling land use changes impacts on the silting of small agricultural water harvesting reservoirs

Water harvesting with Small Agricultural Reservoirs (SmAR) represents a solution for sustainable water management in different contexts. However, many technical challenges are still open despite its widespread application. One of the most relevant, for the sustainable management of SmAR, is represented by the loss of storage volume caused by the inflow of sediments, but the analysis of the dynamics of sedimentation for such small structures has received relatively little interest so far. This study aims to implement a validated model simulating the hydrology and erosion dynamics of the catchment upstream of a SmAR in the Mediterranean basin, specifically in the hilly area of Crete Senesi, Tuscany Region (Italy). Here, wine production is particularly developed, but not within the catchment of study, where the cultivation of cereals and forage is practiced. Our analysis aimed at estimating how much the rate of sediment accumulation in the reservoir would vary with the replacement of currently arable land with vineyards. A model was implemented on the HEC-HMS software, maximizing the value of existent low-cost data (Google Earth imagery and regional erosion maps) for its validation. The validated model was then used to test alternative land use scenarios in the upstream catchment, showing its flexibility for supporting decision-making over SmAR management. The model performed with an error always below 5% on the SmAR area detected by satellite. Erosion values calculated with HEC-HMS were in line, but lower than the estimation made by the Tuscany region with a GIS-based procedure. The scenario analysis showed that the simulated land use change led to a high value of annual sediment accumulation in the reservoir (216% of the original value of erosion obtained with cereals and other crops), showing the indirect cost of changing the cropping patterns to vineyard production. The approach can be replicated at the local scale in all other contexts where similar, and relatively easy-to-get, data are available. Further development of the present approach can include the replication of similar low-cost methodologies on other case studies, refinement of the erosion modelling and sensitivity analysis.