Assessing the impact of rainfall seasonality anomalies
on catchment-scale water balance components
Abstract. Water balance components at catchment scale are strongly related to annual rainfall amount. Nonetheless, water resources availability in Mediterranean catchments depends also on rainfall seasonality. Indeed, a high percentage of annual rainfall occurs between late fall and early spring and feeds natural and artificial water reservoirs. This amount of water stored in the mild-rainy season is used to offset rainfall shortages in the hot-dry season (between late spring and early fall). Observed seasonal anomalies in historical records are quite episodic, but an increase of their frequency might exacerbate water stress or water excess if the rainy season shortens or extends its duration, e.g. due to climate change. Hydrological models are useful tools to assess the impact of seasonal anomalies on the water balance components and this study evaluates the sensitivity of water yield, evapotranspiration and groundwater recharge on changes in rainfall seasonality by using the Soil Water Assessment Tool (SWAT) model. The study area is the Upper Alento River Catchment (UARC) in southern Italy where a long time-series of daily rainfall is available from 1920 to 2018. To assess seasonality anomalies, we compare two distinct approaches: a static approach based on the Standardized Precipitation Index (SPI), and a dynamic approach that identifies the rainy season by considering rainfall magnitude, timing, and duration. The former approach rigidly selects three seasonal features, namely rainy, dry, and transition seasons, the latter being occasionally characterized by similar properties to the rainy or dry periods. The dynamic approach, instead, is based on a time-variant duration of the rainy season and enables to corroborate the aforementioned results within a probabilistic framework. A dry seasonal anomaly is characterized by a decrease of 241 mm in annual average rainfall inducing a concurrent decrease of 116 mm in annual average water yield, 60 mm in actual evapotranspiration and 66 mm in groundwater recharge. We show that the Budyko curve is sensitive to the seasonality regime in UARC by questioning the implicit assumption of temporal steady-state between annual average dryness and evaporative index. Although the duration of the rainy season does not exert a major control on water balance, we have been able to identify seasonal-dependent regression equations linking water yield to dryness index over the rainy season.