This research studies the effect of climate change on the hydrological behavior of two semi-arid basins. For this purpose, the Soil and Water Assessment Tool (SWAT) model was used with the simulation of two future climate change scenarios, one Representative Concentration Pathway moderate (RCP 4.5) and the other extreme (RCP 8.5). Three future periods were considered: close (2019–2040), medium (2041–2070), and distant (2071–2100). In addition, several climatic projections of the EURO-CORDEX model were selected, to which different bias correction methods were applied before incorporation into the SWAT model. The statistical indices for the monthly flow simulations showed a very good fit in the calibration and validation phases in the Upper Mula stream (NS = 0.79–0.87; PBIAS = −4.00–0.70%; RSR = 0.44–0.46) and the ephemeral Algeciras stream (NS = 0.78–0.82; PBIAS = −8.10–−8.20%; RSR = 0.4–0.42). Subsequently, the impact of climate change in both basins was evaluated by comparing future flows with those of the historical period. In the RCP 4.5 and RCP 8.5 scenarios, by the end of the 2071–2100 period, the flows of the Upper Mula stream and the ephemeral Algeciras stream will have decreased by between 46.3% and 52.4% and between 46.6% and 55.8%, respectively.

This study introduces a time-lag-informed Random Forest (RF) framework for streamflow time-series prediction across diverse catchments and compares its results against SWAT predictions. We found strong evidence of RF’s better performance by adding historical flows and time-lags for meteorological values over using only actual meteorological values. On a daily scale, RF demonstrated robust performance (Nash–Sutcliffe efficiency [NSE] > 0.5), whereas SWAT generally yielded unsatisfactory results (NSE < 0.5) and tended to overestimate daily streamflow by up to 27% (PBIAS). However, SWAT provided better monthly predictions, particularly in catchments with irregular flow patterns. Although both models faced challenges in predicting peak flows in snow-influenced catchments, RF outperformed SWAT in an arid catchment. RF also exhibited a notable advantage over SWAT in terms of computational efficiency. Overall, RF is a good choice for daily predictions with limited data, whereas SWAT is preferable for monthly predictions and understanding hydrological processes in depth.

Ephemeral streams are highly dependent on rainfall and terrain characteristics and, therefore, very sensitive to minor changes in these environments. The western Mediterranean area exhibits a highly irregular precipitation regime with a great variety of rainfall events driving the flow generation on intermittent watercourses, and future climate change scenarios depict a lower magnitude and higher intensity of precipitation in this area, potentially leading to severe changes in flows. We explored the rainfall–runoff relationships in two semi-arid watersheds in southern Spain (Algeciras and Upper Mula) to model the different types of rainfall events required to generate new flow in both intermittent streams. We used a non-linear approach through generalized additive models at event scale in terms of magnitude, duration, and intensity, contextualizing resulting thresholds in a long-term perspective through the calculation of return periods. Results showed that the average ∼ 1.2 d and < 1.5 mm event was not enough to create new flows. At least a 4 d event ranging from 4 to 20 mm, depending on the watershed, was needed to ensure new flow at a high probability (95 %). While these thresholds represented low return periods, the great irregularity of annual precipitation and rainfall characteristics makes prediction highly uncertain. Almost a third of the rainfall events resulted in similar flow to or lower flow than the previous day, emphasizing the importance of lithological and terrain characteristics that lead to differences in flow generation between the watersheds.

This study has evaluated the runoff and erosion rates in torrents of Southern Italy, two forested headwaters with very similar climatic, hydrological and geomorphological characteristics; in one headwater, 15 check dams were installed in the mid-1950s, while the other is not regulated with engineering works. To this aim, the hydrological variables have been modeled over 15 years after check dam installation using the HEC-HMS (Hydrologic Engineering Center-Hydrologic Modeling System) model coupled to the MUSLE (Modified Universal Soil Loss Equation) equation. The model simulations have shown that check dams have not played a significant role in reducing the surface runoff compared to the unregulated torrent; in both catchments, the well-developed forest cover determined very low runoff coefficients (lower than 0.3%) with a scarce runoff generation capacity. Additionally, the reduction in peak flow due to the check dams was not significant, on average −7.4% compared to the unregulated headwater. Check dams have retained sediments for about 8–10 years after their installation, reducing erosion by about 35%, although soil loss was much lower than the tolerance limit in both catchments. After the sediment retention capacity of the dam sediment wedge was depleted, the sediment yield in the regulated torrent was even higher (by about 20%) compared to the unregulated catchment. Overall, the study has shown that the use of check dams as a catchment management strategy of forested headwaters under semi-arid Mediterranean conditions should be considered with caution, since the structures could be ineffective to reduce water and sediment flows during floods or, in some cases, check dams may increase erosion rates.