Simulating ecologically relevant hydrological indicators in a temporary river system

Hydrological indicators (HIs) are commonly used in eco-hydrological studies (i.e. environmental flow and hydrological status assessment). Their computation is based on streamflow data, and if measured data are not available, hydrological models can be used to generate flow data. The present paper describes a study that aimed to predict streamflow in a temporary river and to analyze the general reliability of some hydrological indicators evaluated by using simulated data instead of measured flow data. The SWAT model was used to predict daily streamflow in a river section of the Celone river (Puglia, Italy). Several HIs characterizing the patterns of river flow or specific hydrological components were evaluated using observed and simulated streamflow. The results show that the SWAT model is able to simulate streamflow in temporary river systems, but its performance under extreme low flow conditions may be a weak point. When simulated streamflow time series were used, the replicability of the HIs evaluated using a rigorous statistical methodology ranged from good to limited. Good performance was found for the magnitude of discharge in wet months (average monthly flow from November to May), for the high flow indicators (annual maxima, 1-, 3-, 7-, 30-, 90-day mean flow) and timing, while limited performance was detected for low flow indicators (annual minimum 1-, 3-, 7-, 30-, 90-day mean flow) and the number of zero flow days. Better performance for low flow indicators was found after introducing the zero-flow threshold. This type of eco-hydrological study may contribute to characterizing the flow regime and its alterations in regions with scarce data.