Monitoring and Modeling Drainage Network Contraction and Dry Down in Mediterranean Headwater Catchments

Understanding the expansion and contraction dynamics of flowing drainage networks is important for many research fields like ecology, hydrology, and biogeochemistry. This study analyzes for the first time the network shrinking and dry down in two seasonally dry hot‐summer Mediterranean catchments (overall area 1.15 km2) using a comprehensive approach based on monitoring and modeling of the flowing network. A field campaign consisting of 19 subweekly visual surveys was carried out in the early summer of 2019. These observations were used to calibrate and validate an integrated model aimed to estimate the time evolution of the total flowing drainage network length based on meteorological drivers and define the position of the stretches with flowing water based on topographic and geological information. We used a statistical model to describe the observed variations in the total flowing length based on the accumulated difference between antecedent precipitation and evapotranspiration. The study emphasizes the relevant role of evapotranspiration in the seasonal network contraction. Then, we modeled spatial patterns of the flowing channels using an empirical approach based on topographic data, achieving satisfactory performances. Nevertheless, the performance further increased when site‐specific geological information was integrated into the model, leading to accuracies up to 92% for cell‐by‐cell comparisons. The proposed methodology, which combines meteorological, topographic, and geological information in a sequential manner, was able to accurately represent the space‐time dynamics of the flowing drainage network in the study area, proving to be an effective and flexible tool for investigating network dynamics in temporary streams. Key Points For the first time, the network shrinking and dry down in two seasonally dry hot‐summer Mediterranean headwater catchments are analyzed Network length dynamics depend on the difference between antecedent accumulated rainfall and evapotranspiration The modeling of the flowing network spatial patterns based on topographic and geological information achieved accuracies up to 92%