Numerical modeling of the expansion and contraction dynamics of the Turbolo creek

Rivers intermittency is becoming a crucial topic in the environmental sciences because of its important role in several key bio-geochemical and ecological processes. River network dynamics are usually studied looking at the temporal co-evolution of the active network length L and the flow rate Q at the outlet during expansion-contraction cycles, triggered by rain events. Each L(Q) relationship is valid and unique for a peculiar catchment area and it has been described by several analytical and numerical models. The present thesis deals with the numerical modeling of the river network dynamics on the Turbolo creek, a small temporary catchment in the Calabria region (South Italy). In the Turbolo creek area, the average annual precipitation is approximately 1100 mm and the typical Mediterranean climate strongly affects the seasonal water budget of this catchment. In collaboration with the University of Calabria, which provided the relevant input data, the L(Q) relationship has been analyzed and described as a power law, taking into account measurements uncertainty and operational issues in the data collection. Then, the spatially distributed integrated numerical model CATHY has been calibrated, based on the temporal evolution at the outlet discharge and the L(Q) relationship. The CATHY model required knowledge of the DEM (digital elevation model), the peculiar soil and vegetation parameters of the catchment area and the atmospheric forcing. CATHY has been calibrated with the flow rate and the active network length measurements, considering the spatial and temporal distribution of the river network dynamics (expansion and contraction). The model was able to reproduce with a reasonable accuracy the temporal evolution of the discharge at the outlet in the Turbolo catchment (NSE = 0.5945), and provided a modeled L(Q) relation similar to the observed one, but with a more pronounced knee. The study provides a basis for the development of spatially distributed numerical models, which are able to reproduce the dynamics of the river networks in headwater catchment.