On the water resources availability in the Dora Baltea River: a modelling experiment with the GEOframe system

The Dora Baltea River is one of the main tributaries of the Po River. Its basin includes the entire Valle d’Aosta region, where it originates, and a small part of the Piemonte region. The Dora Baltea plays a crucial role in providing water to the downstream regions because most of its flow is regulated by many dams that collet and divert the water to meet several needs downstream. In addition, during summer, snow and glacial melting upstream contribute to the river discharge. The flow of the Dora Baltea is vital for various - often conflicting – purposes including agriculture, hydroelectric power production, tourism, sports and civil uses. Therefore, quantifying the water availability generated in this area is fundamental to assess a sustainable water-use strategy. This thesis contributes to an international project aimed at supporting stakeholders in the development of adaptation strategies to climate change, providing information about the availability of water resources for specific uses in Piemonte and in the Po River basin. In this work the semi-distributed hydrological modelling system GEOframe is used to model the Dora Baltea Basin and extract the discharge arriving in Piemonte in a period of 27 years (1996-2023). The entire model has been built to be run with a daily time resolution in order minimize the computational demand while maintaining good modelling results. The goal is to evaluate the water reaching the main withdrawal points, such as the onset of Canale Farini - an important in-take structure designed to increase and regulate the water flow available in the Canale Cavour, made for irrigation purposes of the fertile lands located in Piemonte. Since a gauging station is necessary for model calibration and none exists at the Canale Farini confluence, the entire basin has been closed about 5 kms downstream, near Verolengo. At the end, to estimate the discharge reaching Canale Farini, a post-processing evaluation has been conducted to quantify the amount of water diverted. Six calibrations have been carried out - four upstream and two downstream - to improve the simulated discharge. Each calibration has been done twice: first with one round of 100 iterations, then with three rounds of 200 iterations. In the end, the values of the objective function – Kling-Gupta efficiency (KGE)- are higher compared to the initial round of calibrations, demonstrating significant improvements through the calibration process. Considering all six calibrations, an average KGE range between 0.60 and 0.80 has been achieved, with the lower values in upstream which are among the most challenging to model considering all the snow melting related processes. Furthermore, the sum of all observed and simulated discharge values is compared, yielding reasonable results, with ∑Qobs/∑Qsim = 1.05 at the basin outlet. The overall result, including all the hydrological processes, are acceptable since, comparing the results with studies that used finer sub-basin discretization, the difference in KGE is about 0.05-0.10. The study demonstrates that it is possible to reduce the number of sub-basins while still achieving good model accuracy and maintaining low computational demand. The model simulates all natural flows, but it could be improved in the future by incorporating anthropogenic changes aimed to flow regulation, such as the dams in the Valle d’Asota region which play a crucial role in the hydrological management of the territory.