moradi, fatemeh (2025) Regionalization of Flow Duration Curves In Southern Peninsular Italy. [Tesi di dottorato]

Documento PDF Fatemeh_Moradi_Thesis_3.pdf Download (8MB)

Abstract

Prediction of streamflow characteristics in ungauged basins is a central challenge in hydrology, particularly in regions with sparse or fragmented monitoring networks. In such contexts, regionalization methods provide a practical solution by transferring hydrological information from gauged to ungauged sites through empirical relationships between streamflow indices and measurable catchment descriptors. A key tool in this process is the Flow Duration Curve (FDC), which condenses the variability of streamflow regimes into a single statistical representation, capturing flows from high extremes to persistent low values. However, the application of conventional FDCs in Mediterranean catchments is particularly complex. These basins are strongly influenced by seasonal rainfall, prolonged summer droughts, and intermittent streamflow regimes, which generate long sequences of zero-flow days. In such contexts, traditional FDCs must be reconstructed to distinguish between zero-flow and non-zero-flow components, ensuring that both flow intermittency and discharge magnitude are properly represented. This methodological adjustment is essential in southern Italy, where climatic gradients, geomorphological complexity, heterogeneous geology, and discontinuous records make regionalization both challenging and necessary. This thesis addresses these challenges by developing a statistical framework for the regionalization of FDC parameters across a study domain that represents nearly one-fifth of southern Italy. The focus is placed on three indices that capture complementary aspects of flow regimes. The first is the mean of the log-normal distribution (μ), which represents the central tendency of normalized flows and reflects the internal structure of the flow regime relative to the mean discharge. The second is the mean non-zero discharge (Qnz), which characterizes the average flow magnitude during periods when water is present. The third is the duration of non-zero flow (τ), which quantifies hydrological persistence and continuity. In addition to these FDC-based indices, the Baseflow Index (BFI) is included as a complementary measure, offering insight into subsurface contributions to streamflow. Importantly, this study is the first to incorporate a permeability map of the region into the regionalization framework, thereby improving the representation of geological controls on runoff generation and groundwater–surface water interactions. The methodology integrates exploratory correlation analysis, stepwise regression in both linear and log-transformed space, jackknife cross-validation, and outlier detection. Particular emphasis is placed on balancing statistical performance with hydrological interpretability, avoiding models that are statistically efficient but physically implausible. Furthermore, the framework explicitly tests the sensitivity of regionalization outcomes to streamflow record length, thereby assessing the reliability of models derived from shorter or incomplete data series.

Downloads

Actions (login required)

Modifica documento