Study of Hydrologically Homogeneous Regions in the Rio das Velhas Basin

Código

I-EBHE0159

Autores

Vitoria Ellen da Silva Oliveira, Eber José de Andrade Pinto, FABIO HENRIQUE RODRIGUES DA SILVA

Tema

WG 2.6: Water systems analysis for integrated planning & management

Resumo

The flow rate in a watershed corresponds to a given value integrated across its entire drainage area, influenced by the combined effects of all internal factors in relation to its physical and climatic characteristics (Tucci, 2017). According to Castellarin et al. (2013), the concept of hydrological similarity helps us group similar watersheds, defining a hydrologically homogeneous region. Naghettini and Pinto (2007) consider the identification and delineation of hydrologically homogeneous regions (HHRs) a challenging step prone to subjectivity. Hosking and Wallis (1997) recommend that the identification of homogeneous regions be conducted in two consecutive steps. The first step is based on local characteristics, and the second involves a statistical test built solely on local statistics to verify the preliminary results. Among the methods for delineating HHRs, clustering analyses stand out, commonly referred to as cluster analyses, a method of multivariate statistical analysis where each station/object of study is associated with a data vector containing local characteristics and/or statistics. These are grouped and regrouped to identify the degree of similarity among their data vectors (Naghettini and Pinto, 2007). This study aimed to apply and test combined methods for obtaining hydrologically homogeneous regions in the Rio das Velhas watershed, an important tributary of the São Francisco River. The first method involved applying cluster analysis to long-term dimensionless flow duration curves derived from average flow rates at each monitored data series in the basin. The second method involved vector association of local watershed characteristics. The use of flow duration curves is justified as they graphically represent a basin's hydrological regime, reflecting variability in flows across a cross-section, based on physical, climatic, morphological, and vegetation cover characteristics (Costa, 2012). The technique employed in this study was Ward's method (Ward, 1963), a hierarchical clustering method where similarity between clusters is measured by the sum of squares of differences across all variables. In this study, these variables corresponded to flow duration curves from 0 to 99 obtained at 18 monitoring stations in the watershed between 1994 and 2014. Subsequently, a critical analysis of the clustering was conducted based on local characteristics (geology, slope, average annual precipitation, evapotranspiration, hydrogeology, and land use/land cover). The clustering method resulted in a dendrogram with 4 well-defined trees and 11 derived branches. After evaluating the physical and climatic characteristics of the region, 5 hydrologically homogeneous regions were identified, showing strong alignment with the inflection points of the flow duration curves. The methodology and delineation obtained proved effective, as evidenced by good performance metrics in regionalization studies conducted in the basin, as highlighted in Oliveira et al. (2023).