LAUSR

Abstract Water resources management is meant to solve problems caused by the intensive use of this resource which is the consequence of economic and population growth. Water resources management is also responsible for preserving hydrological, biological and chemical functions of ecosystems, as well as ensuring that this resource will be maintained with adequate supply for future generations. In Brazil, a country with continental dimensions, streamflow gauge coverage is far from satisfactory. Therefore, the hydrological regionalization techniques are an option to estimate hydrological information in regions with few or no data. This methodology is based on the spatial similarity of functions, variables and parameters that tolerate this transference. The number of regionalization models and the need for ease of access to hydrological data transformed comparative studies of streamflow regionalization methodologies into an important topic of hydrology in Brazil. For this reason, the objective of this paper is to assess the models settled by Liazi et al. (1988), IGAM (2012) and Wolff et al. (2014) related to the variables Q7,10, Q90, Q95 and Q ‾ in the upper Jaguari River basin, which is located between Minas Gerais and Sao Paulo States, and represents an important affluent to the Cantareira System. Firstly, Q7,10, Q90, Q95 and Q ‾ were calculated using the historical flow data from five fluviometric stations located in Jaguari River basin. Then, these same variables were estimated using the regionalization methodologies proposed by Liazi et al. (1988), IGAM (2012) and Wolff et al. (2014). In order to determine the best hydrological regionalization method for the upper Jaguari River basin, the flow rates results calculated with the historical series were compared to those estimated by the models. Liazi et al. (1988) presented the best performance when compared to the other two methods. Its behavior was classified as very good by the index of agreement (d), the Nash-Sutcliffe efficiency (NSE) and the coefficient of determination (R2). IGAM (2012) exhibited an unsatisfactory fit for the statistical indicators of NSE and PBIAS. On the other hand, this method was classified as very good when the coefficient of determination (R2) was analyzed. The proposal of hydrological regionalization of Wolff et al. (2014) presented an adjustment, classified by the index of agreement (d), Nash-Sutcliffe efficiency (NSE), percentage bias (PBIAS), and coefficient of determination (R2) as unsatisfactory, in general. The graphical behavior, however, demonstrated a good potential for future use of this model. In conclusion, it is recommended to apply similar comparative studies in other important sub-basins located between Minas Gerais and Sao Paulo States in Brazil to corroborate or not with the presented results.