Avaliação de métodos para a regionalização de curvas de permanência de vazões para a Bacia do Rio das Velhas

Abstract

With the increasing demand for water due to population growth, industrialization and expansion of the agricultural frontiers, the development of tools of analysis and simulation becomes necessary. These tools are the basis for planning and management and, consequently, making decisions about how much water should be used by the multiple users of water resources within a catchment.The knowledge of the minimum flows is of upmost importance in this context. A frequentlyused low flow index refers to the flow wich is equaled or exceeded in the high percentage of time (e.g. 95%), as given by the so-called flow duration curve (FDC). The FDC is the complement of the cumulative distribution function of daily, monthly, or annual streamflows and represents the magnitude and frequency with which these flows areequaled or exceeded in a certain period of time. Frequently, the construction of the FDC is affected by the inexistence or restricted data availability at the places of interest in a river basin. In this MS thesis, two methods have been evaluated, namely, Tucci`s method (1991), which isused for the regionalization of long term FDC and the Claps and Fiorentinos method (1997), which offers a different approach. In the latter, FDCs are constructed on a yearly-based interval and the resulting FDCs are treated statistically, in the same way as annual minimumor maximum streamflow data. The procedure used for evaluation of the methods was the Jack-knife cross validation procedure, where a station with availability of data is excluded from the regional analysis and later included, with the exclusion of another station of the regional analysis. In this way, the observed data in a station are compared with computed values for the same station. The method of Tucci (1991) showed a good adjustment in all the regional analysis carried out for all the stations in three homogeneous regions for FDC for the Das Velhas river basin. TheClaps and Fiorentino´s method (1997) showed a good performance. However, it is necessary to further explore this method, with other application studies in other regions. As there is a strong relation among minimum flows and hydrogeologic characteristic of a basin, and giventhe need for inclusion of variables related with hydrogeologic characteristic, it was also used a method of computing the BFI (base flow index) for places with no data or scarce data, which did not show satisfactory results for the study area, probably due to the local geologic nature and possibly, to the quality of the available information.