Regionalização de vazões na bacia hidrográfica do rio Paranapanema

Abstract

This scientific work had the objective of identifying hydrologically homogeneous regions and regionalizing the streamflow of the Paranapanema river basin. In the first Article, the hydrologically homogeneous regions of the Paranapanema river basin were identified, for which seven hierarchical algorithms were tested: Average-linkage, Centroid-linkage, Complete-linkage, Median-linkage, Ward method, Mcquitty method, and Single-linkage and k-means, a clustering algorithm. Different input combinations were also tested, with different combinations of the variables: flow, physical characteristics and location, of 98 fluviometric stations which belong to the basin under study. Considering the results, it can be verified that the input constituted by the attributes of flow and location proved to be more adequate for the grouping process associated with the k-means partition algorithm. Three hydrologically homogeneous groups or regions with distinct characteristics were identified. In Article 2, the Traditional and Mass Conservation methods were employed to regionalize the flow rates. For the implementation of the mass conservation method, the use of a conditioned adjustment routine and the use of a point of support was proposed. With the results, it was found that in situations where the tributaries have a fluviometric station located close to the point of drainage, the equation produced cannot represent the phenomenon, or does not even yield regionalization equations, limiting the use for this method. In other cases, the conditioned adjustment, with or without the fulcrum, proved to be efficient. With the application of the Traditional method, it can be concluded that the regression models that best fit the flow data are the simple and multiple potentials. The drainage area and the length of the main river are the best explanatory variables for the estimation of flows in group 2. In the case of group 3, the total length of the rivers and the drainage area were the response variables for the estimation of flows with 90 and 95% of permanence; the main river length and the drainage area for the minimum duration flow of seven days and return period of ten years; and the drainage area to estimate the average long period flow. Comparing the estimates of the Mass conservation and Traditional methods, it can be concluded that the estimates of the streamflow using the method of Mass conservation were as good as those provided by the Traditional Method.