http://lattes.cnpq.br/9410392008100308; CRISPIM, Andréa Motta Coelho.
Résumé:
The watersheds located in the semi-arid northeast of Brazil present,
In general, small and / or discontinuous hydrological series. For
To overcome this problem, well-calibrated and validated hydrological models can be
used to generate long period series. On the other hand, regionalization techniques
can be used in homogeneous regions to determine a given variable
of interest (e.g., maximum flow rate, average, etc.) at sites without information. In this
research, hydrological regionalization techniques were employed to establish
of mathematical functions for the determination of maximum daily average flow rates,
long-term averages and 95% guaranteed flow rates based on simulated data with
NAVMO model in nine sub-basins of the homogeneous region of the upper Piranhas River - PB
(-15000 km2). The model was calibrated for each sub-basin and the parameters
used in the simulation of long period series. Parametric and nonparametric tests
were employed to verify the consistency and homogeneity of the simulated series and
probability distributions, indicated in the literature, were tested with the
Kolmogorov-Smirnov (K-S test) Correlation and regression analyzes (single and multiple)
between the variables of interest and the physical and climatic characteristics of the basins allowed
establish functions for the calculation of hydrological variables anywhere in the region
of the study. The results showed that the chosen model was reasonably calibrated.
based on annual flows and volumes and simulated series with calibrated parameters
presented reasonable consistency and homogeneity. K-S test results
showed that the Pearson III distribution adjusted the frequencies of the flow data
long-term maximums and average flows at all locations. The Gamma distributions of
Gumbel and Normal had some restrictions. Correlation and regression analyzes
with linear and potential functions showed that the two physical variables that best
explained the hydrological variables investigated were the basin area (simple regression)
and the basin area and the length of the main river (multiple regression). The inclusion of
other physical and climatic variables (eg, slope of the drainage density basin and
precipitation) improved the results, but not markedly. In general the
Linear equations represented variables better than potential equations.
