Estudo do regime hidrológico no semi-árido brasileiro por modelagem dinâmica acoplada: aplicação em gerenciamento de reservatórios

Abstract

The premiss for the rational use of water has demanded information on smaller space and time scales. This study investigates the sensitivity of climate simulation of a dynamic downscaling of precipitation (1971-2000) coupled to a rain-flow model and its application to the operation of a water reservoir in the semi-arid region of Ceará State. Emphasis was given to intra seasonal scale (smaller than a month) in years of climatic contrasts classified with La Niña, El Niño and Normals (averaged) in the Tropical Pacific Ocean. We used atmospheric general circulation model ECHAM4.5, and the Regional Spectral Model (RMS), version 97 of National Center for Atmospheric Prediction-NCEP. The RSM was nested with data simulated by ECHAM4.5, taking as boundary condition the sea surface temperature observed in the period January to June 1971-2000. For estimating streamflow we rate used the SMAP model. Results show that after the technique of Probability Density Functions (PDFs) correction, the simulation by RSM, best captured daily variability of precipitation at fluviometric post of Iguatu, average, from 01 January to 30 June. The explaned variance before the correction was less than 5%, rising to around 40% after the correction by PDFs. The simulations of streamflow, analyzed by the explaned variance, square of correlation coefficient, shows that there is a two or more gain of explaned this variance, as compared with explanation to streamflow in relation to rainfall. This gain is observed for all periods (daily and accumulated five and fifteen days), and for the entire period (1971-2000). More than 70% pentads anomalies of stramflow were well simulated. Simulations of volume of a hypothetical reservoir (with the characteristics of the Orós Dam) with initial volum V0 = 0.25k ( maxim capacity - 1940 hm3), 0.50k and 0.75k, showed that there is the potential use of that cascade information of precipitation with rain-flow model, in particular for averaged years classified. The explanation of the variance between simulated and observed daily volumes for the months of January to June, for these years Normal results was more expressive. This was about 80%. In summary, the results showed the potential for use of coupled simulation model between atmospheric modeling water associated with the reservoir operation.