Incertezas em modelos hidrológicos: o caso do modelo SCS/CN aplicado à Bacia Experimental de Aiuaba

Abstract

Most of the hydrological processes that occur in a watershed are, somehow, random. Similarly, the expected impacts are uncertain. Reports of the models used to predict the various impacts (hydrological, economic, environmental, and social) are often uncertain as well. So ignore this uncertainty is ignoring the reality of nature. The objective of this research was (i) to quantify the uncertainties of the λ and CN parameters of the SCS/CN model when applied to an experimental watershed located in the semi-arid region with preserved Caatinga (10 years of rainfall/runoff data, 2005- 2014) and (ii) to propose a method for evaluating the performance of hydrological models. It was evidenced a bimodal behavior in the histograms of the two parameters. This shows that both CN and λ vary due to soil macroporosity conditions provoked by root dynamics in the Caatinga Biome. These facts substantially alter the initial abstractions and ratify the hypothesis that the CN parameter varies from event to event. Therefore, the initial abstractions, influenced by the soil dynamics, contribute significantly to the uncertainties in the hydrological modeling of this watershed. The uncertainty analysis methodology of hydrological models developed here was adequate to the objectives of the study. As long as some fundamental statistical information of the variables, such as the probability density function, is known, it can be reproduced satisfactorily with different models and regions. The simulations based on the Monte Carlo Method hit 5 to 10% of the predictions. This number is not satisfactory and the justifications lie in the simplicity of the SCS/CN model and in the natural uncertainties of the hydrological processes. Due to the wide use of SCS/CN, we suggest that its application is always associated with a stochastic approach. We also recommend that this methodology be replicated in the Aiuaba Experimental Basin with more sophisticated hydrological models in order to validate their use also in regions of different hydrological regime of the study area of this research.