Dinâmica da umidade do solo em bacia experimental do nordeste brasileiro

Abstract

This study aimed at identifying and characterize the different soil wetting processes through rain events in the Guaraira river experimental basin, located in the southern coast of Paraiba, as well as to investigate the factors that influence these processes and determine the most important Minimum Inter-event Time (MIT). For this, it was used data collected from soil moisture probes (TDR) with a high temporal resolution (1 hour) and from rain gauges installed in four distinct sites of the experimental basin. The study period was from 2004 to 2015, varying according to the availability of TDR probes data from each monitoring station. First, it was found that the variation of the MIT reduces the number of events up to 45%. Then, it was applied a methodology proposed by Lozano-Parra et al. (2015a) that it is based on temporal derivation of the soil moisture variation known as maximum wetting slope, and the types of soil wetting processes were classified and quantified. The results indicate a general or standard behavior of the soil moisture dynamics in the experimental basin with a dominant occurrence of the slow wetting processes caused by the matrix flow and a small occurrence of the fast flow, preferential flow processes. The multivariate statistical analysis based on data mining techniques proved that the slow flows are mostly related to the factors: initial soil moisture, precipitation accumulated, maximum precipitation in 30 minutes and the duration of the events. On the other hand, the fast flows are mainly related to the vegetation of each station studied. Regarding the validation of the model, it was observed that for the slow flows the coefficient of determination were very significant (38% - 73%), as well as an overestimation of the predicted values. For the fast flow, it was obtained a coefficient of determination equal to 38%, as well as an overestimation of the data, however for this process it become necessary a greater amount of information for a better calibration of the model. An alternative is the study of the relationship between maximum slope and soil moisture variation, where it was found that soil wetting processes can be explained by total soil moisture variation in the studied event.