Perda de solo e aporte de nutrientes e metais em reservatórios do Semiárido Brasileiro

Abstract

The forms of soil use and management have been associated with the occurrence of erosion which is one of the major forms of soil degradation. In semi-arid regions, erosion is relevant due to a combination of factors such as uneven distribution of rainfall, shallow soils and deciduous vegetation, which favor the erosion processes, especially when associated with the misuse of the soil. In addition to this, with erosion, sediments are transported into reservoirs, causing the eutrophication of their waters. In such context, water reservoirs in the semi-arid deserve special attention due to their importance to water supply for human consumption and for agricultural activities. The objective of the proposed research is to assess the loss of soil for different uses as well as the contribution of eroded soil to the input of loads (nutrients and metals) and to changes in water quality of the Castanhão reservoir (Jaguaribara-CE). To accomplish that, the use and occupation of land surrounding the reservoir were identified and mapped. Subsequently, entry points of loads were selected based on these uses; and erosion plots were installed in those areas. Sampling of the eroded material was conducted at every rainfall event (24h) at the erosion plots, being quantified the level of metals (N, P, Fe, Al, Zn and Cu) in the sediments and water. Soil losses were estimated by means of the Universal Soil Loss Equation (USLE); sediment and water samplings were seasonally captured; and hydrochemical data were collected in situ at seven points (P1, P2, P3, P4, P5, P6 and P7) within the reservoir, aiming at the evaluation of its eutrophication level and its relationship with the identified metals in soils and sediments in the erosion plots. The water and sediment data processing and integration were carried out through multivariate analysis techniques (PCA and Clustering), descriptive statistics and Pearson and Spearman correlations. In the mapping of the uses it has been verified that the Decision Tree algorithm (DT), applied to the sensor with the best spatial resolution (WordView-2), resulted in better accuracy when considering the irrigated areas, ones which were overrated by the ML (Maximum Likelihood). It was found soil losses only in the irrigated area, with a value of 2032 kg ha-1 of soil for the whole evaluation period. The estimation of soil losses through USLE resulted in values of 17.46, 29.1 and 1662.8 kg ha-1 for the areas of native forest, degraded area and irrigated area, respectively. The calculated values for measured loads in the irrigated area were 398 kg km-2 year-1 (N), 621 kg km-2 year-1 (Cu) and 5.462 kg km-2 year-1 (Zn), which suggests that such area has a high potential in the load transportation into the reservoir due to, above all, the characteristics of the soil that has an impedance layer which favors fast saturation of the surface layer. Associated with this, the irrigated area soils have lower levels of total organic carbon and high levels of Zn, revealing, thereby, the need for preventive measures and conservation practices in the evaluated agricultural areas. As for water and sediment quality, it was verified that attention should be given to the areas with greatest anthropogenic pressures, especially the irrigated area. For Cu and Zn, their enrichment factors are well pronounced in the sediments on the bottom of the reservoir over time. The sedimentation profiles support the significant input of nutrients and especially of Zn, with mean sedimentation rate ranging from 1.2 to 2.5 cm year-1. Therefore, the agriculture activity developed in the irrigated perimeter of Curupati already represents a negative influence to the water quality.