Uso de geoprocessamento para o aperfeiçoamento da gestão da qualidade da água em reservatório de região semiárida

Abstract

Semi-arid regions face high risk of drought, being one of the most peopled regions in the world. The loss of water quality in surface reservoirs is a serious problem in those regions. Changes in land use and land cover can uncontrollably intensify water-degrading processes, especially in Permanent Preservation Areas (APPs). In this context, the use of remote sensing contributes to improved water management by allowing the integration of scattered data in geographic into geographical information systems. Moreover, in semi-arid regions, hydrographic basins stand out as an important management system, due to their role in the maintenance of water resources, since there are the source of nutrient loads which represent a complex factor in the water quality degradation. The objective of this study was to determine the correlation between the limnological attributes of the reservoir and land use in Permanent Preservation Areas (APP's) within the basin and to estimate the nutrient, nitrogen (N) and phosphorus (P) produced by the main natural and anthropogenic sources in a Brazilian semiarid basin. The water samples were collected at four points in the General Sampaio reservoir, Ceará, Brazil, from 2014 to 2016. The limnological attributes used were chlorophyll-a, total nitrogen, total phosphorus and transparency. For the years 2009 to 2013 the water quality data were acquired at the Foundation Cearense de Meteorologia – FUNCEME homepage. Data were collected from the Institute of Research and Economic Statistics of Ceará - IPECE, the Brazilian Agricultural Research Company - Embrapa and the Brazilian Institute of Geography and Statistics - IBGE, as well as data from the scientific literature. Landsat 5 and 7/TM images and Landsat-8/OLI images were used in the land use and land cover classification. The delimitation of the APPs based on the Shuttle Radar Topography Mission (SRTM). The orbital data were processed with ArcGis 9.3 and ENVI 4.5 , whereas the GWLF model was applied to determine e daily and annual loads of N and P. This is a concentrated hydrological model with a GIS interface ( MapShed software). Emission factors were used to estimate the annual N and P loads. The input data for the two models were: SRTM Digital Elevation Model, satellite image (Landsat 8), soil map Brazil, municipal data on rainfall, livestock, agriculture, water consumption and population, as well as data from literature on soil losses, nutrient concentration by soil type, nutrient concentration in animal excreta and nutrient concentration in soil derived from overland flow of precipitation. Land use and land cover assessment in the APP showed that there is a significant and inverse correlation between the area occupied by bare soil and water transparency (-0.53), and between a significant and direct correlation between the area occupied by bare soil and the chlorophyll-a index (0.43). The difference between the size of the buffering zone (30 and 100 meters) of riparian forest indicated that the larger buffer is more effective in the presence of vegetation, whereas in the lack of vegetation, the smaller buffer has large effects. GWLF model results indicated that N loads from natural sources accounted for 34.5% of the total load whereas for the emission model factors they accounted 25.4%. In relation to anthropogenic sources, this percentage reaches 65.5% and 74.6% for GWLF and emission factors, respectively. The models, in general, showed that the nutrient inputs (N and P) from anthropic activities exceeded the natural inputs, with agriculture and livestock being the main sources of N and P. Therefore, changes in the agricultural model adopted in the region need changes so as to improve management practices in these areas to promote the reduction of nutrient and sediment loads to the reservoir.