Análise temporal do uso da terra em uma microbacia hidrográfica no município de Piracicaba, SP, por meio de técnicas de geoprocessamento

Abstract

ln order to answer some questions related to the usual methodology of temporal land use data analysis using geographic information systems (GIS) technology, this study was developed at Ribeirão Pau d' Alhinho watershed in Piracicaba, SP. It were used conventional aerial photographs for 1995, 1978, 1962 and 1940 as data sources and GIS techniques to data conversion and analysis. The land use categories were delineated onto clear mylar overlays and then were converted to digital format employing a digitizing table. The vector files created were concatenated and edited, and later rasterized by Idrisi for Windows 2. O software, also used in data analysis. The multitemporal analysis was developed by the crosstabulation, area determination and reclassification GIS functions and priorized examine recent and past land use data. The land use data layers were overlaid to soil-type and slope combined data layer to verify land use adequacy and degradation potential. Natural vegetation data and preservation areas were investigated, too. It was shown that the use of 1940 aerial photographs, despite of their technical limitations, improved historical land use and trends analysis. x v i i The sets of control points elected to coordinates transformation had errors less than the allowable RMS (root mean square) errors, and so all maps meet the U.S. National Map Accuracy Standards. However, the correction of errors visible after land use vector files concatenation consumed a lot of time. Thus, the land use categories delineated on the mylar overlays should be transferred to base maps before data conversion. Another problem was to visualize the land use category boundaries during digitizing process. Y et, the 5m pixel size produced the best results regarding features representation, time processing and data volume. The multitemporal land use data analysis indicated that Pau d' Alhinho watershed was predominantely covered by sugar cane in 1995, regardless its low agricultura! suitability, demonstrating land use inadequacy. Simultaneously, arable lands were not entirely cultivated. Since 1940, a significant portion of land was already cultivated, in general, the sarne land which are nowadays tilled. 10.25% of study area was always inadequately used, suggesting land degradation. Natural vegetation category was covering approximately 8% and 12% of the watershed in 1995 and 1940, respectively, but its spatial distribution changed greatly. Preservation areas were misused specially as pastures. It was confirmed in this research that the usual methodology of temporal land use data analysis is adequate, although it could be improved enlarging the scales of overlays and transferring land use data to base maps prior to data conversion. The multitemporal land use data analysis adopted was advantageous since allowed to investigate past referenced to recent land use.