Agricultural land use and cover change in the Cerrado/Amazon ecotone: a case study of the upper Teles Pires River basin.

• Autor(a) principal: ZAIATZ, A. P. S. R.
• Data de Publicação: 2018
• Outros Autores: ZOLIN, C. A., VENDRUSCULO, L. G., LOPES, T. R., PAULINO, J.
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Institucional da EMBRAPA (Repository Open Access to Scientific Information from EMBRAPA - Alice)
• Texto Completo: http://www.alice.cnptia.embrapa.br/alice/handle/doc/1091418; http://dx.doi.org/10.1590/1809-4392201701930
• topic: Gestão territorial; Territorial management; Amazon Basin; Bacia hidrográfica do alto Teles Pires; Bacia Amazônica; Sensoriamento Remoto; Sistema de Informação Geográfica; Remote sensing; Geographic information systems

The upper Teles Pires River basin is a key hydrological resource for the state of Mato Grosso, but has suffered rapid land use and cover change. The basin includes areas of Cerrado biome, as well as transitional areas between the Amazon and Cerrado vegetation types, with intensive large-scale agriculture widely-spread throughout the region. The objective of this study was to explore the spatial and temporal dynamics of land use and cover change from 1986 to 2014 in the upper Teles Pires basin using remote sensing and GIS techniques. TM (Thematic Mapper) and TIRS (Thermal Infrared Sensor) sensor images aboard the Landsat 5 and Landsat 8, respectively, were employed for supervised classification using the "Classification Workflow" in ENVI 5.0. To evaluate classification accuracy, an error matrix was generated, and the Kappa, overall accuracy, errors of omission and commission, user accuracy and producer accuracy indexes calculated. The classes showing greatest variation across the study period were "Agriculture" and "Rainforest". Results indicated that deforested areas are often replaced by pasture and then by agriculture, while direct conversion of forest to agriculture occured less frequently. The indices with satisfactory accuracy levels included the Kappa and Global indices, which showed accuracy levels above 80% for all study years. In addition, the producer and user accuracy indices ranged from 59-100% and 68-100%, while the errors of omission and commission ranged from 0-32% and 0-40.6%, respectively.