Original title: Reconstructing Three Decades of Land Use and Land Cover Changes in Brazilian Biomes with Landsat Archive and Earth Engine
The researchers built yearly maps of how land in Brazil was used from 1985 to 2017. They used satellite images and a computer method called random forestA computer method that combines many decision trees to sort examples into categories. to sort land into types such as forest, farming, and water. Each map shows detail down to 30-metre pixels, and checks against sampled locations gave an average accuracy of 89%. The maps show that natural vegetation was lost, mainly to cattle ranching and farming, while some converted land was regrowing.
Imagine looking through old family photos to see how a neighbourhood changed. Some homes and gardens have disappeared, while other places have grown green again. In this paper, the photos are satellite images from many years, and the map tracks what each patch of Brazilian land became over time.
People making environmental, farming, and land-use policy need to know not just what Brazil’s landscapes look like now, but how they changed over time. These maps reveal both large losses of natural vegetation and areas where converted land is regrowing.
The abstract reports map accuracy averaging 89%, with results ranging from 73% to 95% across biomes, so accuracy was not uniform. It does not provide class-by-class accuracy, explain the causes of the mapped changes in detail, or establish the ecological outcomes of regrowth. The mapping covers Brazil from 1985 to 2017.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
Reconstructing Three Decades of Land Use and Land Cover Changes in Brazilian Biomes with Landsat Archive and Earth Engine — Carlos Moreira de Souza Júnior, Julia Z. Shimbo, Marcos Reis Rosa et al. — Remote Sensing — doi:10.3390/rs12172735
Abstract licence: CC BY
Explanation generated by Synthesis from the abstract — it is not the original paper.