MapBiomas initiative: Mapping annual land cover and land use changes in Brazil from 1985 to 2017.
Abstract
The human-induced changes in the earth's surface have caused significant transformations on land cover regarding the structure and functioning of ecosystems, with severe implications for the environmental sustainability and people's livelihood. Thus, our ability to detect these changes constitutes a major research challenge, in both the environmental sciences and humanities.
Brazil has already advanced on process of monitoring some of these changes, but mostly forest change and concentraded in the Amazon biome, which covers almost half of the country. However, there are other 5 biomes (Atlantic Forest, Cerrado, Caatinga, Pantanal, and Pampa) with scarse or lacking land cover and land use (LULC) information to support planning and decision making about how to guide these transformations to reduce the impact in the environment and its people. MapBiomas is a collaborative and opensource monitoring initiative created in 2015 to fill this gap of information in Brazil. It innovates by working in a network formed by NGOs, Universities, and private companies organized by biomes and crosscutting themes (Pasture, Agriculture, and Coastal Zones). The MapBiomas network has produced annual LULC maps for the entire country from 1985 to 2017 using Earth Engine cloud computing technology to process Landsat data archive. The LULC maps were based on the random forest algorithm. This presents the most comprehensive information of LULC for the country enabling the characterization and understanding of transitions amongst land cover class for several application such as estimating carbon emissions, monitoring forest and land use dynamics and supporting decision making processes. All the data, digital maps, statistics and code scripts developed by the MapBiomas initiative are available in a web dashboard application. This initiative is being replicated in another countries in South America making possible to fill information gap and improve the understand of LULC at continental scale as well.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFM.B22A..04A
- Keywords:
-
- 1632 Land cover change;
- GLOBAL CHANGEDE: 1640 Remote sensing;
- GLOBAL CHANGEDE: 1855 Remote sensing;
- HYDROLOGYDE: 1942 Machine learning;
- INFORMATICS