Sim, as coleções do MapBiomas estão disponíveis como assets na plataforma Google Earth Engine e podem ser acessados, processados e analisados diretamente na plataforma sem necessidade de baixar os dados. Estão disponíveis 3 coleções de dados principais: > mapas anuais de cobertura e uso da terra no Brasil entre 1985 a 2023 > mapas de transição para períodos selecionados entre 1985 a 2023 > mosaicos de imagens Landsat para cada ano da série entre 1985 a 2023. Para obter os IDs das coleções do MapBiomas no Google Earth Engine acesse: Codes and Tools

Eight collections of MapBiomas maps have been produced since the beginning of the project in July 2015.

The collections represent changes in the coverage periods of the annual map, changes in the legend or corrections to the previous version.

Seven collections of data were published:

• Collection 1 - published in April 2016 with simplified legend, including 7 classes and covering the period 2008-2015

• Collection 2 - published in April 2017, including 13 legend classes and covering the period 2000-2016.
  • Collection 2.3, published in January 2018, with improvements from random forest application.

• Collection 3 - published in August 2018, including 19 legend classes and covering the period 1985-2017.
  • Collection 3.1, published in April 2019, with improvements in classification.

• Collection 4 - published in August 2019, including 19 legend classes and covering the period 1985-2018.
  • Collection 4.1, published in March 2020, with improvements in classification.

• Collection 5 - published in August 2020, including 21 legend classes and covering the period 1985-2019.

• Collection 6 - published in August 2021, including 25 legend classes and covering the period 1985-2020.

• Collection 7 - published in August 2022, including 27 legend classes and covering the period 1985-2021.
  • Collection 7.1 - published in April 2023, with improvements in classification.

• Collection 8 - published in August 2023, with 29 legend classes covering the period from 1985 to 2022.

• Coleção 9 – publicada em agosto de 2024, com 29 classes de legenda cobrindo o período de 1985 – 2023

Each MapBiomas collection reprocesses the entire time series.

The MapBiomas project is an initiative of the Climate Observatory co-created and developed by a multi-institutional network involving universities, NGOs, and technology companies with the purpose of annually map Brazil's land use and land cover and monitor the changes in the territory.

Land use and land cover maps made on the MapBiomas scale (30m pixel) and large time series are impractical in vector format. All data processing is performed in raster format, pixel by pixel.

It is not possible in the MapBiomas platform. The vectorization of maps is extremely costly and impractical for the whole territory. The recommended solution, if you want to work with vector, is to download the data, perform the territorial or spatial clipping of your interest and then vectorize using your favorite software.

MapBiomas data is public and free of charge by merely referring to the source in the following format: "MapBiomas Project - Collection [version] of the Annual Land Use Land Cover Maps of Brazil, accessed on [date] through the link: [LINK]"

Or access here the scientific article: Souza et al. (2020) – Reconstructing Three Decades of Land Use and Land Cover Changes in Brazilian Biomes with Landsat Archive and Earth Engine – Remote Sensing, Volume 12, Issue 17, 10.3390/rs12172735

The complete project description can be found at http://brasil.mapbiomas.org“

Landsat has an average resolution of 30m, so it is common to associate the area of ​​a pixel to 900 m². But, since the original MapBiomas data is created following the GEE standard representation (Lat/Long and WGS84), it does not use a native equal-area projection. Thus, the distance from the target to the Equator Line influences the pixel size. Therefore, in the continental scale of Brazil area calculation counting the pixels and multiplying by 900 m² should be avoided. In MapBiomas we apply two methods for area calculation.

(i) When performed outside Google Earth Engine, we reproject MapBiomas data to the UTM system and calculate the metric value of the central pixel, located at the intersection between the 1: 250,000 (IBGE standard) chart and area of ​​interest. Next, we count all pixels within the area of ​​interest and multiply by the reference value, in m², as previously calculated. This method has been applied, for example, throughout all the statistics calculation in Collections 2 and 2.3.

(ii) When the calculation is performed within Google Earth Engine we apply the function ee.Image.pixelArea() that generates an image in which the value of each pixel is the area of that pixel in square meters, taking into consideration eventual cartographic distortions. This method is being applied from Collection 3.

This difference occurs because the area calculation in MapBiomas is done by the sum of the areas of each pixel. The edge pixels are distributed between the border areas. In general, the difference in areas is less than 0.2%, and in cases of smaller municipalities, it can reach 0.5%.

Classes named as beta are the ones in their first version mapped, still in a preliminary form, for a better version in the next MapBiomas' collection. In the same time, some of the beta classes have not been mapped for the entire brazilian territory yet, but have already shown a good result in the main occurence areas of these classes.