OVERVIEW

GET TO KNOW THE STEPS OF THE MAPBIOMAS ALERT METHODOLOGY

Below you will find the main features, structure, and methodology of the MapBiomas Alert (v.1) to validate and refine deforestation alerts with high-resolution images in Brazilian biomes.

GENERAL CHARACTERISTICS

All alerts and respective deforestation reports are produced from supervised analysis and classification of Planet satellite images with 3 m resolution and daily frequency.

The whole process is done with extensive use of machine learning algorithms through the Google Earth Engine platform that offers tremendous processing power in the cloud.

The work is organized in tiles of 5 x 5 km. The definition of tiles to be analyzed is based on the deforestation alerts generated by the systems DETER / INPE (Amazon and Cerrado), SAD / IMAZON SimparSAR/Censipam(Amazon) and GLAD / University of Maryland (other biomes).

HOW WE ORGANIZE

To validate and refine alerts, teams of programmers, specialists in remote sensing, conservation and land use are organized into teams for each biome and experts in technology and system support.

 

OVERVIEW OF THE METHOD

The diagram below illustrates the main steps in the validation and refinement of deforestation alerts in Brazilian biomes.

SOURCE OF CROSS DATA

The alerts are crossed with land tenure and fiscalization information, such as indigenous territories, protected areas, settlements, Rural Environmental Registry (CAR) areas, including Permanent Preservation Area (APP) and Legal Reserve (RL), as well as areas of embargo, suppression authorizations and the forest management plan of Sinaflor from IBAMA. Alerts are also located in geographical boundaries such as municipalities, states, biomes, and watersheds.

Alerts can be viewed on different base maps (backgrounds) such as Google Earth images (non-associated date), Google Maps 2D, and Brazil's land cover and land use maps between 2012 and 2017 from MapBiomas (Collection 3.1).

This information qualifies the alerts and complements the reports with relevant information to the users.


Layer Year Description Reference Site to download
Countries 2015 Map of Brazil (political boundary) IBGE, 2015 ftp://geoftp.ibge.gov.br/organizacao_do_territorio/malhas_territoriais/malhas_municipais/municipio_2015/Brasil/BR/
States 2015 Map of the States of Brazil IBGE, 2015 ftp://geoftp.ibge.gov.br/organizacao_do_territorio/malhas_territoriais/malhas_municipais/municipio_2015/Brasil/BR/
Municipalities 2015 Map of the Municipalities of Brazil IBGE, 2015 ftp://geoftp.ibge.gov.br/organizacao_do_territorio/malhas_territoriais/malhas_municipais/municipio_2015/Brasil/BR/
Biomes 2016 Map of Brazilian biomes in 1: 5,000,000 (IBGE, 2004) refined RADAM vegetation map to scale 1: 1,000,000 and state boundaries in 1: 250,000 (IBGE, 2015) MapBiomas, 2016 https://drive.google.com/open?id=0Byp5eRWoQ-PkWUpXNHRGMTdkcTg
Watersheds level 1   Map of watersheds level 1, scale 1: 1,000,000 ANA http://dadosabertos.ana.gov.br/datasets?group_ids=084346aa5c18467782432f48bb687f83
Watersheds level 2   Map of the Indigenous Lands of Brazil ANA http://dadosabertos.ana.gov.br/datasets?group_ids=084346aa5c18467782432f48bb687f83
Indigenous Lands 2017 Mapa das Terras Indígenas do Brasil

ISA, 2017

https://terrasindigenas.org.br/pt-br/
Protected area 2019 Map of Conservation Units of Brazil MMA, 2019 http://mapas.mma.gov.br/i3geo/ogc.htm
Rural settlements  

2017

Map of settlements INCRA, 2017 http://acervofundiario.incra.gov.br/acervo/acv.php
CAR 2018 Areas Rural Environmental Registry Serviço Florestal, 2018 https://sistemas.florestal.gov.br/geoserver
Embargo 2019 Embargo areas Sinaflor, http://siscom.ibama.gov.br/geoserver/
Authorizations 2019 Permits for suppression and forest management Sinaflor, SEMA-MT, SEMA-PA

http://siscom.ibama.gov.br/geoserver/

https://monitoramento.semas.pa.gov.br/monitoramento/#/sig

http://monitoramento.sema.mt.gov.br/simlam/

Rural real estate - INCRA SIGEF 2019 Delimitation of rural real estate

SIGEF/

INCRA

https://sigef.incra.gov.br/



WHAT ARE THE PLANET IMAGES?


PlanetScope images
come from a constellation of +130 satellites capable of daily recording the entire earth surface (equivalent to a collection capacity of 340 million km² / day) with A a spatial resolution of 3 m and four spectral bands.

For each alert, dozens of images are selected and activated in a 5 x 5 km cell. From this total, two images are selected, one before and other after deforestation, recording the satellite identifier, date and time of acquisition.