GET TO KNOW THE STEPS OF THE MAPBIOMAS ALERT METHOD

Below you will find the main characteristics, the way of organizing the work, and the overview of the MapBiomas Alert method to validate and refine, with daily satellite images in high spatial resolution, deforestation alerts of Brazilian biomes since 2019. The alerts are crossed with geographic territories (such as biomes, states, municipalities, and watersheds), land tenure (such as Rural Environmental Registry, Conservation Units, and Indigenous Lands), and administrative actions (such as the existence of authorization, assessment or embargo). The result is a complete report for each deforestation event detected in Brazil.

All data and reports are publicly, openly, and freely available on a web platform so that enforcement agencies, financial agents, companies, and civil society can act to reduce illegal deforestation.

GENERAL CHARACTERISTICS

All alerts and respective deforestation reports are produced from the analysis and supervised classification of PlanetScope satellite images of 3-5 m resolution and daily frequency.

The entire process is done with extensive use of machine learning algorithms through the Google Earth Engine platform, which offers immense cloud processing capacity.

Deforestation alerts detected by monitoring systems operating in Brazil from the following responsible institutions are validated and refined, according to the table below:

HOW WE ORGANIZE

To carry out the validation and refinement of alerts, teams of programmers, remote sensing specialists and conservation and land use specialists are organized into teams for each biome, in addition to the technology and systems support team.

METHOD OVERVIEW

The diagram below illustrates the main steps in the process of validating and refining deforestation alerts in Brazilian biomes, going through the stages of compilation/aggregation, pre-validation, validation, refinement, auditing, analysis/crossing and publication.

SOURCE OF CROSS DATA

Once validated and refined with high-resolution images, the alerts are cross-referenced with land tenure and enforcement information, such as indigenous territories, conservation units, rural settlements, afro-descendent territories, areas of the Rural Environmental Registry (CAR), including Permanent Preservation Areas (APP) and Legal Reserves (RL), in addition to embargo areas, suppression authorizations and forest management plan of the Sinaflor do Ibama. Alerts are also located in geographic limits such as biomes, states, municipalities, and watersheds.

In the MapBiomas Alerta Platform, alerts can be visualized in different base maps (backgrounds), such as Google Earth images (without associated date), Google Maps 2D, monthly mosaics of Planet images, and the most current collection of annual maps of land use and land cover in Brazil of MapBiomas.

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

HIGH-RESOLUTION IMAGES

MapBiomas Alert uses PlanetScope (3 meters) high spatial resolution satellite images. For each alert, dozens of daily images are checked to select two images, one before and one after deforestation, recording the identification of the images and the date of acquisition in the alert report.

The process of validating and refining deforestation alerts includes automatic steps and operational steps involving remote sensing specialists with knowledge of the deforestation dynamics associated with each Brazilian biome.

The automatic steps are responsible for discarding false positives (e.g. polygons detected over forestry harvesting areas) and for transferring the data necessary for the refinement process between the different platforms used by the project (e.g. Planet/SCCON Platform, Google Cloud Platform, Google Cloud Storage and Google Earth Engine).

Specialists are responsible for evaluating the alerts produced by the monitoring systems and thus activating Planet images for the regions that will be analyzed, selecting the best before and after deforestation images, and spatially refining each of these deforestation alerts validated in the Planet images activated, with a spatial resolution of 3-5 meters.

Below is a description of each step of this process.

1. Data collection and aggregation: at this step, deforestation alerts of different monitoring systems (DETER/INPE, SAD/IMAZON, GLAD/Univ. Maryland, SIRAD-X/ISA, SAD Caatinga/Geodatin-UEFS, SAD Pantanal/SOS Pantanal-ArcPlan and SAD Mata Atlântica/SOS Mata Atlântica-ArcPlan) are downloaded and consolidated monthly, aggregating alerts whose polygons have some degree of overlap. In this process, the unique identifier (ID) of the alert is generated, which will be followed until the end of the validation, refinement and publication process. At this stage, auxiliary administrative, land and inspection territorial databases are also acquired and imported (eg embargoed areas, deforestation authorizations and forest management plan, Rural Environmental Registry - CAR, Conservation Units, Indigenous Territories and other territorial limits ).

2. Pre-validation: includes two steps. The first is done in an automated way, eliminating all alerts that overlap with reforestation and agriculture areas present in the annual maps of the most recent MapBiomas Collection or that have already been detected in previous surveys. After this process, the alerts maintained are visually evaluated by analysts with the support of PlanetScope high-resolution monthly tiles (daily images with 3 m resolution). The main objective of this step is to remove alerts that are not real, avoiding unnecessary activation of Planet images for regions with false deforestation alerts. At this point, the fraction of alerts that correspond to cases of false positives is discarded, with the corresponding record of the reason for rejection (forestry, agriculture, seasonality, etc.).

Sources of deforestation alerts used in MapBiomas Alerta and the respective classes of native vegetation cover monitored:

In addition to these systems, analysis tests are being carried out with other sources of annual deforestation: (a) INPE/SOS Mata Atlântica monitoring and (b) INPE's PRODES Amazônia and Cerrado, to identify omissions in the monthly monitoring systems of the biomes.

3. Pre-activation: Step operated by analysts to identify which of the alerts should be removed because they are false positives for deforestation, performed by visually inspecting the PlanetScope monthly mosaics. The main objective of this step is to remove alerts that are not real, avoiding unnecessary activation of Planet images for regions with false deforestation alerts. When alerts are not validated, the reason for rejection is recorded (eg duplicate, reforestation, seasonality, agriculture, relief shade, etc.).

3. Planet Activation: Step responsible for activating/acquiring Planet images for regions with valid alerts, according to the analysts' review, for further refinement of the spatial delimitation of these deforestation alerts. The activation of the images takes place on the Planet/SCCON Platform (Planet's representative in Brazil), which cuts the images according to the selected area and stores this result in the Google Cloud Storage platform, ingesting images in Google Earth Engine. Analysts identify images from before and after the deforestation event ('before' and 'after' images). When using Sentinel images the selection is made in the toolkit in Google Earth Engine. Planet images are stored with all spectral bands (e.i. blue, green, red and near infrared), unusable data mask (e.i. Unusable Data Mask-UDM) and respective metadata.

4. Alert Refinement: Step responsible for refining deforestation alerts with reference to activated Planet images, performed by analysts in the Google Earth Engine platform environment, called (Alerts Workspace), as described below:

1. Collection of deforestation and non-deforestation samples within the region of interest;

2. Redesign of the alert polygon containing the effective area of ​​deforestation, through supervised classification using the selected samples and the Random Forest algorithm;

3. Simplification, fine-tuning of polygon geometry and removal of overlap with previously refined polygons;

4. Approval of the refined alert and respective "before" and "after" images, selecting a possible vector of deforestation pressure.

5. Audit: In this step, each refined alert is evaluated by the biome supervisor to ensure consistency of the final result. If necessary, the alert can return to any of the previous steps to be redone.

6. Spatial Analysis: with operations within the database, the spatial crossings of alerts with various thematic maps (e.g. CAR, Conservation Units, Indigenous Lands, Settlements, deforestation authorizations and embargoes) are made available through the dashboard, access APIs remote control and report generation.

7. Publication of reports: For each CAR that spatially intersects with an alert, a report is produced containing: (i) CAR code, (ii) alert source, (iii) before and after deforestation images, (iv) location of the property and location of the alert within the property, (v) location of the alert/property in the FU, (vi) overlapping data with TI, UC, settlements, legal reserve, APP and Nascentes, (vii) existence of embargo, plan of management or suppression authorization on the property, (viii) history of coverage of the area in previous years (based on the MapBiomas Collection) and (ix) descriptive memorial of the alert area on the property. In the case of alerts that do not cross CAR properties, a simplified report is generated with items (ii), (iii), (v), (vi), (vii), (viii) and (ix).

8. Publishing on the dashboard: all alerts >0.3 ha and their respective reports are published on the free access web platform (https://plataforma.alert.mapbiomas.org/), where it is possible to view each alert, filter by territorial (eg states, municipalities, protected areas) or administrative (eg CAR, with or without authorization), in addition to accessing alert statistics (eg number, area, average speed, alert size classes) for the period selected. Each alert can be viewed with before and after images of deforestation with date and affected area and a link to related reports. In a restricted access area, dashboards are set up for consulting reports and customized reports for different public bodies or those of public interest (eg IBAMA, ICMBio, SFB, MPs and OEMAs).

9. Publishing in Service APIs: application programming interfaces (APIs) are built for consultation and access to alerts and customized reports for different public and public interest agencies.

Data can also be accessed and downloaded via spreadsheet, shapefile and alert reports, or via Plug-In for QGIS.

Developed by Luiz Motta (QGIS Contributor), this plugin allows visualizing the MapBiomas Alert data directly on the QGIS.

The minimum Windows version for installing this plugin is 3.14.12 or later.

To install the plugin, you must enable the experimental plugins.

The plugin displays only alerts that are within the QGIS drawing area.

Access the video tutorial to learn how to operate the plugin:

The Law Enforcement Dashboard uses public databases of deforestation permits and law enforcement operations such as environmental infractions and embargoes. Data accessed through active transparency are used, which means public data provided by state agencies, most likely online, without an official request.

Access the full Methodological Note here.