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 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 for spatially refining each of these real deforestation alerts, using activated Planet images, with spatial resolution of 3-5 meters. Below is a description of each step of this process.
1. Data organization: Alerts generated by different platforms (SAD, DETER, GLAD and etc) are downloaded and organized monthly in the database where the different auxiliary databases are also connected (e.g. CAR, Conservation Units, Indigenous Lands and other territorial limits). The deforestation polygons identified annually (PRODES and Atlantic Forest Atlas) are also being analyzed on the platform.
2. Pre-processing: Initial step responsible for selecting which deforestation alerts can be considered valid, having as reference the monitoring systems of each biome, as well as their respective classes of native vegetation cover, according to the most recent MapBiomas map available (see table below). At this stage, some alerts can be ignored (e.g. GLAD alerts in the Amazon) or automatically discarded (e.g. GLAD alerts about forestry areas).
|Biome||Monitoring systems||Native vegetation class(es)|
|Amazon||DETER-B-Amazônia, PRODES, SAD and SIPAM-SAR (for areas with high cloud cover) and SIRADX||Forest Formation|
|Forest Formation; Savanna Formation; Grassland Formation|
|Caatinga||SAD Caatinga||Forest Formation; Savanna Formation; Grassland Formation|
|Pampa||GLAD-Alerts||Forest Formation; Wetland; Grassland Formation|
|Pantanal||GLAD-Alerts||Forest Formation; Savanna Formation; Grassland Formation; Other Non-Forest Formation|
|Atlantic Forest||SAD Mata Atlântica,
GLAD-Alerts, Atlas da Mata Atlântica
|Forest Formation; Mangrove; Savanna Formation; Grassland Formation|
Data from DETER-B AMAZÔNIA, DETER-CERRADO, GLAD, SAD and SIPAM-SAR are monthly, while data from PRODES-AMAZÔNIA and PRODES-CERRADO are annual.
3. Pre-activation: Stage operated by experts to identify which of the alerts should be removed because they are false positives for deforestation, carried out 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.
4. Planet Activation: Stage responsible for activating/acquiring Planet images for regions with valid alerts, according to the experts' 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. 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.
5. Alert Refinement: Stage responsible for the refinement of deforestation alerts with reference to activated Planet images. The refinement process begins when the pre-approved alerts at SCCON and their respective Planet images are available in the MapBiomas Alerta Workspace, a tool that facilitates the operation of experts in Google Earth Engine, as described below:
- Selection of the alert that will be refined with images of before and after deforestation of deforestation;
- Delimitation of the region of interest based on observed deforestation;
- Delimitation of deforestation and non-deforestation samples within the region of interest;
- Redesign of the alert polygon containing the effective area of deforestation, through supervised classification using the selected samples and the Random Forest algorithm;
- Fine adjustment of the tracing of the polygon resulting from the classification of the deforestation alert;
- Approval of the refined polygon associating it with the possible vector of vegetation pressure.
6. Audit: In this step, the 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.
7. Spatial Analysis: with operations within the database, the spatial crossings of the alerts with various thematic maps (e.g. CAR, Conservation Units, Indigenous Lands, Settlements, deforestation authorizations and SINAFLOR/IBAMA embargoes) are made available to the Dashboard, remote access APIs and report generation.
8. Publication of Reports: For each CAR that spatially crosses an alert, a Report is produced containing the (i) CAR ID, (ii) source of the alert, (iii) images of before and after deforestation, (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).
9. Publishing on the Dashboard: all alerts >0.2 ha are published on the online platform (http://alert.mapbiomas.org) organized in such a way as to enable the visualization of alerts on the map and statistics on the number of alerts and deforested area by biome, state and municipality for the selected period. 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 agencies or those of public interest (ex IBAMA).
10. Publication in Service APIs: APIs are built for consultation and access to alerts and customized reports for different public and public interest agencies.
OBS.: All alerts that are not validated and refined have the reason for rejection or non-validation recorded (eg positive lack, not observed due to excess clouds)