This is not a new alert system, but an effort to enhance the usability and effectiveness of alerts already generated. This initiative is a collective construction involving users (MMA, IBAMA, SFB, ICMBio, MPF, etc.) and alert providers (INPE, University of Maryland, IMAZON, ISA, etc).
Existing systems identify alerts from the large scale analysis of medium resolution satellite images (20-250 m resolution). These images have weekly or biweekly frequency and allow to identify deforestation, but without many options to analyze at the level of properties due to resolution and low frequency of acquisition of the images.
In MapBiomas Alerts the deforestation alerts produced by the various monitoring initiatives are gathered, validated and detailed based on daily images of 3 m resolution. Each alert generated is accompanied by a high resolution image before and after deforestation (which produces an effect similar to the vehicle license plate photo of a traffic violation).
Daily images of 3 m resolution and daily frequency of Planet's 200+ micro-satellites constellation, a San Francisco-based start-up represented commercially in Brazil by SCCON (Santiago & Cintra Consultoria), are used.
Most MapBiomas validated deforestation alert polygons are produced from training machine learning algorithms that accurately capture deforestation and natural cover changes. During the processing of alerts, filtering operations are performed to smooth the perimeter and reduce the number of vertices.
The Deforestation Report is a document that aggregates information to the alert to allow a direct and rapid analysis on the nature of deforestation observed. For each CAR (Rural Environmental Cadastre), or another territorial unit in its absence, the alert is identified with high resolution images of before and after deforestation followed by complementary information, such as:
- overlapping of deforestation alert with protected areas (Conservation and Indigenous Territories), APP (Permanent Preservation Area) and RL (Legal Reserve);
- overlap alert with embargo area;
- presence of an existing vegetation suppression authorization;
- land use history in MapBiomas over the past 6 years;
- vertices of the alert polygon.
The data of the real estate owners registered in the CAR, such as name and CPF / CNPJ, are not available to the public in SICAR (National System of Rural Environmental Cadastre).
Not directly. This function can only be attributed to the governmental competent bodies of SISNAMA (National Environmental System). Onotherhand, the laudo contains information that assist the identification of the illegal areas, for example:
- identifies whether there is authorization suppression area associated with existing vegetation;
- identifies whether there is deforestation in declared areas of APP or RL;
- identifies whether there is overlap with protected areas.
Not yet. In the MapBiomas Alerts beta version, a pre-selection of alerts is made based on a series of criteria such as size, location and land use history to prioritize the alerts to be validated and refined.
For each selected alert, it is fully evaluated in a cell / tile of 5 x 5 km dimages satellite dailyup to six months prior to detection of the alert by existing systems.