1.Scope

This chapter describes the spatial and temporal reference scope for each country or region to implement the mapping initiatives for land cover and land use, fire scars, water, etc., in different collections, as well as general guidelines for publishing the related products.

This section provides guidelines for the choice of images and cautions related to it

• Map collections can be produced from different images (e.g. Landsat, Sentinel) with time series amplitude and spatial resolution appropriate to the type of satellite;

• The use of different sensors must take into account the need to produce compatible and comparable data throughout the time series.

This section provides guidelines for the choice of time step and time series interval considering different initiatives and images.

• Land cover and land use maps must be annual, comparable in extent and spatial resolution, and consistent throughout the time series;

• Other monitoring initiatives (e.g. fire, water, etc.) can be generated with specific methodologies and periodicities;

• Use the entire time interval that presents a good quantity and quality of images available. It is recommended, for the Landsat sensor, the beginning of the time series in 1985 and, for the Sentinel sensor, the beginning of the time series in 2016.

This section addresses MapBiomas spatial scope and issues related to the regionalization of mapping.

• MapBiomas' mission focuses on tropical and subtropical regions, starting in South America and Indonesia and aiming to expand to other territories in Asia and Africa;

• Although the organization of the network and the publication of data are focused on countries, the production of data, that is, the images classification, should be based on ecoregions, biomes or other subregions adapted to the context of each country, considering the heterogeneity of landscapes and types of natural formations. The regionalization of mappings can be adjusted to the ease, knowledge of the local context and quality of the classification carried out by local experts;

• Regionalization by biomes or ecoregions is recommended for mapping vegetation. However, transversal themes (urban areas, agriculture, forestry, mangroves, among others) may require specific methodologies and regionalization.

This section provides guidelines for launching, updating and reviewing collections of land cover and land use and other initiatives.

• Annual maps and statistics of land cover and land use are published by administrative and natural territories, adjusted to the reality of each country or region (e.g. states, municipalities, biomes, ecoregions, river basins);

• Additionally, annual land cover and land use maps are used to calculate transitions between classes at predefined time intervals, which can be standardized (e.g. every 10 years) or defined according to the interests of the country or region. For transition analyses, classes can be hierarchically grouped, according to local interest.

• Each new collection is published with the reprocessing of all years of the time series, inclusion of the most recent year and with possible improvements to the content, such as greater detail in the legend, methodological advances, improvements in the samples and reference data used;

• If an error is found after publishing the collection (see item 4.2, in the Quality Control Chapter) that justifies a revision, it is possible to launch another version of the same collection, with special attention to communicating resolved problems (e.g. collection 7.1 from Brazil);

• It must be taken into account that the launch of a new version involves almost the same effort as the launch of the original version, including updating the platform, the factsheet, communication with users, and therefore it must be discussed and planned with the network secretariat;

• Mapping of other initiatives is dependent on the local priorities and context. However, once a new initiative or module is established in a country (e.g., fire), the same general guiding principles as for land cover and land use mapping are relevant.