MAPBIOMAS

Overview

The MapBiomas project provides comprehensive satellite-based data on land cover and land use changes across Brazil. The project uses machine learning algorithms applied to Landsat satellite imagery to classify and track changes in biomes, land cover types, and economic activities over time. This dataset is essential for understanding deforestation, agricultural expansion, conservation outcomes, and landscape dynamics in the Amazon and other biomes.

MapBiomas represents one of the most detailed and scientifically rigorous sources of remote sensing data on Brazilian land cover, with coverage spanning from 1985 to present and annual updates.

Data Coverage

The MapBiomas dataset includes:

  • Time Period: Annual data from 1985 onwards; the exact final year varies by dataset and even by geo_level within the same dataset (most run through 2024, a few still stop at 2023 or earlier) – see Measurement Units and Notes below for how to check the year range for a specific (dataset, geo_level) pair before assuming it matches another
  • Geographic Coverage: All Brazilian biomes (Amazon, Cerrado, Atlantic Forest, Caatinga, Pampa, Pantanal)
  • Spatial Resolution: 30 meters (Landsat resolution)
  • Temporal Frequency: Annual maps with seasonal detail
  • Update Schedule: New data typically released annually with recent years added progressively
  • Aggregation Levels: Municipalities, indigenous lands, states, biomes, and protected areas

Available Datasets

1. Land Cover ("mapbiomas_cover")

Annual maps of land cover types including: - Forest Types: Native forest, forest formation, forest plantation - Non-Forest Vegetation: Grassland/pasture, shrub vegetation, herbaceous - Agricultural Land: Temporary crops (annual/seasonal), perennial crops (sugarcane, coffee, cocoa) - Urban Areas: Urban/built-up land and infrastructure - Water Bodies: Rivers, lakes, reservoirs, aquaculture - Non-vegetated: Mining, bare soil, rock outcrops - Other Categories: Cloud coverage, nodata

Key Applications: Deforestation monitoring, agricultural expansion, urbanization patterns, forest conservation

2. Land Cover Transitions ("mapbiomas_transition")

Year-to-year changes in land cover including: - Deforestation: Conversion from forest to other uses - Reforestation: Conversion to forest types - Agricultural Transitions: Changes between crop types or from other uses to agriculture - Degradation: Forest to degraded/shrub vegetation - Regeneration: Abandoned areas returning to vegetation

Key Applications: Deforestation tracking, regeneration assessment, agricultural dynamics, conservation effectiveness

3. Deforestation ("mapbiomas_deforestation_regeneration")

Specific focus on forest loss: - Deforestation: Permanent loss of forest cover - Degradation Signals: Forest areas showing stress indicators - Cumulative Deforestation: Total forest loss since baseline

Note: despite the name, this dataset covers deforestation only. Regeneration is a separate dataset, "mapbiomas_secondary_vegetation" (below).

Key Applications: Amazon monitoring, conservation evaluation, climate impact studies

4. Secondary Vegetation ("mapbiomas_secondary_vegetation")

Forest regrowth and natural recovery on previously deforested land: - Forest Regeneration: Secondary forest regrowth and natural recovery - Regeneration Extent: Total area of returning vegetation

Key Applications: Regeneration assessment, conservation evaluation, climate impact studies

5. Mining Activities ("mapbiomas_mining")

Areas used for mining operations: - Active Mining: Currently operational extraction sites - Abandoned Mining: Previous mining areas, potentially available for rehabilitation - Mining Extent: Total area impacted by mining activities - Mining Types: Surface mining, artisanal mining, infrastructure

Key Applications: Environmental impact assessment, land degradation, restoration planning, environmental compliance

6. Irrigation ("mapbiomas_irrigation")

Note: Temporarily unavailable – no irrigation statistics file exists on MapBiomas’s Dataverse archive at any collection. docs_url is left pointing at MapBiomas’s documentation for whoever investigates a replacement.

Provides: - Irrigated Areas: Extent of irrigation in agricultural systems - Irrigation Type: Drip, center-pivot, flood irrigation - Crop Types Under Irrigation: Which crops receive irrigation

7. Water Bodies ("mapbiomas_water")

"municipality" and "biome" work. "state" is unavailable: no MapBiomas collection publishes a state-level water sheet upstream.

  • Surface Water Extent: Permanent and seasonal water bodies, by municipality or biome
  • Seasonal Variation: Not broken out at annual granularity in the current source

8. Wildfire Burn Scars ("mapbiomas_fire")

Areas affected by wildfires: - Burn Scars: Areas burned in recent fires - Fire Extent: Total area impacted per year - Fire Frequency: Repeated burning in same areas (indicating high fire risk) - Fire Season: Temporal pattern of fires

Key Applications: Fire monitoring, ecosystem vulnerability, climate impacts, fire management

Function Parameters

Options:

  1. dataset: Eight possible choices (irrigation temporarily unavailable)
    • "mapbiomas_cover": Land cover types
    • "mapbiomas_transition": Changes in land cover
    • "mapbiomas_deforestation_regeneration": Deforestation only (see note above – regeneration is a separate dataset)
    • "mapbiomas_secondary_vegetation": Forest regeneration/regrowth
    • "mapbiomas_mining": Mining areas
    • "mapbiomas_irrigation": Irrigated areas (temporarily unavailable)
    • "mapbiomas_water": Water bodies ("state" unavailable; municipality/biome work)
    • "mapbiomas_fire": Wildfire burn scars
  2. raw_data:
    • TRUE: Data in original format from MapBiomas
    • FALSE: Cleaned and standardized version (recommended for analysis)
  3. geo_level: Varies by dataset
    • For "mapbiomas_cover": "municipality" or "indigenous_land"
    • For "mapbiomas_transition": "municipality" or "biome" (faster)
    • For "mapbiomas_deforestation_regeneration": "municipality" only
    • For "mapbiomas_secondary_vegetation": "municipality" only
    • For "mapbiomas_mining": "indigenous_land" or "municipality"
    • For "mapbiomas_irrigation": "state" or "biome"
    • For "mapbiomas_water": "municipality" or "biome" ("state" currently unavailable)
    • For "mapbiomas_fire": "state" only
  4. language:
    • "pt": Portuguese language labels
    • "eng": English language

Measurement Units and Notes

  • All Areas: Measured in hectares (ha)
  • 1 hectare: Approximately 2.47 acres
  • Validation: MapBiomas uses field validation and visual interpretation to ensure accuracy
  • Updates: Historical data may be revised with new collections and improved algorithms
  • Collections differ across datasets, and even across geo_levels of the same dataset. MapBiomas does not release every dataset on the same collection at the same time, and this package sometimes deliberately pins one geo_level behind its siblings when a newer collection is missing a sheet it needs (e.g. "mapbiomas_mining"’s "indigenous_land" row is one collection behind its "municipality" row – verified live; MapBiomas dropped that sheet from the newer file). Don’t assume one row’s collection or year range applies to another row of the same dataset.

Examples

# download treated MapBiomas land cover data by municipality
data <- load_mapbiomas(
  dataset = "mapbiomas_cover",
  raw_data = FALSE,
  geo_level = "municipality",
  language = "eng"
)

# download treated data on mining on indigenous lands
data <- load_mapbiomas(
  dataset = "mapbiomas_mining",
  raw_data = FALSE,
  geo_level = "indigenous_land",
  language = "eng"
)

# download treated wildfire burn scar data by state
data <- load_mapbiomas(
  dataset = "mapbiomas_fire",
  raw_data = FALSE,
  geo_level = "state",
  language = "eng"
)