Overview

Land Cover Maps for 2018-2023

Impact Observatory creates on-demand, up-to-date, 10m Land Cover maps for your custom area of interest.

The 10m Land Cover maps are available when you need more than our annual global 9-class Sentinel-2 10m land use/land cover time series maps that have had well over 1.3 million accesses from the Esri Living Atlas.

Your custom maps are available for either sub-annual or annual time periods, from the present back to the beginning of 2018, using the most recently available satellite imagery. Impact Observatory uses a unique machine learning approach to classify land use and land cover (LULC) categories globally using ESA Sentinel-2 imagery. 

Order now

Note: Select a location and date range to order a map. Maps will be delivered to you for download within one day. Introductory pricing is available. 

SPECIFICATION

The LULC Map on Demand provides users with a custom map of land use/land cover for a user-specified area of interest and time period (2018-2022). The map is derived from ESA Sentinel-2 imagery at 10m resolution. It is a composite of LULC predictions for 9 classes over the specified time period (3 months or more is usually required for sufficient cloud-free scenes),  generating a representative snapshot of LULC.

TECHNICAL DETAILS

Data Projection: Universal Transverse Mercator (UTM)
Mosaic Projection: WGS84
Available Extent: Global
Source imagery: Sentinel-2
Cell Size: 10m (0.00008983152098239751 degrees)

CLASSES

Water Channel Extent – Full extent of water under normal circumstances (maximum long-term footprint outside of flooding or other exceptional events). Covers areas like lake beds and arroyos where water might be present regularly, but only seasonally.

Variable Water – Intermittent water flow or standing water, representing seasonal fluctuation in cover, weather events, or human activities. At times this class may occur due to frequent turbidity, algal blooms, ice cover, pollution, or glare.

Persistent Water – Sustained water flow or standing water, representing permanent or sustained seasonal cover.

Dense Tree – Healthy, closely packed tree canopy cover (vegetation over 5m high) with very little to no vegetation or bare ground cover. Defined as a tree prediction with a maximum Normalized Difference Vegetation (NDVI) value greater than or equal to a biome-specific threshold during the given time period.

Sparse Tree – Vegetation that is predominantly tree canopy cover (vegetation over 5m high) with some mixing of other vegetation or bare ground due to relatively thin leaf cover and/or dispersed distribution of trees. In some circumstances, this class can also indicate trees gaining/losing leaves or dying/growing during the given time period. Defined as a tree prediction with a maximum NDVI value lower than a biome-specific threshold during the given time period.

Dense Rangeland – Healthy, closely packed vegetation that is predominantly dense, short (under 5m) woody shrubs with very little to no mixed grass or bare ground cover. May contain small isolated trees. Within and around areas classified as built, this class can also include highly manicured lawns or fields. Defined as a rangeland prediction with a maximum NDVI value greater than or equal to a biome-specific threshold during the given time period.

Sparse Rangeland – Vegetation that is some mix of grasses and/or dispersed, short, woody scrub, with or without some bare ground cover. May contain small isolated trees. Defined as a rangeland prediction with a maximum NDVI value less than a biome-specific threshold during the given time period.

High Density Built – Artificial, impervious surfaces in the form of individual features, parts of features, or tight clusters of features with little to no mixed vegetation or bare ground. Isolated clusters of these detections represent small buildings or construction, while larger clusters are indicative of commercial, industrial, or high-density residential areas. Defined as built area with a maximum NDVI value less than a biome-specific threshold.

Low Density Built – Artificial impervious surfaces, buildings, and structures mixed with vegetation that cannot be easily detangled at 10m-resolution. These areas typically represent low-density residential areas, suburban areas, large path networks or other dispersed human construction. Defined as built area with a maximum NDVI value greater than or equal to a biome-specific threshold.

Crops – Human planted/plotted cereals, grasses, and crops not at tree height; examples: corn, wheat, soy, and fallow plots of parceled land.

Flooded Vegetation – Areas of vegetation with obvious intermixing of water throughout the majority of the given time period; mostly herbaceous (non-woody) vegetation and/or scattered tree (mangrove)/scrub/shrub cover. Thick mangrove, swamp, or seasonal wetland may fall under their next-best-fit vegetation classes if water is not observed during the given time period.

Bare Ground – Areas of rock or soil with very sparse to no vegetation for the entire year; large areas of sand with little to no vegetation; examples: exposed rock or soil, desert, dry salt flats, dry lake beds

Snow/Ice – Large, homogenous areas of persistent snow or ice, typically only in mountain areas or high latitudes.

Clouds – No land cover information due to continual cloud cover.

ABOUT IMPACT OBSERVATORY

Impact Observatory brings AI-powered algorithms and on-demand data to sustainability and environmental risk analysis for governments, industries, and markets. Impact Observatory empowers decision-makers with the timely, actionable, science-based geospatial insights they need to succeed.

Founded in 2020 and based in Washington, DC, Impact Observatory’s machine learning (ML) and science teams blend capabilities ranging from data science, software engineering, ML, science communications, program management, and strategic planning to provide the science-based insights you need for your business, mission or impact goals.

Learn more about us.