Explore our Data
Land Cover Change Monitoring
Global Land Cover Change
Geographic coverage: Global | Resolution: 30x30 meters |
Temporal coverage: 2000, 2005, 2010, 2015, 2020 | Data partner: University of Maryland
Tracking how land cover and land use are changing across the globe is crucial to understanding, and solving, the global land squeeze.
Land & Carbon Lab partnered with the University of Maryland to produce the first ever map of global land cover change over the last 20 years. The map distinguishes eight general land cover classes: forest, cropland, built-up area, water, bare ground, short vegetation, snow/ice, and wetlands.
This data provides us with valuable insights into how human activity is transforming our planet, including where land use is changing and what is driving this change. It also provides us with valuable insights into the long-term impacts of these continued changes for people, nature and climate, and how we can manage land more sustainably in future.
This data set was developed using state-of-the-art, locally and regionally calibrated machine learning tools, and validated using statistical sampling that confirms its high accuracy. It is currently available at a 30-meter resolution at five-year intervals between 2000 and 2020. We are now working to make the data available annually, expected to be available in 2025.
OPERA Land Surface Disturbance Product (DIST-ALERT)
Geographic coverage: Global | Resolution: 30x30 meters |
Temporal coverage: 2023-present | Data partner: University of Maryland
A groundbreaking new land surface disturbance monitoring system developed by NASA’s OPERA project and the University of Maryland detects disturbances to any kind of vegetation cover, including forests, grasses, shrubs and even crops, occurring anywhere on Earth in near real-time — Learn more.
Using harmonized data from the NASA/USGS Landsat and ESA Sentinel-2 satellites, the new system, known as OPERA Land Surface Disturbance Alert from Harmonized Landsat Sentinel-2, or DIST-ALERT, can identify and continuously monitor vegetation in 30-meter pixels across the globe.
This new system will fuel future alert products that Land & Carbon Lab and the University of Maryland will develop for many possible applications, such as tracking agricultural expansion through the conversion of natural lands, detecting degradation dynamics such as logging, and monitoring short-lived vegetation changes such as from drought and fire.
Explore the data on Google Earth Engine, or download on the NASA EarthData Catalog
Dynamic World
Geographic coverage: Global | Resolution: 10x10 meters |
Temporal coverage: 2016 - present | Data partner: Google
Dynamic World – developed in partnership with Google - is a flexible data set of global land cover that uses AI to do the hard work of turning satellite images into understandable data on land cover.
Users can map land cover and changes as they happen, build custom maps and refine and analyze data to inform decisions related to land cover such as the conversion and conservation of natural ecosystems, changes in agricultural intensity and food security, climate-related changes to landscapes and the integrity of protected areas.
First, Google Earth Engine receives all imagery from the European Space Agency’s Sentinel 2 satellites. These images come in every 5-10 days at a 10-meter resolution — providing high spatial and temporal resolution. All of the data, which dates back to June 2015, is hosted openly on Google Earth Engine, removing the burden of computing power and storage for users.
Dynamic World then uses AI to detect the likelihood of each of the nine types of land cover (trees, shrubs, grass, crops, water, flooded vegetation, bare ground, snow/ice and buildings/roads) for each pixel. Users can refine the outputs to detect other land cover categories relevant to their goals.
Forest Monitoring
Global Tree Canopy Height
Geographic coverage: Global | Resolution: 1x1 meter |
Temporal coverage: 2020 | Data partner: Meta
Other existing satellite data, which is normally at 10- or 30-meter resolution, is good at understanding dense often tropical forests that have consistent canopy heights. It isn’t granular enough, however, to see the details of more dispersed forest systems such as agroforestry, drylands forests, and alpine forests. Together these constitute more than a third of the world's forests.
By partnering with Meta, and using a groundbreaking new AI foundational model, we’ve produced the world’s first global map of tree canopy height at a 1-meter resolution.
This new high-resolution data sets a baseline for remotely monitoring changes at the level of individual trees, making it a critical advancement for measuring land use emissions and tracking progress on conservation and restoration projects, which are essential for achieving the world’s goals for climate, nature and people.
This project demonstrates a new path towards AI driven earth monitoring. Not only have we made the data open access, we have also made the underlying foundational model free and publicly available, democratizing access to new AI technology.
You can find the data on AWS and Google Earth Engine, the model on GitHub and the paper here.
Tree Cover Gain and Net Change
Geographic coverage: Global | Resolution: 30x30 meters |
Temporal coverage: 2000 - 2020 | Data partner: University of Maryland
Over the past 20 years, geospatial monitoring has made it possible to see where deforestation is happening. A more elusive question, however, has been where new forests are growing. Recent data from the University of Maryland and available on Global Forest Watch, now offers more information on tree cover change between 2000 and 2020. By using tree cover height to measure tree cover gain and net change in tree cover worldwide this data allows users and decision makers to see a fuller picture of forest change dynamics. The tree cover gain data shows areas where tree height increased to at least five meters every five years between 2000 and 2020 to provide users with the opportunity to understand where and why gain is happening. The net change data maps overall tree height change (factoring in loss and gain) down to the district/municipality level between 2000 and 2020. This information can then be used to monitor and inform restoration efforts across the world.
In addition, tracking tree height is crucial to calculating forests’ biomass and carbon storage. Tracking net changes in tree area along with tree height allows us to estimate forests’ resulting emissions and removals of greenhouse gases (GHGs) from the atmosphere with greater precision, improving GHG accounting.
We are now working with the University of Maryland to expand this data set to include annual gain and net change data.
Tropical Tree Cover
Geographic coverage: Tropics | Resolution: 10x10 meters |
Temporal coverage: 2020
Geospatial monitoring is good at mapping dense forests, but what about trees outside of these forests? Until recently, billions of trees in open canopy forests and savannas were invisible to governments, investors and the public using geospatial monitoring. These trees play a vital role within communities providing food, shade, biodiversity and climate benefits, but because they couldn’t be monitored, they were excluded from inventories of tree cover and hence undervalued.
Land & Carbon Lab’s Tropical Tree Cover data set, developed in collaboration with World Resource Insitute's Global Restoration Initiative, makes these trees visible for the first time through geospatial monitoring. This allows decision makers to better understand trees outside of humid forests. Importantly, the data set also allows local communities who protect and restore these ecosystems to get credit for the work they do.
Tropical Tree Cover maps trees inside and outside of forests using 10-meter resolution imagery. Previous data sets were only available at 30-meter resolution, meaning smaller patches of trees were missed. This higher-resolution tree cover data improves our ability to quantify tree cover on non-forest lands, such as urban areas and cropland, and monitor trees at small spatial scales.
The current data set shows tropical tree cover in the year 2020. Land & Carbon Lab and World Resources Institute are now working on a method to monitor annual changes in tropical tree cover (loss and gain) at a 10-meter scale for 2017 onwards.
Agriculture and Food Systems Monitoring
Global Cropland Change
Geographic coverage: Global | Resolution: 30x30 meters |
Temporal coverage: 2000 - 2019 | Data partner: University of Maryland
Food crops are critical for feeding a growing global population, but much of this expansion comes at the expense of forests and other natural ecosystems that are critical to biodiversity and keeping global warming below 1.5 degrees C.
Land & Carbon Lab has partnered with the University of Maryland to produce the first global map of cropland extent and change at 30-meter resolution for 2000 through 2019. The cropland extent maps are available for five time periods (2000-2003, 2004-2007, 2008-2011, 2012-2015, and 2016-2019) and show gross change between time periods as well as net change between the first and last time periods.
This research shows that over the last 20 years, 102 million hectares of land — an area the size of Egypt — has been converted to crops. It also provides valuable insight into where croplands are expanding the fastest and where they are expanding into natural ecosystems, and illustrates the need to decouple crop production from ecosystem conversion to meet climate and biodiversity goals.
Global Grassland and Pasture Monitoring
Coming soon
Grasslands are the single most extensive land cover on Earth’s ice-free land surface, and grazing is the most extensive land use. These ecosystems are incredibly biodiverse, play an important role in mitigating climate change and offer sustenance, income, cultural identity and essential ecosystem services to many local communities. Despite their significance, these lands are currently poorly understood and often overlooked in global discussions on nature-based solutions.
The Global Pasture Watch consortium, led by Land & Carbon Lab and made up of experts in geospatial monitoring, machine learning, ecology and agriculture across an array of institutions, is actively developing 30-meter resolution mapping products for pasture areas and livestock from 2000 onwards.
This research will help decision makers protect, restore and sustainably manage these lands and enhance our understanding of land use conversion, food production, climate change and land productivity at a global scale.
The inaugural collection of global pasture monitoring data is due to be released in late 2024. Prior to this, we will release a beta version of the data for testing and feedback to select partners. If you are interested in beta testing or if you have specific challenges our data can help address, please contact us.
Global Cropland Yield and Area
Coming soon
Croplands cover 1.2 billion hectares of land, approximately 12% of Earth’s surface. The crops they produce are essential for communities around the globe — from direct food to animal feed, seeds, industrial processing and more. But to sustain a growing population without expanding croplands into vital natural ecosystems, we need to develop more efficient crop systems that create more crops on less land.
Land & Carbon Lab has partnered with experts from the University of Minnesota and the International Food Policy Research Institute (IFPRI) to develop, for the very first time, a globally consistent timeseries of crop area, yield, use and change for the years 2010-2020 for 25 of the world’s most common crops. This research will allow companies, government, and researchers to improve food security while protecting valuable ecosystems, improve food production models and monitor their commitments to creating conversion free supply chains. It will also improve land sector greenhouse gas accounting.
This work will build on the success of IFPRI’s Spatial Production Allocation Model. Initially introduced in 2008, using data from 2000, this innovative product has since undergone multiple updates. The latest iteration, developed with support from Land & Carbon Lab, features data from 2020 and represents a significant leap forward. It encompasses 46 crops and crop groups, meticulously mapped at a ten-kilometer resolution globally. The result is global data showing patterns of crop areas and yields, which provides critical information for sustainable supply chains, water stress models and our understanding of agricultural production worldwide.
Carbon Monitoring
Forest Carbon Fluxes
Geographic coverage: Global | Resolution: 30x30 meters |
Temporal coverage: 2000 - 2023
Forests play a critical role in our fight against climate change, absorbing about twice as much carbon dioxide as they emitted globally between 2001 and 2019. But we cannot maintain or enhance the forest carbon sink and reduce emissions from deforestation without knowing where these fluxes are occurring.
Land & Carbon Lab’s global forest carbon flux model provides spatially explicit data on forests’ emissions, removals (sequestration) and net carbon fluxes. Whereas previous models only estimate forest carbon fluxes for entire countries or regions or just the net change in forest carbon stocks, the 30-meter resolution of this data allows users to estimate gross emissions and removals in forests down to the local scale.
The model applies the Intergovernmental Panel on Climate Change’s (IPCC) guidelines for national greenhouse gas inventories to geospatial data. The inputs to the model include ground, airborne and satellite observations of forest change, disturbances (like deforestation, harvest or burning) and the type, age and condition of forests being lost or gained.
Researchers can use the resulting maps, available on Global Forest Watch, to generate local, national, regional or global-scale estimates of forest carbon fluxes. This data will be updated annually, and we aim to expand this data set to include all land-related carbon emissions and removals.
Global Forest Carbon Stocks: Past, Present and Future
Coming soon
Forests are a key component of the global carbon cycle, removing approximately 30% of anthropogenic carbon emissions every year. Managing forests has immense potential for climate change mitigation, but requires precise knowledge about where and how quickly forests are removing carbon from the atmosphere over time.
Land & Carbon Lab is working with the Forest Advanced Computing and Artificial Intelligence (FACAI) Lab at Purdue University to develop an AI-based forest growth model deployed at a global scale. The project will generate global maps of past, present and future tree densities, basal areas and carbon stocks in the world’s forests annually from the year 2000 to 2020, and project those forest properties through the year 2070 under different climate change scenarios.
The maps will be developed using state-of-the-art models by combining detailed ground measurements from more than 1.2 million forest inventory plots worldwide with geospatial data and machine learning algorithms. Beyond accurately quantifying carbon dynamics, this forest growth model will also delineate forest types and capture the dynamics of tree species diversity, helping us better understand the role our planet’s forests play in local nature-based solutions to mitigate global climate change.
A first release of this data for North America is expected to be completed later this year. Global data will be released in 2025.
Cropland and Livestock Emissions
Coming soon
Agriculture – including croplands and livestock – is a major source of global greenhouse gas emissions, but is essential for feeding a growing global population. As the population continues to grow, we need to find ways of growing food more sustainably while at the same time reducing the greenhouse gas emissions associated with crop and livestock systems.
Land & Carbon Lab is working with leading scientists at Cornell University to develop updated global crop and livestock emissions maps. This data will provide critical information for developing targeted, sustainable solutions and more accurate greenhouse gas accounting for the agricultural sector and its wide-ranging contribution to the global food system.
This work will build on a previously published approach and will include maps of global livestock emissions including methane, nitrous oxide for the year 2020. These emissions maps will also consider feed consumption, meat and milk production, manure, nitrogen excretion, GHG efficiency, production value and nutritional value.
It will also provide updated maps of global annual cropland emissions and emissions/production intensity for 42 of the world’s most common crop types, including factors such as fertilizer rates, yields, soil maps, climate variables and peatland drainage.
This research is ongoing. An updated collection of the data for 2020 is expected to be completed by fall 2024.
Natural Ecosystem Monitoring
Natural Lands Map
Geographic coverage: Global | Resolution: 30x30 meters |
Temporal coverage: 2020 | Data partner: Science Based Targets Network
Natural lands around the world are being converted and degraded at unprecedented levels. Three-quarters of land has been significantly impacted by humans through pollution, urban expansion, conversion to crop or livestock production, intensive logging in natural forests and other activities. These changes contribute to climate change, threatening biodiversity and disproportionately impacting vulnerable communities.
To help solve this problem, companies that directly operate within or source products from land need to set ambitious targets to avoid the conversion of natural ecosystems in their supply chains. To facilitate this, Land & Carbon Lab has partnered with the Science Based Targets Network (SBTN) to develop a Natural Lands Map that distinguishes natural from non-natural land cover and provides companies with a common baseline to assess whether they may be responsible for conversion since 2020, and to identify which lands should be protected from future expansion.
For measuring conversion of natural lands, we have used Accountability Framework Initiative (AFi) definitions of natural ecosystems and natural forests to align with standard-setting initiatives and monitoring tools such as Global Forest Watch Pro and the Greenhouse Gas Protocol, as well as translating definitions of natural ecosystems into quantifiable and mappable delineations of natural and non-natural lands.
This map is currently being tested by companies as part of SBTN’s initial target validation group. We are working to release a new version of the map in 2024 alongside SBTN’s guidance around land targets.
Stay up to date
Join our mailing list to stay up to date on the latest research and innovations.