Modeled dataset layers

Via the Lens Library, you have access to several modeled datasets that can be added to your account as layers. These are layers derived from remotely sensed images using various methods to classify or quantify for a specific purpose (examples: land use, tree canopy height, etc.). This article will summarize the available modeled dataset layers in Lens. To learn more about the Lens Library, see our article here.

This article covers:

Biodiversity Intactness

Created by Impact Observatory and Vizzuality, this data layer depicts the intactness of biodiversity globally for the years 2017-2020 at a 100-meter resolution. The intactness of biodiversity is based both on the abundance of individual species as well the species composition compared to an intact baseline. This dataset can help identify areas of critical remaining intact habitat and help support spatial planning and management. This layer is terrestrial and won't include data for areas of water. Like the MoBI layer, this will load beyond property boundaries so you can see the context of the larger area. Learn more about the methodology in this white paper.

Canopy Height

The Global Canopy Height dataset provides 1m resolution maps of canopy height, developed using self-supervised learning models trained on aerial imagery and GEDI Lidar. This dataset was created using sub-meter satellite imagery from Maxar between 2009 through 2020, making it a helpful supplement to baselines. This high-resolution data was developed to drive transparency and accountability in carbon markets and conservation efforts. More information can be found in this research paper.

Global Forest Loss

The University of Maryland Global Forest Change layer displays 30m data showing gross forest loss for each year from 2001 to 2022, derived from Landsat imagery. Areas where forest has transitioned to non-forest are during a given year are displayed in red on the map. Forest loss could be a result of natural or human-induced causes. This dataset makes it possible to assess the timing and annual extent of forest loss over the last two decades, which can support site assessments and baselining. Data and methodology were originally  published by Hansen et. al and this work is a collaboration between the Global Land Analysis & Discovery Lab at the University of Maryland, Google, USGS, and NASA.

Landcover Data from NLCD

This layer shows the National Land Cover Database (NLCD) classifications of land cover at 30m resolution based on a 16-class legend. The NLCD is developed by a partnership of multiple United States Federal agencies and the most recent available layer in Lens is from 2016. You can visualize this public dataset and landcover classifications for properties in Lens but Upstream Tech is not responsible for the accuracy of the classifications. The legend for this layer will appear on your screen when the layer is selected.

Landcover Data from ESA WorldCover

The ESA WorldCover global landcover layer from 2020 is based on data from Sentinel-1 and 2, and provides land cover classifications at 10m resolution globally. This data layer from the European Space Agency was developed by a consortium major European service providers and research organizations and includes 12 categories of land cover. You can visualize this public dataset and landcover classifications for properties in Lens but Upstream Tech is not responsible for the accuracy of the classifications. The legend for this layer will appear on your screen when the layer is selected.

Impact Observatory's Land Use Land Cover

This dataset, produced by Impact Observatory, Microsoft, and Esri, displays a global map of land use/land cover (LULC) derived from ESA Sentinel-2 imagery at 10m resolution for 2017 - 2022. Each year is generated from Impact Observatory’s deep learning AI land classification model using a massive training dataset of billions of human-labeled image pixels. The global maps were produced by applying this model to every Sentinel-2 scene, processing over 400,000 Earth observations per year. The algorithm generates LULC predictions for 9 classes globally. These classifications include Built, Crops, Trees, Water, Rangeland, Flooded Vegetation, Snow/Ice, Bare Ground, and Clouds. Additional information about Impact Observatory's global maps is available here and you can reach out to their team directly with any questions about the data and methodology. The legend for this layer will appear on your screen when the layer is selected.

Custom land cover data from Impact Observatory

In addition to the free Impact Observatory Landcover dataset, we also offer a premium version with more classifications and recent data. Impact Observatory's innovative AI-powered methods categorize land cover into 14 classes: Water Channel Extent, Variable Water, Persistent Water, Dense Trees, Sparse Trees, Dense Rangeland, Sparse Rangeland, Flooded Vegetation, Crops, High Density Built, Low Density Built, Bare Ground, Snow/Ice, Clouds. This custom data can be requested directly through Lens and costs $.02/acre, with a minimum spend of $500. The data will include the time period requested and one year prior for baseline comparison. To initiate a data request, click the "Request" button on the Custom Landcover tile in the Layers Library and fill out the linked form to get a quote for your area and time frame. 

Landcover (Brazil)

The MapBiomas Brasil dataset provides annual land use and land cover maps for Brazil from 2000 to 2022, created through a collaboration between NGOs, universities and technology companies. This dataset includes dozens of land cover categories and offers a comprehensive view of how Brazil's landscapes have transformed over time, making it an essential resource for monitoring environmental changes. The project is continually updated to improve accuracy and coverage. Detailed accuracy information and methodological insights are available here.

Landcover (Chaco)

The MapBiomas Chaco dataset offers annual land cover and land use maps for the Gran Chaco Americano, covering Argentina, Bolivia, Brazil, and Paraguay from 2000 to 2022. This initiative is a collaborative effort involving local experts in land use, satellite remote sensing, and GIS to understand the dynamic changes in this unique region. Using advanced cloud processing and automated classifiers on the Google Earth Engine platform, the project provides an open, accessible, and updated resource for monitoring environmental transformations in the Gran Chaco. More information about this dataset can be found here.

Natural Lands Classification (WRI)

The WRI Natural Lands Classification layer, part of the Science Based Targets Network (SBTN), provides a comprehensive global map of natural lands as of 2020, designed to support companies in setting "no conversion of natural ecosystems" targets. Developed by WRI's Land & Carbon Lab in collaboration with the World Wildlife Fund and Systemiq, this dataset combines the best available global and local geospatial data to establish a common baseline for visualizing natural land extent. The map helps companies assess their responsibility for land conversion post-2020 and make informed decisions about their operations and supply chains. See technical documentation here for more detail.

NatureServe's Map of Biodiversity Importance (MoBI)

The Map of Biodiversity Importance (MoBI) Layer provides information about the number of species in the continental US that are protected by the Endangered Species Act and/or considered to be in danger of extinction. This dataset is free and available to all Lens customers, and will be displayed beyond the property boundaries to enable you to contextualize your site within a larger area. The MoBI layer visualizes NatureServe's All Species Richness dataset which covers the continental United States - it's useful to understand the richness of species including vertebrates, freshwater invertebrates, pollinators, and plants. The richness value represents the number of species with habitat overlapping a given 990 meter cell. As a synthesis of predictive models, MoBI cannot guarantee either the presence or absence of imperiled species at a given location.

NatureServe has been an authoritative source for biodiversity data over many decades, collecting, analyzing, and delivering biodiversity knowledge that informs conservation action. To learn more about their MoBI dataset, see here. 

Precipitation Data from PRISM Climate Group

This data layer is available for Standard and Enterprise customers.

Monthly precipitation data from the PRISM Climate Group at Oregon State University offers insights into monthly patterns across the United States from 2010-present. The PRISM data uses a weighted regression scheme to account for complex climate regimes associated with orography, rain shadows, temperature inversions, slope aspect, coastal proximity, and other factors. By assimilating station data from various networks nationwide, PRISM provides a detailed understanding of precipitation which can be used to contextualize other Lens datasets, identify risks such as drought conditions, and support water resource management decisions. Learn more about PRISM datasets here.

Tropical Moist Forest Landcover

The annual change collection dataset, developed by the European Commission’s Joint Research Centre (JRC), provides annual classifications of tropical moist forest (TMF) land cover from 1990 to 2023, available in tropical forest regions only. It classifies the forest into six categories: intact/undisturbed, degraded, deforested, regrowing, water, and other. The dataset is a valuable tool for tracking changes in tropical forest conditions and monitoring changes such as deforestation, forest degradation, and regrowth over time. See here for more information.