Species Distribution Modeling Data

Biological Diversity and Ecological
Forecasting Data Pathfinder

Whether observing on the ground or from the air, it is almost impossible to count every member of a species. Because it is difficult to ascertain actual species distributions, another approach is to predict species distributions by identifying key environmental (i.e., abiotic) characteristics of suitable species habitats and then use models that incorporate both information on known occurrences of a species (i.e., presence data) along with certain environmental characteristics of those known habitats in which the species occurs to estimate potential habitats in locations where species occurrence data are lacking. In some cases, these models are augmented with information on known absences of a species, i.e., places where it is certain that the species of concern does not exist. The combination of presence and absence species occurrence data makes for even more powerful predictive models.

Species distribution models (SDMs) estimate the relationship between observed, in-situ species occurrences and the environmental and/or spatial characteristics of those locations. SDMs use raster-based layers such as land use/land cover, elevation, precipitation, temperature, and vegetation indices, as predictors of suitable habitats; this information is then combined with ground-collected presence data in statistical models to determine if a habitat is ideal for a particular species.

Conceptual diagram of how data are used in modeling to produce a species distribution map.

Conceptual diagram of how data are used in modeling to produce a species distribution map. Credit: Image used by permission of Dr. Jenica M. Allen, Mount Holyoke College

Numerous raster-based datasets (below) are available for model input. Note that, ideally, environmental data need to be in the same projection to get the most accurate results.

Species Occurrence Data

Species Occurrence Data

Photograph of a mule deer doe with twin fawns

Mule deer does usually give birth to one or two fawns in May or June. Although the fawns will be weaned by autumn, they will remain with their mother through their first winter. (Courtesy Lassen National Park Service)

There is a wealth of species occurrence data available; in fact, the Global Biodiversity Information Facility (GBIF), an international network and research infrastructure funded by the world's governments and aimed at providing anyone, anywhere, open access to data about all types of life on Earth, has close to 1.5 billion occurrence records. Even so, spatial, temporal, and taxonomic gaps, as well as other limitations, mean that much more data are needed. For example, GBIF records represent roughly 60% of named species, and many of these species either do not have a sufficient number of observations, or the observations are not usable, to support SDMs. Researchers using SDMs filter the records carefully so that only those of sufficient quality are used in their models.

Recently, developments in in-situ data collection have changed the way ecologists can detect species. “Transmission tags on terrestrial and aquatic animals, camera traps, sound recording devices, remotely piloted drones, and collections of environmental DNA in fresh and salt waters and soils are directly observing organisms, even the microbial components of biodiversity.” (Turner, 2014).

A relatively new source of observations comes from citizen science. In particular, the smartphone application iNaturalist has led to a surge in contributed observations and now account for a large percentage of new species occurrence observations. These observations are vetted for quality by experts.

Some additional sources of NASA-supported in-situ data, categorized by variable, are shown below:

Variable Data Source Repository
Distribution Telemetry Movebank
Acoustics Wildlife Insights
eDNA USGS Aquatic Invasive Species eDNA Database - in Development
Cal eDNA
Camera Trap Wildlife Insights
Field Observations iNaturalist
eBird
Map Of Life
Movement Mark-Recapture Movebank
Telemetry Movebank
Biologger/Biotags Movebank
Camera Trap Wildlife Insights
Accelerometers Movebank
Abundance/Demographics Acoustics Wildlife Insights
Physiology Accelerometers Movebank
Photos/Field Spectrometers EcoSIS

Back to Top

Land Surface Characteristics

Land Surface Characteristics

The Modern-Era Retrospective analysis for Research and Applications version 2 (MERRA-2) is a NASA atmospheric reanalysis for the satellite era using the Goddard Earth Observing System Model, Version 5 (GEOS-5) with its Atmospheric Data Assimilation System (ADAS). The MERRA project focuses on historical climate analyses for a broad range of weather and climate time scales and places the NASA suite of observations in a climate context. MERRA covers the period 1979–present, continuing as an ongoing climate analysis as resources allow. For more information on the various parameters in the land surface diagnostics product, see the Merra-2 File Specification.

Rootzone soil wetness from the Modern-Era Retrospective analysis for Research and Applications version 2 (MERRA-2) land surface diagnostics data product.

Rootzone soil wetness from the Modern-Era Retrospective analysis for Research and Applications version 2 (MERRA-2) land surface diagnostics data product.

Research quality data products can be accessed directly from Earthdata Search or the Goddard Earth Sciences Data and Information Services Center (GES DISC) data pool:

Data products can be visualized as a time-averaged map, an animation, seasonal maps, scatter plots, or a time series through an online interactive tool, Giovanni. Follow these steps to plot data in Giovanni: 1) Select a map plot type. 2) Select a date range. Data are in multiple temporal resolutions, so be sure to note the start and end date to ensure you access the desired dataset. 3) Check the box of the variable in the left column that you would like to include and then plot the data. For more information on choosing a type of plot, see the Giovanni User Manual.

Data can be visualized in Worldview:

Back to Top

Land Cover Type/Dynamics

Land Cover Type/Dynamics

The Terra and Aqua MODIS Land Cover Type data product provides global land cover types at yearly intervals, derived from six different classification schemes (the MODIS Land Cover User Guide provides additional information on those schemes). The product is derived using supervised classifications of MODIS Terra and Aqua reflectance data. The supervised classifications then undergo additional post-processing that incorporate prior knowledge and ancillary information to further refine specific classes.

MODIS Land Cover Type as seen in the visualization tool Worldview.

MODIS Land Cover Type as seen in the visualization tool Worldview.

The Terra and Aqua MODIS Land Cover Dynamics data product provides global land surface phenology metrics at yearly intervals from 2001 to 2017, derived from time series of the 2-band EVI (EVI2). Vegetation phenology metrics at 500 m spatial resolution are identified for up to two detected growing cycles per year.

The Terra MODIS Vegetation Continuous Fields (VCF) data product is a global representation of surface vegetation cover as gradations of three ground cover components: percent tree cover, percent non-tree cover, and percent non-vegetated (bare). VCF products provide a continuous, quantitative portrayal of land surface cover at 250 m pixel resolution.

The Land Processes Distributed Active Archive Center’s (LP DAAC) Application for Extracting and Exploring Analysis Ready Samples (AρρEEARS) offers the ability to extract subsets, transform, and visualize MODIS and VIIRS vegetation-related data products.

The Oak Ridge National Laboratory DAAC (ORNL DAAC) subsetting tools provide a means to simply and efficiently access and visualize MODIS and VIIRS land cover types, phenology and VCF data products as well.

Back to Top

Land Surface Temperature

Land Surface Temperature

Satellite images show the relationship between the characteristics of a landscape, and day and night surface skin temperature. Heavily forested areas remain relatively cool throughout the day, while barren and arid areas can be tens of degrees warmer. These images were acquired in the early morning and afternoon of July 6, 2011.

Satellite images show the relationship between the characteristics of a landscape, and day and night surface skin temperature. Heavily forested areas remain relatively cool throughout the day, while barren and arid areas can be tens of degrees warmer. These images were acquired in the early morning and afternoon of July 6, 2011. Credit: NASA Earth Observatory

Land surface temperature is useful for monitoring changes in weather and climate patterns which can impact the suitability of an area for a specific species to live and function.

Research quality land surface temperature data products can be accessed directly from Earthdata Search or the LP DAAC Data Pool; MODIS and ASTER data are available as HDF and VIIRS and ECOSTRESS are available as HDF5:

To quickly extract a subset of ECOSTRESS, MODIS, or VIIRS data for your region of interest, use the LP DAAC AppEEARS tool or the ORNL DAAC subsetting tools.

Landsat data from USGS’ Earth Explorer are available via Earthdata Search. Note that you will need a USGS login to download the data.

Data can be visualized in Worldview:

Back to Top

Precipitation

Precipitation

Near real-time IMERG Early Run Half-Hourly Image, acquired on November 12, 2019.

Near real-time IMERG Early Run Half-Hourly Image, acquired on May 7, 2020. Credit: NASA.

NASA’s Precipitation Measurement Missions (PMM) provide a continuous long-term record (over 20 years) of precipitation data through the Tropical Rainfall Measuring Mission (TRMM) and the Global Precipitation Measurement (GPM) mission. GPM, a follow-on mission for TRMM, provides even more accurate measurements, improved detection of light rain and snow, and extended spatial coverage.

The products from TRMM and GPM are available individually and have also been integrated with data from a global constellation of satellites to yield improved spatial/temporal precipitation estimates providing a temporal resolution of 30 minutes (in the case of GPM). The integrated products are the TRMM Multi-satellitE Precipitation Analysis (TMPA) and the Integrated Multi-satellite Retrievals for GPM (IMERG). IMERG’s multiple runs accommodate different user requirements for latency and accuracy (Early = 4 hours, e.g., for flash flood events; Late = 12 hours, e.g., for crop forecasting; and Final = 3 months, with the incorporation of rain gauge data, for research).

NASA, in collaboration with other agencies, has also developed models of precipitation, incorporating satellite information with ground-based data when available. These models are part of the Land Data Assimilation System (LDAS), of which there is a global collection (GLDAS) and a North American collection (NLDAS). LDAS takes inputs of measurements like precipitation, soil texture, topography, and leaf area index and then uses those inputs to model output estimates.

Science quality data products can be accessed via Earthdata Search:

  • TMPA from Earthdata Search
    Rainfall estimate at 3 hours, 1 day, or near real-time (NRT) and accumulated rainfall at 3 hours and 1 day. Data are in HDF format and can be opened using Panoply. Data are available from 1997.
  • IMERG from Earthdata Search
    Early, Late, and Final precipitation data on the half hour or 1-day timeframe. Data are in NetCDF or HDF format and can be opened using Panoply. Data are available from 2000.

Data products can be visualized as a time-averaged map, an animation, seasonal maps, scatter plots, or a time series through an online interactive tool, Giovanni. Follow these steps to plot data in Giovanni: 1) Select a map plot type; 2) Select a date range. Data are in multiple temporal resolutions and multiple temporal coverages, so be sure to note the start and end date to ensure you access the desired dataset; 3) Check the box of the variable in the left column that you would like to include and then plot the data. For more information on choosing a type of plot, see the Giovanni User Manual.

Near real-time data can be accessed from Worldview.

Daymet is a collection of gridded estimates of daily weather parameters. It is modeled on daily meteorological observations. Weather parameters in Daymet include daily minimum and maximum temperature, precipitation, vapor pressure, radiation, snow water equivalent, and day length at 1 km resolution over North America, Puerto Rico, and Hawaii.

Daymet data can be retrieved in a variety of ways, including: Earthdata Search; an ORNL DAAC API; ORNL DAAC tools; and through LP DAAC AppEEARS.

Back to Top

Soil Moisture

Soil Moisture

Soil moisture is important for understanding ecosystem productivity, as it controls the amount of water that can infiltrate the ground, replenish aquifers, and contribute to excess runoff. Current ground measurements of soil moisture are sparse and have limited coverage; satellite data help fill in those gaps. On the other hand, satellite data are limited by their relatively coarse resolution; the preferred measurement should be chosen based upon your needs. Utilizing a combination of both ground-based and remote sensing data provides for spatial and temporal data continuity.

NASA's Soil Moisture Active Passive (SMAP) satellite measures the moisture in the top five cm of soil globally, every 2–3 days, at a resolution of 9–36 km. NASA, in collaboration with other agencies, has also developed models of soil moisture content, incorporating satellite information with ground-based data when available. These models are part of the LDAS, of which there is a global collection (GLDAS) and a North American collection (NLDAS). LDAS takes inputs of measurements like precipitation, soil texture, topography, and leaf area index and then uses those inputs to model output estimates of soil moisture and evapotranspiration.

Science quality data products can be accessed via Earthdata Search; datasets are available as HDF5 (SMAP) files which are also customizable to GeoTIFF.

Soil moisture as visualized with ORNL DAAC Soil Moisture Visualizer. The map shows a flight path over Arizona in 2013. In the graph, AirMoss rootzone soil moisture data is plotted with SMAP rootzone soil moisture. Root zone soil moisture (RZSM) is the daily average of measurements at 0-100 cm depth.

Soil moisture as visualized with ORNL DAAC Soil Moisture Visualizer. The map shows a flight path over Arizona in 2013. In the graph, AirMoss rootzone soil moisture data is plotted with SMAP rootzone soil moisture. Root zone soil moisture (RZSM) is the daily average of measurements at 0-100 cm depth.

ORNL DAAC's Soil Moisture Visualizer integrates ground-based, SMAP, and other soil moisture data into a visualization and data distribution tool. LP DAAC's AppEEARS offers another option to simply and efficiently extract subsets, transform, and visualize SMAP data products. See the Tools for Data Access and Visualization section for additional information.

Data products can be visualized as a time-averaged map, an animation, seasonal maps, scatter plots, or a time series through an online interactive tool, Giovanni. Follow these steps to plot data in Giovanni: 1) Select a map plot type. 2) Select a date range. Data are in multiple temporal resolutions and multiple temporal coverages, so be sure to note the start and end date to ensure you access the desired dataset. 3) Check the box of the variable in the left column that you would like to include and then plot the data. For more information on choosing a type of plot, see the Giovanni User Manual.

Data can be visualized in Worldview:

Back to Top

Snow Cover

Snow Cover

Snow cover is the presence of snow on land or bodies of water; measurements are acquired during the daytime and under cloud-clear conditions. The Terra and Aqua MODIS instruments measure snow cover.

Research quality data products can be accessed via Earthdata Search; datasets are available as HDF files which can be opened using Panoply.

Data can be visualized in Worldview:

LP DAAC AppEEARS offers another option to simply and efficiently extract subsets, transform, visualize, and download MODIS Snow Cover data products.

Back to Top

Topography/Elevation

Topography/Elevation

Topography can play an important role in species distribution as biodiversity changes with elevation. According to Biodiversity Atlas of Los Angeles, “common patterns of biodiversity can be observed with changes in elevation, including: a) decreasing species richness (i.e. the number of species) as elevation increases and b) a peak in richness at intermediate elevations, with lower richness at the base and top of the mountain. An increase in species richness with elevation is very rare.”

An ASTER GDEM image of Mt. Raung and the surrounding area.

An ASTER GDEM image of Mt. Raung and the surrounding area. Image Credit: Land Processes Distributed Active Archive Center

A method for delineating topography is the Shuttle Radar Topography Mission (SRTM). SRTM provides a digital elevation model of all land between 60 degrees north and 56 degrees south, about 80% of Earth’s landmass. The spatial resolution is 30 m in the horizontal plane. The ASTER Global Digital Elevation Model (GDEM) coverage spans from 83 degrees north latitude to 83 degrees south, encompassing 99% of Earth's landmass. The spatial resolution is 30 m in the horizontal plane.

On average, compared to geodetic points over the U.S., SRTM data has a lower root mean square error (RMSE); RMSE is a commonly used method to express vertical accuracy of elevation datasets. Digital elevation model data accuracy is typically very sensitive to vegetation cover, however. ASTER tends to perform better over certain landcover types.

February 2020, LP DAAC released a new data product, NASADEM, available at 1 arc second resolution. NASADEM extends the legacy of the SRTM by improving the DEM height accuracy and data coverage as well as providing additional SRTM radar-related data products. The improvements were achieved by reprocessing the original SRTM radar signal data and telemetry data with updated algorithms and auxiliary data not available at the time of the original SRTM processing.

Back to Top

Sea Surface Temperature/Sea Surface Salinity

Sea Surface Temperature/Sea Surface Salinity

Satellite imagery can be used in the monitoring of marine ecosystems, from algal bloom development to animal migrations to coastal habitat destruction. There are specific Data Pathfinders pertinent to some of these topics. The Water Quality Data Pathfinder provides information on chlorophyll concentration or ocean color data. The forthcoming Disasters-Cyclones Data Pathfinder will provide information on the parameters in cyclone formation, as well as available observations assessing impacts post-storm.

Certain parameters, like sea surface temperature and salinity, are useful for determining habitat suitability for marine life as well as mapping ocean current flow patterns. These data can also be incorporated into general circulation models and biogeochemistry models important for forecasting animal movement, harmful algal blooms, and changes in climate.

Research quality data products, can be accessed via Earthdata Search:

Data products can be visualized as a time-averaged map, an animation, seasonal maps, scatter plots, or a time series through an online interactive tool, Giovanni. Follow these steps to plot data in Giovanni: 1) Select a map plot type. 2) Select a date range. Data are in multiple temporal resolutions, so be sure to note the start and end date to ensure you access the desired dataset. 3) Check the box of the variable in the left column that you would like to include and then plot the data. For more information on choosing a type of plot, see the Giovanni User Manual.

  • Aqua MODIS SST data in Giovanni
    Data products from MODIS on the Aqua satellite at 4 km resolution provided at both 8-day and monthly temporal resolutions.

Data can be visualized in Worldview:

NASA's SMAP satellite delivers derived sea surface salinity (SSS) observations for the world’s oceans, in addition to soil moisture. Algorithm development from the Aquarius mission is applied to SMAP to retrieve SSS. The primary SMAP salinity datasets include a Level 2 orbital dataset, in which data granules contain both the ascending and descending arcs of the orbit, and two Level 3 gridded datasets: an 8-day running average (linked to the day repeat cycle of SMAP) and monthly average.

Research quality (higher-level “standard”) data products, can be accessed via Earthdata Search:

Data can be visualized in Worldview:

Back to Top

Tools for Data Access and Visualization

Tools for Data Access and Visualization

Earthdata Search | Panoply | Giovanni | Worldview | AppEEARS | Soil Moisture Visualizer | MODIS/VIIRS Subsetting Tools Suite | Spatial Data Access Tool (SDAT)| Sentinel Toolbox

Earthdata Search is a website for data discovery of Earth Observation data collections from NASA’s Earth Observing System Data and Information System (EOSDIS), as well as U.S and international agencies across the Earth science disciplines. Users (including those without specific knowledge of the data) can search for and read about data collections, search for data files by date and spatial area, preview browse images, and download or submit requests for data files, with customization for select data collections.

Screenshot of the Search Earthdata site.


In the project area, you can customize your granule. You can reformat the data and output as HDF, NetCDF, ASCII, KML, or a GeoTIFF. You can also choose from a variety of projection options. Lastly you can subset the data, obtaining only the bands that are needed.

Earthdata Search customization tools diagram.

Panoply

HDF and NetCDF files can be viewed in Panoply, a cross-platform application that plots geo-referenced and other arrays. Panoply offers additional functionality, such as slicing and plotting arrays, combining arrays, and exporting plots and animations.

Giovanni

Giovanni is an online environment for the display and analysis of geophysical parameters. There are many options for analysis. The following are the more popular ones.

  • Time-averaged maps are a simple way to observe the variability of data values over a region of interest.
  • Map animations are a means to observe spatial patterns and detect unusual events over time.
  • Area-averaged time series are used to display the value of a data variable that has been averaged from all the data values acquired for a selected region for each time step.
  • Histogram plots are used to display the distribution of values of a data variable in a selected region and time interval.

For more detailed tutorials:

  • Giovanni How-To’s on NASA's GES DISC YouTube channel.
  • Data recipe for downloading a Giovanni map, as NetCDF, and converting its data to quantifiable map data in the form of latitude-longitude-data value ASCII text.

Worldview

NASA’s Worldview mapping application provides the capability to interactively browse over 900 global, full-resolution satellite imagery layers and then download the underlying data. Many of the available imagery layers are updated within three hours of observation, essentially showing the entire Earth as it looks “right now.” This supports time-critical application areas such as wildfire management, air quality measurements, and flood monitoring. Imagery in Worldview is provided by NASA’s Global Imagery Browse Services (GIBS). Worldview now includes nine geostationary imagery layers from GOES-East, GOES-West and Himawari-8 available at ten minute increments for the last 30 days. These layers include Red Visible, which can be used for analyzing daytime clouds, fog, insolation, and winds; Clean Infrared, which provides cloud top temperature and information about precipitation; and Air Mass RGB, which enables the visualization of the differentiation between air mass types (e.g., dry air, moist air, etc.). These full disk hemispheric views allow for almost real-time viewing of changes occurring around most of the world.

Worldview data visualization of the nighttime lights in Puerto Rico pre- and post- Hurricane Maria, which made landfall on September 20, 2017. Post-hurricane image shows widespread outages around San Juan, including key hospital and transportation infrastructure.

Worldview data visualization of the nighttime lights in Puerto Rico pre- and post- Hurricane Maria, which made landfall on September 20, 2017. The post-hurricane image on the left shows widespread outages around San Juan, including key hospital and transportation infrastructure.

AppEEARS

AppEEARS, from LP DAAC, offers a simple and efficient way to access and transform geospatial data from a variety of federal data archives. AppEEARS enables users to subset geospatial datasets using spatial, temporal, and band/layer parameters. Two types of sample requests are available: point samples for geographic coordinates and area samples for spatial areas via vector polygons.

Performing Area Extractions

After choosing to request an area extraction, you will be taken to the Extract Area Sample page where you will specify a series of parameters that are used to extract data for your area(s) of interest.

Spatial Subsetting

You can define your region of interest in three ways:

  • Upload a vector polygon file in shapefile format (you can upload a single file with multiple features or multipart single features). The .shp, .shx, .dbf, or .prj files must be zipped into a file folder to upload.
  • Upload a vector polygon file in GeoJSON format (can upload a single file with multiple features or multipart single features).
  • Draw a polygon on the map by clicking on the Bounding box or Polygon icons (single feature only).

Select the date range for your time period of interest.

Specify the range of dates for which you wish to extract data by entering a start and end date (MM-DD-YYYY) or by clicking on the Calendar icon and selecting dates a start and end date in the calendar.

Adding Data Layers

Enter the product short name (e.g., MOD09A1, ECO3ETPTJPL), keywords from the product long name, a spatial resolution, a temporal extent, or a temporal resolution into the search bar. A list of available products matching your query will be generated. Select the layer(s) of interest to add to the Selected layers list. Layers from multiple products can be added to a single request. Be sure to read the list of available products available through AppEEARS.

Extracting an area in AppEEARS

Selecting Output Options

Two output file formats are available:

  • GeoTIFF
  • NetCDF-4

If GeoTIFF is selected, one GeoTIFF will be created for each feature in the input vector polygon file for each layer by observation. If NetCDF-4 is selected, outputs will be grouped into .nc files by product and by feature.

If GeoTIFF is selected, you must select a projection

Interacting with Results

Once your request is completed, from the Explore Requests page, click the View icon in order to view and interact with your results. This will take you to the View Area Sample page.

The Layer Stats plot provides time series boxplots for all of the sample data for a given feature, data layer, and observation. Each input feature is renamed with a unique AppEEARS ID (aid). If your feature contains attribute table information, you can view the feature attribute table data by clicking on the Information icon to the right of the Feature dropdown. To view statistics from different features or layers, select a different aid from the Feature dropdown and/or a different layer of interest from the Layer dropdown.

Interpreting Results in AppEEARS

Be sure to check out the AppEEARS documentation to learn more about downloading the output GeoTIFF or NetCDF-4 files.

Soil Moisture Visualizer

ORNL DAAC has developed a Soil Moisture Visualizer tool (read about it at Soil Moisture Data Sets Become Fertile Ground for Applications) that integrates a variety of different soil moisture datasets over North America. The visualization tool incorporates in-situ, airborne, and remotely-sensed data into one easy-to-use platform. This integration helps to validate and calibrate the data, and provides spatial and temporal data continuity. It also facilitates exploratory analysis and data discovery for different groups of users. The Soil Moisture Visualizer offers the capability to geographically subset and download time series data in .csv format. For more information on the available datasets and use of the visualizer, view the Soil Moisture Visualizer Guide.

To use the visualizer, select a dataset of interest under Data. Depending on the dataset chosen, the visualizer provides the included latitude/longitude or an actual site location name and relative time frames of data collection. Upon selection of the parameter, the tool displays a time series with available datasets. All measurements are volumetric soil moisture. Surface soil moisture is the daily average of measurements at 0-5 cm depth, and root zone soil moisture (RZSM) is the daily average of measurements at 0-100 cm depth. Lastly it provides data sources for download.

ORNL DAAC Soil Moisture Visualizer

The Soil Moisture Visualizer allows users to compare soil moisture measurements from multiple sources (figure legends, top left and bottom right) at the same location. In this screenshot, Level 4 Root Zone Soil Moisture (L4 RZSM) data from NASA’s Soil Moisture Active Passive (SMAP) Observatory are shown with data from in situ sensors across the 9-kilometer Equal-Area Scalable Earth (EASE) grid cell encompassing the Tonzi Ranch Fluxnet site in the Sierra Nevada foothills of California. Daily precipitation values for the site (purple spikes) are also provided for reference.

MODIS/VIIRS Subsetting Tools Suite

ORNL DAAC also has several MODIS and VIIRS Subset Tools for subsetting data.

  • With the Global Subset Tool, you can request a subset for any location on earth, provided as GeoTiff and in text format, including interactive time-series plots and more. Users specify a site by entering the site's geographic coordinates and the area surrounding that site, from one pixel up to 201 x 201 km. From the available datasets, you can specify a date and then select from MODIS Sinusoidal Projection or Geographic Lat/Long. You will need an Earthdata account to request data.
  • With the Fixed Subsets Tool, you can download pre-processed subsets for 3000+ field and flux tower sites for validation of models and remote sensing products. The goal of the Fixed Sites Subsets Tool is to prepare summaries of selected data products for the community to characterize field sites. It includes sites from networks such as NEON, ForestGeo, PhenoCam and LTER that are of relevance to the biodiversity community.
  • With the Web Service, you can retrieve subset data (in real-time) for any location(s), time period, and area programmatically using a REST web service. Web service client and libraries are available in multiple programming languages, allowing integration of subsets into users' workflow.

Directions for subsetting data with the ORNL DAAC MODIS and VIIRS subset tool

Spatial Data Access Tool (SDAT)

ORNL DAAC’s SDAT is an Open Geospatial Consortium (OGC) standards-based Web application to visualize and download spatial data in various user-selected spatial/temporal extents, file formats, and projections. Several data sets including land cover, biophysical properties, elevation, and selected ORNL DAAC archived data are available through SDAT. KMZ files are also provided for data visualization in Google Earth.

Within SDAT, select a dataset of interest. Upon selection, the map service will open displaying the various measurements, with the associated granule, and a visualization of the selected granule.

Canopy Height, Kalimantan Forests, Indonesia, 2014 from the Oak Ridge National Laboratory Distributed Active Archive Center Spatial Data Access Tool.

Canopy Height, Kalimantan Forests, Indonesia, 2014 from the Oak Ridge National Laboratory Distributed Active Archive Center Spatial Data Access Tool.

You can then select your spatial extent, projection, and output format for downloading.

Canopy Height, Kalimantan Forests, Indonesia, 2014 from the Oak Ridge National Laboratory Distributed Active Archive Center Spatial Data Access Tool with various output options.

Canopy Height, Kalimantan Forests, Indonesia, 2014 from the Oak Ridge National Laboratory Distributed Active Archive Center Spatial Data Access Tool with various output options.

Sentinel Toolbox

Before choosing data, it’s important to determine which band meets your needs, as radar signals penetrate deeper as the sensor wavelength increases. This difference in penetration is due to the dielectric properties of a given medium, which dictate how much of the incoming radiation scatters at the surface, how much signal penetrates into the medium, and how much of the energy gets lost to the medium through absorption.

SAR signal penetration by sensor wavelength λ

SAR signal penetration by sensor wavelength λ. Credit: NASA SAR Handbook.

Note that for biomass estimation, L-band and P-band sensors are preferred over higher frequencies and smaller wavelengths for two reasons: 1) at these bands, the radar waves or energy can penetrate the tree canopy and scatter from larger woody components of the forest, and 2) the scattering from larger tree components, unlike leaves, are more stable temporally and remain highly coherent over the acquisition period in the case of repeated measurements for change detection or interferometric applications (adapted from SAR Handbook, 2019).

The C-band can be used for low-vegetation biomass such as grasslands, shrublands, sparse woodlands, young secondary regeneration, and low-density wetlands.

Another important parameter to take into consideration when choosing a dataset is the polarization, or the direction in which the signal is transmitted and/or received: horizontally or vertically. Dual polarization, for example, refers to two different signal directions, horizontal/vertical and vertical/horizontal (HV and VH). Knowing the polarization from which a SAR image was acquired is important, as signals at different polarizations interact differently with objects on the ground, affecting the recorded radar brightness in a specific polarization channel.

Strong scattering in HH indicates a predominance of double-bounce scattering (e.g., stemmy vegetation, manmade structures), while strong VV relates to rough surface scattering (e.g., bare ground, water), and spatial variations in dual polarization indicate the distribution of volume scatterers (e.g., vegetation and high-penetration soil types such as sand or other dry porous soils).

Strong scattering in HH indicates a predominance of double-bounce scattering (e.g., stemmy vegetation, manmade structures), while strong VV relates to rough surface scattering (e.g., bare ground, water), and spatial variations in dual polarization indicate the distribution of volume scatterers (e.g., vegetation and high-penetration soil types such as sand or other dry porous soils). Credit: NASA SAR Handbook.

SAR data are complex, requiring a certain level of processing skill.

The European Space Agency (ESA) Sentinel-1 Mission consists of two satellites, Sentinel-1A and Sentinel-1B, with synthetic aperture radar (SAR) instruments operating at a C-Band frequency. They orbit 180° apart, together imaging the entire Earth every six days. SAR is an active sensor which can penetrate cloud cover and vegetation canopy and can observe at night; therefore it is ideal for flood inundation mapping. It also provides information useful in detecting the movement of Earth material after an earthquake, volcanic eruption, or landslide. SAR data are very complex to process, but ESA has developed a Sentinel-1 Toolbox to aid with processing and analysing Sentinel-1 data. For more information on active sensors, see What is Remote Sensing?; for more information on SAR specifically, see What is Synthetic Aperture Radar?

Once you have downloaded the needed SAR data, it must be calibrated to account for distortion in the data. The objective in performing calibration is to create an image where the value of each pixel is directly related to the backscatter of the surface. So calibration takes into account radiometric distortion, signal loss as the wave propagates, saturation, and speckle. This process is critical for analyzing images quantitatively; it is also important for comparing images from different sensors, modalities, processors, and different acquisition dates.

Screenshot of the Sentinel-1 toolbox

Important note: DO NOT unzip the downloaded SAR file. Open the .zip file from within the Sentinel Toolbox. When you expand the Bands folder, you will see an amplitude and an intensity file for each polarization option. (The intensity band is a virtual one and is the square of the amplitude.) Open the amplitude file. Subset the data by zooming in to the area of interest and right-clicking on “Spatial Subset from View.”

Calibration is done by following these steps:

  1. Radiometric calibration is performed by selecting Radar/Radiometric/Calibration (leave parameters as default)
  2. Geometric correction is done next to fix the main geometric distortions, due to: Slant Range, Layover, Shadow, and Foreshortening. Terrain correction can be performed by selecting Radar/Geometric/Terrain Correction/ Range-Doppler Terrain Correction. This requires a digital elevation model (within the processing parameters, SRTM is the default selection). You can also specify a map projection in the processing parameters.

Sentinel-1 Toolbox Geometric Correction

Another characteristic of SAR images that must be accounted for is speckle. Speckle is the grey level variation that occurs between adjacent resolution cells, creating a grainy texture. Within the Toolbox, speckle can be removed by selecting “Radar/Speckle Filtering/Single Product Speckle Filter,” and then choosing a type of filter; “Lee” is one of the most common.

Comparison of speckle in SAR imagery within Sentinel-1 Toolbox

Change Detection

One approach for monitoring change detection, caused by forest degradation or deforestation, is the log-ratio scaling method. You will need two images for which you have completed the steps above. The images must be from the same season. This is important for change detection operations as it avoids seasonal changes and focuses on true environmental changes in a change detection analysis.

Log-ratio image with the ArcMap Imagery basemap

The resulting log-ratio image over Huntsville, Alabama, was created from a pair of images acquired on 7/17/2009 and 9/04/2010, approximately one year apart. In the log-ratio image, unchanged features have intermediate gray tones (gray value around zero) while change features are either bright white or dark black. Black features indicate areas where radar brightness decreased while in white areas, the brightness has increased. Credit: ASF DAAC 2017

For further information on SAR change detection, see the Alaska Satellite Facility DAAC (ASF DAAC) change detection recipe for QGIS or change detection recipe for ArcGIS. The SERVIR SAR Handbook also contains tutorials on change detection, developing time series and making RGB composites; these are provided as Python scripts in Jupyter notebooks.

Often-used color scheme for multi-dimensional false color SAR composites

Another option for change detection is to create an RGB composite. When creating RGB composites using SAR data, the example color-scheme is often used. Note that for forest applications in particular, it is always useful to assign cross-polarized (HV/VH) data to the green band as these data are more related to volume scattering of the canopies. Co-polarized data (VV or HH) are suited for the red band, where surface scattering components are more pronounced. When only dual-polarimetric data are available (HH/HV or VV/VH), a color SAR image is often constructed by assigning the ratio of co-polarized to cross-polarized data to the blue channel. For more information on this procedure, read the SAR Handbook Chapter 3.

Sentinel-1 C-band dual polarimetric VV and VH data: (a) VV, (b) VH, (c) VV/VH ratio, and (d) SAR false color composite with RGB = VV/VH/ratio channel assignment. Image acquired on May 31, 2018.

Sentinel-1 C-band dual polarimetric VV and VH data: (a) VV, (b) VH, (c) VV/VH ratio, and (d) SAR false color composite with RGB = VV/VH/ratio channel assignment. Image acquired on May 31, 2018. Credit: NASA SAR Handbook

Forest Biomass

There is strong synergy between ground and remote sensing measurements for quantifying Above Ground Biomass (AGB). However, before biomass estimation from SAR measurements can be acquired, SAR data must be processed such that pixel size, geometric attributes, and environmental effects are all normalized and radiometrically calibrated. In addition to the Sentinel toolbox, the ASF DAAC also has a free software tool, MapReady, which accepts Level 1 detected SAR data, single look complex SAR data, and optical data. MapReady can terrain-correct, geocode, and save the data file in several common imagery formats including GeoTIFF.

The SAR Handbook’s APPENDIX D Mapping Forest Biomass with Radar Remote Sensing – Chapter 5 Training Module provides detailed instructions for mapping forest biomass.

Back to Top

Published April 7, 2020

Page Last Updated: May 15, 2020 at 7:54 AM EDT