forest

                                Explaining Global Forest Watch's Tree Cover data sets 

                                                  

                                         

                                image source - Trees outside of dense forests in Ethiopia

New Earth observation data have significantly enhanced our ability to map tree cover, and automated processes and cheaper processing power have sped up the creation of data sets to help us better understand the distribution of trees around the world. Three tree cover data sets are hosted by Global Forest Watch (GFW), including the World Resources Institute's (WRI) most recent Tropical Tree Cover data (formerly known as Trees in Mosaic Landscapes).

These data sets differ in terms of resolution, geographic and temporal coverage, how they define tree cover, and other qualities that can make them each the most suitable for particular use cases. Here, we outline these variations and show how each of the three data sets contributes significantly to our understanding of the global and tropics-specific forest coverag

Canopy of Trees

The Tree Canopy Cover data collection, which maps tree canopy cover globally at a 30-meter resolution and is available on GFW for the years 2000 and 2010, was released in 2013 by scientists at the University of Maryland's (UMD) GLAD Lab. This data set was one of the first to use medium resolution imagery to map worldwide tree cover. Additionally, it created a baseline for the data on annual tree cover reduction; both data sets together formed the basis for the establishment of Global Forest Watch in 2014.

The density of tree canopy cover inside a 30-meter pixel is measured by the Tree Canopy Cover layer, which calculates % tree cover. The data set, which is obtained from Landsat imagery, defines trees as all vegetation that is at least five meters tall.  

Height of Tree Cover

The Tree Cover Height data collection, which maps tree cover height globally at a 30-meter resolution and is accessible on GFW for the years 2000 and 2020, was released in 2021 by experts at the UMD's GLAD Lab. The 20-year net change statistics on GFW are derived from this data collection, which enables more accurate detection of tree cover gain. Tree height fluctuates gradually and is linearly connected to forest biomass and carbon storage, in contrast to tree canopy cover, which reaches its maximum immediately after tree growth begins.  

The Tree Cover Height layer calculates the height of the tree canopy within a 30-meter pixel, which is measured in meters. The Global Ecosystem Dynamics Investigation (GEDI) LiDAR measurements and Landsat imagery were used to create the data collection, which defines trees as woody vegetation that is at least three meters tall. The upper bounds of height for pixels with partial tree cover or pixels with variable canopy height will be represented by this data collection, which predicts the 95th percentile of tree canopy height.

Coconut Tree Cover

 

The Tropical Tree Cover data collection, which maps tree cover throughout the tropics at a 10-meter resolution and is accessible on GFW for the year 2020, was revised this year by WRI (and is now available for download). The Tropical Tree Cover data set, in contrast to the other data sets, is based on pictures from Sentinel satellites, which were launched in 2015 and raised the resolution of openly available remote sensing data to 10 meters. In order to better detect tree cover in non-forested areas, particularly in drylands, urban areas, and on cropland, and our capacity to monitor trees at small spatial (local) scales, this data collection took use of the development in resolution.

On GFW, we display the tree extent data with a probability threshold of 40%. The Tropical Tree Cover calculates tree extent, which is the likelihood that one or more tree canopies intersect the middle point of a 10-meter pixel. By averaging tree extent probabilities at a half-hectare scale, "percent tree cover" is calculated. Users of GFW can access the Tropical Tree Cover data collection in both the 10-meter and half-hectare versions. Trees are defined as woody vegetation with a minimum canopy diameter of five meters and a height of more than five meters in the Tropical Tree Cover data set. Images from Sentinel 1 and Sentinel 2 were used to create the data. To learn more about Tropical, go to the Help Center.

What distinctions exist between these data sets, and what does this actually mean?

Various Spatial Resolutions  

While both the Tree Canopy Cover and Tree Cover Height data sets have a 30-meter resolution, the Tropical Tree Cover data collection has a better spatial resolution at 10 meters. The Tropical Tree Cover data set is better suited for monitoring tree cover in areas with dispersed tree cover, including dryland regions like the Sahel, forest perimeters and riparian zones, as well as natural trees in agricultural and urban areas. At 30-meter resolution, it can be challenging to distinguish small patches of trees. Since trees that are undiscovered at 30-meter resolution may be recognized at 10-meter resolution, these resolution discrepancies will lead to varied estimations of the amount of tree cover in a given area of interest.

DETECTION OF TREE COVER AT 10 AND 30 METER RESOLUTION

Geographic and Temporal Coverage Variations

The three data sets' annual availability and geographic coverage are two significant differences. There are data sets for the years 2000 and 2010 in the Tree Canopy Cover category. Originally created for the year 2019, the updated Tree Cover Height data collection is now accessible on GFW for the years 2000 and 2020. The year 2020 is the only one currently included in the Tropical Tree Cover data collection. Users may want to emphasize historical record (Tree Canopy Cover and Tree Cover Height) over resolution (Tropical Tree Cover) given the longer historical availability of the Landsat archive.

The Tree Cover Height and Tree Canopy Cover data sets have global coverage, whereas the Tropical Tree Cover data collection exclusively includes the tropics. For applications in temperate and boreal forests, the Tree Cover Height and Tree Canopy Cover data sets should be used, whereas the Tropical Tree Cover data set offers a higher resolution view of tropical forests.

Variations in Change Data

Since they are available for numerous years, GFW's forest change data sets are built upon the Tree Canopy Cover and Tree Cover Height data sets.

The Tree Canopy Cover data for the year 2000 can be used as the baseline for GFW's annual tree cover loss statistics. In order to meet definitions of forest that vary by tree canopy density, tree cover loss is estimated for various tree canopy density thresholds in the year 2000. Users can change this criteria to best fit their needs. GFW bases its data and maps on a minimum canopy density of 30% in the year 2000.The baseline for the 20-year tree cover gain and net change data sets that are available on GFW is the data on tree cover height.

Tropical Tree Cover is a static map for 2020 at this time, making it impossible to evaluate change directly. However, WRI intends to produce a yearly Tropical Tree Cover product on GFW in the future.

To avoid any discrepancies in definitions and methodologies, users whose main goal is quantifying tree cover change should utilize the baseline data set linked to the pertinent forest change product.

Tree Definitions with Variations

The Tropical Tree Cover data set features a variable height requirement of three to five meters based on canopy diameter, unlike the Tree Canopy Cover and Tree Cover Height data sets, which are based on minimum height criterion. As a result, tall herbaceous plants that are classified as trees in other data sets, such as sugarcane, bananas, and cactus, are not included in the Tropical Tree Cover data set.

The extent of tree cover observed by the three data sets will vary significantly for users who are measuring tree cover in regions with significant presence of these crops. The Tree Canopy Cover and Tree Cover Height data sets do not quantify forest area using the same standards as National Forest Inventories or the Food and Agriculture Organization's Forest Resource Assessment (FRA). This is an important distinction to make. Read our blog to find out more about the variations between how the FRA and GFW assess forests and tree cover, as well as other aspects of both data sets' purposes, horizons, and methodologies.

Which data set ought I to employ for this particular application?

In the future, how will these data sets be updated?

Future integrated tree canopy structure data sets published by UMD will allow users to evaluate yearly variations in canopy height and cover.

Users will be able to examine changes in tree cover across the tropics at the 10-meter scale thanks to WRI's ongoing development of yearly Tropical Tree Cover data for 2017–2022. Additionally, WRI is working on new pilot projects to enhance the identification of tree crops and generate carbon estimates for scaled-down restoration initiatives.

The types, heights, and densities of woods vary greatly from location to location. Different tree cover data sets can be better suited to achieve your aim depending on the region, date, and particular forest type. Using new developments in satellite technology and processing approaches, remote sensing data can assist us in monitoring a range of forest properties. As a result, there are many data sets that may be combined to better understand the distribution of trees around the world and offer special chances to promote reforestation and deforestation monitoring.

link source - https://www.globalforestwatch.org/blog/data-and-research/tree-cover-data-comparison/