Ecosystems are dynamic systems comprising living organisms that interact with each other and their physical environment (Odum, 1953). Ecosystems vary in scale, from microecosystems (for example, tidepools) to macroecosystems (for example, boreal taiga forests), and their global distributions are shaped by environmental drivers such as climate, terrain, and substrate properties (Bailey, 2009). Ecosystems can be mapped as areas with unique combinations of environmental factors, and the U.S. Geological Survey, Esri, and The Nature Conservancy (TNC) have taken this approach to develop a comprehensive map of World Terrestrial Ecosystems (WTEs; Sayre and others, 2020). This 250-meter spatial resolution map is globally standardized and is based on data that characterize land areas with unique combinations of climate regimes, landforms, and vegetation or land-cover types (Sayre and others, 2020).

In 2021, the United Nations (UN) Statistical Commission adopted the UN System for Environmental and Economic Accounting (UN SEEA; United Nations, 2021) as a standard for national ecosystem accounting; this standard emphasizes measuring ecosystem extents for policy and decision-making purposes. Ecosystem accounting requires area-based measurements of ecosystem distributions, commonly referred to as ecosystem extent. The UN SEEA guidance encourages the use of national ecosystems maps and data for country-level ecosystem accounting. Although some countries have detailed national ecosystem maps for conservation planning and territorial management (for example, Colombia; IDEAM and others, 2017), others may need to develop a national map from the bottom-up or use a national map derived from a regional or global ecosystems map.

National ecosystems maps may vary considerably based on mapping concept (structure-based mapping versus function-based mapping versus composition-based mapping), use of proxies (for example, land cover), and spatial and classification resolution, among other factors (Geller and others, 2023). To allow for standardized reporting and comparison of national ecosystem accounts, the UN SEEA has recommended using a reference ecosystem classification known as the International Union for Conservation of Nature (IUCN) Global Ecosystem Typology (GET). The GET consists of a list of global ecosystems and their descriptions that is hierarchical, based on ecosystem functional properties, and comprehensive across terrestrial, freshwater, and marine domains (Keith and others, 2022). The UN SEEA guidance identifies the third level in the GET hierarchy, Ecosystem Functional Groups (EFGs), as the standardized reporting unit recommended for ecosystem accounting. As such, if a country has an existing national ecosystem map it uses for ecosystem accounting, the results should be summarized as EFGs. If a country does not have a baseline national ecosystems map, the UN SEEA guidance suggests other resources that may be consulted. Notably, the WTEs produced collectively by the U.S. Geological Survey, Esri, and TNC (Sayre and others, 2020) are the first alternative, top-down, ecosystem data sources that the UN SEEA guidance suggests consulting (United Nations, 2021).

The United Nations Convention on Biological Diversity’s Kunming-Montreal Global Biodiversity Framework (CBD GBF) has also adopted the GET EFGs as its reference ecosystem conservation status reporting unit. Countries reporting progress towards achieving the “30 by 30” conservation goal (30 percent of the area of all ecosystems conserved by 2030) are expected to use the GET EFGs as the reporting unit, enabling consistent characterization of the conservation status of all of Earth’s ecosystems (CBD, 2024). Additionally, developing GET-consistent, country-level ecosystem maps is a priority activity of the Group on Earth Observation (GEO) Global Ecosystem Atlas Initiative (https://globalecosystemsatlas.org/).

The GET resource is an unmapped “comprehensive classification framework for Earth’s ecosystems that integrates their functional and compositional features”; it was designed to support conservation planning and management, research, and other applications (https://global-ecosystems.org/). Because the GET EFGs are the reference classification for UN SEEA national ecosystem accounting and CBD GBF national ecosystem conservation status reporting, if nations wish to participate then they will need to develop EFG national maps by a number of methods—(1) by creating them bottom-up from scratch, (2) by crosswalking from existing national maps, or (3) by deriving them from regional or global EFG maps. In support of extracting national EFG maps from global EFG maps, we developed a global EFGs map as a crosswalk from the existing global 2015 WTEs map.

For data and mapping applications, crosswalking refers to identifying relations between data elements, in this case, ecosystem types, and matching these from one source schema to another target schema (Wu and others, 2023). Crosswalking attempts to make an output from one classification or map compatible with the units from another classification or map.

Crosswalking between ecosystems classifications and maps is rarely straightforward because of differing spatial and classification resolutions. These maps are typically purpose-driven in development and are almost always valid if they are fit-for-purpose and fit-for-scale. However, the more similar classifications or maps are in thematic nature, spatial resolution, currency, and other characteristics, the easier they are to crosswalk (Hearn and others, 2011).

We developed a global terrestrial EFGs map by crosswalking between the IUCN GET and WTE classifications, focusing on the similarities of defining characteristics used to map the WTEs. We associated WTEs with “most probable” EFGs by examining similarities in climate setting, terrain types, and land-cover characteristics. The characteristics of the GET classifications were derived from a set of summary descriptions for each ecosystem (Keith and others, 2020). We also used the biogeographic distribution of WTEs to associate them with generalized locator maps of EFG distributions (Keith and others, 2020). The WTE characteristics were available as attributes in a geographic information system (GIS) data layer (https://data.usgs.gov/datacatalog/data/USGS:6296791ed34ec53d276bb293). In this report, we describe the crosswalking approach in more detail and present the global EFG map produced from the reconciliation of the WTE and EFG resources.

The 2015 World Terrestrial Ecosystems (WTEs) map, developed by a top-down (globally comprehensive), standardized approach, provides a global resource for ecosystem characterization. However, no equivalent global top-down classification of Ecosystem Functional Groups (EFGs) exists. Some countries (for example, Myanmar; Murray and others, 2020) have developed bottom-up (derived from field observations or local knowledge), national EFG classifications and maps. By matching detailed WTEs with policy relevant EFG units we aim to facilitate use of the IUCN GET classification for ecosystem conservation status reporting and ecosystem accounting assessments. A global map of EFGs, derived as a crosswalk from the global map of WTEs, could be a useful interim product while bottom-up national ecosystem maps are developed.

The geographic scope for this new EFGs map is global and represents EFG distributions from 2015 that align with the WTE source data. This work focuses exclusively on terrestrial ecosystems, excluding freshwater aquatic and marine systems, to provide a preliminary global map of terrestrial EFGs aligned with the WTEs.

The crosswalking was accomplished by a single human analyst with appropriate technical knowledge to associate each WTE with the most probable EFG. The input WTE data and the target EFGs are described below, followed by a characterization of the crosswalking process.

Maps of the crosswalking results are shown (1) for the northeastern United States and Canada, (2) for Australia, and (3) for the world in figures 3, 4, and 5, respectively. Assignments of WTEs into EFGs are given in the appendix available online at https://doi.org/10.3133/ofr20251043. A plot showing the number of WTE ecosystems associated with each EFG is in figure 6.

We made a preliminary best attempt to match the WTEs with a corresponding EFG. The crosswalk from this study resulted in predominantly many-to-one relations (fig. 6), where multiple WTEs corresponded with a single EFG. For example, the EFG T1.2 (Tropical-subtropical Dry Forests and Scrubs) aggregated 16 distinct WTEs differentiated by landform type and climatic variation. One-to-many relations were observed for 70 percent of the WTEs. The large percentage of many-to-one relations was expected due to the significantly larger number of WTEs compared to EFGs.

In general, the EFG maps (figs. 3, 4, and 5) appear relatively homogenized compared with the map of WTEs (fig. 1), again to be expected given the magnitude of order reduction in number of resulting EFG classes (41) compared with the original number of unstratified WTEs (421).

The IUCN GET hierarchy includes six levels, and the WTE-to-GET crosswalking exercise was done at a relatively coarse level of the GET hierarchy (the level 3 EFGs), resulting in significant generalization. The classification resolution of the WTEs was likely too fine for the relatively coarser EFGs, and a crosswalking attempt of the WTEs with GET ecosystem units at level 4, level 5, or level 6 might be more appropriate in terms of better-matched classification and spatial resolution. However, the GET typology is not complete beyond level 3. In the future, when GET ecosystem types (a) exist and (b) are mapped for the lower levels of the hierarchy, the WTE-to-GET crosswalk might produce fewer many-to-one relations.

We crosswalked a global set of 1,781 World Terrestrial Ecosystems (WTE) with 41 International Union for the Conservation of Nature Global Ecosystem Typology (GET) Ecosystem Functional Groups. The higher classification resolution of the source data resulted in a considerably generalized product, and many-to-one (WTE-to-Ecosystem Functional Group) relations were common. Although the crosswalking process is potentially valuable for producing standardized, GET-compatible maps for ecosystem accounting and monitoring applications, the resulting homogenization and loss of geographic detail that happens during the process may represent a significant tradeoff. However, the WTEs are likely at a classification and spatial resolution that, for crosswalking purposes, would better match lower levels of the GET hierarchy that have not yet been developed.