Ecoregional distributions of the world’s freshwater vertebrate species

The foundational freshwater vertebrate database introduced here, FreshVerts v1.0, has been assembled by refining the checklist of freshwater vertebrate species, compiling information about their geographical distribution, assigning them to freshwater ecoregions, and validating records and taxonomy. The data sources used are cited throughout the text and in the database documentation. An overview of the methods used can be found in Fig. 1.

Flow diagram of the main methodological steps involved in compiling, harmonizing, and reviewing data for the creation of FreshVerts v1.0. See main text for more details. Data sources are described in the main text and database documentation.

We retained the same 426 ecoregional spatial units developed for the FEOW database so that we could focus on enriching the taxonomic and temporal coverage of species distributions (https://www.feow.org/download). Ecoregions have a mean area of 301,901 km² (SD 483,603 km²), and some are sub-basins of larger basins, some are entire watersheds, and some are amalgamations of multiple small watersheds that were believed to have similar freshwater fish faunas. We assigned freshwater ecoregions to major WWF biogeographic realms to facilitate representation analyses⁴. We updated the distributional data for tetrapods (mammals, birds, reptiles, and amphibians) by drawing on the public IUCN Red List database, and focused our efforts on resolving which species should be considered to be freshwater-dependent. Updating the fish species distributions contained in the FEOW dataset required both expanding the species list and drawing upon information from many sources, including obscure primary literature, owing to the comparatively limited information available for fish compared to other vertebrate groups²⁸.

Freshwater tetrapods

A list of described freshwater tetrapod species was obtained from the IUCN Red List database in July 2022²⁴, and then classified into four major taxonomic groups: mammals, birds, reptiles, and amphibians. The IUCN Red List contains the most comprehensive data for tetrapods and represents the current understanding of both taxonomy and distribution (https://www.iucnredlist.org/resources/spatial-data-download). We included only species listed in the presence categories of “extant” and “probably extant” and in the origin categories of “native” and “reintroduced”. For many tetrapods, the geographic range includes areas of migratory passage or seasonal presence, which we considered to be essential portions of their distribution in terms of life cycle for at least some populations. For a description of the IUCN Red List categories of origin, presence, and seasonality, see the metadata document Digital Distribution Maps on The IUCN Red List of Threatened Species²⁹.

A key aspect of the study was the definition of “freshwater” for tetrapods. In this case, we chose to follow the definitions given by Vences & Köhler³⁰ to include as freshwater species only those tetrapods for which freshwater environments are required to complete their life history, including life cycle and a close dependence for habitat or food. This definition excludes species that can occasionally be found in freshwaters but do not heavily rely on them for completing their life cycle or obtaining food. We revised the IUCN Red List freshwater attribution of each tetrapod class.

Mammals

From those species classified as freshwater by the IUCN Red List, we discarded those being primarily marine species or that are only found occasionally in freshwaters (e.g., Marine otter Lontra felina, or Harbor Seal Phoca vitulina). In total, we compiled 1,604 records of 127 freshwater-dependent mammal species, 50 more than in the previous study from the Freshwater Animal Diversity Assessment³¹. However, this study does not provide a clear definition of what criteria were used to classify mammal species as aquatic³¹. The number of freshwater-dependent mammals in our checklist (127) is almost identical to that provided in Veron et al.³¹ for the combination of “aquatic” plus “freshwater-dependent” species (124), so our definition broadly comprehends both categories.

Birds

Discrepancies between the bird species classified as freshwater-dependent under the IUCN Red List (following BirdLife International assessments) versus the classifications of other conservation organizations, such as The Nature Conservancy (TNC), were concerning. We decided to follow the more conservative approach of TNC’s classification for freshwater birds, given the fact that IUCN Red List classifies entire families as freshwater obligate species (i.e. kingfishers – Alcedinidae), rather than investigating whether each individual species (across all bird species) is dependent on freshwater habitats for breeding (i.e. many ducks, herons) or feeding (i.e. birds are entirely dependent on freshwater habitats, such as freshwater fish, mollusks, and crustaceans). By doing so, for instance, only one-third of kingfisher species are freshwater obligates³². This conservative approach (compared to BirdLife’s) is also shared with previous freshwater bird checklists³³, although our checklist contains 200 more species than Dehorter & Guillemain³³. This difference is due to a larger number of bird species described in 2024 (1,200 more species compared to 2008, the date of³³), while can also indicate a less conservative approach in our checklist. Our list includes 26,753 ecoregional attributions of 764 freshwater-dependent bird species.

Reptiles

Similar discrepancies appeared in the IUCN Red List classification of freshwater reptiles. We decided to exclude species of reptiles that the IUCN Red List classified as found in freshwater, but which are primarily terrestrial, fossorial, or arboreal, or habitat generalists, with only facultative use of freshwater habitats, or where the species was known from very few specimens, so confidence in habitat use is low. Moreover, based on the authors’ expertise, we included those species that are freshwater-dependent according to our prior definition but were not classified as so in the IUCN Red List. The final list includes 4,298 ecoregional records of 639 freshwater-dependent reptile species. These results agree with the criteria used for freshwater-dependence with previous studies from the Freshwater Animal Diversity Assessment, which accounts for 250 freshwater-dependent turtles³⁴ (253 in our checklist) and 24 crocodiles³⁵ (23 in our checklist). However, our freshwater-dependent squamata (lizards and snakes) list is larger than that of^36,37 (365 vs 226, respectively), as a result of the increased number of species among squamata compared to 2008 and a less conservative approach in our selection of freshwater-dependent squamata species. In the case of reptiles, the spatial information was obtained from the Global Assessment of Reptile Distributions (GARD)³⁸ (http://www.gardinitiative.org/data.html), a more comprehensive data source than the IUCN Red List, since many species identified as freshwater dependent were not evaluated in the IUCN Red List.

Amphibians

In the case of amphibians, we followed the classification of freshwater-dependent species elaborated by Vences & Kölher³⁰, only including the genera they classified as dependent on freshwater ecosystems, since it agrees with our previous definition of freshwater-dependent. For genera not included in that study (because of taxonomic reclassification or being newly described genera), we determined their agreement to the freshwater-dependence definition by looking at their life history and taxonomic relationships with existing genera (e.g. new genera that emerge from the taxonomic division of the genera Rana and Hyla were considered to be freshwater-dependent since these are so). In total, we included 15,885 records of 5,089 freshwater-dependent amphibians.

Finally, we made a spatial intersection between the 426 freshwater ecoregions and the spatial ranges of freshwater tetrapod species to identify the ecoregions occupied by each species. Spatial calculations were used with the package sf³⁹ from R⁴⁰. Since freshwater tetrapod distribution maps do not follow HydroSHEDS contours, we tested the false attribution of tetrapod species to ecoregions caused by spatial disagreement by discarding those intersections between the species distribution maps and the ecoregions that both (a) represented less than 10% of a species’ range and (b) represented less than 10% of an ecoregion’s area. Please see the Technical Validation section for a detailed explanation of this threshold calculation and its impact on the resulting data. In total, we obtained 48,540 records corresponding to 6,618 freshwater tetrapods.

Freshwater fishes

To update the list of freshwater fishes, we relied upon data from FishBase (https://fishbase.org), the most complete database that is openly accessible²⁵. FishBase has a habitat classification, and we extracted only species with the attribute “Freshwater”.

Comparing the original FEOW fish list (13,802 species) against the 2022 FishBase list indicated 4,027 additional freshwater fishes in need of inclusion.

To obtain spatial information for every species, we performed a series of steps utilizing multiple databases. First, we extracted distribution maps for as many species as possible from the IUCN Red List database²⁴. Similarly to tetrapods, we only included the spatial information of species in the presence categories “extant”, “probably extant”, and “reintroduced”. We found spatial distributions for an additional 1,386 freshwater fish species in the IUCN Red List.

Since not all freshwater fish species were present in the IUCN Red List database, we searched for information on the remaining ones in the Global Biodiversity Infrastructure Facility (GBIF, https://www.gbif.org/). Using GBIF API, we obtained species records as point coordinates⁴¹ for 1,162 species. For the remaining species, we accessed FishBase records through consultations with FishBase website managers (pers. comm). They were able to provide us with spatial information for 520 freshwater fish species as point coordinates.

After these subsequent queries, there was a final subset of 958 freshwater fish species with no spatial information available in standard repositories, so we sought to fill these gaps based on scientific literature. When possible, we assigned each species’ spatial information as points (e.g., species descriptions with coordinates or specific locations, usually belonging to the holotype). Other species were directly assigned to the ecoregions corresponding with their distribution range based on verbal descriptions, since the information retrieved could not be translated into point coordinates. After performing the bibliographic search, we were able to retrieve reliable information for 848 additional freshwater fish species. The remaining species (see list in Literature search.xlsx) were described using geographic information so vague that we could not reliably determine their distribution, taxonomic validity, or both. These 110 species are not included in the v1.0 database.

Each of the freshwater fish species not included in FEOW’s original checklist was assigned to one or more freshwater ecoregions following the same procedure as for tetrapods. Finally, all species taxonomic information was updated to the 2024 nomenclature. This involved discarding 1,2,273 freshwater fish records corresponding to 1,206 fish species owing to taxonomic changes and redundancies (see Technical Validation section for details). In total, 47,589 ecoregional records of 16,512 freshwater fish species are included in FreshVerts v1.0.

Our final database comprises 96,129 ecoregional population records representing 23,130 freshwater vertebrate species (see Technical Validation section for details).