What ‘unexplored’ means: mapping regions with digitized natural history records to look for ‘biodiversity blindspots’

South America

Publicly available collections in South America are more numerous in mountainous regions and well-known biodiversity hotspots (i.e., the Andes, Atlantic Forest, and the Guiana Highlands; Fig. 3). This pattern is likely a result of collecting efforts being centered in well-known areas of high species richness that are relatively accessible. Similarly, digitized collections along the Andean mountains show a latitudinal gradient in sampling effort. This pattern is consistent across taxonomic groups and is likely linked to a well-documented latitudinal biodiversity gradient (Fine, 2015). Every group examined appears to have a high number of collections in the Atlantic Forest of Brazil and Uruguay relative to dryer adjacent regions including the savannas of the Cerrado in eastern Brazil. However, estimates of terrestrial vertebrate diversity made using publicly available data have identified hotspots like the Andean mountains as among the regions with the fewest voucher specimens (Šmíd, 2022). Assessing the extent of the scarcity of data collection in highly biodiverse regions remains a challenge. There are relevant records stored in private collections, as well as specimens at local institutions which have not, or have only partially been aggregated into digital repositories (e.g., Museum of Zoology of the University of São Paulo, Cartagena Botanical Garden ‘Guillermo Piñeres’, the National Museum of Brazil, the Museo de Historia Natural UNMSM in Lima), which remain obscured to the international scientific community.

The Amazon Basin, one of the largest and most biodiverse regions of the world (holding 10% of all named plant and vertebrate species; (Nelson et al., 1990; Schulman, Toivonen & Ruokolainen, 2007; Feeley, 2015; Winemiller et al., 2016; Stropp et al., 2020; Albert et al., 2023)), has lower sampling across taxa (Fig. 1). Large sections of the Amazon show very few plant samples, with botanical census data covering only a very small portion of the region (Feeley, 2015). Other lowland areas with fewer publicly available records across taxa include the Orinoco Basin, the Atacama Desert, the southern portion of the Chaco, and the Patagonian grasslands. While fewer digitized collections in the Atacama may be representative of the region’s low diversity for the six taxonomic groups examined here, accessibility and armed conflict may be relevant factors limiting exploration in the other regions. Scarcity of roads or navigable water bodies are a challenge to transportation to and from large sections of those regions, increasing the costs, risks, and logistics for field exploration (Daru et al., 2018; Hijmans et al., 2000). Political and civil unrest in Latin America have historically limited collecting in these and other regions and may have kept many collections there from being part of a global digitized database. This lack of digitization may be reflected in the dearth of collections represented in what are very biodiverse countries. In fact, such biodiversity is related in part to a positive relationship between forest cover and the intensity of armed conflict (Álvarez, 2020; McNeely, 2003; Hanson et al., 2009; Negret et al., 2017).

Specimen data for plants are the most abundant compared to the other taxonomic groups examined here. However, it has been estimated that there are zero publicly available botanical records from about 10% of tropical South America (Feeley, 2015). A higher number of botanical collections may be due to higher plant species richness compared to other groups.

Freshwater fishes are the most speciose group of vertebrates, and their species richness peaks in South America, specifically along the main tributaries of the Amazon River Basin (circa 7,000 species; (Albert & Reis, 2011)). The Amazon Basin and major tributaries appear much better collected (and digitized) than the Orinoco River Basin which covers much of Venezuela and Colombia. In Venezuela, only the Universidad Central has published data in GBIF (for insects), and many are privately held. The ‘La Plata’ region (including the Parana and Paraguay) in east central South America is also poorly sampled and/or digitized relative to the Amazon. Notably, the Deseado River estuary, Lake Musters, and Nuevo Gulf in Argentina are well-sampled for fishes. Compared to the other groups examined, there are comparatively fewer records of fungi. Intensively sampled regions for this group may be a consequence of the presence of institutions where this taxonomic group has been well studied and where collections have been digitized. For instance, southeastern Brazil appears more intensively sampled because the Herbarium at Federal University of Pernambuco is located in this area and has the largest collection of fungi in Brazil (GBIF, 2015). Relative to northern South America, Brazil appears sparsely sampled for specimens of mammals, herps, and birds, especially relative to plants and fishes.

North and Middle America (including landmasses in the Caribbean)

North America is one of the most consistently sampled continents with all taxonomic groups exhibiting similar decreases in sampling effort along a south-to-north latitudinal gradient and few areas devoid of specimen samples of any group south of Canada (Fig. 4). Excluding Greenland, interior Canada is the largest gap for all plant and animal groups examined, from Ontario to the Northwest Territories and east to Newfoundland and Labrador. Canada may show relatively few natural history samples in part because of its enormous size and climate (creating areas of remoteness); the lack of knowledge about its abundant freshwater resources has been previously noted (Desforges et al., 2022). Relative to adjacent countries it appears Nicaragua is poorly sampled for all vertebrate groups and fungi despite having an abundance of biodiversity-rich regions including the Mosquito Coast, Lake Nicaragua, and the largest tropical rainforest north of the Amazon (Weaver, Lombardo & Martinez-Sanchez, 2003). Otherwise, most of Central America and the Caribbean/Greater Antilles appear well sampled with a clear gap for Cuba, due more so to conflicting politics and policies (at least with the United States) than to a lack of biodiversity (Denis, Cruz-Flores & Testé, 2018). Conservation work in Cuba is among the best in the region (Goulart et al., 2018) and the Museo Nacional de Historia Natural de Cuba in Havana has digitized collections, but these are not publicly available. Perhaps not surprisingly the United States is very well sampled, and climatic differences between the wetter, more humid east vs. the more arid west can be seen when comparing fungi and fishes from those regions; with the more temperate Pacific Northwest being an exception.

Mammals have weaker collection records from the southeast and midwest United States than other vertebrate groups. Specimens of amphibians and reptiles (herpetofauna) display the sharpest decline moving northward of all examined groups, with few specimens collected in and east of the Rocky Mountains from Montana to Illinois (most likely do a drop in taxonomic diversity because of a climatic shift). Fungi and fishes are, unsurprisingly, sparsely collected from the northern Rocky Mountains south to central Mexico including adjacent desert regions. Although most vertebrate groups display strong holdings from the southeastern United States, non-coastal southern Appalachia is relatively poorly represented in bird collections, as are Montana, Wyoming, and Utah in the west. Plants are the most consistently sampled group across North America with few, if any, regions that could be called a blindspot, while fungi have the highest number of areas lacking sampling according to GBIF, especially in and around the Chihuahuan Desert.

Africa (including Madagascar and adjacent landmasses)

Perhaps more than any other region, Africa’s ‘biodiversity blindspot’ areas likely reflect the sparsity of digitization of regional collections as well as the general lack of collections (Fig. 5). A wide variation in climate such as the dry Sahara region vs. sub-Saharan tropical regions, may be important explanations for variation in collecting effort within the continent, as these climatic factors are linked to both biodiversity and accessibility. For example, the remoteness of many places within the large Sahara Desert (a significant global ‘biodiversity blindspot’ across taxa; Fig. 2) create accessibility issues combined with low species richness for many taxonomic groups. Collections for most groups track with humidity and rainfall, save for fungal diversity which is likely higher than represented by the few collections that have been made across the continent, including Madagascar. Notably, Cape Verde is a bright spot for fungal collections on the West Coast of Africa almost entirely accounted for by collections from lichenologists at the Senckenberg Research Institute and Natural History Museum in Frankfurt, Germany, demonstrating the outsized impact a single research program can have on perceptions of whether an area has been studied (Büdel & Mies, 1993). A striking juxtaposition in known collections is seen between the relatively well-sampled island of Madagascar and adjacent islands such as the Seychelles. This dichotomy is particularly clear for plants and is likely a result of sampling efforts from the Missouri Botanical Garden which has a long history of botanical research on Madagascar: accounting for an estimated 72% of vascular plant collections from Madagascar. Unfortunately, this means the largest collection of Malagasy plants exists well outside of the island which would hinder the growth of local knowledge on these organisms (Phillipson et al., 2006). For mammals, the Great Rift Valley region is particularly well sampled, as are the southern reaches of West Africa (from Liberia to Benin). Madagascar has relatively few mammal specimens in collections despite its high proportion of endemic mammal species; this is likely due to the conservation status of many of the species, including the endangered endemic lemurs.

Despite being largely tropical, Somalia, the Central African Republic, and large areas in the Democratic Republic of the Congo, Angola and Botswana are significant ‘biodiversity blindspots’ across taxonomic groups (Fig. 2). The sparsity of collections in these regions can be attributed to a combination of complex factors including a history of colonization, lack of scientific infrastructure and institutional support, geographic accessibility, and historical and present-day political instability (particularly in central Africa where plant collections have been noted to decline during periods of war (Sosef et al., 2017)). While not exclusive to these poorly collected regions, cultural factors are also an important consideration (e.g., in some areas there may be cultural beliefs that discourage the collection of natural history specimens (Stropp et al., 2020; Sosef et al., 2017)). One additional factor is that the Royal Museum for Central Africa (Belgium) reports 10 million biological specimens in its collections from the Congo and adjacent regions, but only a small percentage (<0.5 million) of these are on GBIF (E. Greenbaum, 2024, personal communication).

We call for collections in Africa to join the GBIF network but also call on institutions outside of Africa with African collections to publish their records (some Western museums may have collections from the region that are not yet digitized–as noted above). Other countries, such as the Central African Republic, have no digitized collections, and all the digitized voucher records from that country are housed in other countries (Table 1). Egyptian collections from Cairo, Alexandria, and along the Red Sea for several groups are notable and likely were a target for several important historical collections. Interior reaches of Africa, outside of the desert regions, including the Okavango Delta still have relatively few collections digitized (Tolley et al., 2016; Greenbaum, 2017). Regional museums such as the Port Elizabeth Museum (South Africa) that have known collections from the Okavango and surrounding regions are not currently linked to GBIF. Given that region’s celebrated conservation status, efforts to collect in this area should be well-regulated. Notably many collections from this region are in Europe and the United States, and efforts to digitize existing collections and create new regional collections should be considered as they would benefit local knowledge.

Australia and Adjacent Islands

Unsurprisingly, the large swath of arid land through central Australia is less densely sampled than tropical northern regions and the temperate south, due to the relative dearth of species biodiversity in this arid climate, particularly for large vertebrates (Fig. 6; Dickman, 2018). This portion also remains logistically difficult to access for collecting, due to its desert climate and relative lack of travel infrastructure (i.e., roads). Tasmania is a significant global hotspot for specimens across taxonomic groups, as is the region near Perth in the southwest, due in part to their more tropical climates, easier accessibility, and the presence of natural history museums in each locale (Dickman, 2018; Fig. 2). The lack of mammal and herp collections from New Zealand is not surprising because of the relatively few species of these groups known to exist there.

Eurasia

Much of Europe has been well collected and the number of digitized specimens from the region is among the highest on the planet; in stark contrast to the rest of Eurasia except the far East (particularly Japan and Taiwan which have substantial collections and active natural history researchers; Fig. 7). The relative absence of digitized collections from India for plants, fungi and mammals is unexpected, given the region’s rich biodiversity. Sri Lanka appears well-sampled relative to India for several groups which may reflect a greater accessibility to collections and digitization of specimens from that island nation. South Korea, Nepal, and Taiwan appear to have a high number of collections of birds, mammals, plants, and herps relative to much of ‘mainland’ China, although the warm temperate areas of Southern China appear well collected for some groups such as plants and herps making the absence of mammals and birds from this region even more striking. Much of the cold remote vastness of Russia lacks GBIF representation for natural history specimens. Yet bright spots on the Russian map shed light on the importance of local museums in the pursuit of cataloging biological diversity. For example, the high concentration of fungi samples near the western central part of the country is not a centroid artifact but the Yugra State University Biological Collection: an institution that has been collecting mycological specimens since the early 20th century (Filippova et al., 2020). Lack of digitization, collecting efforts, climate, and politics probably have a great deal to do with the lack of vouchered samples from the world’s largest country.

For some taxa, a sharp contrast exists in New Guinea between New Guinea (the western Indonesian side often called ‘Irian Jaya’) and the Eastern side of the island, Papua New Guinea, which has been a target for many natural history expeditions (Mayr & Gilliard; Webb, 1995; Cookson, 2000). Another noticeable contrast occurs between north and south of the New Guinea Highlands, with the south being poorly collected versus its northern counterpart. This is the case for both the Indonesian and Papuan halves of New Guinea, most likely due to the heavy monsoon flooding that seasonally submerges areas in that portion of the island (Tanaka, 1994).

Despite its tropical climate and abundance of biodiversity, there appear to be few collections from Indonesia, particularly for fungi. Contrastingly, the Philippines are densely collected relative to Indonesia, again showing how Indonesia appears poorly collected for natural history specimens. It is also possible that Indonesia only appears as a ‘biodiversity blindspot’ because of a dearth of accessible digitized specimens (likely in collections in the West). Indonesia has been a target for museum-based collections since occupation by the Dutch and Alfred Russel Wallace’s early expeditions to the Indo-Australian Archipelago. Few if any of the specimens from southeast Asia collected by Wallace or used as type material by Western zoologists are curated by local institutions. Specimens collected from this region more recently are curated in large institutions like the Museum Zoologicum Bogoriense in Java, Indonesia, which has yet to have its extensive collections digitized. The absence of a fungarium (despite there being herbaria) is an oversight and can be linked to why there are few fungal collections in an area that should be a hotspot for mycologists. Borneo appears remarkably underrepresented in natural history collections for herps and mammals, particularly relative to nearby Java.

Mammals and herps are particularly well collected in the Philippines and Papua New Guinea making adjacent areas even more glaring for their lack of representative samples on GBIF. Birds and plants appear to be the best sampled taxa from Indonesia, perhaps as a result of historic collecting efforts but gaps remain in New Guinea, Borneo, Sulawesi, and Sumatra.

Notably, no museum collection from Oceania (the scattered islands of the Pacific) is listed among the 73 of the world’s largest natural history museums and herbaria from 28 countries, with the nearest of these surveyed collections being in India and Australia (Johnson, Owens & Global Collection Group, 2023). Representation of people from this region and digitization of local collections should be prioritized to further our understanding of the biodiversity of the region.

Antarctica

Antarctica is notable only in the absence of many collections for obvious reasons. There are no collections of freshwater fishes, amphibians or reptiles because none survive on the continent. Large mammal and bird collections do exist but are generally restricted to historic collections or from areas near the coasts including some not yet recorded on GBIF. Similarly, fungi and plants are generally restricted to ice-free areas of the continent (Fig. 8). Observations of Antarctic fauna have been made but many of these are of marine invertebrates and fishes below the sea ice (Chakrabarty et al., 2021).