Using historical and palaeoecological data to inform ambitious species recovery targets

Abstract

Historical data are a valuable resource for addressing present-day conservation issues, for example by informing the establishment of appropriate recovery targets. However, while the recovery of threatened species is the end goal of many conservation programmes, data made available through the efforts of palaeoecologists and historical ecologists are rarely consulted. The proposal of a ‘Green List of Species’ by the International Union for Conservation of Nature (IUCN) will soon change this. The Green List of Species measures recovery against historical baselines; in particular, the method requires estimates of species range and abundance in previous centuries. In this paper, we present the case for why setting species recovery against a historical baseline is necessary to produce ambitious conservation targets, and we highlight examples from palaeoecology and historical ecology where fossil and archival data have been used to establish historical species baselines. Finally, we introduce Conservation Archive (https://conservationarchive.shinyapps.io/ConservationArchive/), a database of resources that can be used to infer baseline species conditions, and invite contributions to this database.

This article is part of a discussion meeting issue ‘The past is a foreign country: how much can the fossil record actually inform conservation?’

1. Introduction

Human impacts on the world have ushered in a new ‘age of extinction’ [1], with the rate of disappearing species reaching levels previously only seen during the five mass extinctions [2–4]. Therefore, it is no surprise that advocates for threatened species conservation have called for quick and decisive action to stop species declines and prevent extinctions, and indeed such aspirations are enshrined in internationally-agreed targets under the Convention of Biological Diversity and the Sustainable Development Goals.

Currently, the most-widely used tool to help guide priorities for species conservation investment is The IUCN red list of threatened species™ (hereafter IUCN Red List or Red List) [5,6], founded in 1965, which classifies species according to their extinction risk. Identifying the species at the highest risk of extinction highlights those in the most dire circumstances, and helps inform where and how conservation actions should be prioritized to prevent permanent losses of biodiversity. But once conservation efforts have halted the decline, and a species has stabilized at a lower risk of extinction, what happens next?

There is a growing awareness in conservation that preventing extinction, while vital, is only one necessary step [7–10]. The maintenance of biological diversity depends on the maintenance of functioning biological communities and ecosystems. Species pulled back from the brink, while still surviving on the planet, are often absent from large parts of their former range, and/or at greatly reduced densities and so are no longer playing their ecological roles. Therefore, a long-term goal of conservation should be to restore species so they are fulfilling ecological roles in systems across their historical range, and to devise plans for more complete species recoveries. However, while extinction is a clearly defined state, there is no consensus among conservationists on the definition of the ‘recovered’ state.

Despite the lack of a clear definition of recovery, it is nonetheless a stated goal of many threatened species conservation programmes. Currently, achieving species recovery generally means a relatively short-term increase in some measure of abundance (e.g. number of individuals, range size). However, there is a growing realization that these increases must be placed in historical context to fully conceptualize recovery [11,12]. Conservation biology as a discipline only emerged in the 1980s; many policies formalizing the protection of species are also relatively recent (e.g. U.S. Endangered Species Act, 1973; CITES, 1973). By the time these protections were put in place, human exploitation of the natural world had long been occurring, meaning that in the twenty-first century we have many shifted baselines for species presence and abundance [13].

Marine turtle recovery provides an excellent example of this phenomenon: since the 1980s, numbers of green turtles, Chelonia mydas, and hawksbill turtles, Eretmochelys imbricata, have been on the rise owing to concerted conservation efforts, with an estimated increase of about 300% in hawksbill numbers over that time period [11,14]. However, these gains have restored only a minuscule fraction of the species' estimated abundance from the late nineteenth century, perhaps as small as 7% (green turtles) or 1% (hawksbill turtles) [11]. The puma Puma concolor provides another example: while it is currently assessed as Least Concern for extinction risk on the Red List, over the past few centuries hunting has extirpated this apex predator from at least a third of its previous range, including the entire eastern half of the United States save a tiny remnant population in southern Florida [15].

In order to create ambitious conservation targets that avoid the trap of shifting baselines, evaluation of historical conditions is necessary. This has recently been advanced with an IUCN Resolution (https://portals.iucn.org/library/node/44008) to develop a Green List of Species. This grew out of the recognition that while the field of conservation currently has a standardized way to assess what we want to avoid—extinction—there is no such method to assess what we want to achieve—recovery. The Green List of Species aims to measure species progress towards recovery. This is posited as being achieved when the species is ‘viable and ecologically functional in every part of its indigenous and projected [future] range’ [10]. Identifying a species' indigenous range is a process that clearly must be informed by historical evidence. Such evidence can also help in assessing whether or not a species is ‘ecologically functional’.

In this paper, we:

• (1)review how historical and palaeoecological data have been used to delineate species' historical ranges,
• (2)discuss the challenges of defining ‘indigenous range’ in the practical context of the IUCN Green List of Species,
• (3)highlight examples of how archival abundance records can provide key evidence of species functional densities, and
• (4)propose a database that can be used to help establish historical baselines of distribution and abundance for the purposes of the Green List of Species, and also to promote greater transmission of information from historical ecology and palaeoecology to modern-day conservation efforts.

2. The mismatch between current metrics of conservation success and a historical perspective

Current conservation practice does not encourage the taking of an historical approach [16]. A potential criticism of the proposed Green List of Species, or indeed of any effort to integrate historical baselines into conservation practice, is that finding reliable information on species' distributions hundreds of years ago will be a major limiting factor. However, this limitation may be less acute than many conservationists realize, owing to their general lack of familiarity with archival data sources. This might be because evaluating progress against the current most prevalent metric of conservation success, extinction risk, rarely requires historical thinking. Sometimes extinct range (since 1500 AD) is carefully documented as part of the Red List assessment process, although this is neither a formal requirement for assessment nor consistently done across species; of the 68 007 current assessments [17] that include mapping data, only 1112 (1.6%) document extinct range. The Red List does require assessors wishing to list species under Criterion A (population decline) to evaluate species population trends over the past 10 years or three generations, whichever is the longer [18]. Hence, for some long-lived species, the assessment of extinction risk could already span hundreds of years or more; for example, for redwoods Sequoia sempervirens, three generations may represent thousands of years [19]. Many plant species that are listed under Criterion B, reflecting small geographical range, are assessed based on herbarium specimens that might be several centuries old owing to a lack of contemporary data [20].

However, for the majority of species, assessment of extinction risk does not require the collation of retrospective data; at most it requires going back one or two decades. Therefore, concern about the amount of work involved in collating such data for a Green List of Species assessment is understandable. However, these data are often more readily available than ‘neoecologists’ (those who study present biological communities; [21]) might think, thanks to the efforts of palaeoecologists (those who study biological communities of the past; [22]) and historical ecologists (those who study the ‘interconnectedness of humans and nature’, drawing from the fields of history, ecology, geography and anthropology; [16,23]).

3. Using historical and palaeoecological data to reconstruct historical ranges

Conservation scientists might not often consider the fossil record when determining contemporary species ranges, but the concept of ‘applied palaeoecology’ [24] is not new. Animal fossils and fossil pollen from the Holocene Epoch (the current geological epoch, which began approximately 11 700 years ago) can provide a starting point when considering the distribution of species prior to major anthropogenic disturbance. The use of fossil evidence to establish baselines against which to measure biodiversity loss is perhaps ‘the most important contribution that palaeo-ecology can make to neo-ecology’ [25, p. 59]. Fossil pollen, as well as plant fragments found in fossilized packrat middens, have been used to reconstruct Late Quaternary plant species distribution for large areas of the world (discussed in [26–28]). Turvey et al. [29] used fossils and archaeological evidence dated from the early Holocene until the Ming Dynasty (a period spanning approximately 11 700 BP–1644 AD) to infer historical ranges of 34 mammal species in China. The fossil record can also be combined with methods from historical ecology to infer historical distributions, for example using official records and documents from hundreds of years ago (e.g. [30]), or with genetic analysis and simulation modelling to infer the contributions of intrinsic and extrinsic factors to species or population extinctions [31]. Manlius [32] used a variety of sources in addition to fossil remains, including prehistoric cave drawings and visual motifs of Pharaonic tomb monuments, to reconstruct the distribution of the common hippopotamus, Hippopotamus amphibius, in Egypt, a country where the species has been extinct since the early 1800s.

For commercially or culturally important species, records of trade can be invaluable for estimating historical extent of occurrence and relative abundances. Owing to the longstanding cultural significance of plants in the genus Agave to the peoples of the Americas, academic interest has motivated scholars to visit accounts from as far back as the 1500s in order to determine which species, using current taxonomy, were referenced (e.g. [33]). The logbooks of commercial fisheries or whaling operations can be used to inform the calculation of geographically-specific changes in abundance over hundreds of years (e.g. Atlantic cod, [34]), and to parametrize models to infer historical range (e.g. North Atlantic right whale Eubalaena glacialis, [35]). Other perhaps-unexpected sources of historical data, including menus, can be analysed to infer range and aid conservation efforts [36]. More generally, historical information related to human impacts, including maps of human density and land use, and records of agricultural expansion and invasive species occurrences, would provide invaluable data for hindcasting species distributions.

Archaeological remains can also provide evidence for past occurrence in areas where a species has never been recorded in writing. Bone fragments found in archaeological excavations of Roman sites recently provided the first evidence that North Atlantic right whales and grey whales, Eschrichtius robustus, were once present in the Strait of Gibraltar and possibly in the Mediterranean, which had never been considered a part of their historical range [37]. Archaeological remains in northern Borneo also showed that Bornean orangutans, Pongo pygmaeus, were once much more widespread than they are today [38].

Of course, historical accounts of species tend to be biased towards culturally important, useful or high-profile species [39] and against species found in remote and/or inaccessible areas [40]. In some cases, human interest in a species makes it more difficult to determine indigenous range, for example plant species that have been cultivated and traded so extensively that the native range is no longer obvious [41]. Other species will not be found in the fossil record owing to taphonomic bias (i.e. the factors that make some remains more likely to persist in the fossil record than others; discussed in [25]). In such cases, it is still possible to infer indigenous range via hindcasting using climate envelope or habitat suitability models [42]. Such models can use historical climate, landcover and land use data, compared against current species occurrences, to create a potential species distribution at a temporal benchmark. For example, Fouquet et al. [43] used ecological niche modelling to hindcast the expected distribution of Hochstetter's frog, Leiopelma hochstetteri, in New Zealand at 1300 AD (pre-human arrival). Another technique that could potentially be applied to estimate indigenous range is the estimation of ‘potential current natural range’ in the absence of human influence [44,45]. Finally, different sources of information can be combined to make inferences about the impact that humans have had on species distributions. Gibson et al. [46] used models informed by the fossil record and environmental variables to show that human presence has had a much greater impact on the distribution of the Hispaniolan solenodon, Solenodon paradoxus, than climatic changes since the last glacial maximum.

4. At what point in time should the indigenous range be evaluated?

In 2018, the IUCN Species Survival Commission's Species Conservation Success Task Force proposed a Green List of Species framework [10]. The authors recommended that a fixed temporal benchmark should be chosen for the delineation of indigenous range, in order to make assessments comparable. They suggested that the fixed benchmark should be as ‘early as feasible’ [10] in order to avoid the shifting baseline trap, but recent enough that sufficient evidence will be available for the majority of species (either direct observations or evidence to facilitate inference).

Akçakaya et al. [10] proposed two potential temporal benchmarks: the year 1500 AD and the year 1750 AD. Species that went extinct before 1500 AD are not assessed for the IUCN Red List, so choosing 1500 AD is a natural link between the two assessment frameworks. Regarding anthropogenic impact, 1500 AD is also the approximate start date of widespread global movements by Europeans [47], which had a significant impact on global diversity. Restoration efforts in the Galápagos Islands use a temporally similar date, 1535 (the year the islands were first visited by Europeans), as a benchmark for restoration of species' distribution and abundance [48]. The second proposed date, 1750 AD, reflects the start of the global industrial revolution and has been used to define the end of the pre-industrial period by the Intergovernmental Panel on Climate Change (IPCC) since the Fourth Assessment Report [49]. One advantage of this date is that records of species occurrence are likely to be more readily available than for 1500 AD.

The concept of a temporally fixed benchmark at which to evaluate indigenous range has come under criticism. Sanderson [50] argues not only that the proposed benchmarks of 1500 AD and 1750 AD take a Eurocentric perspective, but that the timing of human impacts vary so widely across the globe that choosing one date does not make sense. Sanderson [50, p. 1209] suggests that rather than at a standardized baseline, the Green List of Species method should stipulate that the indigenous range be evaluated at ‘a time before human beings were the most important element limiting species' distributions’, which would vary from species to species. If such an approach were taken, this would increase the scope for fossil and subfossil records from the Pleistocene and Holocene to inform the delineation of indigenous range. Indeed, there is evidence that human impacts on mammal ranges within Europe were operating even in prehistoric times [51].

The discussion of benchmark dates for the Green List of Species will surely continue in coming years. Nonetheless, it is important to remember that choosing a benchmark for a Green List of Species assessment does not imply that the baseline conditions at this chosen point represent the ‘desired’ or ‘pristine’ state; rather, the benchmark date is provided as a practical compromise between the desire to avoid shifting baselines, the need for data, and the ultimate goal of inspiring assessors to think more ambitiously about species recovery in a standardized way.

5. Using historical abundance data to inform the definition of ‘ecologically functional’

There has long been a desire to ensure the conservation of ‘ecologically functional populations’ of species, carrying out their roles in the ecosystem, rather than just the minimum amount of individuals required for species persistence (e.g. [7–9,52]). The Green List of Species framework considers ecologically functional populations to be a key indicator of species recovery [10]. However, it is difficult to determine exactly how many individuals, of what age classes and density, are required in order for a population to be carrying out the species' function in the environment [53]. As a workaround, the Green List of Species framework suggests that one solution for assessing ecological function and creating a recovery target beyond mere viability is to use population levels assessed prior to human impact as a proxy [10]. Fossil and historical evidence will be invaluable here.

Indeed, archival data are already being used to set demographic targets for conservation planning. Oysters Crassostrea virginica in the U.S. Mid-Atlantic have been in decline for centuries [54] and, given their role in regulating water quality, shoreline protection, and providing habitat for other marine organisms [55–57], the potential loss of ecological function is non-trivial. Fossil data from the Pleistocene, samples taken during the American colonial period, and modern records were compared to reveal that oyster abundance has declined dramatically with harvesting effort, and that harvesting is also correlated with changes in population structure, with loss of the oldest and largest classes of individuals [58]. Restoring Pleistocene age and size structure is therefore recommended as a management goal in the U.S. Mid-Atlantic region [58].

Pre-impact carrying capacity is also suggested as a potential proxy for determining if a population is ecologically functional [10,53]. This can also be informed using historical data; as an example, historical catch records and habitat preferences were used in a modelling framework to estimate the pre-whaling carrying capacity for right whales in the North Atlantic [12].

Finally, palaeo and historical records can demonstrate changes in species interactions and behaviour after the introduction of human exploitation. Jackson et al. [55] used palaeoecological sediment records, archaeological records, historical trade documents and ecological data from the past century to identify changes in trophic interactions between functional groups in several marine habitats post-introduction of fishing. Human presence can also alter habitat selection behaviour, making it more difficult (if not impossible) to restore certain species to functionality in a given area. For example, records dating from the fall of the Roman Empire indicate that prior to the start of human persecution, Mediterranean monk seals, Monachus monachus, were known to form large colonies on beaches; today, this is rarely observed and the species is largely confined to caves [59].

6. A database to bring past insights into modern conservation

It is clear that there is a treasure trove of data about past environmental conditions and species distributions that neoecologists have thus far had little reason to delve into. With the developing IUCN Green List of Species, conservation scientists may soon find themselves starting to look for and use fossil data, historical records and long-term climatic/landcover datasets to determine the indigenous range of a species. In many cases, assessors will be looking for proxies, so consultation of multiple data sources to increase precision is recommended [55]. To accelerate and encourage this effort, we have created Conservation Archive (https://conservationarchive.shinyapps.io/ConservationArchive/), a searchable list of such resources, arranged by taxon, time period covered and data type (formatting of sample entries visible in table 1). Conservation Archive was built using the R package ‘shiny’ [60]. This database, like the Green List of Species concept, is still under development and barely begins to scratch the surface of the datasets available. Therefore, we invite suggestions of relevant open datasets that the authors, as neoecologists, are currently unaware of, as well as the contact details of those who manage private datasets but would be willing to share data on a case-by-case basis. Those interested in contributing to the database may do so directly online, or may contact the corresponding author with suggestions.

Table 1.

A sample of datasets currently listed in Conservation Archive. Readers can submit new datasets at https://conservationarchive.shinyapps.io/ConservationArchive/.

resource name	data type	earliest data	latest data	taxa	geographical range	summarya
GBIF, the Global Biodiversity Information Facility (https://www.gbif.org/)	species occurrence	1600–1700 CE	2000–2100 CE	multiple taxa	global	‘GBIF—the Global Biodiversity Information Facility—is 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.’
The Paleobiology Database (PBDB) (https://paleobiodb.org/)	species occurrence	Palaeozoic	2000–2100 CE	multiple taxa	global	‘The Paleobiology Database (PBDB) is a non-governmental, non-profit public resource for palaeontological data … Its purpose is to provide global, collection-based occurrence and taxonomic data for organisms of all geological ages.’
Neotoma Paleoecology Database and Community (https://www.neotomadb.org/)	species occurrence	Cenozoic: Pre-Quaternary	2000–2100 CE	multiple taxa	North America	‘Neotoma Paleoecology Database and Community is an online hub for data, research, education, and discussion about paleoenvironments. Anyone with an Internet connection can access Neotoma.’
PHYLACINE, the Phylogenetic Atlas of Mammal Macroecology (https://megapast2future.github.io/PHYLACINE_1.2/)	species occurrence	Quaternary: Pleistocene	2000–2100 CE	mammals	global	‘…contains phylogenies, range maps, trait data, and threat status for all 5831 known mammal species that lived since the last interglacial.’
IUCN Red List of Threatened Species (https://www.iucnredlist.org/)	species occurrence	1500–1600 CE	2000–2100 CE	multiple taxa	global	‘…the world's most comprehensive information source on the global conservation status of animal, fungi and plant species.’
Anthromes v2: The Global Ecological Patterns Created by Humans (http://ecotope.org/anthromes/v2/maps/)	landcover	1700–1800 CE	2000–2100 CE	n.a.	global	‘Aim: To map and characterize anthropogenic transformation of the terrestrial biosphere before and during the Industrial Revolution, from 1700 to 2000.’
NOAA Paleoclimatology Datasets (https://www.ncdc.noaa.gov/data-access/paleoclimatology-data/datasets)	landcover	Mesozoic	2000–2100 CE	n.a.	global	‘…time series of geophysical or biological measurements and some include reconstructed climate variables such as temperature and precipitation.’
Harmonized Global Land Use for Years 1500–2100, V1 (https://daac.ornl.gov/cgi-bin/dsviewer.pl?ds_id=1248)	climate	1500–1600 CE	2000–2100 CE	n.a.	global	‘…fractional land use and land cover patterns annually for the years 1500–2100 for the globe at 0.5-degree (∼50 km) spatial resolution.’
NOAA Paleoclimatology Datasets (https://www.ncdc.noaa.gov/data-access/paleoclimatology-data/datasets)	climate	Mesozoic	2000–2100 CE	n.a.	global	‘…time series of geophysical or biological measurements and some include reconstructed climate variables such as temperature and precipitation.’

^aDirect quotes from resource.

Not every Green List of Species assessment requires a large-scale effort of data collation and analysis to determine an accurate indigenous range; pragmatic considerations may dictate that large-scale assessment initiatives take a lighter touch approach, just as happened for the Red List [61]. However, the Green List of Species can provide a catalyst for developing a better understanding of the indigenous range of species that are the subject of conservation attention. We hope that Conservation Archive will accelerate the use of archival data not only for the purposes of the Green List of Species, but also to help achieve a range of conservation goals.

7. Future directions

The Green List of Species framework, while still in its infancy, offers a vital opportunity to put the current status of species in the broader context of a longer-term goal for recovery that is informed by what we know from history. Indeed, a limitation of the Red List is that because the criteria are designed to detect species currently at risk of extinction, they do not identify species that were once much more numerous and/or widespread; hence, individual species Red List assessments do not reflect the general status of biodiversity within a full historical context [61]. This is particularly the case for species that were historically exploited but have since stabilized at lower levels, such as marine turtles and large marine mammals.

In such cases, the future-focused Conservation Gain and Recovery Potential elements of the Green List of Species framework provide the ideal contextual counterpart to the Red List ([10]; figure 1) because they provide a mechanism to recognize and quantify improvements or declines in species recovery status separate from changes in extinction risk. This gives a strong message about the need and potential for conservation, highlighting that even if a species may no longer be at risk of extinction, concerted and sustained action may still be needed to support it towards recovery. The ‘Recovery Potential’ metric (figure 1), in particular, provides a realistic assessment of the degree to which the species could return to functionality within its entire indigenous range, and therefore could be particularly informed by palaeoecological analyses. As discussed in this paper, determining species' functionality requires an understanding of a species' historical abundance and role within an ecosystem, not just the extent of its historical range.

Figure 1.

Illustration of the application of the Green List of Species framework to an assessment of a depleted species and how it complements current measures of species status. Panel (a) (adapted from [61]) depicts a hypothetical species (generation length 10 years) that has gone through a period of rapid decline (of either population size or geographical range size, here represented by extent of occurrence) but has now stabilized at a much-reduced level. Because the rate of decline no longer meets Criterion A thresholds and the population has stabilized above thresholds for Criteria B and C, at seven generations before present the species no longer qualifies as threatened under Red List Criteria (b). If the species were to begin to recover in the future, this would not cause a change in Red List status (LC, Least Concern; NT, Near Threatened; VU, Vulnerable; EN, Endangered; CR, Critically Endangered) (b). Under the Green List of Species approach, a Conservation Gain score is produced, representing the recovery the species is expected to achieve within three generations assuming conservation efforts are sustained and/or increased (G, c), and a longer-term Recovery Potential score, which represents the maximum expected recovery within 100 years (P, c). Note that the maximum expected recovery remains some distance from the historical abundance index and indeed may never achieve that level. (Online version in colour.)

For too long, palaeoecology and neoecology have worked in isolation from one another [62]. The IUCN Green List of Species presents a catalyst for palaeoecologists and neoecologists to start talking to each other, and its potential to influence policy and practice might help usher in an era of truly applied palaeoecology.

Data accessibility

The database referenced in this paper is available online at https://conservationarchive.shinyapps.io/ConservationArchive/. If you wish to see the underlying code for the database, contact the corresponding author.

Authors' contributions

All authors are members of the IUCN SSC Species Conservation Success Task Force and have worked together to develop the Green List of Species Concept. M.G. was the primary author of this paper, and M.H. contributed greatly to early drafts. All other authors contributed equally in contributing to later drafts and revisions.

Competing interests

The authors of this paper are directly involved with the development of the Green List of Species as members of the IUCN SSC Species Conservation Success Task Force.

Funding

NERC Knowledge Exchange Fellowships NE/R002614/1 and NE/S006125/1 supported M.G. Stony Brook University OVPR Seed Grant Program supported H.R.A.