The past is a foreign country: how much can the fossil record actually inform conservation?

Humans have driven biodiversity loss and modified ecosystem structure for millennia. Using modern ecological data therefore has the risk of considerably affecting our understanding of ecological patterns of “natural” species distributions and of the dynamics and drivers of past extinctions. This bias ultimately narrows our perception of the options available for conservation and the potential for species’ recovery. Long-term archives are often perceived as unsuitable for ecological analyses because of the substantial levels of bias and error they may contain. However, by extending the timeline usually considered in ecology, they can provide unique new insights into extinction dynamics and changing species distribution through time and represent a unique opportunity to better inform regional environmental management. Assessment and utilization of long-term archives thus represent both a methodological challenge and an important conservation research priority. Here, we integrate fossil, historical and modern occurrence records of large mammals in Southern Africa from the Holocene to the present and explore the challenges and opportunities associated with using long-term biodiversity data in ecological analyses. We present some limitations of fossil and historical records and highlight the importance of considering these long-term archives to estimate species’ historical distribution and natural habitat requirements. We evidence local extinctions and changes in community composition, consistent with a response to the demographic expansion of European colonists in South Africa. These results contribute to novel baselines for conservation and provide a strengthened evidence-base for understanding long-term faunal responses to human pressures. It also allows the “shifted baselines” around modern mammal distributions to be identified, providing an avenue for new analyses of large mammal biogeographic patterns in Southern Africa.

Long-term data are vital for tracking changes in biodiversity through time. The Holocene contains a rich fossil record of global biodiversity for the last 11,700 years to the present day. It is therefore considered a particularly robust data source for researching dynamics and drivers of species extinction, and for informing our understanding of the current extinction crisis. However, as with all fossil datasets it nonetheless suffers from incompleteness and bias. Without knowledge of the extent of these biases we risk making false inferences, which in turn limit the use of this data for conservation. Patterns of bias and/or incompleteness were investigated in two Holocene fossil data sets collected at different spatiotemporal scales. Firstly, factors associated with the description date of 256 extinct mammals were analysed in order to assess bias in our knowledge of global, species-level mammal extinctions. Secondly, sources of bias were identified in a dataset of nearly 20,000 fossil mammal records used to reconstruct population-level range change and extinction. A number of biases were found in both datasets, with the Holocene fossil record general biased towards the discovery and description of larger-bodied mammals in continental regions, compared with smaller-bodied mammals in insular regions. The implications of these results in shaping our understanding of past biodiversity loss and our ability to use the Holocene fossil record in conservation-related research are discussed.

Understanding past environmental baselines and the extent that human activities have already affected species and ecosystems is crucially important for today’s “conservation hotspots”, which contain disproportionate amounts of threatened biodiversity. It is necessary to evaluate the unique information content of different long-term archives, and determine what novel insights are available on past ecosystem structure and change, in order to understand how past baselines can inform current conservation research and management. Eastern and Southeast Asia contain the world’s highest numbers of threatened vertebrates and plants, and the region has a long history and prehistory of human occupation. Within this region, China (a huge, ‘megadiverse’ country) possesses diverse ecological archives including a millennial-scale historical record and rich Late Quaternary palaeontological and zooarchaeological records. By integrating and analysing a series of complementary, temporally non-overlapping archives (fossils, a ~400-year gazetteer record, and ethnic minority group oral traditions and local knowledge), the history of a regional Chinese mammal fauna is reconstructed across the Holocene, revealing that Hainan Island has experienced an ‘extinction filter’ associated with protracted, ongoing depletion of large mammal diversity over recent millennia, rather than recent ecological collapse of a faunally intact system. However, China’s Quaternary faunal archives are all incomplete and biased in their information content on key ecological parameters, and differ from each other in their representativeness of past diversity. Whereas palaeontological, zooarchaeological and historical records are an invaluable resource for reconstructing pre-human environments and biodiversity loss through time, interpreting and extrapolating what they show therefore requires both caution and context.

Conservation biologists face a daunting challenge: conserving biological diversity and services during a carbon release event potentially as big as the Paleocene-Eocene Thermal Maximum, rises in global mean temperatures to Pliocene-like levels, at rates as fast or faster than those during the abrupt events of the last glacial period and deglaciation, and with hydrological extreme events likely to be as or more severe than those of the Holocene. The emergence of novel 21st-century climate rates and states challenges the predictive ability of ecological forecasting models; hindcasting tests against benchmark fossil records from the last deglaciation show a steep drop-off in predictive ability as climatic novelty increases. Improved ecological forecasting, however, can be achieved by the close integration of predictive models with neo- and paleoecological observations encompassing a broad range of timescales, processes, and system states. Advances in data-model assimilation are particularly promising and require a tight and iterative coupling between data and models. This in turn requires the assembly of paleoecological data into open data systems and with data quality improved on an on-going basis by expert curation, new data generation, and synthesis. Such efforts can also indicate key areas of model uncertainty that can be the target of new data-gathering campaigns.

Most current real-world ventures named as rewilding fall under the concept of trophic rewilding, as do most discussion of rewilding. Trophic rewilding is an ecological restoration strategy that uses species introductions to restore top-down trophic interactions and associated trophic cascades to promote self-regulating biodiverse ecosystems (http://bit.ly/rewildingPNAS). Trophic rewilding is often focused on large-bodied animals (both herbivores and carnivores) due to their ecological and societal importance and the widespread, strong historical and recent prehistorical losses of megafauna in most regions around the world. In this presentation, I will first discuss the concept of trophic rewilding, and then outline the scientific background for trophic rewilding, integrating paleoecology and contemporary ecology. Key themes here will the fact that megafauna-rich ecosystems have been the norm on evolutionary timescales, with the widespread modern-day megafauna-poor ecosystems being a novel situation, and the emerging evidence for the strong ecological importance of large-bodied animals. I will also discuss the current role of trophic rewilding in nature conservation and ecological restoration as well as its potential, now and in the future. Linked to this, I will also discuss tricky issues such as base-lines, relations to other land uses, perceived risks to people and ecosystems, and the use of exotic species in rewilding. Finally, I will outline key priorities and possibilities for this research field.

Ancient human impacts to ecosystems are often underestimated because such changes took place gradually and left little recorded evidence. This translates into distorted perceptions of what natural ecosystems correspond to – i.e., into shifted ecological baselines – that have pervasive effects on management and conservation decisions. For example, biodiversity indicators built from recent data on species’ distributions, population sizes and trends – such as the Red List Index and the Living Planet Index – underestimate the extent of human impacts and may translate into unambitious conservation goals and recovery targets. It is however not straightforward to obtain biodiversity indicators that take into account ancient impacts, given the incompleteness of the historical record. Furthermore, variations in the timeline of impacts across species and regions mean that there is not a single appropriate baseline date that is applicable to all. This presentation will introduce a new approach for assessing populations according to the extent to which they have been impacted by past human activities. It will illustrate its application to large whale species, to show how such assessments can be applied consistently across species with very different impact histories and with wide variations in the depth of available historical data, by making the best use of the information available in each case. Multi-species assessments can then be aggregated into temporal and spatial indices of biodiversity that are less prone to the shifting baseline syndrome.

The most widely-used global metric of species conservation status today is the IUCN Red List of Threatened Species, which assesses species’ extinction risk. The timescales over which changes in species status are evaluated must necessarily be relatively short (10 years or 3 generations, depending on the species) in order to highlight rapid declines in population size or range which may signal acute threats to species persistence. However, this relatively narrow temporal focus presents a danger of shifting baselines, since declines which have occurred in previous centuries often do not fall within the relevant time window for assessment.

This talk outlines the opportunities and difficulties in using historical data (accounts, museum records, zooarchaeological and fossil records, etc.) to inform Red List assessments and also introduces the IUCN’s response to the challenge of shifting baselines: The Green List of Species. The Green List of Species, currently in development, will be used to measure species recovery against a baseline set centuries ago, rather than decades. The Green List will allow long-term ambitions to restore species to historical levels to work in parallel with the crucial short-term focus on preventing extinction, providing a more comprehensive conservation story that includes a species’ past, present, and potential future.