Timeline to collapse

Lead Research Organisation: University of Bristol
Department Name: Biological Sciences

Abstract

Biological systems on which humanity relies for food, fresh water, and clean air are becoming increasingly stressed. One way in which such stress manifests is through loss of resilience: ecosystems are increasingly at risk of rapid change, characterised by sudden collapses in the abundance of populations. Increasing anthropogenic stressors thus leave us at a critical point for ecosystem management, where to preserve biodiversity and the services on which society relies, we need to reliably detect what systems are at most risk of collapse, and thus where conservation effort should be targeted.

Such predictions are hampered not only by the nature of ecological systems, which are inherently high dimensional and non-linear, but also by a lack of long-term high-resolution data for many populations and communities. Such limitations mean that process-to-pattern methods such as mechanistic modelling approaches - whereby detailed models on a specific system are parameterised and perturbed to assess whether that system is at risk - are unsuitable for predicting the fates of most systems.

An alternative school of thought has suggested a pattern-to-process approach, whereby signals observed in data are used to infer changes in the structure of a system which may lead to its collapse. These pattern-to-process phenomenological approaches required fewer data and make fewer assumptions about the structure of the system, providing generalisable rule-of-thumb methods to warn of approaching disaster. Were they to be reliable, such methods would be an invaluable tool to help manage and conserve biodiversity in a rapidly changing world.

However, the promise of such phenomenological methods is negated by the lack of basic testing. For example, whilst many warning methods have been developed (based fluctuations in population abundances, declines in body size, and changes in the spatial distribution of populations) thus far no work has compared the relative predictive efficacy of these methods. Moreover, we know nothing about how key drivers of stability such as community and landscape complexity affect our ability to make reliable predictions, catastrophic failings if such methods are to be used to inform the management of heterogeneous populations and communities in the real world.

In this project, we will take a bold new approach to the advancement and testing of phenomenological warning signals by developing a spatially-explicitly multi-species experimental system which will allow high-resolution, high-dimensional data to be generated on multiple species at the landscape scale. We will use this system to (i) provide the first simultaneous test of currently proposed phenomenological warning methods, (ii) develop and test novel warning signals derived from ecological theory based on changes in the behaviour of individuals and assess how these perform in relation to previously proposed methods, and (iii) assess the reliability of warning signal methods across various spatial and community complexities, and thus their suitability for informing conservation decision making.

The work we are proposing is fundamental not only to or understanding of resilience loss in biological systems, but also to practical on-the-ground management of key resources on which human society relies. Our focus during this work will be not only to understand how integral drivers of population dynamics such as spatial and community complexity affect our ability to make reliable predictions, but how reliable the developed methods will be when subjected to the vagaries and inconsistencies seen in real-world conservation data. Thus, this project spans a range of disciplines, generating important insights in fields including demography and community ecology, whilst targeting significant downstream socio-economic impact by developing robust predictive frameworks to help minimise biodiversity loss in the face of anthropogenic forcing.

Planned Impact

In a rapidly changing world characterised by increasing anthropogenic pressures the maintenance of biodiversity is a key target not only governments worldwide who have committed to minimizing the loss of species to maintain critical ecosystem functions, but those who manage natural resources for economic and cultural reasons. Key to doing this effectively and efficiently are reliable predictive frameworks which can allow targeted intervention when necessary. Thus, a clear output of this project must be to disseminate the findings of our research not only to academics, but also to sectors where the methods developed can have real-world socio-economic impacts.

One sector to which the warning signals approaches to be developed and tested in this proposal are particularly relevant is in the management of natural capital. Indeed, PI Clements has recently published work on warning signals in fisheries stocks, a high economic value sector which could benefit significantly from robust predictive frameworks. Data-efficient warning signal tools would be invaluable to developing nations which do not have the data or resources to build mechanistic models of their fisheries but want to manage their stocks with a view to the well-being of future generations. For this reason, funding for a visit and to organise a workshop in Bangladesh is included in this proposal (4th quarter, year 2). We will organise a multi-stakeholder workshop and liaise directly with those who work on policy, resource management, and sustainable growth, particularly fisheries. This visit will occur in the second half of the project to give the opportunity to feedback the needs of those working on the ground who would implement such methods to imperfect data. In this way we can design methods which will be useful in real-world management situations and open up the potential for collaboration and future grant applications to apply the developed methods to on-the-ground management issues.

Beyond fisheries, such predictive tools offer an opportunity to bring a quantitative approach to conservation prioritisation, a key goal in the light of stretched budgets and the ever-increasing number of threatened ecosystems. To ensure that conservation practitioners are aware and able to use the quantitative methods developed, PI Clements will make use of his network of collaborators at the Zoological Society of London's (ZSL's) Institute of Zoology (IoZ) - where he carried out research during his PhD - to disseminate information about the findings of this project. The IoZ is the research wing of ZSL and runs and finances a large number of on-the-ground conservation actions globally across a broad range of taxa and ecosystems, many of which could benefit directly from the research to be carried out in this proposal. IoZ also has strong ties to policy makers, both in the UK and internationally, making them an ideal channel through which to pass information. Funding is included for yearly visits to discuss the results of this project and how to tie these in to policy and conservation decision making protocols.

We believe that central to the work to be carried out in this proposal is an understanding of the needs of those who may implement the developed methods and will assess the ongoing success of our impact programme during the funding period to ensure that these outreach goals are achieved.

Publications

10 25 50
 
Description University of Sheffield Partnership 
Organisation University of Sheffield
Country United Kingdom 
Sector Academic/University 
PI Contribution We are providing an experimental system to test theory being developed by the sheffield portion (Dr Dylan Childs and PDRA) of this grant.
Collaborator Contribution Dylan Childs (UoS) is contributing significantly to the in review publication and design of statistical methods/experiments being carried out.
Impact Currently in review publication
Start Year 2022