Managing landscapes for biodiversity during rapid climate change.

As the climate warms, all wild species will have to adjust or adapt if they are to survive. We know from studying climate change after previous ice ages, that species all over the temperate parts of the earth adapted by shifting their geographic ranges, following where the climate led them. We also have evidence that numerous species are shifting their ranges now, but most do not seem to be "keeping up" with the climate.

The problem is that natural (or seminatural) habitat for the majority of species is now rare and highly fragmented compared to the era before agricultural intensification and industrialisation. When a species' habitat is restricted to a tiny fraction of the land area, it becomes very difficult for it to invade new territory at the cool edge of its range: it could easily get stuck in isolated fragments of habitat and eventually go extinct as the climate becomes unsuitable. This is a major stimulus for conservation organisations and governments to recreate and restore habitats in order to reconnect potentially isolated populations. However it is unrealistic to suppose we could restore the areas of some habitats to their prehistoric levels. For example, in the UK woodland used to be the dominant habitat. To provide effective conservation of biodiversity in the UK, we need to make informed decisions about (a) how much woodland is enough to preserve the majority of species and (b) how to target woodland creation where it will be most beneficial in allowing range shifting to occur. We also have to find a way of proceeding when we don't know all the factors that limit each species' distribution now (factors ranging from temperature and rainfall to interactions with other species), and cannot reliably predict the direction and distance they will need to shift.

This piece of research will assist this decision-making by finding a general-purpose strategy for habitat creation that will work for most species in most landscapes. We will test a number of strategies - all in the form of a numerical rule for deciding, given an existing landscape, which land parcel to choose next for conversion. As examples of existing landscapes we will use real landscapes in the UK, with habitat information from our project partner Forest Research and from the satellite derived "Land Cover Map 2007". We will judge the strategies based on simulations of animal and plant species invading the landscape (assuming they start from a small population at any one edge). The variation in traits of our simulated species (e.g. habitat specialism, reproduction rate and dispersal ability) will be based on the variation observed across all species for which data exists. We hope that this will produce results that are representative of the whole community, rather than focussing just on a few well-studied species. We will also include genetic information in our simulations - genetic diversity is an important, and sometimes overlooked, aspect of biodiversity, and one which may be dramatically reduced as the population passes through habitat "bottlenecks" during range expansion.

Based on the simulation results we will be able to rank the landscapes both in terms of the proportion of species that survive, and the proportion of genetic diversity that survives in these populations. Our results will be useful to planners and conservation organisations, but also to any land owners who are interested in creating habitat to benefit wildlife. The ranking tool we will develop will be able to estimate the future value of a proposed habitat patches (e.g. areas of woodland) to biodiversity, and therefore strengthen the case for acquiring and maintaining the best sites when there are other competing demands on land use.

We envisage three key impacts of this project. The most substantial would be a change in the way landscapes are designed such that biodiversity is better able to cope with climate change. To achieve this impact, key user groups have been identified and we plan for effective two-way knowledge exchange between representatives of these user groups and the project team. Also, the tool produced by the project will be freely available and we will provide initial training in appropriate use. A second impact that we expect the project to have is changing the direction taken within the academic community. The approach we will be taking is fundamentally different to that being pursued by the great majority of the community, which typically focuses on projecting (or predicting) where species will be in the future. We aim to shift the emphasis of research away from prediction using climate envelope modelling (and away from species-specific approaches) towards developing robust tools for management that are effective for a broad range of biodiversity. The third important impact is in training a highly skilled postdoctoral researcher such that they are able to integrate ecological and genetic models with real landscape data. Not only will this have a short-term impact in terms of the work detailed within this project, but promises downstream impact in terms of the work that the PDRA will contribute to in the future. There is a real lack of people with the skills to integrate dynamic models with GIS, yet this is a direction ecological modelling needs to take if it is to become more useful as a land-management tool. Thus, we place considerable importance on enabling the PDRA to acquire these skills. Below, we provide further information on how we plan to achieve our desired impacts.

The main output of our research will be a set of rules of thumb for deciding which habitat creation strategy is likely to deliver the greatest benefit in terms of wild species surviving climate change. The users will be able to take this estimated benefit, and set it against the costs of alternative conservation actions, to arrive at an informed decision. Information about the costs of conservation, and indeed the benefits and costs of competing land uses, will have to be sought from other sectors and therefore it is appropriate that much of this analysis happens at a high level in local or national government. However, all the habitat creation strategies that we investigate will work incrementally, i.e. they will answer the question "which land parcel should next be chosen for conversion?" Because of this, the strategies could be applied by individual landowners and need not be implemented "top-down", hence their particular value to the second user group. Additionally, our new dynamic methods have the potential to be linked to existing static multi-habitat optimisation algorithms and used to determine effective landscape planning strategies across multiple habitat types.

Some key milestones and measures of success are listed below:
- Is our work quoted in policy documents? - Are our models used by Forest Research and other end-user groups to inform landscape management?
- How many people request or download a Handbook describing use of the model?
- Is our work accepted for publication in high quality peer-reviewed journals?
- Destination of the postdoc - is their skill set sought after and do they progress to develop novel methodologies linking dynamic modelling and GIS that meet end-user needs and inform policy?