Biodiversity scenarios neglect future land-use changes
Abstract
Efficient management of biodiversity requires a forward-looking approach based on scenarios that explore biodiversity changes under future environmental conditions. A number of ecological models have been proposed over the last decades to develop these biodiversity scenarios. Novel modelling approaches with strong theoretical foundation now offer the possibility to integrate key ecological and evolutionary processes that shape species distribution and community structure. Although biodiversity is affected by multiple threats, most studies addressing the effects of future environmental changes on biodiversity focus on a single threat only. We examined the studies published during the last 25 years that developed scenarios to predict future biodiversity changes based on climate, land-use and land-cover change projections. We found that biodiversity scenarios mostly focus on the future impacts of climate change and largely neglect changes in land use and land cover. The emphasis on climate change impacts has increased over time and has now reached a maximum. Yet, the direct destruction and degradation of habitats through land-use and land-cover changes are among the most significant and immediate threats to biodiversity. We argue that the current state of integration between ecological and land system sciences is leading to biased estimation of actual risks and therefore constrains the implementation of forward-looking policy responses to biodiversity decline. We suggest research directions at the crossroads between ecological and environmental sciences to face the challenge of developing interoperable and plausible projections of future environmental changes and to anticipate the full range of their potential impacts on biodiversity. An intergovernmental platform is needed to stimulate such collaborative research efforts and to emphasize the societal and political relevance of taking up this challenge.
Introduction
Biodiversity plays an important role in the provision of ecosystem functions and services (Mace et al., 2012; Bennett et al., 2015; Oliver et al., 2015a). Yet, it is undergoing important decline worldwide due to human-induced environmental changes (Collen et al., 2009; Pimm et al., 2014). Governance and anticipative management of biodiversity require plausible scenarios of expected changes under future environmental conditions (Sala et al., 2000; Pereira et al., 2010; Larigauderie et al., 2012). A forward-looking approach is essential because drivers of biodiversity decline and their associated impacts change over time. In addition, delayed mitigation efforts are likely more costly and time-consuming than early action and often fail to avoid a significant part of the ecological damage (Cook et al., 2014; Oliver & Roy, 2015). Hence, biodiversity scenarios are on the agenda of international conventions, platforms and programmes for global biodiversity conservation, such as the Convention on Biological Diversity (CBD) and the Intergovernmental Platform for Biodiversity & Ecosystem Services (IPBES) (Leadley et al., 2010, 2014; Secretariat of the Convention on Biological Diversity, 2014; Díaz et al., 2015; Kok et al., 2016).
An increasing number of ecological models have been proposed over the last decades to develop biodiversity scenarios (Evans et al., 2013; Kerr & Dobrowski, 2013; Thuiller et al., 2013). They integrate and predict the effects of the two main factors that will determine the future of biodiversity: (1) the nature, rate and magnitude of expected changes in environmental conditions and (2) the capacity of organisms to deal with these changing conditions through a range of ecological and evolutionary processes (Fig. 1). Most modelling approaches rely on strong assumptions about the key processes that shape species distribution, abundance, community structure or ecosystem functioning (Kearney & Porter, 2009; Evans, 2012; Thuiller et al., 2013), with only few studies considering the adaptation potential of the species. Hence, recent work has mainly focused on improving the theoretical foundation of ecological models (Evans et al., 2013; Thuiller et al., 2013; Harfoot et al., 2014a; Zurell et al., 2016).
Yet, the credibility of developed biodiversity scenarios remains severely limited by the assumptions used to integrate the expected changes in environmental conditions into the ecological models. Biodiversity scenarios draw upon narratives (storylines) of environmental change that project plausible socio-economic developments or particularly desirable future pathways under specific policy options and strategies (van Vuuren et al., 2012; O'Neill et al., 2015) (Fig. 1). Although biodiversity is affected by multiple interacting driving forces (Millennium Ecosystem Assessment, 2005; Mantyka-Pringle et al., 2012; Settele & Wiemers, 2015), most biodiversity scenarios are based on environmental change projections that represent a single threat only (Bellard et al., 2015). With a literature survey on the biodiversity scenarios published during the last 25 years, we show here a dominant use of climate change projections and a relative neglect of future changes in land use and land cover. The emphasis on the impacts of climate change reflects the urgency to deal with this threat as it emerges from studies, data and reports such as those produced by the Intergovernmental Panel on Climate Change (IPCC) (Tingley et al., 2013; Settele et al., 2014). The direct destruction or degradation of habitats is, however, among the most significant threats to biodiversity to date (Millennium Ecosystem Assessment, 2005; Leadley et al., 2014; Newbold et al., 2015, 2016) and not including them raises concerns for the credibility of biodiversity scenarios. Habitat destruction and degradation result from both changes in the type of vegetation or human infrastructures that cover the land surface (i.e. land cover) and changes in the manner in which humans exploit and manage the land cover (i.e. land use) (Verburg et al., 2012; van Asselen & Verburg, 2013). The lack of coherent and interoperable environmental change projections that integrate climate, land use and land cover across scales constitutes a major research gap that impedes the development of credible biodiversity scenarios and the implementation of efficient forward-looking policy responses to biodiversity decline. We identify key research challenges at the crossroads between ecological and environmental sciences, and we provide recommendations to overcome this gap.
Climate and land-use/cover changes are important drivers of biodiversity decline
Biodiversity decline results from a number of human-induced drivers of change, including land-use/cover change, climate change, pollution, overexploitation and invasive species (Pereira et al., 2012; Leadley et al., 2014). Ostberg et al. (2015) have recently estimated that climate and land-use/cover changes have now reached a similar level of pressure on the biogeochemical and vegetation-structural properties of terrestrial ecosystems across the globe, but during the last three centuries land-use/cover change has exposed 1.5 times as many areas to significant modifications as climate change. The relative impacts of these driving forces on biodiversity have also been assessed at the global scale. In its volume on state and trends, the Millennium Ecosystem Assessment (2005) reported that land-use/cover change in terrestrial ecosystems has been the most important direct driver of changes in biodiversity and ecosystem services in the past 50 years. Habitat destruction or degradation due to land-use/cover change constitutes an on-going threat in 44.8% of the vertebrate populations included in the Living Planet Index (WWF, 2014) for which threats have been identified, whereas climate change is a threat in only 7.1% of them. A query performed on the website of the IUCN Red List of Threatened species (assessment during the period 2000–2015) indicates that more than 85% of the vulnerable or (critically) endangered mammal, bird and amphibian species in terrestrial ecosystems are affected by habitat destruction or degradation (i.e. residential and commercial development, agriculture and aquaculture, energy production and mining, transportation and service corridors, and natural system modification) and less than 20% are affected by climate change and severe weather conditions (see also Pereira et al., 2012). Interactions between multiple driving forces, such as climate, land-use and land-cover changes, may further push ecological systems beyond tipping points (Mantyka-Pringle et al., 2012; Oliver & Morecroft, 2014) and are key to understanding biodiversity dynamics under changing environmental conditions (Travis, 2003; Forister et al., 2010; Staudt et al., 2013; Mantyka-Pringle et al., 2015).
Emphasis on climate change impacts in biodiversity scenarios
Available projections of climate and land-use/cover changes (van Vuuren et al., 2012; O'Neill et al., 2015) are used to inform on future environmental conditions for biodiversity across a variety of spatial and temporal scales (de Chazal & Rounsevell, 2009) (Fig. 1). Many studies have predicted the consequences of expected climate change on biodiversity (Bellard et al., 2012; Staudinger et al., 2013; Pacifici et al., 2015). For instance, future climate change is predicted to induce latitudinal or altitudinal shifts in species ranges with important effects on ecological communities (Maes et al., 2010; Barbet-Massin & Jetz, 2015), to increase the risks of species extinction (Thomas et al., 2004; Urban, 2015) or to reduce the effectiveness of conservation areas (Araújo et al., 2011). Projections of land-use/cover change have been used to predict future changes in suitable habitats for a number of species (Martinuzzi et al., 2015; Newbold et al., 2015), to predict future plant invasions (Chytrý et al., 2012), to estimate potential future extinctions in biodiversity hotspots (Jantz et al., 2015) or to highlight the restricted potential for future expansion of protected areas worldwide (Pouzols et al., 2014). Visconti et al. (2015) estimated the coverage of suitable habitats for terrestrial mammals under future land-use/cover change and based on global protected areas expansion plans. They showed that such plans might not constitute the most optimal conservation action for a large proportion of the studied species and that alternative strategies focusing on the most threatened species will be more efficient.
Climate and land-use/cover change projections have also been combined in the same modelling framework to address how climate change will interplay with land-use/cover change in driving the future of biodiversity (Jetz et al., 2007; Martin et al., 2013; Ay et al., 2014; Saltré et al., 2015; Visconti et al., 2016). For instance, future refuge areas for orang-utans have been identified in Borneo under projected climate change, deforestation and suitability for oil-palm agriculture (Struebig et al., 2015). Alkemade et al. (2009) used land-use/cover change, climate change and projections of other driving forces to predict the future impacts of different global-scale policy options on the composition of ecological communities. Recently, it has been shown that the persistence of drought-sensitive butterfly populations under future climate change may be significantly improved if semi-natural habitats are restored to reduce fragmentation (Oliver et al., 2015b).
We searched published literature from 1990 to 2014 to estimate the yearly number of studies that developed biodiversity scenarios based on climate change projections, land-use/cover change projections or the combination of both types of projections. A list of 2313 articles was extracted from the search procedure described in Table 1. We expected a number of articles within this list would only weakly focus on the development of biodiversity scenarios based on climate and/or land-use/cover change projections and therefore, we randomly sampled articles within this list (sample size: N = 300). We then carefully checked their titles and abstracts to allocate each of them to one of the following categories:
| Query | Field | Parameters | Motivation |
|---|---|---|---|
| #1 | Year | (1990–2014) | Restricts the time period of the results to the last 25 years |
| #2 | Topic | (scenari* OR projecti* OR predicti* OR simulati* OR forecast* OR foresight* OR storyline*) AND future* | Captures studies addressing predictions in the future |
| #3 | Topic | (“climat* chang*” OR “chang* climat*”) OR (“land use chang*” OR “land cover chang*” OR “land* chang*”) | Identifies studies focusing on climate or land-use/cover changes |
- We used Boolean operators “AND” to combine the different queries and we refined the obtained results using “Articles” as Document Type and using “Ecology” or “Biodiversity conservation” as Web of Science Categories. We also tested if the parameters that we used in the query #3 might potentially underestimate the number of studies focusing on land-use/cover change. To do so, we tried to capture land-use/cover change in a broader sense and we included additional parameters in the query #3 as follows: (“climat* chang*” OR “chang* climat*”) OR (“land-use chang*” OR “land-cover chang*” OR “land* system* chang*” OR “land* chang*” OR “habitat loss*” OR “habitat degradation*” OR “habitat chang*” OR “habitat modification*”). We refined the results as described above and we obtained a list of 2388 articles, that is, only 75 additional articles compared to the search procedure with the initial query #3 (see main text). Hence, the well-balanced design of the search procedure as described in the table does not underestimate the use of land-use/cover change projections compared to climate change projections in biodiversity scenarios studies.
- Article reporting on the development of biodiversity scenarios based only on climate change projections.
- Article reporting on the development of biodiversity scenarios based only on land-use/cover change projections.
- Article reporting on the development of biodiversity scenarios based on the use of climate and land-use/cover change projections.
- Article reporting on the development of biodiversity scenarios based on other types of environmental change projections.
- Article not reporting on the actual development of biodiversity scenarios.
We considered that articles reported on the development of biodiversity scenarios when they produced predictions of the response of biodiversity to future changes in environmental conditions.
We calculated for each year between 1990 and 2014 the proportions of studies allocated to each of the five categories among the random sample of articles. We used a window size of 5 years and we calculated two-sided moving averages of the yearly proportions along the 25-year long time series. With this approach, we smoothed out short-term fluctuations due to the limited sample size and we highlighted the long-term trend in the proportions of articles allocated to the different categories.
We used these smoothed proportions estimated from the sample of articles and the total number of 2313 articles extracted from the search procedure to estimate the yearly numbers of articles during 1990–2014 that reported on the development of biodiversity scenarios and that used climate change projections (category 1), land-use/cover change projections (category 2) and both types of environmental change projections (category 3).
Our survey revealed that the number of studies that have included the expected impacts of future land-use/cover change on biodiversity falls behind in comparison with the number of studies that have focused on the effects of future climate change (Fig. 2). Among the studies published during the period 1990–2014 and that drew upon at least one of these two driving forces to develop biodiversity scenarios, we estimated that 85.2% made use of climate change projections alone and that 4.1% used only projections of land-use/cover change. Climate and land-use/cover change projections were combined in 10.7% of the studies. A sensitivity analysis was carried out and indicates that the number of articles for which we checked the titles and abstracts was sufficient to reflect those proportions in a reliable way (Appendix S1 and Fig. S1). The imbalance between the use of climate and land-use/cover change projections has increased over time in the last 25 years and has now reached a maximum (Fig. 2).
Where biodiversity scenarios lack credibility
Disregarding future changes in land use or land cover when developing biodiversity scenarios assumes that their effects on biodiversity will be negligible compared to the impacts of climate change. Two main reasons are frequently brought forward when omitting to include the effects of land-use/cover change in biodiversity scenarios: (1) the available representations of future land-use/cover change are considered unreliable or irrelevant for addressing the future of biodiversity (e.g. Stanton et al., 2012) and (2) climate change could outpace land use and land cover as the greatest threat to biodiversity in the next decades (e.g. Bellard et al., 2012). Here, we build on these two lines of arguments to discuss the lack of credibility of assuming unchanged land use/cover in biodiversity scenarios and to stress the need for further development of land-use/cover change projections.
Available large-scale land-use/cover change projections are typically associated with a relatively coarse spatial resolution and a simplified thematic representation of the land surface (Verburg et al., 2012). This is largely due to the fact that most of these projections have been derived from integrated assessment models which simulate expected changes in the main land cover types and their impacts on climate through emission of greenhouse gases (de Chazal & Rounsevell, 2009; Verburg et al., 2012; Harfoot et al., 2014b). A strong simplification of the representation of land use and land cover is inevitable due to the spatial extent and computational complexity of these models. Some studies have implemented downscaling methods based on spatial allocation rules to improve the representation of landscape composition in large-scale projections (Verburg et al., 2006). Because their primary objective is to respond to the pressing need to assess future changes in climatic conditions and to explore climate change mitigation options, such downscaled projections use, however, only a small number of land cover types and are, consequently, of limited relevance for addressing the full impact of landscape structure and habitat fragmentation on biodiversity (de Chazal & Rounsevell, 2009; Verburg et al., 2012; Harfoot et al., 2014b).
In addition, much of land system science has focused on conversions between land cover types (e.g. from forest to open land through deforestation), but little attention has been paid to capture some of the most important dimensions of change for biodiversity that result from changes in land use within – and not only between – certain types of land cover (de Chazal & Rounsevell, 2009; van Asselen & Verburg, 2013; Stürck et al., 2015). Changes in land management regimes (e.g. whether grasslands are mown or grazed) and intensity of use (e.g. through wood harvesting or the use of fertilizers, pesticides and irrigation in cultivated areas) are known to strongly impact biodiversity (Pe'er et al., 2014) and are expected to cause unprecedented habitat modifications in the next decades (Laurance, 2001; Tilman, 2001). For instance, management intensification of currently cultivated areas (Meehan et al., 2011) rather than agricultural surface expansion will likely provide the largest contribution to the future increases in agricultural production (van Asselen & Verburg, 2013). These aspects of land-use change remain poorly captured and integrated into currently available projections (Rounsevell et al., 2012; Verburg et al., 2012; Stürck et al., 2015). Furthermore, the frequency and sequence of changes in land use and land cover, or the lifespan of certain types of land cover, interact with key ecological processes and determine the response of biodiversity to such changes (Kleyer et al., 2007; Watson et al., 2014). Although methods have become available to represent the dynamics and the expected trajectories of the land system (Rounsevell et al., 2012), these temporal dimensions of change are still rarely incorporated in land-use/cover change projections (de Chazal & Rounsevell, 2009; Harfoot et al., 2014b).
This lack of integration between ecological and land system sciences limits the ability to make credible evaluations of the future response of biodiversity to land-use and land-cover changes in interaction with climate change (de Chazal & Rounsevell, 2009; Harfoot et al., 2014b). In turn, this makes it hazardous to speculate that the expected rate and magnitude of climate change will downplay the effects of land-use/cover change on biodiversity in the future. There is no consensus on how the strength of future climate change impact should be compared to that of other threats such as changes in land use and land cover (Tingley et al., 2013). Some of the few studies that included the combined effect of both types of drivers in biodiversity scenarios have stressed that, although climate change will severely affect biodiversity at some point in the future, land-use/cover change may lead to more immediate and even greater biodiversity decline in some terrestrial ecosystems (Jetz et al., 2007; Pereira et al., 2010; Visconti et al., 2016). For example, considerable habitat loss is predicted in some regions during the next few decades due to increasing pressures to convert natural habitats into agricultural areas (Lambin & Meyfroidt, 2011). The rapid conversion of tropical forests and natural grasslands for agriculture, timber production and other land uses (Laurance & Edwards, 2014) is expected to have more significant impacts on biodiversity than climate in the near future (Jetz et al., 2007; Laurance et al., 2012). Again, most of these studies focused on changes that will emerge from conversions between different types of land cover and only few of them addressed the future impacts of land-use change within certain types of land cover. For instance, the distribution changes of broad habitat types were predicted under future climate, land-use and CO2 change projections in Europe and it was shown that land-use change is expected to have the greatest effects in the next few decades (Lehsten et al., 2015). In this region, effects of land-use change might lead to both a loss and a gain of habitats benefitting different aspects of biodiversity. This will likely happen through parallel processes of intensification and abandonment of agriculture that offer potential for recovering wilderness areas (Henle et al., 2008; Queiroz et al., 2014). These immediate effects of land-use/cover changes on biodiversity deserve further attention with regard to the ecological forecast horizon, i.e. how far into the future useful predictions can be made (Petchey et al., 2015). Immediate changes in land use/cover may significantly alter the ability of ecological systems to deal with the impacts of climate change that are expected to be increasingly severe in the future (Tingley et al., 2013). Hence, ecological predictions that neglect the immediate effects of land-use/cover changes and only focus on the effects of climate change in a distant future may be largely uncertain. It is therefore needed to identify appropriate time horizons for biodiversity scenarios, with increased reliance on those associated with greater predictability and higher policy relevance (Petchey et al., 2015).
Climate change will exert severe impacts on the land system, but the way humans are managing the land will also influence climatic conditions, so that both processes interact with each other. For instance, deforestation and forest management constitute a major source of carbon loss with direct impacts on the carbon cycle and indirect effects on climate (Pütz et al., 2014; Naudts et al., 2016). Climate change mitigation strategies include important modifications of the land surface such as the increased prevalence of biofuel crops. This mitigation action may pose some conflicts between important areas for biodiversity conservation and bioenergy production (Alkemade et al., 2009; Fletcher et al., 2011; Meller et al., 2015). In integrated assessment models or other global land-use models, such interactions are often restricted to impacts of climate change on crop productivity and shifts in potential production areas. These models neglect a wide range of human adaptive responses to climate change in the land system (Rounsevell et al., 2014), such as spatial displacement of activities (Lambin & Meyfroidt, 2011) that may pose a significant threat to biodiversity (Estes et al., 2014). Increased attention to the feedback effects between climate and land-use/cover changes is therefore needed to help assessing the full range of consequences of the combined impacts of these driving forces on biodiversity in the future.
Both climate and land-use/cover changes are constrained or driven by large-scale forces linked to economic globalization, but the actual changes in land use/cover are largely determined by local factors (Lambin et al., 2001; Lambin & Meyfroidt, 2011; Rounsevell et al., 2012). Modifications in the land system are highly location-dependent and a reflection of the local biophysical and socio-economic constraints and opportunities (Rounsevell et al., 2014). In Europe, observed changes in agricultural practices in response to increased market demands and globalization of commodity markets include the intensification of agriculture, the abandonment of marginally productive areas, and the changing scale of agricultural operations. These processes occur at the same time but at different locations across the continent (Henle et al., 2008; Stürck et al., 2015; van Vliet et al., 2015). Hence, land-use/cover change and its impacts on biodiversity are highly scale-sensitive processes: they show strongly marked contrasts from one location to the other (Tzanopoulos et al., 2013). Many subtle changes that are locally or regionally significant for biodiversity may be seriously underestimated in the available land-use/cover change projections because they are occurring below the most frequently used spatial, temporal and thematic resolution of analysis in large-scale land-use models (de Chazal & Rounsevell, 2009). Most statistical downscaling approaches based on spatial allocation rules neglect such scale-sensitivity issues and therefore fail to represent landscape composition and structure to appropriately address the local or regional impacts of land-use/cover changes on biodiversity (Verburg et al., 2012).
A multi-scale, integrated approach is therefore required to unravel the relative and interacting roles of climate, land use and land cover in determining the future of biodiversity across a range of temporal and spatial scales. A good example of this need is the prediction of the impacts of changes in disturbance regimes, such as fire, for which idiosyncratic changes may be expected in particular combinations of future climate and land-use/cover changes (Brotons et al., 2013; Regos et al., 2016).
A way forward for biodiversity scenarios
Most large-scale land-cover change projections are derived from integrated assessment models. They are coherent to some extent with climate change projections because they are based on the same socio-economic storylines. This coherence is useful for studying the interplay between different driving forces. Integrated assessment models capture human energy use, industrial development, agriculture and main land-cover changes within a single modelling framework. However, their original, primary objective is to provide future predictions of greenhouse gas emissions. It is therefore important to recognize that these models are not designed to describe the most relevant aspects of land-use and land-cover changes for (changes in) biodiversity (de Chazal & Rounsevell, 2009; Harfoot et al., 2014b). Here, we provide two recommendations to increase the ecological relevance of land-use/cover change projections: (1) reconciling local and global land-use/cover modelling approaches and (2) incorporating important ecological processes in land-use/cover models.
Novel and flexible downscaling and upscaling methods to reconcile global-, regional- and local-scale land-use modelling approaches are critically required and constitute one of the most burning issues in land system science (Letourneau et al., 2012; Rounsevell et al., 2012; Verburg et al., 2012). An important part of the land-use modelling community focuses on the development of modelling and simulation approaches at local to regional scales where human decision-making and land-use/cover change processes are incorporated explicitly (Rounsevell et al., 2014). These models offer potential to include a more detailed representation of land-use/cover trajectories than integrated assessment models. Beyond the classification of dominant land cover types, they inform on land use, intensity of use, management regimes and other dimensions of land-use/cover changes (van Asselen & Verburg, 2013). An integration of scales will provide the opportunity to better represent the interactions between local trajectories and global dynamics (Kok et al., 2016) and to deal more explicitly with scale-sensitive factors such as land-use/cover changes (Tzanopoulos et al., 2013). To achieve this integration, a strengthened connection between ecological and land-use modelling communities is needed as it would ensure that the spatial, temporal and thematic representation of changes in land-use models matches with the operational scale at which biodiversity respond to these changes. Harfoot et al. (2014b) recently suggested development needs for integrated assessment models and recommended the general adoption of a user-centred approach that would identify why ecologists need land-use/cover change projections and how they intend to use them to build biodiversity scenarios. Although we believe such an approach will also be needed to ensure the ecological relevance of integrating the different scales of analysis in land-use models, this will only be successful if ecologists increase their use of already available land-use/cover change projections and suggest concrete modifications to improve their ecological relevance (de Chazal & Rounsevell, 2009; Martin et al., 2013). To address the scale-sensitivity issue thoroughly, we should also move beyond the current emphasis on large and coarse scale of analysis in global change impact research and increase our recognition for studies examining the local and regional effects of climate and land-use/cover changes on biodiversity.
Ecological processes in marine, freshwater or terrestrial ecosystems remain poorly incorporated in existing integrated assessment models and other land-use models (Harfoot et al., 2014b). Ecological processes in natural and anthropogenic ecosystems provide essential functions, such as pollination, disease or pest control, nutrient or water cycling and soil stability, that exert a strong influence on land systems through complex mechanisms (Sekercioglu et al., 2004; Klein et al., 2007). Incorporating these processes at appropriate spatial and temporal scales in land-use models constitutes an important challenge, but it would considerably increase the ecological realism of these models and, in turn, their ability to predict emergent behaviour of the future ecosystems and the related biodiversity patterns (Harfoot et al., 2014b). Therefore, we urge the need for strengthened interactions between different scientific communities to identify (1) which ecological processes are relevant in driving land-use/cover dynamics and (2) how and at which scales these processes could be incorporated in land-use models to predict the trajectories of socio-ecological systems.
A successful implementation of our two recommendations does not solely depend on collaborative scientific efforts, but it also requires societal agreement and acceptance. The dialogue with and engagement of stakeholders, such as policy advisers and NGOs, within a participatory modelling framework (Voinov & Bousquet, 2010) will be key to agreeing on a set of biodiversity-oriented storylines and desirable pathways at relevant spatial and temporal scales for decision-making processes in biodiversity conservation and management. An improved integration of the expertise and knowledge from social science into the development and interpretation of the models may allow a better understanding of likely trajectories of land-use/cover changes. Moreover, such an integration would provide a better theoretical understanding and practical use of social-ecological feedback loops in form of policy and management responses to changes in biodiversity and ecosystem services, which in turn will impact future land-use decisions and trajectories.
The priority given to investigating future climate change impacts on biodiversity most likely reflects how the climate change community has attracted attention during the last decades. The availability of long-term time series of climatic observations in most parts of the world and the increasing amount of science-based, spatially explicit climatic projections derived from global and regional circulation models have clearly stimulated the development of studies focusing on the impacts of climate change (Tingley et al., 2013; Harfoot et al., 2014b). Under the World Climate Research Programme (WCRP), the working group on coupled modelling has established the basis for climate model diagnosis, validation, intercomparison, documentation and accessibility (Overpeck et al., 2011). The requirements for climate policy, mediated through the IPCC, have further mobilized the use of a common reference in climate observations and simulations by the scientific community. The set of common future emission scenarios (SRES) released in 2000 (Nakicenovic et al., 2000), the more recent representative concentration pathways (RCPs) (van Vuuren et al., 2011), and the fact that these can be shared easily have played a major role in mobilizing the scientific community to use climate change projections in biodiversity scenarios. Work is underway to facilitate open access to land-use/cover change time series and projections, but clear and transparent documentation of land-use model representations, uncertainties and differences is also needed and should be understandable and interpretable by a broad interdisciplinary audience (Harfoot et al., 2014b).
The IPCC has also clearly demonstrated that an independent intergovernmental body is an appropriate platform for attracting the attention of the nonscientific community. Many actors now perceive climate change as an important threat to ecosystem functions and services. This emphasis can be heard in the media and among policy-makers, such as during the United Nations conferences on climate change. As a response to the increasing societal and political relevance of climate change, research efforts have been mostly directed towards climate change impact assessments (Herrick et al., 2013). From this observed success of the IPCC and the climate change community, it becomes evident that an independent body is needed for mobilizing the scientific and nonscientific communities to face the significant challenge of developing biodiversity-oriented references for land-use and land-cover change projections. With its focus on multiscale, multidisciplinary approaches, the working programme of the Intergovernmental Platform on Biodiversity and Ecosystem Services (IPBES) (Inouye, 2014; Díaz et al., 2015; Lundquist et al., 2015) is offering a suitable context to stimulate collaborative efforts for taking up this challenge. In line with Kok et al. (2016), we therefore encourage IPBES to strengthen its investment in the development and use of interoperable and plausible projections of environmental changes that will allow to better explore the future of biodiversity.
Conclusion
Neglecting the future impacts of land-use and land-cover changes on biodiversity and focusing on climate change impacts only is not a credible approach. We are concerned that such an overemphasis on climate change reduces the efficiency of identifying forward-looking policy and management responses to biodiversity decline. However, the current state of integration between ecological and land system sciences impedes the development of a comprehensive and well-balanced research agenda addressing the combined impacts of future climate, land-use and land-cover changes on biodiversity and ecosystem services. We recommend addressing two key areas of developments to increase the ecological relevance of land-use/cover change projections: (1) reconciling local and global land-use/cover modelling approaches and (2) incorporating important ecological processes in land-use/cover models. A multidisciplinary framework and continuing collaborative efforts from different research horizons are needed and will have to build on the efforts developed in recent years by the climate community to agree on a common framework in climate observations and simulations. It is now time to extend these efforts across scales in order to produce reference environmental change projections that embrace multiple pressures such as climate, land-use and land-cover changes. IPBES offers a timely opportunity for taking up this challenge, but this independent body can only do so if adequate research efforts are undertaken.
Acknowledgements
N.T., K.H., J.B.M., I.R.G., W.C. and L.B. acknowledge support from the EU BON project (no. 308454, FP7-ENV-2012, European Commission, Hoffmann et al., 2014). N.T. and L.B. were also funded by the TRUSTEE project (no. 235175, RURAGRI ERA-NET, European Commission). N.T., A.R. and L.B. were also supported by the FORESTCAST project (CGL2014-59742, Spanish Government). I.R.G. and W.C. contribute to the Labex OT-Med (no. ANR-11-LABX-0061) funded by the French Government through the A*MIDEX project (no. ANR-11-IDEX-0001-02). P.H.V. received funding from the GLOLAND project (no. 311819, FP7-IDEAS-ERC, European Commission). We thank Piero Visconti and one anonymous reviewer for useful comments on a previous version of this paper.
