
==== Front
Philos Trans R Soc Lond B Biol Sci
Philos Trans R Soc Lond B Biol Sci
RSTB
royptb
Philosophical Transactions of the Royal Society B: Biological Sciences
0962-8436
1471-2970
The Royal Society

10.1098/rstb.2023.0011
rstb20230011
100160207Articles
Research Articles
Projected future climatic forcing on the global distribution of vegetation types
Projected future climatic forcing on the global distribution of vegetation types
http://orcid.org/0000-0003-0282-6407
Allen Bethany J. Conceptualization Formal analysis Investigation Methodology Project administration Visualization Writing – original draft Writing – review & editing bethany.allen@bsse.ethz.ch
1 2
http://orcid.org/0000-0001-5492-3925
Hill Daniel J. Conceptualization Formal analysis Investigation Methodology Visualization Writing – original draft Writing – review & editing 3
http://orcid.org/0000-0002-7033-7798
Burke Ariane M. Supervision Writing – review & editing 4
http://orcid.org/0000-0001-7161-7751
Clark Michael Supervision Writing – review & editing 5 6 7
http://orcid.org/0000-0001-5013-4056
Marchant Robert Supervision Writing – review & editing 8
http://orcid.org/0000-0003-0017-1654
Stringer Lindsay C. Supervision Writing – review & editing 8 9 10
http://orcid.org/0000-0002-0379-1800
Williams David R. Supervision Writing – review & editing 3
http://orcid.org/0000-0003-2319-2933
Lyon Christopher Conceptualization Funding acquisition Project administration Writing – review & editing christopher.lyon@york.ac.uk
8 9
1 Department of Biosystems Science and Engineering, ETH Zurich, Basel 4056, Switzerland
2 Computational Evolution Group, Swiss Institute of Bioinformatics, Lausanne 1015, Switzerland
3 School of Earth and Environment, University of Leeds, Leeds, LS2 9JT, UK
4 Département d'Anthropologie, Université de Montréal, Montréal, Quebec, H3C 3J7, Canada
5 Smith School of Enterprise and the Environment, University of Oxford, Oxford, OX1 3QY, UK
6 Oxford Martin School, University of Oxford, Oxford, OX1 3BD, UK
7 Department of Biology, University of Oxford, Oxford, OX1 3RB, UK
8 Department of Environment and Geography, University of York, York, YO10 5NG, UK
9 Leverhulme Centre for Anthropocene Biodiversity, University of York, York, YO10 5DD, UK
10 York Environmental Sustainability Institute, University of York, York, YO10 5DD, UK
One contribution of 16 to a theme issue ‘Ecological novelty and planetary stewardship: biodiversity dynamics in a transforming biosphere’.

Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.7131291.

27 5 2024 May 27, 2024
8 4 2024 April 8, 2024
8 4 2024 April 8, 2024
379 1902 Theme issue ‘Ecological novelty and planetary stewardship: biodiversity dynamics in a transforming biosphere’ compiled and edited by Jens-Christian Svenning, Melodie A. McGeoch, Signe Normand, Alejandro Ordonez and Felix Riede 202300117 7 2023 July 7, 2023
7 3 2024 March 7, 2024
© 2024 The Authors.
2024
https://creativecommons.org/licenses/by/4.0/ Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited.

Most emissions scenarios suggest temperature and precipitation regimes will change dramatically across the globe over the next 500 years. These changes will have large impacts on the biosphere, with species forced to migrate to follow their preferred environmental conditions, therefore moving and fragmenting ecosystems. However, most projections of the impacts of climate change only reach 2100, limiting our understanding of the temporal scope of climate impacts, and potentially impeding suitable adaptive action. To address this data gap, we model future climate change every 20 years from 2000 to 2500 CE, under different CO2 emissions scenarios, using a general circulation model. We then apply a biome model to these modelled climate futures, to investigate shifts in climatic forcing on vegetation worldwide, the feasibility of the migration required to enact these modelled vegetation changes, and potential overlap with human land use based on modern-day anthromes. Under a business-as-usual scenario, up to 40% of terrestrial area is expected to be suited to a different biome by 2500. Cold-adapted biomes, particularly boreal forest and dry tundra, are predicted to experience the greatest losses of suitable area. Without mitigation, these changes could have severe consequences both for global biodiversity and the provision of ecosystem services.

This article is part of the theme issue ‘Ecological novelty and planetary stewardship: biodiversity dynamics in a transforming biosphere’.

climate change
, global warming
, biosphere
, biome
, biodiversity
, anthropogenic
Eidgenössische Technische Hochschule Zürich http://dx.doi.org/10.13039/501100003006 ETH+ BECCY Leverhulme Trust http://dx.doi.org/10.13039/501100000275 White Rose University Consortium cover-dateMay 27, 2024
==== Body
pmc1. Introduction

Biomes represent a well-established scheme for categorizing ecosystems, and particularly the vegetation within them, at the global scale. Biomes have previously been defined in myriad ways [1–5], but are broadly recognized as structurally and functionally (but not necessarily taxonomically) consistent plant communities, sometimes associated with a specific set of climatic conditions and soil states [2,3,6,7]. Animal communities are also typically assumed to be well described by these delineations (e.g. [3,5]), and in some cases can also play a role in determining biome composition, structure and distribution [8]. Continental drift and climate change have governed the shifting of biomes over geological timescales, as evidenced in the fossil record (e.g. [9–11]).

Humans have also shaped, and will continue to shape, the distribution of biomes across the world. Around three-quarters of terrestrial area has been populated for at least 12 000 years [12], with accompanying changes in flora and fauna, and the creation of anthropogenic biomes (anthromes) [13]. Intensification of land use, invasive species and pollution are driving both extinction and extirpation, altering the composition of local communities at an alarming rate [14,15]. Anthropogenic climate change is also shifting the ranges of species [16], and has the potential to significantly impact the geographical distribution of biomes in the near future [7], including risks of abrupt shifts [17]. These vegetation transitions will result in feedback through changed land cover, probably contributing to additional climate warming [18–20].

Modelling what the future distribution of vegetation types might look like will be fundamentally important for conservation planning and management, particularly if certain biomes—and the species within them—are to be protected [7,21–24]. In addition, half of the world's human population live in areas which are highly vulnerable to biome change [7]. Such perturbations have the potential to negatively impact human wellbeing, through alterations or reductions in the ecosystem services they produce or receive [7,25,26]. This could prompt changes to the locations and forms of settlements, and their associated political economies [27].

Previous attempts to model vegetation change into the future have generally ended their projections at 2100 (e.g. [5,7,19]). However, climate change is expected to extend well beyond this interval [28], and will continue to reconfigure global environments for centuries to come [27,29]. Here, we build on the approach taken in Lyon et al. [27], who modelled climate change between 2000 and 2500 CE using different representative concentration pathway (RCP) scenarios [28]. We combine HadCM3 global climate model outputs with BIOME4, a vegetation model [30], to infer how the area suited to different biomes will change over time under these climate scenarios. We use these projections to infer how biomes could be geographically distributed in the future, and the impact this might have on their total area over time. We then investigate the feasibility of vegetation migrating to these locations, taking into account potential overlap with human land use using the History Database of the global Environment (HYDE) anthromes [31].

2. Methods

(a) Climate models

Projected future climate change was modelled under three different RCP scenarios: 2.6, 4.5 and 6.0 [32]. RCP2.6 represents a strong reduction in anthropogenic greenhouse gas emissions. RCP4.5 and RCP6 are medium stabilization scenarios, with moderate and limited levels of mitigation, respectively. RCP8.5 describes unmitigated growth in greenhouse gas emissions, but has been argued to be problematic and outdated [33]: this scenario would require the repeal of recent policy changes and commitments, and requires actions, such as a fivefold increase in coal burning by 2100, that have subsequently been demonstrated to be unrealistic [34]. We therefore chose to use RCP6 as the highest atmospheric greenhouse gas scenario.

Three simulations were run using the HadCM3 version of the UK Met Office Unified Model [35], each driven by atmospheric greenhouse gas concentrations from a different RCP scenario (2.6, 4.5 and 6.0), as outlined in Meinshausen et al. [28], using the procedure described in Lyon et al. [27]. The HadCM3B-M2.1a version of the model [36], used here, includes the top-down representation of interactive flora and foliage including dynamics (TRIFFID) dynamic global vegetation model (DGVM; [37]). This vegetation model simulates the fractional coverage of plant functional types (PFTs) within each climate model gridbox, based on a carbon balance approach, which actively updates the physical parameters in the land surface scheme. TRIFFID enables the climate model to incorporate many vegetation-climate feedbacks and outputs, as seen in Lyon et al. [27].

Each of these simulations were run from spun-up pre-industrial initial conditions at 1850 CE out to 2500 CE, with greenhouse gas concentrations fixed at 2300 CE levels from the extended RCP scenarios [28]. These simulations include the atmospheric greenhouse gas forcing of future climate change from carbon dioxide, methane and nitrogen dioxide, but not including other human-made pollutants, such as aerosols. Potential changes in ice sheets and sea-level rise, which could become important on these longer, multi-century timescales, are also not modelled. HadCM3 produces climate fields on a regular longitude-latitude grid of 3.75° × 2.5° (approx. 1.16 × 106 km2 at the equator). The monthly mean temperature, precipitation and cloud amount fields were averaged over each 20 year period of the simulation from the most recent two decade period, 2000–2020 CE. These means were then downscaled with a standard anomaly method applied, from pre-industrial (1880–1899 CE mean) to observed reference fields [38] to produce suitable inputs for the BIOME4 vegetation model with a spatial resolution of 0.5° in latitude and longitude (approx. 3100 km2 at the equator). Downscaling to this extent provides a higher spatial resolution of climate variation, but without introducing unrealistic interpolation artefacts [30,39]. Higher-resolution climate datasets are beginning to be available (e.g. [40]), but these have yet to be fully tested with BIOME4.

(b) Biome model

We converted the climate parameters from HadCM3 into projected climatic forcing on vegetation using BIOME4, a global vegetation model [30]. BIOME4 is a coupled carbon and water balance model that predicts the steady state, global vegetation distribution. It requires information on temperature, sunshine, precipitation, soil parameters (which are kept constant, as we do not model how they might change) and atmospheric CO2 concentration, for which we take the value at the midpoint of each 20 year period. The model maximizes the net primary productivity for each of 12 possible PFTs and then, based on a ranked combination of these PFTs, assigns each gridpoint to one of 26 biomes (table 1; [30]). The biomes inferred were also converted into nine ‘megabiomes' using the scheme of Harrison & Prentice [39]; this reduction in biome resolution was used to quantify more extreme projected floral transitions. Table 1. The biomes [30], and the megabiomes to which they belong [39], as used in this study.

biome name	abbreviation	megabiome name	
tropical evergreen broadleaf forest	TrEB	tropical forest	
tropical semi-evergreen broadleaf forest	TrSB	tropical forest	
tropical deciduous broadleaf forest and woodland	TrDB	tropical forest	
temperate deciduous broadleaf forest	TeDBF	temperate forest	
temperate evergreen needleleaf forest	TeENF	temperate forest	
cool mixed forest	CoM	temperate forest	
cool evergreen needleleaf forest	CoEN	temperate forest	
cool-temperate evergreen needleleaf and mixed forest	CoTeEN	temperate forest	
warm-temperate evergreen broadleaf and mixed forest	WTeEB	warm-temperate forest	
cold evergreen needleleaf forest	CdEN	boreal forest	
cold deciduous forest	CdD	boreal forest	
tropical savannah	TrS	savannah and dry woodland	
temperate sclerophyll woodland and shrubland	TeS	savannah and dry woodland	
temperate deciduous broadleaf savannah	TeDBS	savannah and dry woodland	
temperate evergreen needleleaf open woodland	TeENW	savannah and dry woodland	
tropical xerophytic shrubland	TrX	grassland and dry shrubland	
temperate xerophytic shrubland	TeX	grassland and dry shrubland	
cold parkland	CdP	grassland and dry shrubland	
tropical grassland	TrG	grassland and dry shrubland	
temperate grassland	TeG	grassland and dry shrubland	
desert	De	desert	
graminoid and forb tundra	GFTu	dry tundra	
low and high shrub tundra	ShTu	tundra	
erect dwarf-shrub tundra	EDTu	tundra	
prostrate dwarf-shrub tundra	PDTu	tundra	
cushion-forb tundra	CFTu	tundra	

HadCM3 and BIOME4 are commonly used together for modelling climate and biomes over longer-term scenarios, both in deep time and into the future (e.g. [41–44]). However, BIOME4 models biomes assuming that both climate and vegetation have reached an equilibrium state. Our input data represents the average simulated climate during each 20 year interval, and therefore does not capture any changes in climate within this time (although general trends over the total 500 years should be adequately captured). In addition, 20 years is probably insufficient time for vegetation to reach equilibrium in response to the modelled climate. As a result, our analyses describe climatic forcing on vegetation, but not necessarily realizable biome states, as transient ecosystem dynamics may be inadequately modelled.

(c) Human footprint

All subsequent data cleaning, manipulation and analysis was carried out in R, version 4.2.2 [45] using the geosphere [46], landscapemetrics [47], raster [48] and tidyverse [49] packages.

In addition to our standard biome projections, we also conducted sensitivity analyses examining how interaction with human activity could impact biome realization. To do this, we created datasets in which we removed grid cells associated with human land use from our biome projections, under the assumption that human occupation could inhibit the migration of biomes into these cells. Human impact was assessed using the HYDE 3.2 anthromes dataset for the year 2017 [31]; in the absence of future projections, we assumed that these anthromes remained the same from the present day through to 2500. Two tiers of human influence were used: in the first, more relaxed tier, only ‘urban' and ‘dense settlement' environments were removed from the projections, while in the more stringent second tier, all environments except those classified as ‘wild' or ‘semi-natural' were removed.

The HYDE dataset also uses a latitude-longitude grid, but to a much higher resolution than our biome models (1/12° latitude and longitude). In order to compare the two datasets, the biome models were disaggregated to create smaller grid cells of the same size as those in the HYDE dataset. Cells with a human presence in the HYDE dataset could then be removed from the projected biome areas in each 20 year time slice.

(d) Quantifying potential biome change

The nature of projected vegetation changes over time, between the three emissions scenarios, and using the different severities of human footprint, were quantified in five ways. Collectively, these statistics provide a summary of the distribution and coverage of area suited to each biome over time, until 2500. The area covered by each grid cell was calculated using the equation of Santini et al. [50], in order to convert biome projections for each grid cell into total terrestrial area.

Firstly, the biome projected to occupy each grid cell was compared between adjacent 20 year time slices, and with the present day (the 2000–2020 CE time slice). This was conducted both at a global scale and within low (30° N–30° S), mid (30°–60° N and S) and high (60°–90° N and S) latitude bands, to quantify the rate of change of climatic forcing over time, and the cumulative areal extent of climatic forcing towards a different biome compared to the present, respectively.

Secondly, the total area climatically suitable for each biome was calculated and compared between time slices. This indicates which biomes may gain total area over time under each scenario, and which biomes are more likely to lose area.

Thirdly, the proportion of area projected to be suitable for each biome within the same grid cells as in the previous time slice was calculated. One minus this value indicates the proportion of area which would require the colonization of a new region in order to be realized. As an extension of this, cells not adjacent to a cell of their projected biome in the previous time slice were also counted. These grid cells would require long-distance migration, or artificial transplantation, to successfully develop into their projected biome in that time slice.

Fourthly, the latitude–longitude coordinates of the centroid of each biome's projected area was found for each time slice. This was then used to determine the distance and bearing moved by each biome's centroid over time. This approach aimed to quantify the overall projected rate and direction of movement of the area suitable for each biome over time.

Finally, the number of patches (spatially segregated areas) suited to each biome was quantified using landscapemetrics::get_patches(), to determine the spatial cohesiveness of area suitable for each biome over time. This provides an indication of how easily migration, and genetic mixing, might occur between projected fragments of a single biome in each time slice.

3. Results

BIOME4 was used to infer climatic forcing towards different biomes across 1.31 × 108 km2 of terrestrial area (all area which is ice-free and not described as ‘barren'; figure 1). Applying the HYDE dataset to this map increased the resolution of the coastlines, resulting in a reduced 1.25 × 108 km2 of terrestrial area. Removing the area associated with more relaxed (urban) and more stringent (non-wild) anthromes left 1.24 × 108 km2 and 6.18 × 107 km2 of terrestrial area, respectively. The biome summaries given here were calculated as a proportion of these areal values. Figure 1. Maps of the area climatically suited to different megabiomes simulated with BIOME4, using the megabiome scheme of table 1. Present-day megabiomes are inferred using the mean climate of 2000–2019 CE (a). The other panels show the projected megabiomes in areas where a different megabiome is climatically suitable in 2480–2499 CE compared to the first time slice, under RCP2.6 (b), RCP4.5 (c) and RCP6 (d).

Even under RCP4.5, which is often considered a ‘business-as-usual' scenario, around 35% of terrestrial area is expected to be climatically suitable for a biome different to its present state, and 25% to a different megabiome, by 2150, with change continuing at a lower rate after this time (figure 2a; electronic supplementary material, S1a). This is matched by a high, but falling, rate of change between 20 year time slices from the present to 2150: 12% of land area is expected to experience climatic forcing towards a different biome per 20 years (and 7% to a different megabiome) by 2040, but by 2160, this falls to around 5%, which then remains consistent until 2500 (figure 3a; electronic supplementary material, S2a). The extent of human footprint applied does not substantially impact this result (figure 2a), although a slightly smaller proportion of wild and semi-natural regions is expected to experience climatic forcing towards a different biome over each 20 year period compared to all terrestrial environments (figure 3a). The mid- to high-latitudes are expected to experience more change, proportional to their area, than the low-latitudes (electronic supplementary material, figures S3a and S4a). Figure 2. (a) For each RCP and human footprint, the amount of terrestrial area which is expected to change climatically suitable megabiome compared to the first time slice (2000–2020) over time. The area is quantified as all terrestrial land, all land except for urban conurbations, and only wild or semi-natural land. (b) For each megabiome and time slice, the proportion of climatically suitable terrestrial area which overlaps with that of the same megabiome in the first time slice (2000–2020).

Figure 3. (a) For each RCP and human footprint, the amount of terrestrial area which is expected to change climatically suitable megabiome compared to the previous 20 year time slice. The area is quantified as all terrestrial land, all land except for urban conurbations, and only wild or semi-natural land. (b) For each megabiome and time slice, the proportion of climatically suitable terrestrial area which overlaps with that of the same megabiome in the previous time slice.

Under RCP6, similar but more extreme trends are seen to those of RCP4.5, with up to 50% of terrestrial area expected to experience climatic forcing towards a different biome than its present state, and 40% to a different megabiome, with this plateau only being reached between 2400 and 2500 (figure 2a; electronic supplementary material, S1a). By contrast, under RCP2.6, nearly 20% of land area is expected to be climatically suitable to a different biome, and 13% to a different megabiome, by 2080. However, some of this area will later revert, stabilizing at about 12% of terrestrial area being climatically forced away from its current biome, and 9% from its current megabiome, between 2250 and 2500 (figure 2a; electronic supplementary material, S1a). The rate at which area is climatically forced towards different biomes is also expected to fall faster under RCP2.6, reaching a plateau around 2100 (figure 3a; electronic supplementary material, S2a), but this plateau is at a higher rate than the other two emissions scenarios; while 5% of land area is expected to experience climatic forcing towards a different biome per 20 years by 2500 under RCP4.5 and RCP6, this is expected to be 6–7% under RCP2.6 (figure 3a; electronic supplementary material, S2a).

Unsurprisingly, it is biomes found in cold environments which are expected to lose the most total area of appropriate climate forcing. Area suitable for boreal forest biomes, both cold evergreen needleleaf forest and cold deciduous forest, is expected to decline under RCP4.5 and RCP6 (electronic supplementary material, figures S6 and S7). This is also true of cool mixed forest, although the area suited to temperate forest overall is predicted to increase. Tundra biomes are particularly hard hit, with prostrate dwarf-shrub tundra expected to have no climatically suitable area by 2160, and cushion-forb tundra predicted to completely lose climatically suitable area by 2300, under RCP6 (electronic supplementary material, figures S6 and S7). By contrast, grassland and dry shrubland biomes, including tropical and temperate xerophytic shrubland and tropical and temperate grassland, as well as desert, are expected to gain climatically suitable area over the next 500 years.

The proportion of overlap of area climatically suited to each individual biome, both with its present-day distribution and its distribution in the previous time slice, varies between biomes and over time (figures 2b and 3b; electronic supplementary material, S1b and S2b). For most biomes and megabiomes, the proportion of overlap is reduced at higher RCPs, implying a larger extent of loss or movement of suitable environmental conditions for that biome. For example, large swathes of area suitable for cool evergreen needleleaf forest are expected to shift even under RCP2.6, with only around 50% of its projected distribution in 500 years overlapping with its present distribution (electronic supplementary material, figure S1b). At higher RCPs, wholly different areas are expected to be climatically suitable for this biome in as little as 100 years. By contrast, the area suited to desert is projected to mostly maintain its present distribution throughout the next 500 years, with little change expected between 20 year time slices (figures 2b and 3b). Some biomes are projected to have abundant climatically suitable area that is not even adjacent to that in the previous 20 year interval (electronic supplementary material, figure S5). For example, under RCP4.5 and 6, up to 50% of the projected area suited to cool-temperate evergreen needleleaf and mixed forest in each time slice may not be adjacent to that in the previous 20 years.

Under RCP4.5, the latitudinal ranges of area suited to each biome in each hemisphere are expected to remain largely consistent over the next 500 years (electronic supplementary material, figure S8a). For most biomes, this includes no clear poleward shift in distributions, also confirmed by the bearing of movement of centroids of the area suited to each biome over time, which generally show no clear pattern (electronic supplementary material, tables S1-S3). There is also no clear trend in how connected the area suited to each biome is expected to be, either through time or across RCPs (electronic supplementary material, figure S9). However, many biomes do show evidence of strong fluctuation in the latitudinal ranges of their suitable area between adjacent time slices, such as temperate sclerophyll woodland and shrubland and temperate grassland (electronic supplementary material, figure S8a). For several biomes, including cold evergreen needleleaf forest, cold deciduous forest and low and high shrub tundra, the latitudinal ranges of their suitable area in the Southern Hemisphere are expected to strongly reduce.

4. Discussion

Our climate models indicate that under a ‘business-as-usual' scenario, climate change in response to anthropogenic emissions is expected to continue well beyond 2100 [27]. Here, we demonstrate that this may also result in substantial changes in the amount and distribution of area suited to different biomes, continuing long beyond the end of this century. Our RCP4.5 models show that up to 40% of terrestrial area will no longer be climatically suitable for its present biome by 2500 (figure 2a), with many cold-adapted biomes, particularly boreal forest and dry tundra, losing a large proportion of their suitable area (figure 2b). A higher proportion of area in the mid- to high-latitudes is expected to experience change than the low-latitudes (electronic supplementary material, figures S3a and S4a), particularly in the Southern Hemisphere (electronic supplementary material, figure S8a).

Our results show broad agreement with previous, similar analyses of changing biomes conducted on shorter timescales. Huntley et al. [5] predicted that 52% of terrestrial land will have changed biome under RCP4.5 by 2100, slightly higher than our estimate of 35%. Across the three RCPs, we show that between 12% and 50% of land area will be suited to a different biome in the year 2500 compared to the present-day; this is comparable to the range estimated by Gonzalez et al. [7], who used a variety of climate models and DGVMs to infer that between one-tenth and one-half of global land area may currently be highly vulnerable to biome change.

Even under RCP2.6, we expect 6–7% of land area to become climatically suited to a different biome every 20 years from 2100 onwards (figure 3a). This is owing to fluctuations in the area suitable to some biomes, which can also be seen in their expected latitudinal distributions (figure 3b; electronic supplementary material, S8a). Such fluctuation has previously been recognized in studies of present biome distributions: for example, Friedl et al. [51] estimated that around 10% of terrestrial area currently undergoes biome transition per year. Our RCP4.5 and RCP6 projections reach a slightly lower (5%) plateau in the rate of becoming suitable to a different biome after 2200 (figure 3a), however in these cases, it is associated with continued divergence from the biome distribution of the present-day (figure 2a). This suggests that it will take longer to reach a constant rate of change under these scenarios, and that once reached, they may have a lower rate of natural fluctuation compared to the RCP 2.6 scenario, and the recent past.

In most of our analyses, the impact of removing the human footprint on the proportional loss of area suited to each biome was minimal, particularly in comparison with the difference between emissions scenarios (e.g. figure 2a). This was also true for Higgins et al. [2]. However, wild and semi-natural regions were projected to experience a slightly lower rate of change over time than all terrestrial environments collectively (figure 3a). This difference was most pronounced at low-latitudes (electronic supplementary material, figures S3b and S4b). This finding indicates that proportionally more change might be expected to take place within regions which are currently within the human footprint, particularly agricultural land. However, we modelled anthromes as being static over the next 500 years, and human impacts on the environment have increased in area and severity over the last 30 years [52], and are expected to continue to do so. As such, it is likely that land use change will become an increasingly important factor in limiting biome migration in future.

Important questions arise as to whether the biome changes projected here are ecologically possible. We describe patches of terrestrial area which would be climatically amenable to a particular biome, but there are many barriers to these patches being realized, including limited dispersal, geographical and anthropogenic barriers, and competitive exclusion [7]. Dispersal rates, in particular, are a major concern given that we expect many patches to arise a substantial distance from the area previously suited to that biome (electronic supplementary material, figure S5; [53]). Tree migration rates are estimated to be only 100 m to 10 km per year [54,55], and many boreal and temperate tree species may only be able to migrate to between a quarter and half of their newly arising suitable habitat over the next century [56–58]. However, addressing questions about migration feasibility in our models is difficult, as dispersal ability is highly heterogeneous between different species within a biome, compounded by local variability in climate change and terrain [59]. Regardless, it is clear that for some biomes, we infer that suitable climatic conditions will arise in regions too far from their prior distributions to reach by migration. Human intervention serves as a potential solution: relocating species to more amenable habitat as a means of conserving them, or to provide benefits to humans, is a key action proposed for climate adaptation and practised today [60,61]. However, this must be carefully planned, particularly given the uncertainties and unintended consequences of introducing non-historically native species into new ecosystems [61,62].

Even if dispersal to a newly arising area of suitable climatic conditions is possible, there are questions about how the process of transition takes place, and the time needed for a new biome to establish (e.g. [63]). Functional convergence of the plants within a patch towards a particular biome is typically assumed to be owing to environmental filtering of those poorly adapted to those conditions, and selection of those which are well adapted [3]; these are processes which probably require more time than the 20 year intervals we model here. In some cases, it is possible that the biome changes we predict would not take place. Biome delineation is both less exact and more complex than our models suggest, with some biomes occupying a range of environmental conditions, and groupings of environmental conditions being associated with multiple biomes [2,3]. It is also not well understood how historical contingency can play a role in determining biomes [3,11,64]. Further, disturbances such as herbivorous grazing and fires can 'consume' existing vegetation, accelerating the realization of a new biome [3], while in some cases, anthropogenic changes in fire regimes, whether via hunting-gathering or agriculture, have been found to increase or maintain open woodland and savannah biomes [65]. Gaining more knowledge on how biome transitions take place, and the timescales on which they occur, is an important next step in validating our model projections.

Large areas of land transitioning biomes over the next 500 years would have wide-reaching impacts on both biodiversity and society. The projected reduction in suitable area of some biomes, if realized, would lead to extinction for many plants adapted to those environments (e.g. [66,67]). As climatic conditions become unsuitable for the survival of current vegetation, a reduction of plant biomass and terrestrial carbon storage is expected, either gradually or through sudden, destructive events such as wildfires [5,7,19,20]. Biome stability is also strongly correlated with regional biodiversity, indicating that more recently perturbed regions are less diverse [5,68]. Faunal biodiversity will face the double threat of changing climate and the loss or migration of vegetation on which it relies, and will need to either migrate in concert or adapt to survive [69].

As biomes change, the stocks and flows of ecosystem services they provide will come into question. These include services fundamental to human survival and habitability, such as food and water [70]. They also include protecting human populations against the increasing frequency and severity of various weather events; for example, tropical mangrove forests reduce the severity of storm surges [71]. Such events might otherwise require adaptations that may face significant social, technological, and physiological challenges or limits, or could even force human relocations [72,73]. However, as some regions become less habitable, others may open [73], and society must develop plausible, desirable and equitable strategies to share landscapes, access ecosystem services, and avoid conflict while learning to live with these changes [74,75].

Our results agree with previous studies that boreal and tundra biomes, which are mostly found at mid- to high-latitudes, are likely to lose a considerable proportion of their climatically suitable area over the next few centuries (figures 2b and 3b) [2,7,19]. The boreal biome provides a number of ecosystem services, including timber forest products, amenities and carbon sequestration, which stand to become depleted as the biome wanes. It is also highly prone to disturbances from fire and pests [76,77], including recent record wildfires that are projected to intensify [77–79] as higher latitudes warm much faster than the global average, threatening regions of the biome that remain. Increased efforts should therefore be made to protect boreal and tundra biomes specifically, but this will need to take into account the geographical shifts these biomes may make over the next 500 years.

This study presents, to our knowledge, the first simulations of potential climate-driven biome changes to 2500 CE and, as such, has used available climate model simulations and an equilibrium biome model often associated with this general circulation model. Future climate and biodiversity are highly uncertain, and the testing of additional models, both for climate and vegetation, would be beneficial, ideally at a higher spatial and temporal resolution [7,57,80]. Our climate modelling approach used TRIFFID, which allows interaction between the vegetation and local climate to be taken into account [3], but mismatches between the vegetation inferred by TRIFFID and the biome projections produced using BIOME4 are possible (e.g. [11]). The use of a DVGM to infer biomes, rather than an equilibrium model, would also allow migration to be simulated within the model (e.g. [81]). Terrestrial ice area was fixed, so the greening of currently frozen land could not be taken into account, and corresponding sea-level rise was not modelled. Our climate models use cells delineated by degrees, which vary considerably in their surface area, and this means the spatial resolution of our results are variable with latitude. Our use of a static human footprint for the year 2017 [31] is also suboptimal, as the space occupied by anthropogenic activities is likely to increase in future [52]. However, the exact geography of this expansion is difficult to predict. While we consider anthropogenic greenhouse gas emissions to represent the biggest source of uncertainty in our results, investigation of these other factors may reveal more insights to future biome change which could prove useful in planning for mitigation and adaptation.

5. Conclusion

By modelling how climatic forcing towards different biomes may manifest in future, our research contributes to a new but growing body of work examining climate change impacts after 2100. We show that across climate scenarios, significant changes in the global distribution and areal extent of land amenable to different biomes will probably occur, particularly over the next two centuries. Cold-adapted biomes are expected to lose the greatest amount of climatically suitable area, particularly impacting mid- to high-latitude regions. This will probably have strong implications for biodiversity, and the ecosystem services that humans rely on. Our results highlight the necessity of better understanding climate and biome change past 2100, in order to mitigate and adapt to the many challenges these changes will present for ecosystems and humanity.

Acknowledgements

The climate modelling was undertaken on ARC3, part of the high performance computing facilities at the University of Leeds, UK. Post-processing analyses were undertaken on Euler, ETH Zurich's high performance computer, Switzerland. We thank Erin Saupe, Alex Dunhill, Andrew Beckerman and members of the Leverhulme Centre for Anthropocene Biodiversity, University of York, and Computational Evolution Group, ETH Zurich, for discussion. This manuscript was also greatly improved by feedback from three anonymous reviewers and the editors of this theme issue.

Ethics

This work did not require ethical approval from a human subject or animal welfare committee.

Data accessibility

The data and code are provided in the Zenodo repository: https://doi.org/10.5281/zenodo.10820536 [82], and the supplementary figures in the electronic supplementary material [83].

Declaration of AI use

We have not used AI-assisted technologies in creating this article.

Authors' contributions

B.J.A.: conceptualization, formal analysis, investigation, methodology, project administration, visualization, writing—original draft, writing—review and editing; D.J.H.: conceptualization, formal analysis, investigation, methodology, visualization, writing—original draft, writing—review and editing; A.M.B.: supervision, writing—review and editing; M.C.: supervision, writing—review and editing; R.M.: supervision, writing—review and editing; L.C.S.: supervision, writing—review and editing; D.R.W.: supervision, writing—review and editing; C.L.: conceptualization, funding acquisition, project administration, writing—review and editing.

All authors gave final approval for publication and agreed to be held accountable for the work performed therein.

Conflict of interest declaration

We have no competing interests.

Funding

B.J.A.'s contributions were funded by an ETH + grant (BECCY). C.L. is grateful for funding from Leverhulme Trust (grant no. RC-2018-021), through the Leverhulme Centre for Anthropocene Biodiversity, University of York. This research is an output of the Refugia of Futures Past project supported by the White Rose Collaboration Fund.
==== Refs
References

1. Schimper AFW. 1903 Plant-geography upon a physiological basis. Oxford, UK: Clarendon Press.
2. Higgins SI, Buitenwerf R, Moncrieff GR. 2016 Defining functional biomes and monitoring their change globally. Glob. Change Biol. 22 , 3583-3593. (10.1111/gcb.13367)
3. Moncrieff GR, Bond WJ, Higgins SI. 2016 Revising the biome concept for understanding and predicting global change impacts. J. Biogeogr. 43 , 863-873. (10.1111/jbi.12701)
4. Hunter J, Franklin S, Luxton S, Loidi J. 2021 Terrestrial biomes: a conceptual review. Vegetation Classif. Surv. 2 , 73-85. (10.3897/VCS/2021/61463)
5. Huntley B, Allen JRM, Forrest M, Hickler T, Ohlemüller R, Singarayer JS, Valdes PJ. 2021 Projected climatic changes lead to biome changes in areas of previously constant biome. J. Biogeogr. 48 , 2418-2428. (10.1111/jbi.14213)
6. Walter H. 1973 Vegetation of the earth in relation to climate and the eco-physiological conditions. New York, NY: Springer.
7. Gonzalez P, Neilson RP, Lenihan JM, Drapek RJ. 2010 Global patterns in the vulnerability of ecosystems to vegetation shifts due to climate change. Global Ecol. Biogeogr. 19 , 755-768. (10.1111/j.1466-8238.2010.00558.x)
8. Bond WJ. 2005 Large parts of the world are brown or black: a different view on the ‘Green World’ hypothesis. J. Vegetation Sci. 16 , 261-266. (10.1111/j.1654-1103.2005.tb02364.x)
9. Crisp MD et al. 2009 Phylogenetic biome conservatism on a global scale. Nature 458 , 754-756. (10.1038/nature07764)19219025
10. Chaboureau A-C, Sepulchre P, Donnadieu Y, Franc A. 2014 Tectonic-driven climate change and the diversification of angiosperms. Proc. Natl Acad. Sci. USA 111 , 14 066-14 070. (10.1073/pnas.1324002111)
11. Zhu D, Ciais P, Chang J, Krinner G, Peng S, Viovy N, Peñuelas J, Zimov S. 2018 The large mean body size of mammalian herbivores explains the productivity paradox during the Last Glacial Maximum. Nat. Ecol. Evol. 2 , 640-649. (10.1038/s41559-018-0481-y)29483680
12. Ellis EC et al. 2021 People have shaped most of terrestrial nature for at least 12,000 years. Proc. Natl Acad. Sci. USA 118 , e2023483118. (10.1073/pnas.2023483118)33875599
13. Ellis EC, Klein Goldewijk K, Siebert S, Lightman D, Ramankutty N. 2010 Anthropogenic transformation of the biomes, 1700 to 2000. Global Ecol. Biogeogr. 19 , 589-606. (10.1111/j.1466-8238.2010.00540.x)
14. Pereira HM et al. 2010 Scenarios for global biodiversity in the 21st Century. Science 330 , 1496-1501. (10.1126/science.1196624)20978282
15. Newbold T et al. 2015 Global effects of land use on local terrestrial biodiversity. Nature 520 , 45-50. (10.1038/nature14324)25832402
16. Scheffers BR, Pecl G. 2019 Persecuting, protecting or ignoring biodiversity under climate change. Nat. Clim. Change 9 , 581-586. (10.1038/s41558-019-0526-5)
17. Willcock S, Cooper GS, Addy J, Dearing JA. 2023 Earlier collapse of Anthropocene ecosystems driven by multiple faster and noisier drivers. Nat. Sustain. 6 , 1331-1342. (10.1038/s41893-023-01157-x)
18. Friedlingstein P et al. 2006 Climate-carbon cycle feedback analysis: results from the C4MIP model intercomparison. J. Clim. 19 , 3337-3353. (10.1175/JCLI3800.1)
19. Sitch S et al. 2008 Evaluation of the terrestrial carbon cycle, future plant geography and climate-carbon cycle feedbacks using five Dynamic Global Vegetation Models (DGVMs). Glob. Change Biol. 14 , 2015-2039. (10.1111/j.1365-2486.2008.01626.x)
20. Jones C, Lowe J, Liddicoat S, Betts R. 2009 Committed terrestrial ecosystem changes due to climate change. Nat. Geosci. 2 , 484-487. (10.1038/ngeo555)
21. Hannah L, Midgley GF, Millar D. 2002 Climate change-integrated conservation strategies. Global Ecol. Biogeogr. 11 , 485-495. (10.1046/j.1466-822X.2002.00306.x)
22. Brooks TM, Mittermeier RA, Da Fonseca GAB, Gerlach J, Hoffmann M, Lamoreux JF, Mittermeier CG, Pilgrim JD, Rodrigues ASL. 2006 Global biodiversity conservation priorities. Science 313 , 58-61. (10.1126/science.1127609)16825561
23. Moore JW, Schindler DE. 2022 Getting ahead of climate change for ecological adaptation and resilience. Science 376 , 1421-1426. (10.1126/science.abo3608)35737793
24. Zani PA. 2023 What does the scientific community owe future generations of biologists? Biology in the era following centuries-long anthropogenic change. BioScience 73 , 331-332. (10.1093/biosci/biad023)
25. Schröter D et al. 2005 Ecosystem service supply and vulnerability to global change in Europe. Science 310 , 1333-1337. (10.1126/science.1115233)16254151
26. Folke C et al. 2021 Our future in the Anthropocene biosphere. Ambio 50 , 834-869. (10.1007/s13280-021-01544-8)33715097
27. Lyon C et al. 2022 Climate change research and action must look beyond 2100. Glob. Change Biol. 28 , 349-361. (10.1111/gcb.15871)
28. Meinshausen M et al. 2011 The RCP greenhouse gas concentrations and their extensions from 1765 to 2300. Clim. Change 109 , 213. (10.1007/s10584-011-0156-z)
29. Mengel M, Nauels A, Rogelj J, Schleussner C-F. 2018 Committed sea-level rise under the Paris Agreement and the legacy of delayed mitigation action. Nat. Commun. 9 , 601. (10.1038/s41467-018-02985-8)29463787
30. Kaplan JO et al. 2003 Climate change and Arctic ecosystems: 2. Modeling, paleodata-model comparisons, and future projections. J. Geophys. Res. 108 (D19 ), 8171. (10.1029/2002JD002559)
31. Klein Goldewijk K, Beusen A, Doelmann J, Stehfest E. 2017 Anthropogenic land use estimates for the Holocene - HYDE 3.2. Earth Syst. Sci. Data 9 , 927-953. (10.5194/essd-9-927-2017)
32. Riahi K et al. 2017 The shared socioeconomic pathways and their energy, land use, and greenhouse gas emissions implications: an overview. Global Environ. Change 42 , 153-168. (10.1016/j.gloenvcha.2016.05.009)
33. Hausfather Z, Peters GP. 2020 RCP8.5 is a problematic scenario for near-term emissions. Proc. Natl Acad. Sci. USA 117 , 27 791-27 792. (10.1073/pnas.20171241117)
34. Ritchie J, Dowlatabadi H. 2017 The 1000 GtC coal question: are cases of vastly expanded future coal combustion still plausible? Energy Econ. 65 , 16-31. (10.1016/j.eneco.2017.04.015)
35. Gordon C, Cooper C, Senior CA, Banks H, Gregory JM, Johns TC, Mitchell JFB, Wood RA. 2000 The simulation of SST, sea ice extents and ocean heat transports in a version of the Hadley Centre coupled model without flux adjustments. Clim. Dyn. 16 , 147-168. (10.1007/s003820050010)
36. Valdes PJ et al. 2017 The BRIDGE HadCM3 family of climate models: HadCM3@Bristol v1.0. Geosci. Model Dev. 10 , 3715-3743. (10.5194/gmd-10-3715-2017)
37. Cox P. 2001 Description on the TRIFFID dynamic global vegetation model, Technical Report 24. Exeter, UK: Met Office Hadley Centre.
38. Haxeltine A, Prentice IC. 1996 BIOME3: an equilibrium terrestrial biosphere model based on ecophysiological constraints, resource availability, and competition among plant functional types. Global Biogeochem. Cycles 10 , 693-709. (10.1029/96GB02344)
39. Harrison SP, Prentice IC. 2003 Climate and CO2 controls on global vegetation distribution at the last glacial maximum: analysis based on palaeovegetation data, biome modelling and palaeoclimate simulations. Glob. Change Biol. 9 , 983-1004. (10.1046/j.1365-2486.2003.00640.x)
40. Hersbach H et al. 2020 The ERA5 global reanalysis. Q. J. R. Meteorol. Soc. 146 , 1999-2049.
41. Salzmann U, Haywood AM, Lunt DJ. 2008 The past is a guide to the future? Comparing Middle Pliocene vegetation with predicted biome distributions for the twenty-first century . Phil. Trans. R. Soc. A 367 , 189-204. (10.1098/rsta.2008.0200)
42. Hoogakker BAA et al. 2016 Terrestrial biosphere changes over the last 120 kyr. Clim. Past 12 , 51-73. (10.5194/cp-12-51-2016)
43. Prescott CL, Dolan AM, Haywood AM, Hunter SJ, Tindall JC. 2018 Regional climate and vegetation response to orbital forcing within the mid-Pliocene Warm Period: a study using HadCM3. Global Planet. Change 161 , 231-243. (10.1016/j.gloplacha.2017.12.015)
44. Krapp M, Beyer RM, Edmundson SL, Valdes PJ, Manica A. 2021 A statistics-based reconstruction of high-resolution global terrestrial climate for the last 800,000 years. Sci. Data 8 , 228. (10.1038/s41597-021-01009-3)34453060
45. R Core Team. 2022 R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing . Seehttps://www.R-project.org/.
46. Hijmans RJ, Karney C, Williams E, Vennes C. 2022 geosphere: spherical trigonometry. Version 1,18. See https://cran.r-project.org/web/packages/geosphere/index.html.
47. Hesselbarth MHK, Sciaini M, With KA, Wiegand K, Nowosad J. 2019 Landscapemetrics: an open-source R tool to calculate landscape metrics. Ecography 42 , 1648-1657. (10.1111/ecog.04617)
48. Hijmans RJ et al. 2023 raster: geographic data analysis and modeling. Version 3,20. See https://cran.r-project.org/web/packages/raster/index.html.
49. Wickham H et al. 2019 Welcome to the Tidyverse. J. Open Source Software 4 , 1686. (10.21105/joss.01686)
50. Santini M, Taramelli A. 2010 ASPHAA: a GIS-based algorithm to calculate cell area of a latitude-longitude (geographic) regular grid. Trans. GIS 14 , 351-377. (10.1111/j.1467-9671.2010.01200.x)
51. Friedl MA, Sulla-Menashe D, Tan B, Schneider A, Ramankutty N, Sibley A, Huang X. 2010 MODIS Collection 5 global land cover: algorithm refinements and characterization of new datasets. Remote Sens. Environ. 114 , 168-182. (10.1016/j.rse.2009.08.016)
52. Williams BA et al. 2020 Change in terrestrial human footprint drives continued loss of intact ecosystems. One Earth 3 , 371-382. (10.1016/j.oneear.2020.08.009)
53. Fricke EC, Ordonez A, Rogers HS, Svenning J-C. 2022 The effects of defaunation on plants' capacity to track climate change. Science 375 , 210-214. (10.1126/science.abk3510)35025640
54. Clark JS, Lewis M, Horvath L. 2001 Invasion by extremes: population spread with variation in dispersal and reproduction. Am. Nat. 157 , 537-554. (10.1086/319934)18707261
55. Svenning J-C, Skov F. 2007 Could the tree diversity pattern in Europe be generated by postglacial dispersal limitation? Ecol. Lett. 10 , 453-460. (10.1111/j.1461-0248.2007.01038.x)17498144
56. Epstein HE, Kaplan JO, Lischke H, Yu Q. 2007 Simulating future changes in Arctic and Subarctic vegetation. Comput. Sci. Eng. 9 , 12-23. (10.1109/MCSE.2007.84)
57. Morin X, Thuiller W. 2009 Comparing niche- and process-based models to reduce prediction uncertainty in species range shifts under climate change. Ecology 90 , 1301-1313. (10.1890/08-0134.1)19537550
58. Lischke H. 2020 Simulation der Baumartenmigration im Klimawandel. Schweizerische Zeitschrift für Forstwesen 171 , 151-157. (10.3188/szf.2020.0151)
59. Platts PJ, Gereau RE, Burgess ND, Marchant R. 2013 Spatial heterogeneity of climate change in an Afromontane centre of endemism. Ecography 36 , 518-530. (10.1111/j.1600-0587.2012.07805.x)
60. Hällfors MH, Vaara EM, Hyvärinen M, Oksanen M, Schulman LE, Siipi H, Lehvävirta S. 2014 Coming to terms with the concept of moving species threatened by climate change - a systematic review of the terminology and definitions. PLoS ONE 9 , e102979. (10.1371/journal.pone.0102979)25055023
61. Butt N, Chauvenet ALM, Adams VM, Beger M, Gallagher RV, Shanahan DF, Ward M, Watson JEM, Possingham HP. 2020 Importance of species translocations under rapid climate change. Conserv. Biol. 35 , 775-783. (10.1111/cobi.13643)33047846
62. McLachlan JS, Hellmann JJ, Schwartz MW. 2006 A framework for debate of assisted migration in an era of climate change. Conserv. Biol. 21 , 297-302.
63. Blonder B et al. 2017 Predictability in community dynamics. Ecol. Lett. 20 , 293-306. (10.1111/ele.12736)28145038
64. Haywood AM et al. 2019 What can palaeoclimate modelling do for you? Earth Syst. Environ. 3 , 1-18. (10.1007/s41748-019-00093-1)
65. Phelps LN et al. 2020 Asymmetric response of forest and grassy biomes to climate variability across the African Humid Period: influenced by anthropogenic disturbance? Ecography 43, 1118–1142. (10.1111/ecog.04990)
66. Giam X, Bradshaw CJA, Tan HTW, Sodhi NS. 2010 Future habitat loss and the conservation of plant biodiversity. Biol. Conserv. 143 , 1594-1602. (10.1016/j.biocon.2010.04.019)
67. Corlett RT. 2016 Plant diversity in a changing world: status, trends, and conservation needs. Plant Divers. 38 , 10-16. (10.1016/j.pld.2016.01.001)30159445
68. Colville JF, Beale CM, Forest F, Altwegg R, Huntley B, Cowling RM. 2020 Plant richness, turnover, and evolutionary diversity track gradients of stability and ecological opportunity in a megadiversity center. Proc. Natl Acad. Sci. USA 117 , 20 027-20 037. (10.1073/pnas.1915646117)
69. Doherty S, Saltré F, Llewelyn J, Strona G, Williams SE, Bradshaw CJA. 2023 Estimating co-extinction threats in terrestrial ecosystems. Glob. Change Biol. 29 , 5122-5138. (10.1111/gcb.16836)
70. Scholes RJ. 2016 Climate change and ecosystem services. WIREs Clim. Change 7 , 537-550. (10.1002/wcc.404)
71. Zhang K, Liu H, Li Y, Xu H, Shen J, Rhome J, Smith III TJ. 2012 The role of mangroves in attenuating storm surges. Estuarine Coastal Shelf Sci. 102–103 , 11-23. (10.1016/j.ecss.2012.02.021)
72. Dow K, Berkhout F, Preston BL. 2013 Limits to adaptation to climate change: a risk approach. Curr. Opin. Environ. Sustain. 5 , 384-391. (10.1016/j.cosust.2013.07.005)
73. Lenton TM et al. 2023 Quantifying the human cost of global warming. Nat. Sustain. 6 , 1237-1247. (10.1038/s41893-023-01132-6)
74. Mach KJ et al. 2019 Climate as a risk factor for armed conflict. Nature 571 , 193-197. (10.1038/s41586-019-1300-6)31189956
75. Beardsworth R. 2020 Climate science, the politics of climate change and futures of IR. Int. Relations 34 , 374-390. (10.1177/0047117820946365)
76. Volney WJA, Fleming RA. 2000 Climate change and impacts of boreal forest insects. Agricul. Ecosyst. Environ. 82 , 283-294. (10.1016/S0167-8809(00)00232-2)
77. Wang X, Studens K, Parisien M-A, Taylor SW, Candau J-N, Boulanger Y, Flannigan MD. 2020 Projected changes in fire size from daily spread potential in Canada over the 21st century. Environ. Res. Lett. 15 , 104048. (10.1088/1748-9326/aba101)
78. Boucher D, Gauthier S, Thiffault N, Marchand W, Girardin M, Urli M. 2020 How climate change might affect tree regeneration following fire at northern latitudes: a review. New Forests 51 , 543-571. (10.1007/s11056-019-09745-6)
79. Quilcaille Y, Gudmundsson L, Seneviratne SI. 2023 Extending MESMER-X: a spatially resolved Earth system model emulator for fire weather and soil moisture, EGUsphere (preprint), 1–35. (10.5194/egusphere-2023-589)
80. Thuiller W, Guéguen M, Renaud J, Karger DN, Zimmermann NE. 2019 Uncertainty in ensembles of global biodiversity scenarios. Nat. Commun. 10 , 1446. (10.1038/s41467-019-09519-w)30926936
81. Lehsten V, Mischurow M, Lindström E, Lehsten D, Lischke H. 2019 LPJ-GM 1.0: simulating migration efficiently in a dynamic vegetation model. Geoscientific Model Dev. 12 , 893-908. (10.5194/gmd-12-893-2019)
82. Allen BJ, Hill DJ, Burke AM, Clark M, Marchant R, Stringer LC, Williams DR, Lyon C. 2024 Data from: Projected future climatic forcing on the global distribution of vegetation types. Zenodo. (10.5281/zenodo.10820536)
83. Allen BJ, Hill DJ, Burke AM, Clark M, Marchant R, Stringer LC, Williams DR, Lyon C. 2024 Projected future climatic forcing on the global distribution of vegetation types. Figshare. (10.6084/m9.figshare.c.7131291)
