Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Climate and air pollution impacts on habitat suitability of Austrian forest ecosystems

  • Thomas Dirnböck ,

    Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Writing – original draft, Writing – review & editing

    thomas.dirnboeck@umweltbundesamt.at

    Affiliation Department for Ecosystem Research and Environmental Information Management, Environment Agency Austria, Spittelauer Lände 5, Vienna, Austria

  • Ika Djukic,

    Roles Data curation

    Affiliation Department for Ecosystem Research and Environmental Information Management, Environment Agency Austria, Spittelauer Lände 5, Vienna, Austria

  • Barbara Kitzler,

    Roles Data curation

    Affiliation Austrian Research Centre for Forests—BFW, Vienna, Austria

  • Johannes Kobler,

    Roles Data curation, Methodology

    Affiliation Department for Ecosystem Research and Environmental Information Management, Environment Agency Austria, Spittelauer Lände 5, Vienna, Austria

  • Janet P. Mol-Dijkstra,

    Roles Software

    Affiliation Alterra, Wageningen UR, Wageningen, the Netherlands

  • Max Posch,

    Roles Conceptualization, Data curation, Methodology, Software

    Affiliation Coordination Centre for Effects (CCE), RIVM, Bilthoven, the Netherlands

  • Gert Jan Reinds,

    Roles Conceptualization, Data curation, Methodology, Software

    Affiliation Alterra, Wageningen UR, Wageningen, the Netherlands

  • Angela Schlutow,

    Roles Data curation, Software

    Affiliation OEKO-DATA—Ecosystem Analysis and Environmental Data Management, Strausberg, Germany

  • Franz Starlinger,

    Roles Data curation

    Affiliation Austrian Research Centre for Forests—BFW, Vienna, Austria

  • Wieger G. W. Wamelink

    Roles Conceptualization, Methodology, Software, Writing – original draft

    Affiliation Alterra, Wageningen UR, Wageningen, the Netherlands

Abstract

Climate change and excess deposition of airborne nitrogen (N) are among the main stressors to floristic biodiversity. One particular concern is the deterioration of valuable habitats such as those protected under the European Habitat Directive. In future, climate-driven shifts (and losses) in the species potential distribution, but also N driven nutrient enrichment may threaten these habitats. We applied a dynamic geochemical soil model (VSD+) together with a novel niche-based plant response model (PROPS) to 5 forest habitat types (18 forest sites) protected under the EU Directive in Austria. We assessed how future climate change and N deposition might affect habitat suitability, defined as the capacity of a site to host its typical plant species. Our evaluation indicates that climate change will be the main driver of a decrease in habitat suitability in the future in Austria. The expected climate change will increase the occurrence of thermophilic plant species while decreasing cold-tolerant species. In addition to these direct impacts, climate change scenarios caused an increase of the occurrence probability of oligotrophic species due to a higher N immobilisation in woody biomass leading to soil N depletion. As a consequence, climate change did offset eutrophication from N deposition, even when no further reduction in N emissions was assumed. Our results show that climate change may have positive side-effects in forest habitats when multiple drivers of change are considered.

Introduction

Climate change combined with excess deposition of nitrogen (N) are among the main stressors of biodiversity, at least in Europe, North America and parts of Asia [1]. Climate warming has caused phenological, physiological and genetic adaptations and changes of spatial distribution patterns of plant and animal species [2, 3]. Nutrient enrichment in response to N deposition has also caused changes in the structure of communities and declines in biodiversity [4]. According to large-scale modelling, major parts of the European Natura 2000 habitats, which are at the core of the European Habitat Directive [5], are under threat [6]. If both greenhouse gas emissions, causing climatic warming, and N emissions from fossil fuel burning and agriculture, leading to the deposition of reactive N, are not mitigated more effectively in the future, even harsher effects are to be expected [79]. Some success in reducing N emissions has already been achieved in North America and Europe. As a result, total N in wet deposition in Europe decreased on average by 2.7% between 2001–2002 and 2005–2007 [10]. However, no climate change mitigation measures are in place at a scale necessary to avoid further damage [11] nor will N emission reductions be able to fully release sensitive habitats from chronic N stress [6]. Moreover, N deposition in South America, Africa and most prominently in Asia are expected to raise further [10]. That is why studying the magnitude of these threats to protected habitats is important.

To date, assessments of potential future risks to biodiversity have largely neglected interactions between climate change, N deposition, and biodiversity [12]. Since most plant species have a distinct climate niche, climate warming and changes in the precipitation regime alter the distribution pattern of their suitable habitats and eventually lead to the loss of species [1315]. The main effect due to increased N deposition in terrestrial ecosystems is enhanced growth of some species, effectively using the additional N [4, 16]. In forests, increasing availability of soil N for plants might lead to more homogenous forest floor plant communities and hence to biodiversity loss [17] which in turn might have consequences for the food web structure and ecosystem functioning [18]. Beside deposition of inorganic N compounds, plant N availability is controlled by the effects of temperature and moisture on litter decomposition rates and mineralization of organic N soils [19]. Whereas N deposition leads to an increase in N availability [2024], expected climate change may have different effects. Warming will lead to enhanced N mineralization and nitrification and eventually to higher N uptake in response to higher tree growth [25, 26]. McDonnell et al. [27] showed in a modelling study that tree growth stimulation by climate change can considerably offset negative effects of N deposition on biodiversity. Wamelink et al. [28] showed that a higher N deposition may lead to devastating effect on biodiversity, especially in heathland and semi-natural grasslands.

Although we focus on eutrophication effects of N deposition and climate change induced impacts, soil acidification (through N and sulphur deposition) is an important additional driver. Sulphur (S) deposition (and to some extent N) has led to large-scale soil acidification during the second half of the 20th century, but significantly decreased in Europe since the late 1980s [10] followed by increasing soil pH values [29, 30] and a decrease in the prevalence of plant species adapted to acid soils [3133]. Climate warming can potentially accelerate soil recovery from acidification, because base cation input to the soil increases with an increase in weathering and litter decomposition [34, 35].

Coupled soil-plant models capable of quantifying potential interacting effects of climate change and N deposition on biodiversity are just emerging [36] and are being increasingly applied [27, 35, 3739]. Here we used the dynamic biochemical soil model VSD+ [40] together with two plant response models, PROPS [41] and BERN (for evaluation only) [39] with data from 18 long-term forest ecosystem observation sites distributed across Austria. Both plant models use empirical environmental response functions to predict species occurrence based on its realized niche [42]. We focused on habitat suitability, i.e. the capacity of a site (habitat) to host its characteristic plant species relying on the concept of potential natural vegetation [43]. We do so because the European Habitat Directive [5] adheres to the concept of “favourable conservation status” which can be measured by habitat suitability [44]. For future predictions three different N and S deposition and four different climate scenarios were used. Specifically, we hypothesized 1) that climate change will weaken the suitability of forest habitats to host their characteristic plant species because major changes in temperature and precipitation will push these species out of their optimum range, 2) that N deposition will also lower the habitat suitability due to soil N enrichment and a concomitant increase in the occurrence probability of nitrophilic plant species on the expense of oligotrophic species, and 3) that when taking interactions into account, eutrophication in forest floor species composition due to N deposition effects will be mitigated by climate driven increased N immobilisation in wood biomass.

Methods

The field sites are all part of the Austrian federal long-term monitoring program of air pollution effects of the Austrian Ministry for Agriculture, Forestry, Environment and Water Management. Owners of the land gave permission to conduct monitoring. Our study did not involve field studies on endangered species other than observation no involvement of human participants, specimens or tissue samples, or vertebrate animals, embryos or tissues.

Site characteristics

We used 18 forest sites which are part of ICP Forests and ICP Integrated Monitoring program within the framework of effects monitoring of the UNECE Convention on Long-range Transboundary Air Pollution [45] (Table 1). These sites were located to represent the variation in the more widespread forest ecosystems in Austria and hold long-term measurements of climate parameters, stand properties and soil condition [4648].

thumbnail
Table 1. Study sites, their code and name according to the European Habitat Directive (Directive 92/43/EEC), and their potential natural vegetation types.

Characteristic plant species are listed in S1 Table.

https://doi.org/10.1371/journal.pone.0184194.t001

The sites comprise 5 forest types protected under the European Habitat Directive in Austria. These are:

  • Luzulo-Fagetum beech forests (EU code 9110) with Fagus sylvatica L. as the dominant tree species, found at submontane to montane altitudes on acid soils in the Austrian Alps, its foothills, and the Bohemian Massive;
  • Asperulo-Fagetum beech forests (EU code 9130) with Fagus sylvatica as the dominant tree species found on neutral or near-neutral soils with mull humus;
  • Medio-European limestone beech forests of the Cephalanthero-Fagion (EU code 9150) with Fagus sylvatica as the dominant tree species found on dry south slopes at submontane to lower montane altitudes with base-rich soils on carbonate bedrock;
  • Pannonic woods with Quercus petraea (Matt.) Liebl. and Carpinus betulus L. (EU code 91G0), which are mixed deciduous forests found at colline to submontane altitudes of eastern Austria with dry continental climate;
  • Acidophilous Picea forests of the montane to alpine levels (Vaccinio-Piceetea) (EU code 9410) with Picea abies (L.) H.Karst. as the dominant tree species found at montane and subalpine altitudes of the Austrian Alps.

Soil-vegetation model

We use the dynamic geochemical soil model VSD+ (version 5.4, [40]) together with a newly developed plant response models PROPS [41]. For validation purpose, we also applied the plant model BERN [39] (Fig 1). The VSD+ model includes cation exchange (Gaines-Thomas or Gapon) and organic C and N dynamics according to the RothC-Model (version 26.3, [49]). VSD+ is driven by time series of N and S deposition as well as temperature and hydrology to predict soil solution chemistry and C and N pools. In its current version, the PROPS model is a database holding statistical niche functions for 4053 plant species occurring in Europe that were derived from a huge set of vegetation relevés together with associated soil data [41]. The outputs of PROPS are probabilities of species occurrences as a function of precipitation, temperature, N deposition, soil C:N ratio and soil pH. As the final habitat response indicator, we applied the Habitat Suitability Index (HSI) that describes the degree of suitability of site conditions for a set of typical species to co-occur. The HSI is defined as the arithmetic mean of the normalised probabilities of occurrence of these species [6]. In our study, we adopted the common approach taken in the EU [50] to use phytosociological plant community descriptions to define distinctive plant species for each of the 18 forest ecosystems (see Table 1 and S1 Table). These distinctive species are the characteristic species and all constant attendant species that can be found with a similar abundance in more than 70% of all vegetation relevés representing the plant community.

thumbnail
Fig 1. Model chain.

Model chain with VSD+ as the dynamic soil chemistry model and PROPS model for predicting the probability of occurrence for all characteristic species (BERN model for validation) and its respective Habitat Suitability Index (HSI). Climate driven inputs to VSD+ came from the hydrological model MetHyd, litter element input and plant uptake were calculated with GrowUp. Climate, N and S deposition scenario data was input to the model chain.

https://doi.org/10.1371/journal.pone.0184194.g001

Model setup

We had climate measurements with a daily resolution at the sites (6 sites) or from the closest meteorological station representative for the site (12 sites). These time series comprised the last 20–27 years. By using the weather generator ClimGen [51] and this measured data (precipitation, maximum temperature, minimum temperature, air humidity, global solar radiation), we derived the baseline climate time series for 1950–2100. Then, scenario data were synthesized by means of anomalies gathered from the A1B, A2 and B1 scenarios [52]. Newer, downscaled climate scenarios were not available at the start of the project. Parameter-specific monthly climate change anomalies for the study site were derived from the respective grid cell of the regional climate model COSMO-CLM [53]. The A1B, A2, and B1 scenarios were based on the global circulation model ECHAM5 and the A1B scenario was also available from the HadCM3 model [52] (Table 2).

thumbnail
Table 2. Climate change, N and S deposition scenarios.

Mean change and standard error in temperature (T) and precipitation (P) for three time slices (2030: 2025–35; 2050: 2045–55; 2100: 2090–2100) as compared to 2010 (2005–15). N and S deposition for 2020 and after 2030 with three scenarios given as B10: N and S deposition in the year 2010 with no further reduction; CLE: N and S deposition under revised Gothenburg Protocol emissions; MFR: N and S deposition under the technically maximum feasible emission reductions.

https://doi.org/10.1371/journal.pone.0184194.t002

We calculated total deposition of N and S for each site as the sum of throughfall and canopy exchange. The method is based on a canopy exchange model according to [54] with sodium as the tracer ion. In order to get long-term deposition of N and S we scaled large-scale modelled data to the measurements. We used reconstructed deposition from 1880 to 2000 [55] and deposition values for 2005, 2010, 2020 and 2030 from the latest EMEP model version [56] using the current legislation (CLE) scenario with revised Gothenburg Protocol emissions and the technically maximum feasible emission reduction scenario (MFR). A third deposition scenario was kept constant after 2010 (B10) (Table 2).

Soil data were taken from soil inventories in 2005 (IM_AT01, IM_AT02) or 2008 (all other sites). Mineral soils were sampled at depths of 0–5, 5–10, 10–20, 20-(35)40, and (35)40-(50)80 cm from a number (n≥3) of soil pits per plot. The organic layer was sampled and analysed separately. Specimen preparation and analysis followed the methods described in the ICP Forests monitoring program (http://icp-forests.net/page/icp-forests-manual). The calculation of input parameters for VSD+ followed the recommendations given in [40] and are described in more detail in S2 Table (input files in S1 File). Soil solution element concentrations were available for the 6 evaluation sites and were aggregated to mean annual values for the years 1998 to 2012.

Uptake of C, N and base cations as well as above- and belowground litterfall was calculated using GrowUp, version 1.3.2 [40]. Since both climate change and nitrogen deposition are supposed to change tree growth and thereby the uptake of N by trees and input via litterfall, a scaling of these input variables was done comparable to [57]. N and base cation uptake as well as C and N in litterfall were scaled according to a reference situation, i.e. mean values between 1970 and 1990 where the core of the forest yield tables [58] was obtained (see details in S2 File).

For 6 sites detailed data on forest floor vegetation was available for several years between 1990 and 2010 [31]. Forest vegetation at the study sites was recorded on permanent plots using sub-sampling units which were selected randomly or were distributed in a regular grid. Permanent plot size was 0.5 ha and sub-sampling unit size was 4 m2. For the ICP Integrated Monitoring plots IM_AT01 and IM_AT02 8 subsamples per plot were recorded, for ICP Forests sites 10 subsamples were available.

Model calibration and evaluation

The soil model VSD+ was calibrated using measured C and N pools, base saturation and soil solution chemistry (pH for all sites; NO3- and SO42- for 6 sites) using Markov Chain Monte Carlo method [59]. Linear regression, mean error and RMSE between measured and modelled values were calculated for soil C pools, C:N ratios, and soil solution pH, SO42- and NO3-. We calculated observed HSI by using the frequency of characteristic species in vegetation records of the sites. Additionally, we compared results from PROPS with the HSI calculated with the BERN model. BERN differs from PROPS as to the underlying relevé data and regarding the statistics used in deriving niche functions [36, 39]. Linear regressions between the two models were calculated.

Statistical analyses

In total, three N and S deposition scenarios (B10, CLE, MFR) and 5 climate scenarios (including baseline) were modelled for 18 sites resulting in 270 datasets.

First, we applied a two-way ANOVA with F-test statistics to test whether an overall trend exists over time in temperature (T), precipitation (P), N and S deposition, soil C:N ratio, soil solution pH and HSI and whether this trend is different between the climate scenarios and the deposition scenarios. To account for repeated measures over time, the scenarios were nested into four time slices (2010: average of 2005–15; 2030: 2025–35; 2050: 2045–55; 2100: 2090–2100) and interactions between climate and deposition scenarios were taken into account. Second, we used the same ANOVA design for each site. The ANOVA coefficients were used to determine effect sizes of the climate and deposition scenarios on soil C:N ratio, soil solution pH and HSI. Differences in effect sizes by 2100 between EU protection habitat types were tested using one-way ANOVA and Tukey’s HSD multi-comparison test. All statistical analyses were carried out with the package R, version 3.2.3 [60].

Results

Model calibration and evaluation

Using linear regression, R2 between modelled and measured HSI was 0.36 but not significant (p = 0.201) (Fig 2). Mean error was 0.05 and RMSE was 0.12. HSI values as derived from VSD+ and PROPS were comparable with those derived from VSD+ and BERN (R2 = 0.70, p = 0.037) but they differed in magnitude, with PROPS giving lower values (mean error was 0.22 and RMSE was 0.26) (Fig 2).

thumbnail
Fig 2. Modelled versus observed Habitat Suitability Index (HSI).

Comparison of the modelled Habitat Suitability Index (mean and standard deviation) using PROPS with A) observed data and B) with modelled HSI from BERN (using the same soil and climate data as PROPS). Six sites with vegetation records between 1996 and 2007 were used. The 1:1 line is dashed, the regression line is solid.

https://doi.org/10.1371/journal.pone.0184194.g002

Modelled and measured forest growth achieved an R2 of 0.63 (p<0.001), mean error was 527 kg m-2 and RMSE was 5519. VSD+ could be calibrated to the observed C pool and C:N-ratios resulting into R2 of modelled versus measured values of 0.86 and 0.99, respectively (p-values < 0.001). The mean error for the C pool was 1.05 kg m-2, RMSE was 2.19. The mean error for the C:N ratio was 0.13, RMSE was 0.68. As to soil solution composition, only pH values were significantly related with observations (R2 = 0.57, p<0.001; mean error = 0.41, RMSE = 1.32). SO42- and NO3- had mean errors of 0.15 and 0.05 eq m-3 and RMSE of 0.21 and 0.07 (S3 File).

Future trends in climate and deposition

Future T and P changed significantly over time in the applied climate scenarios (S4 File). Whereas T increased by 2.7°C on average between 2010 and 2100, P increased slightly towards 2050 and decreased by 76 mm in the year 2100 in comparison to 2010 (Table 2). Only T trends were significantly different between the climate scenarios. The ECHAM5-A1B, A2 and HADCM3-A1B scenario showed the highest T increases (3.4–4°C by 2100 on average), whereas the ECHAM5-B1 scenario showed the lowest increase (2.1°C). Only in ECHAM5-A2 P decreased continuously. ECHAM5-A1B and HADCM3-A1B were relatively similar regarding their T trajectories (the latter showed a faster increase) but differed as to P. Neither increase in T until 2100 nor change in P were significantly different between habitat types (ANOVA p = 0.979 and 0.364).

In both, the current legislation scenario (CLE) and the maximum feasible reduction scenario (MFR), N and S deposition decrease significantly over time and significantly differed from each other (S4 File). Averaged over all sites, N deposition decreased by 2 kg N.ha-1.yr-1 (CLE) and 4 kg N.ha-1.yr-1 (MFR) compared to the deposition in 2010 (Table 2). Differences in the decrease in N deposition until 2100 between habitat types due to regional differences in deposition trajectories were marginally significant (ANOVA p = 0.099) due to habitat type 9410 which experienced stronger decrease than 9130 (Tukey’s HSD p = 0.053).

Effects of climate, N and S deposition

The average increase of the soil C:N ratio units until 2100 over all scenarios was 3.4 (p<0.001) (S4 File). When compared to the baseline climate, climate scenarios had an increasing effect on soil C:N ratios in 82% of the sites and years (2030, 2050 and 2100). HADCM3-A1B had the strongest effect on C:N ratios (median: 0.72; 90th percentile: 4.29). Climate effects on C:N ratios by the year 2100 were different between habitat types (ANOVA p<0.001, Fig 3). B10 and CLE deposition scenarios, when compared to the lowest emission scenario (MFR), had decreasing effects on C:N ratios in 29% of all scenario runs and increasing effects in 69% (S3 Table). The effects of N deposition were an order of magnitude lower than the climate effects (B10: median = 0.03 and 90th percentile = 0.15). Deposition effects on C:N ratios by the year 2100 were mostly negative and differed between habitat types (ANOVA p<0.001, Fig 3).

thumbnail
Fig 3. Effects of climate change and N deposition scenarios.

Effects of climate change (A-C) and N and S deposition (D-F) on soil C:N ratio, soil pH, and the Habitat Suitability Index (mean and standard deviation of effects in the year 2100 as derived from 5 different climate change scenarios). Means with different letters are significant different (Tukey’s HSD p < 0.05). Effects are given in the form of ANOVA coefficients describing the difference between the mean values of all baseline climate model runs and the respective climate change scenario by the year 2100 and the difference between the mean values of all MFR deposition scenarios and the respective CLE and B10 deposition scenario by the year 2100 at each site. Positive coefficients represent increasing, negative coefficients decreasing effects. Note that MFR scenarios have the lowest N deposition. 9110: Luzulo-Fagetum beech forests, 9150: Medio-European limestone beech forests of the Cephalanthero-Fagion, 9130: Asperulo-Fagetum beech forests, 91G0: Pannonic woods with Quercus petraea and Carpinus betulus, 9410: Acidophilous Picea forests of the montane to alpine levels (Vaccinio-Piceetea).

https://doi.org/10.1371/journal.pone.0184194.g003

The average soil solution pH over all scenarios increased slightly between 2010 and 2100, but not significantly (p = 0.209) (S4 File). When compared to the baseline climate, climate scenarios had an increasing effect on soil solution pH in 50% of the sites and years (2030, 2050 and 2100) whereas 22% remained unaffected and in 28% we found a decreasing effect (S3 Table). Climate effects on soil pH by the year 2100 were indifferent between habitat types (ANOVA p = 0.136, Fig 3). Higher future N deposition had decreasing effects on soil pH in 57% of all model runs and no effect in 42%. Deposition effects on soil pH by the year 2100 were different between habitat types (ANOVA p<0.028, Fig 3). Overall, the effects of different climate and deposition scenarios on soil solution pH were rather small (10% percentile -0.11–0.02; 90% percentile: 0–0.18) (S3 Table).

The average HSI over all scenarios decreased significantly from 2010 to 2100 (from 0.2 to 0.14, p<0.001) (S4 File). When compared to the baseline climate, climate scenarios had a negative effect on HSI in 43% of the sites and years (2030, 2050 and 2100), 21% had a positive effect and 36% had no effect (S3 Table). Both A1B scenarios had the highest proportion of negative effects, ECHAM5-B1 the lowest. Positive and negative effects increased over the years so that only in 11–22% of the climate scenario runs in 2100 no effects were found. With higher deposition in B10 and CLE deposition scenario as compared to MFR, HSI was negatively affected in only 15%, positively affected in 45% and 40% of the model runs did not cause an effect. The effects of different climate and deposition scenarios on HSI were rather small (10% percentile of all model runs: -0.06 to -0.01; 90% percentile: 0.01–0.04) (S3 Table).

Climate effects on HSI by the year 2100 were indifferent between habitat types (ANOVA p = 0.255). However, for sites belonging to the EU habitat type 91G0, HSI exhibited the strongest negative climate effects (Fig 3) resulting from a disproportional number of plant species with decreasing occurrence probability, most strongly Stellaria holostea L. and Carpinus betulus (S5 File). Also sites belonging to EU habitat types 9110, 9150, and 9410 HSI experienced negative climate change effects on average. The plant species experiencing the strongest negative effects were (in descending order) Linnaea borealis L. and Lonicera caerulea L. (EU habitat type 9410), Hieracium murorum L. and Luzula luzuloides (Lam.) Dandy & Wilmott. (EU habitat type 9110), and Mercurialis perennis L. (EU habitat type 9150). In contrast to these habitats, mean effect on HSI in the EU habitat 9130 was small, and in some cases even positive (ECHAM5-B2 and HADCM3-A1B) (Fig 3). There, a small number of plant species showed strongly increasing occurrence probability, particularly in the habitat type 91G0 (Rosa canina L., Hedera helix L., Quercus petrea). The strongest negative effects were found for Senecio ovatus (G.Gaertn. et al.) Willd. and Sorbus aucuparia L., the strongest positive effects for Huperzia selago (L.) Bernh. and Vaccinium myrtillus L. (S5 File).

As mentioned above, HSI showed a positive response to higher N and S deposition in the B10 and CLE scenario as compared to the MFR scenario. Deposition effects on HSI by the year 2100 were indifferent between habitat types (ANOVA p = 0.742, Fig 3). The differences in the occurrence probabilities of plant species between the baseline climate and climate scenarios was much higher (between -60 and +30%) than the differences between deposition scenarios. MFR versus B10 scenarios resulted into differences in the range of -4 to +16%. The species with the strongest increases per habitat type were Hieracium murorum and Vaccinium myrtillus (EU habitat type 9110), Carpinus betulus and Corylus avellana L. (EU habitat type 91G0), Milium effusum L. and Senecio ovatus (EU habitat type 9130), Carex sylvatica Huds. and Athyrium filix-femina (L.) Roth (EU habitat type 9410), and Fagus sylvatica and Mercurialis perennis (EU habitat type 9150) (S5 File).

Discussion

Climate change effects on soil chemistry and HSI

According to our climate change scenarios, T increased and P decreased towards 2100 and these changes did not differ between forest habitat types albeit they differed between scenarios. When singled out, i.e. calculating differences between the baseline and climate change scenarios, climate strongly increased the soil C:N ratio. This increase in soil N deficiency was mostly triggered by N immobilisation in woody biomass as derived from empirical relationships between climate and forest growth [57]. Interactions of temperature and precipitation were taken into account insofar as to limit tree growth, particularly in relatively dry areas in eastern Austria which are likely becoming even more water deficient under expected climate change. As a consequence, the C:N ratio of oak forest (habitat 91G0) soils in these areas were negatively affected by climate change.

We expected that these climate changes should directly increase the occurrence probability of the most thermophilic plant species among the distinctive species of a habitat while the most cold-tolerant species should decrease. Indeed, the species experiencing the strongest negative climate effect were cold-tolerant species (e.g. Linnaea borealis, Lonicera caerulea) and many of those increasing were thermophilic (Rosa canina, Hedera helix). Additionally, and except for the dry forest habitat 91G0, oligotrophic species should increase while species preferring sites with higher nutrient availability should decrease. Likewise, oligotrophic species increased (e.g. Vaccinium myrtillus) but not so in the dry habitat 91G0. There, the thermophilic species of nutrient rich sites, Stellaria holostea L., increased. This is in accordance with decreasing soil C:N ratios indicating higher nutrient availability. Species such as Hieracium murorum and Luzula luzuloides decreased in acidophytic beech dominated forests (habitat 9110). These are typical species of these forests but future N availability may decline to an extent which lowers their abundance. These changes in species occurrence probabilities decreased the HSI in all habitats with no significant differences in the magnitude between them because direct climate effects on the habitat suitability predominated. To our knowledge, no comparable study exists so far. However, impacts studies have shown the decisive role of climate for the future of tree species in Austria [61] and neighbouring countries [62, 63].

N and S deposition effects on soil chemistry and HSI

Contrary to our expectations, the response of the HSI to higher N and S deposition in the B10 and CLE scenario as compared to the MFR scenario was positive. We found a slightly increasing overall trend in soil solution pH, corroborating observations from the Austrian forest soil monitoring [64] and other long-term observations across Europe [29, 30, 65, 66]. The small change in soil pH was no surprise since some of the sites are well buffered, S deposition declined most strongly in the 1980s and is on a rather low level since several years [10], and N deposition continued to acidify soils. When singling out deposition effect on soil pH from this long-term trend, beech dominated forests on acidic soils (9110), were more affected than all other habitats. These soils particularly acidified during the past period of high S deposition, and recovery from acidification might be more efficient than in spruce dominated forests. We note that climate warming can, in addition, increase base cation input to the soil via accelerated weathering and litter decomposition [34, 35], but significant effects of climate change on soil pH were only found in some of our study sites. Higher N deposition predominantly led to increasing soil C:N ratios until 2030 and 2050, but by 2100 most sites experienced negative effects. The effects were quite small (means < 1) and very likely irrelevant for plant species occurrence. A decrease in the soil C:N ratio in response to N deposition was found in numerous studies [21, 24]. However, N deposition may also stimulate plant growth, causing increases in soil C:N ratios [67]. Most of the study sites were N limited and exposed to relatively low N deposition, so that the implemented tree growth response to N deposition seems reasonable. As to plant species responses, we found what we expected: plant species preferring nutrient-rich sites increased to some extent (e.g. Corylus avellana, Milium effusum, Senecio ovatus, Carex sylvatica, Athyrium filix-femina, Mercurialis perennis). However, the resulting increase in the HSI means that additional N improved habitat suitability at these forest sites which might be due to the fact that N deficiency is still widespread in Austria, because of historic overuse and acidification during the last part of the 20th century [64]. Notwithstanding the difficulty to explain the increase in HSI, deposition effects on species occurrence probabilities, and hence the HSI, were much smaller than effects from climate causing negative total changes in habitat suitability.

Interactions of climate and N deposition

First, we note that the forest sites used for these analyses represent the major forest habitats in Austria with a wide distribution but do not cover many rarer habitats under conservation protection. Hence, only one of totally 8 of the EU directive Annex I priority habitats [5] could be included due to a lack of data. Our results suggests that while climate change will clearly lower the “favourable conservation status” of these Austrian forest habitats N deposition effects will be comparably weak. The reasons are twofold. First, N deposition in most of these forests will not exceed loads at which major changes in the soil chemistry occur, and, secondly, climate driven increase in N immobilisation in woody biomass will offset soil N enrichment, a result which is in line with Butler et al. [26]. Our results are also in line with observations showing that during the last decades European forests have become more nutrient deficient albeit N deposition was relatively high [68]. They showed that enhanced forest growth due to N deposition, climate warming and, possibly, CO2 fertilization increased the demand for N and other nutrients rendering soils more nutrient limited, underlining earlier hypothesis on progressive N limitation [69]. Indeed, incorporation of nutrient limitation other than N is a clear research field for the future. Observations are meagre but suggest that in some areas N (and S) deposition has caused changes in forest floor species composition [69, 70], while soil inventory data showed increases in pH but, together with needle leaf N concentrations, did not give an indication of large-scale soil N enrichment [64].

Conclusions

Measures of habitat suitability, such as the HSI, have the advantage that they are directly relevant for habitat management and conservation policies [44, 71, 72]. According to the European Union conservation legislation, a protected habitat is considered to have “favourable conservation status” if “the specific structure and functions which are necessary for its long-term maintenance exist” [5]. The HSI is based on predictions from empirical niche functions of distinctive species in a habitat [41]. While plant species’ niche functions have been used extensively in climate impact assessments and seem to be well defined [71, 73], their usefulness in studying air pollution effects is just being discussed [74]. Here we applied niche functions implemented in the newly developed model PROPS, which uses mean annual T, annual total P, annual N deposition, soil C:N ratio, and the soil solution pH as predictor variables. We were able to model the soil chemistry in a reasonable way using the dynamic soil model VSD+ [40]. When comparing the results of PROPS with those from a second plant response model (BERN), which is based on an independent empirical data set and which is rather different as to its statistical approach [39], they showed a high correlation. Hence, the niche functions, which are implemented in the current version of PROPS are, in general, reliable and this is obvious, because they are statistical representations of observed species occurrence data. However, on a site scale, and particularly when management comes into play, predictions are inherently difficult and hard to validate, a fact that has caused considerable discussion in climate impact studies before [75] as has the applicability of empirical niche functions to perform future impact assessments [42] Nevertheless, we could show that climate and air pollution effects on habitat suitability significantly interact yet often in an idiosyncratic way. Owing to this complexity and the still high uncertainty in present knowledge, adaptation to protect forest habitat biodiversity, such as defined under the European Habitat Directive, will be challenging. Dynamic soil-plant models can play an important role as supporting tools to assess possible future trajectories.

Supporting information

S1 Table. List of all distinctive plant species for the study sites.

https://doi.org/10.1371/journal.pone.0184194.s001

(PDF)

S2 Table. Methods used for soil and climate input data for VSD+.

https://doi.org/10.1371/journal.pone.0184194.s002

(PDF)

S3 Table. Site- and scenario-specific effects of climate change and N deposition.

https://doi.org/10.1371/journal.pone.0184194.s003

(PDF)

S2 File. Methods and results for forest growth.

https://doi.org/10.1371/journal.pone.0184194.s005

(PDF)

S4 File. Future changes at the site level.

https://doi.org/10.1371/journal.pone.0184194.s007

(PDF)

S5 File. Future changes in the occurrence probability of plant species.

https://doi.org/10.1371/journal.pone.0184194.s008

(PDF)

Acknowledgments

The study was funded by the Austrian Climate Research Program (ACRP5—CCN-Adapt—KR12AC5K01308) and by the Austrian Federal Ministry of Agriculture, Forestry. We are grateful to the data providers of the UNECE-CLRTAP International Cooperative Programmes ICP Forests and ICP Integrated Monitoring.

References

  1. 1. Assessment ME. Ecosystems and human well-being: biodiversity synthesis. Washington, DC: World Resources Institute; 2005.
  2. 2. Peñuelas J, Sardans J, Estiarte M, Ogaya R, Carnicer J, Coll M, et al. Evidence of current impact of climate change on life: a walk from genes to the biosphere. Global Change Biology. 2013;19(8):2303–38. pmid:23505157
  3. 3. Parmesan C. Ecological and evolutionary responses to recent climate change. Annual Review Ecol Evol Syst. 2006;37:637–69.
  4. 4. Bobbink R, Hicks K, Galloway J, Spranger T, Alkemade R, Ashmore M, et al. Global assessment of nitrogen deposition effects on terrestrial plant diversity: a synthesis. Ecological Applications. 2010 Jan;20(1):30–59. pmid:20349829
  5. 5. European Commission. Council Directive 92/43/EEC of 21 May 1992 on the conservation of natural habitats and of wild fauna and flora (1992).
  6. 6. Posch M, Hettelingh JP, Slootweg J, Reinds GJ. Deriving critical loads based on plant diversity targets. In: Slootweg J, Posch M, Hettelingh JP, Mathijsssen L, editors. Modelling and Mapping the impacts of atmospheric deposition on plant species diversity in Europe. Bilthoven, NL: RIVM; 2014. p. 41–53.
  7. 7. Bleeker A, Hicks WK, Dentener F, Galloway J, Erisman JW. N deposition as a threat to the World’s protected areas under the Convention on Biological Diversity. Environmental Pollution. 2011;159(10):2280–8. pmid:21122958
  8. 8. Pereira HM, Leadley PW, Proença V, Alkemade R, Scharlemann JPW, Fernandez-Manjarrés JF, et al. Scenarios for Global Biodiversity in the 21st Century. Science. 2010;330(6010):1496–501. pmid:20978282
  9. 9. Bellard C, Bertelsmeier C, Leadley P, Thuiller W, Courchamp F. Impacts of climate change on the future of biodiversity. Ecology Letters. 2012;15(4):365–77. pmid:22257223
  10. 10. Vet R, Artz RS, Carou S, Shaw M, Ro C-U, Aas W, et al. A global assessment of precipitation chemistry and deposition of sulfur, nitrogen, sea salt, base cations, organic acids, acidity and pH, and phosphorus. Atmospheric Environment. 2014;93:3–100.
  11. 11. OECD. Climate Change Mitigation: Policies and Progress. Paris: OECD Publishing; 2015.
  12. 12. Maes SL, De Frenne P, Brunet J, de la Peña E, Chabrerie O, Cousins SAO, et al. Effects of enhanced nitrogen inputs and climate warming on a forest understorey plant assessed by transplant experiments along a latitudinal gradient. Plant Ecology. 2014;215(8):899–910.
  13. 13. Engler R, Randin CF, Thuiller W, Dullinger S, Zimmermann NE, Araújo MB, et al. 21st century climate change threatens mountain flora unequally across Europe. Global Change Biology. 2011;17(7):2330–41.
  14. 14. Lenoir J, Gégout JC, Marquet PA, de Ruffray P, Brisse H. A significant upward shift in plant species optimum elevation during the 20th century. Science. 2008;320:1768–71. pmid:18583610
  15. 15. Thuiller W, Lavorel S, Araújo MB, Sykes MT, Prentice IC. Climate change threats to plant diversity in Europe. Proceedings of the National Academy of Sciences. 2005;102(23):8245–50.
  16. 16. Suding KN, Collins SL, Gough L, Clark C, Cleland EE, Gross KL, et al. Functional- and abundance-based mechanisms explain diveristy loss due to N fertilization. Proceedings of the National Academy of Sciences. 2005;102(12):4387–92.
  17. 17. Gilliam FS. Response of the herbaceous layer of forest ecosystems to excess nitrogen deposition. Journal of Ecology. 2006;94:1176–91.
  18. 18. Meunier CL, Gundale MJ, Sánchez IS, Liess A. Impact of nitrogen deposition on forest and lake food webs in nitrogen-limited environments. Global Change Biology. 2016;22(1):164–79. pmid:25953197
  19. 19. Butterbach-Bahl K, Gundersen P. Nitrogen processes in terrestrial ecosystems. In: Sutton MA, Howard CM, Erisman JW, Billen G, Bleeker A, Grennfelt P, et al., editors. The European Nitrogen Assessment: Cambridge University Press; 2011. p. 99–124.
  20. 20. Sardans J, Rivas-Ubach A, Peñuelas J. The C:N:P stoichiometry of organisms and ecosystems in a changing world: A review and perspectives. Perspectives in Plant Ecology, Evolution and Systematics. 2012;14(1):33–47.
  21. 21. Cools N, Vesterdal L, De Vos B, Vanguelova E, Hansen K. Tree species is the major factor explaining C:N ratios in European forest soils. Forest Ecology and Management. 2014;311(0):3–16.
  22. 22. Templer PH, Mack MC, Iii FSC, Christenson LM, Compton JE, Crook HD, et al. Sinks for nitrogen inputs in terrestrial ecosystems: a meta-analysis of 15N tracer field studies. Ecology. 2012;93(8):1816–29. pmid:22928411
  23. 23. Zechmeister-Boltenstern S, Keiblinger KM, Mooshammer M, Peñuelas J, Richter A, Sardans J, et al. The application of ecological stoichiometry to plant–microbial–soil organic matter transformations. Ecological Monographs. 2015;85(2):133–55.
  24. 24. Mulder C, Hettelingh JP, Montanarella L, Pasimeni MR, Posch M, Voigt W, et al. Chemical footprints of anthropogenic nitrogen deposition on recent soil C: N ratios in Europe. Biogeosciences. 2015;12(13):4113–9.
  25. 25. Porter EM, Bowman WD, Clark CM, Compton JE, Pardo LH, Soong JL. Interactive effects of anthropogenic nitrogen enrichment and climate change on terrestrial and aquatic biodiversity. Biogeochemistry. 2013;114(1):93–120.
  26. 26. Butler SM, Melillo JM, Johnson JE, Mohan J, Steudler PA, Lux H, et al. Soil warming alters nitrogen cycling in a New England forest: implications for ecosystem function and structure. Oecologia. 2012;168(3):819–28. pmid:21983640
  27. 27. McDonnell TC, Belyazid S, Sullivan TJ, Sverdrup H, Bowman WD, Porter EM. Modeled subalpine plant community response to climate change and atmospheric nitrogen deposition in Rocky Mountain National Park, USA. Environmental Pollution. 2014;187:55–64. pmid:24448482
  28. 28. Wamelink GWW, van Dobben HF, Mol-Dijkstra JP, Schouwenberg EPAG, Kros J, de Vries W, et al. Effect of nitrogen deposition reduction on biodiversity and carbon sequestration. Forest Ecology and Management. 2009;258(8):1774–9.
  29. 29. Akselsson C, Hultberg H, Karlsson PE, Pihl Karlsson G, Hellsten S. Acidification trends in south Swedish forest soils 1986–2008—Slow recovery and high sensitivity to sea-salt episodes. Science of The Total Environment. 2013;444(0):271–87.
  30. 30. Cools N, De Vos B. Availability and evaluation of European forest soil monitoring data in the study on the effects of air pollution on forests. iForest—Biogeosciences and Forestry. 2011 2011-11-03;4(5):205–11.
  31. 31. Dirnböck T, Grandin U, Bernhardt-Römermann M, Beudert B, Canullo R, Forsius M, et al. Forest floor vegetation response to nitrogen deposition in Europe. Global Change Biology. 2014;20(2):429–40. pmid:24132996
  32. 32. Reinecke J, Klemm G, Heinken T. Vegetation change and homogenization of species composition in temperate nutrient deficient Scots pine forests after 45 yr. Journal of Vegetation Science. 2014;25(1):113–21.
  33. 33. Vanhellemont M, Baeten L, Verheyen K. Relating changes in understorey diversity to environmental drivers in an ancient forest in northern Belgium. Plant Ecology and Evolution. 2014;147(1):22–32.
  34. 34. Aherne J, Posch M, Forsius M, Lehtonen A, Härkönen K. Impacts of forest biomass removal on soil nutrient status under climate change: a catchment-based modelling study for Finland. Biogeochemistry. 2011;107(1):471–88.
  35. 35. Gaudio N, Belyazid S, Gendre X, Mansat A, Nicolas M, Rizzetto S, et al. Combined effect of atmospheric nitrogen deposition and climate change on temperate forest soil biogeochemistry: A modeling approach. Ecological Modelling. 2015;306:24–34.
  36. 36. De Vries W, Wamelink GWW, Dobben Hv, Kros J, Reinds GJ, Mol-Dijkstra JP, et al. Use of dynamic soil–vegetation models to assess impacts of nitrogen deposition on plant species composition: an overview. Ecological Applications. 2010;20(1):60–79. pmid:20349830
  37. 37. Belyazid S, Kurz D, Braun S, Sverdrup H, Rihm B, Hettelingh J-P. A dynamic modelling approach for estimating critical loads of nitrogen based on plant community changes under a changing climate. Environmental Pollution. 2011;159(3):789–801. pmid:21145634
  38. 38. Rizzetto S, Belyazid S, Gégout J-C, Nicolas M, Alard D, Corcket E, et al. Modelling the impact of climate change and atmospheric N deposition on French forests biodiversity. Environmental Pollution. 2016;213:1016–27. pmid:26809502
  39. 39. Schlutow A, Dirnböck T, Pecka T, Scheuschner T. Use of an empirical model approach for modelling trends of ecological sustainability. In: De Vries W, Hettelingh JP, Posch M, editors. Critical Loads and Dynamic Risk Assessments: Nitrogen, Acidity and Metals in Terrestrial and Aquatic Ecosystems. Dordrecht, NL: Springer; 2015. p. 381–400.
  40. 40. Bonten LTC, Reinds GJ, Posch M. A model to calculate effects of atmospheric deposition on soil acidification, eutrophication and carbon sequestration. Environmental Modelling & Software. 2016;79:75–84.
  41. 41. Reinds GJ, Mol-Dijkstra JP, Bonten L, Wamelink GWW, De Vries W, Posch M. VSD+ PROPS: Recent Developments. In: Slootweg J, Posch M, Hettelingh JP, Mathijssen L, editors. Modelling and Mapping the impacts of atmospheric deposition on plant species diversity in Europe. Bilthoven, NL; 2014. p. 47–53.
  42. 42. Elith J, Leathwick JR. Species distribution models: ecological explanation and prediction across space and time. Annual Review of Ecology, Evolution, and Systematics. 2009;40:677–97.
  43. 43. Tüxen R. Die heutige potentielle natürliche Vegetation als Gegenstand der Vegetationskartierung. Angewandte Pflanzensoziologie (Stolzenau). 1956;13:5–42.
  44. 44. Rowe EC, Ford AES, Smart SM, Henrys PA, Ashmore MR. Using Qualitative and Quantitative Methods to Choose a Habitat Quality Metric for Air Pollution Policy Evaluation. PLoS ONE. 2016;11(8):e0161085. pmid:27557277
  45. 45. UNECE. The 1999 Gothenburg Protocol to abate acidification, eutrophication and ground-level ozone. Geneva, Switzerland: UNECE1999.
  46. 46. Neumann M, Schnabel G, Gärtner M, Starlinger F, Fürst A, Mutsch F, et al. Forest Condition Monitoring in Austria—Results of the permanent observation plots (Level II). Vienna: Austrian Research Centre for Forests 2001.
  47. 47. Mutsch F, Leitgeb E, Hacker R, Amann C, Aust G, Herzberger E, et al. Projekt BioSoil–Europäisches Waldboden-Monitoring (2006/07) Datenband Österreich. Vienna: Austrian Research Centre for Forests 2013.
  48. 48. Jost G, Dirnböck T, Grabner M-T, Mirtl M. Nitrogen leaching of two forest ecosystems in a karst watershed. Water Air Soil Pollut. 2011;218(1):633–49.
  49. 49. Coleman K, Jenkins DS. RothC-26.3. A model for the turnover of carbon in soil. Model description and users guide. Harpenden, UK: IACR Rothamsted 2005.
  50. 50. European Commission DG Environment. Interpretation Manual of European Union Habitats—EUR28. Brussles: European Commission DG Environment 2013.
  51. 51. Stöckle CO, Campbell G, Nelson S. ClimGen Manual. Pullman, WA, USA: Biological Systems Engineering Department, Washington State University; 1999.
  52. 52. IPCC AR4 WG1. Climate Change 2007: The Physical Science Basis, Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. S. S, D. Q, M. M, Z. C, M. M, K.B. A, et al., editors. Cambridge, United Kingdom and New York, NY, USA 2007.
  53. 53. Loibl W, Formayer H, Schöner W, Truhetz H, Anders I, Gobiet A, et al. reclip:century 1: Models, Data and GHG-Scenarios, Simulations. Vienna: AIT Austrian Institute of Technology 2011.
  54. 54. Adriaenssens S, Staelens J, Baeten L, Verstraeten A, Boeckx P, Samson R, et al. Influence of canopy budget model approaches on atmospheric deposition estimates to forests. Biogeochemistry. 2013;116(1):215–29.
  55. 55. Schöpp W, Posch M, Mylona S, Johansson M. Long-term development of acid deposition (1880–2030) in sensitive freshwater regions in Europe. Hydrol Earth Syst Sci. 2003;7(4):436–46.
  56. 56. Simpson D, Benedictow A, Berge H, Bergström R, Emberson LD, Fagerli H, et al. The EMEP MSC-W chemical transport model and technical description. Atmos Chem Phys. 2012;12(16):7825–65.
  57. 57. De Vries W, Posch M. Modelling the impact of nitrogen deposition, climate change and nutrient limitations on tree carbon sequestration in Europe for the period 1900–2050. Environmental Pollution. 2011;159(10):2289–99. pmid:21163561
  58. 58. Marschall J. Hilfstafeln für die Forsteinrichtung. 5 ed. Vienna, Austria: Österreichischer Agrarverlag; 2011.
  59. 59. Reinds GJ, van Oijen M, Heuvelink GBM, Kros H. Bayesian calibration of the VSD soil acidification model using European forest monitoring data. Geoderma. 2008;146(3–4):475–88.
  60. 60. Team RC. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing2015.
  61. 61. Lexer MJ, Hönninger K, Scheifinger H, Matulla C, Groll N, Kromp-Kolb H, et al. The sensitivity of Austrian forests to scenarios of climatic change: a large-scale risk assessment based on a modified gap model and forest inventory data. Forest Ecology and Management. 2002;162(1):53–72.
  62. 62. CH2014-Impacts. Toward quantitative scenarios of climate change impacts in Switzerland Bern, Switzerland 2014.
  63. 63. Wohlgemuth T. Climate change and tree responses in Central European forests. Annals of Forest Science. 2015;72(3):285–7.
  64. 64. Jandl R, Smidt S, Mutsch F, Fürst A, Zechmeister H, Bauer H, et al. Acidification and Nitrogen eutrophication of Austrian forest soils. Applied and Environmental Soil Science. 2012;2012:9.
  65. 65. Matzner E, Zuber T, Lischeid G. Response of soil solution chemistry and solute fluxes to changing deposition rates. Ecological Studies: Biogeochemistry of forested catchments in a changing environment. 2004;172:339–60.
  66. 66. Kirk GJD, Bellamy PH, Lark RM. Changes in soil pH across England and Wales in response to decreased acid deposition. Global Change Biology. []. 2010;16(11):3111–9.
  67. 67. Rowe EC, Jones L, Dise NB, Evans CD, Mills G, Hall J, et al. Metrics for evaluating the ecological benefits of decreased nitrogen deposition. Biological Conservation. 2016.
  68. 68. Jonard M, Fürst A, Verstraeten A, Thimonier A, Timmermann V, Potočić N, et al. Tree mineral nutrition is deteriorating in Europe. Global Change Biology. 2015;21(1):418–30. pmid:24920268
  69. 69. Zechmeister HG, Dirnböck T, Hülber K, Mirtl M. Assessing airborne pollution effects on bryophytes—Lessons learned through long-term integrated monitoring in Austria. Environmental Pollution. 2007;147:696–705. pmid:17084007
  70. 70. Hülber K, Dirnböck T, Kleinbauer I, Willner W, Dullinger S, Karrer G, et al. Long-term impacts of nitrogen and sulphur deposition on forest floor vegetation in the Northern limestone Alps, Austria. Applied Vegetation Science. 2008;11:395–404.
  71. 71. Guisan A, Rahbek C. SESAM–a new framework integrating macroecological and species distribution models for predicting spatio-temporal patterns of species assemblages. Journal of Biogeography. 2011;38(8):1433–44.
  72. 72. Delbaere B, Nieto Serradilla A, Snethlage M. BioScore: A tool to assess the impacts of European Community policies on Europe’s biodiversity. Tilburg, the Netherlands: ECNC; 2009.
  73. 73. Thuiller W, Lafourcade B, Engler R, Araújo MB. BIOMOD–a platform for ensemble forecasting of species distributions. Ecography. 2009;32(3):369–73.
  74. 74. Rowe EC, Wamelink GWW, Smart SM, Butler A, Henrys PA, van Dobben HF, et al. Field survey based models for exploring nitrogen and acidity effects on plant species diversity and assessing long-term critical loads. In: De Vries W, Hettelingh JP, Posch M, editors. Critical Loads and Dynamic Risk Assessments: Nitrogen, Acidity and Metals in Terrestrial and Aquatic Ecosystems. Dordrecht, NL: Springer; 2015. p. 297–326.
  75. 75. Randin CF, Dirnböck T, Dullinger S, Zimmermann NE, Zappa M, Guisan A. Are niche-based species distribution models transferable in space? Journal of Biogeography. 2006;33(10):1689–703.