全球变化知识资源中心

Natural disturbances, that is, large pulses of tree mortality from agents such as wildfire, insect outbreaks or strong winds, are integral drivers of forest dynamics⁶ and contribute to the diversity and adaptive capacity of ecosystems⁷. Yet, forest disturbance regimes have changed considerably in recent years. The frequency and severity of large wildfires, for instance, has increased around the globe in past decades^{1, 8, 9}. In addition, recent bark beetle outbreaks, for example, in North America and Central Europe^{10, 11}, have reached unprecedented levels. A continuation of this trend towards more frequent and severe disturbances is also presumed for the coming decades. Bark beetles are, for instance, expected to colonize previously unsuitable habitats in higher latitudes and mountain forests¹², and large wildfires occurring only rarely in the past are predicted to return with higher frequency under climate change¹³.

Intensifying disturbance regimes are thus expected to be among the most severe impacts of climate change on forest ecosystems, raising concerns that disturbances might increasingly interfere with a continuous and sustainable provisioning of ecosystem services to society^{14, 15}. With regard to forest carbon (C) storage, which is an increasingly important ecosystem service in the context of climate change mitigation, forests are ‘slow in, rapid out’ systems¹⁶, with disturbance being a major pathway of fast, large-scale ecosystem C loss³. Forest-related climate change mitigation policies are thus highly sensitive to disturbance regimes⁴. Intensifying disturbance regimes have already been associated with a weakening of the European forest C sink recently⁵. A further increase in disturbance damage in the future might thus pose a major risk for Europe’s climate change mitigation efforts, as it could counteract the efforts to offset anthropogenic climate change through enhanced C storage in forest ecosystems⁴. Yet, consistent continental-scale assessments of potential changes in the forest disturbance regime under climate change are still missing so far. Furthermore, it is as of yet unclear how alternative European forest policies¹⁷ will influence disturbance regimes, and to which degree increasing disturbances might offset the potential of these policies to enhance Europe’s forest C storage capacity.

Using a new combination of Europe-wide disturbance observations (>29,000 records), scenario simulations of future forest development, and statistical disturbance modelling, we here report a continental-scale disturbance time series from 1971 to 2030 (Supplementary Information). Our analysis focused on an area of 131.6 × 10⁶ hectares of forests available for wood supply in 2005, covering 29 European countries in eight different ecoregions. We address the three most detrimental forest disturbance agents in Europe, which are wind, bark beetles and forest fires. Future disturbance damage for 2011–2030 is projected under four alternative forest management strategies¹⁷ for an ensemble of climate scenarios (continental-scale ensemble median warming of +1.1 °C, precipitation increase of +2.8%, and an increase in maximum daily windspeed by +0.7% until 2030 compared with 1971–2001, see Supplementary Table 4). To investigate potential impacts of disturbance on climate change mitigation efforts we quantified the impact of future disturbance regimes on the C storage capacity of Europe’s forests by means of an analytical C cycle model¹⁸. We evaluated how the effects of changing climate and disturbance regimes interact with alternative management strategies with regard to their implications on forest C storage.

We found that the disturbance intensification previously reported for the second half of the twentieth century^{1, 2} accelerated in the first decade of the twenty-first century. With damage of 32.3 × 10⁶ m³ yr⁻¹ (wind), 14.5 × 10⁶ m³ yr⁻¹ (bark beetles) and 9.4 × 10⁶ m³ yr⁻¹ (forest fires), the disturbance levels observed for the first decade of the twenty-first century were the highest of the past 40 years for all three disturbance agents (increasing by +139.6%, +601.9% and +231.1% relative to 1971–1980, respectively). The total disturbance damage from these three agents increased on average by +1.06 × 10⁶ m³ yr⁻¹ between 1971 and 2001. This rate of increase rose to +1.60 × 10⁶ m³ yr⁻¹ in 2002–2010.

Assuming a continuation of current forest management (reference strategy), projections under climate change resulted in a further increase in disturbance damage in all scenarios and for all agents (Fig. 1). The ensemble median (and interquartile range, IQR) wind damage for Europe was estimated to 44.5 × 10⁶ m³ yr⁻¹ (IQR: 6.5 × 10⁶ m³ yr⁻¹) in 2021–2030 (+229.4% compared with 1971–1980). Bark beetle damage increased to 17.9 × 10⁶ m³ yr⁻¹ (IQR: 1.4 × 10⁶ m³ yr⁻¹) in 2021–2030 (+763.7% relative to 1971–1980). For the same period, the timber volume damaged from forest fires was predicted to increase to 11.7 × 10⁶ m³ yr⁻¹ (IQR: 0.6 × 10⁶ m³ yr⁻¹) (+313.9% relative to 1971–1980). Over all agents, the annual rate of increase in disturbance damage predicted for 2011–2030 was +0.91 × 10⁶ m³ yr⁻¹ (median over the ensemble of climate change scenarios). If stable climatic conditions were assumed, however, disturbance damage remained constant or decreased moderately (Supplementary Fig. 9). This documents that climate change is the key driver of the disturbance changes projected for the coming decades (see also Supplementary Information).

Values are disturbance damage in millions of cubic metres of timber per year—note the different scales on the individual panels. Predictions (lighter hues) assume a continuation of business-as-usual forest management (reference strategy) and represent the median over an ensemble of scenarios of future climate and forest growth. Missing agents—such as bark beetles in the Northern and Atlantic ecoregions—indicate that they were not modelled in these particular ecoregions owing to a lack of historical data and/or relevance of the agent in these areas (Supplementary Table 5). The x axes are the same in all panels.

Assessment tools.

To project trajectories of future forest development under different management strategies and future climates we used the European Forest Information SCENario model EFISCEN. EFISCEN is a large-scale forest simulation model that projects forest resource development at regional to European scale²⁸. In EFISCEN, the state of the forest is described as an area distribution over age- and volume-classes in matrices, based on national forest inventory data. Transitions of area between matrix cells represent different natural processes in the simulation (for example, growth, mortality), and are influenced by management regimes and scenario changes. The effects of climatic changes are implemented in EFISCEN through deriving response functions for key processes such as forest growth from detailed process-based models.

We used empirically parameterized disturbance models to project the future damage by wind, bark beetles and wildfires in Europe’s forest ecosystems. The empirical relationships were developed using country-scale disturbance data for the period 1958–2001¹, and are based on a compilation of >29,000 disturbance records across Europe²⁹. The disturbance agents wind, bark beetles and forest fire were represented by individual models at country scale. Using unsupervised machine learning, indicators of climate as well as forest extent, structure and composition were selected as predictor variables². Their influence strength on disturbance damage was determined through structural equation modelling². For prediction, we used climate scenario data in combination with the respective forest structure and composition projections of EFISCEN.

To estimate the impact of disturbance on the forest C budget we used the REGIME model¹⁸, which provides a general framework to quantitatively assess disturbance effects on ecosystem carbon storage capacity at large spatial scales. The main constituents of the C cycle in REGIME are net primary production, the size of the ecosystem C pools, and their residence times. The disturbance regime is characterized by the fraction of live biomass C removed per unit of time. We used simulation results of undisturbed EFISCEN runs to parameterize the REGIME model, and estimated disturbance levels by means of empirical disturbance models. The overall C effect of disturbance in Europe’s forests was calculated by additively combining the three disturbance agents studied. Details and evaluations of the applied approaches can be found in the Supplementary Information.

Scenario analysis.

For the period 2011–2030 we studied 14 different scenarios of future climate change and tree growth (Supplementary Table 4). These scenarios cover three different storylines of future global development as outlined by the Intergovernmental Panel on Climate Change (A1B, B1 and B2). The corresponding changes in the climate system were derived from runs with three different sets of global circulation models and regional climate models, and climate change signals were calculated by standardizing the prediction period to a past baseline period. For each of these scenarios (with the exception of the B2 storyline) two alternative assumptions with regard to the fertilizing effect of CO₂ on tree growth were studied, ranging from full acclimation (no fertilizing effect) to a persistent growth enhancement through elevated atmospheric CO₂ levels³⁰. For the ensemble analysis of future trajectories we focused on the nine unique (non-replicated) scenarios of future climate change and tree growth, and report the ensemble median except where stated otherwise. A scenario of stable climate (assuming the climate conditions of 1971–2001) was used to isolate the role of climate change in our projections.

For all climate scenarios, four different management strategies were simulated with EFISCEN (ref. 17). The reference strategy describes the future trajectory of Europe’s forests under a continuation of current forest policies (business-as-usual management). Both forest area and growing stock increase distinctly under this strategy, whereas the proportion of conifers on growing stock increases only slightly and the median forest age decreases. The three alternative management strategies represent a shift in forest policy towards carbon management, biodiversity conservation, and wood energy production, respectively¹⁷. By the end of the study period in 2030 these alternative strategies considerably alter the forest extent, structure and composition relative to reference management (Supplementary Table 7). The strategies with the most profound changes in forest structure, composition and extent are the biodiversity and carbon strategies. The former assumes that priority is given to the protection of biological diversity, resulting in less wood removals, longer rotation periods and increasing growing stock. In addition, a sizeable percentage of the forest area is taken out of management. As the focus of this study is on forests available for wood supply¹⁷, the forest area is decreasing under the biodiversity strategy in our analysis. In the carbon strategy, rotation periods were increased and thinning regimes were optimized for increased C storage. The fourth strategy, promoting wood energy, showed comparatively little contrast to reference management with regard to indicators relevant in the context of disturbance damage, as many of its assumptions relate to increased removal of harvest residues as well as market changes and policy responses.

Corrected online 04 September 2014

In the version of this Letter previously published, the value given for net ecosystem productivity was incorrect, and should have read 99.2 Tg C yr^–1; this has no impact on the reported results. These corrections have been made in the online versions of the Letter.

Corrected online 08 August 2014

In the version of this Letter originally published, the received date was incorrect and should have read 14 October 2013. This error has now been corrected in all versions of the Letter.

1. Schelhaas, M-J., Nabuurs, G. & Schuck, A. Natural disturbances in the European forests in the 19th and 20th centuries. Glob. Change Biol. 9, 1620–1633 (2003).
   • Article
2. Seidl, R., Schelhaas, M-J. & Lexer, M. J. Unraveling the drivers of intensifying forest disturbance regimes in Europe. Glob. Change Biol. 17, 2842–2852 (2011).
   • Article
3. Kurz, W. A. et al. Mountain pine beetle and forest carbon feedback to climate change. Nature 452, 987–990 (2008).
   • CAS
   • ISI
   • PubMed
   • Article
4. Le Page, Y. et al. Sensitivity of climate mitigation strategies to natural disturbances. Environ. Res. Lett. 8, 015018 (2013).
   • Article
5. Nabuurs, G-J. et al. First signs of carbon sink saturation in European forest biomass. Nature Clim. Change 3, 792–796 (2013).
   • CAS
   • Article
6. Franklin, J. F. et al. Distu