Access the full text.
Sign up today, get DeepDyve free for 14 days.
References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.
International Journal of Biodiversity Science, Ecosystem Services & Management, 2014 Vol. 10, No. 3, 187–197, http://dx.doi.org/10.1080/21513732.2014.939719 The importance of climate variability changes for future levels of tree-based forest ecosystem services Livia Rasche* Research Unit Sustainability and Global Change, University Hamburg, Grindelberg 5, 20144 Hamburg, Germany The climate is changing, yet to which degree and in what pattern remains uncertain in many areas. In forest science, mainly impacts of long-term mean changes in temperature and precipitation distributions are studied. This paper therefore presents a sensitivity analysis to determine the importance of accounting for variability changes. A 10*10 matrix of different mean annual temperatures and precipitation sums is set up, covering the temperate forest zone of Europe, and the current level of several ecosystem services (ES) is calculated. Subsequently, mean and variability of temperature and precipitation distribu- tions are changed in different steps, and new ES levels calculated. The results show that for the study of climate change impacts on forest ES, climate parameter variability is of secondary importance. The trends are well represented with scenarios of mean climate parameter changes only; however, on moisture-limited and heat-stressed sites impacts of changes in variability gain in importance. Most of these impacts are negative, and can be observed not only in monocultures already at their physiological limit, but also in diverse stands. Different ES, however, show different sensitivities towards changes in mean and variability, underlining the need to develop adaptation measures tailored to the sites and ES of interest. Keywords: environmental gradient; monoculture; sensitivity study; structural diversity; temperate forest 1. Introduction 2006) and thus affect species’ ranges. Species distribution models using measures of extremes – in terms of inter- Climate models driven by different emission scenarios annual variability of climate parameters – thus show a have repeatedly been applied to the question of how our clear improvement over models using only means when climate will change in future. The results clearly show that explaining species’ range limits (Zimmermann et al. there will be changes in the spatial and temporal patterns 2009). Species distribution modelling and related fields of temperature and precipitation distributions (Cassman & are special cases, however, because a change in variability Wood 2005). How the patterns will change, and to which has to be included explicitly as an explanatory variable, degree, however, is still subject to large uncertainties in whereas modellers in other fields just use climate data series many cases (IPCC 2007b). This poses some problems to as input to run their models, and thus simulate changes in impact analysts, as even small shifts in the mean or varia- mean and variability as present in the data automatically. bility of climatic parameter distributions can lead to very The problem with this is the aforementioned uncertainty large percentage changes in extremes (Trenberth 2012). In about variabilities in the generated climate data series. crop science, for example, not only the occurrence, but In this paper, I therefore want to carry out a sensitivity also the timing of extreme events is crucial to determine analysis to determine how important it is for impact ana- the accurate prediction of yields, as unseasonal extreme lyses in the forest sector to account for the uncertainty in events have much stronger impacts (Collier et al. 2008; variability changes. Biomass, for example, decreases with Luo 2011). These very immediate and sometimes cata- an increasing variability in the precipitation distribution, strophic impacts have made the study of the relationship and pine trees gain dominance over spruce trees on more between weather variation and growth responses of plants water-limited sites (Bugmann & Pfister 2000). How sensi- the topic of many talks and publications (Monteith 2007). tive other forest ecosystem services (ES) are to changes in In forest science, the focus lies more on the impacts of variability has not been explored systematically yet, how- long-term changes in the mean of temperature and precipi- ever. I thus want to expand on current knowledge, and tation distributions on factors such as species composition study the influence of changes in climate variability on the and timber yield (e.g. Solomon 1986; Shugart et al. 2003; provision of regulating, cultural, and supporting ES in the Eggers et al. 2008; Kellomaki et al. 2008;Daleet al. 2010). temperate forest zone of Europe. Furthermore, I want to This is hardly surprising, considering the relative immunity determine whether diverse stands and monocultures of of mature trees to changes in climate. Variability gains in Fagus sylvatica L. and Picea abies Karst. show different importance, however, the closer species are to their physio- sensitivities, and whether there are thresholds in the sever- logical tolerances. Even short amounts of time beyond ity of change that may lead to drastically different certain thresholds may limit successful regeneration responses. And lastly, I want to identify a possible ‘worst (Honnay et al. 2002) or cause diebacks (Bigler et al. *Email: firstname.lastname@example.org © 2014 Taylor & Francis 188 L. Rasche case’ and a corresponding ‘best case’ scenario of change 2.2. Gradient of environmental conditions for each forest type and ES. For the systematic analysis of changes in ES provision The results of this study may lead to a clearer under- caused by changes in temperature and precipitation dis- standing of the importance of accounting for uncertainties tributions, an artificial gradient of environmental condi- in climate parameter variability and thus to a better under- tions covering the temperate zone of Europe was standing of the ranges of change in ES provision that may constructed. The site-specific weather variables for each be expected in future. month were determined as follows: First, a previously utilized gradient spanning environmental conditions from the cold to the dry tree line in Central Europe (e.g. 2. Material and methods Bugmann & Solomon 2000) was scrutinized for its cli- matic range, and the coldest measured long-term mean 2.1. Ecosystem services monthly temperature for each month was chosen to form Forests provide many goods for human use, such as tim- the base series T1. Values in this series were then ber, fibres, and non-wood forest products, but they also increased nine times in steps of 1 degree Celsius, resulting provide many services that ultimately benefit humankind: in 10 temperature series (Table 1). The same was carried habitat for more than 50% of the world’s known terrestrial out for precipitation, with the lowest measured long-term plant and animal species, storage of about 50% of the monthly precipitation sums chosen for series P1, and then world’s terrestrial carbon stock, and protection of more increased nine times in steps of 24 cm per year (2 cm per than 75% of the world’s freshwater (Shvidenko et al. month), also resulting in 10 series of baseline climate 2005). A classification of these goods and services was conditions (Table 2). attempted in the Millennium Ecosystem Assessment report Each temperature series was subsequently combined (Millennium Ecosystem Asssessment 2003), which distin- once with each precipitation series, resulting in weather guishes provisioning, regulating, cultural, and supporting data for 100 distinct sites. Average soil characteristics with services. As only natural forests were simulated to avoid a water-holding capacity (‘bucket size’) of 15 cm and an confounding impacts of forest management with impacts intermediate nutrient availability of 70 kg nitrogen per of climate change, no provisioning service (of timber) was hectare were universally assumed. In a first run of the −1 −1 included in the evaluation. Biomass per hectare (t∙ha ∙a ) 100 sites, all 30 native species currently parameterized was chosen to represent regulating services (potential for for Europe (Table 2 in Rasche et al. 2012) were allowed carbon storage and thus climate regulation), and cultural to establish, in a second run the species list was limited to services were represented by a structural diversity index only F. sylvatica, and in a third one to P. abies, the two (Staudhammer & Lemay 2001), illustrating the common economically most important species in Europe. perception that diverse forests are more aesthetically pleas- ing than even-aged stands (e.g. Carvalho-Ribeiro & Lovett 2011). Shannon’s diversity index is based on the diameter 2.3. Weather data for current and future climatic at breast height (DBH = p ) of each species-specific cohort conditions (all cohorts = s): As described in Section 2.2, I used a previously defined environmental gradient as basis for the artificial gradient. The weather data for the original gradient were obtained H ¼ p ðlnðp ÞÞ structure i i from the Landscape Dynamics Unit at the Swiss Federal i¼1 Institute for Forest, Snow and Landscape Research (WSL), Lastly, productivity in terms of volume growth who used the DAYMET model (Thornton et al. 1997)to 3 −1 −1 (m ∙ha ∙a ) was chosen to represent the supporting interpolate daily climate data to a resolution of 1 ha for the services that underpin the provision of all other ES. time period 1930–2006. The same institution also Table 1. Artificial gradient of long-term monthly mean temperatures, with series T1 based on the minimum values of an actual environmental gradient located in Switzerland. For series T2–T10 the mean monthly temperature was increased stepwise by one degree. [°C] January February March April May June July August September October November December T1 −5 −6 −6 −4 −23 7 9 9 7 4 −2 T2 −4 −5 −5 −3 −14 8 10 10 8 5 −1 T3 −3 −4 −4 −2 0 5 9 11 11 9 6 0 T4 −2 −3 −3 −11 6 10 12 12 10 7 1 T5 −1 −2 −2 0 2 7 11 13 13 11 8 2 T6 0 −1 −1 1 3 8 12 14 14 12 9 3 T7 1 0 0 2 4 9 13 15 15 13 10 4 T8 2 1 1 3 5 10 14 16 16 14 11 5 T9 3 2 2 4 6 11 15 17 17 15 12 6 T10 4 3 3 5 7 12 16 18 18 16 13 7 International Journal of Biodiversity Science, Ecosystem Services & Management 189 Table 2. Artificial gradient of long-term monthly precipitation sums, with series P1 based on the minimum values of an actual environmental gradient located in Switzerland. For series P2–P10 the monthly precipitation sum was increased stepwise by 2 cm per month (24 cm per year). [cm] January February March April May June July August September October November December P1 3 3 3 3 3 4 5 5 4 3 4 4 P2 5 5 5 5 5 6 7 7 6 5 6 6 P3 7 7 7 7 7 8 9 9 8 7 8 8 P4 9 9 9 9 9 10 11 11 10 9 10 10 P5 11 11 11 11 11 12 13 13 12 11 12 12 P6 13 13 13 13 13 14 15 15 14 13 14 14 P7 15 15 15 15 15 16 17 17 16 15 16 16 P8 17 17 17 17 17 18 19 19 18 17 18 18 P9 19 19 19 19 19 20 21 21 20 19 20 20 P10 21 21 21 21 21 22 23 23 22 21 22 22 provided several scenarios of climate change for the gra- conditions were changed linearly over a period of dient (Zimmermann et al. 2013), from which the trends 100 years, until they reached the desired magnitude of and ranges of change were adopted. change. Simulations were continued for another 1400 For the scenarios of climate change, I first increased years, to again reach a quasi-equilibrium with the new only the mean of the temperature distribution four times climatic conditions. Values for biomass, productivity, and in steps of 1 degree Celsius (+0.5 in the winter and spring structural diversity were sampled from these two months, and +1.5 in the summer and fall ones; and thus equilibria. on average +1 per annum), then only the standard devia- tion of the distribution four times in steps of 0.5, and then 2.4. The forest model mean and standard deviation of the temperature distribu- tion combined (Figure 1). Second, I changed the precipi- The concept of the model ForClim (Bugmann 1996), tation distribution in steps of 5% (+5% in the winter and which was used in this study, is based on the theory of spring months and –5% in the summer and fall ones; and forest succession (Watt 1947), which states that stand thus on average ±0), then only the standard deviation in dynamics as a whole can be represented by the individual steps of 0.5, and then both combined. Lastly I changed succession of trees on small patches of land (Shugart both temperature and precipitation distributions simulta- 1984). Based on this reasoning, ForClim simulates estab- neously, first both means, then both standard deviations, lishment, growth, and mortality of trees on typically 50– and lastly all together (Table 3), thus resulting in 36 200 patches, each with the size of 800 m , small enough to scenarios of climate change, and 10,800 simulations be potentially dominated by a single tree. overall, when applied to each of the 100 site files and The biotic processes simulated in ForClim are driven each of the three species scenarios. Each simulation was by the abiotic environment, represented for each site by initialized under current climate conditions and run for the soil variables water-holding capacity and available 1500 years to reach a quasi-equilibrium. Then climatic nitrogen, and the monthly weather variables long-term Δμ Δσ Δμ + Δσ T T T T current distribution scenarios of change 1 2 3 4 0 5 10 −5 0 5 10150 5 101520 Mean annual temperature [°C] Figure 1. Presumed changes in the distribution of temperatures (adapted from IPCC 2012), with a) changes in the mean (μ), b) changes in the variability (σ), and c) changes in both. Numbers 1–4 refer to the different degrees of change (+1 to +4°C, +0.5 to +2 standard deviations). 190 L. Rasche Table 3. Simulated changes in seasonal climate variables 2012). A detailed mathematical description of the model (μ average, σ standard deviation, T Temperature, and P precipita- can be found in Bugmann (1994). tion). The following combinations were considered: Δμ , Δσ , T T (Δμ + Δσ ), Δμ , Δσ ,(Δμ + Δσ ), (Δμ + Δμ ), (Δσ + Δσ ), T T P P P P T P T P and (Δμ + Δσ + Δμ + Δσ ). T T P P 3. Results 12 34 The importance of changes in climate parameter variability can only be determined in relation to the impacts of Δμ (°C) Winter/spring +0.5 +1.5 +2.5 +3.5 changes in the mean. For this reason, both cases are Summer/fall +1.5 +2.5 +3.5 +4.5 discussed in the following. Δσ (°C) Winter/spring +0.5 +1 +1.5 +2 Summer/fall +0.5 +1 +1.5 +2 Δμ (%) Winter/spring +5 +10 +15 +20 Summer/fall −5 −10 −15 −20 3.1. Diverse forests Δσ (%) Winter/spring +5 +10 +15 +20 An increase in mean temperature has mainly positive Summer/fall +5 +10 +15 +20 effects on all three ES in diverse forests (Figure 2). The main cause of this effect is a gain in productivity on currently temperature-limited sites through a general mean of temperature and its standard deviation, long-term lengthening of growing season and a shift in species mean sum of precipitation and its standard deviation, and composition towards more productive species. With cross-correlation of both temperature and precipitation. increasing severity of change, however, negative effects Monthly values are necessary to account for different increase in number on warm and dry sites (> 9°C, growing season lengths of deciduous and evergreen spe- < 116 cm) when considering productivity and structural cies. Actual values for each month are generated by sto- diversity. A shift in mean precipitation results in a rising chastically sampling from the respective long-term number of negative effects on biomass and productivity on statistics of monthly temperature means and precipitation the drier sites. Structural diversity is almost insensitive to sums. A different realization is generated for each of the shifts in mean precipitation. In the combination simula- 500 patches simulated per site. This implementation facil- tions of both mean temperature and precipitation changes, itates an easy manipulation of the mean of the distribution, the effects of temperature shift dominate for structural the standard deviation, or both, but it should also be diversity and productivity, whereas on the drier sites of mentioned that it makes it impossible to portray very the gradient the response of biomass is directed by the short, very extreme events. Only extreme years or clusters negative influence of the precipitation shift. of extreme years can be simulated. Considering the mod- An increase in temperature variability mainly has el’s time step of one year and the long simulation time, negative effects, and most strongly influences structural however, this limitation can be considered as negligible. diversity, to a lesser degree productivity, and, with increas- The variables of the abiotic environment are translated ing severity of change, biomass on the drier sites. An into bioclimatic variables that influence establishment and increase in precipitation variability negatively influences growth of trees. Establishment is determined by minimum biomass on the drier half of the gradient and productivity winter temperature, growing season temperature, growing on the warmer sites, increasing in extent with increasing season soil moisture, and light availability, all of which severity of change, and positively influences structural have to be favourable for establishment to succeed. Tree diversity on the drier sites, and productivity on the very growth is simulated according to Moore’s(1989) carbon dry (44 cm) sites. A closer analysis of the results revealed budget approach, modified by Risch et al. (2005), Didion that the dominant species on these sites was simulated to et al. (2009), and Rasche et al. (2012), where an optimal be Pinus sylvestris L., the most drought-tolerant species growth rate is reduced based on light and nitrogen avail- currently parameterized for ForClim. As the trees were ability, growing season temperature, growing season soil already drought stressed in the simulations under current moisture, and crown length. The resulting volume growth climate, their productivity was not significantly reduced is allocated dynamically to height and diameter growth, under the scenario of change. It actually increased when based on available light and the shade tolerance of the the greater variability in precipitation occasionally pro- trees. Tree mortality is triggered by an age-related and a duced values above the species’ drought tolerance. In the stress-induced component. combination simulations of temperature and precipitation The model was tested and validated against a variety variability changes, the results indicate a true combination of empirical data (e.g. national forest inventories, long- of both effects, with neither temperature nor precipitation term growth, and yield research plots), in terms of both dominating. general applicability and precision of single processes, and In combination simulations of mean and variability showed good results in terms of simulated species compo- changes, the effect of a shift in the mean dominates the sition, growth rates, biomass, harvested timber, diameter effect of a shift in variability, at least concerning produc- and height distributions, and several other tree-based vari- tivity and structural diversity. Biomass, on the other hand, ables (see, e.g., Bugmann & Cramer 1998; Wehrli et al. is strongly influenced by an increase in the variability of 2007; Didion et al. 2009; Heiri 2009; Rasche et al. 2011, precipitation, and consequently the effects of this change International Journal of Biodiversity Science, Ecosystem Services & Management 191 Annual precipitation sum [cm] ΔT ΔP ΔT+ ΔP 12 3 4 12 3 4 12 3 4 biomass 61 8 10214 –1 –1 0 [t·ha ·a ] –40 0 +40 –1 –1 [Δ t·ha ·a ] 92 productivity 61 8 10214 0 0.45 3 –1 –1 [m ·ha ·a ] –0.5 0 +0.5 3 –1 –1 [Δ m ·ha ·a ] structural diversity 61 8 10214 0 4 [–] 236 –4 0 +4 [–] 5171 9 13 61 810214 5 7 9 11 13 6 8 10 12 14 5 7 9 11 13 61 8 10214 Mean annual temperature [°C] Figure 2. (colour online) Simulated current levels of ES (small plots on left-hand side) and significant changes in ES levels (main plots) on multi-species sites. Each square represents 100 sites with different environmental conditions, ranging from cold and dry (5°C, 44 cm) to warm and moist (14°C, 260 cm). Simulated were changes in the mean (μ), the variability (σ), and a combination of both of only the temperature distribution (T), only the precipitation distribution (P), and both simultaneously. Numbers on the column tops refer to the severities of change (see Table 3). White squares did not experience a significant change (Student’s t-test), green squares indicate significant positive changes, and red squares with points indicate significant negative changes. Δμ Δμ Δμ + Δσ Δσ Δμ + Δσ Δσ Δμ + Δσ Δσ Δμ 192 L. Rasche dominate the results in the combination simulations. In warmer sites (Figure 4). The negative changes on the both cases, there is no cumulation of the effects of changes warm sites are due to failed establishments caused by a in mean and variability. lack of chilling in winter, which indicates that the distribu- tion boundary of P. abies was reached. The effects on structural diversity are overwhelmingly negative, leaving 3.2. F. sylvatica monocultures almost no site untouched. A shift in mean precipitation has An increase in mean temperature has mainly positive effects few impacts; however, on the dry sites (<92 cm) there are on all three ES, with the exception of biomass and produc- increasingly negative impacts for all three ES, and occa- tivity on the driest sites warmer than 8°C (Figure 3). A shift sional positive impacts on structural diversity. Owing to in mean precipitation has only few significant effects on all the overall low impact precipitation changes have, in the three ES, but those are almost exclusively negative: on dry combination simulations of both temperature and precipi- sites (68 cm and lower) values of ES decrease significantly, tation changes the changes in mean temperature dominate affecting more sites with an increasing severity of change. the overall effects. In the combination simulations of both temperature and An increase in temperature variability has negative precipitation mean changes the effects are dominated by impacts on the very cold sites (<6°C) for biomass and the changes in mean temperature. structural diversity and an increasingly positive effect on An increase in temperature variability has positive warm sites (>10°C) for all three ES. An analysis of the effects on the colder sites of the gradient (<8°C), as more results revealed that lower simulated maximum winter years reaching the species-specific required minimal num- temperatures were responsible for this, as they allowed ber of degree days for growth are simulated, thus resulting higher establishment rates, and as a consequence higher in significant positive changes. Negative impacts are more values of ES. This shows the importance of extreme abundant, however, beginning at sites with average annual events on the lower end of the temperature distribution, temperatures of 8°C and continuing towards sites with as successful establishment and sapling growth of P. abies higher temperatures as the severity of change increases. trees have to occur in years colder than the future norm. An analysis of simulation results revealed that this slightly An increase in precipitation variability has positive effects surprising pattern could be explained by a higher abundance on the very dry sites (44 cm) for all three ES, showing that of winters with temperatures below what can be tolerated on sites where drought is severe already in the beginning, by F. sylvatica seedlings for successful establishment. The a change in variability results in more years where drought bigger the variability was simulated to be, the warmer the is not as extreme and thus growth is enhanced, and occa- sites that occasionally experienced very cold winters tended sionally positive impacts on the rest of the sites for struc- to get. Colder sites were not concerned, as winter tempera- tural diversity. Negative effects are overall more abundant, tures already influenced establishment numbers under the however, growing in number as the severity of change current climate. An increase in precipitation variability increases and affecting the dry/warm sites for biomass negatively influences all three ES on the drier sites, increas- and productivity. When both temperature and precipitation ing inextentwithincreasingstrength, as the drought toler- variabilities are changed simultaneously, precipitation ance of F. sylvatica trees is exceeded more often. On the variability dominates the response of biomass, whereas very dry sites (<68 cm), however, an increase is simulated temperature variability dominates the response of produc- for all three ES, showing that on sites where drought is tivity and structural diversity. severe already in the beginning, a change in variability When both the mean and the standard deviation of the results in more years where drought is not as extreme, and climate variable distributions are changed, the effect of a shift thus growth is enhanced. This shows the importance of in mean dominates the effect of a shift in variability, at least extreme events on the upper end of the precipitation dis- concerning productivity and structural diversity. Biomass is tribution. In the combination simulations of temperature and again more strongly influenced by an increase in variability. precipitation variability change, the influence of tempera- On the warmer sites (>10°C), a higher variability in tempera- ture variability dominates the results, but the positive ture can slightly ameliorate the negative effects a higher changes on very dry sites due to precipitation variability mean temperature has on the level of ES. changes are also visible. In combination simulations of mean and variability 3.4. Comparison of the impact on different ES changes, a shift in the mean dominates over a shift in variability; yet if the effect of increasing variability is In diverse stands, productivity and structural diversity were negative (e.g. for biomass), this negative effect shows negatively affected mostly on hot sites by an increase in as well. temperature variability and positively by an increase in mean temperature on colder ones. Biomass, on the one hand, reacted most strongly to a change in precipitation. In 3.3. P. abies monocultures beech monocultures, ES were affected in the same broad An increase in mean temperature has positive effects for patterns. In spruce monocultures, biomass and productivity biomass and productivity on the colder and moister part of reacted similarly as well, yet biomass clearly experienced a the gradient (<9°C, >68 cm) and negative effects on the higher impact through precipitation changes than International Journal of Biodiversity Science, Ecosystem Services & Management 193 Annual precipitation sum [cm] ΔT ΔP ΔT + ΔP 12 3 4 1 2 34 1 2 34 biomass 260 164 212 116 164 68 61 8 10214 –1 –1 0 [t·ha ·a ] –40 0 +40 –1 –1 [Δ t·ha ·a ] productivity 61 8 10214 0 3 –1 –1 0.45 [m ·ha ·a ] –0.5 0 +0.5 3 –1 –1 92 [Δ m ·ha ·a ] structural diversity 61 8 10214 0 [–] 236 –4 0 +4 [–] 5171 9 13 61 810214 5171 9 13 61 810214 5171 9 13 61 810214 Mean annual temperature [°C] Figure 3. (colour online) Simulated current levels of ES (small plots on left-hand side) and significant changes in ES levels (main plots) in F. sylvatica monocultures. See caption Figure 2 for more details. Δμ Δσ Δμ Δμ Δμ + Δσ Δσ Δμ + Δσ Δμ + Δσ Δσ 194 L. Rasche Annual precipitation sum [cm] ΔT ΔP ΔT + ΔP 12 3 4 1 234 12 3 4 biomass 61 8 10214 –1 –1 0 [t·ha ·a ] –40 0 +40 –1 –1 92 [Δ t·ha ·a ] productivity 61 8 10214 0 0.45 3 –1 –1 [m ·ha ·a ] 236 –0.5 0 +0.5 3 –1 –1 [Δ m ·ha ·a ] structural diversity 61 8 10214 0 [–] 4 –4 0 +4 [–] 5171 9 1 3 61 810214 5171 9 13 61 810214 5171 9 13 61 810214 Mean annual temperature [°C] Figure 4. (colour online) Simulated current levels of ES (small plots on left-hand side) and significant changes in ES levels (main plots) in P. abies monocultures. See caption Figure 2 for more details. Δμ Δσ Δμ Δμ + Δσ Δσ Δμ Δμ + Δσ Δμ + Δσ Δσ International Journal of Biodiversity Science, Ecosystem Services & Management 195 productivity. Structural diversity, on the other hand, was temperature and precipitation changed, as even today, much more sensitive to changes in mean temperature than several P. abies stands are at their natural limit in Europe the other two ES. Nearly every site experienced a significant (Kölling et al. 2009; Yousefpour et al. 2010), and are loss in structural diversity, even on sites where biomass and pushed beyond this threshold in the simulations of climate productivity were simulated to rise. An analysis of the dia- change. Although drought is an important factor in this meter distributions before and after climate change revealed development, failed establishment on very warm sites is that a reduction of trees in the smaller diameter classes was simulated to be the more important one overall (cf. Young responsible for this. On formerly temperature-limited sites, & Hanover 1978; Dumais & Prévost 2007). Only natural trees grew to be bigger and prohibited larger establishment succession was simulated in this study, but other studies numbers through shading, and on drought-limited sites lack suggest that planted trees may be affected as well, going of precipitation was the cause. so far as to warn that forest management may in future require repeated replanting of sensitive seedlings, regard- less of how well-adapted they might be as adults (Fuhrer 4. Discussion et al. 2006). However, if only the strength of extreme frost The results show that changes in variability yield a lower events decrease without too high winter temperatures hin- number of significant changes in forest ES levels than dering establishment, an increase in productivity is simu- lated, and observed (e.g. Rammig et al. 2010). changes in the mean, regardless of forest type. Yet with the exception of P. abies monocultures, these changes are Some of the ‘best case’ scenarios are of course highly projected to be almost exclusively negative and to increase unlikely. It is certain that mean temperatures have already in number the greater the change in variability becomes. risen and will further rise in future (IPCC 2007a), so that a Negative changes are also not exclusively focused on the change in variability only is improbable. It is also unlikely distributional edges of the species; in the diverse forest that only a shift in mean temperature will occur. The tem- scenario, losses in structural diversity and productivity can peratures of Europe’s2003 heat summer, forexample, can be observed over a broad gradient of environmental condi- only be explained by a combined shift of the statistical tions. Almost the exact opposite is true for changes in the distribution towards warmer temperatures and an increase mean of climate parameter distributions: A high number of in variability (Schär et al. 2004). In Austria, however, tem- significant and often positive changes are projected to take perature variability has evolved independently of mean tem- place, which are focused mainly on currently temperature- perature in the last 140 years and decadal trends of variability limited sites, indicating a gain in growth potential due to a hardly exceed + /- 1 standard deviation (Hiebl & Hofstätter lengthening of the growing season and shifts towards more 2012). Precipitation patterns are subject to even more uncer- productive species in diverse stands (cf. e.g. Eggers et al. tainty than temperature. It is assumed that there will be a 2008; Albert & Schmidt 2010; Lindner et al. 2010). greatly increased risk of extreme precipitation events The implication of this result for impact analyses (Coumou & Rahmstorf 2012), yet which changes in the regarding forest ES is that the overall trend in ES provi- rainfall distribution will be responsible for this is not sure. sion is well represented even if only the means of climate Studies of the past climate indicate that the shape parameter of the precipitation distribution may remain relatively stable, parameter distributions are changed, but that at the phy- whereas the scale parameter varies spatially and temporally siological limits of species, changes in variability have to (Groisman et al. 1999; Alexander et al. 2006). be accounted for in order to realistically assess impacts (cf. In addition to changes in temperature and precipita- Zimmermann et al. 2009). tion, the level of CO in the atmosphere is projected to rise This sensitivity study also highlights possible best and worst case scenarios for the level of ES provided by stands (IPCC 2007a), a feature not considered in this study to of different species. For the ES produced by diverse and facilitate the exact attribution of the results to either tem- pure F. sylvatica stands, it would be fortuitous if climate perature or precipitation effects, and thus to the underlying change mainly involved a change in the mean of the growth processes. If this factor and other factors such as temperature distribution (cf. Bergh et al. 2003; Albert & pathogens or mechanical disturbances were included, Schmidt 2010), and less favourable if a change in varia- results may look different (cf. Dale et al. 2001; Seidl bility was involved – temperature variability for structural et al. 2008; Bugmann & Bigler 2011). It should also be diversity, and variability of the precipitation distribution mentioned that neither heat waves nor extremes in preci- for biomass (cf. Bugmann & Pfister 2000) and productiv- pitation would most likely pose the most serious risk to ity. However, there is some evidence that species’ genetic forests, but winter storms will (Fuhrer et al. 2006). diversity – a factor not included in the ForClim model – However, wind is not conclusively coupled to climate may be sufficient for a good performance even under change (Coumou & Rahmstorf 2012), and was therefore extreme conditions (Beierkuhnlein et al. 2011). not incorporated into this study. In P. abies stands, biomass and productivity are pro- jected to benefit most from a change in temperature varia- 5. Conclusions bility, whereas structural diversity would benefit most from a change in both temperature and precipitation varia- For the study of climate change impacts on the levels of bility. The worst case would be if both the mean of forest ES in the temperate zone of Europe, uncertainty 196 L. Rasche about climate parameter variability is of secondary impor- Bugmann H, Bigler C. 2011. Will the CO fertilization effect in forests be offset by reduced tree longevity? Oecologia. tance. The trend in the provision of ES is well represented 165:533–544. doi:10.1007/s00442-010-1837-4 with scenarios of mean climate parameter changes only, Bugmann H, Cramer W. 1998. Improving the behaviour of forest but on moisture-limited and heat-stressed sites impacts of gap models along drought gradients. Forest Ecol Manage. changes in variability gain in importance. Most of these 103:247–263. doi:10.1016/S0378-1127(97)00217-X impacts are negative, and can be observed not only in Bugmann H, Pfister C. 2000. Impacts of interannual climate variability on past and future forest composition. Reg stands populated with a single species already at its phy- Environ Change. 1:112–125. doi:10.1007/s101130000015 siological limit, but also in diverse stands, indicating that Bugmann H, Solomon AM. 2000. Explaining forest composition the current species diversity of temperate forests in Europe and biomass across multiple biogeographical regions. Ecol is not enough to retain the current levels of ES provision. Appl. 10:95–114. doi:10.1890/1051-0761(2000)010[0095: The sensitivity to the changes is different for different ES, EFCABA]2.0.CO;2 Carvalho-Ribeiro SM, Lovett A. 2011. Is an attractive forest also with biomass, for instance, being more influenced by considered well managed? Public preferences for forest cover changes in precipitation, and productivity and structural and stand structure across a rural/urban gradient in northern diversity by changes in temperature. Changes in tempera- Portugal. Forest Policy Econ. 13:46–54. doi:10.1016/j. ture variability ameliorate the negative impacts on struc- forpol.2010.09.003 tural diversity, whereas they mostly aggravate impacts on Cassman KG, Wood S. 2005. Cultivated systems. In: Millenium Ecosystem Assessment Board, editor. Millenium ecosystem the other ES. assessment: global ecosystem assessment report on conditions The results underline the need to develop adaptation and trends. Washington (DC): Island Press; p. 745–787. measures tailored to the sites and ES of interest, and also Collier R, Fellows JR, Adams SR, Semenov M, Thomas B. 2008. show that although trends are sufficiently described by Vulnerability of horticultural crop production to extreme changes in the mean of climate variables, for the fine- weather events. Aspects Appl Biol. 88:3–14. Coumou D, Rahmstorf S. 2012. A decade of weather extremes. tuning of measures the consideration of climate variability Nat Clim Change. 2:491–496. is required. Dale VH, Joyce LA, McNulty S, Neilson RP, Ayres MP, Flannigan MD, Hanson PJ, Irland LC, Lugo AE, Peterson CJ, et al. 2001. Climate change and forest disturbances. Bioscience. 51:723–734. doi:10.1641/0006-3568(2001)051 Acknowledgements [0723:CCAFD]2.0.CO;2 Dale VH, Tharp ML, Lannom KO, Hodges DG. 2010. Modeling The author is grateful to Harald Bugmann (ETH Zurich), Kerstin transient response of forests to climate change. Sci Total Jantke (University of Hamburg), the editors of this journal, and Environ. 408:1888–1901. doi:10.1016/j.scitotenv.2009.11.050 two anonymous reviewers for comments on the manuscript. Didion M, Kupferschmid A, Lexer MJ, Rammer W, Seidl R, Bugmann H. 2009. Potentials and limitations of using large- scale forest inventory data for evaluating forest succession models. Ecol Model. 220:133–147. doi:10.1016/j. References ecolmodel.2008.09.021 Dumais D, Prévost M. 2007. Management for red spruce con- Albert M, Schmidt M. 2010. Climate-sensitive modelling of site- servation in Québec: the importance of some physiological productivity relationships for Norway spruce (Picea abies and ecological characteristics – a review. Forest Chron. (L.) Karst.) and common beech (Fagus sylvatica L.). Forest 83:378–391. doi:10.5558/tfc83378-3 Ecol Manage. 259:739–749. doi:10.1016/j. Eggers J, Lindner M, Zudin S, Zaehle S, Liski J. 2008. Impact of foreco.2009.04.039 changing wood demand, climate and land use on European Alexander LV, Zhang X, Peterson TC, Caesar J, Gleason B, Tank forest resources and carbon stocks during the 21st century. AMGK, Haylock M, Collins D, Trewin B, Rahimzadeh F, Glob Change Biol. 14:2288–2303. doi:10.1111/j.1365- et al. 2006. Global observed changes in daily climate 2486.2008.01653.x extremes of temperature and precipitation. J Geophys Res- Fuhrer J, Beniston M, Fischlin A, Frei C, Goyette S, Jasper K, Atmos. 111. doi:10.1029/2005JD006290 Pfister C. 2006. Climate risks and their impact on agriculture Beierkuhnlein C, Thiel D, Jentsch A, Willner E, Kreyling J. and forests in Switzerland. Clim Change. 79:79–102. 2011. Ecotypes of European grass species respond differently doi:10.1007/s10584-006-9106-6 to warming and extreme drought. J Ecol. 99:703–713. Groisman PY, Karl TR, Easterling DR, Knight RW, Jamason PF, doi:10.1111/j.1365-2745.2011.01809.x Hennessy KJ, Suppiah R, Page CM, Wibig J, Fortuniak K, Bergh J, Freeman M, Sigurdsson B, Kellomäki S, Laitinen K, et al. 1999. Changes in the probability of heavy precipitation: Niinistö S, Peltola H, Linder S. 2003. Modelling the short- important indicators of climatic change. Clim Change. term effects of climate change on the productivity of selected 42:243–283. doi:10.1023/A:1005432803188 tree species in Nordic countries. Forest Ecol Manage. Heiri C. 2009. Stand dynamics in Swiss forest reserves: an 183:327–340. doi:10.1016/S0378-1127(03)00117-8 analysis based on long-term forest reserve data and dynamic Bigler C, Bräker OU, Bugmann H, Dobbertin M, Rigling A. modeling [dissertation]. Zürich: Swiss Federal Institute of 2006. Drought as an inciting mortality factor in Scots pine Technology, Diss. No. 18388. stands of the Valais, Switzerland. Ecosystems. 9:330–343. Hiebl J, Hofstätter M. 2012. No increase in multi-day tempera- doi:10.1007/s10021-005-0126-2 ture variability in Austria following climate warming. Clim Bugmann H. 1994. On the ecology of mountainous forests in a Change. 113:733–750. doi:10.1007/s10584-011-0389-x changing climate: a simulation study. Zürich: Swiss Federal Honnay O, Verheyen K, Butaye J, Jacquemyn H, Bossuyt B, Institute of Technology. Hermy M. 2002. Possible effects of habitat fragmentation Bugmann H. 1996. A simplified forest model to study species and climate change on the range of forest plant species. Ecol composition along climate gradients. Ecology. 77:2055– Lett. 5:525–530. doi:10.1046/j.1461-0248.2002.00346.x 2074. doi:10.2307/2265700 International Journal of Biodiversity Science, Ecosystem Services & Management 197 IPCC. 2007a. Climate change 2007: synthesis report. Seidl R, Rammer W, Jäger D, Lexer MJ. 2008. Impact of bark Contributions of working groups I, II and III to the fourth beetle (Ips typographus L.) disturbance on timber production assessment report of the intergovernmental panel on climate and carbon sequestration in different management strategies change. Geneva: IPCC. under climate change. Forest Ecol Manage. 256:209–220. IPCC. 2007b. Contribution of working group I to the fourth doi:10.1016/j.foreco.2008.04.002 assessment report of the intergovernmental panel on climate Shugart H. 1984. A theory of forest dynamics: the ecological change. Cambridge: CU Press. implications of forest succession models. New York (NY): IPCC. 2012. Managing the risks of extreme events and disasters Springer. to advance climate change adaptation. A special report of Shugart HH, Sedjo RA, Sohngen BL. 2003. Forests & global working groups I and II of the intergovernmental panel on climate change: potential impacts on US forest resources. climate change. New York (NY): CU Press. Arlington (TX): Pew Center on Global Climate Change. Kellomaki S, Peltola H, Nuutinen T, Korhonen KT, Strandman Shvidenko A, Barber CV, Persson R. 2005. Forest and woodland H. 2008. Sensitivity of managed boreal forests in Finland to systems. In: de los Angeles M, Sastry C, editors. Ecosystems climate change, with implications for adaptive management. and human well-being: current state and trends. Volume 1, Phil Trans R Soc B: Biol Sci. 363:2339–2349. doi:10.1098/ Washington (DC): Island Press; p. 585–621. rstb.2007.2204 Solomon AM. 1986. Transient response of forests to CO - Kölling C, Knoke T, Schall P, Ammer C. 2009. Überlegungen induced climate change: simulation modeling experiments zum Risiko des Fichtenanbaus in Deutschland vor dem in eastern North America. Oecologia. 68:567–579. Hintergrund des Klimawandels. Forstarchiv. 80:42–54. doi:10.1007/BF00378773 Lindner M, Maroschek M, Netherer S, Kremer A, Barbati A, Staudhammer CL, Lemay VM. 2001. Introduction and evaluation Garcia-Gonzalo J, Seidl R, Delzon S, Corona P, Kolström M, of possible indices of stand structural diversity. Can J Forest et al. 2010. Climate change impacts, adaptive capacity, and Res. 31:1105–1115. doi:10.1139/x01-033 vulnerability of European forest ecosystems. Forest Ecol Thornton PE, Running SW, White MA. 1997. Generating sur- Manage. 259:698–709. doi:10.1016/j.foreco.2009.09.023 faces of daily meteorological variables over large regions of Luo Q. 2011. Temperature thresholds and crop production: a complex terrain. J Hydrol. 190:214–251. doi:10.1016/S0022- review. Clim Change. 109:583–598. doi:10.1007/s10584- 1694(96)03128-9 011-0028-6 Trenberth KE. 2012. Framing the way to relate climate extremes Millennium Ecosystem Asssessment. 2003. Ecosystems and to climate change. Clim Change. 115:283–290. doi:10.1007/ human well-being: a framework for assessment. s10584-012-0441-5 Washington (DC): Island Press. ISBN 1559632283. Watt AS. 1947. Pattern and process in the plant community. J Monteith JL. 2007. Climatic variation and the growth of crops. Q J Ecol. 35:1–22. doi:10.2307/2256497 Roy Meteorol Soc. 107:749–774. doi:10.1002/qj.49710745402 Wehrli A, Weisberg PJ, Schönenberger W, Brang P, Bugmann H. Moore AD. 1989. On the maximum growth equation used in 2007. Improving the establishment submodel of a forest forest gap simulation models. Ecol Model. 45:63–67. patch model to assess the long-term protective effect of doi:10.1016/0304-3800(89)90100-2 mountain forests. Eur J Forest Res. 126:131–145. Rammig A, Jönsson AM, Hickler T, Smith B, Bärring L, Sykes doi:10.1007/s10342-006-0142-6 M. 2010. Impacts of changing frost regimes on Swedish Young E, Hanover JW. 1978. Effects of temperature, nutrient and forests: incorporating cold hardiness in a regional ecosystem moisture stresses on dormancy of blue spruce seedlings model. Ecol Model. 221:303–313. doi:10.1016/j. under continuous light. Forest Sci. 24:458–467. ecolmodel.2009.05.014 Yousefpour R, Hanewinkel M, Le Moguédec G. 2010. Rasche L, Fahse L, Zingg A, Bugmann H. 2011. Getting a virtual Evaluating the suitability of management strategies of forester fit for the challenge of climatic change. J Appl Ecol. pure Norway spruce forests in the Black Forest area of 48:1174–1186. doi:10.1111/j.1365-2664.2011.02014.x Southwest Germany for adaptation to or mitigation of cli- Rasche L, Fahse L, Zingg A, Bugmann H. 2012. Enhancing gap mate change. Environ Manag. 45:387–402. doi:10.1007/ model accuracy by modeling dynamic height growth and s00267-009-9409-2 dynamic maximum tree height. Ecol Model. 232:133–143. Zimmermann NE, Schmatz D, Psomas A. 2013. Climate change doi:10.1016/j.ecolmodel.2012.03.004 scenarios to 2100 and implications for forest management. Risch AC, Heiri C, Bugmann H. 2005. Simulating structural forest Adapting to climate change in European forests – results of patterns with a forest gap model: a model evaluation. Ecol the MOTIVE project. Sofia: Pensoft Publishers; p. 9–14. Model. 181:161–172. doi:10.1016/j.ecolmodel.2004.06.029 Zimmermann NE, Yoccoz NG, Edwards TC, Meier ES, Thuiller Schär C, Vidale PL, Lüthi D, Frei C, Häberli C, Liniger MA, W, Guisan A, Schmatz DR, Pearman PB. 2009. Climatic Appenzeller C. 2004. The role of increasing temperature extremes improve predictions of spatial patterns of tree spe- variability in European summer heatwaves. Nature. cies. Proc Natl Acad Sci USA. 106:19723–19728. 427:332–336. doi:10.1038/nature02300 doi:10.1073/pnas.0901643106
International Journal of Biodiversity Science, Ecosystem Services & Management – Taylor & Francis
Published: Jul 3, 2014
Keywords: environmental gradient; monoculture; sensitivity study; structural diversity; temperate forest
Access the full text.
Sign up today, get DeepDyve free for 14 days.