Regional Environmental Change

, Volume 17, Issue 1, pp 17–32 | Cite as

Ecosystem service provision, management systems and climate change in Valsaín forest, central Spain

  • Marta Pardos
  • Susana Pérez
  • Rafael Calama
  • Rafael Alonso
  • Manfred J. Lexer
Original Article

Abstract

This study addresses the impact of climate change and management approach on the provision of four ecosystem services (ES) (timber production, protection against gravitational hazards, carbon sequestration and biodiversity) in Valsaín forest in central Spain. The hybrid forest patch model PICUS v1.6 was used to simulate the development of 24 representative stand types over 100 years (2010–2110) in a full factorial simulation experiment combining three management regimes [“business as usual” management (BAU) and two alternatives to BAU (AM1 and AM2)], a no-management scenario (NOM) and six climate scenarios (historic climate represented by the period 1961–1990 and five transient climate change scenarios). Simulations indicated relatively small differences as regards the impact of the different management alternatives (BAU, AM1 and AM2) on the provision of ES as well as a clear improvement in biodiversity, protection and carbon storage under the no-management regime (NOM). Although timber production indicators were the most sensitive to climate change scenarios, biodiversity-related indicators responded fastest to the management regimes applied. Indicators of protection against rockfall and landslides were affected by both management and climate change. The results indicate substantial vulnerability of ES provisioning under the more extreme climate change scenarios at low elevations (1250 m). At higher elevations, the productivity of Scots pine stands may show a moderate decrease or increase, depending on the climate change scenario.

Keywords

Timber production Biodiversity Carbon storage Forest modelling Climate change vulnerability Resilience 

Introduction

Forests provide a multitude of ecosystem services (ES) (e.g. MEA 2005; Ojea et al. 2012). A key issue in sustainable forest management is how to maintain the functioning of forests and how to ensure thereby the continued long-term provision of ES (Fürst et al. 2013; Ray et al. 2014). The structure and processes of ecosystems and therefore the future provisioning of ES will be affected not only by climate change, but also by increasing demands for bioenergy production and other ES provision along with strong interests in nature conservation which lead to the setting aside of forests as strict reserves. Decision-making about future forest management will thus involve substantial uncertainties (Deal and White 2012).

The latest climate change scenario projections for Europe suggest a temperature increase and changes in the precipitation regime that will be especially evident in the Mediterranean region (Lindner et al. 2010; Lindner and Calama 2013; Spathelf et al. 2014). Although there is still much uncertainty surrounding climate projections for the twenty-first century with respect to precipitation and extreme events such as storms and droughts, it is expected that exposure to climatic change in the Mediterranean region will be greater than in other regions of Europe (Lindner et al. 2010). Potential impacts on ES include a decline in the production of wood and non-wood products, changes in carbon sequestration rates and the net carbon balance of forests, shifts in vegetation composition and irrecoverable biodiversity losses (Lindner and Calama 2013). Adapting forests and forestry in the Mediterranean region to climate change is therefore a major challenge (Palahí et al. 2008). A prerequisite for successful adaptation is understanding and forecasting the consequences of climatic changes on the provision of ES (Allen et al. 2010). The way in which climate change will affect forests and related ES in mountain regions of the Mediterranean and sub-Mediterranean bioclimatic region has so far not been studied in detail.

The impact of climate change on the provision of ES will vary depending on the different types of forest ecosystem, their current structure and species composition and the ways in which they are managed. Planning and implementing multifunctional forest management is challenging because ES may be affected differently by changes in climate and management. Due to the synergies (e.g. carbon sequestration and timber production, Duncker et al. 2012) and trade-offs that exist between ES (e.g. fostering biodiversity may reduce timber production), planning and implementing multifunctional forest management can be challenging. Maximizing one ES may cause substantial declines in other ES (Wang and Fu 2013). A key question is how to enhance or sustain the provision of different ES while at the same time maintaining forest resilience in the face of projected climate change and related biotic and abiotic risks (Mason and Mencuccini 2014). Potential impacts of climate change are better understood in the case of wood production and carbon sequestration (e.g. Seidl et al. 2011a), while for other ES such as non-wood products and protection against gravitational hazards, much less knowledge is available (e.g. Maroschek et al. 2014). Until now, process-based forest ecosystem models have mainly been developed for predicting wood production and carbon sequestration (e.g. Poetzelsberger et al. 2015).

However, sustainable forest management needs to consider stakeholder demands for various ES in a transparent way. This calls for an expansion of forest management planning approaches including the development and use of predictive forest models which are capable of responding to climatic changes and management interventions, and providing a level of structural and functional resolution which allows us to consider numerous ES (Mäkelä et al. 2012). The hybrid forest patch model PICUS v.1.6 (e.g. Seidl et al. 2005; Maroschek et al. 2014) simulates stand-level (i.e. up to 20 ha in size) forest development with a structural resolution of individual trees under different climate change conditions and forest management regimes and provides related ecosystem service indicators (Seidl et al. 2011a, b).

The main objective of this work is to use simulation-based scenario analysis to explore the provision of four main ES of mountain forests in the Sierra de Guadarrama in Spain (timber production, protection against the gravitational hazards of rockfall and erosion, carbon sequestration, and biodiversity and nature conservation) according to different management regimes (“business as usual” management and three alternatives, including a no-management option) under historic climate and five transient climate change scenarios. The spatial scale of the analysis is stand level (10-20 ha).

The specific objectives are (1) to assess the combined effect of forest management and climate on the provision of ES; (2) to reveal synergies and trade-offs in the provision of these ES; and (3) to assess and understand the future functional resilience of these Mediterranean mountain forests.

Materials and methods

Valsaín forest

Valsaín is a public forest of 10,668 ha which is owned and managed by the Spanish National Parks Agency. It is located on the North-facing slopes of the Sierra de Guadarrama at elevations ranging from 1180 to 2130 m a.s.l. in the Central Mountain Range of Spain (40° 49′ N, 4° 1′ W). The climate in Valsaín is sub-Mediterranean, with mean monthly temperatures between 1.2 °C (January) and 18.3 °C (July). Mean annual temperature at 1500 m is 8.5 °C, average annual rainfall is 1275 mm, and precipitation between May and September is 651 mm. Moderately deep dystric cambisols and ferric luvisols have developed over acidic bedrock as major soil types. Sandy loam is the prevailing soil texture type.

According to the most recent forest inventory, eighty-two per cent of the forested area is occupied by pure, even-aged Pinus sylvestris forest (between 1400 and 1900 m a.s.l.), while mixed Pinus sylvestris-Quercus pyrenaica stands (10 % of the forested area) and pure Quercus pyrenaica stands (8 % of the forested area) occur at altitudes below 1400 m (Fig. 1). Above 1900 m, alpine shrubs are the prevailing vegetation type.
Fig. 1

Schematic representation of the representative stand types (RST) in Valsaín forest. RSTs differed in species composition, stand developmental stage, altitude and site type (plant available nitrogen N in kg ha−1 year−1) and water holding capacity WHC in mm). Ps: Pinus sylvestris, Qp: Quercus pyrenaica. For each altitudinal zone, species and mixture shares, WHC and N values (WHC/N) and number of RSTs are shown. The two sides of the triangle do not represent slope orientation. Mean annual temperature and mean annual precipitation in the altitudinal zones are based on the period 1961–1990

Timber harvesting and livestock grazing have taken place in Valsaín forest for centuries without any formal planning. This has led to overexploitation, lack of regeneration and severe degradation. Since the first management plans were adopted in 1889, even-aged forest management based on natural regeneration has been the common practice in Pinus sylvestris stands. Since it is one of the most productive Pinus sylvestris forests in Spain, the main emphasis of forest policies over many decades was timber production. This is reflected by the use of a uniform shelterwood system with a rotation of 120 years and a 20-year regeneration period since the 1940s. Beginning in the 1980s, the concept of multifunctionality has been progressively introduced and the silvicultural system was changed to a shelterwood group system. This avoids standwise shelterwood cutting and extends the regeneration period to 40 years, to assure sufficient natural regeneration. At the moment, this is the “business as usual” management approach for Pinus sylvestris stands. Quercus pyrenaica stands have traditionally been managed as coppice; however, due to falling demand for charcoal and fuelwood, management operations in these stands have frequently not been implemented. Based on Serrada et al. (2008) and in accordance with the current management plan for Valsaín, business as usual management is aimed at conversion to high forest.

The main ecosystem services currently demanded from Valsaín forest are timber production (in pine stands), fuelwood and biomass (in oak stands), carbon storage, biodiversity and habitat conservation (focusing on the imperial eagle, black stork, black vulture and wolf), recreation (Madrid, with 5 million inhabitants, is less than 60 min away by car), game (roe deer and wild boar), livestock grazing (cattle) and regulation of water run-off. Since 2013, the upper parts of the Valsaín forest (above 1875 m) are part of the “Sierra de Guadarrama” National Park, where management is highly regulated and even restricted in some areas.

Climate scenarios

A historic baseline climate (C0) and five transient climate change scenarios (C1C5), each consisting of a 100-year time series of daily temperature, precipitation, radiation and vapour pressure deficit, were used in the analysis. The baseline climate was generated using available daily instrumental data (1961–1990) from the meteorological station Puerto de Navacerrada (4.01°W, 40.79°N) and adjusted for different altitudes, slopes and aspects (Fig. 1) using the approach proposed by Thornton and Running (1999). The five climate change scenarios selected for the current study were based on regional climate model simulations from the ENSEMBLES project (Hewitt and Griggs 2004; www.ensembleseu.org) and meant to represent the range of potential future climates for the region. The climate change scenario data were statistically downscaled by means of quantile mapping. Each of the downscaled climate change scenarios is represented by time series of 100 years (2010–2110) with daily data for temperature, precipitation, water vapour deficit and solar radiation. For details on the downscaling approach, see Bugmann et al. (2015). In all climate change scenarios, temperature increased (+3.7 °C in C1, +3.8 °C in C2, +3.9 °C in C3, +4.6 °C in C4 and +5.9 °C in C5), summer precipitation (May to September) decreased in all scenarios (−25 % in C1, −30 % in C2, −24 % in C3, −38 % in C4 and −58 % in C5), while changes in annual precipitation were not as drastic (−17 % in C1, −10 % in C2, +4 % in C3, −11 % in C4 and −17 % in C5). Mean anomalies were calculated from the baseline climate (see section Valsain forest above) along with climate change scenario periods 2081–2110 for each climate change scenario. The figures presented correspond to an elevation of 1500 m.

Representative stand types (RST)

For the purposes of the analysis, the current forest conditions in Valsaín were characterized according to 24 representative stand types (RST) which had been defined based on forest inventory data from the current management planning period. In general, each RST was defined by species mixture, stand development stage and site type. Twelve RSTs were pure Pinus sylvestris stands including thicket (40 years), pole (60 years), mature (80 years) and overmature (>120 years) stand development stages; 6 RSTs were pure Quercus pyrenaica coppice stands at age 60; and 6 RSTs were composed of Pinus sylvestris (Ps) andQuercus pyrenaica (Qp) where the mixture is in patches (thicket stage: Ps 40 years, Qp 30 years; pole stage: Ps 60 years, Qp 40 years; mature stage: Ps 80 years, Qp 60 years). A DBH distribution (stem numbers in 5 cm diameter classes above a caliper threshold of DBH = 0 cm) and a standwise height–diameter equation at species level were generated using data from local forest inventories for each RST. This information was used to generate initial stand structures for the model simulations. Figure 1 shows a schematic representation of RST occurrence in Valsaín. Soil depth (60–100 cm), soil texture and stoniness (45–60 %) remained fairly constant across RSTs.

Forest management alternatives

According to a questionnaire completed by local forest managers and forest stakeholders, climate change mitigation, nature conservation and recreation are expected to gain importance, while game management and hunting are expected to be of less importance in the future. Accordingly, the management alternatives in the current study explored coppice management for Quercus pyrenaica in pure and mixed stands due to the renewed interest in fuelwood (i.e. bioenergy) as well as different thinning regimes. In the case of Scots pine, the focus of alternative management was on changes in the thinning regime to favour more diverse stand structures and to promote tree vigour while promoting quality timber for wood products with long life spans.

For each of the major stand types (Pinus sylvestris, Quercus pyrenaica, mixed stands), the business as usual management regime, two alternatives as well as a “no-management” scenario were defined. An overview is shown in Table SM1.

Business as usual (BAU)

In Pinus sylvestris stands, the focus is on the production of valuable timber while maintaining a satisfactory level of other ES. In pure Ps stands, three light thinnings from below are applied at ages 40, 60 and 80 years. An irregular shelterwood approach is used for stand regeneration. Regeneration fellings start by opening patches of 0.5–1.0 ha spread over the compartment, taking into account existing advance regeneration. In total, four regeneration fellings are applied, starting with a seeding cut in the selected patches (80 % of volume removed). In a second step, the gaps are enlarged, removing 30 % of the residual volume in the initial gaps, 70 % in the enlargements and 40 % in the rest of the compartment. In the third step, volume is generally reduced by 50 % in the gap areas and the rest of the compartment. In the final regeneration cut after 20 years, a few residual trees per hectare are left standing to provide a nesting habitat for birds. Rotation length is 120 years.

Pure Quercus pyrenaica coppice stands are being converted into high forests. Light thinnings from below removing 15–20 % of the standing volume are applied to promote growth of individual shoots while keeping resprouting at low levels due to limited light availability. Up to 5 thinnings are implemented, beginning at stand age of 30 years. The rotation length is increased to 120 years. For stand regeneration, a uniform shelterwood approach with four entries spread out over 20 years is used. The seeding cut removes 20 % of standing volume, attempting to maintain resprouting at low levels and to favour seed regeneration.

In mixed stands, patches of Quercus pyrenaica coppice are being transformed into high forest as in pure Quercus pyrenaica stands. Pinussylvestris is treated as in pure Pinus sylvestris stands, the rotation length being 120 years for both species.

Alternative management AM1

The management objectives are similar to BAU, promoting quality timber from both species. In contrast to BAU, in pure stands of Pinussylvestris and Quercus pyrenaica selective crown thinnings (35–40 % of standing volume removed) are applied to promote growth and vigour of good quality trees. Quercuspyrenaica coppice stands are maintained as coppice with a rotation length of 70 years. The rotation of Pinussylvestris stands is 120 years, employing the irregular shelterwood system as in BAU for natural regeneration.

In mixed stands, the general concept is as BAU for both species. However, thinnings in Pinussylvestris as well as in Quercus pyrenaica are heavier and from above.

Alternative management AM2

The main difference in AM2 compared to BAU is that Quercuspyrenaica in pure and in mixed stands is maintained as coppice. Rotation length is 70 years for Quercuspyrenaica and 120 years for Pinus sylvestris. Thinning regimes in Quercus pyrenaica and Pinus sylvestris as well as the regeneration method for Pinus sylvestris is as in BAU.

No-management scenario NOM

The main objective of NOM is to allow natural processes including mortality and natural succession to develop without management intervention to create natural, ecologically valuable habitats.

Simulation model PICUS v1.6

The model used in the study is the hybrid forest patch model PICUS v1.6. This model combines elements of a 3D gap model in capturing forest dynamics (Lexer and Hönninger 2001) using process-based approaches to estimate annual stand-level primary production derived from the 3PG model (Landsberg and Waring 1997). The stand-level NPP is then redistributed to individual trees in the simulated forest depending on their relative competitive status. PICUS allows the simulation of individual trees within a 3D frame of 10 × 10 m patches extended in a vertical direction by crown cells of 5 m height. Forest dynamics are driven by key processes of growth, mortality and regeneration, as described in Lexer and Hönninger (2001) and Seidl et al. (2005).

The model follows a modular structure, which includes a management module enabling the user to prescribe almost any silvicultural activity. The PICUS version 1.6 used for this study includes a more detailed representation of the water balance of a simulated forest and the water relations of forest vegetation. To better capture the intra-seasonal dynamics of the soil water regime, this model version uses daily climate data of minimum, maximum and mean temperature, precipitation, radiation and vapour pressure deficit. A brief outline of the core logic of the new water-related modules is presented here (compare Figure SM1, Supplementary Material). For equations and details, see the Supplementary Material SM1.

Soil layers

In PICUS 1.6, the soil consists of several stacked soil layers which replace the simple bucket model of earlier versions of the model. Each soil layer is defined via depth and attached hydraulic properties and extends across the entire simulated area. Precipitation is mediated by evaporation of water intercepted by the canopy; it is temporarily stored in the snow pack and distributed vertically between the soil layers. Vertical unsaturated flow of soil water is calculated at hourly resolution based on Darcy’s law; in saturated soils water percolation depends on the hydraulic conductivity of the respective layer. The soil water content (SWC) is an important variable for various ecosystem processes and is updated daily for each soil layer. The fine root biomass (rooting depth, vertical distribution) of a tree depends on species and tree size, and its vertical distribution over soil layers is updated annually (Eq. SM1-1, Eqs. SM1-2a-c).

Interception

The interception of precipitation depends on whether it is rainfall or snow. The actual interception of rainfall is limited by the storage capacity of the canopy. The actual storage capacity is limited by the maximum storage capacity, which is interpolated between 2 mm (broadleaved) and 4 mm (conifers) based on LAI shares. The interception of snow is determined by species-specific interception efficiency and a maximum snow load as well as the eventual unloading of snow from the canopy on days following snowfall events (Eqs. SM1-3a-g).

Evaporation

The calculation of evaporation from the canopy uses the Penman–Monteith (PM) equation as suggested by Landsberg and Gower (1997) (Eq. SM1-4, Eqs. SM1-5a-b). Intercepted water that is not evaporated within one day is added to throughfall and infiltrates the soil. Snow on the canopy and in the snowpack on the ground melts at temperatures >0 °C at a daily rate of 0.7 mm/°C. In calculating the evaporation of water from the soil with the PM equation, soil conductivity depends on the actual soil water content and the soil texture type of the uppermost soil layer (Eq. SM1-4, Eqs. SM1-5a-b).

Transpiration

The transpiration (T) is calculated with a canopy conductance obtained from species-specific stomatal conductance weighted with the species’ LAI. Transpiration estimates scale with leaf area index linearly to a maximum of 6 mm per day (Eqs. SM1-6a-d).

The stomatal conductance gs(j) results from a species-specific maximum stomatal conductance reduced by modifiers for light, temperature, vapour pressure deficit and soil water content as well as a vegetation period modifier for broadleaved species (Eqs. SM1-7a-h).

Tree species response to soil water content is calculated for each soil layer and is then aggregated to the soil water response at species level by weighting the layer-specific responses with the species’ share of fine roots in each layer.

Response of stand-level NPP

In the formula used to estimate photosynthetically active utilizable radiation (see Landsberg and Waring 1997), the terms for VPD and soil moisture were replaced by a modifier that represents the response of stomatal conductance to soil moisture and VPD (Eq. 1).
$$\varPhi_{{{\text{p}} . {\text{a}} . {\text{u}} .}} = \varPhi_{{{\text{p}} . {\text{a}} .}} \cdot {\text{resp}}_{\text{stomata}} \cdot {\text{resp}}_{\text{temp}} \cdot {\text{resp}}_{\text{frost}}$$
(1)
where respstomata denotes the average fraction of realized to potential transpiration rates [Eq. (2)].
$$E_{\text{rel}} = \frac{{E_{T} }}{{E_{\text{pot}} }}$$
(2)
For Epot a modified gs is used in the PM equation where f(VPD) and f(SWC) are set to 1, i.e. no limitation of stomatal conductance due to soil water status and saturation deficit of the atmosphere is assumed.

Differences in the vertical distribution of fine roots result in varying soil water responses of individual trees and a competitive advantage of larger trees since they are able to tap into deeper soil layers. For the establishment and regeneration processes, the soil moisture response values for the uppermost soil layer are used.

Model calibration

Following the generic concept of species parameterization (see Seidl et al. 2005), all standard species parameters for Pinus sylvestris available in the model were used except as described below. The standard potential height development equation for Pinus sylvestris did not compare well with height data for the dominant size classes in the Pinus sylvestris RSTs and seemed to overestimate potential tree height at older ages. As tree height is relatively insensitive to management, the age-specific maximum tree height data from the available inventory were used to adjust the potential height development for older Pinus sylvestris trees in Valsain. Species parameters required to calculate the water relations were taken from the literature, as shown in Supplementary Material SM1.

Quercus pyrenaica had not been included in the PICUS species set previously and had to be set up accordingly. Species parameters other than those shown in SM1 were taken from Quercus petraea (response to light and Nitrogen supply) and Quercus pubescens (response to soil water supply and temperature, allometric equations for leaves and branches) based on qualitative information about the ecology of Quercus pyrenaica. The potential height development was derived from height information in regional yield tables for Quercus pyrenaica.

Earlier versions of the PICUS model have been evaluated successfully in temperate and alpine forests (Lexer and Hönninger 2001; Seidl et al. 2005, 2011a, b; Didion et al. 2009; Huber et al. 2013). Recently, the model has also been applied successfully to Mediterranean Pinus pinea stands in Spain (Pardos et al. 2015). In the Valsaín study region, no empirical time series growth data were available for model validation. Thus, the adapted model version was evaluated against chronosequences of Pinus sylvestris stands and regional yield tables for Quercus pyrenaica (see Model Evaluation in the results section).

Initializing representative stands for the model simulations

For each RST, a generic initial state was generated using the diameter distribution and tree height in each diameter class. If no information on stand structure and mixing pattern was available, trees were randomly distributed over the simulated area (1–3 ha depending on the silvicultural regime applied) starting with the largest trees from the diameter distribution and assigned to one of the 10 × 10 m patches. In a trial pre-run, the matter of whether the generated stand structure allowed all trees to survive at least 5 years was tested. If not, dying trees were iteratively reallocated to another patch until the stand structure allowed the tree population to be sustained (compare Maroschek et al. 2014).

For each RST, the three alternative management regimes (BAU, AM1, AM2) as well as the NOM scenario were simulated under the baseline climate and the 5 transient climate change scenarios over 100 years (period 2010–2110). The management prescriptions used within PICUS were written using a model-specific scripting language which defined the timing, the tree subpopulations that were to be harvested (species, dimensions, position) and the operation type (e.g. cut and extract).

Analysis

We assess four main ES [timber production (TP), habitat and biodiversity conservation (NCBD), protection against gravitational hazards, rockfall and erosion (PGH) and carbon sequestration (CS)] through different ES indicators.

Based on a survey of 34 regional stakeholders (Mariñas 2014), 27 indicators of TP, NCBD, PGH and CS were initially chosen for the current study (Table 1). TP was represented by harvested timber volume (TVH), forest productivity (VI) and forest stocking (V). Indicators of NCBD were related to (1) living trees [tree size diversity (H) and abundance of large living trees (LLTN)], (2) dead trees [deadwood volume (SDWV), abundance (LSDTN) and volume of large standing dead trees (LSDTV)] and (3) gaps in crown cover (COV). Protection against gravitational hazards includes protection against rockfalls (RPIi, i = 1–10), protection against the release of snow avalanches (API) and protection against landslides and erosion (LPI). RPI and API integrate stem density and mean diameter; LPI is based on crown cover. CS was represented by total carbon stored in aboveground and belowground tree biomass. Table 1 shows all used indicators. For details on the calculation of the ES indicators, we refer to Bugmann et al. (2015).
Table 1

Factor loadings and communalities for the 27 indicators of timber production (TP), nature conservation and the maintenance of biodiversity (NCBD) and protection against gravitational hazards (PGH)

Significant loadings (p < 0.01) are bolded. The 12 indicators selected for TP, NCBD and PGH based on the factor analysis are shaded

PICUS provided model outputs for all the indicators on an annual basis over the entire simulation period (2010–2110). A factor analysis was performed to identify those indicators among the 27 that best explained timber production (TP), biodiversity conservation (NCBD) and protection (PGH) and that provided a clear and comprehensive insight into the structure of the data for these three ES. The number of factors extracted in each of these three ES was decided by comparing the scree of factors of the observed data with that of a random data matrix of the same size as the original (Dillon and Goldstein 1984). The varimax rotation was used in the analysis, and factors were extracted using the principal factor method. Polytomous methods were employed with the categorical indicator LPI (protection against landslides and erosion).

The reduced set of ES indicators for TP, NCBD and PGH plus CS was then included in a second factor analysis in order to assess the interdependencies among ES and to understand the different patterns of common variance in the four management scenarios. The analysis was performed using the psych (Revelle 2014) and GPArotation (Bernaards and Jennrich 2005) packages of R (R Core Team 2014).

The vulnerability of ES provision to climatic changes was evaluated by analysing the sensitivity of the selected indicators to each of the climate change scenarios relative to the baseline climate C0 (i.e. the ratio between the expected value under baseline conditions and under the different climate change scenarios, Ci/C0, i = 1–5) in four 25-year periods. ANOVAs were used to identify significant main effects of species composition and management on ES provisioning as well as their first-order interactions. For the Pinus sylvestris RSTs, significance of altitude, management and their interactions were also tested. All ANOVAs were performed using SAS (SAS v. 9.2, SAS Institute, Cary, NC, USA).

Results

Model evaluation

In model simulations, it was possible to reproduce the chronosequence data of the Pinus sylvestris RSTs very well (see Supplementary Material SM2). Simulated mean productivity over 100 years (current climate) in the Pinus sylvestris RSTs was 8.6 m3 ha−1 year−1 at 1500 m, 5.2 m3 ha−1 year−1 at 1750 m and 3.9 m3 ha−1 year−1 at 2000 m a.s.l. For Quercus pyrenaica, all RSTs were of similar age; therefore, regional yield tables were used to evaluate the simulated temporal development of Quercus pyrenaica tree populations. Simulated mean biomass production at stand level over a 60-year rotation ranged from 2.1 to 3.0 t ha−1 year−1 and compared well with low to intermediate yield classes (see SM2). Please note that although the potential height development for oak and pine trees was based on regional data, the challenge for PICUS was to capture the response of several biometric stands and tree attributes to the site gradients in Valsaín. Overall, the response to the site gradients appeared plausible and consistent with observations.

Factor analysis

The results of the first factor analysis performed separately for each ES group of indicators are given in Table 1. Based on these results, we reduced the number of ES indicators to 12, by dropping species-specific volume indicators and all RPI indicators except those highly correlated with factors 1 and 2 (RPI1 and RPI9). Carbon storage (above- and belowground living tree biomass) was added to this list of indicators; thus, a subset of 13 indicators was used for further analysis. A second factor analysis was run with these 13 indicators in order to assess the interdependences among ES and to understand the different patterns of common variance over the four management regimes (Fig. 2). Three factors were extracted, accounting for 79.1 % of the total variance. The first factor (41.7 % of the common variance) represents protection against gravitational hazards (highly and significantly correlated with all RPIs and LPI). The second factor (38.9 % of the common variance) correlates positively with stand volume (V), carbon (C) and indicators of diversity (H, LSDTV, LSDTN and COV and SDWV). The third factor (19.4 %) correlates with timber productivity (VI and TVH). The results were very similar for all analysed climate scenarios.
Fig. 2

Results of the factor analysis based on the indicators used, for the baseline climate (C0) (a, d) and climate change scenarios C2 (b, e) and C5 (c, f). The three management alternatives (BAU, AM1 and AM2) and NOM are plotted as circles, indicators as triangles. Factor 1 correlates positively with protection against rockfall (RPI(i=1–10)). Factor 2 correlates positively with stand volume (V) and indicators of diversity (H, LSDTN and COV and SDWV). Factor 3 correlates positively with timber production (TVH and VI)

Provision of ES under baseline climate

The results of the simulations for all RSTs showed that the indicators of CS, NCBD and PGH (rockfall, landslides and erosion) were significantly better in the no-management alternative (NOM) from 2060 onwards (periods 3 and 4 in Table 2), with not much difference among the other management regimes (see Table 2). In general, BAU also performed better than AM1 and AM2. With regard to timber production, TVH and VI significantly decreased in the last 50 years of the simulation, due to the renewal of overmature Quercus pyrenaica stands at the beginning of the analysis period (TVH) and ageing tree populations (VI). Most of the analysed indicators (except V, LLTN, COV and LPI) decreased in the last period under the management alternatives, while they increased under the no-management alternative (significant management x time period interaction).
Table 2

Effects of three management alternatives (BAU, AM1, AM2) and the no-management variant NOM on the provision of ecosystem services under the baseline climate (C0) for all RSTs in Valsaín forest

 

Forest management alternatives by period

Period 1

Period 2

Period 3

Period 4

AM1

AM2

BAU

NOM

AM1

AM2

BAU

NOM

AM1

AM2

BAU

NOM

AM1

AM2

BAU

NOM

TP

V

225.5

221.6

254.2

285.8

148.1b

156.5b

219.8b

392.1a

109.28b

125.7b

138.3b

457.6a

102.4b

117.8b

132.4b

467.3a

VI

6.3

6.2

7.0

6.8

5.0

5.1

5.6

6.1

5.3

5.6

4.5

5.0

5.1

5.4

4.9

4.0

TVH

5.2a

5.2a

3.7a

0.0b

5.3a

4.5a

5.6a

0.0b

3.3a

3.5a

4.8a

0.0b

3.3a

3.0ba

2.0b

0.0c

CS

CS

94.8

93.0

109.0

121.3

63.5b

68.5b

92.9b

155.7a

49.8b

58.6b

59.0b

171.5a

45.6b

53.0b

57.0b

168.5a

NCBD

LLTN

24.9

25.1

27.0

33.1

16.3b

18.9b

25.1ba

50.9a

7.0b

9.1b

13.5b

63.4a

1.3b

1.7b

5.0b

69.3a

SDWV

5.79

5.56

5.85

8.57

6.88b

6.79b

7.97b

22.34a

5.25b

4.80b

9.34b

31.61a

4.3b

4.5b

7.6b

39.7a

LSDTN

3.6

3.5

3.6

5.0

3.65b

3.38b

4.68b

12.05a

2.01b

1.94b

4.82b

17.45a

1.19b

1.01b

2.93b

18.33a

LSDTV

4.6

4.4

4.5

6.9

5.3b

5.1b

6.1b

17.5a

4.1b

3.4b

7.3b

25.7a

2.4b

2.0b

4.3b

34.3a

COV

0.24

0.24

0.20

0.20

0.18b

0.19b

0.22ba

0.28a

0.18b

0.20b

0.24ba

0.28a

0.25b

0.26b

0.25b

0.24a

H

0.88b

0.83b

1.06ba

1.11a

0.79cb

0.69c

0.98b

1.24a

0.76b

0.65b

0.68b

1.24a

0.61b

0.55b

0.68b

1.28a

PGH

RPI1

0.91b

0.92b

0.96ba

0.99a

0.93

0.94

0.92

0.99

0.92ba

0.93ba

0.87b

0.99a

0.81b

0.88ba

0.83b

0.99a

RPI9

0.46

0.45

0.54

0.57

0.29b

0.31b

0.49ba

0.70a

0.26b

0.32b

0.37b

0.72a

0.24b

0.28b

0.28b

0.69a

LPI

2

2

3

3

2

2

2

3

2

2

2

3

2

2

2

3

TP timber production, CS carbon sequestration, NCBD nature conservation and biodiversity, PGH protection against gravitational hazards

V, standing volume (m3 ha−1); VI, current annual volume increment (m3 ha−1 year−1); TVH, volume of timber harvested (m3 ha−1 year−1); CS, total carbon storage (t.ha−1); LLTN, abundance of large living trees (n ha−1); SDWV, dead wood volume (m3 ha−1); LSDTN, abundance of large standing dead trees (n ha−1); LSDTV, volume of large standing dead trees (m3 ha−1); COV, canopy cover; H, tree size diversity; RPI1, RPI9, protection against rockfall (1 = small boulders, 9 = large boulders); LPI, protection against landsides and erosion

Results are shown for four 25-year periods from 2010 to 2110. Different letters show significant differences between forest management alternatives within each period. Only when differences are significant, letters are shown

Vulnerability of ES provision to climate change scenarios

When considering all RSTs, the analysis of the sensitivity of the studied indicators to scenarios of climate change relative to the baseline climate C0 (i.e. Ci/C0, i = 1–5) showed three general trends (Table SM2 and Fig. 3): (1) a change in the provision of the ES appeared in the last 50 years of the 100-year simulation period (periods 3 and 4); (2) climate scenarios C2 and C3 had a positive impact, while climate scenarios C4 and particularly C5 had a negative impact, and climate scenario C1 had a positive or neutral impact, depending on the indicator analysed; and (3) distinguishable effects of the NOM alternative were visible for volume, productivity, carbon storage, protection, canopy cover and volume of dead trees; the other management alternatives (BAU, AM1, AM2) did not cause any major differences.
Fig. 3

Sensitivity of standing volume (V), tree size diversity (H), protection against rockfall (RPI9) and current annual volume increment (VI) to climate change scenarios relative to the baseline climate C0 (Ci/C0, i = 1–5), for the four 25-year periods of the simulation period (2010–2110) for the forest management alternatives (NOM, BAU, AM1 and AM2). Asterisks show significant differences between managements within each period (**P < 0.01; *P < 0.05). Straight dotted line represents Ci/C0 = 1

When compared to the baseline climate, standing volume (V) in the last period of the simulation increased or at least stabilized under climate change scenarios C1, C2 and C3 (3–11 %) in a similar way for all management alternatives, while it decreased under climate change scenarios C4 (−9 %) and particularly C5 (−14 %) for AM1, AM2 and BAU, and to a lesser extent for NOM (−3 %). When compared to the baseline climate, carbon storage also increased under climate scenarios C1, C2 and C3, while it decreased under management alternatives in C4 and C5 (Table SM2). The decrease in standing volume (−13 to −29 %) under climate scenario C5 was especially visible in the lower parts of the forest (1250 m a.s.l.), mainly for AM1, AM2 and BAU, for both mixed and pure Quercus pyrenaica stands.

The indicators of biodiversity related to large standing trees (dead and alive: LLTN, LSDTV, LSDTN) showed minor differences among climate change scenarios. Climate scenarios C1 and C2 had a positive impact on COV under NOM and BAU alternatives (40.8 and 12.6 % increase in C1; 6.3 and 26.8 % increase in C2, respectively, compared to the baseline climate) and a negative impact under climate C5 (9.6 % decrease) for all management alternatives. For all RSTs analysed, climate change scenarios C1, C2, C3 and C5 had a positive impact on tree size diversity index (H) for the AM1 alternative and climate scenarios C1 and C3 for AM2. Results for the ES “protection against gravitational hazards” showed greater protection against rockfall (RPI) for the NOM alternative under climate C1. Climate scenario C5 had a negative impact on RPI under AM1 and AM2. Neither climate change nor management scenarios significantly affected LPI.

Table SM3 summarizes the effect of species composition and management on the provision of ES for all climate scenarios in the last 25 years of the simulation. The initial differences in V, VI, TVH and carbon storage according to species composition were maintained throughout the 100-year simulation period; thus, results are only shown for the last 25 years. For all climate scenarios, volume, productivity and carbon storage were lowest in the pure Quercus pyrenaica stands and highest in the mixed or pure pine stands. Under management regimes (BAU, AM1 and AM2), standing volume and carbon storage in pure pine stands and mixed stands stabilized in the last 50 years of the simulation period, while these indicators significantly decreased in the Quercus pyrenaica stands in comparison with NOM in the last 25 years.

BAU and NOM offered better protection against rockfall compared to AM1 and AM2. NOM also showed lower variation between climate scenarios. LPI was relatively insensitive to climatic changes and also responded less to management compared to API.

As regards diversity indicators, compared to other stand types, pure Quercus pyrenaica stands showed significantly lower values for all indicators except COV. The NOM alternative had a positive impact on the diversity indicators, regardless of species composition, mainly over the last 50 years of the simulation. In addition, throughout the simulation, pine stands favoured standing deadwood (SDWV) and the abundance of large standing dead trees (LSDTN) under all climate scenarios, while there was a clear differentiation in tree size diversity (H) related to species composition, with descending order from mixed to pure pine to pure Quercus pyrenaica stands.

Table SM4 summarizes the effect of altitude (2000 m–1750 m–1500 m) and management on ES provisioning for all climate scenarios in pure pine stands. Under all climate scenarios and throughout the simulation period, timber, carbon storage and biodiversity indicators were highest at 1500 m, intermediate at 1750 m and lowest at 2000 m, while the protection against rockfall was highest at 2000 m. The NOM alternative favoured standing volume, carbon storage and most of the indicators related to biodiversity (LSDTV, LLTN, SDWV, LSDTN and BA) at all altitudes. At 2000 m and for the last years of the simulation, both NOM and BAU alternatives increased standing volume, carbon storage, protection and biodiversity, while they showed the lowest timber volume harvested (TVH). In the last years of the simulation, under climate scenarios C4 and C5, there was also a decrease in standing volume and carbon storage in the unmanaged stands at altitudes of 1750 m and above.

Discussion

Model validity

As in any model-based study, model reliability is a crucial issue. For the current analysis, we employed a mature model that has been successfully applied in many climate change-related studies (e.g. Seidl et al. 2011a, b; Maroschek et al. 2014). Recently, it has also been applied for continental Mediterranean conditions, and it performed well against empirical data (Pardos et al. 2015). In addition, the model version in the current study included improved algorithms for calculating the water balance and plant water relations under dry conditions. Although an established species parameter set was used in the case of Pinus sylvestris, Quercus pyrenaica had to be defined as a completely new species within PICUS. It must be acknowledged that many species-specific parameters were based on available information about Quercus petraea and Quercus pubescens which were transferred to Quercus pyrenaica based on the literature regarding the ecology of oak species (Sánchez-Palomares et al. 2012). When new ecophysiological knowledge concerning the environmental response of Quercus pyrenaica becomes available, model formulations should be carefully scrutinized and updated.

The comparison of simulated development of Pinus sylvestris and Quercus pyrenaica stands over a rotation cycle with observed chronosequence data and yield tables indicated plausible responses to the site gradients involved. Please note that the use of regional data on maximum tree height did not preclude a meaningful evaluation of the model behaviour along site gradients (temperature, water supply, nitrogen). However, there are two issues which should be taken into consideration when interpreting the results of this study: (1) disturbances, particularly from biotic agents, were not considered explicitly; (2) while the reliability of physiological parameter values for transpiration processes is high (see SM1), the limits of drought tolerance for Quercus pyrenaica are prone to substantial uncertainty. Results for the low elevation sites should therefore be interpreted with care.

Vulnerability to climate change

Based on our results, we may conclude that the choice of climate change scenario exerted a moderate influence on the provision of ES at sites above 1500 m a.s.l., with no apparent risk of forest dieback in the simulated 100-year period. Nevertheless, the exploration of the impacts of climate change scenarios when compared to the baseline climate revealed a negative effect of climate scenarios C4 and particularly C5 (due to the interactive effect of warmer temperatures and reduced precipitation), and a positive or neutral effect of climate scenarios C1 to C3 on the provision of all analysed ES. Under climate scenarios C4 and C5, the decrease in the provision of timber and carbon storage over the last 50 years of the simulation period, particularly (although not exclusively) in the case of BAU, suggested an increasing risk of declining vitality and related increase in tree mortality beyond the 100-year simulation period, particularly at low elevations (1250 m). For instance, in a recent study, Gea-Izquierdo et al. (2014) found symptoms of pine mortality at the lower sites (below 1100 m), but not at 1500 m. This finding is in agreement with results of empirical analyses indicating that the productive optimum for P. sylvestris under historic climate was at altitudes between 1200 and 1600 m (Montero 1994).

Climate-driven large-scale shifts from forest to non-forest ecosystem types seem unlikely in Valsaín, at least during the 100-year simulation period, because water availability in the climate scenarios was sufficient even under the driest scenario C5 (mean annual precipitation between 875 and 1172 mm). These sub-Mediterranean conditions in Valsaín are responsible for the contrasting results in typical continental Mediterranean forests, where different studies point to severe impacts of climate change (e.g. Sabaté et al. 2002). Carbon storage and timber supply even increased under climate scenarios C1 to C3 in comparison with the baseline climate for the management alternatives considered. This is a frequent prediction at sites with sufficient water availability (Lindner and Calama 2013). Thus, a warmer climate could potentially increase net primary production and carbon stocks in Valsaín (Lindner et al. 2010), as long as the decrease in precipitation remains small and does not limit the growth response to warmer temperatures (Turner et al. 2013).

No-management implications

The no-management alternative could buffer the impact of climate scenarios C4 and C5, maintaining an ES provision similar to that of the baseline climate. However, NOM fails to initiate regeneration and maintains a high share of ageing, single-layered mature stands with lower adaptive capacity to environmental changes compared to younger stand development stages, which have a higher plasticity in terms of structural and compositional changes (Carrer and Urbinati 2004; Vieira et al. 2008). It is increasingly acknowledged that biodiversity plays an important role in the functioning of forests. For instance, the ability of forests to resist environmental changes and to recover from disturbances is related to biodiversity from genetic to landscape scales (Brockerhoff et al. 2013). Our results indicate improved biodiversity indicators under no management, which is related to the presence of large standing trees and deadwood, while the increase in COV pointed to less gaps and open canopy conditions under no management (McElhinny 2002) and thus lower microclimatic variation. Tree size diversity increased mainly due to continuing differentiation of trees, while regeneration was at low levels due to closed canopies.

Old stands with overmature trees (Lindenmayer et al. 2000) provide more opportunities for nesting birds and animal shelter (van den Meerschaut and Vandekerkhove 1998). Deadwood may also be a good indicator of functional attributes such as decomposition and nutrient cycling processes, while large trees create important niches for invertebrates, birds, mammals, fungi and epiphytes, thus contributing to biodiversity (McElhinny 2002). For instance, large pine trees constitute important niches for the imperial eagle, black stork and black vulture in Valsaín forest. In contrast, these well-established biodiversity indicators did not develop favourably in coppice stands. The no-management alternative obviously had a negative impact on the provision of timber, although the persistence of the stands was not negatively affected, at least not during the 100-year simulation period. As regards carbon storage, the no-management alternative performed better than the other three management regimes. Under the management alternatives, there will be a trade-off between storage within the forest and ex situ carbon sinks, while the assimilated carbon will be released again due to respiration and decomposition in the unmanaged stands (Duncker et al. 2012).

Forest composition response

Scientists have begun to explicitly investigate the role of tree species diversity in the simultaneous provision of multiple ES (Gamfeldt et al. 2013). The species effect in our study was confounded by several factors. Pure Quercus pyrenaica stands decreased their productivity in the last 25 years of the simulation period, with a drastic decrease under the BAU alternative. This behaviour is mainly due to the legacy effect of the coppice system which has generated the current Quercus pyrenaica stands. In these stands, most of the trees are coppice shoots which, according to the model algorithms, cannot maintain their growth rate for as long as trees from seed regeneration (compare Serrada et al. 2008). Another reason was that Quercus pyrenaica stands mainly occur at less productive sites (compare Fig. 1). At the low elevation sites (1250 m) in particular, Quercus pyrenaica was very sensitive to drastic climate change (climate scenario C5). This may result in loss of vigour and related increase in the susceptibility of Quercus pyrenaica to biotic (pest and disease) stresses. Other studies have suggested that climate change may lead to Quercus pyrenaica being replaced in the long term by the sclerophyllous and more drought-tolerant Quercus ilex (Sánchez-de-Dios et al. 2009). In the Pyrenees, for instance, this shift in species composition seems to have already begun (Galiano et al. 2010; Gea-Izquierdo et al. 2014). However, in our model analysis, management prescriptions in all actively managed scenarios prevented “uncontrolled” species change. Although shifts among pine and Quercus pyrenaica were potentially possible in the no-management scenario, Q. ilex was not considered in the simulations.

Management trade-offs and synergies

Our simulations indicate a relatively small difference among the analysed management alternatives (BAU, AM1 and AM2) as regards the provision of ES and a clear improvement in diversity indicators, protection and carbon storage through the no-management alternative (NOM) (Table SM3). The NOM alternative is interesting as it is the scenario which will be closest to reality within the framework of the recently declared National Park for the upper vegetation belt in Valsaín. However, this alternative does not support the forestry sector in the region because it implies no commercial timber production, fewer silvicultural operations and less direct employment in forestry. In contrast, if the managed alternatives are applied (BAU, AM1 and AM2), Valsaín forest is likely to remain productive (with a loss of 6.8 and 12.6 % in standing volume under climate scenarios C4 and C5) while maintaining an acceptable provision of diversity and protection (compare Hunter 2001; Gamborg and Larsen 2003).

Protection against rockfall was best in the mixed stands and was favoured by the no-management alternative (NOM). This result is due to the combined occurrence of high-density Quercus pyrenaica patches and larger pine trees. Whereas mixed stands favoured tree size diversity, and pure pine stands favoured large standing trees and deadwood, pure Quercus pyrenaica stands presented higher canopy cover (COV). Multi-layered forests, typical of mixed stands, create a greater number and diversity of niches which in turn may accommodate more animals and plant species (Van den Meerschaut and Vandekerkhove 1998) and maintain macrofungal diversity and abundance (Deal et al. 2014). However, although the biodiversity indicators employed are well established in the literature, care should be taken before drawing final conclusions. An evaluation with local data would be extremely valuable.

In addition, a combined adaptation strategy that creates improved resilience through greater structural and species diversity as well as promoting increased drought tolerance through suitable mixtures (Jucker et al. 2014) seems effective in preventing loss of productivity and maintaining timber supply (Steenberg et al. 2011), as occurred in the mixed stands.

The results also clearly show that it is not possible to maximize all ES simultaneously at stand level. However, stand management concepts can be tailored to specific combinations of ES. For instance, using a targeted approach, deadwood management can improve the biodiversity profile of timber production regimes. It must be noted that the differences between the analysed managed alternatives (BAU, AM1, AM2) were fairly small, with no direct implications for forest structure and composition. On the other hand, mixing different management alternatives at landscape scale is a straightforward solution to maintaining a high diversity of multifunctional forests.

Conclusions

More diverse (mixed-species) forests may be able to withstand and adapt to climate change because there will be a greater probability that part of the tree population in a stand will be able to persist as climate is changing (Brockerhoff et al. 2013). In this regard, the mixed stands in Valsaín (1250–1500 m a.s.l.) showed the best performance in providing the four analysed ES. Moving towards multi-species management can better realize the full potential of economically, ecologically and culturally valuable ES (Gamfeldt et al. 2013). The enrichment of species composition appears to be a good alternative at medium altitudes (1500 m), while above 1500 m, productive pines stands can be left as they would benefit from warmer temperatures in winter under climate change. In simulations at lower sites (1250 m), Quercus pyrenaica was already showing decline in growth, indicating increasing risk of failing to deliver ES in pure coppice stands. The conversion of coppice stands to high forest of Quercus pyrenaica stands will improve the rejuvenation of the species and increase the genetic adaptation potential. However, the conversion strategy should be complemented by admixing other more drought-tolerant species.

Notes

Acknowledgments

This study was financed by the EU FP7 project ARANGE- 289437. We are especially grateful to Javier Donés, Director of Valsaín forests and to Miguel Cabrera for facilitating the access to data. We are grateful to three anonymous reviewers who provided thoughtful comments on earlier versions of the manuscript. English was revised by Adam Collins.

Supplementary material

10113_2016_985_MOESM1_ESM.docx (81 kb)
Supplementary material 1 (DOCX 81 kb)
10113_2016_985_MOESM2_ESM.docx (43 kb)
Supplementary material 2 (DOCX 42 kb)
10113_2016_985_MOESM3_ESM.docx (14 kb)
Supplementary material 3 (DOCX 14 kb)
10113_2016_985_MOESM4_ESM.docx (19 kb)
Supplementary material 4 (DOCX 18 kb)
10113_2016_985_MOESM5_ESM.docx (24 kb)
Supplementary material 5 (DOCX 24 kb)
10113_2016_985_MOESM6_ESM.docx (25 kb)
Supplementary material 6 (DOCX 24 kb)
10113_2016_985_MOESM7_ESM.docx (117 kb)
Supplementary material 7 (DOCX 116 kb)

References

  1. Allen CD, Macalady AK, Chenchouni H, Bachelet D, McDowell N, Vennetier M, Kitzberger RA, Breshears DD, Hogg EH, Gonzalez P, Fensham R, Zhang Z, Castro J, Demidova N, Lim JH, Allard G, Running SW, Semerci A, Cobb N (2010) A global overview of drought and heat-induced tree mortality reveals emerging climate change risks for forests. For Ecol Manag 259:660–684. doi:10.1016/j.foreco.2009.09.001 CrossRefGoogle Scholar
  2. Bernaards CA, Jennrich RI (2005) Gradient projection algorithms and software for arbitrary rotation criteria in factor analysis. Educ Psychol Meas 65: 676–696. doi:10.1177/0013164404272507 http://www.stat.ucla.edu/research/gpa
  3. Brockerhoff EG, Jactel H, Parrotta JA, Ferraz SFB (2013) Role of eucalypt and other planted forests in biodiversity conservation and the provision of biodiversity-related ecosystem services. For Ecol Manag 301:43–50. doi:10.1016/j.foreco.2012.09.018 CrossRefGoogle Scholar
  4. Bugmann H, Cordonnier Th, Truhetz H, Lexer MJ (2016) Impacts of business-as-usual management on ecosystem services in European mountain ranges under climate change: introduction. Reg Environ Change (this volume)Google Scholar
  5. Carrer M, Urbinati C (2004) Age-dependent tree-ring growth responses to climate in Larix decidua and Pinus cembra. Ecology 85(3):730–740. doi:10.1890/02-0478 CrossRefGoogle Scholar
  6. Deal RL, White R (2012) Integrating forest products with ecosystem services: a global perspective introduction. For Policy Econ 17:1–2. doi:10.1016/j.forpol.2012.02.014 CrossRefGoogle Scholar
  7. Deal RL, Hennon P, O’Hanlon R, D’Amore D (2014) Lessons from native spruce forests in Alaska: managing Sitka spruce plantations worldwide to benefit biodiversity and ecosystem services. Forestry 87:193–208. doi:10.1093/forestry/cpt055 CrossRefGoogle Scholar
  8. Didion M, Kupferschmid AD, 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(2):133–147. doi:10.1016/j.ecolmodel.2008.09.021 CrossRefGoogle Scholar
  9. Dillon W, Goldstein M (1984) Multivariate analysis: methods and applications. Wiley, New YorkGoogle Scholar
  10. Duncker PS, Raulund-Rasmussen K, Gundersen P, Katzensteiner K, De Jong J, Ravn HP, Smith M, Eckmüllner O, Spiecker H (2012) How forest management affects ecosystem services, including timber production and economic return: synergies and trade-offs. Ecol Soc 17(4):50. doi:10.5751/ES-05066-170450 Google Scholar
  11. Fürst C, Frank S, Witt A, Koschke L, Makeschin F (2013) Assessment of the effects of forest land use strategies on the provision of ecosystem services at regional scale. J Environ Manag 127:S96–S116. doi:10.1016/j.jenvman.2012.09.020 CrossRefGoogle Scholar
  12. Galiano L, Martinez-Vilalta J, Lloret F (2010) Drought-induced multifactor decline of scots pine in the pyrenees and potential vegetation change by the expansion of co-occurring oak species. Ecosystems 13:978–991. doi:10.1007/s10021-010-9368-8 CrossRefGoogle Scholar
  13. Gamborg C, Larsen JB (2003) ‘Back to nature’—a sustainable future for forestry? For Ecol Manag 179:559–571. doi:10.1016/S0378-1127(02)00553-4 CrossRefGoogle Scholar
  14. Gamfeldt L, Snäll T, Bagchi R, Jonsson M, Gustafsson L, Kjellander P, Ruiz-Jaen MC, Fröberg M, Stendahl J, Philipson CD, Mikusiński G, Andersson E, Westerlund B, Andrén H, Moberg F, Moen J, Bengtsson J (2013) Higher levels of multiple ecosystem services are found in forests with more tree species. Nat Comm 4:1340. doi:10.1038/ncomms2328 CrossRefGoogle Scholar
  15. Gea-Izquierdo G, Viguera B, Cabrera M, Cañellas I (2014) Drought induced decline could portend widespread pine mortality at the xeric ecotone in managed mediterranean pine-oak woodlands. For Ecol Manag 320:70–82. doi:10.1016/j.foreco.2014.02.025 CrossRefGoogle Scholar
  16. Hewitt CD, Griggs DJ (2004) Ensembles-based predictions of climate changes and their impacts (ENSEMBLES). Eos Trans. 85(52):566. doi:10.1029/2004EO520005 CrossRefGoogle Scholar
  17. Huber MO, Eastaugh CS, Gschwantner T, Hasenauer H, Kindermann G, Ledermann T, Lexer MJ, Rammer W, Schörghuber S, Sterba H (2013) Comparing simulations of three conceptually different forest models with National Forest Inventory data. Environ Model Softw 40:88–97. doi:10.1016/j.envsoft.2012.08.003 CrossRefGoogle Scholar
  18. Hunter M (2001) Maintaining biodiversity in forest ecosystems. Cambridge University Press, CambridgeGoogle Scholar
  19. Jucker T, Bouriaud O, Avacaritei D, Dănilă I, Duduman G, Valladares F, Coomes DA (2014) Competition for light and water play contrasting roles in driving diversity–productivity relationships in Iberian forests. J Ecol 102:1202–1213. doi:10.1111/1365-2745.12276 CrossRefGoogle Scholar
  20. Landsberg JJ, Gower ST (1997) Applications of physiological ecology to forest management, Physiological Ecology Series. Academic Press, London. ISBN 0-12-435955-8Google Scholar
  21. Landsberg JJ, Waring RH (1997) A generalized model of forest productivity using simplified concepts of radiation-use efficiency, carbon balance and partitioning. For Ecol Manag 95:209–228. doi:10.1016/S0378-1127(97)00026-1 CrossRefGoogle Scholar
  22. Lexer MJ, Hönninger K (2001) A modified 3D-patch model for spatially explicit simulation of vegetation composition in heterogeneous landscapes. For Ecol Manag 144(1–3):43–65. doi:10.1016/S0378-1127(00)00386-8 CrossRefGoogle Scholar
  23. Lindenmayer DB, Margules CR, Botkin DB (2000) Indicators of biodiversity for ecologically sustainable forest management. Conserv Biol 14(4):941–950. doi:10.1046/j.1523-1739.2000.98533.x CrossRefGoogle Scholar
  24. Lindner M, Calama R (2013) Climate change and the need for adaptation in Mediterranean forests. In: Lucas-Borja NE (ed) Forest Management of Mediterranean forests under the new context of climate change. Building alternatives for the coming future. Nova Science Pub, New York, pp 13–30Google Scholar
  25. Lindner M, Maroschek M, Netherer S, Kremer A, Barbati A, Garcia-Gonzalo J, Seidl R, Delzon S, Corona P, Kolström M, Lexer MJ, Marchetti M (2010) Climate change impacts, adaptive capacity, and vulnerability of European forest ecosystems. For Ecol Manag 259:698–709. doi:10.1016/j.foreco.2009.09.023 CrossRefGoogle Scholar
  26. Mäkelä A, del Río M, Hynynen J, Hawkins MJ, Reyer C, Soares P, van Oijen M, Tomé M (2012) Using stand-scale forest models for estimating indicators of sustainable forest management. For Ecol Manag 285:164–178. doi:10.1016/j.foreco.2012.07.041 CrossRefGoogle Scholar
  27. Mariñas G (2014) Análisis de las opiniones de los stakeholders sobre la gestión forestal en sistemas forestales de montaña. El caso del Sistema Central, Proyecto Fin de CarreraGoogle Scholar
  28. Maroschek M, Rammer W, Lexer J (2014) Using a novel assessment framework to evaluate protective functions and timber production in Austrian mountain forests under climate change. Reg Environ Change. doi:10.1007/s10113-014-0691-z Google Scholar
  29. Mason B, Mencuccini M (2014) Managing forests for ecosystem services—can spruce forests show the way? Forestry 87(2):189–191. doi:10.1093/forestry/cpu002 CrossRefGoogle Scholar
  30. McElhinny C (2002) Forest and woodland structure as an index of biodiversity: a Review. Department of Forestry, Australian National University, CanberraGoogle Scholar
  31. Millenium Ecosystem Assessment (MEA) (2005) Ecosystems and human well-being: current state and trends. Island Press, WashingtonGoogle Scholar
  32. Montero G (1994) Generalities on silviculture of Pinus sylvestris L. in Spain. Investigación Agraria. Sistemas y Recursos Forestales, Fuera de serie 3:251–257Google Scholar
  33. Ojea E, Martín-Ortega J, Chiabai A (2012) Defining and classifying ecosystem services for economic valuation: the case of forest water services. Environ Sci Policy 19–20:1–15. doi:10.1016/j.envsci.2012.02.002 CrossRefGoogle Scholar
  34. Palahí M, Mavsar R, Gracia C, Birot Y (2008) Mediterranean forests under focus. Int For Rev 10(4):676–689. doi:10.1505/ifor.10.4.676 Google Scholar
  35. Pardos C, Calama R, Maroschek M, Rammer W, Lexer MJ (2015) A model-based analysis of climate change vulnerability of Pinus pinea stands under multi-objective management in the Northern Plateau of Spain. Ann For Sci. doi:10.1007/s13595-015-0520-7 Google Scholar
  36. Poetzelsberger E, Wolfslehner B, Hasenauer H (2015) Climate change impacts on key forest functions of the Vienna Woods. Eur J For Res 134(3):481–496. doi:10.1007/s10342-015-0866-2 CrossRefGoogle Scholar
  37. R Core Team (2014) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org/
  38. Ray D, Bathgate S, Moseley D, Taylor P, Nicoll B, Pizzirani S, Gardiner B (2014) Comparing the provision of ecosystem services in plantation forests under alternative climate change adaptation management options in Wales. Reg Environ Change. doi:10.1007/s10113-014-0644-6 Google Scholar
  39. Revelle W (2014) Procedures for personality and psychological research. Northwestern University, Evanston. http://CRAN.R-project.org/package=psych
  40. Sabaté S, Gracia CA, Sánchez A (2002) Likely effects of climate change on growth of Quercus ilex, Pinus halepensis, Pinus pinaster, Pinus sylvestris and Fagus sylvatica forests in the Mediterranean region. For Ecol Manag 162:23–37. doi:10.1016/S0378-1127(02)00048-8 CrossRefGoogle Scholar
  41. Sánchez-de-Dios R, Benito-Garzón M, Sainz-Ollero H (2009) Present and future extension of the Iberian submediterranean territories as determined from the distribution of marcescent oaks. Pl Ecol 204(2):189–205. doi:10.1007/s11258-009-9584-5 CrossRefGoogle Scholar
  42. Sánchez-Palomares O, López-Senespleda E, Roig-Gómez S, Vázquez de la Cueva A, Gandullo-Gutiérrez JM (2012) Las estaciones ecológicas actuales y potenciales de los rebollares españoles. Monografías INIA: Serie Forestal 17:343Google Scholar
  43. Seidl R, Lexer MJ, Jäger D, Hönninger K (2005) Evaluating the accuracy and generality of a hybrid patch model. Tree Physiol 25:939–951. doi:10.1139/x10-235 CrossRefGoogle Scholar
  44. Seidl R, Rammer W, Lexer MJ (2011a) Climate change vulnerability of sustainable forest management in the eastern Alps. Clim Change 106(2):225–254. doi:10.1007/s10584-010-9899-1 CrossRefGoogle Scholar
  45. Seidl R, Rammer W, Lexer MJ (2011b) Adaptation options to reduce climate change vulnerability of sustainable forest management in the Austrian Alps. Can J For Res 41(4):694–706. doi:10.1139/x10-235 CrossRefGoogle Scholar
  46. Serrada R, Montero G, Reque JA (eds) (2008) Compendio de Selvicultura en España. INIA-Fundación Conde del Valle de Salazar, 1178 pGoogle Scholar
  47. Spathelf P, van der Maaten E, van der Maaten-Theunissen M, Campioli M, Dobrowolska D (2014) Climate change impacts in European forests: the expert views of local observers. Ann For Sci 71(2):131–137. doi:10.1007/s13595-013-0280-1 CrossRefGoogle Scholar
  48. Steenberg JWN, Duinker PN, Bush PG (2011) Exploring adaptation to climate change in the forests of central Nova Scotia, Canada. For Ecol Manag 262:2316–2327. doi:10.1016/j.foreco.2011.08.027 CrossRefGoogle Scholar
  49. Thornton PE, Running SW (1999) An improved algorithm for estimating incident daily solar radiation from measurements of temperature, humidity, and precipitation. Agric For Meteor 93(4):211–228. doi:10.1016/S0168-1923(98)00126-9 CrossRefGoogle Scholar
  50. Turner MG, Donato DC, Romme WH (2013) Consequences of spatial heterogeneity for ecosystem services in changing forest landscape: priorities for future research. Landscape Ecol 28:1081–1097. doi:10.1007/s10980-012-9741-4 CrossRefGoogle Scholar
  51. Van den Meerschaut D, Vandekerkhove K (1998) Development of a stand-scale forest biodiversity index based on the state forest inventory. In: Hansen M, Burk T (eds) Integrated tools for natural resources inventories in the 21st century. USDA, Boise, pp 340–349Google Scholar
  52. Vieira J, Campelo F, Nabais C (2008) Age-dependent responses of tree-ring growth and intra-annual density fluctuations of Pinus pinaster in Mediterranean climate. Trees 23(2):257–265. doi:10.1007/s00468-008-0273-0 CrossRefGoogle Scholar
  53. Wang S, Fu B (2013) Trade-offs between forest ecosystem services. For Policy Econom 26:145–146. doi:10.1016/j.forpol.2012.07.014 CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2016

Authors and Affiliations

  • Marta Pardos
    • 1
  • Susana Pérez
    • 2
  • Rafael Calama
    • 1
  • Rafael Alonso
    • 1
  • Manfred J. Lexer
    • 2
  1. 1.Department of Silviculture and Forest System ManagementINIA-CIFORMadridSpain
  2. 2.Institute of SilvicultureUniversity of Natural Resources and Life Sciences, ViennaViennaAustria

Personalised recommendations