Modeling the links between biodiversity, ecosystem services and human wellbeing in the context of climate change: Results from an econometric analysis of the European forest ecosystems

Ecological Economics 97 (2014) 60–73

Contents lists available at ScienceDirect

Ecological Economics journal homepage: www.elsevier.com/locate/ecolecon

Methodological and Ideological Options

Modeling the links between biodiversity, ecosystem services and human wellbeing in the context of climate change: Results from an econometric analysis of the European forest ecosystems Helen Ding a,⁎, Paulo A.L.D. Nunes b,c a b c

Biodiversity Governance Research Unit (BIOGOV), Center for Philosophy of Law (CPDR), Université catholique de Louvain, Belgium WAVES — Wealth Accounting and Valuation of Ecosystem Services, Agriculture and Environmental Services Department, The World Bank, Washington, DC 20433, USA Department of Agricultural and Resource Economics — TESAF, University of Padova, Campus di Agripolis, Viale dell'Università, 16, 35020 Legnaro, Pd, Italy

a r t i c l e

i n f o

Article history: Received 9 November 2012 Received in revised form 14 September 2013 Accepted 9 November 2013 Available online 1 December 2013 Keywords: Natural capital index Composite biodiversity indicator European forest ecosystem services Simultaneous equation modeling 3SLS Nature-based policy for climate change mitigation

a b s t r a c t This paper constitutes a first attempt to model the relationship between climate change, biodiversity, and ecosystem services, with a specific emphasis on European forests. Firstly, we construct a composite biodiversity indicator that integrates quantitative and qualitative changes of biodiversity projected to 2050 for the EU-17 under future IPCC scenarios. Secondly, this indicator is integrated into two simultaneous equation models to capture the marginal impacts of changes in biodiversity on the value of ecosystem goods and services (EGS) due to climate change. Our estimation results confirm the role of biodiversity as a nature-based policy solution for climate change mitigation, shedding light on the policy actions that generate co-benefits by enhancing ecosystems' capacity to mitigate climate change impacts, while conserving biodiversity and sustaining the flows of EGS for human livelihoods. Especially, nature-based mitigation policies are more cost-effective and better at coping with the ethic and inequality issues associated with distributional impacts of the policy actions, compared to the pure technical solutions to improving energy efficiency and reducing emissions. However, the strength of biodiversity as a nature-based policy option for climate change mitigation depends on both the nature of the EGS and the geographical area under consideration. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Current model projections have consistently indicated that biodiversity would continue to decline over the 21st century, under different socioeconomic scenarios with trajectories of key indirect drivers of ecological changes, such as human population growth and greenhouse gas emissions (Leadley et al., 2010; Pereira et al., 2010). This in turn will impose threats to the benefits of future humanity and result in a change in our production and consumption patterns in the long run (Martens et al., 2003), as biodiversity underpins a variety of ecosystem goods and services (EGS) that are vital to human well-being. Biodiversity by definition encompasses the variety of life on earth from genes to species, through to the broad scale of ecosystems across time and space (Loreau et al., 2001). It is important in terms of determining the health of ecosystem, ensuring the stability and productivity of ecosystem, as well as contributing directly or indirectly to human wellbeing ⁎ Corresponding author at: BIO IS-Deloitte, Deloitte's Sustainability Services in France, Deloitte Touche Tohmatsu Limited, 185 avenue Charles De Gaulle, 92524 Neuilly, Ile-deFrance, France. Tel.: +33 652422258. E-mail address: [email protected] (H. Ding). 0921-8009/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ecolecon.2013.11.004

(MEA, 2005). In this regard, the term “biodiversity” is used broadly as an assumed foundation for ecosystem processes, rather than simply the changing number of species on a species list (Loreau et al., 2001). The relationship between biodiversity and ecosystem functioning or primary productivity has been of long-standing interest to ecologists (Cameron, 2002; Kinzig et al., 2001; Loreau et al., 2001, 2002). Over the past years, the subject has been researched in various ways via experimental field research, the formulation of mechanistic theories and quantitative field observation, most of which have led to a common conclusion that a large variety of species has a positive influence on the productivity and stability of ecosystems, as greater biodiversity can cope with various circumstances in a given habitat and thus lead to the more efficient use of available natural resources (Loreau et al., 2001; Martens et al., 2003). Nonetheless, quantifying the link between biodiversity and ecosystem goods and services remains a major scientific challenge to date (Pereira et al., 2010), because there does not exist a general ecological relationship between ecosystem function and diversity owing to species-specific effects and important tropic links (Paine, 2002; Willims et al., 2002). Certainly, biodiversity loss will negatively affect ecosystem functioning by changing the composition and distribution of species (Bloger, 2001; Giller and O'Donovan, 2002; Loreau et al.,

H. Ding, P.A.L.D. Nunes / Ecological Economics 97 (2014) 60–73

2001; Schmid et al., 2000), which may have far-reaching socioeconomic consequences in the future, through the provision of ecosystem services to human society (Martens et al., 2003). In order to quantify the loss of biodiversity, scientists rely on the use of biodiversity indicators to measure and monitor the different dimensions of biodiversity and to predict the future trends of biodiversity and ecosystems. At a global scale, there are roughly 40 potential measures being developed for the Convention on Biological Diversity (CBD) and about 26 indicators being considered in the Streaming Biodiversity Indicators in Europe 2010 process (Mace and Baillie, 2007). Each of these indicators has been developed to reflect a specific attribute and/ or issue of concern. Nevertheless, for the purpose of public and business decisions and effective communication with broader audience, there is a general need of creating a single, simple or composite biodiversity measure that can both encompass essential biological information and incorporate socio-economic impacts. Moreover, such kind of indicator may also have broader application in the socio-economic research in terms of explicitly quantifying and evaluating the effect of biodiversity loss on human welfare. In fact, the economics literature has shown many attempts to both conceptualize and value biodiversity, exploring the use of stated- and revealed-preference valuation methods, both of which intend to estimate the marginal impact of biodiversity loss on utility (Kontoleon et al., 2007; Nunes and van den Bergh, 2001). These methods have been largely used to estimate the nonmarket values of biodiversity. On the other hand, biodiversity also has considerable market value through the supply of important inputs for economic production. Thus, when estimating (at the margin) the economic value of biodiversity, this exercise should encompass biodiversity's impact, on, or biodiversity's contribution to the ecosystems capacity to provide goods and services, including provisioning, cultural, regulating and supporting services1 (see Chiabai et al., 2011; Ding et al., 2010). Nonetheless, numerical analysis of the links between biodiversity and human wellbeing remains exceptional in the literature. In this regard, only two studies have attracted our particular attention, both of which exploring the use of different biodiversity indicators, i.e. species richness and threatened flora and fauna indexes in modeling the effect of biodiversity loss in the value of ecosystem services or ecosystem productivity. The first refers to a recent study conducted by Costanza et al. (2007), who numerically demonstrated a positive relationship between species richness and net primary production (NPP) for the US, followed by Ojea et al. (2010), who employed the use of metaanalysis to greatly extend the regional forest ecosystem valuation studies to a global scale. These indicators partially explain (not sufficiently enough) the causality between biodiversity loss and changes in ecosystem services and human welfare, but some other important information may be lost as most of the individual biodiversity indicators deal with only one biodiversity attribute or a specific policy target. In this context, the present paper constitutes a first attempt to model, and empirically estimate, the relationship between climate change, biodiversity and EGS, by constructing a news composite biodiversity indicator that integrates essential information of species changes (e.g. change in species richness and abundance) and ecosystem changes (e.g. change in the area of particular biomes). This indicator is expected to be a simple, but comprehensive, measure to map quantitative and qualitative changes of biodiversity projected to 2050 for seventeen European countries under future climate-change scenarios. Furthermore, this indicator is integrated into a set of constructed simultaneous 1 As defined by the Millennium Ecosystem Assessment (MEA) (2005), provisioning services are the goods obtained from ecosystems and they include food, fiber, fresh water, and genetic resources. Regulating services include benefits obtained from the regulation of ecosystem processes, including air quality regulation, climate regulation, water regulation, erosion regulation, pollination and natural hazard regulation. Cultural services are the nonmaterial benefits that people obtain from the ecosystem through esthetic experience, reflection, recreation and spiritual enrichment. Supporting services refer to an ecosystem's life supporting function, which will ultimately influence the provision of other three types of ecosystem goods and services.

61

equation systems to allow formally estimating the marginal impacts of changes in biodiversity on the value of ecosystem goods and services due to climate change. Data availability with regards to both biological species and economic values of the ecosystem services leads us to focus on the forest ecosystems in Europe. The organization of the paper is as follows. Section 2 introduces the concept of climate-change scenarios and describes the source of data used in the research. Section 3 presents the empirical and innovative approach that is characterized by the creation of a composite biodiversity indicator. Section 4 presents and discusses the theoretical model, which is characterized by the use of simultaneous equations. Section 5 presents and compares the empirical results for both model specifications, i.e. a European-aggregated, and a European-regional model specification, respectively. Section 6 discusses the impact of the estimation results on the design and implementations of the EU environmental policies. Section 7 concludes. 2. Data Description 2.1. Future Climate Change Scenarios For a comprehensive interpretation of climate change scenarios and the respective socio-economic and biological impacts, it is an essential first step to understand the underlying assumptions of the scenarios under consideration. Scenarios do not predict the future, but rather paint pictures of possible futures and explore the various outcomes that might result if certain basic assumptions are changed. In order to explore the possible future patterns of biodiversity in Europe, the scenarios are developed based on the recent efforts of the Intergovernmental Panel on Climate Change (IPCC) (2000), which explore the global and regional dynamics that may result from changes at a political, economic, demographic, technological and/or social level. The distinction between classes of scenarios is broadly structured by defining them ex ante along two dimensions. The first dimension relates to the extent of economic convergence, and of social and cultural interactions across regions; the second has to do with the balance between economic objectives and environmental and equality objectives. This process therefore leads to the creation of four climate change scenarios, namely A1, A2, B1 and B2. Hereafter, we call them IPCC storylines throughout the paper. Table 1 below summarizes the political, economic, demographic, technological and social assumptions made in each of the IPCC storyline and analyzes their potential impacts on the future patterns of global biodiversity. As we can see, scenario A1 and A2 are both economic-oriented scenarios, but with differentiated focuses on global and regional economic development, respectively. In particular, scenario A2 represents a world differentiated into a series of consolidated economic regions characterized by low economic, social, and cultural interactions, uneven economic growth and with the income gap between industrialized and developing countries that does not narrow. Alternatively, the B-type scenarios depict a world, where economic objectives and environmental and equity objectives are more balanced. In particular, scenario B1 shows that environmental and social consciousness can be combined in a more sustainable manner at global scale, offering a more favorable perspective for biodiversity than the A-type scenarios. Moreover, technological development is expected to shift towards renewable energy and higher productivity and consequently reduce the pressure on natural ecosystems from decreased pollution and land conversion. Finally, biodiversity will also benefit from lower pressure of global population growth and improved ecological capital. Similar to scenario B1, the B2 scenario is environmentally oriented with a focus on local environmental and social sustainability. In this scenario, average education level and degrees of organization within communities are high and energy and material efficiency can be achieved. All these social and technological achievements can reduce the pressure on natural ecosystem.

62

H. Ding, P.A.L.D. Nunes / Ecological Economics 97 (2014) 60–73

Table 1 IPCC scenarios of future global biodiversity patterns. Source: adapted from Martens et al. (2003) and ATEAM model assumption. Storyline

Key assumptions

Summary of major effects of the scenario

Impacts on biodiversity

A1 (offers an unfavorable perspective for biodiversity)

Slight population increase till 2050, then decrease; very rapid economic growth; high level of income; a global mean increase in temperature of at least 4.4 °C (std 0.9) toward 2080; forest area is stable due to increasing timber demand and recreational land use pressure; and significant conversion of agricultural land from food to bioenergy production. Continually growing human population (15 billion by 2100); slow economic growth; economic development is primarily oriented and uneven; regional self-reliance in terms of resources; weak global environmental concern; Total consumption of natural resources is considerable; a global mean increase in temperature of at least 3.5 °C (std 0.7) toward 2080; slightly decrease of forest area; and significant conversion of agricultural land from food to bioenergy production and human settlement. A sharp reduction in arable farming and cattle breeding acreage due to a strong increase in productivity; the estimated temperature increase is about 2.7 °C (std 0.6) toward 2080; pressure from population growth is considerably lower; forest area increases; and significant conversion of agricultural land from food to bioenergy production and human settlement. The pressure on natural system is greatly reduced due to high average education levels and high degree of organization within communities; stable population; relatively slow economic development; regionally and locally oriented environmental policies are successful; a global mean increase in temperature of at least 2.0 °C (std 0.7) toward 2080; and land-use changes from food to bioenergy production or forestry.

Many pristine natural areas are converted into man-made areas; costs of preserving natural areas are very high due to increase in land prices; reduced ecosystem quality due to increased population densities, increased tourism, etc.; and higher concentrations of GHG due to a substantial increase in energy use and land conversion

Patterns of bird and herptile species richness will not change dramatically; and plant and tree species richness will decrease in the southern part of Europe but increase in central and Scandinavian Europe.

Sharply increasing demand for foods, water, energy and land will result in a significant loss of natural ecosystems and species; regional competition for good-quality natural resources will negatively affect the economic conditions in these countries and reduce attention for the preservation of natural resources; and an increasing number of people will compete for a declining number of natural resources at the cost of quantity and quality of those remaining resources.

Patterns of bird and herptile species richness will not change dramatically; and plant and tree species richness will decrease in the southern part of Europe but increase in central and Scandinavian Europe.

A lot is done to improve ecological capital and therefore reduce threatening factors and prospects for biodiversity; cropland production is concentrated in optimal locations; and grassland is protected by policy.

Natural ecosystems are less affected both in quantity and quality

The general picture of biodiversity in the future largely depends on the introduction of socioeconomic policies that support local and regional initiatives to achieve structural solutions.

Hard to estimate global biodiversity trend due to the high heterogeneity

A2 (offers a heterogeneous world)

B1 (offers a more favorable perspective for biodiversity)

B2 (very locally concentrated social, economic and ecological problems)

2.2. Socio-economic, Climatic and Ecological Data Projected Under Future IPCC Scenarios Under different IPCC storylines, projections have been developed to describe possible outcomes of different political, economic, demographic, technological and social assumptions for the future development. These include the projected trends of GDP, population, incremental temperature, ecosystem productivity, and distribution of species and so on, subject to the changes in a set of key assumptions on which the IPCC storylines are based. In this study, we explore the use of climatic, socio-economic and ecological projections to investigate the pressure on biodiversity and to quantify the consequent quality and quantity changes of terrestrial biodiversity following the four future development paths defined by IPCC. Since we have empirical evidence showing that the impacts of changing climate conditions are highly spatially heterogeneous, as organisms, populations and ecological communities do not respond to approximated average of global warming (Walther et al., 2002), we opt to work at the country level. This however, requires a strong investment in data, both in terms of data from earth observation systems (e.g. current land use patterns) as well as from simulation model architectures (e.g. projection of species diversity for 2050). In this context, we opt to work with all the European countries that report (projected) values for biological species under the four IPCC storylines under consideration. To account for regional climate differences/ commonalities, where similar climatic patterns and taxa might be identified, we propose to map the EU-17 in terms of three geoclimatic clusters, including the Mediterranean Europe (Greece, Italy, Portugal, Spain), the Central North Europe (Austria, Belgium, France, Germany, Ireland, Luxembourg, Netherlands, Switzerland, United Kingdom), and the Scandinavian Europe (Denmark, Finland, Norway,

Sweden). The data used are independently published by a number of IPCC data distribution centers across the world for 2050, downscaled at country level. The demographic and economic trends represented by the future per capita GDP, population density are projected and distributed by the Center for International Earth Science Information Network (CIESIN) (2002) at Columbia University. The annual mean temperature are projected by the Tyndall Centre in the UK (www. tyndall.ac.uk), who combines the use of Global Circulation Models/ SRES (including CGCM2, CSIRO2, HadCM3 and PCM) to estimate the possible increase of temperature in degrees Celsius for each country under different IPCC scenarios. The biophysical changes of biodiversity comprise the quantitative changes measured in terms of changes in the area of forest habitat, and the qualitative change indicated by changes in the number of terrestrial species (including plant, tree, bird and herptile). The future trends of these changes under IPCC scenarios are projected in the frame of the Advanced Terrestrial Ecosystem Analysis and Modeling (ATEAM) project (Schröter et al., 2004). In particular, species richness under current and future conditions are projected taking into account total 383 bird species, 108 reptile and amphibian species, 1350 plant species and 125 tree species appeared in the EU. To keep the consistency across a large range of data sources, we derive all data from projections that represent a combination with the HadCM32 model, which directly relates socioeconomic changes to climatic changes through greenhouse gas concentration, and relates land use changes through climatic and socioeconomic drivers such as the demand for food (Schröter et al., 2004).

2 HadCM3, Hadley Centre Couplet Model Version 3 is a coupled atmosphere–ocean GCM developed at the Hadley Centre and described by Gordon et al. (2000).

H. Ding, P.A.L.D. Nunes / Ecological Economics 97 (2014) 60–73

63

Fig. 1. Trends of ecosystem quality and quantity using NCI. Source: Ten Brink (2000) pp. 2.

Our knowledge about to what extent biodiversity can respond to climate change is limited and the quantification of associated economic gains or losses to human welfare cannot be straightforward but through valuing biophysical changes of ecosystem services under future climate conditions. In this study, values of EGS provided by the European forests are taken from Ding et al. (2010), who provided detailed projections of ecosystem values following four future IPCC storylines vis-à-vis to the baseline year of 2000. The valuation exercises were conducted separately for three types of forest ecosystem services defined in Millennium Ecosystem Assessment, i.e. provisioning, regulating and cultural services (MEA, 2005). More specifically, forest provisioning services contained the benefits derived from the production of timber and other wood forest products, regulating services provided non-monetary benefits from CO2 sequestration in the forest, and cultural services provided humans with direct incomes from the related tourism industries and non-monetary benefits from the enjoyment of existing forests. All values were first projected to 2050 and then adjusted to 2005 US$. All original data used in the present study are presented in the Annex — see also Ding et al. (2010) and Schröter et al. (2004) for more details. 3. Biodiversity Metrics 3.1. The Rationale for the Use of a Composite Indicator As previously mentioned, there is a growing interest among scientists to develop a single, simple or composite biodiversity measure that is able to encompass essential biological information, incorporate socioeconomic impacts, and guide policy interventions towards more effective biodiversity management. A simple format may allow it to be more influential in the public and business decision-making and more effective in communications, just like the use of Gross Domestic Production (GDP) in economic analysis and the Dow Jones indicator in stock market (Balmford et al., 2005; Halpern et al., 2012; Mace and Baillie, 2007). To this extent, the existing biodiversity data will be useful for developing quantitative scenarios of the future trajectories of biodiversity (Pereira et al., 2010). Moreover, from a methodological perspective, there is a general need of creating a workable “calculus” of biodiversity that allows not just global summation, but also estimation of the more localized marginal gains and losses from global changes induced by socioeconomic development and land use changes in different places (Faith, 2005). This is particularly the case when assessing climatechange-induced biodiversity changes, because individual indicators alone, such as species richness or abundance of a certain species, do not provide sufficient information to enable a better understanding of the impacts of increased temperature or precipitation rate on the ecosystem functioning and overall performance. Therefore, this section will focus on the development of a methodological approach to construct such a composite biodiversity indicator, which will serve as a biodiversity variable in the econometric model later on. By far, a number of composite indicators have been developed in the literature. For example, the Natural Capital Index (NCI) is constructed as a weighted sum of the product of the extent of each ecosystem (relative

to a baseline) with the condition of the ecosystem, where the condition is measured as the population size of a group of indicator species relative to a baseline (ten Brink, 2000). A similar indicator is the Biodiversity Intactness Index recently developed by Scholes and Biggs (2005), who also takes into account different ecosystems being weighted by their species richness and population size being estimated for each land-use class in each ecosystem. Apparently, the latter requires more detailed information of species under each type of land-use. Our biodiversity data are limited, as we have only the country data on species richness projected under future climate scenarios. Thus we propose to adopt the NCI approach to construct an NCI-like composite biodiversity indicator for analyzing climate change impacts on the biodiversity and ecosystem services in Europe. This way, the future state of ecosystem in both quality and quantity terms under different climate change scenarios may be assessed with respect to a selected baseline. NCI framework considers biodiversity as a natural resource containing all species with their abundance, distribution, and natural fluctuations. Human direct and indirect interference may affect ecosystem size (through land conversion) and exert pressures on ecosystem quality (such as over-exploitation and fragmentations). As a result, both decreased ecosystem quantity and quality will lead to the loss of biodiversity. In this context, the development of NCI framework aims at providing a quantitative and meaningful picture of the state of and trends in biodiversity to support policymakers in a similar way as GDP, employment and Price Index do in economics. Moreover, the structure of NCI also allows the analysis of socio-economic scenarios on their effect on biodiversity. In technical terms, NCI is the product of changes in the size of ecosystems (“ecosystem quantity”) and the changes in abundance of a core set of species (“ecosystem quality”) within the remaining ecosystem, where both quality and quantity are expressed relative to an “optimal” or “intact” baseline (ten Brink, 2000).

Equation of the NCI: NCI = ecosystem quality (% of species abundance) × ecosystem quantity (% area of the country) (Source: ten Brink, 2000)

Originally, the NCI chooses the use of less modified “pre-industrial baseline” so that major anthropogenic impacts on the changes of biodiversity quality (e.g. loss of species abundance) and quantity (e.g. loss of natural habitat) can be observed and compared. The NCI score ranges from 0 to 100% representing an entire deteriorated (0%) and intact ecosystem (100%), respectively. It summarizes the extent to which a landscape has preserved its original (baseline) natural capital and enables the analysis of biodiversity effects in different socio-economic scenarios. Obviously, one of the advantages of the NCI is that it allows us to aggregate many biodiversity parameters to a few or perhaps a single, more or less representative biodiversity index for the entire ecosystem — see Fig. 1.

64

H. Ding, P.A.L.D. Nunes / Ecological Economics 97 (2014) 60–73

Species richness in 2000

Forest biological diversity

IPCC scenarios in 2050

Country average biodiversity quality

Country ecosystem quantity

Constructed NCIlike indicator

Plant

A1

SBIA1

x

ForestA1

CFBIA1

Tree

A2

SBIA2

x

ForestA2

CFBIA2

Bird

B1

SBIB1

x

ForestB1

CFBIB1

Herptile

B2

SBIB2

x

ForestB2

CFBIB2

Baseline (2000)

Target (2050)

Fig. 2. Constructing a NCI-like indicator to estimate the trend of biodiversity in future IPCC scenarios (note: SBI refers to the aggregated average score of species richness of plant, tree, bird and herptile species).

3.2. Constructing a Composite Forest Biodiversity Indicator In the present study, an NCI-like composite biodiversity indicator, i.e. Composite Forest Biodiversity Indicator (CFBI) is constructed following a three-step approach, to measure the overall change of forest biodiversity in Europe between 2000 and 2050 under different IPCC scenarios. The first step is to compute the average changes of forest ecosystem quality, which contains the changes of total numbers of four core species, i.e. tree, plant, bird and herptile, for each of the EU-17 countries, under different climate change scenarios. Note that for each country, the change of individual species under each scenario is expressed as the ratio between species richness of the species in 2050 and that of the baseline. Furthermore, we aggregated individual percentage changes of species richness for tree, plant, bird and herptile to get a country average score, which depicts the changes of country's forest ecosystem quality under each IPCC scenario with respect to the baseline. The second step is to calculate the forest ecosystem quantity, which is expressed as the percentage of a country's forest coverage to its total land area under different IPCC scenarios. The third step is to produce the CFBI, which is the product of percentage changes of forest ecosystem quality (calculated in step 1) and the percentage changes of forest ecosystem quantity (calculated in step 2). Therefore, we may expect that the computed CFBI score can also reflect the direct land-

use change impacts on biodiversity. In particular, the expansion of forest area in many parts of Europe can have a positive impact on the CFBI score. Fig. 2 presents a flow chart to visualize how our NCI-like composite biodiversity indicator is constructed. It is important to note that our data cover only the post-industrial era, during which many stringent environmental policies have been successfully implemented among the 17 most developed European economies, in terms of pollution reduction, sustainable resource management and greening economy. As a result, many countries are found to have a stable increase in either forest area or increased species richness or both, and thus the original NCI score range cannot apply. Instead, we set up two intervals to indicate the future state of the forest biodiversity: (1) 0 b CFBI b 100% indicates a degradation of forest biodiversity in quantity and/or quality terms in 2050, with respect to the baseline year 2000. This implies that the state of biodiversity in a country is deteriorated because of the decreased forest coverage as a result of land conversion to agricultural production or human settlement, and/or because of the decreased species richness in the country. (2) CFBI N 100% indicates an improved state of forest biodiversity in 2050 with respect to the baseline year 2000. However, such

Fig. 3. Computed CFBI score for the EU-17 under four IPCC storylines.

H. Ding, P.A.L.D. Nunes / Ecological Economics 97 (2014) 60–73

improvement may not be necessarily caused by the increased species richness as a result of some effective conservation measures. It may be due to the increased ecosystem coverage/area as an outcome of national conservation policies. Moreover, it is also important to note that a CFBI score greater than 100% does not necessarily mean that the individual species are not under threats, rather it indicates an overall improvement of the biodiversity status. Thus, to better interpret the CFBI score, we need to look closely at the national/regional forest management policies and their effectiveness. 3.3. Mapping Composite Forest Biodiversity Indicator Across Different IPCC Scenarios The calculated CFBI scores for the EU-17 under four different IPCC scenarios are presented in Fig. 3. Overall, the CFBI score decreases when moving towards the economic oriented development paths (as represented by the A1 and A2 scenarios), as increased pressures from the fast growing economy and population, rising global temperature and unbalanced land-use conversion will most likely worsen future state of forest biodiversity across Europe. Among all others, the warmer region, i.e. Mediterranean Europe suffers the greatest loss of biodiversity quantity and quality in both scenarios compared to the other colder regions. On the contrary, the CFBI score increases following the environmental oriented development paths (as represented by B1 and B2 scenarios) in most of the European countries, except Portugal, Spain, Finland, Norway and Sweden. This implies that the adoption of sustainable forest management practices in Europe is successful in general. However, given the relatively high level of forest status in the Scandinavian countries in the baseline year, we will not foresee any significant increase in forest ecosystem quantity over the next 50 years, independent from the future standpoints. Finally, although the Mediterranean forests appear to be the most vulnerable to global change (Lindner et al., 2010; Sala et al., 2000), for instance, in Portugal and Spain three out of four selected species are projected to decrease by 2050 (see Table A6 in Annex), we can still observe a general improvement of the future state of forest biodiversity in Italy and Greece owing mainly to a mixed effect of increasing in certain species richness and slowing down economic growth and land conversion in the region.

65

Table 2 Descriptive statistics (summary). Variables

Obs Mean

Forest area (fa) Population density (pop_dens) GDP Number of tree species (nts) Number of bird species (nbs) Number of plant species (nps) Number of herptile species (nhs) The composite forest biodiversity indicator (CFBI) Temperature (t) Economic value of provisioning services (EVPS) Economic value of cultural services (EVCS) Economic value of regulating services (EVRS)

68 68 68 68 68 68 68 68

7.02 7.36 0.07 1.24 1.02 0.08 1110.28 1310.00 22.38 38.42 13.51 10.96 130.26 13.58 106.47 259.64 36.52 199.61 20.00 11.04 1.72 1.08 0.34 0.47

68 68

3.69 1.22 1.5 6.9 4776.07 5214.79 100.95 17,600

68 68

454.07

Std. dev.

Min

Max 25.88 3.33 5569.02 70.96 154.31 361.78 39.39 2

568.80

3.13

2615.14

2041.77 2023.33

71.39

7465.75

In this context, we propose to model the relationship between biodiversity, ecosystem and human welfare in a simultaneous equation system. More specifically, we choose to run a three-stage-least-squares (3SLS) regression instead of a two-stage-least-squares (2SLS) regression. Theoretically, 3SLS performs in a similar manner as 2SLS in terms of regressing endogenous regressors against all predetermined variables of the system to get theoretical values (Verbeek, 2000), but the main difference is that 2SLS focuses on individual equations, which causes inefficiency. 3SLS is preferred in this study because it considers not only the simultaneous correlations between various equations' error terms, but also the inter-temporal (and not simultaneous) correlations between error terms, which has an obvious advantage in assessing climate change impacts at different points in time. In the present study, the simultaneous structural system contains the following three equations: ln ðEV i Þ ¼ β10i þ β11i ln ðfaÞ þ β12i ln ðt Þ þ β13i CFBI þ ε1i

ð1Þ

ln ðfaÞ ¼ β20 þ β21 ln ðGDPÞ þ β22 ln ðt Þ þ β23 ln ðpopdens Þ þ ε2

ð2Þ

2

CFBI ¼ β30 þ β31 t þ β32 t þ β33 nts þ β34 nbs þ β35 nps þ β36 nhs þβ37 ln ðpopdens Þ þ β38 ln ðGDP Þ þ ε3

ð3Þ

where 4. The Econometric Model EV 4.1. European-aggregated Model Specification This section focuses on econometric model specification so as to capture the marginal impacts of changes in biodiversity on the value of EGS due to climate change. The CFBI will be a constituent of our model. In particular, we would like to test: (1) Whether climate change, here expressed as an increase in temperature, will alter the pattern of biodiversity distribution and species richness presented in a geo-climatic zone, which is measured by the CFBI. In particular, we want to test (i) whether increases in temperature will have effects over the CFBI and across geo-climatic clusters. (2) Whether the climate-change-induced CFBI changes will further affect the ecosystem's capacity of providing goods and services and their respective values. Similarly, we want to test whether climate-change-induced CFBI changes will have effects over the economic value of EGS. In particular, we want to assess (ii) whether this effect will change across the geo-climatic regions, and (iii) whether it will differ depending on the type of EGS under consideration.

the estimated economic value of the ith type of ecosystem service by 2050 under future IPCC scenarios (in million $) i 1, 2 and 3 types of ecosystem services under consideration, i.e. provisioning services (PS), cultural services (CS), and regulating services (RS) fa projected forest area (million ha) in 2050 under future IPCC scenarios t increased Celsius degrees of local temperature by 2050 under future IPCC scenarios CFBI Composite Forest Biodiversity Indicator is measured as the changes (%) of average biodiversity status between 2000 and 2050 GDP projected gross domestic production (billion $) in 2050 under IPCC scenarios Pop_dens projected population density (heads/ha) in 2050 under IPCC scenarios nts number of tree species projected in 2050 under IPCC scenarios nbs number of bird species projected in 2050 under IPCC scenarios nps number of plant species projected in 2050 under IPCC scenarios nhs number of herptile species projected in 2050 under IPCC scenarios.

66

H. Ding, P.A.L.D. Nunes / Ecological Economics 97 (2014) 60–73

Table 3 Aggregated model — estimation results. Provisioning services Eq. (1) (2) (3)

“R-sq” 0.401 0.534 0.615

Cultural services chi2

P-value

Eq.

54.25 78.04 141.77

0.000 0.000 0.000

(1) (2) (3)

Regulating services “R-sq” 0.931 0.537 0.624

chi2

P-value

Eq.

548.13 78.80 138.74

0.000 0.000 0.000

(1) (2) (3)

Eq. (1)

Eq. (1)

Eq. (1)

Dep. var.: lnEVi

Dep. var.: lnEVi

Dep. var.: lnEVi

Var. lnfa lnt cfbi

Coef.

z

0.671 1.032 2.299

5.68 2.04 3.94

P N |z|

Var.

Coef.

z

P N |z|

Var.

0.000 0.041 0.000

lnfa lnt cfbi

1.060 −0.664 −0.895

22.19 −3.24 −3.73

0.000 0.001 0.000

lnfa lnt cfbi

Eq. (2)

Eq. (2)

Dep. var.: lnfa

“R-sq” 0.839 0.536 0.636

Coef.

chi2

P-value

198.57 79.36 135.69

0.000 0.000 0.000

P N |z|

z

0.740 0.670 1.202

12.27 2.59 3.97

0.000 0.010 0.000

Eq. (2)

Dep. var.: lnfa

Dep. var.: lnfa

Var.

Coef.

z

P N |z|

Var.

Coef.

z

P N |z|

Var.

Coef.

z

P N |z|

lnGDP lnt lnpop_dens

0.850 0.854 −0.453

8.00 2.16 −3.65

0.000 0.030 0.000

lnGDP lnt lnpop_dens

0.837 0.819 −0.539

7.85 2.07 −4.27

0.000 0.038 0.000

lnGDP lnt lnpop_dens

0.836 0.813 −0.555

7.84 2.06 −4.40

0.000 0.040 0.000

Eq. (3)

Eq. (3)

Dep. var.: CFBI

Dep. var.: CFBI

Eq. (3) Dep. var.: CFBI

Var.

Coef.

z

P N |z|

Var.

Coef.

z

P N |z|

Var.

Coef.

z

P N |z|

t t2 nts nbs nps nhs lngdp lnpop_dens

−0.492 0.054 0.016 0.004 −0.001 0.001 0.022 0.046

−4.48 4.01 5.03 1.72 −0.91 0.72 1.01 1.85

0.000 0.000 0.000 0.085 0.363 0.474 0.311 0.064

t t2 nts nbs nps nhs lngdp lnpop_dens

−0.519 0.058 0.017 0.001 −0.001 −0.000 0.024 0.022

−4.51 4.12 5.33 0.45 −0.60 −0.04 1.12 0.86

0.000 0.000 0.000 0.653 0.548 0.972 0.264 0.391

t t2 nts nbs nps nhs lngdp lnpop_dens

−0.494 0.055 0.020 −0.000 −0.001 0.005 0.032 0.014

−4.28 3.90 6.20 −0.02 −1.07 2.11 1.46 0.55

0.000 0.000 0.000 0.986 0.286 0.035 0.145 0.585

Nr. of observations: 68. Endogenous variables: lnEVi, lnfa and CFBI. Exogenous variables: lnt, lngdp, lnpop_dens, t, t2, nts, nbs, nps and nhs.

We assume that EV, CFBI and fa are endogenous variables in the system and ε1, ε2 and ε3 are the stochastic disturbance terms that capture all unobservable factors, which may influence the dependent variables. In the first two equations, all variables, except CFBI are in their log-transformations indicating that the estimated beta coefficients measure the elasticity of dependent variables with respect to the changes in a set of explanatory variables. As for the warming impact on biodiversity, it is estimated using Eq. (3) by regressing CFBI on temperature variables (t and t2), along with other biological and socio-economic variables that may explain the trends of biodiversity changes in the future scenarios. In particular, the temperature variable t will capture the marginal impact of climate change on biodiversity with increment of 1 °C in the temperature and the squared t is introduced to capture the rate of this change. Finally, the impact of GDP growth in the EU-17 on biodiversity and forest ecosystems is mixed. On the one hand, it may positively affect biodiversity if the future GDP growth goes hand-inhand with enhanced public awareness of environmental and biodiversity protection, as demonstrated by the Environmental Kuznets Curve (EKC) that in the early stages of economic growth degradation and pollution increase, but beyond a turning point, the trend reverses, so that at high-income levels economic growth leads to environmental improvement (De Bruyn, 2000). On the other hand, this impact may be negative, if the future GDP growth is associated with overexploitation of natural capital, which can directly result in the reduction of the ecosystem quantity indicator (i.e. forest land cover) and/or the degradation of the ecosystem quality indicator (i.e. species richness of core species). Ex ante, the final impact is therefore ambiguous and will be depend on the result of the empirical exercise.

In Table 2, we summarize the descriptive statistics of all the variables. For each variable, we have four observations under four IPCC storylines for total 17 countries under consideration, which gives rise to total 68 observations. Next, we proceed with running a 3SLS regression in a global condition, where all data are pooled together without considering the regional heterogeneity of climate change impacts. This allows us to estimate simultaneously (1) the determinants of economic value of ecosystem services; (2) the determinants of land-use changes (i.e. the changes of forest cover); and (3) the determinants of changes in biodiversity. Eq. (1) attempts to explain the economic value of ecosystem services as a function of forest area, average annual temperature and the state of forest biodiversity. We simultaneously test the hypotheses that the enlarged forest area and improved state of biodiversity will positively affect the ecosystem values, whereas the rising temperature may have a negative impact. Since the ecosystem values vary greatly depending on the types of ecosystem services under consideration, we shall treat separately the three types of values to reduce bias. Eq. (2) attempts to explain that the change of forest coverage is determined by three variables: the wealth of the nation (expressed as GDP), the population density and the average annual temperature, but in opposite directions. We expect that GDP growth in the EU-17 is beyond the EKC turning point assumption and will positively affect forest cover, as in these wealthy states increasing demand for forest EGS, including timber product consumption and recreational use of forests (especially natural forests), which in turn will direct the forest management practices towards more sustainable-sound use and management of forest resources. On the contrary, the mounting population

H. Ding, P.A.L.D. Nunes / Ecological Economics 97 (2014) 60–73

in all future IPCC scenarios is assumed to increase the pressure on natural forest land, and lead to more severe competition between natural forests occupation and other land uses, such as agricultural lands or human settlement. Finally, we assume that temperature may play a role in affecting the forest natural regeneration process, but the direction of its impact on forest coverage is ambiguous. Eq. (3) tests statistically whether the RHS variables, such as rising temperature, changes of species richness in terms of a set of key species, and changes of socio-economic and demographic conditions under different climate change scenarios can influence the dependent variable CFBI in different future states. Especially, we are interested in testing whether warmer conditions will negatively affect forest biodiversity, which is integral part of the EGS and the value it provides. Moreover, high population density is expected to impose high pressure on biodiversity through the intensive land conversion from forest habitats to agricultural lands or human settlements. Furthermore, we would like to empirically test how GDP growth can directly affect biodiversity. 4.2. European-regional Model Specification In addition to the “global effect” tested above, this section will focus on testing the presence of heterogeneous effects of regional climatechange-induced CFBI impact on EGS. To do this, the model specification is modified in two steps to count for the three geo-climatic zones, i.e. Mediterranean (M), Central-North (C) and Scandinavian (S) Europe, respectively. First, we introduce a cross-effect between CFBI and regional temperature variables to generate a matrix of CFBI_Tregion, which contains three region specific CFBI variables, namely cfbi_ts, cfbi_tm and cfbi_tc for the S, M, and C regions, respectively. This gives rise to Eq. (4), in which CFBI is replaced by CFBI_Tregion to capture the indirect impacts of climate change on the value of ecosystem services through associated biodiversity effect in the regional forests (i.e. the climatechange-induced biodiversity effect, CCIBE). Second, in Eq. (6), the regional warming effects on biodiversity are captured by introducing a matrix of regional temperature variable, tregion, which consists of three temperature variables ts, tm and tc to represent changing temperatures in the S, M, and C regions, respectively. As a result, we can obtain a modified structural simultaneous system below: ln ðEV i Þ ¼ β10i þ β11i ln ðfaÞ þ β12i ln ðt Þ þ β13i CFBIT region þ ε1i

ð4Þ

ln ðfaÞ ¼ β20 þ β21 ln ðGDPÞ þ β22 ln ðt Þ þ β23 ln ðpopdens Þ þ ε2

ð5Þ

2

CFBI ¼ β30 þ β31 t region þ β32 t þ β33 nts þ β34 nbs þ β35 nps þβ36 nhs þ β37 ln ðpopdens Þ þ β38 ln ðGDPÞ þ ε3 :

ð6Þ

The regional effects can be estimated by repeating the 3SLS regression in this modified structural system. At this stage, we are in conditions to proceed with the empirical estimation of the two simultaneous equation systems, and to estimate the respective beta coefficients. The results are presented and discussed in the following section. 5. Empirical Results 5.1. European-aggregated Model Estimates Table 3 below reports the 3SLS results of the European-aggregated model Eqs. (1)–(3). The goodness of the linear approximation in the structural simultaneous system is assessed based on the coefficients of determination (R2). For most of the simultaneous equations, across the three different types of forest ecosystem services — including provisioning, cultural and regulating services, the R2 statistic is larger than 0.5 (with P N 0.0000) and thus confirm the overall goodness of fit of the performed model specifications. Moreover, most of the estimated beta coefficients carry the expected sign.

67

Firstly, in Eq. (1), it shows that the value of forest ecosystem services is statistically significantly related to the forest size. In other words, an additional hectare of forest will be associated with an increase in the economic value across all the forest ecosystem services under consideration. Bearing in mind the estimation results, expressed in terms of elasticity, vary from 0.67 for the provisioning services, to 0.74 for the regulating services, and to 1.06, for the cultural services. This means that an increase in forest area is always associated with an increase in the economic value of the forest EGS. In particular, 10% increase of the forest size is associated with an increase of 10.6%, 6.7%, and 7.4% increase in the economic value of cultural services, provisioning services, and regulating services, respectively. From a strict cost–benefit analysis perspective, the cultural value of forest is ranked the most valuable forest EGS ceteris paribus, and shall be given careful consideration while making the forest land use planning. Moreover, the estimated coefficients of CFBI reveal to be statistically significant for all ecosystem services, implying that biodiversity has marginal impact on mapping the economic value of forest EGS. The direction of the impact is, however, mixed. On the one hand, the marginal impact of CFBI is found positive for both regulating and provisioning services, which is estimated between 1.202 and 2.299, respectively. On the other hand, the marginal impact of CFBI is found negative for culture services, which is estimated to be 0.895. This means that, on average, a marginal
increase of one-unit of CFBI is associated with 229.9% (¼ eβ3 −1 ñ100percentage) change in the economic value of forest provisioning services and 120.2% change in the economic value of regulating services, ceteris paribus, but with 89.5% decrease in the economic value of cultural services. Since this is an aggregated global effect, we suspect this is due to the fact that pooling data across geoclimatic regions may lead to the neglected spatial heterogeneity of the impacts. We will further investigate on this point using a modified regional model specification. Finally, as expected, changes in the temperature are statistically significant, at the margin, in explaining the economic value of the three forest ecosystem services under consideration. According to the elasticity estimates, we can see that an increase of 10% of the average temperature may result in 10.32% increase of the value of forest provisioning services and 6.7% increase of the value of forest regulating services. These results reconfirm the recent finding in scientific research that the changing climate can increase both forest productivity and carbon stock in the boreal forests in the Scandinavian Europe for at least the short run (Garcia-Gonzalo et al., 2007). However, our estimation results show that temperature impacts are negatively associated with the economic value of forest cultural services. According to the parameter estimate, a 10% increase in the temperature causes, on average, a decrease of 6.64% in the economic value of cultural services provided by the European forests. Secondly, Eq. (2) shows that all selected explanatory variables are statistically significantly related to changing forest coverage. The estimated coefficients of each variable are found similar across all ecosystem services, suggesting the robustness of our results. In particular, our estimation results show that GDP growth is positively correlated with the extension of forest areas. As argued previously, this result is consistent with the phenomena explained by Environmental Kuznets Curve. This indicates that economic growth in the EU-17 will further stimulate the social desire for improved environment and extended forest coverage in Europe. Moreover, the rising temperature is found positively correlated with the forest area, which might be explained by the fact that warmer climate will result in increased timber productivity due to prolonged growing season for boreal forests in Scandinavian Europe (Garcia-Gonzalo et al., 2007), which in turn may stimulate the decision of extending forested area for timber production. Next, the population density is found to have a negative impact on the forest area, independent from the type of ecosystem services under consideration. According to the estimation results, an increase of 10% in the population density causes, on average, a decrease of 4.53% to 5.55% in the

68

H. Ding, P.A.L.D. Nunes / Ecological Economics 97 (2014) 60–73

Table 4 Regional model — estimation results. Provisioning services Eq. (1) (2) (3)

“R-sq” 0.582 0.533 0.643

Cultural services chi2

P-value

Eq.

111.16 77.07 154.25

0.000 0.000 0.000

(1) (2) (3)

Regulating services “R-sq” 0.985 0.537 0.643

chi2

P-value

Eq.

3704.47 79.38 152.49

0.000 0.000 0.000

(1) (2) (3)

Eq. (4)

Eq. (4)

Eq. (4)

Dep. var.: lnEVi

Dep. var.: lnEVi

Dep. var.: lnEVi

Var.

Coef.

Z

P N |z|

Var.

Coef.

lnfa lnt cfbi_ts cfbi_tm cfbi_tc

0.863 0.193 −0.041 −0.493 0.062

8.19 0.41 −0.27 −2.50 0.57

0.000 0.680 0.786 0.012 0.571

lnfa lnt cfbi_ts cfbi_tm cfbi_tc

1.011 −0.290 −0.059 0.279 −0.027

Eq. (5)

Eq. (5)

Dep. var.: lnfa

Dep. var.: lnfa

z 43.18 −2.77 −1.74 6.31 −1.10

“R-sq” 0.874 0.537 0.642

chi2

P-value

345.85 79.37 157.07

0.000 0.000 0.000

P N |z|

Var.

Coef.

z

P N |z|

0.000 0.006 0.082 0.000 0.272

lnfa lnt cfbi_ts cfbi_tm cfbi_tc

0.769 −0.156 0.085 0.251 0.259

13.50 −0.62 1.04 2.38 4.38

0.000 0.536 0.296 0.018 0.000

Eq. (5) Dep. var.: lnfa

Var.

Coef.

Z

P N |z|

Var.

Coef.

lnGDP lnt lnpop_dens

0.844 0.859 −0.446

7.94 2.18 −3.56

0.000 0.030 0.000

lnGDP lnt lnpop_dens

0.846 0.821 −0.524

z 7.93 2.08 −4.14

P N |z|

Var.

Coef.

z

P N |z|

0.000 0.038 0.000

lnGDP lnt lnpop_dens

0.838 0.820 −0.532

7.89 2.08 −4.26

0.000 0.038 0.000

Eq. (6)

Eq. (6)

Eq. (6)

Dep. var.: CFBI

Dep. var.: CFBI

Dep. var.: CFBI

Var.

Coef.

Z

P N |z|

Var.

Coef.

ts tc tm t2 nts nbs nps nhs lngdp lnpop_dens

−0.536 −0.513 −0.575 0.061 0.017 −0.001 −0.000 0.007 0.035 −0.008

−4.68 −4.40 −4.73 4.27 5.11 −0.43 −0.42 1.73 1.56 −0.28

0.000 0.000 0.000 0.000 0.000 0.669 0.674 0.083 0.119 0.781

ts tc tm t2 nts nbs nps nhs lngdp lnpop_dens

−0.538 −0.514 −0.578 0.061 0.017 −0.001 −0.000 0.007 0.037 −0.018

z −4.70 −4.40 −4.76 4.29 5.11 −0.60 −0.38 1.72 1.64 −0.57

P N |z|

Var.

Coef.

z

P N |z|

0.000 0.000 0.000 0.000 0.000 0.550 0.702 0.086 0.102 0.566

ts tc tm t2 nts nbs nps nhs lngdp lnpop_dens

−0.503 −0.483 −0.553 0.057 0.018 −0.001 −0.001 0.009 0.038 −0.022

−4.46 −4.19 −4.61 4.07 5.44 −0.65 −0.57 2.11 1.69 −0.71

0.000 0.000 0.000 0.000 0.000 0.513 0.570 0.035 0.091 0.477

Nr. of observations: 68. Endogenous variables: lnEVi, lnfa, and CFBI. Exogenous variables: lnt, cfbi_ts, cfbi_tm, cfbi_tc, lnpop_dens, lnpd, ts, tc, tm, t2, nts, nbs, nps and nhs.

total forest area of the EU-17 countries. We can interpret this estimation result as signaling that higher population density in the European countries may accelerate the conversion of land uses from forested lands to agriculture and human settlements. Finally, estimation results of Eq. (3) show that an increase in the temperature has always a negative impact on the CFBI, independent from the type of forest ecosystem services under consideration, On average, a 1 °C increase in the temperatures can lead to about 0.5 decrease in the CFBI value, and at an accelerating rate, which is explained by the positive and statistically significant t2 parameter estimate. And again, this parameter estimate is found consistent across different ecosystem services, suggesting the robustness of our results. Next, the number of different trees species (nts) is revealed to be the most significant constituent of the CFBI. In particular, changes in one-unit of the nts is associated with an increase of 0.016 in the CFBI value. In contrast, changes in the richness of other species are not statistically significant in explaining CFBI, so as the two socio-economic variables, including GDP and population density. All in all, the proposed model specification was successful in estimating empirically the relationship between climate change, biodiversity, and the value of forest ecosystem services, with a specific emphasis on European forest ecosystems. Note that the investigation is conducted in a European-aggregated model, where a global effect of climate change was considered. Now, we will move on to disentangling the climate change effects across three European regions, including

Mediterranean Europe, Central Europe, and Scandinavian Europe, so as to capture the potential geo-climatic distributional impacts. This model is presented and discussed in the next section. 5.2. European-regional Model Estimates The 3SLS regression results for the European-regional model, as described by Eqs. (4)–(6), are presented in Table 4. As we can see, the introduction of the region specific climate change impacts improves significantly the overall goodness of fit for the proposed 3SLS routine. In fact, both Eqs. (4) and (6), which now have also explanatory variables that are region specific, present higher R2 statistics when compared with Eqs. (1) and (3) respectively. These results signal the success of the regional model, when compared to the European-aggregated model, in explaining the variation of the data. In particular, Eq. (4) shows that the elasticity estimates for the forest area are positive, same as in the European-aggregated model specification. Furthermore, the respective empirical magnitudes are also similar, ranging from 0.769 to 0.863, respectively for the economic value of the forest regulating and provisioning services. Again, forest area elasticity is estimated higher than 1 for forest cultural services, reiterating the economic significance of this services provided by the European forests. With respect to the climate change impacts on the economic value of forest EGS, we have now two types of impact (or value transmission mechanisms). First, we have the direct climate change impact, which

H. Ding, P.A.L.D. Nunes / Ecological Economics 97 (2014) 60–73

69

Table 5 Climate change marginal impacts on the economic value of European forest's ecosystem services. Provisioning

a

Europe Scandinavianb Centralb Mediterraneanb a b

Cultural

Regulating

Aggregated model

Regional model

Aggregated model

Regional model

Aggregated model

Regional model

1.032 – – –

0.193 n.s.s. n.s.s. –0.493

−0.664 – – –

−0.290 –0.059 n.s.s. 0.279

0.670 – – –

n.s.s. n.s.s. 0.251 0.259

Expressed in terms of elasticity. Expressed in terms of perceptual change in the economic value of the forest ecosystem services under consideration.

is captured by the beta coefficient estimated for the temperature variable. Second, we have the indirect regional climate-change-inducedbiodiversity effect (CCIBE), which is modeled as a cross-effect between the regional temperature and the biodiversity indicator, CFBI. All in all, estimation results show that the estimated directions of direct climate change impacts on the economic value of provisioning and cultural services remain the same as in the European-aggregated model, whereas the regional CCIBE show a great heterogeneity across geo-climatic regions and forest EGS. First, for provisioning services, the model estimates show that the direct climate change impact is now 0.193, still positive but smaller than in the European-aggregated model, ceteris paribus. Furthermore, the CCIBE is found statistically significant only in the Mediterranean Europe, with an estimated marginal impact of − 0.493. This means that an increment of 1 °C in the Mediterranean temperature can cause at the margin, a reduction of 57% in the CFBI score (estimated in Eq. (6)), which together with the direct impact of rising temperature (estimated in Eq. (1)), will be responsible for at least 49% loss in every one-dollar value derived from the provisioning services. In other words, the cross-effect of rising temperature and biodiversity loss in the Mediterranean forests goes hand-in-hand, causing a net negative impact on the total economic value generated from this EGS. Second, for the cultural services, the model estimates show that the direct climate change impact is −0.290, which is still negative, but at a smaller magnitude, when compared to the European-aggregated model specification, ceteris paribus. Furthermore, the CCIBE is estimated to be − 0.059, negative and statistically significant for the Scandinavian Europe. This means that CCIBE will accelerate the loss of economic value of cultural services provided by the Scandinavian forests, and this effect is estimated to be nearly 6% of every additional one-dollar reduction of the value of cultural services provided. On the contrary, CCIBE will increase, at the margin, the economic value of cultural services provided by the Mediterranean forests, whose effect amounts to nearly 28% of every additional one-dollar increase of the value of cultural services provided. Finally, for the regulating services, the model estimates show that the direct climate change impact is not statistically significant. However, the CCIBE is revealed to be positive and statistically significant for both forests located in Mediterranean and Central Europe. According to estimation results, these impacts are estimated to be 0.251 and 0.259, respectively. This means that for the forest in the Mediterranean and Central Europe, very one-unit biodiversity change caused by the increment of the regional temperature will correspond to, at the margin, 25% of every additional one-dollar increase of the value of these forest regulating services. 5.3. Comparative Analysis By comparing the results obtained from European-regional specific climate model with the European-aggregated model, we are able to highlight the following two key findings: (1) Climate-change-induced biodiversity effect, which is captured by the cross-effect between regional temperature and

the composite biodiversity indicator, has clear heterogenic characteristics. In particular, it is characterized by the reduction of the estimated magnitude of average effect for the EU-17, and thus can hamper the direct temperature effect at regional level. As we can see from Table 5, the average direct temperature effect is reduced from 1.032 (aggregated model) to 0.193 (regional model) and from − 0.664 (aggregated model) to − 0.290 (regional model) respectively for the economic value of the European forest's provisioning and cultural services. In addition, as far as the regulating service is concerned, on average, the estimated direct temperature impact on the economic value of these forest ecosystem services is reduced from 0.670 (aggregated model) to an impact that is not statistically different from zero (regional model). (2) The impacts of CCIBE on the economic value of EGS are mixed. In some cases, this impact amplifies the (reduction of the positive) direct temperature impact. This is the case for the cultural services provided by the Scandinavian forests, where the CCIBE imposes a negative impact on the economic value of the cultural services and thus play as an additional Keynesian multiplicative effect of the aggregated climate change impact on the EGS in this geo-climatic region of Europe (adding up to the also negative direct impact of temperature). A similar situation is observed for the provisioning services provided by the Mediterranean forest, where the CCIBE also negatively affect the economic value of these services. Furthermore, the estimated magnitude of this regional marginal effect suggests that the aggregated climate change impact for this EGS in this geo-climatic region of Europe is most likely to be negative (reversing the marginal and positive direct impact of temperature). However, in other cases, we observe the opposite cross-effects of biodiversity and temperature change on the value of ecosystem services. This is the case for the cultural services provided by the Mediterranean forests. Here, the CCIBE can lead to a marginal positive impact on the economic value of the services and therefore will play as buffering impact to the negative direct impact of the temperature. Furthermore, the estimated magnitude of this regional marginal impact suggests an offset effect of the marginal negative direct impact of temperature. In other words, the aggregated climate change impact for this EGS in this geo-climatic region of Europe is most likely to be null. A similar situation is observed for the regulating services provided by the Central and Mediterranean forests, where the CCIBE is responsible for a positive aggregate impact on the economic value of this service. 6. Policy Recommendations: the Role of Biodiversity as a Nature-based Policy Option for Climate Change Mitigation These estimation results confirm the role of biodiversity as a naturebased policy option for mitigating climate change impacts. Policies as such can generate co-benefits by enhancing ecosystems' capacity to mitigate climate change impacts, while conserving biodiversity and sustaining the flows of EGS for human livelihoods. In particular, well-

70

H. Ding, P.A.L.D. Nunes / Ecological Economics 97 (2014) 60–73

defined natural resource management and biodiversity conservation policies may provide sustained flows of EGS for local livelihoods and cost-effective options for natural adaptation and mitigation of climate change impacts. The policy implication of our model results is that, when the marginal benefits derived from biodiversity conservation efforts are large enough to compensate the marginal loss of ecosystem productivity and values as a result of direct negative impacts of global warming (i.e. the CCIBE), policy-making should be directed to favor the actions/measures that improve the overall biodiversity conditions, for example by either increasing the species richness or enlarging the total area of natural habitats. Thus the ecosystem functioning can be strengthened and serve as a natural mitigation means to sequester CO2 emissions, while providing shelter and other ecosystem services to safeguard human livelihoods. Estimation results show, however, that the strength of biodiversity as a nature-based policy solution to climate change mitigation will be depending on both the nature of the EGS as well as the geo-climatic region under consideration. Moreover, estimation results inform us, that the potential of the biodiversity as a nature-based policy option for climate change mitigation will be of particular interest to the management of Mediterranean forests. It is important to highlight the impacts of data limitation on the model results. As mentioned earlier, the use of time-span of 50 years between 2000 and 2050 to describe the evolution of number of species under climate change scenarios has resulted in an observed increase of species richness as well as forest areas in many EU-17 countries. This gives rise to a very large range of CFBI scores (between 0% and 200%), which increases the difficulty to perceive the negative impacts of global warming on biodiversity. However, the robustness of the econometric model results can be improved by selecting the “pre-industrial baseline” as the reference year may allow us to observe more sensitive reaction of biodiversity (in terms of estimated CFBI scores) to the changes of temperature over time. In particular, this effect may be more significant when conducting sensitivity analysis of the heterogeneity of regional climate change impacts and help us to better understand the cross-effects between biodiversity and temperature as well as the pattern in which they affect the value of ecosystem services. 7. Conclusions and Further Research This paper attempted to model the relationships between climate change, biodiversity and the value of ecosystem services with a specific emphasis on the climate change included biodiversity effects in European forests. To our knowledge, this represented one of the first attempts in the literature to formally model and empirically test the strength of biodiversity as a nature-based policy option for climate change mitigation. The research begun with the construction of a Composite Forest Biodiversity Indicator (CFBI) that integrated quantitative and qualitative changes of biodiversity projected under different future climate scenarios. The model results suggested that CFBI could serve its multiple design purposes better than other individual biodiversity indicators for two reasons. First, it was capable of aggregating the core biological information to measure and predict the trends of biodiversity changes in response to both global warming and socio-economic changes over a period of time under different future climate change scenarios and then link the biodiversity state to its capacity of providing EGS. Second, the computation of the CFBI was standard and simple, which led to a simple format of biodiversity indicator with the scales of measurement that could be both easily understood by, and effectively communicated with a variety of stakeholders, including nature scientists, social scientists, economists, policymakers and the broader audience. We also made the best use of existing data released by a large number of IPCC data distribution centers, regarding the projected trends of

population and economic growth, future species richness and increase in local temperature under different future climate scenarios. Economic values of ecosystem services were derived from a most recent assessment study on the climate change impacts on forest ecosystems in Europe (Ding et al., 2010). In this data setting, we used 3SLS regression to simultaneously estimate (1) the determinants of economic value of ecosystem services; (2) the determinants of land-use changes (i.e. the changes of forest land cover); and (3) the determinants of changes in biodiversity. The investigation was conducted first in a baseline model at an aggregated European level, where a global effect of climate change was assessed, followed by a regional model, where the marginal magnitudes of CCIBEs were assessed at three specific geo-climatic regions. Despite the data limitation, our preliminary results from a 3SLS regression were promising. First, European-aggregated model specification results confirmed that rising temperature negatively affected biodiversity conditions at an accelerating rate across geo-climatic regions in Europe by 2050. Second, we also found a strong relationship between temperature and the value of EGS, but the direction of this relationship depended on the type of EGS under consideration. For example, this relationship was estimated to be positive for provisioning and regulating services, but negatively related to cultural services. Third, the regional model specification results suggested that the negative impacts of climate change on biodiversity (i.e. CCIBE) could go against the positive direct climate change impact on forest growth and generate a net negative impact on total value of EGS, such as for the provisioning services in the Mediterranean Europe. However, the model estimates also showed that these impacts were region specific and shall be assessed accordingly. In some cases, this marginal impact could amplify the (reduction of the positive) direct temperature impact, e.g. for the cultural services provided by the Scandinavian forests; whereas in some other cases, we could observe an opposite cross-effect of biodiversity and temperature change on the value of ecosystem services, such as for the cultural services provided by the Mediterranean forests. In the latter case, the climate-change-induced biodiversity effect was responsible for a marginal positive impact on the economic value of the cultural services, through its buffering impact on the negative direct impact of the temperature. To conclude, independent from the sign and magnitude of the effects, these estimation results confirmed the role of biodiversity as a nature-based policy solution for climate change, shedding light on the policy actions that generate co-benefits by enhancing ecosystems' capacity to mitigate climate change impacts, while conserving biodiversity and sustaining the flows of EGS for human livelihoods. Especially, nature-based mitigation policies could be more cost-effective and better at coping with the ethic and inequality issues associated with distributional impacts of the policy actions, compared to the pure technical solutions to improving energy efficiency and reducing emissions. Finally, notwithstanding the success in modeling the relationship between biodiversity, ecosystem services and human wellbeing in the context of climate change, we acknowledge that several aspects deserve future investigation for the future development of the model. For instance, the degree of heterogeneity within the three geo-climatic zones is not tackled in the present model. A further extension of this model may focus on investigating the marginal difference of climate change impacts on forest EGS across countries within the same geoclimatic zone, and the results may be useful for guiding the design of specific national forest management strategies. In addition, a sensitivity analysis of the model's outputs to the changes of the policy-mix underpinning the scenarios shall be conducted, so as to allow policymakers to compare the marginal impacts of environmental policies implemented in different socio-economic contexts. Finally, future research may also be extended to other countries outside Europe, preferably to areas in the globe, where present a wider biological diversity with respect to their geo-climatic conditions as well as the economic significance of forest EGS to their economies. We are convinced that these

H. Ding, P.A.L.D. Nunes / Ecological Economics 97 (2014) 60–73

roadmaps for potential future research will present promising, follow-up economic analysis, with which we can benchmark the results presented here. Acknowledgment This research received financial support from the European Investment Bank University Research Sponsorship Programme (EIBURS). Particular thanks go to Rafat Alam, Jonah Busch and Peter Carter and participants of the 12th annual BIOECON conference held in Venice, Italy, for their comments on a previous version of the paper. The authors are also grateful to Laura Onofri for the discussions on the econometric work. Annex

Table A1 Trends of European forest area projected under IPCC scenarios (estimates in 1000 ha). Country

2005a

2050 A1FIb

2050 A2b,c

2050 B1b

2050 B2b

Greece Italy Portugal Spain Austria Belgium France Germany Ireland Luxembourg Netherlands Switzerland UK Denmark Finland Iceland Norway Sweden

3752 9979 3783 17,915 3862 667 15,554 11,076 669 87 365 1221 2845 500 22,500 46 9387 27,528

2292 8346 2170 12,052 5298 526 15,094 10,049 442 80 151 1985 1986 414 18,224 30 6478 22,704

2360 8253 2174 11,969 5177 545 16,056 10,075 379 78 421 1913 2145 677 17,999 29 6277 22,198

3762 11,677 3254 17,389 5199 698 20,080 12,696 638 103 333 2113 2780 434 16,517 28 5141 25,884

3598 11,893 3283 17,633 5471 842 21,926 14,033 656 94 413 2121 3476 839 17,079 28 5761 22,704

Source: a data from FAO; b projections by ATEAM and CLIBIO on the basis of the Integrated Model to Assess the Global Environment (IMAGE), developed by Netherlands Environmental Assessment Agency; c interpreted by the European Commission as the baseline scenario, i.e. the scenario characterized by policy inaction. NB: For the detailed data explanation, please refer to Ding et al. (2010).

Greece Italy Portugal Spain Austria Belgium France Germany Ireland Luxembourg Netherlands Switzerland United Kingdom Denmark Finland Norway Sweden

Table A3 Projected total value of forest carbon stocks in EU-17 by 2050. Country

Greece Italy Portugal Spain Austria Belgium France Germany Ireland Luxembourg Netherlands Switzerland United Kingdom Denmark Finland Norway Sweden

2005

2050 A1FI

2050 A2

2050 B1

2050 B2

9052 4768 614 2911 3372 364 7020 6703 198 197 184 1035 1232 186 5487 1740 7816

2695 2617 273 1796 3690 185 6408 3972 140 89 71 1349 668 111 2459 693 3879

2775 2628 264 1784 3748 203 6750 4144 136 87 166 1357 796 208 2429 670 4043

4424 3236 364 2269 3985 222 7466 4969 169 115 114 1502 913 135 2831 731 5746

4230 3075 337 2218 3900 212 7097 4752 174 104 124 1428 924 160 2539 724 4370

Note: value estimates are derived from Ding et al. (2010).

Table A4 Projected total cultural value of forest in EU-17 by 2050. Country

(Million US$ 2005)

Greece Italy Portugal Spain Austria Belgium France Germany Ireland Luxembourg Netherlands Switzerland United Kingdom Denmark Finland Norway Sweden

2005

2050 A1FI

2050 A2

2050 B1

2050 B2

390 1039 394 1864 402 69 1619 1153 70 9 38 127 296 52 2342 977 2865

239 869 226 1254 222 22 632 421 19 3 6 83 84 17 462 164 576

247 863 227 1251 206 22 640 402 15 3 17 76 86 27 459 160 566

566 1756 489 2615 308 41 1191 753 38 6 20 125 194 30 1039 323 1629

490 1619 447 2401 218 34 872 558 26 4 16 84 156 38 833 281 1107

Source: value estimates are derived from Ding et al. (2010).

Country

(Million US$ 2005) 2005

2050 A1FI

2050 A2

2050 B1

2050 B2

141 3225 1859 3337 5990 4807 7204 16,636 506 216 3693 2003 2665 465 12,067 1863 13,200

101 1465 1760 2212 7510 4832 4909 12,741 299 107 2568 2120 2997 439 15,913 2021 17,606

104 1447 1844 2197 7236 3343 5281 12,712 250 104 9289 2039 2925 1067 15,333 1625 16,984

166 1884 2279 2870 5186 3513 5684 12,620 304 137 5134 2095 2543 410 12,985 1476 17,310

158 2082 2301 3233 6897 4306 6211 14,906 384 125 6375 1847 3361 714 14,183 1708 16,052

Note: value estimates are derived from Ding et al. (2010).

(Million US$ 2005)

Table A5 Trends of GDP and population in IPCC scenarios (2050).

Table A2 Projected total value of wood forest products in EU-17 by 2050. Country

71

Greece Italy Portugal Spain Austria Belgium France Germany Ireland Luxembourg Netherlands Switzerland United Kingdom Denmark Finland Norway Sweden

Population density (head/ha)a

GDP per capita (000'US$)

A1FI

A2

B1

B2

A1FI

A2

B1

B2

0.41 1.06 0.63 0.57 0.87 3.33 1.10 2.12 0.53 2.99 3.169 0.93 1.66 0.14 0.12 0.10 2.11

0.41 1.06 0.63 0.57 0.87 3.33 1.10 2.12 0.53 2.99 3.19 0.93 1.66 0.14 0.12 0.10 2.11

0.41 1.06 0.63 0.57 0.87 3.33 1.10 2.12 0.53 2.99 3.19 0.93 1.66 0.14 0.12 0.10 2.11

0.37 0.92 0.56 0.51 0.89 3.05 0.93 1.85 0.35 1.78 2.73 1.02 1.49 0.13 0.12 0.09 2.32

27.38 73.52 23.17 47.14 72.06 59.96 57.47 70.11 26.44 45.75 54.83 117.04 49.00 77.26 85.15 69.32 88.51

21.36 57.36 18.08 36.78 56.21 46.78 44.83 54.70 20.63 35.69 42.77 91.30 38.23 60.27 66.43 54.08 69.05

21.87 58.73 18.51 37.66 57.56 47.90 45.91 56.01 21.12 36.55 43.80 93.49 39.14 61.71 68.02 55.37 70.70

19.30 53.65 16.70 33.13 44.74 41.50 42.67 50.67 25.44 48.56 40.53 67.46 34.44 52.19 54.17 50.38 50.90

Source: CIESIN (2002).

72

H. Ding, P.A.L.D. Nunes / Ecological Economics 97 (2014) 60–73

Table A6 Trends of tree, bird, plant, and herptile for the EU17 projected between 2000 and 2050. Country Greece

Italy

Portugal

Spain

Austria

Belgium

France

Germany

Ireland

Luxembourg

Netherlands

Switzerland

Denmark

Biodiversity indicators

Baseline 2000

A1FI 2050

A2 2050

No. of tree species No. of bird species No. of plant species No. of herptile species No. of tree species No. of bird species No. of plant species No. of herptile species No. of tree species No. of bird species No. of plant species No. of herptile species No. of tree species No. of bird species No. of plant species No. of herptile species No. of tree species No. of bird species No. of plant species No. of herptile species No. of tree species No. of bird species No. of plant species No. of herptile species No. of tree species No. of bird species No. of plant species No. of herptile species No. of tree species No. of bird species No. of plant species No. of herptile species No. of tree species No. of bird species No. of plant species No. of herptile species No. of tree species No. of bird species No. of plant species No. of herptile species No. of tree species No. of bird species No. of plant species No. of herptile species No. of tree species No. of bird species No. of plant species No. of herptile species No. of tree species No. of bird species

24 98.28 262.40

22 108.71 236.66

30 26 31 111.09 110.62 110.44 247.95 214.89 228.45

38.85

38.24

33 104.30 330.55

34 116.41 273.96

30.80

29.51

18 89.10 237.24

11 106.47 242.43

38.76

37.84

23 103.56 284.12

16 121.16 243.05

35.00

34.81

46 141.60 257.74

52 146.29 305.52

20.32

19.71

30 125.77 281.29

31 129.21 230.98

22.04

23.08

32 126.36 307.62

20 130.55 227.17

24.28

24.33

32 157.08 309.47

36 154.31 263.87

22.86

22.86

21 113.97 236.79

28 119.16 227.62

6.65

7.28

29 114.92 262.75

21 122.17 203.00

20.83

21.25

28 145.43 281.45

44 145.86 261.31

19.39

20.70

47 130.76 344.60

41 135.23 225.68

19.01

18.30

31 137.97

41 145.13

38.52

B1 2050

39.39

31.01

39.22

37.77

35.70

Finland

Norway

37.78

19 20 21 123.11 124.07 122.28 246.59 231.15 230.89 34.18

United Kingdom

30.68

14 13 13 110.33 111.13 110.11 231.88 215.35 204.92 37.51

Country

B2 2050

43 41 45 117.08 117.85 116.51 303.47 269.50 287.84 30.31

Table A6 (continued)

Sweden

34.95

69 70 71 145.87 146.32 145.19 361.78 338.09 348.90 19.68

19.86

Biodiversity indicators

Baseline 2000

A1FI 2050

A2 2050

No. of plant species No. of herptile species No. of tree species No. of bird species No. of plant species No. of herptile species No. of tree species No. of bird species No. of plant species No. of herptile species No. of tree species No. of bird species No. of plant species No. of herptile species No. of tree species No. of bird species No. of plant species No. of herptile species

308.02

298.86

318.13 359.58 337.47

17.35

16.85

22 127.05 231.94

30 126.74 245.09

9.69

9.64

21 141.18 129.59

31 143.73 240.21

5.58

5.42

25 113.91 171.62

36 116.51 291.60

2.00

1.94

24 140.70 165.85

33 142.26 250.59

6.14

5.34

16.77

B1 2050

16.82

B2 2050

16.63

34 37 37 127.79 128.39 126.33 260.11 254.19 251.73 10.03

10.44

9.95

33 36 36 143.94 144.04 144.70 211.84 204.93 199.61 5.41

5.71

5.43

40 41 42 116.46 117.01 116.82 256.03 243.24 249.37 1.82

1.85

1.72

36 39 38 142.82 143.52 143.37 235.68 226.14 230.55 5.20

5.66

5.37

19.62

46 52 52 128.42 129.12 128.55 256.88 261.99 252.35

Source: Schröter et al. (2004).

References 23.62

23.81

23.37

31 34 38 129.93 130.26 129.97 265.41 268.48 262.96 25.02

25.11

25.03

45 50 49 153.61 153.98 152.97 281.88 297.25 279.91 22.50

22.57

22.64

33 37 37 121.29 120.55 118.62 243.49 253.40 252.58 7.63

7.70

7.53

37 45 46 119.25 120.67 117.92 230.00 255.00 243.00 21.42

22.67

21.83

48 51 51 145.51 146.65 145.04 283.51 279.56 274.16 20.85

20.17

20.15

60 63 63 134.40 135.08 133.73 281.27 282.64 273.33 18.22

18.24

18.04

45 49 48 146.93 148.69 146.66

Balmford, A., Bennun, L., ten Brink, B., Cooper, D., Côté, I.M., Crane, P., Dobson, A., Dudley, N., Dutton, I., Green, R.E., Gregory, R.D., Harrison, J., Kennedy, E.T., Kremen, C., LeaderWilliams, N., Lovejoy, T.E., Mace, G., May, R., Mayaux, P., Morling, P., Phillips, J., Redford, K., Ricketts, T.H., Rodriguez, J.P., Sanjayan, M., Schei, P.J., van Jaarsveld, A.S., Walther, B.A., 2005. The convention on biological diversity's 2010 target. Science 307, 212–213. Bloger, T., 2001. The functional value of species biodiversity: a review. Biol. Environ. Proc. R. Ir. Acad. 101 B (3), 199–224. Cameron, T., 2002. 2002: the year of the ‘diversity-ecosystem function’ debate. Trends Ecol. Evol. 17 (11), 495–496. Chiabai, A., Travisi, C., Markandya, A., Ding, H., Nunes, P.A.L.D., 2011. Economic assessment of forest ecosystem services losses: cost of policy inaction. J. Environ. Resour. Econ. 50 (3), 405–445. CIESIN, 2002. Environmental sustainability index. The Center for International Earth Science Information Network. 296 (February). Costanza, R., Fisher, B., Mulder, K., Liu, S., Christopher, T., 2007. Biodiversity and ecosystem services: a multi-scale empirical study of the relationship between species richness and net primary production. Ecol. Econ. 61, 478–491. De Bruyn, S.M., 2000. Economic growth and the environment: an empirical analysis. Economy & Environment, 18. Kluwer Academic Publishers, Dordrecht, Boston and London. Ding, H., Silvestri, S., Chiabai, A., Nunes, P.A.L.D., 2010. A hybrid approach to the valuation of climate change effects on ecosystem services: evidence from the European forests. FEEM Working Paper. No. 2010.050. Faith, D.P., 2005. Global biodiversity assessment: integrating global and local values and human dimensions. Glob. Environ. Chang. 15, 5–8. Garcia-Gonzalo, J., Peltola, H., Briceño-Elizondo, E., Kellomäki, S., 2007. Changed thinning regimes may increase carbon stock under climate change: a case study from a Finnish boreal forest. Clim. Change 81, 431–454. Giller, P.S., O'Donovan, G., 2002. Biodiversity and ecosystem function: do species matter? Biol. Environ. Proc. R. Ir. Acad. 102B (3), 129–139. Gordon, C., Cooper, C., Senior, C.A., Banks, H., Gregory, J.M., Johns, T.C., Mitchell, J.F.B., Wood, R.A., 2000. The simulation of SST, sea ice extents and ocean heat transports in a version of the Hadley Centre coupled model without flux adjustments. Climate Dynam. 16, 147–168. Halpern, B.S., Longo, C., Hardy, D., McLeod, K.L., Samhouri, J.F., Katona, S.K., Kleisner, K., Lester, S.E., O'Leary, J., Ranelletti, M., Rosenberg, A.A., Scarborough, C., Selig, E.R., Best, B.D., Brumbaugh, D.R., Chapin, F.S., Crowder, L.B., Daly, K.L., Doney, S.C., Elfes, C., Fogarty, M.J., Gaines, S.D., Jacobsen, K.I., Karrer, L.B., Leslie, H.M., Neeley, E., Pauly, D., Polasky, S., Ris, B., Martin, K.S., Stone, G.S., Sumaila, U.R., Zeller, D., 2012. An index to assess the health and benefits of the global ocean. Nature 488, 615–620. IPCC, 2000. Special Report on Emissions Scenarios.Cambridge University Press, Cambridge.

H. Ding, P.A.L.D. Nunes / Ecological Economics 97 (2014) 60–73 Kinzig, A., Pacala, S.P., Tilman, D. (Eds.), 2001. The Functional Consequences of Biodiversity. Princeton University Press, Princeton, New Jersey, USA. Kontoleon, A., Pascual, U., Swanson, T., 2007. Introduction. In: Kontoleon, A., Pascual, U., Swanson, T. (Eds.), Biodiversity Economics: Principles, Methods and Applications. Cambridge University Press, Cambridge, p. 692. Leadley, P., Pereira, H.M., Alkemade, R., Fernandez-Manjarrés, J.F., Proença, V., Scharlemann, J.P.W., Walpole, M.J., 2010. Biodiversity scenarios: projections of 21st century change in biodiversity and associated ecosystem services. Secretariat of the Convention on Biological Diversity, Montreal. . Technical Series No. 50. Lindner, M., Maroschek, M., Netherer, S., Kremer, A., Barbati, A., Garcia-Gonzalo, J., Seidl, R., Delzon, S., Corona, P., Kolstroma, M., Lexer, M.J., Marchetti, M., 2010. Climate change impacts, adaptive capacity, and vulnerability of European forest ecosystems. For. Ecol. Manage. 259, 698–709. Loreau, M., Naeem, S., Inchausti, P., Bengtsson, J., Grime, J.P., Hector, A., Hooper, D.U., Huston, M.A., Raffaelli, D., Schmid, B., Tilman, D., Wardle, D.A., 2001. Biodiversity and ecosystem functioning: current knowledge and future challenges. Science 294, 804–808. Loreau, M., Naeem, S., Inchausti, P. (Eds.), 2002. Biodiversity and Ecosystem Functioning: Synthesis and Perspectives. Oxford University Press, Oxford, UK. Mace, G.M., Baillie, J.E.M., 2007. The 2010 biodiversity indicator: challenges for science and policy. Conserv. Biol. 21 (6), 1406–1413. Martens, P., Rotmans, J., de Groot, D., 2003. Biodiversity: luxury or necessity? Glob. Environ. Chang. 13, 75–81. Millennium Ecosystem Assessment (MEA), 2005. Ecosystems and Human Well-being: Synthesis.Island Press, Washington, DC. Nunes, P.A.L.D., van den Bergh, J.C.J.M., 2001. Economic valuation of biodiversity: sense or nonsense. Ecol. Econ. 39 (2), 203–222. Ojea, E., Loureiro, M., Nunes, P.A.L.D., 2010. Mapping of forest biodiversity values: a plural perspective. Environ. Resour. Econ. 47, 329–347. Paine, R.T., 2002. Tropic control of production in a rocky intertidal community. Science 296, 736–739. Pereira, H.M., Leadley, P.W., Proença, V., Alkemade, R., Scharlemann, J.P.W., FernandezManjarrés, J.F., Araújo, Miguel B., Balvanera, P., Biggs, R., Cheung, W.W.L., Chini, L.,

73

Cooper, H.D., Gilman, E.L., Guénette, S., Hurtt, G.C., Huntington, H.P., Mace, G.M., Oberdorff, T., Revenga, C., Rodrigues, P., Scholes, R.J., Sumaila, U.R., Walpole, M., 2010. Scenarios for global biodiversity in the 21st century. Science 330, 1496–1501. Sala, O.E., Chapin III, F.S., Armesto, J.J., Berlow, E., Bloomfield, J., Dirzo, R., HuberSanwald, E., Huenneke, L.F., Jackson, R.B., Kinzig, A., Leemans, R., Lodge, D.M., Mooney, H.A., Oesterheld, M., LeRoy, Poff N., Sykes, M.T., Walker, B.H., Walker, M., Wall, D.H., 2000. Global biodiversity scenarios for the year 2100. Science 287, 1770–1777. Schmid, B., Joshi, J., Schläpfer, F., 2000. Empirical evidence for biodiversity-functioning relationships. In: Kinzig, A.P., Tilman, D., Pacala, S.W. (Eds.), Functional Consequences of Biodiversity: Experimental Progress and Theoretical Extensions, Monographs. Popul. Biol., 33. Princeton University Press, Princeton, NJ, pp. 120–150. Scholes, R.J., Biggs, R., 2005. A biodiversity intactness index. Nature 434 (7029), 45–49. Schröter, D., Acosta-Michlik, L., Arnell, A.W., Araújo, M.B., Badeck, F., Bakker, M., Bondeau, A., Bugmann, H., Carter, T., de la Vega-Leinert, A.C., Erhard, M., Espiñeira, G.Z., Ewert, F., Fritsch, U., Friedlingstein, P., Glendining, M., Gracia, C.A., Hickler, T., House, J., Hulme, M., Kankaanpää, S., Klein, R.J.T., Krukenberg, B., Lavorel, S., Leemans, R., Lindner, M., Liski, J., Metzger, M.J., Meyer, J., Mitchell, T., Mohren, F., Morales, P., Moreno, J.M., Reginster, I., Reidsma, P., Rounsevell, M., Pla, E., Pluimers, J., Prentice, I.C., Pussinen, A., Sánchez, A., Sabaté, S., Sitch, S., Smith, B., Smith, J., Smith, P., Sykes, M.T., Thonicke, K., Thuiller, W., Tuck, G., van der Werf, G., Vayreda, J., Wattenbach, M., Wilson, D.W., Woodward, F.I., Zaehle, S., Zierl, B., Zudin, S., Cramer, W., 2004. ATEAM (Advanced Terrestrial Ecosystem Analyses and Modelling) Final Report. Potsdam Institute for Climate Impact Research. Ten Brink, B., 2000. Biodiversity indicators for the OECD environmental outlook and strategy. GLOBIO Report Series No. 25. (RIVM). Verbeek, M., 2000. A Guide to Modern Econometrics.John Wiley & Sons, Ltd., England. Walther, G., Post, E., Convey, P., Menzel, A., Parmesan, C., Beebee, T.J.C., Fromentin, J., Hoegh-Guldbergy, O., Bairlein, F., 2002. Ecological responses to recent climate change. Nature 416, 389–395. Willims, W.D., Dormaar, J.F., Adams, B.W., Douwes, H.E., 2002. Response of the mixed prairie to protection from grazing. J. Range Manage. 55, 210–216.