Scale mismatches and their ecological and economic effects on landscapes: A spatially explicit model

Scale mismatches and their ecological and economic effects on landscapes: A spatially explicit model

ARTICLE IN PRESS Global Environmental Change 18 (2008) 768–775 Contents lists available at ScienceDirect Global Environmental Change journal homepag...

663KB Sizes 0 Downloads 3 Views

ARTICLE IN PRESS Global Environmental Change 18 (2008) 768–775

Contents lists available at ScienceDirect

Global Environmental Change journal homepage: www.elsevier.com/locate/gloenvcha

Scale mismatches and their ecological and economic effects on landscapes: A spatially explicit model Akiko Satake a,, Thomas K. Rudel b,c, Ayumi Onuma d a

Creative Research Initiative Sousei, Hokkaido University, Sapporo 001-0021, Japan Department of Human Ecology, Rutgers University, New Brunswick, NJ 08901, USA Department of Sociology, Rutgers University, New Brunswick, NJ 08901, USA d Faculty of Economics, Keio University, Tokyo 108-8345, Japan b c

a r t i c l e in f o

a b s t r a c t

Article history: Received 23 December 2007 Received in revised form 19 July 2008 Accepted 21 July 2008

Mismatches between the spatial scales of human decision-making and natural processes contribute to environmental problems such as global warming and biodiversity losses. People damage the environment through local activities like clearing land or burning fossil fuels, but the damages only become manifest at larger regional or global scales where no one pays for them. Payments for ecological services like carbon sequestration can correct for these damages caused by scale mismatches. This paper presents a spatially explicit land-use model to investigate the consequences of scale mismatches for pollination and carbon storage services and examine the effect of payment for only carbon storage services. The model integrates processes in multiple spatial scales ranging from the parcel level used by landowners’ decision about deforestation, to the larger scale used by animals to pollinate plants, and finally to the global scale where carbon storage services are supplied. We show that payment for carbon storage services can become an effective mechanism to protect forests at the same time that it creates inequities among landowners in income level. These findings suggest that market-based approaches that focus on conservation of a single ecosystem service may reproduce unequal power relations among landowners. & 2008 Elsevier Ltd. All rights reserved.

Keywords: Land-use change Deforestation Markov chain model Equity Payment of environmental services Decision making

1. Introduction Human activity is a major driving force in global environmental change and has caused diverse environmental problems including global warming, degradation of biogeochemical and hydrological dynamics, and losses of biodiversity (Dixon et al., 1994; Vitousek et al., 1997; Houghton et al., 2000), all of which have important consequences for human welfare (Millennium Ecosystem Assessment, 2005). ‘‘Scale mismatches’’ characterize all of these problems; in other words, individual humans do not experience the negative effects of their actions locally. Instead, these effects cumulate and receive expression at a regional or even global scale where no one pays for damages (Folke et al., 1998; Carpenter et al., 2006; Cumming et al., 2006). Scale mismatches occur when the scale of provision for ecosystem services, such as pollinations of crops, purification of water, and stabilization of climate, do not coincide with the scale of decision-making by agents who manage resources. For example,

 Corresponding author. Tel.: +8111706 9182/2261; fax: +8111706 9182.

E-mail address: [email protected] (A. Satake). 0959-3780/$ - see front matter & 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.gloenvcha.2008.07.007

the conversion of land from forest to agriculture provides benefits to local actors, but runoff from the new fields may contain fertilizers and pesticides that compromise the quality of water used by people downstream (WRI, 2000). In the case of greenhouse gas emissions, people emit locally, in their daily round of activities, but the consequences of their actions play themselves out on a global meteorological scale, negatively impacting on vulnerable populations through pathways that are difficult to trace. Land users typically receive no compensation for the ecosystem services they generate for others. As a result, they have little incentive to provide these services, which causes damage to others. The damages caused by scale mismatches can be avoided by payment for environmental services (PES) from those who use these services (or others acting on the behalf of the users of these services) to those whose lands provide these services (Engel et al., 2008). Several programs, the PAS (pagos por servicios ambientales) program in Costa Rica (Sa´nchez-Azofeifa et al., 2007) and the ´tico Carbon project in Mexico (Corbera et al., 2007), Fondo Bioclima now make payments for environmental services provided by forests (Pagiola et al., 2002). In the Costa Rican program, the primary funding source was a 15% consumer tax on fossil fuels.

ARTICLE IN PRESS A. Satake et al. / Global Environmental Change 18 (2008) 768–775

2. The model and methods 2.1. Markovian transition between forest, agricultural, and abandoned land We briefly explain the Markovian model for land-use dynamics that was studied by Satake and Rudel (2007). We assume that a landscape is apportioned into discrete land parcels of equal size and arranged in a two-dimensional square lattice (Fig. 1a). The parcel i is managed by a landowner i, iA{1,y,N}, implying secured property rights for each landowner. Let Si(t) be the state variable at the land parcel i in year t. Si(t) is in a forested (F), agricultural

ΩC

...

...

ΩP F

F

A

E

A

F

F

A

A

F

F

A

F

F

F

A

E

A

F

F

A

A

E

E

F

...

ΩD ...

The recipients of payments were local landowners who managed either planted or natural forests that supplied services like carbon sequestration, hydrological filtering, scenic views, and biodiversity. Approximately 300,000 ha of primary, secondary, or planted forest received funding through the PAS program. The introduction of agri-environment schemes in EU countries is another example in which farmers are paid to modify their farming practices for biodiversity conservation (Kleijn and Sutherland, 2003). At the global level, the Clean Development Mechanism of the Kyoto Protocol provides payments through an international carbon market for environmental services (Fearnside, 1999; Santilli et al., 2005). These programs have focused on four types of forest ecosystem services: carbon sequestration, watershed protection, biodiversity conservation, and landscape aesthetics (LandellMills and Porrai, 2002). Other ecosystem services have been damaged by scale mismatches in which no one created systems for paying providers to preserve the services. For example, pollination services are considered as one of the most important forest ecosystem services (Millennium Ecosystem Assessment, 2005). Its economic value is recognized (Southwick and Southwick, 1992), but markets do not incorporate the value of this service into their evaluations of forests that provide essential habitat to pollinators. Thus, landowners continue to degrade forests in order to extract woods whose value is recognized by the markets. With the decline of the forests, pollinators decline in number and the risks of extinction rise for plant species that rely upon the pollinators (Aizen and Feinsinger, 1994; Groom, 2001). The net profit from forest conservation should therefore be influenced both by payments for a certain type of ecosystem services and value of other types of ecosystem services for which there are no systems for payment. Variations in spatial scales of provision among the different types of ecosystem services complicate the situation further because there will be mismatches in the scale of payments for one service and the scale of damages or benefits for another service. To appreciate the role of scale mismatches in environmental problems and the potential of PES remedies to these problems, this paper presents the model of scale mismatches for two forest ecosystem services, pollination services and carbon storage services, that are supplied at different spatial scales. The model aims to provide a synthetic understanding about the aggregated impacts of damages to pollination services and payments for carbon storage services on decision-making by landowners. It also predicts the consequences of these impacts for landowners’ well-being and equity in a society. The analysis extends a model for coupled social and ecological systems (Satake and Rudel, 2007) by incorporating scale mismatches in a spatially explicit way. Based on results of our analysis, we assess the distributive effects of payments for forest ecosystem services within societies.

769

Forest forest recovery

deforestation r

μ

Abandoned

Agriculture

abandonment η Fig. 1. (a) The spatial scales of provision of pollination services (OP: a dark gray region) and that of carbon storage services (OC: a white region) illustrated on a landscape. The parcel is either forested (F: black cells), agricultural (A: white cells), or abandoned (E: light gray cells). OD represents the spatial scale for landowners decisions to deforest. (b) Land use transitions on a single land parcel. r, Z, and m are the deforestation rate, abandonment rate, and forest recovery rate respectively.

(A), or abandoned (E) state: 8 > :E

if parcel i is forested; if parcel i is agricultural;

(1)

if parcel i is abandoned:

Each land parcel shows Markovian transitions between different states. For example, at the parcel i, the forested parcel (F) changes to the agricultural parcel (A) in year t with probability ri(t). We call ri(t) the deforestation rate per year, and ri(t) is an increasing function of the net gain of deforestation as explained later. Agricultural land changes to abandoned land with probability Z in a year. We call Z the ‘‘abandonment rate’’ per year. Abandonment of agricultural land may result in growth of secondary vegetation, and finally revert back to forested land with probability m in a year. We call m the ‘‘forest recovery rate’’ per year. The transition matrix for land use at the parcel i in year t is thus given as

0 B Pi ðtÞ ¼ B @

F

A

1  r i ðtÞ

r i ðtÞ

0

1Z

m

0

E 0

1

F

C

A.

Z C A

1m

E

(2)

ARTICLE IN PRESS 770

A. Satake et al. / Global Environmental Change 18 (2008) 768–775

More complicated Markovian models have been used to explain land-use dynamics in Brazilian Amazon (e.g. Fearnside, 1996). 2.2. Modelling scale mismatches: pollination services and carbon storage services Land conversion from forest to agriculture is made by a landowner’s decision. Landowners decide whether or not to deforest their land by comparing the expected utility of forest conservation with that of deforestation. Utility represents the magnitude of net benefit (i.e. benefit minus cost) from the consumption of goods. In this section, we define the actual utility given to each land-use state. We focus on two ecosystem services produced by forests: pollination service and carbon storage service. The utility of forested parcel is influenced by these two services. Let OP, OC, and OD be the scale of pollination services, the scale of carbon storage services, and the scale for landowners decisions to deforest, respectively. The scale of pollination services (OP) represents the spatial extent over which the utility of forested parcel is influenced by the level of pollination services. Similarly, the scale of carbon storage services (OC) represents the spatial extent over which the utility of forested parcel is influenced by the level of carbon storage services. (1) The scale of pollination service is larger than that of landowner deforestation decisions, but is smaller than that of carbon storage services (i.e. OD oOP oOC ): The pollination service is supplied at regional scale (OP ) larger than that of landowner deforestation decisions (OD ) because the loss of forest by a landowner has a visible effect on degradation of plant reproduction in neighboring forests managed by others through reduced pollinator diversity and abundance (Aizen and Feinsinger, 1994 ; Klein et al., 2007 ; see Fig. 1 a). Studies in agricultural landscapes show that increased distance from forest fragments can result in a decrease in both abundance and species richness of flower-visiting bees (Steffan-Dewenter and Tscharntke, 1999 ; Ricketts, 2004 ). The carbon storage service is supplied at global scale (OC ; see Fig. 1 a). Carbon storage by local forests benefits global human society by reducing atmospheric carbon dioxide and mitigating global warming (Fearnside, 2000 ). (2) No payment for pollination services: Although the pollination service is considered as one of the most important ecological services provided by pollinators in forests (Millennium Ecosystem Assessment, 2005), markets do not incorporate the value of this service into their evaluations of forest that provides essential habitat to pollinators. The utility of forested parcel caused by pollination service from the neighboring forests is therefore degraded by the destruction of forested land in the neighborhood. To be specific, let Pi(t) be the utility of forested parcel i in year t caused by consumption of nontimber forest products (e.g., firewood, medical products and wild fruits). The utility is influenced by pollination services supplied from the neighborhood because some of non-timber forest products result from animal pollination. When all of land parcels in the neighborhood of scale OP are forested, the landowner i receives the maximum utility of b because pollination services from the neighborhood are at maximum level. However, pi(t) is degraded as neighboring forests decrease due to a decreased production of non-timber forest products or an increased extinction risk of plant species providing non-timber products that lack pollinators. We simply assume that this degradation linearly increases with an increased fraction of non-forested land in the

neighborhood OP: pi ðtÞ ¼ b  dð1  di ðtÞÞ,

(3)

where d is the extent of degradation and di(t) is the forest density in the neighborhood OP of the parcel i (Fig. 1a). We define the regional fraction of forested land given in Eq. (3) as di ðtÞ ¼ ni ðtÞ=Op ,

(4)

where ni(t) is the number of forested parcels in the neighborhood OP. We assume OP ¼ 8, meaning the Moore neighborhood that comprises the eight parcels surrounding a focal parcel on a two-dimensional square lattice (Fig. 1a). (3) Payment for carbon storage services: an income transfer system from developed to developing countries: Forests provide services to store carbon: if forests are cleared, carbon dioxide is released to the atmosphere, which will contribute to the accelerated greenhouse effect. Here we assume an economic system for payment of the carbon storage services: an income transfer system from developed (the North) to developing (the South) countries. This has been originally proposed by Pearce (1991). In the following, we determine the amount of payment of the carbon storage services per hectare per year using the income transfer framework. Let s be the carbon amount that is released from deforestation per hectare. After Brown and Pearce (1994), we consider that s ranges from 100 to 200 metric tons, although we acknowledge that the magnitude of carbon emission by deforestation of tropical forests is the subject of considerable uncertainty and debate (Achard et al., 2002 ; Fearnside and Laurance, 2004 ; Ramankutty et al., 2007 ). The total amount of carbon released from deforestation of size F¯  FðtÞ is given as R ¼ sðF¯  FðtÞÞ, where F¯ and F (t ) represent the initial size of forests and the size of forests in year t . Suppose the global warming damage per year from deforestation is given by F (R ) that satisfies D0 40 and D00 40. We assume that the North aims to mitigate the global warming damages by transferring a part of their national incomes to the South for forest conservation. Let q be the payment of the North to the conserved forests in the South per hectare per year. For the size of conserved forest of F , the total cost for the North amounts to C ¼ qF þ DðsðF¯  FÞÞ. Given q , the North determines the size of conserved forest that minimizes C . That is, the demand of the North for the conserved forests of size Fd is given by solving q ¼ sD0 ðsðF¯  F d ÞÞ. At the demand–supply equilibrium, the current size of forest F (t ) is consistent with Fd , which allows having the equilibrium payment as qn ¼ sD0 ðsðF¯  FðtÞÞÞ.

(5) dqn =dF

2

00

From Eq. (5) , we have ¼ s D , which means that the payment for conserved forest decreases as remaining forest size increases. In other words, the payment rises as remaining forests get scarce. For the convenience, in this paper, we specify D(R) ¼ hR2 where h is positive constant. The equilibrium payment for conserved forests in the South per hectare per year is therefore given by ¯  xðtÞÞ, qn ¼ 2hs2 Fð1

(6)

where xðtÞ ¼ FðtÞ=F¯ representing the fraction of forested area in year t. For example, when we assume that F¯ equals the size of Rondo´nia state in Brazilian Amazon, F¯ is estimated as about 2.43  107 ha, and the fraction of forested area is about 81% in 1995 (Andersen et al., 2002). When we use F¯ ¼ 2:43  107 , s ¼ 150, x(t) ¼ 0.81, and h ¼ 109, the payment from the North to the South amounts to 442$/ha/year. Since h has not been

ARTICLE IN PRESS A. Satake et al. / Global Environmental Change 18 (2008) 768–775

estimated, we change the magnitude of h and investigate its impact on landowners’ deforestation decision and resultant landscape dynamics in the following analysis. (4) The utility of a forest owner is given by the sum of utility of nontimber forest products and carbon storage We assume  payment:  that the net utility of forested parcel i U Fi ðtÞ is given by the sum of utility of non-timber forest products and carbon storage payment:

771

where o is the discount factor that ranges from 0 to 1; r^ is the deforestation rate in the future (0pr^ p1) with which landowners anticipate the likelihood of future deforestation. o is the discount factor defined as 1/(1+^i) where ^i is an interest rate. In a similar manner, the expected discounted utilities of agriculture and abandoned land are given by V Ai ðtÞ ¼ c þ o½ð1  ZÞV Ai ðt þ 1Þ þ ZV Ei ðt þ 1Þ,

(10b)

V Ei ðtÞ ¼ o½ð1  mÞV Ei ðt þ 1Þ þ mV Fi ðt þ 1Þ.

(10c)

U Fi ðtÞ ¼ pi ðtÞ þ qn ¼ b  dð1  di ðtÞÞ þ gð1  xðtÞÞ,

(7)



where g ¼ 2hs F. This formulation expresses the situation where the net value of forest is influenced by actions by others at different scales; loss of forests at scale OP decreases U Fi ðtÞ because of degradation of pollination service, but loss of forests at scale OC increases U Fi ðtÞ because of payment of carbon storage services. In the following, we call d and g the ‘‘coefficient of pollination service damage’’ and the ‘‘coefficient of carbon storage payment’’ respectively. (5) Constant utilities for owners of agriculture and abandoned lands: We simply assume that the utilities of agricultural and abandoned lands are constant: UA ¼ c and UE ¼ 0. When landowners are engaged in agriculture, they receive utility c, which is the total revenue received (e.g. monetary benefits by crop sales) minus the cost incurred (e.g. cultivating, harvesting, and transporting costs). For example, in the state of Rondo´nia in Brazilian Amazon, the maximum utility from agriculture was estimated as 300$/ha/year (Pearce, 1996). The utility of an abandoned parcel is 0, lower than that of forested and agricultural land, because bare land has no economic values. Although we also examined the more realistic situation where the utility of agricultural lands depends on the surrounding forest density or on the total amount of agricultural crops produced, results and conclusions are not significantly different from the simplest situation where the utility is constant. 2.3. Decision about deforestation The deforestation rate by the landowner i in year t (ri(t)) is influenced by the net expected gain of deforestation (DVi(t)). The net expected gain of deforestation is defined as the expected discounted utility received from deforestation minus that of forest conservation lost through deforestation:

DV i ðtÞ ¼ V Ai ðtÞ  V Fi ðtÞ, V Ai ðtÞ

(8)

V Fi ðtÞ

where and are the expected discounted utilities of agricultural and forested parcel at time t as explained later. We assume that deforestation occurs more frequently if it results in a larger DVi(t): r i ðtÞ ¼

1 , 1 þ expðbDV i ðtÞÞ

(9)

where b is the positive constant. b is a parameter that controls the degree of sensitivity to the change of DVi(t). As b becomes infinitely large, the landowner’s behavior resembles a deterministic decision, and he chooses the best land-use state that represents the highest expected utility; otherwise it is stochastic. Satake and Iwasa (2006) give a detailed explanation why we use the decision dynamics in Eq. (9). The expected discounted utility of forested land is given as the cumulative sum of the current and the future utilities that are discounted over time: V Fi ðtÞ

¼

U Fi ðtÞ

þ o½ð1 

r^ ÞV Fi ðt

þ 1Þ þ

r^ V Ai ðt

þ 1Þ,

(10a)

When landowners assume that the extent of forest cover at global (x(t)) and regional scale (di(t)) changes slowly and stays at almost the present level in the future as well, the expected discounted utility becomes independent of time. Given this assumption, we obtain the net expected gain of deforestation for each parcel i (DVi) as

DV i ¼ K½f1  oð1  mÞgc  f1  oð1  Z  mÞgU Fi ,

(11)

where K ¼ 1=½1  oð2  r^  Z  mÞ  o2 fr^ ð1  ZÞ þ Zð1  mÞ þ mð1  r^ Þ  1g.

When b-N, landowners decide to deforest if   oZ UF . c4 1 þ 1  oð1  mÞ i

(12)

The above equation tells us how three parameters, the discount factor (o), the abandonment rate (Z), and the forest recovery rate (m), jointly influences the deforestation decision. Satake and Rudel (2007) derived the same formula in Eq. (11). Parameters used in the model are summarized in Table 1. 2.4. Analysis of scale mismatch across three different situations To highlight the impact of scale mismatch and the payment for carbon storage services on landscape patterns, we make a comparison between three different situations on scale mismatches in the following analysis. (1) No scale mismatch (d ¼ g ¼ 0): Forest utility is simply given by the basic utility from non-timber forest products (b) regardless of the size of remaining forests in scales OP and OC. This situation is similar to a simple Markov model analyzed by Satake and Iwasa (2006) where the utility of forest is independent of actions by others. (2) Scale mismatch only for pollination services (d40, g ¼ 0): The utility of forests is degraded by deforestation in neighboring parcels through the destruction of pollination services, but the payment for carbon storage services is not taken into account. Table 1 summary of parameters b c

o Z m d

g b di(t) xi(t) V Fi ðtÞ

Snap-shot utility of forest conservation Snap-shot utility of agriculture Discount factor Rate of agricultural abandonment Rate of forest recovery Coefficient of pollination service damage Coefficient of carbon storage payment Degree of stochasticity in decision making Density of forested parcel in the neighborhood Density of forested parcel as a whole Expected discounted utility of forest conservation

V Ai ðtÞ

Expected discounted utility of agriculture

V Ei ðtÞ

Expected discounted utility of abandonment

ARTICLE IN PRESS A. Satake et al. / Global Environmental Change 18 (2008) 768–775

For each situation, we investigate the macroscopic spatial patterns that emerge from aggregated dynamics of land-use change at each parcel by computer simulations. We then assess how well-being at individual and societal levels are influenced by the different scale mismatch situations. To generate spatial landuse patterns, we simulate the process of land-use change in a two dimensional lattice of 128  128. Note that the scale for carbon storage services is consistent with the system size (i.e. OC ¼ 128  128). We assumed a periodic boundary condition, i.e. a lattice of a torus shape in which the far-right column is the nearest neighbor of the far-left column, and the top row is nearest neighbor of the bottom row. Using the spatial land-use pattern generated by the model, we calculated landowners’ well-being. We assume that the wellbeing of the landowner i in year t is equivalent to the utility that the landowner i receives from land management in year t (i.e. U Si ðtÞ, SA{F,A,E}). The net well-being by the landowner i is calculated by taking average of U Si ðtÞ over long run (we sampled 1000 time points). We denote the net well-being by the landowner i as Wi. Wi may be significantly different among landowners in a society. To investigate the inequality regarding the well-being among landowners, we calculated the mean ¯ and and standard deviation of Wi , which is given by W SD respectively.

3. Results 3.1. Land-use patterns in spatial settings When there is no scale mismatch (d ¼ g ¼ 0), the landscape pattern converges to either forested or deforested landscapes depending on the inequality specified in Eq. (12). From Eq. (12), we find that the deforestation rate is suppressed as o and Z increases, while it increases as m increases. This implies that longterm perspective by decision makers (large o) and frequent abandonment of agricultural land (large Z) leads to forested landscape, composed largely of forested parcels. In contrast, fast recovery of forest (large m) favors a deforested landscape that is composed of the mixture of agricultural and abandoned lands. The scale mismatch for pollination services (i.e. d40, g ¼ 0) causes a degradation of pollination services once deforestation occurs on neighboring parcels (Eq. (3)). Starting from a forested landscape, the resultant decline of forest utility reduces the incentive for forest conservation and elevates the deforestation rate, which allows an expansion of agricultural land in the neighborhood of the deforested parcels (Figs. 2a and 3b). Agriculture spreads until most of forests are converted (Fig. 3c). After the fraction of agricultural land attains the peak, it gradually decreases because of agricultural abandonment (Fig. 2a), and converges to the stable landscape where the agricultural and abandoned lands are mixed (i.e. deforested landscape; Fig. 3d). In sum, the degradation of forest ecosystem services caused by the scale mismatch for pollination services triggers a shift from forested to deforested landscape. If we allow payments for carbon storage services (i.e. d, g40), this trend changes. Starting from the forested landscape, the fraction of forested land initially decreases due to agricultural expansion in the neighborhood of deforested parcels because

1.0

fraction of land use

(3) Scale mismatch for pollination and carbon storage services and carbon storage services are paid (d, g40): The utility of forests results from offsetting trends. It declines because deforestation at scale OP causes a loss of pollination services at the same time that it increases because deforestation at scale OC leads to larger payments for carbon storage services.

0.8 agriculture

0.6 0.4 abandoned

0.2

forest 0.0 0

10

20

30

40

50

1.0

fraction of land use

772

0.8 forest

0.6 0.4

agriculture

0.2 abandoned

0.0 0

10

20

30

40

50

time Fig. 2. Temporal change of the fraction of forested, agricultural, and abandoned parcels. (a) A transition from forested to deforested landscape. d ¼ 300, g ¼ 0. (b) A transition from forested to mosaic landscape. d ¼ 300, g ¼ 270. Solid thick lines: the fraction of forested parcels. Solid thin lines: the fraction of agricultural land parcels. Dashed lines: the fraction of abandoned parcels. Other parameters are b ¼ 270, c ¼ 300, Z ¼ 0.03, m ¼ 0.03, o ¼ 0.9, and b ¼ 1.

pollination services are degraded (Figs. 2b and 3e). However, the decline of forested land at global scale increases the value of forest conservation because the related increases in damages from global warming raises the payments for the carbon storage services of the remaining forests. When the loss of forest value through degradation of pollination services is countered by an increase in forest value attributable to carbon storage payments, further deforestation stops (Figs. 2b and 3f). The resulting landscape pattern is a mosaic of forested clusters and deforested clusters (Fig. 3g). Forested clusters emerge because the forested parcels surrounded by forests are conserved with high probability. These forest patches provide rich ecosystem services including healthy pollination service. In these settings the incentives for forest conservation are higher than on lands located close to agriculture or abandoned parcels where pollination services are degraded. Once forested clusters emerge, the landscape pattern does not change much. Only a small number of land clearings occur during each time period at the boundary between forested and deforested clusters. Within the deforested clusters, a cycle of agricultural abandonment, forest recovery, and deforestation happens continuously. 3.2. Impact of scale mismatches on well-being To investigate the impact of scale mismatches on well-being and equity, we calculated the mean and standard deviation of the

ARTICLE IN PRESS A. Satake et al. / Global Environmental Change 18 (2008) 768–775

t=2

t=5

773

t = 50

t=0

Degradation of pollination services t=2

t=5

t = 50

Degradation of pollination services + payment for carbon storage services Fig. 3. Spatio-temporal patterns of landscape dynamics: (a), (b), (c) and (d) A transition from forested to deforested landscape. d ¼ 300, g ¼ 0. (a), (e), (f) and (g) A transition from forested to mosaic landscape. d ¼ 300, g ¼ 270. Black, white and gray cells represent forested, agricultural, and abandoned land parcels respectively. Other parameters are b ¼ 270, c ¼ 300, Z ¼ 0.03, m ¼ 0.03, o ¼ 0.9, and b ¼ 1.

¯ and SD, for various net well-being among landowners, W magnitudes of g/d, the relative magnitude of g (coefficient of carbon storage payment) and d (coefficient of pollination service damage). When the magnitude of g against d is small, degradation of pollination service at regional scale is severe and landowners recoup relatively little of the lost value through payments for carbon storage. Therefore the utility of forest is low, which increases the deforestation rate and deforested landscapes emerge (Fig. 4 a). In the deforested landscape, landowners may receive high utility from agriculture. But once agricultural lands are abandoned, utility level drops off to a level even lower than that of forested land. Landowners then need to suffer from low utility for a long time until abandoned lands finally revert back to forested ¯ is therefore low, and the lands. The mean of net well-being (W) standard deviation (SD) is small (Fig. 4b). The distribution of the net well-being of different landowners in a society (Wi, iA{1,yN}) draws a unimodal distribution (Fig. 5a). As the magnitude of coefficient of payment of carbon storage services (g) against that of degradation of pollination services (d) increases, the fraction of forested parcels increases (Fig. 4 a), because the enhanced payments for carbon storage services ¯ creates an incentive for forest conservation, which increases W (Fig. 4b). SD also increases, but once it attains a peak, it shows a gradual decline (Fig. 4b). A large SD is observed especially when the landscape is composed of the mixture of forested and deforested clusters as exemplified in Fig. 3g. When SD is large, the net well-being of different landowners in a society draws a clear bimodal distribution (Fig. 5b), implying the existence of two groups of landowners: one enjoying high income from the management of healthy forested lands and another receiving low income from the management of a deforested parcel. In this setting, the net well-being by landowners managing forested parcel is approximated as b+g(1x*) where x* is the global density of forests at equilibrium. Their net well-being in

this situation is even higher than their net well-being when most of land parcels are covered with forests (i.e. b). This difference in well-being occurs because forest losses caused by the degradation of pollination service due to a scale mismatch benefit these forest landholders through enhanced payments for carbon storage services from the remaining forests. Once the payments for carbon storage services get high enough (i.e. the magnitude of g against d is large enough), each landowner enhances the incentive to conserve forests, resulting in an increased fraction of forested parcels (Fig. 4a). This would increase the mean of net well-being (Fig. 4b) and reduce the inequities in the society (Fig. 5c). In sum, payments for carbon storage services can increase the mean of net well-being (Fig. 3b), but they also create inequities in well-being among individual landowners when the degradation of pollination services at a regional scale is accompanied by payments for carbon storage services at the global scale.

4. Conclusion We theoretically explored the consequences of scale mismatches involving deforestation decisions, provision of pollination services, and payment for carbon storage services. Our model assumes payments for one environmental service (carbon storage) and no payments for another environmental service (pollination services). The income transfer system for carbon storage (Pearce, 1991) formalized in the paper showed that the payment for conserved forests from developed to developing countries rises as the size of remaining forests decreases. The scale of provision of pollination services was assumed to be intermediate between the local scale of household benefits from decisions to deforest and the global scale of payments for carbon storage services. This causes a situation where the net value of forest is influenced by actions by others at different scales; deforestation at the scale of

ARTICLE IN PRESS 774

A. Satake et al. / Global Environmental Change 18 (2008) 768–775

2000 number of landowners

forested landscape 1.0 forest

landscape

0.6

500

0

agriculture

0

0.2

1.0

1.5

2.0

forested landscape

300

W

deforested landscape

number of landowners

0.5

100

200

2000

abandoned

0.0

mean and SD of net well-being

1000

0.4 300

400

high-income group

1500

1000 low-income group 500

0

200

0

100

0

100

200

300

400

300

400

12000 100 SD 0 0.5

1.0

1.5

2.0

γ/δ Fig. 4. (a) Plot of proportion of land parcels and (b) plot of mean and standard deviation (SD) of net well-being along the relative magnitude of coefficient of carbon storage payment (g) and that of pollination service damage (d). Other parameters are b ¼ 270, c ¼ 300, Z ¼ 0.03, m ¼ 0.05, o ¼ 0.9, and b ¼ 1.

pollination services decreases the utility of forest at parcel level because of degradation of pollination service, but deforestation at global scale increases global warming damages and raises the forest value through increased payment for carbon storage services. The results from analyses of our model indicate that payments for carbon storage services may both protect forests as well as increase the mean well-being of people (Fig. 4). However the model also predicts that the scale mismatch between pollination services and payment of carbon storage services may result in the creation of inequalities among landowners regarding income level (Fig. 5b); landowners with their lands in forest receive significantly larger payments than landowners who have either agricultural or abandoned land. The inequity in economic returns in a society when carbon storage payments exist sounds a cautionary note about the effects of environmental reforms on the social and economic fabric. Equity is a central pillar of sustainable development (World Commission on Environment and Development, 1987), global environmental justice (Schlosberg, 2004), and a key criterion for sustainable environmental governance (Adger et al., 2003). Brown and Corbera (2003) distinguished three elements of equity: equity in access, equity in decision-making and equity in outcome. Our model considered equity in access and equity in decision making: landowners are assumed to have property rights and to

number of landowners

fraction of land uses

deforested 0.8

1500

10000 8000 6000 4000 2000 0 200 net well-being

Fig. 5. Distribution of landowners along the net well being. (a) g ¼ 30; (b) g ¼ 160; (c) g ¼ 300. Other parameters are b ¼ 270, c ¼ 300, Z ¼ 0.03, m ¼ 0.03, m ¼ 0.9, b ¼ 1, and d ¼ 300.

participate equally in the program of payment for carbon storage services. Despite these conditions, unequal outcomes resulted from the heterogeneous landscape created by the scale mismatch between the payment for carbon storage services at global scale and damages for pollination services at regional scale. Because high-income landowners usually have a larger proportion of their lands in forests than do low-income landowners (Barraclough and Collarte, 1973), they would stand to benefit disproportionately from a system of payments for ecological services compared with low-income landowners. This circumstance suggests that market-based approaches for the conservation of single ecosystem service may reproduce unequal power relations among landowners. In the search for policies that produce more equitable outcomes for human welfare, ecologists, social, and political scientists need to work together to map coupled ecological and social conditions at multiple spatial scales. We believe that theoretical knowledge about the landscape dynamics associated with scale mismatches and reforms to correct them should prove useful to policymakers in this search.

ARTICLE IN PRESS A. Satake et al. / Global Environmental Change 18 (2008) 768–775

We acknowledge that our model is too simplified for a comprehensive understanding of the causes and consequences of scale mismatches. An enhanced model would incorporate other factors such as the spatially dependent utility of agricultural and abandoned parcels or the increased rate of agricultural abandonment in areas containing little forested land. Having created a minimum model that incorporates a few essential factors, future investigations might address how the addition of these missing factors would alter the model’s predictions both in theory and in application to empirical cases of scale mismatches.

Acknowledgements This work was partly supported by a Grant-in-Aid from JSPS to A.S. and Keio Gijuku Academic Development Funds to A.O. We thank Frank Wa¨tzold and two anonymous reviewers for many useful comments. References Achard, F., Eva, H.D., Stibig, H.-J., Mayaux, P., Gallego, J., Richards, T., Malingreau, J.-P., 2002. Determination of deforestation rates of the world’s humid tropical forests. Science 297, 999–1002. Adger, W.N., Brown, K., Fairbrass, J., Jordan, A., Paavola, J., Rosendo, S., Seyfang, G., 2003. Governance for sustainability: towards a ‘thick’ analysis of environmental decision making. Environment and Planning A 35, 1095–1110. Aizen, M.A., Feinsinger, P., 1994. Forest fragmentation, pollination, and plant reproduction in a chaco dry forest. Argentina. Ecology 75, 320–341. Andersen, L.E., Granger, C.W.J., Reis, E.J., Weinhold, D., Wunder, S., 2002. The Dynamics of Deforestation and Economic Growth in the Brazilian Amazon. Cambridge University Press, Cambridge. Barraclough, S., Collarte, J.C., 1973. Agrarian Structure in Latin America. Lexington Books, Lexington. Brown, K., Corbera, E., 2003. Exploring equity and sustainable development in the new carbon economy. Climate Policy 3, S41–S56. Brown, K., Pearce, D.W., 1994. The economic value of non-market benefits of tropical forests: carbon storage. In: Weiss, J. (Ed.), The Economics of Project Appraisal and the Environment. Edward Elgar, London, pp. 102–123. Carpenter, S.R., DeFries, R., Dietz, T., Mooney, H.A., Polasky, S., Reid, W.V., Scholes, R.J., 2006. Millennium ecosystem assessment: research NEEDS. Science 314, 257–258. Corbera, E., Kosoy, N., Tuna, M.M., 2007. Equity implications of marketing ecosystem services in protected areas and rural communities: case studies from Meso-America. Global Environmental Change 17, 365–380. Cumming, G.S., Cumming, D.H.M., Redman, C.L., 2006. Scale mismatches in socialecological systems: causes, consequences, and solutions. Ecology and Society 11, 14 [online] /URL: http://www.ecologyandsociety.org/vol11/iss1/ art14/S. Dixon, R.K., Brown, S., Houghton, R.A., Solomon, A.M., Trexler, M.C., Wisniewski, J., 1994. Carbon pools and flux of global forest ecosystems. Science 263, 185–190. Engel, S., Pagiola, S., Wunder, S., 2008. Designing payments for environmental services in theory and practice: an overview of the issues. Ecological Economics 65, 663–674. Fearnside, P.M., 1996. Amazonian deforestation and global warming: carbon stock in vegetation replacing Brazil’s Amazon forest. Forest Ecology and Management 80, 21–34. Fearnside, P.M., 1999. Forests and global warming mitigation in Brazil: Opportunities in the Brazilian forest sector for responses to global warming under the ‘‘clean development mechanism’’. Biomass & Energy 16, 171–189.

775

Fearnside, P.M., 2000. Global warming and tropical land-use change: Greenhouse gas emissions from biomass burning, decomposition and soils in forest conversion, shifting cultivation and secondary vegetation. Climate Change 46, 115–158. Fearnside, P.M., Laurance, W.F., 2004. Tropical deforestation and greenhouse-gas emissions. Ecological Applications 14, 982–986. Folke, C., Pritchard, L., Berkes, F., Colding, J., Svedin, U., 1998. The problem of fit between ecosystems and institutions. IHDP Working Paper Number 2. International Human Dimensions Programme on Global Environmental Change (IHDP), Bonn. Groom, M.J., 2001. Consequences of subpopulation isolation for pollination, hervivory, and population growth in Clarkia concinna concinna (Onagraceae). Houghton, R.A., Skole, D.L., Nobre, C.A., Hackler, J.L., Lawrence, K.T., Chomentowski, W.H., 2000. Annual fluxes or carbon from deforestation and regrowth in Brazilian Amazon. Nature 403, 201–304. Kleijn, D., Sutherland, W.J., 2003. How effective are European agri-environment schemes in conserving and promoting biodiversity? Journal of Applied Ecology 40, 947–969. Klein, A.M., Vaissie´re, B.E., Cane, J.H., Steffan-Dewenter, I., Cunningham, S.A., Kremen, C., Tscharntke, T., 2007. Importance of pollinators in changing landscapes for world crops. Proceedings of the Royal Society B 274, 303–313. Landell-Mills, N., Porrai, I., 2002. Silver Bullet or Fool’s Gold? A Global Review of Markets for Forest Envirronmental Services and their Impacts on the Poor. Instruments for Sustainable Private Sector Forestry Series. IIED, London. Millennium Ecosystem Assessment, 2005. Ecosystems and Human Well-being: Current State and Trends. Island Press, Washington, DC. Pagiola, S., Landell-Mills, N., Bishop, J., 2002. Market-based mechanisms for forest conservation and development. In: Pagiola, S., Landell-Mills, N., Bishop, J. (Eds.), Selling Forest Environmental Services. Earthscan, London. Pearce, D., 1991. An economic approach to saving the tropical forests. In: Helm, D. (Ed.), Economic Policy Towards the Environment. Blackwell Publishers, Oxford, pp. 239–262. Pearce, D., 1996. Global environmental value and the tropical forests: demonstration and capture. In: Adamowicz, W.L., Boxall, P., Luckert, M.K., Phillips, W.E., White, W.A. (Eds.), Forestry, Economics and the Environment. CAB International, Wallingford. Ramankutty, N., Gibbs, H.K., Achard, F., Defries, R., Foley, J.A., Houghton, R.A., 2007. Challenges to estimating carbon emissions from tropical deforestation. Global Change Biology 13, 51–66. Ricketts, T.H., 2004. Tropical forest fragments enhance pollinator activity in nearby coffee crops. Conservation Biology 18, 1262–1271. Sa´nchez-Azofeifa, G.A., Pfaff, A., Robalino, J.A., Boomhower, J.P., 2007. Costa Rica’s payment for environmental services program: intention, implementation, and impact. Conservation Biology 21, 1165–1173. Santilli, M., Moutinho, P., Schwartzman, S., Nepstad, D., Curran, L., Nobre, C., 2005. Tropical deforestation and the Kyoto protocol. Climate Change 71, 267–276. Satake, A., Iwasa, Y., 2006. Coupled ecological and social dynamics in a forested landscape: the deviation of the individual decisions from the social optimum. Ecological Research 21, 370–379. Satake, A., Rudel, T.K., 2007. Modelling the forest transition: forest scarcity and ecosystem service hypotheses. Ecological Applications 17, 2024–2036. Schlosberg, D., 2004. Reconceiving environmental justice: global movements and political theories. Environmental Politics 13, 517–540. Southwick, E.E., Southwick, L., 1992. Estimating the economic value of honey-bees (Hymenoptera; Apidae) as agricultural pollinators in the United States. Journal of Economic Entomology 85, 621–633. Steffan-Dewenter, I., Tscharntke, T., 1999. Effects of habitat isolation on pollinator communities and seed set. Oecologia 121, 432–440. Vitousek, P.M., Mooney, H.A., Lubchenco, J., Melillo, J.M., 1997. Science 277, 400–494. World Commission on Environment and Development, 1987. Our Common Future. Oxford University Press, Oxford. WRI, 2000. World resources 2000–2001: people and Ecosystems: the fraying web of life. United Nations Development Programme, United Nations Environment Programme, World Bank, World Resources Institute, Washington, DC.