Ecosystem services at risk: integrating spatiotemporal dynamics of supply and demand to promote long-term provision

Abstract

Ecosystem services (ESs)—the benefits provided to people by nature—are fundamental to human well-being. The sustainable provision of such services is constrained by both spatial and temporal dynamics of ES supply (S) and demand (D), but the temporal aspect is usually disregarded despite its high relevance in sustainability analyses. Here, we propose a conceptual framework integrating both spatial and temporal dynamics of S and D to forecast trends in ES provision. In this framework, we identify three temporal trends of ES threat: steady undersupply (S decreases through time, and D exceeds it), cyclical highly variable undersupply (S is cyclical through time, and D exceeds S periodically), and stochastic undersupply (unpredictable events might change either S or D). Once the type of S/D trend is identified, landscape management strategies can be designed accordingly to increase flow between S and D areas, reducing the risk of ES extirpation over time. Ecosystem services (ESs)—the benefits provided to people by nature—are fundamental to human well-being. The sustainable provision of such services is constrained by both spatial and temporal dynamics of ES supply (S) and demand (D), but the temporal aspect is usually disregarded despite its high relevance in sustainability analyses. Here, we propose a conceptual framework integrating both spatial and temporal dynamics of S and D to forecast trends in ES provision. In this framework, we identify three temporal trends of ES threat: steady undersupply (S decreases through time, and D exceeds it), cyclical highly variable undersupply (S is cyclical through time, and D exceeds S periodically), and stochastic undersupply (unpredictable events might change either S or D). Once the type of S/D trend is identified, landscape management strategies can be designed accordingly to increase flow between S and D areas, reducing the risk of ES extirpation over time. Ecosystem services (ESs), understood as the benefits humans obtain from ecosystems, are of fundamental importance to human well-being.1Millennium Ecosystem AssessmentEcosystems and Human Well-Being: Synthesis. Island Press, 2005http://www.millenniumassessment.org/en/Synthesis.aspxGoogle Scholar They are classified in provisioning, regulating, and cultural services and include, for example, the provision of food and water, air and water purification, disease regulation, and artistic inspiration.1Millennium Ecosystem AssessmentEcosystems and Human Well-Being: Synthesis. Island Press, 2005http://www.millenniumassessment.org/en/Synthesis.aspxGoogle Scholar Since humans depend directly on these services to survive, they have great potential to influence environmental decisions and landscape management.2Villamagna A.M. Angermeier P.L. Bennett E.M. Capacity, pressure, demand, and flow: a conceptual framework for analyzing ecosystem service provision and delivery.Ecol. Complex. 2013; 15: 114-121Crossref Scopus (344) Google Scholar Since the 1960s, the demand (D) for ES per capita has increased exponentially, very often in detriment of natural capital, as a result of unsustainable practices.1Millennium Ecosystem AssessmentEcosystems and Human Well-Being: Synthesis. Island Press, 2005http://www.millenniumassessment.org/en/Synthesis.aspxGoogle Scholar,3Daily G.C. Nature's Services: Societal Dependence on Natural Ecosystems. Island Press, 1997Google Scholar, 4Palmer M. Bernhardt E. Chornesky E. Collins S. Dobson A. Duke C. Gold B. Jacobson R. Kingsland S. Kranz R. et al.Ecology for a crowded planet.Science. 2004; 304: 1251-1252Crossref PubMed Scopus (377) Google Scholar, 5Vitousek P.M. Mooney H.A. Lubchenco J. Melillo J.M. Human domination of Earth's ecosystems.Science. 1997; 277: 494-499Crossref Scopus (6316) Google Scholar This has resulted in degraded ecosystems unable to provide ES, which could put human well-being at risk. For example, between 1997 and 2011, global land-use changes led to the decline in 62% of ESs6IPBESDíaz S. Settele J. Brondízio E.S. Ngo H.T. Guèze M. Agard J. Arneth A. Balvanera P. Brauman K.A. Butchart S.H.M. Summary for Policymakers of the Global Assessment Report on Biodiversity and Ecosystem Services of the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services. IPBES Secretariat, 2019Google Scholar and in an estimated loss of $4.3 to $20.2 trillion/year of ESs.7Constanza R. de Groot R. Sutton P. can der Ploeg S. Andeson S.J. Kubiszewski I. Farber S. Turner R.K. Changes in the global value of ecosystem services.Glob. Environ Change. 2014; 26: 152-158Crossref Scopus (2507) Google Scholar Therefore, it is of great importance to achieve a sustainable management of ESs to guarantee their provision for future generations. The capacity of an ecosystem to generate services differs from the actual services delivered to society. For ES provision to occur, an interaction between who receives the benefit (ES D; see Note S1 for a glossary of definitions) and what or who can provide it (ES supply [S]) is needed, and thus ES flow from S to D is crucial. Even though previous studies have shown that ES S and D vary spatially,8Syrbe R.U. Grunewald K. Ecosystem service supply and demand – the challenge to balance spatial mismatches.Int. J. Biodiv Sci. Ecosyst. Serv. Manag. 2017; 13: 148-161Crossref Scopus (51) Google Scholar,9Sturk J. Poorting A. Verburg P.H. Mapping ecosystem services: the supply and demand of flood regulation services in Europe.Ecol. Indic. 2013; 38: 198-211Crossref Scopus (149) Google Scholar only a few studies have explored ES relationships across different spatial scales.10Scholes R. Reyers B. Biggs R. Spierenburg M.J. Duriappah A. Multi-scale and cross-scale assessments of social–ecological systems and their ecosystem services.Curr. Opin. Environ. Sustain. 2013; 5: 16-25Crossref Scopus (152) Google Scholar,11Raudsepp-Hearne C. Peterson G. Scale and ecosystem services: how do observation, management, and analysis shift with scale—lessons from Québec.Ecol. Soc. 2016; 21: 16Crossref Scopus (75) Google Scholar Moreover, ES S and D also vary temporally because of natural changes in ecosystem functions or to changes induced by human activity.12Nicholson E. Mace G.M. Armsworth P.R. Atkinson G. Buckle S. Clements T. Ewers R.M. Fa J.E. Gardner T.A. Gibbons J. et al.Priority research areas for ecosystem services in a changing world.J. Appl. Ecol. 2009; 46: 1139-1144Google Scholar In spite of the importance of temporal dynamics evaluation for both short- and long-term projection of ES provision,13Rau A.L. Wehrden H.V. Abson D.J. Temporal dynamics of ecosystem services.Ecol. Econ. 2018; 151: 122-130Crossref Scopus (30) Google Scholar most assessments on ES S and D are performed with little or no consideration of temporal dimensions of change.14Abson D.J. Termansen M. Valuing ecosystem services in terms of ecological risks and returns.Conserv. Biol. 2011; 25: 250-258PubMed Google Scholar Because of the idiosyncratic status of evaluation and understanding of spatiotemporal dynamics on ES provision (see Box 1 for an overview of knowledge gaps on spatiotemporal dynamics of S and D), it is a challenge to adequately integrate ESs into decision-making processes15Bastian O. Grunewald K. Syrbe R.W. Space and time aspects of ecosystem services, using the example of the EU Water Framework Directive.Int. J. Biodivers Sci. Ecosyst. Serv. Manag. 2012; 8: 5-16Crossref Scopus (53) Google Scholar or in landscape management and/or planning. To understand and propose effective efforts for land-use management and long-term sustainability16Burkhard B. Kroll F. Nedkov S. Muller F. Mapping ecosystem service supply, demand and budgets.Ecol. Indic. 2012; 21: 17-29Crossref Scopus (1151) Google Scholar where the provision of ES S is guaranteed over time,17Wu J. Landscape sustainability science: ecosystem services and human well-being in changing landscapes.Landscape Ecol. 2013; 28: 99-1023Crossref Scopus (694) Google Scholar it is crucial to understand how S and D vary spatiotemporally by considering these dynamics together.15Bastian O. Grunewald K. Syrbe R.W. Space and time aspects of ecosystem services, using the example of the EU Water Framework Directive.Int. J. Biodivers Sci. Ecosyst. Serv. Manag. 2012; 8: 5-16Crossref Scopus (53) Google Scholar,16Burkhard B. Kroll F. Nedkov S. Muller F. Mapping ecosystem service supply, demand and budgets.Ecol. Indic. 2012; 21: 17-29Crossref Scopus (1151) Google Scholar In this perspective, we performed a literature review in order to understand the current knowledge gaps on spatiotemporal dynamics of ES S and present a conceptual framework integrating spatiotemporal dynamics of ES S and D to forecast trends in ES provision. Our framework integrates multiple temporal categories of behavior of interaction between S and D according to the nature of ESs. Once the temporal trend of the S/D relationship is identified, we denote how spatial planning can be applied in order to increase ES flow between S and D areas, reducing vulnerability and ensuring long-term ES provision for a variety of ESs. Finally, we discuss some challenges and opportunities when evaluating spatiotemporal dynamics of S/D relationships.Box 1Current knowledge gaps on spatiotemporal dynamics of ES S and DIn a systematic literature review on 1,545 articles published between 1994 and 2019 (see details in the experimental procedures), we found that only 3.2% of the papers performed spatiotemporal evaluations of ES, including S and/or D. Beyond the scarce literature, our review identified a considerable portion (44%; n = 142) of ES evaluations including spatiotemporal relationships of S and/or D performed for provisioning, regulating, and cultural services (34% and 17%, respectively). We also found an increased interest in understanding spatiotemporal changes on fresh water (18%), raw materials (18%), and food (12%). Changes on provision of services related to human safety, such as erosion prevention (12%) and local climate (10%), have also been receiving attention. On the other hand, regulation services such as pollination and pest control and the moderation of natural disasters are still poorly investigated in such spatiotemporal aspects (see figure below). All selected articles can be found in Table S1.Evaluations on spatiotemporal dynamics of ES provision have shown that most of the time (i.e., 84%), S is the only aspect evaluated. Both components together (S and D) appear 14.5% of the time (n = 21), and D alone rarely appears (0.7%; n = 1). Although the current scarcity of studies incorporates spatiotemporal evaluations on ES provision, they already cover a wide variation of temporal windows from a few months up to 200 years. In general, the time windows were mainly lower than 50 years (18.05 ± 8.06 years; Figure S4). Studies using projections or scenarios often have larger time windows (>50 years; 136 ± 41.10 years).We identified four frequency classes in S and D measurements: (1) time series, i.e., measures across temporal windows obtained at equal intervals of time (yearly) (S: 77 evaluations, 53.1%; Figures S5A and S5; D: 12 evaluations, 80%); (2) time intervals, i.e., measures at snapshots across different intervals of time (S: 42, 29%; D: 2, 13.3%; interval between measures: 26.9 ± 27 years; data were mostly taken at three snapshots in time; 36%; Figure S5B); (3) projections, i.e., predictions of a given ES provision at a unique moment in the future (minimum–maximum range: 27–140 years); and (4) seasonal, i.e., evaluations across seasons along 1 year (5.65%).Number of studies identified in the literature review (n = 142) evaluating spatiotemporal aspects of ES S (gray) and S and D (purple) after ES classification.19Maron M. Mitchell M.G.E. Runting R.K. Rhodes J.R. Mace G.M. Keith D.A. Watson J.E.M. Towards a threat assessment framework for ecosystem services.Trends Ecol. Evol. 2017; 32: 240-248Abstract Full Text Full Text PDF PubMed Scopus (41) Google Scholar In a systematic literature review on 1,545 articles published between 1994 and 2019 (see details in the experimental procedures), we found that only 3.2% of the papers performed spatiotemporal evaluations of ES, including S and/or D. Beyond the scarce literature, our review identified a considerable portion (44%; n = 142) of ES evaluations including spatiotemporal relationships of S and/or D performed for provisioning, regulating, and cultural services (34% and 17%, respectively). We also found an increased interest in understanding spatiotemporal changes on fresh water (18%), raw materials (18%), and food (12%). Changes on provision of services related to human safety, such as erosion prevention (12%) and local climate (10%), have also been receiving attention. On the other hand, regulation services such as pollination and pest control and the moderation of natural disasters are still poorly investigated in such spatiotemporal aspects (see figure below). All selected articles can be found in Table S1. Evaluations on spatiotemporal dynamics of ES provision have shown that most of the time (i.e., 84%), S is the only aspect evaluated. Both components together (S and D) appear 14.5% of the time (n = 21), and D alone rarely appears (0.7%; n = 1). Although the current scarcity of studies incorporates spatiotemporal evaluations on ES provision, they already cover a wide variation of temporal windows from a few months up to 200 years. In general, the time windows were mainly lower than 50 years (18.05 ± 8.06 years; Figure S4). Studies using projections or scenarios often have larger time windows (>50 years; 136 ± 41.10 years). We identified four frequency classes in S and D measurements: (1) time series, i.e., measures across temporal windows obtained at equal intervals of time (yearly) (S: 77 evaluations, 53.1%; Figures S5A and S5; D: 12 evaluations, 80%); (2) time intervals, i.e., measures at snapshots across different intervals of time (S: 42, 29%; D: 2, 13.3%; interval between measures: 26.9 ± 27 years; data were mostly taken at three snapshots in time; 36%; Figure S5B); (3) projections, i.e., predictions of a given ES provision at a unique moment in the future (minimum–maximum range: 27–140 years); and (4) seasonal, i.e., evaluations across seasons along 1 year (5.65%). Number of studies identified in the literature review (n = 142) evaluating spatiotemporal aspects of ES S (gray) and S and D (purple) after ES classification.19Maron M. Mitchell M.G.E. Runting R.K. Rhodes J.R. Mace G.M. Keith D.A. Watson J.E.M. Towards a threat assessment framework for ecosystem services.Trends Ecol. Evol. 2017; 32: 240-248Abstract Full Text Full Text PDF PubMed Scopus (41) Google Scholar Most ES studies do not consider the temporal dimension despite the importance of this dynamic for the sustainability of the ES provision (Box 1). We propose a conceptual framework that explicitly integrates both spatial and temporal components with ES S/D dynamics. We used as a baseline for this framework previous proposals from both Maron et al.20Fischer J. Lindenmayer D.B. The conservation value of paddock trees for birds in a variegated landscape in southern New South Wales. Species composition and site occupancy patterns.Biodivers Conserv. 2002; 11: 807-832Crossref Scopus (99) Google Scholar and Rau et al.,13Rau A.L. Wehrden H.V. Abson D.J. Temporal dynamics of ecosystem services.Ecol. Econ. 2018; 151: 122-130Crossref Scopus (30) Google Scholar which advanced ideas regarding the temporal dynamics of ES provision. Maron and colleagues20Fischer J. Lindenmayer D.B. The conservation value of paddock trees for birds in a variegated landscape in southern New South Wales. Species composition and site occupancy patterns.Biodivers Conserv. 2002; 11: 807-832Crossref Scopus (99) Google Scholar presented a framework for classifying risks to adequate ES provision by combining information on both ES S and ES D and categorizing the extent to which they are threatened in a Red List-type system assessment. Thus, if an ES is classified as least concerned, for example, it means that S exceeds or meets D and is predicted to continue to do so in the long term. On the other hand, if an ES is categorized as endangered, it means that D exceeds S (see details in Maron et al.20Fischer J. Lindenmayer D.B. The conservation value of paddock trees for birds in a variegated landscape in southern New South Wales. Species composition and site occupancy patterns.Biodivers Conserv. 2002; 11: 807-832Crossref Scopus (99) Google Scholar). However, since Maron et al.’s20Fischer J. Lindenmayer D.B. The conservation value of paddock trees for birds in a variegated landscape in southern New South Wales. Species composition and site occupancy patterns.Biodivers Conserv. 2002; 11: 807-832Crossref Scopus (99) Google Scholar framework's main goal was simply to classify ES into risk categories, it does not consider the temporal dynamics explicitly. On the other hand, Rau and colleagues13Rau A.L. Wehrden H.V. Abson D.J. Temporal dynamics of ecosystem services.Ecol. Econ. 2018; 151: 122-130Crossref Scopus (30) Google Scholar propose a framework to classify ES temporal dynamics according to different patterns: a linear dynamic (where S and D constantly increase or decrease), a periodic dynamic (where dynamics patterns are regularly repeated through time), or a non-linear dynamic (where patterns have either irregular amplitude or periodicity, directly affected by the unpredictable or erratic events that cause perturbations, such as natural disasters, or changes in agricultural management practices). Yet, in their framework, they do not consider threat categories or the ratio of S over D, i.e., how much of S is meeting D through time, which is crucial for sustainable provision of ESs. Thus, we integrate their ideas in a more complete framework by considering how S, D, and the ratio S/D (see below) may change over time by applying the ES risk category,20Fischer J. Lindenmayer D.B. The conservation value of paddock trees for birds in a variegated landscape in southern New South Wales. Species composition and site occupancy patterns.Biodivers Conserv. 2002; 11: 807-832Crossref Scopus (99) Google Scholar expanding the Rau et al.13Rau A.L. Wehrden H.V. Abson D.J. Temporal dynamics of ecosystem services.Ecol. Econ. 2018; 151: 122-130Crossref Scopus (30) Google Scholar proposal to 25 possible combinations of S and D dynamics, and proposing landscape management strategies most relevant to each combination to promote long-term ES provision. Our framework has five main assumptions: (1) to simplify, we initially consider that landscape structure (composition and configuration) is static, which means that we do not consider (at this point) the complexity of real landscapes changing through time; (2) we assume both S and D are fully expressed at the same spatial extent (i.e., we assume that each unit of S is equivalent to a unit of D); (3) the framework considers only one ES at a time, and so, for simplicity, we are not considering interactions among ESs (however, once the information is available, the same framework can be applied to multiple ESs, and thus ES interactions can be analyzed a posteriori); (4) tendencies for both S and D are analyzed from the same starting point where D is balanced with S even though, in reality, they might start from different points; and (5) we are only considering rival ESs by assuming that the use of a given service makes it unavailable or less available for another user.21Potts S.G. Biesmeijer J.C. Kremen C. Newman P. Schweiger O. Kunin W.E. Global pollinator decline: trends, impacts and drivers.Trends Ecol. Evol. 2010; 25: 345-353Abstract Full Text Full Text PDF PubMed Scopus (3068) Google Scholar A discussion about the limitations of such assumptions can be found in the “framework limitations” section. To access the relationship between S and D, we are using the ratio S/D (the relationship between the amount of S and D, which expresses how much one is bigger than the other). We consider these relationships between five potential trends through time: (1) stochastic behavior, where there is no predictability for S and/or D; (2) periodic behavior, in which there is a cyclical variation for S and/or D; (3) constant increasing, where S or D is growing steadily over time; (4) constant decreasing, where S or D is decreasing steadily over time; and (5) constant over time, in which S and/or D presents no change over time. Thus, we could expect 25 combinations of dynamics (Figure 1) once we have five temporal expected trends for S and five temporal trends for D, which may occur simultaneously. Therefore, adapting the proposal from Maron et al.,20Fischer J. Lindenmayer D.B. The conservation value of paddock trees for birds in a variegated landscape in southern New South Wales. Species composition and site occupancy patterns.Biodivers Conserv. 2002; 11: 807-832Crossref Scopus (99) Google Scholar we expect three outcomes of threat categories: (1) least concern, when S exceeds D (i.e., oversupply; S > D or S/D > 1; filled green spaces in Figure 1), which indicates that there is enough S to meet D and there is no risk of long-term ES provision; (2) vulnerable, when S meets D (S = D or S/D = 1; filled yellow spaces in Figure 1), which indicates that all S is being used by D, putting ES provision in a vulnerable condition in the event of small changes in S or D; and (3) endangered, if S does not meet D (i.e., undersupply or overdemand; S < D or S/D < 1; filled salmon spaces in Figure 1), which represents a situation of extinction risk of S and reducing the capability of nature to recover and meet D. A wide range of dynamics can be expected (Figure 1). For example, if both S and D vary unpredictably over time, the ratio is also expected to vary stochastically (Figure 1A) with uncertainties about when S will be in deficit or attending D. Thus, there will be moments in which S exceeds D and is not under threat (filled green spaces; least concern), while in other moments D will exceed S, putting ES provision in deficit (filled salmon spaces; endangered). A similar trend is expected when S has unpredictable behavior and when D has a constant increasing or decreasing behavior (Figures 1C and 1D). In these cases, the general trend will be led by variation in S. For instance, in a situation where S is unpredictable and D is constantly increasing, the ratio will have a decreasing behavior (Figure 1C) even if unpredictable. The opposite behavior is also expected when S is unpredictable but D presents a constant decrease (Figure 1D). In that case, even with stochasticity, ES provision may not be at risk over time since D is not exceeding S. Smooth trends with adequate S are expected only when both S and D have the same behavior (Figures 1B2, 1C3, 1D4, and 1E5), but any abrupt change in either S or D might put ES provision in debt; thus, ES is classified as vulnerable in all situations. Predictable cyclical tendencies are expected only when S has this kind of trend and D steadily increases, decreases, or remains constant over time or vice versa (Figures 1C2, 1D2, 1E2, and 1B3–1B5). The dynamics represented in Figure 1 are more illustrative than exhaustive: situations in which these responses may differ must occur; for example, when the slope and/or periodicity of S and D are different or when they have cyclical patterns over different time periods. Our framework enables us to clearly identify whether and when intervention is required in order to avoid ES extirpation, maintaining sustainable provision. Thus, in order to avoid ES undersupply, any change in the relationship should be evaluated carefully, and management actions should be taken before D exceeds S. The combinations of S and D through time might result in three main trends for ES threat (when S tends to not meet D), which allows the identification of landscapes that will require management actions in order to avoid or reduce the debt on ES provision: (1) predictable steady undersupply or overdemand, (2) predictable cyclical undersupply or overdemand, and (3) unpredictable or stochastic undersupply or overdemand. A predictable steady undersupply or overdemand trend occurs in situations when S declines and D has a constant or increasing pattern over time (Figures 1C4 and 1E4) or when S is steady and D presents an increasing pattern (Figure 1C5). This is a very common scenario around the world. For example, since 1970, trends in D for agricultural products, fish, or bioenergy production have increased, whereas 75% of contributions from nature have declined.7Constanza R. de Groot R. Sutton P. can der Ploeg S. Andeson S.J. Kubiszewski I. Farber S. Turner R.K. Changes in the global value of ecosystem services.Glob. Environ Change. 2014; 26: 152-158Crossref Scopus (2507) Google Scholar Thus, the declining capability of nature to maintain its provision indicates that gains in material contributions are often not sustainable7Constanza R. de Groot R. Sutton P. can der Ploeg S. Andeson S.J. Kubiszewski I. Farber S. Turner R.K. Changes in the global value of ecosystem services.Glob. Environ Change. 2014; 26: 152-158Crossref Scopus (2507) Google Scholar in the face of increased D of the growing population. A predictable cyclical undersupply or overdemand may occur when S presents a cyclical pattern over time and D has a steady trend (Figures 1C2, 1D2, and 1E2) or vice versa. These cyclical patterns encompass, for instance, natural oscillations such as changes in annual seasons or resource pulses following the life cycle of organisms, or even precipitation and flooding regimes, which are dependent on climate conditions. Given the consistent oscillatory nature of these fluctuations, it is possible to predict the long-term behavior. Finally, an unpredictable undersupply or overdemand situation can occur when S or D varies stochastically (Figures 1A1–1E1 and 1A2–1A5). These situations may result from environmental changes due to sudden and/or catastrophic events, such as flooding and ground slipping due to unusual storms or unpredictable changes in climatic conditions. In order to propose effective landscape strategies to decrease ES deficit across time, it is necessary to recognize in which category described above (predictable steady undersupply or overdemand, predictable cyclical undersupply or overdemand, or unpredictable or stochastic undersupply or overdemand) the related ES fits (see Boxes 2, 3, and 4 for examples of different ESs with varying temporal trends of S and D). A first general rule in order to decrease the S/D unbalance, especially regarding predictive steady or cyclical undersupply or overdemand (see Boxes 2 and 3), is to improve flow between S and D through space. In the case of unpredictable stochastic undersupply or overdemand (see Box 4), the general rule is that actions must be precautionary, much in advance of potential undersupply or overdemand, because it may be difficult to predict declining provision.Box 2Management of predictable steady ESsPollination is typically an example of an ES of predictable steady undersupply or overdemand around the world given the current pollination crisis.22Klein A.M. Vaissiere B.E. Cane J.H. Steffan-Dewenter I. Cunninghan S.A. Kremen C. Tscharntke T. Importance of pollinators in changing landscapes for world crops.Proc. Biol. Sci. 2007; 274: 303-313Crossref PubMed Scopus (3157) Google Scholar Coffee pollination is higher when nearby native habitat patches,23Ricketts T.H. Tropical forest fragments enhance pollinator activity in nearby coffee crops.Conserv Biol. 2004; 18: 1262-1271Crossref Scopus (381) Google Scholar,24Saturni F.T. Jaffé R. Metzger J.P. Landscape structure influences bee community and coffee pollination at different spatial scales.Agr Ecosyst. Environ. 2016; 235: 1-12Crossref Scopus (54) Google Scholar and pollinators presence, might increase coffee fruit set up to 28%.25Boscolo D. Candia-Gallardo C. Awade M. Metzger J.P. Importance of interhabitat gaps and stepping-stones for lesser woodcreepers (Xiphorhynchus fuscus) in the Atlantic forest.Brazil. Biotropica. 2008; 40: 273-276Crossref Scopus (75) Google Scholar Given a hypothetical landscape with the spatial distribution of S areas (sources of pollinators) interleaved with areas of D (coffee crops), we would expect ES provision to occur only in D areas where there is enough flow to connect S and D (i.e., very close to native patches; Figure 2A). Given that management strategies must consider the dispersal capacity of the species providing the service (e.g., pollinators), one low-cost management strategy to increase flow between S and D would be to establish stepping stones across the landscape, facilitating pollinators' movements (Figure 2B). The implementation of field margins, windbreaks, and hedgerows are other feasible successful alternatives to increase structural connectivity.26Prevedello J.A. Almeida-Gomes M. Lindenmayer D.B. The importance of scattered trees for biodiversity conservation: a global meta-analysis.J. Appl. Ecol. 2018; 55: 205-214Crossref Scopus (60) Google Scholar,27Baum K.A. Haynes K.F. Dillemuth F.P. Haynes K.J. The matrix enhances the effectiveness of corridors and stepping-stones.Ecology. 2004; 85: 2671-2676Crossref Scopus (199) Google Scholar Although this alternative can facilitate pollinators' movements across the landscape, its implementation alone is not enough to provide long-term nesting and feeding resources,28Roubik D.W. Ecology and Natural History of Tropical Bees. Cambridge University Press, 1992Google Scholar and consequently, it will have little impact on changing S/D trend in the long term. Creating new habitat patches for pollinators (S areas) and increasing the size of existing ones are two potentially more efficient strategies (Figure 2C). An increased amount of S areas coul