Science of the Total Environment 593–594 (2017) 599–609
Contents lists available at ScienceDirect
Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv
Distinction, quantification and mapping of potential and realized supply-demand of flow-dependent ecosystem services Romain Goldenberg a,⁎, Zahra Kalantari a, Vladimir Cvetkovic b, Ulla Mörtberg b, Brian Deal c, Georgia Destouni a a b c
Department of Physical Geography & Bolin Center for Climate Research, Stockholm University, SE-10691 Stockholm, Sweden Division of Land and Water Resources, KTH Royal Institute of Technology, SE-10044 Stockholm, Sweden Department of Urban and Regional Planning, University of Illinois at Urbana-Champaign, IL-61820 Champaign, USA
H I G H L I G H T S
G R A P H I C A L
A B S T R A C T
• Quantitative distinction of potential and realized ecosystem service supplydemand • Quantification and mapping of flows of ecosystem services through the landscape • Simple and scalable quantification and mapping methodology • Local climate regulation and storm water regulation in urban and peri-urban areas
a r t i c l e
i n f o
Article history: Received 18 October 2016 Received in revised form 14 March 2017 Accepted 14 March 2017 Available online xxxx Editor: D. Barcelo Keywords: Ecosystem services Flow dependence Potential and realized service Local climate regulation Storm water regulation Stockholm region
a b s t r a c t This study addresses and conceptualizes the possible dependence of ecosystem services on prevailing air and/or water flow processes and conditions, and particularly on the trajectories and associated spatial reach of these flows in carrying services from supply to demand areas in the landscape. The present conceptualization considers and accounts for such flow-dependence in terms of potential and actually realized service supply and demand, which may generally differ and must therefore be distinguished due to and accounting for the prevailing conditions of service carrier flows. We here concretize and quantify such flow-dependence for a specific landscape case (the Stockholm region, Sweden) and for two examples of regulating ecosystem services: local climate regulation and storm water regulation. For these service and landscape examples, we identify, quantify and map key areas of potential and realized service supply and demand, based for the former (potential) on prevailing relatively static types of landscape conditions (such as land-cover/use, soil type and demographics), and for the latter (realized) on relevant carrier air and water flows. These first-order quantification examples constitute first steps towards further development of generally needed such flow-dependence assessments for various types of ecosystem services in different landscapes over the world. © 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/).
1. Introduction ⁎ Corresponding author. E-mail address:
[email protected] (R. Goldenberg).
Ecosystem services are defined as “the direct and indirect benefits people obtain from ecosystems” with different types of such services
http://dx.doi.org/10.1016/j.scitotenv.2017.03.130 0048-9697/© 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
600
R. Goldenberg et al. / Science of the Total Environment 593–594 (2017) 599–609
distinguished: supporting services, provisioning services, regulating services and cultural services (MEA, 2005). Overall, the ecosystem service framework is an anthropocentric utilitarian concept, with the value of services provided by ecosystems depending on the utility that people derive from their consumption, either directly or indirectly (UNEP, 2011). As such, there is also increasing recognition that scientific assessments of ecosystem services need to facilitate closer stakeholder engagement (Daily et al., 2009). Regulating services are of particular importance in urban and peri-urban areas (Bolund and Hunhammar, 1999), and advancing their quantification and mapping is essential for consideration and integration in urban planning (Gómez-Baggethun and Barton, 2013; Mörtberg et al., 2017). However, most ecosystem service studies still focus primarily on the supply side, without considering the water and/or air flow processes that carry some ecosystem services through the landscape, from source areas with excess service supply to other areas with excess human needs (demands) that can be met by the service supply flow from the source areas. For example, Burkhard et al. (2012) define the supply and demand of a service over a given time period as: “the capacity of a particular area to provide a specific bundle of ecosystem goods and services” and “the sum of all ecosystem goods and services currently consumed or used in a particular area”, respectively. However, the ecosystem service supply may be linked to its human beneficiaries by a service flow that occurs in the landscape between an area of excess service supply and an area of excess service demand, with the excess service supply from the first area then possibly only reaching, and thus only being able to meet (realize), a part of the excess service demand in the latter area (Fig. 1). The total ecosystem service supply (potential supply in Fig. 1) in one area of the landscape may thus differ and needs to be distinguished from the service supply that is ultimately consumed (realized supply in Fig. 1). The service consumption may occur partly within the supply source area itself and partly in some other area, to which the service is carried, e.g., by air or water flow (see example services and their carrier flows in Fig. 1). In analogy, the total ecosystem service demand in an area (potential demand in Fig. 1) may differ and needs to be distinguished from the demand part that is actually met (realized demand in Fig. 1) by a corresponding realized supply. Some studies have considered these supply and demand links, so far mostly in conceptual terms, or for some specific ecosystem service example. In particular, Syrbe and Walz (2012) introduced the concepts of service providing area (SPA), benefiting area (SBA) and connecting area (SCA). These terms relate to the supply (SPA) and the demand (SBA) of services, with SCA then representing an area of required service flow in order to link the two. SPA and SBA can be identical, overlap, or be separated and thus in need of being linked through some type of service
flow, as illustrated here explicitly in Fig. 1. Turner et al. (2012) further classified three main service-flow models: proximal (on different scales, e.g., pollination and food production), global (e.g., global climate regulation) and slope dependent (e.g., water flow regulation). More concretely, Bagstad et al. (2013) developed a model for linking SPA and SBA, through a network of service source, sink and use regions, through which a beneficial carrier (useful service) or detrimental carrier (disservice) may travel. Serna-Chavez et al. (2014) have also proposed an indicator for the proportion of SBA that is supported through flows of services from SPA. Furthermore, the importance of flow processes for actual ecosystem service realization has been concretely quantified for the specific regulatory service example of nutrient retention by wetlands in the landscape (Quin et al., 2015). The actual trajectories of water flow through the landscape are for this example shown to largely determine the realized nutrient retention on the scale of whole catchments. In this paper, we further conceptualize, concretize and quantify flows of ecosystem services through the landscape and their key implications for distinction, quantification and mapping of main differences between potential and realized service supply and demand, with main focus on air and water carrier flows. Differences between potential and realized service supply and demand may depend considerably on such air and water flows between corresponding areas in the landscape (Fig 1). For concrete exemplification and quantification of such service flows, we consider here two examples of regulating ecosystem services (also exemplified in Fig. 1): the service of local climate regulation (carried by air flow processes), and the service of natural storm water regulation and associated flood protection (carried by water flow processes). Natural storm water regulation is then distinguished from engineered such regulation, by dams and reservoirs; in the following we will use the short term storm water regulation to mean natural such regulation. The first ecosystem service example of local climate regulation is selected in view of the impacts that ecosystems may have on local temperature (as well as on wind, radiation balance, and precipitation) through biogeophysical flow processes; for comparison, biogeochemical processes affect global climate through greenhouse gas dynamics. These biogeophysical processes may be important for avoiding local climate stress, not least due to effects of urban heat islands that may, for instance, in turn also influence human health (Rizwan et al., 2007). Furthermore, we consider the second service example of storm water regulation in view of the temporary storage of water that occurs in some ecosystems, which can reduce peak flows and mitigate high-flood events in times of intense precipitation. Such mitigation can prevent associated damages to e.g. infrastructures (Kalantari and Folkeson, 2013; Van der Sande et al., 2003) and also reduce the amount of polluted runoff from cities to nearby waterways (Pitt et al., 1995).
Fig. 1. Schematic illustration and distinction of areas of potential and realized supply and demand of an ecosystem service, and of the service flow between such areas in the landscape.
R. Goldenberg et al. / Science of the Total Environment 593–594 (2017) 599–609
These two service examples are thus both of the regulating type and are selected here because they depend on some fundamental air or water flow processes occurring in the landscape, with our aim being to investigate key implications of such service-carrying flows for distinction, quantification and mapping of potential and realized service supply and demand. In addressing this aim, we also exemplify the development of relatively simple and readily scalable methods for quantifying such service flows in a concrete landscape case. The urban and peri-urban regional landscape that includes and surrounds the Swedish capital, Stockholm, is chosen as this case because it is relatively wellinvestigated and data-rich (e.g., Destouni et al., 2013; Elmhagen et al., 2015; Quin et al., 2015; Queiroz et al., 2015; Meacham et al., 2016; Kaczorowska et al., 2016). For this landscape case and the two regulating service examples, we quantify and map the flow-dependent spatiotemporal service distributions and differences between potential and realized service supply and demand at various scales of interest over the considered regional landscape. 2. Materials and methods In the following, the first section (2.1) describes the specific regional landscape example for the present concrete quantification, and the following sections (2.2 and 2.3) outline the joint data and methods used for quantification of potential service supply and demand in both considered service examples. The service-specific quantifications of relevant flows and associated realized ecosystem services for each of the two service examples are further presented in the last section (2.4).
601
2.1. Regional case study To concretely quantify the present two examples of regulating ecosystem services, we consider here the well-investigated and relatively data-rich regional case example of Stockholm County (Sweden), with particular focus on the city of Stockholm and its surroundings (59°N, 18°E; Fig. 2a). Stockholm is the capital of Sweden and the most populous city in Sweden and Scandinavia. In 2015, approximately 923,516 people lived in the municipality and 2,231,439 lived in the county, to be compared with the total Swedish population of 9,851,017 inhabitants at the time (Statistics Sweden, http://www.scb.se). For this region, many previous studies have focused on investigating in detail its land-use, soils, climate and hydrology, as well as their interlinkages (see, e.g., Destouni et al., 2013; Destouni and Verrot, 2014; Elmhagen et al., 2015, and many other references therein) and various ecosystem services (provisioning, regulating and cultural) within the region, their quantification, spatial distribution or dependence on socio-ecological drivers, as well as urban professionals' use of the concept (Quin et al., 2015; Queiroz et al., 2015; Meacham et al., 2016; Kaczorowska et al., 2016). The present study area within this whole region is approximately 2500 km2 (including surface water area), with smaller-scale exemplification also focusing on the central urban part of Stockholm. Stockholm County is located in the boreo-nemoral mixed-forest biome (Elmhagen et al., 2015). The physical landscape mostly includes open water (both lakes and sea, approximately 23% of the total area; 532 km2 in absolute terms), coniferous (24%; 566 km2) and mixed (coniferous/deciduous; 4%; 88 km2) forests, arable land
Fig. 2. (a) Location of the study area in the region of Stockholm. Panels (b), (c) and (d) show mapped assessment results for the ecosystem service of local climate regulation in the whole region and an enlarged sub-region within it. The latter panels show: (b) the potential service demand and (c) the potential service supply over the whole region, and (d) the net remaining excess demand or supply of the service for an enlarged central part of the region. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)
602
R. Goldenberg et al. / Science of the Total Environment 593–594 (2017) 599–609
(7%; 172 km2) as well as urban green spaces (7%; 150 km2). The main types of soils in the area are bare rocks (31%; 718 km2), glacial and post-glacial layered clay (depending if these layers were deposed during or after the period of deglaciation; 19%; 440 km2), sandy moraines (12%; 279 km2), or post-glacial sand (4%; 94 km2). The Swedish Meteorological and Hydrological Institute (http://www.smhi.se) reports for this area an average annual temperature of 6.6 °C (8.7 °C in 2015) and average annual rainfall of 539 mm (656 mm in 2015) over the climate reference period 1961–1990. Overall, the greater metropolitan area of Stockholm has undergone major loss of ecosystems due to population increase and associated urban development, with green areas becoming more fragmented and degraded, and associated risk of decline in corresponding ecosystem services (Colding et al., 2003). 2.2. Data requirements and sources Data required for the present concrete quantification of the two examples of flow-dependent regulatory ecosystem services in the Stockholm area includes Swedish Land Cover Data ((SMD); obtained from the Swedish Mapping, Cadastral and Land Registration Authority (Lantmäteriet, 2013a)), containing 59 different land cover classes as classified by the Swedish Environmental Protection Agency (SEPA, 2004; http://gpt.vic-metria.nu/data/land/SMD_produktbeskrivning_ 20140627.pdf) with a resolution of 25 meters. The data further include orthophotos (Lantmäteriet, 2013b) with a resolution of 0.5 meters and a digital elevation map of the study area (DEM; Lantmäteriet, 2013c) with an original resolution of 2 meters that is rescaled to 25 meters. In addition, soil data (rescaled to 25 meters) has been obtained from the Swedish Geological Survey (SGU, 2010) and habitat classification data (classification of different forest habitats, such as deciduous and coniferous forest, also rescaled to 25 meters) has been obtained from SEPA (2004). The land cover data (SMD) represent the physical land surface for the year 2000, and have here been further merged with the habitat classification information. This merging allows for a better resolved representation of the landscape, in particular for distinguishing forested areas from urban greenbelts at a relatively fine scale. 2.3. Basic indicator methodology To quantify the potential supply and demand of the present two ecosystem service examples in the Stockholm area, we use and further develop as a basis for both service types the method introduced by Burkhard et al. (2012), of a “non-monetary evaluation scheme based on indicators which are categorized and mapped in relation to relative supply/demand scales”. In this method, an assessment matrix is developed for each ecosystem service, with indicator values for potential supply and demand of the service assigned to each grid cell of a spatially discretized landscape based on the cell classification in terms of land cover type (Fig. 3). The assigned indicator values are based on both available quantitative data and the authors' joint expert knowledge and judgment of the landscape; the latter was arrived at through discussion and consensus, based on assessing and comparing the composition and distribution of various land covers in the case study landscape. For the assessment matrix, Burkhard et al. (2012) proposed integers in the value interval 0–5 as indicator scores for different land cover classes. We have here developed this method further for use with the Swedish land cover data base (SMD) because SMD contains more land- cover classes than the CORINE land cover data base considered by Burkhard et al. (2012). Classes of urban land cover are thus more detailed, with a greater number of indicator scores, in the present assessment matrices (Fig. 3) than in those of Burkhard et al. (2012). Score values in Fig. 3 are
based on Burkhard et al. (2012) if there is direct correspondence between the SMD and the CORINE land cover classes, and are otherwise determined by the expert judgment of the authors for land cover types without such direct correspondence. Furthermore, for both service examples, the urban grid cells with lower population density (such as low density built-up areas, corresponding to single houses and farmyards, or places with less than 200 inhabitants for example) are here assigned higher scores of potential supply and lower scores of potential demand than those with higher population density (such as dense urban areas or places with few greenbelts and gardens). The different urban land cover classes from the SMD product are also assessed based on satellite ortophotos (Lantmäteriet, 2013b) with regard to their typical ratio of sealed and green-cover surfaces as basis for assigning relevant scores for their potential service supply and demand. Examples of similar methodology applied also in other case studies can be found in Vihervaara et al. (2010), Nedkov and Burkhard (2012) and Baral et al. (2012). For the ecosystem service of storm water regulation, an additional complementary assessment matrix is also developed (Fig. 4), which considers and combines infiltration capacity and topographic slope, in addition to the land-cover indicator scores in Fig. 3. For the complementary assessment matrix (Fig. 4), soils within the study area (SGU, 2010) have been classified into the following soil type classes: gravel, sand, till, peat, clay and rock, with different water infiltration scores that depend primarily on the particle sizes of each soil type. Moreover, local surface slopes have been computed based on the DEM for the investigated Stockholm area (Lantmäteriet, 2013c). The scores in the supplementary matrix in Fig. 4 are further combined with those in Fig. 3 to produce total scores for potential supply and demand of the ecosystem service of storm water regulation. This combination implies that, e.g., relatively flat forest areas with high-infiltration soils get the highest combined scores for potential supply of the service of storm water regulation. Overall, the indicator score values in the service-relevant assessment matrices (Figs. 3–4) represent an approximate quantification of potential supply and demand for each considered service. Corresponding quantification of realized service supply and demand depend in turn on the service flow that prevails between potential supply and demand areas at the landscape scale of interest (Fig. 1). This service flow further depends on the service-relevant air or water flow trajectories and processes that occur along them; these service-specific flow aspects are in the following quantified, discussed and illustrated separately for each of the two considered service examples. 2.4. Service-specific quantification of flow-dependent service supplydemand 2.4.1. Local climate regulation Local climate regulation depends on near-surface air-flow dynamics and land-atmosphere interactions, requiring elaborate, physically based atmospheric dynamics modeling for detailed process quantification and resolution. To relatively simply estimate and illustrate flow-dependent differences between potential and realized demand-supply of this service, we have spatially discretized the land surface in 150 m wide hexagon units (larger grid cells than for the land-cover classification). This discretization scale represents the estimated spatial reach of local air flow that redistributes heat and mixes local air temperatures, thereby diffusing spatial temperature differences within each such spatial unit (Vercauteren et al., 2013). It is this air mixing that supplies the service of local climate regulation, which is important for humans as well as for other organisms, for example by regulating climate-related phenological behavior (Prieto and Destouni, 2015).
Fig. 3. Assessment matrix, displaying for different land cover classes their estimated supply capacity and human demand for two considered regulating services. Legends below the main table display the interpretation of scores for potential demand and supply of each considered service: local climate regulation, and storm water regulation. Land cover classes highlighted in grey are included in the Swedish Land Cover Data (SMD; Lantmäteriet, 2013a) but not present within the Stockholm study area. The scores are based on Burkhard et al. (2012) and the authors' assessment of additional land cover classes, as explained in the main text (scores from Burkhard et al., 2012 are indicated with a star).
R. Goldenberg et al. / Science of the Total Environment 593–594 (2017) 599–609
603
604
R. Goldenberg et al. / Science of the Total Environment 593–594 (2017) 599–609
To quantify service-flow effects in this case, we first compute and map across the whole Stockholm area the scores for potential service demand (negative score values, Fig. 2b) and potential service supply (positive score values, Fig. 2c), based on the land-cover score values for each grid-cell in the landscape, as obtained from the assessment matrices in Fig. 3. Then, the net budget of potential demand and supply scores is quantified, as illustrated here in detail for the central area of Stockholm (Fig. 2d). This net budget represents the excess demandsupply that remains in each part of the landscape after the potential service supply has met the maximum possible potential demand that it can reach based on the prevailing service-flow trajectories and processes (Fig. 1). In the present, relatively simple service-flow model, the maximum flow reach of this local service supply is assumed to be the 150 m discretization scale of the hexagonal spatial units. In general, for each such spatial unit L containing n and m smaller land-cover grid-cells of net potential service supply and demand, respectively: n
Net budget in L ¼
m
∑i¼1 Ai Si þ ∑ j¼1 A j D j AL
with: Ai and Aj being the areas of grid-cells i and j, respectively Si being the score of net potential supply of element i Dj being the score of net potential demand of element j AL being the area of the hexagonal spatial unit L (here constant for all units L). If the net budget of the score values for potential demand (negative, Fig. 2b) and potential supply (positive, Fig. 2c) in a hexagonal spatial unit is near zero (yellow in Fig. 2d), these potential demand and supply values quantify also the realized demand and supply in this unit. If the net unit budget is negative (red in Fig. 2d), there is a remaining excess service demand that has not been met by the realized service supply; the latter, smaller realized service supply equals then the potential service supply, and also quantifies the realized service demand part in such spatial units (Fig. 1). Conversely, if the net unit budget is positive (green in Fig. 2d), there is remaining excess supply of the service that cannot feed into any reachable service demand; the smaller value of realized service demand equals then the potential service demand, and also quantifies the realized supply part in such spatial units. 2.4.2. Storm water regulation As for the ecosystem service of local climate regulation, also the service of storm water regulation is related to flow processes, in this case of water, which determine the balance between potential and realized service demand and supply. More of the service of storm water regulation is used locally if more local precipitation and snowmelt water infiltrates the soil and recharges groundwater. The latter moves relatively slowly to feed downslope surface water compared to if most of the incoming water instead runs off at or just below the surface and relatively quickly feeds into and flows through the surface water network. We refer to the storm water regulation service that is potentially supplied or demanded locally in this way, for just the locally generated precipitation and snowmelt water, as the local service supply and demand. In addition to this local potential service supply-demand, there is also a non-local contribution of flow from upslope areas to any point in the landscape that further determines a non-local service realization component. Overall, it is the total water flow into each landscape location - that generated locally and that contributed from upslope areas that determines how much of the potential supply of storm water regulation that is actually realized at that location. Also with regard to demand, upstream areas with high remaining local excess supply of the storm water regulation service (such as flat forested areas with highinfiltration soil) can lower the total service demand of downslope areas (such as urban areas with high local potential demand for storm water regulation) by lowering the upslope water flow contributions into the latter.
To assess the net total service demand-supply budget, and thus the total realized and the remaining excess service demand-supply, we thus also need to quantify the added water flow contributions to each point in the landscape from upslope areas. To do this, we use the DEM of the study area to determine flow directions and flow convergence into each grid cell. The flow convergence calculation assumes that one unit of water flow is generated within each grid cell, to which is further added the total sum of additional water flow units converging into that grid cell from other, upslope cells, based on the prevailing flow directions. This flow convergence calculation thus roughly estimates the relative catchment area that contributes water flow to each grid cell in the landscape, by simple number counting of all contributing upslope cells. The flow convergence calculations are performed over the whole landscape to track the spatial flow connection relations between areas with local excess of potential service supply or demand. Thereafter, areas of local excess supply and those of local excess demand are handled differently. The water flow contribution from upslope areas into receiving cells with remaining local excess supply is assumed to realize some or all of the locally remaining excess supply in the receiving cells (by infiltration and recharge of groundwater). To determine the score change due to this service realization, a flow convergence threshold is chosen, such that cells with non-local water flow input from less than 10 upslope cells (corresponding to a catchment area of 6250 m2 in the present case study) are considered as low-convergence cells, and cells with greater water-flow input (from more than 10 upslope cells) are considered as high-convergence cells. For high-convergence cells, the final modified score for their net total realized supply is obtained by increasing their local supply score by one unit. For lowconvergence cells, the local service supply score value is decreased by one unit to represent that some of the potential local service supply is not actually realized. Furthermore, the same flow convergence threshold is used for cells with local excess demand. Also, in this case, cells with non-local water flow input from less than 10 upslope excess demand cells are considered as low-convergence cells, while cells with water-flow input from more than 10 upslope excess demand cells are considered as highconvergence cells. For high-convergence cells, the modified score for their net total demand is obtained by increasing the local demand score by one unit, to represent the demand increase due to particularly high flow convergence from upstream areas into these cells. For lowconvergence cells, the local demand score is instead decreased by one unit, in order to represent that the local excess demand is only partly realized because these cells receive particularly low flow inputs from upstream areas. Also, most cells with local excess demand located downslope and next to cells with local excess supply have their flow convergence value set to zero, to represent that most of the water flow from upstream areas infiltrates and recharges groundwater in the upslope/nearby cells with excess supply. The data processing, calculations and analysis were performed in the ArcGIS (ESRI) environment. 3. Results 3.1. Local climate regulation At the overall regional scale, the urban center of Stockholm (blackline square in Fig. 2b–c) emerges as an area of particularly high potential demand (darkest red, Fig. 2b) and particularly low potential supply (lightest green, Fig. 2c), relative to other, less urbanized parts of the whole study region. The overall net budget of potential demand and potential supply of local climate regulation over the whole region (Fig. 5) also shows this central urban area as one with particularly large remaining excess demand of local climate regulation. This result is consistent with measured surface temperatures being on average about 1 °C higher in the city than in rural areas located around it (Richter et al., 2013). Within this central urban area, the relatively small potential supply is
R. Goldenberg et al. / Science of the Total Environment 593–594 (2017) 599–609
605
Fig. 4. Supplementary assessment matrix, outlining scores of potential supply of the service of storm water regulation, depending on local soil type (determining infiltration rates) and topographic slope (determining main flow trajectories and flow rates).
thus mostly equal to the realized supply, and there is a relatively large remaining excess potential demand that cannot be met by the available service supply. A star shaped spatial pattern in both potential demand (Fig. 2b) and remaining excess demand (Fig. 5) further expands outward from the central Stockholm area (square in Fig. 2b–c and Fig. 5 zoomed into in Fig. 2d) through some less urbanized surrounding parts. The latter parts combine a lower potential service demand (orange, Fig. 2b) with a still (as in the central urban part) relatively low potential service supply (lightest green, Fig. 2c), leading in combination to a remaining excess demand (red-orange parts in Fig. 5). The star-shaped pattern goes further towards and finally shifts into the conditions of the least urbanized periphery of the region. This periphery has relatively small potential demand (mostly yellow to orange, Fig. 2b) and relatively high
potential supply (darker green, Fig. 2c), leading in combination to relatively large remaining excess supply (dark green parts in Fig. 5) of the ecosystem service of local climate regulation. Zooming into the net budget of potential demand-supply in the central urban area of Stockholm (Fig. 2d) further shows that this is far from spatially homogeneous. Overall, different patterns of remaining excess service demand-supply prevail in different parts of the central urban Stockholm area. Heavily built areas have indeed high remaining excess demand even after full supply realization (darkest red grid cells, Fig. 2d), but also such areas may include within them parts with locally well balanced service demand and supply. See examples of the latter in parts close to water within ellipse i and square ii, as well as in square iii close to the major park of Humlegården, with all of these landscape parts showing well-balanced (yellow and lightest orange) spatial units
Fig. 5. Net remaining excess demand or supply of the service of local climate regulation over the whole Stockholm region. The square in the map shows the central part of the region that is enlarged in the corresponding map of net remaining service excess demand or supply in Fig. 2d. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)
606
R. Goldenberg et al. / Science of the Total Environment 593–594 (2017) 599–609
within or around larger areas of large excess demand (dark red) in Fig. 2d. With regard to the major Stockholm park Humlegården (square iii, Fig. 2d), Jansson et al. (2006) has reported measured temperatures that are indeed around 0.5 to 0.8 °C lower than in the surrounding built area during the day; this contrast tends to increase further in the afternoon and reach a maximum difference between the park and surrounding city temperatures of 2 °C at sunset. Furthermore, the results for the island of Södermalm (ellipse i, Fig. 2d), exhibiting locally wellbalanced demand-supply along its shoreline and overall excessdemand in its inner parts, are consistent with previously reported effects of local climate regulation by nearby large water bodies (Vercauteren et al., 2013). The island of Södra Djurgården (ellipse iv, Fig. 2d) is a relatively central city part that is popular for recreation, with major green areas along with historical monuments, museums and other buildings. The results for this island exhibit excess potential supply (green), provided by the surrounding water in combination with a mix of deciduous forest, wetlands and urban greenbelts in the eastern part of the island. For the mix of built and forested areas in the western part of the island, results show predominantly well balanced demand-supply of local climate regulation (mostly yellow and lightest-green). Further out towards and in the periphery of central Stockholm, results show mostly relatively well balanced service demand-supply (e.g., squares v–vi, Fig. 2d), with some main parts, predominantly close to water, exhibiting remaining excess supply (green). In general, residential areas located close to water and green areas benefit from this proximity, so that their total demand for the service of local climate
regulation may be nearly met by the available local service supply. In contrast, residential areas with the same potential demand but instead surrounded by more buildings have remaining excess demand for this service, which cannot be met by any sufficiently close-by ecosystem, even though there are several areas with excess service supply within the overall Stockholm region. 3.2. Storm water regulation Over the whole region, the net sum of local potential demand and supply scores for the service of local storm water regulation (Fig. 6a; obtained from the local assessment matrices in Figs. 3–4), as well as the overall net sum that also accounts for additional service supplydemand realization due to the water flow from upslope towards downslope areas (Fig. 7), exhibit a similar star-shaped pattern as the service of local climate regulation (Fig. 5). The central urban area of Stockholm has a high remaining local excess demand of the storm water regulation service (central dark red cells in the local – Fig. 6a – and the overall – Fig. 7 – net budget), due to its high density of buildings and associated impervious surfaces. The star shaped spatial pattern of remaining excess service demand (red and orange cells in Fig. 6a and Fig. 7) further expands outward from the central Stockholm area towards the least urbanized regional periphery of the region, which has mostly remaining excess service supply (predominantly green cells in Fig. 6a and Fig. 7). Some hotspot examples of relatively high local and final excess supply are identified in areas of gravel-sandy glaciofluvial deposits, including for example (dashed-line squares in Fig. 6a and
Fig. 6. Results for the ecosystem service of local storm water regulation for the whole region of Stockholm and an enlarged sub-region within it. The net sum of net local excess demand or supply scores is shown in (a), with some hotspots of relatively high local excess supply identified (numbers 1 to 5). Panel (b) zooms into the same local service map as in (a) for the selected study area example of Tumba, and (c) shows the overall net total remaining service supply or demand in the selected area example of Tumba after consideration of also the non-local water-flow contribution from upstream areas into each grid cell in the landscape. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)
R. Goldenberg et al. / Science of the Total Environment 593–594 (2017) 599–609
607
Fig. 7. Overall net remaining excess demand or supply of the service of storm water regulation over the whole Stockholm region, after consideration of also the non-local water-flow contribution from upstream areas into each grid cell in the landscape. The solid-line square in the map shows the Tumba part of the region that is enlarged in the corresponding map of overall net remaining service excess demand or supply in Fig. 6c. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)
corresponding parts in Fig. 7): (1) a relatively central park (Hagaparken) and cemetery (Norra Begravningsplatsen) area; (2) forested areas on the island of Ekerö in the regional periphery; (3) another forested periphery area in the municipality of Botkyrka; (4) a central cemetery area (Skogskyrkogården) in southern Stockholm; (5) relatively central forested areas in the Stockholm parts of Bollmora and Älta. In comparison with these examples, parts of central Stockholm that are also located on glaciofluvial deposits do not exhibit such local or overall excess service supply, due to their built urban land surface that limits infiltration into these deposits. Fig. 6b and Fig. 6c further show the same local service and the same overall service results as the regional maps in Fig. 6a and Fig. 7, respectively, zooming into a selected area example of Tumba, located southwest of the city of Stockholm (solid-line square in Fig. 6a and corresponding part in Fig. 7). This area includes a large forested area around an urban center and is chosen for further clarification through exemplification of the effect of the additional service supply-demand realization due to the water flow from upslope towards downslope areas in this land-cover configuration. This additional supply-demand realization can be seen in the decrease of excess service supply in the forest area (lighter green cells in Fig. 6c than in Fig. 6b) as well as in the decrease of excess service demand in a major part of the urban areas (lighter red cells in Fig. 6c than in Fig. 6b) located next to the forest area with excess service supply. However, due to their location in relation to main flow trajectories, other urban area parts exhibit instead higher excess demand in Fig. 6c than in Fig. 6b. This is due to the large water flow convergence into these urban area parts, which is not mitigated by any excess ecosystem service supply along the main flow convergence trajectories.
4. Discussion Understanding the relationship between service supply and demand in urban settings is recognized as an important issue remaining to handle in the framework of ecosystem services (Kremer et al., 2016). Relevant evaluation of ecosystem services requires assessment of the actual needs of human societies in relation to the service supply that is available and able to reach the main demand areas through the prevailing service flows. The pathways and processes of service flows from areas of potential supply to those of potential demand are essential for determining this supply-demand realization. These key flow characteristics are specific to each considered service and spatiotemporal scale of interest. Finding the relevant flow trajectories and their spatial reach, as well as the necessary and appropriate flow process representation and level of complexity in models for service evaluation is not a simple task. Yet the task still needs to be handled and this handling needs to be improved through further research contributions. In our present service and associated flow-quantification examples we have focused on two regulating services. However, also other types of services may be flow-dependent. For example, the freshwater provisioning service of a landscape depends critically on the same local water-partitioning processes (precipitation and snowmelt water infiltrating the soil and recharging groundwater) and overall flowconvergence processes (catchment-scale water flow from upstream to downstream areas) that we have also had to consider here for the service of storm water regulation. Furthermore, also the food provisioning service of a landscape depends greatly on these same local and overall water flow processes, with crop yields of both rain-fed and irrigated agriculture depending critically on prevailing water availability and
608
R. Goldenberg et al. / Science of the Total Environment 593–594 (2017) 599–609
agriculture being the major water-using sector globally. Provisioning of food and other types of biomass for various human uses depends for example also on pollination conditions, which may in turn depend on local climate regulation and thus the same air-flow processes that we have had to consider here for this regulating service. The value of recreational ecosystem services may to some degree also depend on prevailing conditions of local climate regulation and water availability in the landscape, and thus on the same air and water flow processes that determine the regulating service examples considered here. Moreover, the realization of recreational ecosystem services may not only depend on natural flows, but also on accessibility issues relating to distance/time of human transportation, i.e., on human and human-managed flows (depending for instance on transportation networks) to the service realization locations. Also provisioning service realization depends on human-managed product flows in the overall supply chain from primary production to the service realization locations. In general, the main aspect emphasized and to some degree concretized in the present study is that the actual realization of any type of ecosystem service may often depend on various types of natural flows (such as the main air and water flows exemplified and quantified here) as well as on human and/or human-managed flows (not considered here) between areas of ecosystem service supply and human beneficiaries in demand areas. Ecosystem service potential and realization thus need to be more generally distinguished, under consideration of various types of service flows and the relations that these imply between service supply and demand areas, for various types of ecosystem services in different landscapes over the world. We have here contributed some first steps towards such a general development, by conceptualizing (Fig. 1) and quantifying at first order some concrete case examples with focus on air- and water-borne service flows. In general, assessment of flow-dependent ecosystem services and their potential and realized supply and demand will require matching, consistent data and model interpretations, with similar support scales and over the same landscape parts across disciplines. However, recent studies have shown that such consistency is often lacking (Karlsson et al., 2011; Elmhagen et al., 2015). Synthesizing data and model results across different disciplines and their various data-model scale perspectives is a complex task, with comparability, transferability and accumulation of results from a wide-ranging literature on ecosystem services still being a main obstacle (de Groot et al., 2010). Nevertheless, this task must be addressed and handled in order to increase our capability to provide more meaningful and realistic quantifications, maps and valuations of ecosystem services. The present specific use of a similar, extended method to that in Burkhard et al. (2012) has been made in order to concretize and illustrate some quantitative result examples in this general direction, with the present focus being on two different types of natural servicecarrying flows, of air and of water. Through this specific method use, we have been able to provide first-order estimates of such service supply, demand and flow aspects, for the two ecosystem service examples that depend on these two types of flows. Such compatible and scalable score estimates for supply and demand constitute necessary first steps towards more complex and detailed assessments of these ecosystem service relations. The scalability of these estimates lies in that the same principal score-based estimation method can be used on various scales and spatial resolutions of relevance for different services and application cases. Estimation methods can be developed further from the present first-order one, by progressively integrating new and more complex and detailed datasets and flow models for evaluating both sides of the human-nature relations of ecosystem services and the service flows between them. For example, with specific regard to the ecosystem service of storm water regulation, potential supply scores have here been simply set to zero in the water-covered parts of the landscape. This is because it is
difficult to differentiate between sea, coastal lagoon and estuary grid cells in the highly fragmented coastal landscape of the inner archipelago area of the studied Stockholm region. In reality, potential supply and demand of the service of storm water regulation may vary among such water-covered landscape parts. Inland watercourses and water bodies, for instance, can store water and thus provide storm water regulation at some times and/or locations, while being sources of flooding at other times/locations. More complex, process-based hydrological modeling (e.g., Kalantari et al., 2014), which has been outside the scope of the present study, is then needed to realistically differentiate the potential supply and demand of this service by various water bodies in various parts of the landscape and at different times. In general, awareness of and openness with main knowledge limitations and assumptions made for ecosystem service quantification, mapping and visualization are necessary and beneficial for appropriate consideration and use of the ecosystem service concept in real applications. Not least a focus on only the potential supply of ecosystem services, without understanding its difference from and the complexities involved in the actual realization of such supply, can mislead answers to important questions about ecosystem service benefits and beneficiaries. Such questions may, for instance, be: Who can benefit from which ecosystem services? Where should homes, activities and businesses be located in the landscape in order to be able to benefit from various ecosystem services? For urban development planning, the understanding and relevant quantification of the service supply-demand realization and its spatial dependence on service flows can be essential for scenario assessment in such applications; specifically, for making the best possible use of available potential ecosystem-service supply in meeting the needs of various possible beneficiaries over the landscape. For example, a set of hypothetical, yet realistic, scenarios of future urban development trajectories may be formulated and evaluated, showing different implications dependent on various scenario factors, such as density, extent and location of new built environments, representing main areas of increased potential service demand and decreased potential service supply, and their spatial flow-dependent relation to remaining natural areas of relatively large potential service supply. The present quantification and mapping approaches and results for the regional landscape of Stockholm exemplify and concretize key differences between spatial distribution patterns of potential and realized ecosystem service demand and supply, which may be essential for planning of further urbanization developments in this as in other landscapes over the world. The assessment steps and exemplifications in this paper have provided some relatively simple but still realistic estimates of service supplydemand situations in one urban and peri-urban region example. In general, such distinction, quantification and mapping of the spatial pattern of realized service supply-demand are needed to assess and evaluate the usefulness of ecosystem services over various landscape scales and development scenarios.
5. Conclusion The present approach to service-flow quantification, combined with look-up assessment matrices for local service conditions, allows for relatively simple quantification and mapping of the spatial distribution of potential and realized supply-demand of ecosystem services over the landscape. Thereby, areas of relatively good or bad spatial service matching can be identified, and scenarios of future land-use changes can be assessed with regard to their consequences for service potential and realization. The present results also call for more research to improve the service-flow quantifications and the associated distinctions between potential and realized ecosystem services. Many challenges remain for this research, which must handle a wide-ranging problem complexity across different regions, scales and ecosystem services of interest to various stakeholders.
R. Goldenberg et al. / Science of the Total Environment 593–594 (2017) 599–609
Acknowledgments We gratefully acknowledge the data provided by Lantmäteriet (Swedish mapping, cadastral and land registration authority) and Naturvårdsverket (Swedish Environmental Protection Agency). We also thank Jan Haas from the KTH department of Geodesy and Geoinformatics for initial help with the data. We further gratefully acknowledge funding for this research from the strategic program Ekoklim at Stockholm University and from the Formas funded project (grant 2014-754). References Bagstad, K.J., Johnson, G.W., Voigt, B., Villa, F., 2013. Spatial dynamics of ecosytem service flows: a comprehensive approach to quantifying actual services. Ecosystem Services 4 (2013), 117–125. Baral, H., Keenan, R.J., Fox, J.C., Stork, N.E., Kasel, S., 2012. Spatial assessment of ecosystem goods and services in complex production landscapes: a case study from southeastern Australia. Ecol. Complex. 13, 35–45. Bolund, P., Hunhammar, S., 1999. Ecosystem services in urban areas. Ecol. Econ. 29, 293–301. Burkhard, B., Kroll, F., Nedkov, S., Müller, F., 2012. Mapping ecosystem service supply, demand and budgets. Ecol. Indic. 21, 17–29. Colding, J., Elmqvist, T., Lundberg, J., Ahrné, K., Andersson, E., Barthel, S., Borgström, S., Duit, A., Ernstsson, H., Tengö, M., 2003. The Stockholm Urban Assessment (SUASweden) (Millenium ecosystem assessment sub-global report). Daily, G.C., Polasky, S., Goldstein, J., Kareiva, P.M., Mooney, H.A., Pejchar, L., Ricketts, T.H., Salzman, J., Shallenberger, R., 2009. Ecosystem services in decision making: time to deliver. Front. Ecol. Environ. 7, 21–28. De Groot, R.S., Alkemade, R., Braat, L., Hein, L., Willemen, L., 2010. Challenges in integrating the concept of ecosystem services and values in landscape planning, management and decision making. Ecol. Complex. 7 (3), 260–272. Destouni, G., Verrot, L., 2014. Screening long-term variability and change of soil moisture in a changing climate. J. Hydrol. 516, 131–139. Destouni, G., Jaramillo, F., Prieto, C., 2013. Hydroclimatic shifts driven by human water use for food and energy production. Nat. Clim. Chang. 3, 213–217. Elmhagen, B., Destouni, G., Angerbjörn, A., Borgström, S., Boyd, E., Cousins, S.A.O., Dalén, L., Ehrlén, J., Ermold, M., Hambäck, P.A., Hedlund, J., Hylander, K., Jaramillo, F., Lagerholm, V.K., Lyon, S.W., Moor, H., Nykvist, B., Pasanen-Mortensen, M., Plue, J., Prieto, C., van der Velde, Y., Lindborg, R., 2015. Interacting effects of change in climate, human population, land use and water use on biodiversity and ecosystem services. Ecol. Soc. 20 (1), 23. Gómez-Baggethun, E., Barton, D.N., 2013. Classifying and valuing ecosystem services for urban planning. Ecol. Econ. 86, 235–245. Jansson, C., Jansson, P.E., Gustafsson, D., 2006. Near surface climate in an urban vegetated park and its surroundings. Theor. Appl. Climatol. 89 (3), 185–193. Kaczorowska, A., Kain, J.H., Kronenberg, J., Haase, D., 2016. Ecosystem services in urban land use planning: integration challenges in complex urban settings—case of Stockholm. Ecosyst. Serv. 22, 204–212 Part A. Kalantari, Z., Folkeson, L., 2013. Road drainage in Sweden: current practice and suggestions for adaptation to climate change. J. Infrastruct. Syst. 10, 1061. Kalantari, Z., Briel, A., Lyon, S.W., Olofsson, B., Folkeson, L., 2014. On the utilization of hydrological modelling for road drainage design under climate and land use change. Sci. Total Environ. 475, 97–103.
609
Karlsson, J.M., Bring, A., Peterson, G.D., Gordon, L.J., Destouni, G., 2011. Opportunities and limitations to detect climate-related regime shifts in inland Arctic ecosystems through eco-hydrological monitoring. Environ. Res. Lett. 6 (1), 014015. Kremer, P., Hamstead, Z., Haase, D., McPhearson, T., Frantzeskaki, N., Andersson, E., Kabisch, N., Larondelle, N., Lorance Rall, E., Voigt, A., Baró, F., Bertram, C., GómezBaggethun, E., Hansen, R., Kaczorowska, A., Kain, J.H., Kronenberg, J., Langemeyer, J., Pauleit, S., Rehdanz, K., Schewenius, M., van Ham, C., Wurster, D., Elmqvist, T., 2016. Key insights for the future of urban ecosystem services research. Ecol. Soc. 21 (2), 29. Lantmäteriet, 2013a. GSD Swedish Land Cover Data. © Lantmäteriet [i2012/920]. Lantmäteriet, 2013b. GSD Ortofoto © Lantmäteriet [i2012/920]. Lantmäteriet, 2013c. GSD Elevation Data © Lantmäteriet [i2012/920]. MEA, Millennium Ecosystem Assessment, 2005. Ecosystems and Human Well-being: Synthesis. Island Press, Washington DC. Meacham, M., Queiroz, C., Norström, A.V., Peterson, G.D., 2016. Social-ecological drivers of multiple ecosystem services: what variables explain patterns of ecosystem services across the Norrström drainage basin? Ecol. Soc. 21 (1), 14. Mörtberg, U., Goldenberg, R., Kalantari, Z., Kordas, O., Deal, B., Balfors, B., Cvetkovic, V., 2017. Integrating ecosystem services in the assessment of urban energy trajectories — a study of the Stockholm region. Energ Policy 100, 338–349. Nedkov, S., Burkhard, B., 2012. Flood regulating ecosystem services — mapping supply and demand, in the Etropole municipality, Bulgaria. Ecol. Indic. 21, 67–79. Pitt, R., Field, R., Lalor, M., Brown, M., 1995. Urban stormwater toxic pollutants: assessment, sources, and treatability. Water Environment Research 67 (3), 260–275 (16). Prieto, C., Destouni, G., 2015. Climate-driven, phenological change: developing robust spatiotemporal modeling and projection capability. PLoS ONE 10 (11), e0141207. Queiroz, C., Meacham, M., Richter, K., Norström, A.V., Andersson, E., Norberg, J., Peterson, G., 2015. Mapping bundles of ecosystem services reveals distinct types of multifunctionality within a Swedish landscape. Ambio 44, 89–101. Quin, A., Jaramillo, F., Destouni, G., 2015. Dissecting the ecosystem service of large-scale pollutant retention: the role of wetlands and other landscape features. Ambio 44 (1), 127–137. Richter, M., Deppisch, S., Von Storch, H., 2013. Observed changes in long-term climatic conditions and inner-regional differences in urban regions of the Baltic Sea Coast. Atmospheric and Climate Sciences 3, 165–176. Rizwan, A.M., Dennis, Y.C.L., Liu, C., 2007. A review on the generation, determination and mitigation of Urban Heat Island. J. Environ. Sci. 20, 120–128. SEPA, 2004. Kontinuerlig Naturtypskartering [Continuous Habitat Type Mapping]. Swedish Environmental Protection Agency, Bromma. Serna-Chavez, H.M., Schulp, C.J.E., Van Bodegom, P.M., Bouten, W., Verburg, P.H., Davidson, M.D., 2014. A quantitative framework for assessing spatial flows of ecosystem services. Ecol. Indic. 39 (2014), 24–33. SGU, 2010. Geological Map. Swedish Geological Survey. Syrbe, R.U., Walz, U., 2012. Spatial indicators for the assessment of ecosystem services: providing, benefiting and connecting areas and landscape metrics. Ecol. Indic. 21 (2012), 80–88. Turner, W.R., Brandon, K., Brooks, T.M., Gascon, C., Gibbs, H.K., Lawrence, H.K., Mittermeier, R.A., Selig, E.R., 2012. Global biodiversity conservation and the alleviation of poverty. Bioscience 62, 85–92. UNEP, 2011. Towards a Green Economy: Pathways to Sustainable Development and Poverty Eradication — A Synthesis for Policy Makers. www.unep.org/greeneconomy. Van der Sande, C.J., De Jong, S.M., De Roo, A.P.J., 2003. A segmentation and classification approach of IKONOS-2 imagery for land cover mapping to assist flood risk and flood damage assessment. Int. J. Appl. Earth Obs. Geoinf. 4, 217–229. Vercauteren, N., Lyon, S.W., Destouni, G., 2013. Seasonal influence of insolation on fineresolved air temperature variation and snowmelt. J. Appl. Meteorol. Climatol. 53, 323–332. Vihervaara, P., Kumpula, T., Tanskanen, A., Burkhard, B., 2010. Ecosystem services — a tool for sustainable management of human–environment systems. Case study Finnish Forest Lapland. Ecol. Complex. 7, 410–420.
