A multi-criteria, ecosystem-service value method used to assess catchment suitability for potential wetland reconstruction in Denmark

Ecological Indicators 77 (2017) 151–165

Contents lists available at ScienceDirect

Ecological Indicators journal homepage: www.elsevier.com/locate/ecolind

Original Articles

A multi-criteria, ecosystem-service value method used to assess catchment suitability for potential wetland reconstruction in Denmark Mette Vestergaard Odgaard a,∗ , Katrine Grace Turner a,b , Peder K. Bøcher b , Jens-Christian Svenning b , Tommy Dalgaard a a b

Aarhus University, Foulum, Department of Agroecology, Blichers Allé 20, P.O. Box 50, DK-8830 Tjele, Denmark Aarhus University, Department of Bioscience, Ny Munkegade 116, DK-8000 Aarhus C, Denmark

a r t i c l e

i n f o

Article history: Received 23 June 2016 Received in revised form 30 November 2016 Accepted 1 December 2016 Keywords: Catchment screening Reconstructed wetlands Restored wetlands Ecosystem services Hotspot analysis Multi-criteria Scenario mapping

a b s t r a c t Wetlands provide a range of ecosystem services such as drought resistance, flood resistance, nutrient deposition, biodiversity, etc. This study presents a new multi-criteria, ecosystems service value-driven method to drive the optimal placement of restored wetlands in terms of maximizing selected ecosystem services which a wetland can provide or affect. We aim to answer two questions: 1) which of the ecosystem services indicators defines the placement of wetlands today? 2) Based on the ecosystem services indicator assessment, what are the recommendations for future selection of catchments for potential wetland reconstruction (i.e. restoration)? Five key ecosystem services indicators produced or affected by wetlands in Denmark were mapped (recreational potential, biodiversity, nitrogen mitigation potential, inverse land rent, and flash-flood risk). These services were compared to current placements of wetlands. Furthermore, scenario testing and hotspot analysis were combined to provide future recommendations for optimal placements of wetlands. The scenarios investigated were Climate Adaptation and Protection of Aquatic Environment, Land-Based Economy, and Rich Nature. Based on these scenarios, the most suitable areas for wetland reconstruction were mapped, taking both the scenarios and attached weightings of ecosystem services indicators into account. According to statistical results current reconstructed wetlands are situated in catchments with lower biodiversity, higher nitrogen mitigation potential, higher land rent (i.e. agricultural intensive areas), and to some extent higher flash flood risk compared to the median of catchments with wetlands. Hence, recreation potential, high biodiversity, and low land rent has not been prioritized. 35 out of the 3023 catchments investigated were identified with an especially high suitability when optimizing all scenarios. This coincides with a high suitability around peri-urban and urban areas and near natural areas, hence capturing both supply and demand services. Of the 35 identified catchments with potentially high suitability, only 2 actually hold a presently reconstructed wetland. This indicates a prior placement with almost no consideration of maximizing ecosystem services benefits. We recommend a systematic approach, such as the ecosystem service value-driven method demonstrated in the present case study, to target more services and improve the overall benefit from wetlands. This approach seeks to inform decision makers of synergies in the landscape, which is likely to transcend future policy implementations. © 2016 Elsevier Ltd. All rights reserved.

1. Introduction

∗ Corresponding author. E-mail addresses: [email protected] (M.V. Odgaard), [email protected] (K.G. Turner), [email protected] (P.K. Bøcher), [email protected] (J.-C. Svenning), [email protected] (T. Dalgaard). http://dx.doi.org/10.1016/j.ecolind.2016.12.001 1470-160X/© 2016 Elsevier Ltd. All rights reserved.

Ecosystem services are the flow of natural capital to society, such as wood, fiber, food, nutrient cycling, and drinking water (MEA, 2005). Hence, wherever humans live, complex socio-ecological interactions are formed with the surrounding ecologic landscape, affecting and directing the flow of ecosystem services. These interactions can differ between regions and societies depending on local

152

M.V. Odgaard et al. / Ecological Indicators 77 (2017) 151–165

characteristics (MEA, 2005; Costanza et al., 2014). The value of wetland ecosystem services are among the highest of any ecosystem assessed (MEA, 2005; Russi et al., 2013), and wetlands have played a long and important role throughout history to secure human well-being (MEA, 2005; Russi et al., 2013), as well as the social and cultural evolution of humankind (Barbier, 2011; Maltby and Acreman, 2011). Therefore, wetlands might serve as a considerable part of the solution to some of the current and future global-scale problems, e.g. in relation to biodiversity loss, climate change effects, and nutrient displacement (Harrington et al., 2011; Steffen et al., 2015). Wetlands both store excess water and function as a drought resistance as they delay the discharge of water (Barbier et al., 1997). Furthermore, they clean the water and promote deposition of environmentally damaging substances, excessive amounts of nutrients, and carbon (Kayranli et al., 2010; Adhikari et al., 2011). Studies have suggested that especially the increase in droughts, due to global climate changes, is threatening the persistence of wetland ecosystems in many areas around the globe (Okruszko et al., 2011; Pachauri et al., 2014). For Central Europe, a possible 26–46% loss of these ecosystems because of water scarcity and hydrological changes has been estimated (Okruszko et al., 2011). In northern Europe, however, it is projected that there will be more precipitation in the future and thus more surface water and extreme rainfall events (Frei et al., 2006; Brander et al., 2012), potentially increasing the abundance of wetlands. The year 2015 was the original year set by the Water Framework Directive (WFD) (Directive 2000/60/EC) where all the surface and ground water in every European Union Member State, including Denmark, were to reach ‘good’ conditions. This policy represents a large scale attempt to manage water resources across political borders and change the way natural resources are managed (Chave, 2001). The WFD was implemented in 2000 to improve the water quality in designated water bodies across the member states. This included groundwater (enhancing the chemical and quantity status) as well as surface water (enhancing the ecological status and the chemical/nutrient status). Subsequent legislation in Denmark is based on implementing the objectives of the WFD (Dalgaard et al., 2014). These policies are simultaneously directed at improving the water environment among others by using restoration or reconstruction of water ways as a method to reduce nitrate pollution in surface waters (Dalgaard et al., 2014). In Denmark, wetlands have been drained for centuries to increase the area of agricultural land, leading to a 70% reduction in natural wetland areas as compared to the situation in 1800 C.E. (Larsson, 2004). This has severely lowered wetlands’ ecosystem service potential. With future climate changes more extreme weather conditions are expected in Denmark (e.g. more intensive rainfall and storms), potentially increasing material damages to both public and private holdings (Grøndahl et al., 2014). Furthermore, Denmark has set a target to reduce the pressure of active N pollution from human activities, stop the loss of biodiversity and the general decline in nature quality, connectivity, and heterogeneity (Agger et al., 2012). Until now, most policies on the restoration of wetlands have been directed by the WFD to mitigate nitrogen and phosphorus output (Natur og Landbrugskommisionen, 2013). Consequently, the current selection of potential sites for reconstructing (i.e. restoring) wetlands in Denmark has primarily been driven by the leaching of especially nitrate, and the nutrient loss reduction cost-effectiveness of wetland projects and this most likely dominate the placement of these wetlands (Jacobsen, 2012; Danish Ministry of Food, agriculture and Fisheries, 2015). Hence, the placement of wetlands has probably not taken other potential wetland ecosystem services into account. The potential of reconstructed wetlands to increase biodiversity and offer flood protection and recreational value etc. has therefore probably been undervalued.

Therefore, there is a need to increase the awareness and improve methods to secure the high potential of wetlands. This includes long term perspectives, such as mitigation of climate change projections of increased flash floods and biodiversity protection, as well as the possibility of recreational values in the newly reconstructed wetlands. With management of natural resources follows an inevitable trade-off between various types of land use (Turner et al., 2015). A repeatedly reported trade-off occurs between agriculture and natural areas (Raudsepp-Hearne et al., 2010; Bai et al., 2011; Dick et al., 2014; Turner et al., 2014). Earlier research has shown that the present distributions of ecosystem services in Denmark are non-random and appear in distinct groups (clusters) in the Danish landscape (Turner et al., 2014). These are driven by policy and land-use gradients between agriculture- and forestdominated landscapes, in a distinct east-west gradient broadly reflecting increasing demands for cultural and recreational services in the urbanized and densely populated eastern parts, and sociobiophysical drivers of agriculture and natural wetlands’ regulating services in the west (Turner et al., 2014). Sustainable management of trade-offs becomes more key as the value of wetlands increase with human population and anthropogenic pressures on the landscape (Ghermandi et al., 2010). To obtain long-term sustainable land management, policy must encompass the economic, social, and environmental aspects that reflect the conditions in a local geographical context (Cowling et al., 2008). Therefore, it is not possible to generate a single grand scheme for selecting optimal sites for reconstructing wetlands (Harrington et al., 2011), as the optimal situation depends of the weighting of these criteria. Consequently, it is of high priority to develop methods to ensure sustainable management while embracing most interests. Here, we combine two methods to assess catchment suitability for wetland reconstruction; future scenarios and hotspot analysis: • Scenario testing is a common used approach which typically describes the outcome of various policy scenarios, both within the field of nature conservation (Melbourne-Thomas et al., 2011; Okruszko et al., 2011) and in socio-ecology (Willemen et al., 2010; Swetnam et al., 2011; Jarchow et al., 2012; Whitfield and Reed, 2012; Bateman et al., 2013). • Hotspot analyses, on the other hand, is often used to identify hotspots of e.g. threatened species (Grenyer et al., 2006), appropriate areas for targeted conservation (Naidoo et al., 2008; Greve et al., 2013), or to detect areas providing most ecosystem services (Egoh et al., 2008; Raymond et al., 2009; Bai et al., 2011; Fisher et al., 2011). Within the field of socio-ecology, both future scenarios and hotspot analyses have been widely used to describe changes in ecosystem services in for instance management dynamics, but the combination of the two has to our knowledge not been used previously. 1.1. Aim of study This paper will analyze the potential of using ecosystem services indicators, defined as indicators of ecosystem services which a wetland can provide or affect, to locate the related, most suitable catchments for reconstructing wetlands in the Danish landscape. A suitable area is hypothesized to be characterized by:1) low recreation potential, 2) high biodiversity, 3) high nitrogen mitigation potential, 4) low land rent, and 5) high risk of flooding during extreme weather events. We hypothesize that until now the placement of reconstructed wetlands is spatially determined by the funding from policies of nitrogen mitigation in agriculture, and

M.V. Odgaard et al. / Ecological Indicators 77 (2017) 151–165

that other services therefore are underrepresented. Policy scenarios based on different assemblies of ecosystem services will result in differentiated geography (Bateman et al., 2013) and we will use this to locate the optimal placements of reconstructed wetlands. Therefore, we included three different policy scenarios, inspired by the project Future Farming in Denmark (Jørgensen et al., 2015), which represents the three different policy dependent futures in Denmark, focusing on: i) Climate Adaptation and Protection of Aquatic Environment, ii) Land Based Economy, or iii) Rich Nature. By analyzing the potential value of the ecosystem services that are affected (positively or negatively), the analysis will choose hotspots for Danish wetlands based on the system’s ecosystem service profiles. We based the analyses on the following questions: 1) Which of the following selected ecosystem service indicators are priorities when selecting present wetland reconstruction sites: recreation potential, biodiversity, nitrogen mitigation, land rent, and flash flood risk? 2) What are the recommendations for selection of catchments for future reconstruction of wetlands if we were to base it on a wider array of ecosystem services provided or affected by wetlands and policy scenarios and not nitrogen assimilation alone? 2. Materials and methods 2.1. Study area We chose Denmark (43075 km2 ) as a study region (Fig. 1). This European lowland region is dominated by cultural landscapes, with agriculture as the predominant land use, occupying 66% of the land area on average, of which 80% is used as arable land (Eurostat, 2010). The population (5.6 million) has a high Human Development Index of 0.895, 16th among nations (UN Development programme (UNDP, 2011), and approximately 349 billion USD in GNP (Statistics Denmark). The relatively high land use and population pressure makes it an interesting case study because this situation will increase globally in the future according to current projections of such factors (MEA, 2005; Carpenter et al., 2009; Ellis et al., 2010; Pereira et al., 2010; Moore et al., 2012). The soils are dominantly sandy in the west and loamy in the east (Fig. 2). Average annual precipitation is 712 mm, where the western parts receive approximately 900 mm p.a. and the eastern part approximately 600 mm p.a. (Danish Meteorological Institute, 2011; Statistics Denmark). Danish mainland climate is coastal temperate with relatively mild winters and cool summers with a mean temperature of 0.0 ◦ C and 15.6 ◦ C, respectively. 2.2. Data Denmark has been divided into 3124 so-called ID15 water catchments (average area of approx. 15 ha), which act as a central unit for the future Danish aquatic environment policies (Højberg et al., 2015) (Fig. 3). The ID15 catchments scale is based on a natural landscape boundary, which also acts as an administrative border. This limits uncertainties based on modifiable area unit problems, which can create bias with an ad hoc placement of a boundary (Jelinski and Wu, 1996). Hence, we aim this study at national and regional policy makers and have identified the ID15 catchment scale to be of sufficient resolution. Data on five ecosystem service indicators was collected for each Id15 catchment (recreational potential, biodiversity, nitrogen mitigation potential, inverse land rent, and flash flood risk). To have full ecosystem service data coverage the islands Laesø, Samsø, Bornholm, as well as smaller islands were excluded from the final dataset resulting in a final dataset of 3023 catchments. All ecosystem service data was aggregated to

153

fit the ID15 sub-catchments. Furthermore, all data was normalized [(x − x(min) /x(max) − x(min) )] (Fig. 4). 2.2.1. Reconstructed wetlands Data on wetland projects was derived from the Danish Ministry of Food, Agriculture and Fisheries (2014). This is an online-available collection of all private, municipal, and state administered wetland projects that received funding from the Danish State Rural Development program from 2007 to 2013. All catchments containing an established or planned site for reconstructing wetlands were assigned the value 1 and others the value 0 (Fig. 3). 2.2.2. Ecosystem services indicators In this study it is important to note that we do not map actual ecosystem services, but indicators of where there is a demand for reconstructing a wetland based on selected ecosystem services a wetland can provide or affect. More specifically, a wetland can yield higher recreational value, improved or strengthening of already existing biodiversity (for a discussion on biodiversity as an ecosystem service see: Mace et al., 2012), nitrogen mitigation through denitrification, and prevent flooding in case of flash floods. Moreover, constructing a wetland on high value agricultural land would influence the supply of high agricultural yields. Therefore, the indicators include in this study describe a demand for a wetland based on the potential ecosystem services a wetland provide or affect. Data on five different types of ecosystem service indicators produced or affected by the reconstruction of a wetland were included in this analysis (recreational potential Rec Pot, biodiversity Biodiv, nitrogen mitigation potential N Mit, inverse land rent LRent Inv, and flash flood risk FF Risk). The suitability of a catchment for a wetland should increase with a high Rec Pot, a high Biodiv (to enhance already existing biodiversity), high N Mit, high LRent Inv (note, inversed land rent), and high FF Risk. Most reconstructed wetlands of sizes larger than 500m2 are likely established in agricultural areas in Denmark. We therefore included data on the agricultural land rent for all fields in the catchments. 2.2.2.1. Recreational potential (Rec Pot). The indicator for recreational value in the catchments was obtained from the report “Recreation Values of Nature in Denmark”, for the Danish Economic Council (Bjørner et al., 2014). The data represents the Inclusive Value (IV00 ), an estimate utility value derived from a random utility maximization model, of the collected choice (quality and distance) sets of recreational areas to a resident of a given area (Parsons, 2003). Therefore, this data values the proximity and quality of nature recreation for the resident in the given catchment (Bjørner et al., 2014), and thus is a measure of the substitutability between sites of recreation in the given area (Parsons, 2003). We adjusted IV00 for population density with population data from Statistics Denmark (Statistics Denmark) [IV00 (Adjust) = IV00 /population density]. The spatial resolution of the original data was 1 km2 . In the aggregation of the 1 × 1 km2 data and the catchments we accounted for edge effects of the data overlay by calculating the area ratio [Area ratio = area[new] /area[old] ] where area(old) is the area of the 1 × 1 km grid cells and area(new) is the area of each polygon after the overlay (GIS-tool UNION) (Fig. 5). This method eliminates multiple counting of the same [IV00 (Adjust) − value] in the dataset. Finally, the [IV00 (Adjust) × Area ratio] were summed for each catchment and adjusted for catchment area. Recreation should in this research be calculated as a demand, that is the recreational potential, therefore, the inverse value of the normalized IV00 (Adjust) was calculated [Rec Pot = 1–IV00 (Adjust) ]. Rec Pot = 1 indicates a catchment with low recreation and thus have a high recreational potential and vice versa for Rec Pot = 0.

154

M.V. Odgaard et al. / Ecological Indicators 77 (2017) 151–165

Fig. 1. The study area, Denmark.

Table 1 The present prioritization of reconstructed wetlands placement in Denmark, i.e., due to current and past policies. Ecosystem-service-indicator values are the median of the normalized values [(x − x(min) /x(max) − x(min) )] in all catchments without reconstructed wetlands and catchments with reconstructed wetlands. The random value depicts the median of 350 random values ranging from 0 to 1. Catchment

n

Median of normalized values (0–1) Rec Pot

Without RCW With RCW Random

2673 350 350

a

0.9994 0.9996bB 0.487D

Biodiv

N Mit a

0.0621 0.0485bB 0.5219D

LRent Inv a

0.2230 0.2384bB 0.4211D

a

0.7515 0.7384bB 0.4588D

FF Risk 0.0209a 0.0222bB 0.4932D

a, b, and B and D indicate statistically significant difference in column numbers (p < 0.005). RCW is reconstructed wetlands.

2.2.2.2. Biodiversity (Biodiv). Data on biodiversity was obtained from the national biodiversity map (Ejrnæs et al., 2014). This represents a bioscore index, two indicators combining a species score and a proxy score map for the state of nature in the location. The species score is based on 537 endangered species’ distribution maps (score between 1 and 9), whereof approximately 40% of the species are wetland dependent (e.g. species of beetles, dragon flies, butterfly, amphibians, birds, vascular plants, etc.) (Ejrnæs et al., 2014). The proxy score is based on a range of species and habitat indicators and

13 significant proxy species indicators (score between 1 and 11). The two indicators are summarized to a bioscore index between 0 and 20 (Ejrnæs et al., 2014). The spatial resolution was 9.6-m. All intensively cultivated fields have been given a score of 0 by default (Levin et al., 2012; Ejrnæs et al., 2014), and since wetlands are most likely to be placed on agricultural fields there will rarely be a significant trade-off between biodiversity and wetland reconstruction. For each ID15 catchment the mean bioscore was calculated, meaning that much local variation was lost. The logic behind this was

M.V. Odgaard et al. / Ecological Indicators 77 (2017) 151–165

155

Fig. 2. Soil conditions (Greve et al., 2007) and the major urban areas in Denmark (Kort10; Danish Geodata Agency, 2014).

that if reconstructed wetlands are established on intensely managed fields also within landscapes with many natural areas, a high biodiversity score for the overall landscape would be preferable, since the wetland then would increase the connectivity between patches of nature in the landscape. Thus the biodiversity indicator is a proxy for biodiversity in the greater landscape. It could be argued that increased connectivity in areas with a relatively low biodiversity should be preferable to enhance biotic movement across areas of low biotic suitability. In this study we choose not to include connectivity as it would add a lot of extra complexity and uncertainty. Notably, due to the highly variable dispersal capacity among species and organism groups which benefit from increased connectivity, e.g., waterfowl vs amphibians vs grassland plants. While this issue is in fact often considered for waterfowl (Beatty et al., 2014; Merken et al., 2015; Petit et al., 1995), which have high dispersal capacity, most biodiversity is actually characterized by a much lower dispersal capacity (e.g., the other groups mentioned above). Additionally, the effectiveness of corridors in conservation planning is controversial due to doubts about the actual use of the corridors due to pour study design (Harsh, 2015), and spread of “unwanted guests” such as diseases, invasive species, predators etc. (Ogden 2015). Hence, from an overall biodiversity perspective it is justifiable to ignore connectivity and thereby focus on localized benefits to biodiversity. 2.2.2.3. Nitrogen mitigation potential (N Mit). In Denmark, leaching of nitrogen (N) is one of the most important drivers for eutrophication of surface waters, and the denitrification of N in wetlands is

an important measure to mitigate the effect of this severe pollution (Kronvang et al., 2008). For the present study we use the previous modelled N leaching and N retention (calculated for all Id15 catchments) to calculate coastal N load: 1) N leaching (ton/ha/year) is the estimated total N leaching from the root zone and is based on data from 2001 to 2011. It includes added manure and artificial fertilizer, N fixating, percolation, soil type, crop rotation, and agricultural crop cover during summer and winter (Kristensen et al., 2008) and 2) N retention (%) estimates how much of the leached N is retained or removed by denitrification in the pathway from the root zone to the coast. It includes climatic data, surface and groundwater flow information as well as redox potential in the soil (Højberg et al., 2015). Here, the costal N load was used as a rough proxy indicator for in which watersheds new reconstructed wetlands may yield the highest mitigation effect (N Mit): [(N leaching × ((100-N retention)/100))/farm area]. Hence, catchments with high costal N loads yield high N Mit. 2.2.2.4. Inverse land rent (LRent Inv). Land Rent indicates the economic value of cultivation in an area. It is calculated by subtracting the production cost from the crop market price. Production costs include seeds, fertilizer, irrigation water, chemicals, labor hours, and the depreciation and return on machines and equipment. With all the variable costs in the production, this indicator is a measure of how much the crop output contributes to the farm income. This is the price a farmer should ask if he is to rent out the land to cover his costs, or the lost production value if the land is taken out of

156

M.V. Odgaard et al. / Ecological Indicators 77 (2017) 151–165

Fig. 3. Map of the ID15 catchments intersected with the distribution of reconstructed wetlands. (i.e. wetland projects that received funding from the Danish State Rural Development program from 2007 to 2013; Danish Ministry of Food, Agriculture and Fisheries 2014). On the left is highlighted all the catchments with a reconstructed wetland.

production (Olsen et al., 2014). Data of land rent on field polygon level for the year 2011 (Olsen et al., 2014) was extracted for all potential areas for wetland recunstruction (i.e. restored wetlands), based on the areas mapped with organic soils at 250 m resolution (indicating that they may have earlier been water logged wetlands) (Greve et al., 2014). In the aggregation of the field polygons to the ID15 catchments we accounted for edge effects of the data overlay using the same method as described in the calculation of Rec Pot (Fig. 5). Land rent for all areas with organic soils were summarized for the ID15 catchment, and adjusted for agricultural area. Land rent measures the actual economic output of the field and it is thereby an measure of the supply of ecosystem services in the catchment. Reconstruction of a wetland in an area with a high land rent could therefore cause a tradeoff between services such as economic output (land rent) and other services such as biodiversity. Therefore, we calculated the inverse value of the original land rent data [LRent Inv = (1–Land Rent)]. LRent Inv = 1 indicates a catchment with a low land rent and thereby a preferred placement of a wetland to minimize tradeoff.

2.2.2.5. Flash-flood risk (FF Risk). We calculated the area of buildings potentially flooded as a result from a 100 year rainfall event in Denmark (Danish Geodata Agency, 2014). The result was summed for each catchment and adjusted for area of the ID15 catchments. The output expresses total area of flooded buildings relative to total catchment area. A 100 year rainfall event theme of Denmark was calculated (SCALGO, 2014). The amount of rain required for a 100 year event varies across Denmark, and is explicitly mapped in this theme. The theme is based on historical rainfall statistics and does not include future climate changes. Based on the Danish National Elevation Model (Danish Geodata Agency, 2007) with a spatial resolution of 1.6-m and the SCALGO/Hydrology-Flood module (SCALGO, 2014), we computed the terrestrial areas at risk of flooding by such an event. The elevation model was conditioned to account for flow under bridges, in drains and sewer pipes, as far as officially registered (Danish Geodata Agency, 2013). We then selected all buildings spatially intersecting the mask of flooded areas. The buildings were extracted from the Danish National Topographic Map, Kort10 (Danish Geodata Agency, 2014). The flash flood

M.V. Odgaard et al. / Ecological Indicators 77 (2017) 151–165

157

Fig. 4. Workflow of the two parts of the analysis, 1) Present Priorities and 2) Future Recommendations. The weights in the second analysis is adapted from the defined policy scenarios CliEnvi, LandEcon and RichNat (ESI = ecosystem service indicators RCW = reconstructed wetlands).

Fig. 5. Illustration of method used to calculate the ecosystem-service-indicator data for recreational potential (Rec Pot) and inverse land rent (LRent Inv).

158

M.V. Odgaard et al. / Ecological Indicators 77 (2017) 151–165

map is an indicator of the demand for ecosystem services that mitigate the risk of flooding built property.

2.2.3. Policy scenarios In the Future Farming project, policy scenarios were modelled using a multi criteria analysis, based on a relatively fine scale resolution (50 × 50m) geographically specific, environmental, and economic data spanning all of Denmark (Olsen et al., 2014). The aim was to focus on the effects of different types of management on the land use and land cover for Denmark in 2050. The scenarios were designed and then back-casted to explore what it would take to implement the selected possible ‘ideal future landscapes’. Since we based our analysis on the best available data, we adjusted the policy scenarios to meet our data foundation to the best possible degree. This added up to the following three major scenario pathways: i) Climate Adaptation and Protection of Aquatic Environment (CliEnvi), ii) Land Based Economy (LandEcon) and iii) the Rich Nature (RichNat) scenario. Each scenario focused on different, but equally important, aspects of sustainable management (Jørgensen et al., 2015). Each scenario was calculated by multiplication of the normalized ecosystem-service indicator values, with the assigned weights for each scenario (Fig. 4).

2.2.3.1. Climate adaptation and protection of aquatic environment (CliEnvi). The scenario Climate Adaptation and Protection of Aquatic Environment (CliEnvi) was based on the Green Growth scenario described in Olsen et al. (2014) and Jørgensen et al. (2015), which was defined as a management plan that focused on mitigating the nutrient and pesticide runoff from agricultural production, and reducing greenhouse gas emissions. As we, furthermore, had good data on the high risk areas affected by flash floods (related to climate change with more water in the landscape, see 2.2.3.2) we in addition included the FF Risk to this scenario. In this scenario the dominating services were N Mit and FF Risk (Fig. 4).

2.2.3.2. Land-Based economy (LandEcon). The Land-Based Economy scenario (LandEcon) corresponded to a combination of the UrbanRural and the biobased economy scenarios in the Future Farming project (Olsen et al., 2014; Jørgensen et al., 2015). This scenario was defined as a combined demand for rural development and the increasing urban population’s demand for recreation possibilities and local food production. This scenario focused on the markedbased economic possibilities in the landscape and thus favored Rec Pot and LRent Inv (Fig. 4). We furthermore chose to increase N Mit because of the increased agricultural production in this scenario (Fig. 4). In this scenario the dominating services were Rec Pot and LRent Inv followed by N Mit (Fig. 4).

2.2.3.3. Rich nature (RichNat). In the Rich Nature scenario (RichNat), biodiversity had the highest weight. This scenario focused on creating a diverse, balanced, and interconnected nature interrupted by agricultural fields (Jørgensen et al., 2015). As the fairly coarse scale of the ID15 catchments limited the sensitivity of the bioscore values, and the fact that intensely farmed fields had a score of zero by default, we chose to define high Biodiv in the catchment, as the desired areas to reconstruct wetlands (Fig. 4). In this scenario we furthermore enlarged Rec Pot and N Mit weights (Fig. 4), in order to have less available nutrients in the catchment, but retain a possible source of income based on recreation for local stakeholders. In this scenario the dominating services was Biodiv followed by Rec Pot and N Mit (Fig. 4).

2.3. Data analysis The analysis was partitioned into two distinct sections: 1) analyzing the present priorities, and 2) analyzing future recommendations (Fig. 4). 2.3.1. Present priorities In the first section, present priorities, we analyzed the present distribution of ecosystem service indicators by mapping the five selected ecosystem service indicators, and analyzed the mean composition of ecosystem services indicators in the catchments that have a reconstructed wetland. To test whether there was a tendency to prioritize the placement for reconstructing wetlands by their ecosystem services a wetland provide or benefit or by random choice, we conducted a Mann-Whitney Wilcoxon test on the nonparametric ecosystem-service-indicator values of the catchments with reconstructed wetlands vs. catchments without wetlands median value (five tests) as well as vs. a random subset median value (five tests) and reported the U-statistics. The random subset with random numbers from 0 to 1 was produced automatically in Excel 2010 using the function Rand(). This identified the priorities of the present policies on the ecosystem service indicators in the catchments of the reconstructed wetlands. 2.3.2. Future recommendations The second section, future recommendations, consists of identifying catchments where it would be most optimal to reconstruct a wetland based on the three future policy scenarios: i) Climate Adaptation and Protection of Aquatic Environment (CliEnvi), ii) Land Based Economy (LandEcon) and iii) the Rich Nature (RichNat) scenario. The scenarios were built by assigning each ecosystem service indicator a different weight depending on the prioritization represented by the policy scenario (Fig. 4). All weighted services were then summed in each catchment. For each scenario the catchments with the highest 0–10% (90th percentile, n = 302), 10–20% (80th percentile, n = 302), 20–30% (70th percentile, n = 302), and 30–40% (60th percentile, n = 302) suitability for reconstructing a wetland were selected and were examined for spatial overlap. Scenario overlap of the 90 th , 80 th , 70 th , 60 th percentile were computed in a separate map describing the maximum suitability map. The scenario overlaps reveal if some areas have synergies in the landscape that yield high ecosystemservice indicator-values, taking all chosen policy priorities − or scenarios − into account. Hence, the suitability hotspots are the areas where policy targets of all three policy scenarios may best be achieved, and where long term sustainable wetland construction can be achieved. We used ArcGIS 10.3 (ESRI, 2012) as the primary processing software and R for statistical analyses (R Development Core Team, 2008). 3. Results 3.1. Present priorities and general distribution of ecosystem services indicators The distribution of Rec Pot shows generally low variation with mainly high values and a few low values in costal zones (Fig. 6a). Biodiv depict opposite distributions of N Mit (Fig. 6bc) but similar distributions to LRent Inv (Fig. 6bd), with the highest values in areas along the coasts − especially the west coast − the Lake District of central Jutland, urban areas, as well as in Northern Zealand (Fig. 6bd). Overall LRent Inv has higher values compared to Biodiv. FF Risk show a generally low variation with mainly low values and sporadically distributed higher values (Fig. 6e). These high values are mainly situated near urban areas (Figs. 2 and 6 e). For all five services the median of catchments with a reconstructed wetland

M.V. Odgaard et al. / Ecological Indicators 77 (2017) 151–165

159

Fig. 6. Distribution of the five ecosystem service indicators in Denmark: four that gain value from reconstructed wetlands (Rec Pot, Biodiv, N Mit, and FF Risk) and one service trade-off (LRent Inv). The scale is normalized values [(x-x(min) /x(max) -x(min) )].

Table 2 The U-statistics from the Mann-Whitnet Wilcoxon tests. The first five tests are the median values of ecosystem service indicators in catchments with reconstructed wetlands (n = 350) against median values of indicators in catchments without reconstructed wetlands (n = 2673). The last five tests are the median values of ecosystem service indicators in catchments with reconstructed wetlands (n = 350) against median values of random values (n = 350). Test variables

Rec Pot × Rec Pot Biodiv × Biodiv N Mit × N Mit LRent Inv × LRent Inv FF Risk × FF Risk Rec Pot × Random Biodiv × Random N Mit × Random LRent Inv × Random FF Risk × Random

Outputs from the Mann-Whitney Wilcoxon test (U-statistics) U

p values

380820 526130 423140 525500 430760 122300 9459 35171 89711 4351

1.49 × 10−8 1.40 × 10−4 3.65 × 10−3 1.70 × 10−4 1.59 × 10−3 2.20 × 10−16 2.20 × 10−16 2.20 × 10−16 2.20 × 10−16 2.20 × 10−16

is statistically significantly different from both catchments with no reconstructed wetland and the random subset (Tables 1 and 2). Overall, according to the statistical results the present situation indicates that reconstructed wetlands are prioritized in areas with higher Rec Pot, lower Biodiv, higher N Mit, lower LRent Inv, and lower FF Risk compared to catchments with no reconstructed wetland (Tables 1, 2, and Fig. 7).

3.2. Policy scenarios and future recommendations The CliEnvi scenario, which prioritizes high N Mit and areas with high FF Risk (Fig. 4), indicated a high suitability in areas with high nitrogen load to the coast (Figs. 6 c and 8 a) and around urban areas (Figs. 2 and 8 a). The LandEcon scenario which prioritizes rural economic development and urban recreation demand (Fig. 4) indicate a high suitability in areas with a high LRent Inv, whereas Rec Pot is not important (Figs. 6 ad and 8 b) Lastly, the RichNat scenario which prioritizes Biodiv followed by N Mit and Rec Pot (Fig. 4) had substantial hotspots on the coasts, especially the west coast, the Lake District in central Jutland, and northern Zealand (Fig. 8c). This coincides with the distribution of biodiv (Fig. 6b). Of the three policy scenarios CliEnvi and LandEcon have most Id15 catchment overlaps in all percentiles followed by LandEcon and RichNat (Table 3). CliEnvi and RichNat have least overlaps (Table 3). The maximum suitability map indicating the scenario hotspots, shows the overlaps of the highest 90th, 80th, 70th, and 60th percentile of the scenario value (n = 302 for each scenario and each percentile, Fig. 9). The hotspots represent the catchments that hold the most potential for future reconstruction of wetlands, when considering these five ecosystem service indicators and these three policy scenarios. There are 35 catchments that overlap completely with all three scenarios in the 90th percentile (Table 3), which is a relatively low number ∼ 1% of the total catchments. These possible hotspots are mostly located in and around peri- and urban areas, such as North Zealand and Northern Copenhagen city, in Central Jutland and around Aarhus.

160

M.V. Odgaard et al. / Ecological Indicators 77 (2017) 151–165

Fig. 7. Present situation. Ecosystem-service-indicator composition in catchments with reconstructed wetlands (CW). Data was Z-score scaled, thus zero is the mean of catchments without wetlands..

Fig. 8. The three scenarios maps, CliEnvi, LandEcon, and RichNat. The suitability index indicates a catchment’s suitability for a reconstructed wetland, ranging from the lowest suitability (blue) to the 90th percentile suitability (red), based on the proposed future policies. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Table 3 Scenario-overlap in catchment-suitability for wetland reconstruction. Scenario suitability is ranked descending for each scenario and then investigated for overlap between all combinations of scenarios. The 90th percentile, 80th percentile, 70th percentile, and 60th percentile are represented. Values are number of catchments. Scenario overlay

Scenario overlap in 90th percentile

Scenario overlap in 80th percentile

Scenario overlap in 70th percentile

Scenario overlap in 60th percentile

CliEnvi × LandEcon CliEnvi × RichNat LandEcon × RichNat CliEnvi × LandEcon × RichNat

144 45 129 35

67 28 61 10

61 30 44 11

56 35 44 7

Two of these 35 catchments contain reconstructed wetlands today (Figs. 4 and 9). Furthermore, the more remote areas in northwestern Jutland also display great potential for reconstructed wetlands.

4. Discussion 4.1. Present priorities The present priorities establish a baseline for the whole analysis. We conducted this analysis to find out whether there were any

patterns in the catchments that allowed for further analysis and scenario building. The general distribution of the five ecosystem service indicators reflects other underlying drivers. Rec Pot does not show much variation, but the distribution of Biodiv reflect areas with already established nature areas such as national parks or larger forests, and is thus similar to the inversed land rent. LRent Inv reflect soil conditions that place most of the arable soils with high land rent on the dominating loamy and clayey soils in the eastern parts of Jutland and on the islands (Figs. 2 and 6). Sandy soils, however, can also provide high land rents, e.g. for harboring livestock with high needs for roughage production and land for distribution

M.V. Odgaard et al. / Ecological Indicators 77 (2017) 151–165

161

Fig. 9. Maximum suitability-map. Color indicates hotspots found by overlapping the 90th, 80th, 70th, and 60th percentiles of the three policy scenarios. The remaining catchments (grey) have two or less overlaps between scenarios. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

of manures, but also for high value vegetable productions, such as carrots and potatoes (Greve et al., 2007; Olsen et al., 2014). N Mit shows the high potential in, i.a., parts of Jutland, which is most likely because of said soil conditions combined with a high precipitation and a high livestock density (Dalgaard et al., 2014; Højberg et al., 2015). Also the major islands, show high N Mit in some areas, dispte a lower livestock production. This is most probably due to the low local N retetnion. Hence, the spatial distribution of N Mit depends on agricultural N leaching but also the N retention. Therefore, N Mit can also be high in areas of low agricultural pressure if the N retention is low − indicating a high costal N load. FF Risk is directed by heavy populated areas in the areas where it displays high variation (Figs. 2 and 6 e). The ecosystem service indicators vary between catchments with reconstructed wetlands compared to those without, and the indicators also vary from random both for catchments with and without reconstructed wetlands, indicating that wetland reconstruction

sites are not randomly distributed across the Danish region. The catchments with a reconstructed wetland are dominated by a statistically significant higher Rec Pot, lower biodiv, higher N Mit, lower LRent Inv, and low FF Risk than in catchments without reconstructed wetlands. The statistically significant higher potential for recreation in catchments with wetlands compared to without wetlands if probably due to the big sample size in our analysis (sample 1 n = 2673, sample 2 n = 350). Still, the variation in Rec Pot is almost non-existent, and displays mainly high and only few low values (Fig. 6a). Therefore, it should not be interpreted similar to the statistical test. Hence, from a policy making perspective Rec Pot has probably not been taken into account when prioritizing present location of wetlands despite the significance of the Mann-Whitney Wilcoxon test (Tables 1 and 2). This is also partly the case for risk of flooding. FF Risk also show a statistically significant higher risk of flooding in catchments with wetlands compared to catchments without (Table 1). In the areas with a relatively low variation in

162

M.V. Odgaard et al. / Ecological Indicators 77 (2017) 151–165

FF Risk it should not be considered an important indicator, but in areas with more variation such as areas near the big cities it has a greater importance (Figs. 2 and 6 e). We did not expect to find a significant expression of this service as the risk of flash floods has not historically been a large problem in Denmark, where more emphasis has been directed towards coastal water floods (Moeslund et al., 2009). Biodiv is lower in catchments with reconstructed wetlands (Table 1). This is against our hypothesis but is probably due to the fact that reconstructed wetlands have been placed in high intensive farming areas away from larger connected nature areas. This is also reflected in the higher land rent in catchments with reconstructed wetlands which is probably due to the fact that wetlands have previously been located in intensive agricultural areas with high land rent (low LRent Inv) and N Mit potential rather that remote biodiversity rich areas. Also, as expected, N Mit is higher in catchments with reconstructed wetlands compared to those without because as this is the main goal of constructing wetlands in the Water Framework Directive (Dalgaard et al., 2014; Natur og Landbrugskommisionen, 2013). However, we might have expected a more pronounced effect. The higher land rent (low LRent Inv) together with the higher N Mit in areas with reconstructed wetlands compared to areas without reconstructed wetlands might seem surprising since placement of wetlands often results in a trade-off between the economic cost of the project, including the loss of profit from a field in rotation, and the nitrogen mitigation potential, but a high land rent (low LRent Inv) together with high N Mit might lead to lower costs per kg N. Consequently, it is plausible that the value of wetland reconstruction schemes may be higher in intensive agricultural regions that thus display a high land rent. 4.2. Policy scenarios and future recommendations We selected the policy scenarios and hotspot mapping to analyze the future perspectives of priorities in reconstructed wetlands because of the inherent dissemination value. Scenario testing is a great tool for discussing implications for policies for stakeholders (Jørgensen et al., 2015). This enables policy-makers to understand the implications of the policies and managers and users can better decide what a desirable future entails. We chose to use pre-defined policy scenarios, but this analysis has no restrictions on what scenarios to use, nor how many to include. The dominating ecosystem service indicators in each scenario tend to reflect the distribution of the scenario values and their individual 90th percentiles (Figs. 6 and 8). CliEnvi reflects N Mit and FF Risk (Figs. 8 and 6 ce) which are also the dominating indicators (Fig. 4). In the case of FF Risk this is mainly true for the areas containing the major cities where this indicator actually displays a variation. LandEcon reflects mainly the distributions of L Rent and N Mit (Figs. 8 and 6 cd), despite the fact that Rec Pot and N Mit were the dominating indicators (Fig. 4). This reflects the low variation in the values of Rec Pot (Fig. 6a). Still the values in the LandEcon scenario are higher compared to CliEnvi and RichNat because of the generally high values of Rec Pot. Hence, Rec Pot will not influence the placement of the hotspots in the LanEcon scenario but will instead have an overall importance and yield high scenario values. Hence, based on these results Rec Pot should not be taken into account when placing a reconstructed wetland, since wetlands will equally benefit recreational value in the study area. Lastly, RichNat reflects Biodiv (Figs 8 and 6 b) which is also the dominating indicator (Fig. 4). CliEnvi and LandEcon scenarios display the highest occurrence of overlapping catchments (Table 3) of the same hotspots, in both cases generally associated with urban areas. The main difference is the fact that CliEnvi is more evenly spread out across the study region than LandEcon, and the highest values in CliEnvi are situated in and around high population density areas (for instance East Jut-

land), including the major cities, whereas LandEcon has the much higher values in the rural parts of Denmark. This is probably influenced by the FF Risk is weighted higher in CliEnvi and directs the hotspots towards the cities, whereas high LRent Inv and high N Mit drives the LandEcon hotspots towards rural areas (Fig. 8). RichNat is quite different from both other scenarios. The main driver is the Biodiv and thus has a more biophysical supply driver than the other two scenarios. The highest values in this scenario primarily lie along the coast, along conservation areas and their adjacent catchments (Ejrnæs et al., 2014), and thus the hotspots are not as correlated with human population and agricultural production as the other two policy scenarios. However, as the Rec Pot is also dominant in this scenario, the areas in around the major urban and metropolitan areas have a high potential of adopting a RichNat scenario. Hence, despite the fact that Rec Pot is no more important in urban areas compared to other areas (from a policy perspective) it still raise recreational value of an area to construct a wetland (Bjørner et al., 2014), just not more in one area compared to another (Fig. 6a). It is noteworthy to see that catchments in the two largest cities in Denmark, Copenhagen and Aarhus, are both represented in the 90th percentile overlap. However, most of the catchments are associated with a peri-urban area, since there is a high level of urban development throughout Denmark. This suggests that placing nature restoration and conservation with recreation near big cities and peri-urban areas have potential and does not have to solely be in remote and marginal areas. The overall message for the future recommendation is that synergy hotspots on the Maximum Suitability map (Fig. 9) show very few catchments fitting to the priorities in all three scenarios (n = 35). A higher suitability and synergy across scenarios may have been expected, but of course, depends on the nature of the scenarios chosen and the level of orthogonality in them, which reduce overlap (Table 3). This is something to take into account when choosing scenarios for this method of analysis. It is noteworthy though, that of the 35 hotspot catchments we have found to be the best potential wetland reconstruction sites, only two contain a reconstructed wetland, according to the Danish Ministry of Food, Agriculture and Fisheries (2014). This supports our hypothesis that reconstructed wetlands to date have been placed on the basis of the WFD sole focus on water pollution mitigation, and not as such on a wider array of ecosystem services. If more services were taken into account, our results suggest that in particular peri-urban areas as well as larger isolated nature areas have a high and until now not utilized potential value for placement of reconstructed wetlands.

4.3. The multi-criteria, ecosystem-service value method The present study illustrates the ecosystem value method used to analyze how future policy scenarios may shape land use and thus also the landscapes themselves. The framework is, to our knowledge, the first of its kind that focus on ecosystem-serviceindicator values for screening catchments for suitable locations for reconstructing wetlands in Denmark. It is aimed primarily at policymakers and managers to use policy scenarios to enhance communication of policy implementations. Importantly, we included a scenario hotspot analysis to highlight possibilities for landscape and watershed scale synergies of future policy scenarios. Because of the relatively coarse scale and limited number of ecosystems service dimensions, this tool should only be viewed as a national/regional scale screening tool to identify relevant catchments for further planning of new wetlands. It is not a local scale tool to select the specific site within the catchment, which should include more stakeholder involvement, as well as detailed investigations of hydrology, soil conditions, land use, etc. However, it is much easier to perform an in depth analysis of a small selection of

M.V. Odgaard et al. / Ecological Indicators 77 (2017) 151–165

catchments instead of an entire region or even nation, and for this the presented screening tool may be useful. We chose the ecosystem service idicators that were relevant for Denmark, the type of wetland we aimed for, and the quality of data available. This could all be adapted for other systems, e.g., depending on the aim of the study, the type of wetland (e.g. mini-constructed wetlands, or large lake-size reconstructed wetlands), and the data available. In other regions, these services may change with climatic zones, development status, and political systems (Barbier et al., 1997; Turner et al., 2015), and it is, thus, crucial that the right services and indicators are applied (Seppelt et al., 2011; Burkhard et al., 2012). For instance, small scale wetlands are not very effective for biodiversity conservation and flash flood mitigation, as they are mostly in situ nitrogen mitigating, and thereby are too small to mitigate flash floods and has too high concentrations of nitrogen and phosphorous to benefit biodiversity (Eriksen et al., 2014). There are conflicting opinions on whether a reconstructed wetland will increase the overall biodiversity, but the main conclusion is that this will depend on the design of the wetland (Hansson et al., 2005). Here, we assume an overall improvement of biodiversity by converting cultivated area to a reconstructed wetland. Still, it is also beyond the scope of this analysis and is a part of the local solutions after a catchment is selected. The recreation indicator was modelled from a comprehensive set of proxies (Bjørner et al., 2014), but it would have produced a better picture if we had access to an actual demand indicator. A low supply does not equal a high demand if few people are residing in the catchment as most people partake in recreation close to their residence, although large nature areas draw large numbers of visitors from far away (Bjørner et al., 2014). Land rent was chosen as a better indicator than physical crop harvest yield, because it also takes the cost of management inputs into account. Land rent includes the costs and benefits of land based production, including the cost for livestock manure distribution, and the benefits of roughage production for livestock. However, the gross margin from the livestock production itself is not included in this analysis. The influence of crop versus livestock production will likely vary from region to region, where different types of agricultural practices and biophysical spatial relationships are in play (Turner et al., 2015), and the correct indicators for the food provisioning services should be prioritized for the region of interest.

163

igation, and to some extent higher flash flood risk, whereas biodiversity is highly underrepresented compared to catchments without reconstructed wetlands. Future policies of CliEnvi and LandEcon show the highest number of potential synergies, and result in similar distribution of catchments with a high benefit from wetland reconstruction, apparently directed by peri-urban areas. RichNat is complementary to this, and the optimal spots for this strategy lie in and around established and protected nature areas, but also appear in urban and high populated areas. Maximum suitability overlaps for reconstructed wetlands across all three prioritization scenarios occur in urban and peri-urban areas such as the larger cities, Northern Zealand as well as the nature areas in Northwest Jutland. However, only two of the hotspot catchments found to be the maximum suitable wetlands reconstruction sites by using this method contains a reconstructed wetland to date, indicating that present placement do not support a wider array of ecosystem services. A more systematic approach, such as the ecosystem service valuedriven method demonstrated in the present case study, could target more services and improve the overall benefit from wetlands. Furthermore, the method can be used to identify areas for local action and policy scenario evaluation of wetland construction potentials in catchments. Acknowledgements The development and testing of the present method is dependent on previously develop methods and datasets as well as expertise within each of the ecosystem service areas included. Thanks to good colleagues, especially from Aarhus University Departments of Environmental Sciences (Mette Termansen et al.), Agroecology (Christen Duus Børgesen, Inge T. Kristensen et al.), Morten Revsbech at Scalgo.dk, and Bioscience Kalø (Jesper Moeslund et al.) for all their kind help and assistance. Also, a large thanks to both the Velux foundation sponsored www.fremtidenslandbrug. dk project, as well as the Danish Innovation Fund, the dNmark Research Alliance www.dnmark.org, and the www.buffertech.dk project for funding this study. Finally, thanks to the institute of Agroecology and the Velux foundation for funding the Ph.D. of Katrine Grace Turner who contributed substantial to this specific paper. References

5. Conclusions This study presents a new ecosystems service value-driven method for multi-criteria assessment of catchments and their suitability for reconstruction of wetlands. Using Denmark as a study case, we mapped the value of the described five ecosystem service indicators, and conducted: 1) firstly an analysis of the present situation in catchments that harbored a reconstructed wetland and 2) second a production of one maximum suitability map to find the maximum synergy between three policy scenarios. In summary, Biodiv, N Mit, LRent Inv and to some extent FF Risk are non-random spatially clustered whereas Rec Pot show a non-varying spatial distribution. The geography of LRent Inv, Biodiv, and N Mit are determined by land use and biophysical characteristics, whereas FF Risk follows human population densities. The low variation of Rec Pot makes this ecosystem service indicator non-important in a policy decision-making process, despite it showing statistically significant higher values in catchments with reconstructed wetlands compared to catchments without. The statistically results show that current reconstructed wetlands are situated in catchments with statistically significant higher land rent (i.e. agriculture intensive areas), higher nitrogen mit-

Adhikari, A.R., Acharya, K., Shanahan, S.A., Zhou, X.P., 2011. Removal of nutrients and metals by constructed and naturally created wetlands in the Las Vegas Valley, Nevada. Environ. Monit. Assess. 180, 97–113. Agger, P., Baatrup-Pedersen, A., Bregnballe, A.T., Bruun, H., Buttenschøn, R.M., Dalgaard, T., Danielsen, F., Clausen, P., Ejrnaes, R., Fjeldsaa, J. et al. 2012. Danmarks natur frem mod 2020–om at stoppe tabet af biologisk mangfoldighed. Det Grønne Kontaktudvalg, c/o Danmarks Naturfredningsforening. Bai, Y., Zhuang, C.W., Ouyang, Z.Y., Zheng, H., Jiang, B., 2011. Spatial characteristics between biodiversity and ecosystem services in a human-dominated catchment. Ecol. Complexity 8, 177–183. Barbier, E.B., Acreman, M., Knowler, D., 1997. Economic Valuation of Wetlands: a Guide for Policy Makers and Planners. Ramsar Convention Bureau Gland, Switzerland. Barbier, E.B., 2011. Wetlands as natural assets. Hydrol. Sci. J.–Journal Des Sciences Hydrologiques 56, 1360–1373. Bateman, I.J., Harwood, A.R., Mace, G.M., Watson, R.T., Abson, D.J., Andrews, B., Binner, A., Crowe, A., Day, B.H., Dugdale, S., Fezzi, C., Foden, J., Hadley, D., Haines-Young, R., Hulme, M., Kontoleon, A., Lovett, A.A., Munday, P., Pascual, U., Paterson, J., Perino, G., Sen, A., Siriwardena, G., van Soest, D., Termansen, M., 2013. Bringing ecosystem services into economic decision-making: land use in the United Kingdom. Science 341, 45–50. Beatty, W.S., Kesler, D.C., Webb, E.B., Raedeke, A.h., Naylor, L.w., Humburg, D.D., 2014. The Role of protected area wetlands in waterfowl habitat conservation: implications for protected area network design. Biol. Conserv. 176, 144–152. Bjørner T.B., Jensen C.U., Termansen M., 2014. Den rekreative vaerdi af naturområder i Danmark [The recreational value of nature areas in Denmark]. Brander, L.M., Brauer, I., Gerdes, H., Ghermandi, A., Kuik, O., Markandya, A., Navrud, S. Nunes, P.A.L.D, Schaafsma, M., Vos, H., Wagtendonk, A., 2012. Using

164

M.V. Odgaard et al. / Ecological Indicators 77 (2017) 151–165

meta-analysis and GIS for value transfer and scaling up: valuing climate change induced losses of European wetlands. Environ. Resource Econom. 52, 395–413. Burkhard, B., Kroll, F., Nedkov, S., Müller, F., 2012. Mapping ecosystem service supply, demand and budgets. Ecol. Indic. 21, 17–29. Carpenter, S.R., Mooney, H.A., Agard, J., Capistrano, D., Defries, R.S., Diaz, S., Dietz, T., Duraiappah, A.K., Oteng-Yeboah, A., Pereira, H.M., Perrings, C., Reid, W.V., Sarukhan, J., Scholes, R.J., Whyte, A., 2009. Science for managing ecosystem services: beyond the millennium ecosystem assessment. Proc. Natl. Acad. Sci. U. S. A. 106, 1305–1312. Chave, P., 2001. The EU Water Framework Directive. Costanza, R., De Groot, R., Sutton, P., Van der Ploeg, S., Anderson, S.J., Kubiszewski, I., Farber, S., Turner, R.K., 2014. Changes in the global value of ecosystem services. Global Environ. Change 26, 152–158. Cowling, R.M., Egoh, B., Knight, A.T., O’Farrell, P.J., Reyers, B., Rouget, M., Roux, D.J., Welz, A., Wilhelm-Rechman, A., 2008. An operational model for mainstreaming ecosystem services for implementation. Proc. Natl. Acad. Sci. U. S. A. 105, 9483–9488. Dalgaard, T., Hansen, B., Hasler, B., Hertel, O., Hutchings, N.J., Jacobsen, B.H., Jensen, L.S., Kronvang, B., Olesen, J.E., Schjørring, J.K., 2014. Policies for agricultural nitrogen management – trends, challenges and prospects for improved efficiency in Denmark. Environ. Res. Lett. 9, 115002. Danish Geodata Agency, 2007. The Danish National Elevation Model, DK-DEM. DHM2007-DTM. http://download.kortforsyningen.dk/content/dhmterraen-16m-grid, (Accessed 01.12.2015). Danish Geodata Agency 2013. The Danish Hydrology Model − corrections, DHyM/tilpasninger. http://download.kortforsyningen.dk/content/ dhymtilpasninger, (Accessed 01.12.2015). Danish Geodata Agency, 2014. The Danish National Topographic Map, Kort10. http://download.kortforsyningen.dk/content/kort10, (Accessed 01.12.2015). Danish Ministry of Food, Agriculture and Fisheries, 2014. Alle miljø- og naturprojekter på ét kort. Natur og Erhvervstyrelsen, Copenhagen, Denmark. R. Development Core Team, R: A language and environment for statistical computing. R Foundation for Statistical Computing Vienna Austria 2008. Dick, J., Maes, J., Smith, R.I., Paracchini, M.L., Zulian, G., 2014. Cross-scale analysis of ecosystem services identified and assessed at local and European level. Ecol. Indic. 38, 20–30. ESRI. 2012. ArcMap 10.1., Environmental Systems Resource Institute, Redlands, California. Egoh, B., Reyers, B., Rouget, M., Richardson, D.M., Le Maitre, D.C., Van Jaarsveld, A.S., 2008. Mapping ecosystem services for planning and management. Agric. Ecosyst. Environ. 127, 135–140. Ejrnæs, R., Petersen, A.H., Bladt, J., Bruun, H.H., Moeslund, J.E., Wiberg-Larsen, P., Rahbek, C., 2014. Biodiversitetskort for Danmark [Biodiversity Map for Denmark]. Nationalt center for Miljø og Energi. Ellis, E.C., Goldewijk, K.K., Siebert, S., Lightman, D., Ramankutty, N., 2010. Anthropogenic transformation of the biomes, 1700–2000. Global Ecol. Biogeogr. 19, 589–606. Eriksen, J., Jensen, P.N., Jacobsen B.H., (eds.) 2014. Virkemidler til Realisering af 2. Generations Vandplaner og Målrettet Arealregulreing. DCA report No 052. Foulum, Denmark ISBN 978-87-93176-44-7. Eurostat. 2010. EU farm structure survey agricultural survey. http://ec.europa.eu/ eurostat/web/agriculture/farm-structure (Accessed 01.12.2015). Fisher, B., Polasky, S., Sterner, T., 2011. Conservation and human welfare: economic analysis of ecosystem services. Environ. Resource Econom. 48, 151–159. Frei, C., Schöll, R., Fukutome, S., Schmidli, J., Vidale, P.L., 2006. Future change of precipitation extremes in Europe: intercomparison of scenarios from regional climate models. J. Geophys. Res.: Atmos. 111. Ghermandi, A., Van den Bergh, J.C.J.M., Brander, L.M., De Groot, H.L.F., Nunes, P.A.L.D., 2010. Values of natural and human-made wetlands: a meta-analysis. Water Resour. Res. 46. Grøndahl, L., Poulsen, N., Christensen, J.H., Arnbjerg-Nielsen, K., Grindsted, A., Halsnaes, K., Jeppesen, E., Madsen, H., Olesen, J.E., Porter, J.R., 2014. Analyse af IPCC delrapport 2: Effekter, klimatilpasning og sårbarhed-med saerligt Fokus på Danmark. Naturstyrelsen. Grenyer, R., Orme, C.D.L., Jackson, S.F., Thomas, G.H., Davies, R.G., Davies, T.J., Jones, K.E., Olson, V.A., Ridgely, R.S., Rasmussen, P.C., 2006. Global distribution and conservation of rare and threatened vertebrates. Nature 444, 93–96. Greve, M.H., Greve, M.B., Bøcher, P.K., Balstrøm, T., Breuning-Madsen, H., Krogh, L., 2007. Generating a Danish raster-based topsoil property map combining choropleth maps and point information. Geografisk Tidsskrift 107, 1–12. Greve, M., Reyers, B., Lykke, A.M., Svenning, J.C., 2013. Spatial optimization of carbon-stocking projects across Africa integrating stocking potential with co-benefits and feasibility. Nat. Commun. 4. Greve, M., Christensen, O., Greve, M., Bou Kheir, R., 2014. Change in peat coverage in Danish cultivated soils during the past 35 years’. J. Soil Sci. 179, 250–257. Højberg, A., Windolf, J., Børgesen, C.D., Troldborg, L., Tornbjerg, H., Blicher-Mathiesen, G., Kronvang, B., Thodsen, H., Ernstsen, V., 2015. National kvaelstofmodel. Oplandsmodel til belastning og virkemidler Metode rapport. De Nationale Geologiske Undersøgelser for Danmark og Grønland, Klima-, Energi- og Bygningsministeriet. Aarhus Universitet; DCE − Nationalt Center for Miljø og Energi, DCA − Nationalt Center for Fødevarer og Jordbrug. Hansson, L.-A., Brönmark, C., Anders Nilsson, P., Åbjörnsson, K., 2005. Conflicting demands on wetland ecosystem services: nutrient retention, biodiversity or both? Freshw. Biol. 50, 705–714.

Harrington, R., Carroll, P., Cook, S., Harrington, C., Scholz, M., McInnes, R.J., 2011. Integrated constructed wetlands: water management as a land-use issue, implementing the ‘Ecosystem Approach’. Water Sci. Technol. 63, 2929–2937. Harsh, S., 2015. Wildlife corridors: a conservation tool. Int. J. Adv. Res. 3 (3), 115–126. Jørgensen, L.B., Normander, B., Ege, C., Skovsbøl, U., Bech, G., Agger, P., Thorsen, T.S., Dubgaard, A., Olsen, S.B., Vogdrup-Schmidt, M., Dalgaard, T., Kristensen, I.T., Kristensen, D.K., Kjeldsen, C., 2015. Scenarier for Fremtidens LandbrugHovedrapport. Det Økologiske Råd, Københavns Universitet og Aarhus Universitet, København. Jacobsen, B.H., 2012. Analyse af landbrugets omkostninger ved implementering af vandplanerne fra 2011, FOI Udredning; Nr. 2012/6. Jarchow, M.E., Kubiszewski, I., Larsen, G.L.D., Zdorkowski, G., Costanza, R., Gailans, S.R., Ohde, N., Dietzel, R., Kaplan, S., Neal, J., Petrehn, M.R., Gunther, T., D’Adamo, S.N., McCann, N., Larson, A., Damery, P., Gross, L., Merriman, M., Post, J., Sheradin, M., Liebman, M., 2012. The future of agriculture and society in Iowa: four scenarios. Int. J. Agric. Sustain. 10, 76–92. Jelinski, D.E., Wu, J., 1996. The modifiable areal unit problem and implications for landscape ecology. Landsc. Ecol. 11, 129–140. Kayranli, B., Scholz, M., Mustafa, A., Hedmark, Aa., 2010. Carbon storage and fluxes within freshwater wetlands: a critical review. Wetlands 30, 111–124. Kristensen, K., Waage-Petersen, J., Børgesen, C.D., Vinther, F.P., Grant, R., Blicher-Mathiesen, G., 2008. Reestimation and Further Development in the Model N-LES, N-LES3 to N-LES4. Aarhus Universitet, Faculty of Agricultural Science. Kronvang, B., Andersen, H.E., Børgesen, C., Dalgaard, T., Larsen, S.E., Bøgestrand, J., Blicher-Mathiasen, G., 2008. Effects of policy measures implemented in Denmark on nitrogen pollution of the aquatic environment. Environ. Sci. Policy 11, 144–152. Larsson, T., 2004. Vådområder i Norden og Ramsarkonventionen. (Nordic Wetlands and the Ramsar Convention). Levin, G., Jepsen, M.R., Blemmer, M.K., 2012. Basemap: Technical Documentation of a Model for Elaboration of a Land-Use and Land-Cover Map for Denmark. Aarhus University, DCE −Danish Centre for Environment and Energy. MEA, 2005. Ecosystems and Human Well-Being: Synthesis. Island Press. Mace, G.M., Norris, K., Fitter, A.H., 2012. Biodiversity and ecosystem service: a multilayered relationship. Trends Ecol. Evol. 27, 19–26. Maltby, E., Acreman, M.C., 2011. Ecosystem services of wetlands: pathfinder for a new paradigm. Hydrol. Sci. J.–Journal Des Sciences Hydrologiques 56, 1341–1359. Melbourne-Thomas, J., Johnson, C.R., Fung, T., Seymour, R.M., Cherubin, L.M., Arias-Gonzalez, J.E., Fulton, E.A., 2011. Regional-scale scenario modeling for coral reefs: a decision support tool to inform management of a complex system. Ecol. Appl. 21, 1380–1398. Merken, R., Deboelpaep, E., Teunen, J., Saura, S., Koedam, 2015. Wetland suitability and connectivity for trans-saharan migratory waterbirds. PLOS ONE. Moeslund, J.E., Bøcher, P.K., Svenning, JC., Mølhave, T., Arge, L., 2009. Impacts of 21st century sea-level rise on a Danish major city–an assessment based on fine-resolution digital topography and a new flooding algorithm. Page 012022 in IOP Conference Series: Earth and Environmental Science. IOP Publishing. Moore, D., Cranston, G., Reed, A., Galli, A., 2012. Projecting future human demand on the Earth’s regenerative capacity. Ecol. Indic. 16, 3–10. Naidoo, R., Balmford, A., Costanza, R., Fisher, B., Green, R.E., Lehner, B., Malcolm, T.R., Ricketts, T.H., 2008. Global mapping of ecosystem services and conservation priorities. Proc. Natl. Acad. Sci. U. S. A. 105, 9495–9500. Natur og Landbrugskommisionen 2013: http://www.naturoglandbrug.dk/ slutrapport 2013.aspx?ID=52071 (Accessed 01.12.2015). Ogden, L.E., 2015. Do wildlife corridors have a downside. BioBriefs 65 (4), 452. Okruszko, T., Duel, H., Acreman, M., Grygoruk, M., Florke, M., Schneider, C., 2011. Broad-scale ecosystem services of European wetlands-overview of the current situation and future perspectives under different climate and water management scenarios. Hydrol. Sci. J.–Journal Des Sciences Hydrologiques 56, 1501–1517. Olsen, S.B., Vogdrup-Schmidt, M., Dubgaard, A., Normander, B., Jørgensen, L.B., Kristensen, I.T., Dalgaard, T., 2014. Detaljeret beskrivelse af multikriterieanalyse(MCA)-model anvendt i projektet Fremtidens Landbrug [Details of the Multicriteria analysis in the project Farming the Future]. Institut for Fødevare- og Ressourceøkonomi, Københavns Universitet., Frederiksberg. Pachauri, R.K., Allen, M., Barros, V., Broome, J., Cramer, W., Christ, R., Church, J., Clarke, L., Dahe, Q., Dasgupta, P., 2014. Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Parsons, G.R., 2003. The Travel Cost Model Pages 269–329 A Primer on Nonmarket Valuation. Springer. Pereira, H.M., Leadley, P.W., Proenca, V., Alkemade, R., Scharlemann, J.P.W., Fernandez-Manjarres, J.F., Araujo, M.B., Balvanera, P., Biggs, R., Cheung, W.W.L., Chini, L., Cooper, H.D., Gilman, E.L., Guenette, S., Hurtt, G.C., Huntington, H.P., Mace, G.M., Oberdorff, T., Revenga, C., Rodrigues, P., Scholes, R.J., Sumaila, U.R., Walpole, M., 2010. Scenarios for global biodiversity in the 21st century. Science 330, 1496–1501. Petit, L.j., Petit, D.R., Martin, T.E., 1995. Landscape-level management of migratory birds: loking past the trees to see the forest. Wildl. Soc. Bull. 23 (3), 420–429. Raudsepp-Hearne, C., Peterson, G.D., Bennett, E.M., 2010. Ecosystem service bundles for analyzing tradeoffs in diverse landscapes. Proc. Natl. Acad. Sci. U. S. A. 107, 5242–5247.

M.V. Odgaard et al. / Ecological Indicators 77 (2017) 151–165 Raymond, C.M., Bryan, B.A., MacDonald, D.H., Cast, A., Strathearn, S., Grandgirard, A., Kalivas, T., 2009. Mapping community values for natural capital and ecosystem services. Ecol. Econ. 68, 1301–1315. Russi, D., Ten Brink, P., Farmer, A., Badura, T., Coates, D., Förster, J., Kumar, R., Davidson, N., 2013. The Economics of Ecosystems and Biodiversity for Water and Wetlands. Ramsar Secretariat, Gland, IEEP, London and Brussels. SCALGO. 2014. SCALGO/Hydrology-Flood module. www.scalgo.dk. Seppelt, R., Dormann, C.F., Eppink, F.V., Lautenbach, S., Schmidt, S., 2011. A quantitative review of ecosystem service studies: approaches, shortcomings and the road ahead. J. Appl. Ecol. 48, 630–636. Statistics Denmark, Mational Statistical Database (Statistikbanken). Available at: [http://www.dst.dk/da/Statistik/statistikbanken]. Steffen, W., Richardson, K., Rockström, J., Cornell, S.E., Fetzer, I., Bennett, E.M., Biggs, R., Carpenter, S.R., De Vries, W., De Wit, C.A., 2015. Planetary boundaries: guiding human development on a changing planet. Science 347, 1259855. Swetnam, R.D., Fisher, B., Mbilinyi, B.P., Munishi, P.K.T., Willcock, S., Ricketts, T., Mwakalila, S., Balmford, A., Burgess, N.D., Marshall, A.R., Lewis, S.L., 2011. Mapping socio-economic scenarios of land cover change: a GIS method to enable ecosystem service modelling. J. Environ. Manage. 92, 563–574.

165

Turner, K.G., Odgaard, M.V., Bøcher, P.K., Dalgaard, T., Svenning, J.C., 2014. Bundling ecosystem services in Denmark: trade-offs and synergies in a cultural landscape. Landsc. Urban Plann. 125, 89–104. Turner, K.G., Anderson, S., Gonzales-Chang, M., Costanza, R., Courville, S., Dalgaard, T., Dominati, E., Kubiszewski, I., Ogilvy, S., Porfirio, L., Ratna, N., Sandhu, H., Sutton, P.C., Svenning, J.-C., Turner, G.M., Varennes, Y.-D., Voinov, A., Wratten, S., 2015. A review of methods, data, and models to assess changes in the value of ecosystem services from land degradation and restoration. Ecol. Modell. http://dx.doi.org/10.1016/j.ecolmodel.2015.07.017. UN Development Programme (UNDP), 2011 Human Development Report 2011, ISBN 9780230363311, available at: http://hdr.undp.org/en/content/humandevelopment-report-2011 [accessed 7 December 2016]. Whitfield, S., Reed, M.S., 2012. Participatory environmental assessment in drylands: introducing a new approach. J. Arid Environ. 77, 1–10. Willemen, L., Hein, L., Verburg, P.H., 2010. Evaluating the impact of regional development policies on future landscape services. Ecol. Econom. 69, 2244–2254.