Developing a methodology for a species-based and spatially explicit indicator for biodiversity on agricultural land in the EU

Ecological Indicators 37 (2014) 186–198

Contents lists available at ScienceDirect

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

Developing a methodology for a species-based and spatially explicit indicator for biodiversity on agricultural land in the EU Koen P. Overmars a,∗ , Catharina J.E. Schulp a,b , Rob Alkemade a , Peter H. Verburg b , Arnaud J.A.M. Temme c , Nancy Omtzigt b , Joop H.J. Schaminée d,e a

PBL Netherlands Environmental Assessment Agency, P.O. Box 30314, 2500 GH The Hague, The Netherlands Institute for Environmental Studies, Amsterdam Global Change Institute, VU University Amsterdam, De Boelelaan 1087, 1081 HV Amsterdam, The Netherlands c Chair group of Land Dynamics, Environmental Sciences Group, Wageningen University, Droevendaalsesteeg 3, 6708 PB Wageningen, The Netherlands d Environmental Sciences Group, Alterra - Wageningen University and Research, P.O. Box 47, 6700 AA Wageningen, The Netherlands e Institute for Water and Wetland Research, Radboud University Nijmegen, P.O. Box 9010, 6500 GL Nijmegen, The Netherlands b

a r t i c l e

i n f o

Article history: Received 10 May 2012 Received in revised form 25 September 2012 Accepted 7 November 2012 Keywords: Biodiversity indicator European Union Common Agricultural Policy Agriculture Land-use intensity Species richness

a b s t r a c t In Europe agricultural areas are of great importance to biodiversity conservation. One of the aims of the Common Agricultural Policy (CAP) after 2013 is to avoid additional loss of agriculture-related biodiversity. Farmland biodiversity is a public good that provides ecosystem services necessary for the sustainability of agriculture itself as well as for a sustainable environment as a whole. To evaluate policies such as the CAP and to monitor the development of biodiversity in agricultural areas, specifically designed indicators are needed. Current EU-level indicators of agricultural biodiversity are often limited to a specific species group, for example the group of farmland birds, and are not designed for evaluation of future policies. This study presents a methodology for a new indicator that is targeted specifically at biodiversity in agricultural areas, considering a large variety of species and focussing on policy. The methodology combines maps of the potential occurrence of 132 relevant species (plants and vertebrates) on a 50 km grid, with detailed information (1 km grid) on the influence of environmental pressures on these species. A first indicator map on a 1 km grid for the EU is provided, based on available data. This map shows great variety in the state of the biodiversity of agricultural areas in the EU. Generally speaking, biodiversity in agricultural areas in the south and east of the EU is in a better state than in the west and north. However, spatial variability is high between and even within regions. The presented indicator may be used to explore the dynamics of biodiversity following policy interventions, using the biodiversity map or by modelling the effect of policies on the environmental pressures that form the basis of the indicator. © 2012 Elsevier Ltd. All rights reserved.

1. Introduction Europe has an 8000-year agricultural history (Donald et al., 2002). Especially in medieval times large areas were deforested, after an earlier period of clearance in Roman times. Already three centuries ago the majority of current agricultural land was used for agriculture (Klein Goldewijk and Ramankutty, 2004). Because of the historical agricultural expansion a major part of biodiversity in Europe became dependent on agricultural land (Donald et al., 2002). Moreover, species from the Asian steppes and the Mediterranean semi-deserts benefited from the opening of the landscape for agriculture and spread over Europe (Donald et al., 2002).

∗ Corresponding author. Present address: Koen Overmars Consultancy, Jansveld 35A, 3512BE Utrecht, The Netherlands. Tel.: +31 06 48010758. E-mail addresses: [email protected], [email protected] (K.P. Overmars). 1470-160X/$ – see front matter © 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ecolind.2012.11.006

Traditional agricultural practises resulted in highly diverse farmland landscapes (Laiolo et al., 2004), consisting of open areas with extensive grazing or arable farming in combination with shrubs and woodlands. Nowadays a majority of Europe’s biodiversity is associated with agricultural land (Reif et al., 2008; Pocock, 2010). Examples of agricultural landscapes with high levels of biodiversity are the large, extensively used grasslands in parts of northern Europe, silvo-pastoral systems in southern Europe (Paracchini et al., 2008) and extensive, often traditionally managed farming systems in mountainous areas (MacDonald et al., 2000). Generally, extensive agricultural systems are important to the biodiversity heritage of Europe. The area of extensive agricultural systems, such as semi-natural grassland, has decreased significantly during the last century, mainly due to conversion of semi-natural grassland into cropland, intensification, and, since the 1950s, abandonment of agricultural land, leading to encroachment by shrubs and trees (Laiolo et al., 2004; Sjödin et al., 2008). Simultaneously, the biodiversity associated with these semi-natural grasslands

K.P. Overmars et al. / Ecological Indicators 37 (2014) 186–198

and extensive agricultural areas decreased dramatically, which is indicated, for example, by decreasing farmland bird populations (Donald et al., 2002) and plants species (Firbank et al., 2008). The Common Agricultural Policy (CAP) of the European Union (EU) has long contributed to both intensification and abandonment of agricultural land (Donald et al., 2002), thereby contributing to the decrease in biodiversity in agricultural areas. Intensification results in large-scale agricultural production systems, with only a few species tolerant to high levels of inputs, monotonous landscape and other disturbances, whereas abandonment results in a decrease in species associated with farmland and an increase in those associated with forests and shrubland (Laiolo et al., 2004; Firbank et al., 2008). Understanding these processes is key to assessing the impacts of changes in the CAP on farmland biodiversity. Since the 1980s, intensification and abandonment have been recognised as threats to biodiversity on agricultural land in the context of CAP. Farmland biodiversity is recognised as a public good contributing to the ecosystem services that the agricultural system is providing. This includes the sustainability of the agricultural system itself as well as many other services including landscape, recreational services and clean water. Concrete examples are pollination of crops and contribution to pest control (Tscharntke and Brandl, 2004). High biodiversity is often related to higher carbon sequestration, lower erosion risk and higher production (Bullock et al., 2007). In the CAP reforms of the past two decades policies were adopted to counteract these processes. The most important policies are so-called agri-environment schemes (AES) and regulations to retain green structures within the agricultural landscapes. These measures and new, related measures, such as the ‘greening’ of the first pillar of the CAP (i.e. mandatory measures regarding ecological focus areas, crop diversification, and preservation of permanent grassland), are part of current discussions on a CAP reform for the 2014–2020 period (EC, 2010). However, the effectiveness of such measures within the context of continuing intensification of the surrounding agricultural areas is still under debate (Marshall and Moonen, 2002; Berendse et al., 2004; Grashof-Bokdam and Van Langevelde, 2005). Biodiversity indicators suitable to assess the impacts of policy measures on biodiversity in agricultural areas in a comprehensive manner are essential. Biodiversity, generally defined as the variety of all forms of life, has many aspects. At the scale level of this study (Europe) biodiversity certainly includes both species richness and abundance. However, there is no overarching indicator for biodiversity (Gregory et al., 2005). The choice and design of an indicator depends on the purpose of a study. Choices should be made with respect to what the indicator is meant to reflect: the state of a specific species or biodiversity in general; whether the indicator is needed for monitoring purposes or for ex-ante policy evaluation; and whether or not the indicator should be responsive to environmental pressures. Furthermore, the spatial and temporal scale is important in designing an indicator (Gregory et al., 2005; EEA, 2007). Generally, indicators must have certain basic characteristics. They should be sensitive to the impacts addressed, representative, simplify information, be easily understood and policy relevant. Gregory et al. (2005) give a wider set of characteristics important to biodiversity indicators. For this study, the objective was to develop an indicator to evaluate biodiversity effects of land use changes on agricultural land (e.g. more intensive use of inputs, or extensification in inputs or changes in number of cattle per ha), with a special focus on the CAP. Current indicators are not fully suited to this specific purpose, since their initial objectives were different. The focus of many indicators is on biodiversity in natural habitats, and they are unsuited to express the specific character of biodiversity in agricultural areas. Some indicators are primarily based on pressures on natural area and, therefore, lack a connection with species representative of

187

agricultural areas (Alkemade et al., 2009). Many studies focus on one species group, for example, on birds (Donald et al., 2002; Scholefield et al., 2011). Many indicators (e.g. in Scholefield et al., 2011) do not have the spatial resolution necessary in policymaking around the CAP, where sub-national or regional information is needed. The EU uses the European Farm-land Bird Index (EFBI) as a Structural and Sustainable Development Indicator. In this approach identified farmland bird trends are used as a proxy for wider biodiversity trends on farmland (Butler et al., 2010). The spatial resolution of this indicator is the country level and one species group (birds) is included. However, an indicator to evaluate effects of the CAP on biodiversity on agricultural land should be representative of all biodiversity in agricultural areas, not just of one species group. Furthermore, this indicator should be easily understood, in order for it to be useful to policymakers. The indicator has to be linked to the pressures that influence biodiversity, since potential policy measures are aimed to influence these pressures in order to support biodiversity; these policies are not aimed directly at biodiversity itself. Responsiveness to pressures also enables ex-ante evaluation by modelling those pressures. To be able to use the indicator for monitoring and assessment purposes of EU policies, it is important that the indicator consistently covers the entire territory of the EU and that the spatial resolution matches the effects of the policy measures under study. This paper presents the methods for developing an indicator specifically aimed at biodiversity in agricultural areas. The indicator is targeted for ex-ante evaluations of the CAP, but may serve other policy evaluations, for example, of the EU biodiversity strategy (EC, 2011). The indicator is based on relations between species and the pressures of land cover, land-use intensity and fragmentation. The indicator covers the 27 countries of the EU (EU27) and has a resolution of 1 km. The indicator has been applied to the situation of the year 2000 by detailing species occurrence data with pressure data of the year 2000.

2. Data and methods 2.1. Overview The indicator has been based on data from the BIOSCORE project (Delbaere et al., 2009; Louette et al., 2010) on species occurrence and their sensitivity to a variety of pressures. The main data used are species lists with sensitivity scores (not, low, medium, high) for 35 environmental variables (e.g. land use, patch size, soil acidity). Coarse spatial data on species occurrence was used (Section 2.2) and combined with detailed, spatially explicit data (1 km grid) on pressures (Section 2.3). The most important determinant of biodiversity is habitat, which is here represented by land cover. Changing the land cover in a undesired direction for biodiversity is considered as a pressure. We used the CORINE Land Cover 2000 database (CORINE). Originally, CORINE has a resolution of 100 m. Here we use an adapted version (Verburg et al., 2006) with a resolution of 1 km. This data set represents the dominant land cover and does not distinguish between intensive and extensive management of agricultural areas. Therefore, the land-cover data were supplemented with a land-use intensity analysis for arable land and grassland. In this analysis we combined land use intensity data from FSS and point level data on crops from LUCAS to derive relations with spatially explicit data in order to construct an area-covering land use intensity map (more detail in Section 2.3.2). Intensity is pivotal to linking the indicator to a policy such as the CAP, since influencing the level of intensity is an important instrument in this policy. Permanent crops (e.g. fruit trees, olives and vineyards) are also included in the analysis,

188

K.P. Overmars et al. / Ecological Indicators 37 (2014) 186–198

but without having different levels of intensity. The third pressure included is fragmentation. From the BIOSCORE database those species were selected that prefer agricultural areas as their habitat. Per species, the occurrence (including infrequent occurrence) was detailed with the pressure data (land cover, land-use intensity and fragmentation). Subsequently, these maps were combined into the biodiversity indicator (see Section 2.4). An overview of the methodology is depicted in Fig. 1. We compared the biodiversity indicator against the map on High Nature Value farmlands (HNV) (Paracchini et al., 2008). 2.2. Species occurrence and the relationship with environmental pressures The BIOSCORE database integrates data on the impact of a large set of pressures derived from EU policies on various species groups (mammals, reptiles, amphibians, birds, butterflies, vascular plants, freshwater fish and aquatic macro-fauna). Species-specific requirements are used for generating species sensitivity scores, which typically link environmental pressures directly to the ecology of species (Louette et al., 2010). The sensitivity scores in BIOSCORE are based on literature. As it is unfeasible to work with all animal and plant species occurring in Europe, the BIOSCORE database contains a selection of species, representing geographical and environmental variety within Europe. The selection of species is described in Louette et al. (2010) and Delbaere et al. (2009). The BIOSCORE project used a focal species approach, meaning that a set of species was selected of which the management and conservation potentially could be effective for most of the other species occurring within the same landscape. The database contains information on habitat suitability and sensitivity to environmental pressures for over 1000 species.

Within the BIOSCORE project, data on the potential spatial distribution of these species was also indicated. For birds, species occurrence was available from the EBCC atlas (Hagemeijer and Blair, 1997), for mammals, amphibians and reptiles data were available from various sources (Gasc et al., 1997; Mitchell-Jones et al., 1999; IUCN, 2006a,b,c; Linnell et al., 2007; Temple and Terry, 2007). Data on vascular plants was based on a study by Jalas and Suominen (1972–1999). All species occurrence data was available in 50 km × 50 km blocks. We based the indicator on terrestrial vertebrates and vascular plants included in the BIOSCORE database that depend on open grassland or arable land, and on species that have a majority of their habitat in agricultural land. Butterflies and dragonflies were not included because data on the distribution did not have the spatial coverage necessary for this analysis. Fish and aquatic macro-fauna were not included because the analysis of changes in land cover and intensity focused on terrestrial changes only. The following criteria were used for selecting the species: For each vascular plant, the database would indicate whether land-cover types based on CORINE provided a suitable habitat. The area per land cover type suitable for a certain species in the EU27 was calculated form the CORINE database. With this information plants that have a majority of their habitat in agricultural land were selected. For each terrestrial vertebrate, BIOSCORE provided information on the habitat suitability of each CORINE land-cover type. Overall (average) suitability was calculated, by giving the classes a value, for agricultural land and other land. Subsequently, terrestrial vertebrate species were selected that overall were more suited to agricultural land than to land with other types of cover. Based on these criteria, 34 bird species, 12 mammal species, 12 reptile species, 13 amphibian species and 61 vascular plant species were selected. For this list of species we consider that agricultural area is important for their survival and changes in agricultural land use form a potential threat.

Fig. 1. Schematic overview of the method to calculate the biodiversity indicator.

K.P. Overmars et al. / Ecological Indicators 37 (2014) 186–198

189

Fig. 2. Sum of the species occurrence maps for the 132 species included in the analysis per 50 km grid cell.

The final selection of species was compared with other literature to see if all species considered to be important for farmland biodiversity were included (see Discussion). The sum of the original spatial distribution maps from BIOSCORE of the selected species is presented in Fig. 2. 2.3. Pressures For each species considered, the BIOSCORE database indicates the habitat suitability of each land-cover type following the CORINE land-cover map and the species sensitivity to 68 environmental pressures. We included pressures that would affect the selected species that are relevant in agricultural land and that are related to land cover and land-use intensity (Table 1). Below, the three pressure groups as used in this study are presented in detail. Each individual pressure is summarized as a map that indicates if a species can or cannot occur according to this pressure. 2.3.1. Land cover Land cover is included as a pressure in a way that is consistent with the method of species selection (Section 2.2). For each species, suitability scores per land-cover type indicate whether a

land-cover type provides a highly suitable, suitable, moderately suitable or unsuitable habitat. To simplify calculations we assumed that species occur in land-cover types that are suitable or highly suitable, consistent with assumptions made in earlier studies using the BIOSCORE database (Delbaere et al., 2009; Eggers et al., 2009; Louette et al., 2010). A simplified version of the original CORINE land cover was used. The 44 level 3 classes in CORINE were aggregated thematically into 14 new classes, and the resolution was aggregated from 100 m to a 1 km grid for the EU27 (Verburg et al., 2006). One land use type is assigned to each cell. To translate the suitability scores from the extended list of land-cover types in the BIOSCORE database into suitability scores for the simplified version, the most suitable landcover types of the extended list were used. So, if in the extended list a species is highly suitable for land use type A and suitable for land use type B this is translated into highly suitable in case the two are aggregated thematically. Semi-natural areas, as classified by CORINE, were not included in this analysis. These areas area are not considered farmland in CORINE, though some of these areas are in (extensive) agricultural use, mainly as grazing areas. As a result, some extensive agricultural areas with important biodiversity values were not included in our study.

190

K.P. Overmars et al. / Ecological Indicators 37 (2014) 186–198

Table 1 Pressures related to agriculture. Pressure group in this study

BIOSCORE pressure

Explanation

Species group affected

Land cover

Land cover

Suitability of land-cover types as habitat

Fragmentation

Fragmentation

Intensity

Disturbance

Intensity

Harvesting

Intensity

Toxic pollutants

Intensity Intensity

Trampling N input

Disintegration of suitable habitat caused, e.g., by infrastructure, clearcutting Disturbance (e.g. from farming operations, hunting, recreation) Harvesting, mowing and other agricultural operations during breeding period High concentrations of, e.g., persistent organic pollutants, heavy metals, pesticides, alkaline compounds Trampling by livestock Increase in nutrients in the form of nitrogen

Vascular plants, amphibians, reptiles, birds, mammals Amphibians, reptiles, mammals

2.3.2. Land-use intensity Construction of the land-use intensity map Databases of land use normally focus on representing land cover based on remote sensing observations and lack information on land-use intensities (Verburg et al., 2011). To represent the intensity of agricultural practices, a land-use intensity map was constructed. Land use type arable in the land-cover map (Section 2.3.1) was re-classified into 3 intensity classes of agricultural management of arable land and grassland in 2 classes of grassland. Temme and Verburg (2011) proposed to combine European level databases to construct land-use intensity maps using separate methodologies for arable land and grassland. Building on this approach, a land-use intensity map was constructed for the entire EU, which is described in detail below. Additionally, the approach was modified in such a way that the information on intensity would be better tailored to be used in the biodiversity indicator. As an indicator for the intensity of agricultural land management, nitrogen application was selected. Nitrogen application is highly relevant to agro-biodiversity (e.g. Kleijn et al., 2009; Bobbink et al., 2010) and is commonly used as a land-use intensity indicator (Herzog et al., 2006). However, there are no EU-wide data, at high spatial resolutions, showing the spatial distribution of nitrogen application. Data at the highest spatial resolution available to us were those on sub-national administrative units (NUTS2/3 level). For each administrative unit, nitrogen input levels are reported per crop type collected within the Farm Structure Survey (FSS). Another European-level database, the Land Use/Cover Area frame statistical Survey (LUCAS) (Bettio et al., 2002; Jacques and Gallego, 2005), provides point-based observations of crop types from 2003 and 2006, for about 150,000 sample points across agricultural areas in the EU. Each point of the LUCAS data set was assigned the crop-specific nitrogen application rate reported in the FSS data set for the corresponding administrative unit, assuming that variation in nitrogen application within an administrative unit may be approximated by the cropping pattern. Nitrogen application rates were then classified in three classes: low (<50 kg/ha); medium (50–150 kg/ha) and high (>150 kg/ha) based on the relevance of these nitrogen application levels for biodiversity (Kleijn et al., 2009). Using the LUCAS sample locations as individual observations with a specific intensity class, the spatial diversity of these intensity classes was explained by a set of environmental and socioeconomic locational factors (drivers) using multinomial regression. The estimated regression models were used to predict the three intensity classes on all locations classified as arable land, using the socio-economic and physical conditions of these locations. Locational factors included are topographic conditions, soil and climate conditions, population densities and accessibility. A list of all factors included is provided by Temme and Verburg (2011) and Verburg et al. (2006). The regression models were estimated for all countries, separately, to account for country-specific relations

Birds Birds Birds

Birds Vascular plants

between agricultural intensity and locational factors, often originating from different cultural–historical development pathways. For the five countries without LUCAS data (Czech republic, Slovakia, Hungary, Romania and Bulgaria), we used regressions estimated from neighbouring countries with comparable agricultural practices. The regression models provided probability maps for the three arable land intensity classes. A hierarchical procedure was used to allocate the total area of each land-use intensity class, as reported at the level of administrative units to individual pixels (at 1 km grid). First, the area in the highest intensity class was allocated to the pixels with the highest probability for that class. Second, the least-intensive areas were allocated to the pixels with the highest probabilities for that class. The remainder of the locations was allocated to the medium intensity class. For grassland a different approach was taken, similar to Temme and Verburg (2011). For the LUCAS observations of grassland, the nitrogen input was estimated based on local stocking densities. Stocking densities were derived from the livestock maps of Neumann et al. (2009). Data on cows was used as farming of these animals forms the most important land-based animal husbandry in Europe. We assumed a uniform annual quantity of 100 kg N/ha per cow (Van der Hoek, 1998), based on total N intake minus the N in animal products (e.g. milk) and reclassified the observations into two classes: intensive grassland with >50 kg N/ha and extensive grassland with <50 kg N/ha. Similar to the procedure for arable land, country-specific logistic regression models were estimated and used for downscaling the areas of the different intensity classes within the administrative units to the 1 km × 1 km pixel map. Using the land-use intensity map as an indicator of pressure for biodiversity Agricultural intensity is determined by the level of inputs and outputs of an agricultural system (Lambin et al., 2000; Shriar, 2000). The BIOSCORE database subdivides intensity into inputs of nitrogen and pesticides, and disturbances related to grazing and agricultural operations. In the BIOSCORE database, several plants and birds are sensitive to intensity-related pressures. The data used in this study as described above, however, is solely based on N input in the form of manure and fertilizer. As N application is strongly related to other aspects of land-use intensity (Donald et al., 2001; Kleijn et al., 2009), the intensity map described above is assumed to be a proxy for the intensity related pressures included in BIOSCORE. To determine a sensitivity score for birds, we assumed disturbance, harvesting and trampling to occur on grasslands and disturbance, harvesting, and toxic pollutants on arable lands. For each bird species, the number of pressures to which it reacts and the level of sensitivity were translated into a score indicating if the bird could occur (1) or not (0) in each intensity class of land use. For example, if a bird species is highly sensitive to all pressures, it only occurs in extensive grassland and extensive cropland. A bird species

K.P. Overmars et al. / Ecological Indicators 37 (2014) 186–198

that is moderately sensitive to two of the three pressures active in cropland occurs in extensive and moderately intensive croplands. Per plant species, the BIOSCORE database gives the range of Ellenberg values for nitrogen levels (Ellenberg et al., 1992) at which they may occur. Ellenberg values for nitrogen may be interpreted ´ as the total nutrient supply (Schaffers and Sykora, 2000) and range from 1 (extremely infertile conditions) to 9 (extremely rich conditions). In order to match the BIOSCORE data, the intensity classes from the map were also translated into an Ellenberg range. For arable land, extensive lands are assigned an Ellenberg value of 4–5, moderately intensive lands 6–7, and intensive arable land 8–9. Extensive grasslands have Ellenberg values between 4 and 6, intensive grasslands between 7 and 9. We assume that a species can only occur if its Ellenberg value from BIOSCORE overlaps with the Ellenberg range of the intensity class of the map as described above. 2.3.3. Fragmentation Sensitivity for habitat fragmentation by infrastructure and land cover that does not provide a suitable habitat was included in the BIOSCORE database for mammals, reptiles and amphibians. To model the effect of fragmentation, for each mammal, reptile and amphibian species the habitat was mapped using the complete aggregated land-cover map and the suitability scores for the

191

land-cover types. Subsequently, we made an overlay of this map and a road map (PBL, 2011) and identified continuous patches in the habitat map. Finally, patches that would be too small to support a sufficient population were classified unsuitable for species occurrence. We assumed that the species considered would need a continuous area of at least 100 km2 (based on Alkemade et al., 2009). This figure is far less than the >10,000 km2 required for a complete species composition, including large herbivores and top predators (Verboom et al., 2007; Alkemade et al., 2009). However, we assumed patches of continuous habitat of >100 km2 to be sufficiently large to harbour all the species considered here. 2.4. Determining relative species richness 2.4.1. Species number For each species, the maps of land cover, land use intensity and fragmentation were combined into a pressures map, using the BIOSCORE data containing the relation between theoretical occurrence and these pressures. This pressures map is 0 in case one or more of the pressures indicate an unsuitable situation for the particular species and 1 if none of the three pressures is limiting. Subsequently, the occurrence maps at 50 km resolutions were downscaled to 1 km resolution maps indicating whether a

Fig. 3. Species numbers based on 132 species dependent on or mainly occurring in agricultural areas in Europe per 1 km grid cell (agricultural areas are indicated by colour; non-agricultural areas in grey).

192

K.P. Overmars et al. / Ecological Indicators 37 (2014) 186–198

species could occur at a certain location (1) or not (0), by combining the pressure maps and the species occurrence maps. Subsequently, the 132 downscaled maps per species were added together, resulting in a 1 km resolution species richness map. Having 61 vascular plants and 71 species of animals these have approximately the same weight in the analysis. The basic result of this calculation is a species number per (agricultural) grid cell. However, the method included 132 focal species that are assumed to represent the complete species richness of agricultural land. To indicate how the current situation relates to the species number that potentially may occur, in a further step we also present a relative species richness. 2.4.2. Relative species richness The relative species richness in this study was defined as the number of species predicted to occur in a 1 km square on the basis of both the distribution map and pressures map divided by the number of species predicted to occur in that square on the basis of the distribution map alone (Fig. 2) in each grid cell/pixel. With this indicator locations can be compared in terms of actual biodiversity relative to the potential in that specific location. This may be interpreted as the relative influence of the pressures on the number of species.

3. Results 3.1. Biodiversity map for the year 2000 The methodology is applied to construct biodiversity reference maps for the year 2000. Fig. 3 shows the absolute number of simulated species. Fig. 4 shows the relative species richness (see Section 2.4). The difference in pattern between potential occurrence (Fig. 2) and the relative species richness is largely the result of land cover and landuse intensity. Intensive arable farming in northwestern Europe negatively influences biodiversity. A similar effect occurs in irrigated areas in Mediterranean countries. Parts of Mediterranean countries (including France) and the new EU Member States have areas with high relative species richness. Grassland generally has higher biodiversity than arable land. Fig. 5 presents the relative species richness of Fig. 4 in a histogram, identifying the land cover and land-use intensity classes. This figure shows the variability of relative species richness, summarised per type of land cover and land-use intensity. Many species prefer extensive grasslands and permanent crops (e.g. orchards) as their habitats. In areas with these land-use types, 70–90% of the species identified as potentially occurring, were modelled to be

Fig. 4. Relative species richness (agricultural areas are indicated by colour; non-agricultural areas in grey).

K.P. Overmars et al. / Ecological Indicators 37 (2014) 186–198

Fig. 5. Histogram of the relative species richness map for 2000, stratified into land-cover types and land-use intensities.

Fig. 6. Overlay of relative species index of 75% or higher and HNV areas in agricultural land.

193

194

K.P. Overmars et al. / Ecological Indicators 37 (2014) 186–198

present. The agricultural areas with the least biodiversity are the intensively managed arable lands. 3.2. Comparison with High Nature Value farmlands We compared the relative species richness map with that of High Nature Value (HNV) farmland (Paracchini et al., 2008). HNV farmlands include agricultural land with great species and habitat diversities, those that contain species of European conservation concern, or both. The HNV methodology of this dataset is quite different from the biodiversity indicator presented in this and partly based on expert knowledge. Our objective here is to see the agreement in the maps, not to be a validation. In our study, we used the wider definition of HNV areas (constructed from four data layers: CORINE Land Cover selection, NATURA2000, Important Bird Areas, Primary Butterfly Areas) (Paracchini et al., 2008). An overlay was made of the HNV map and a sub-selection of the relative species richness map that included only the areas with a relative species richness greater than or equal to 75% (Fig. 6). In this study, we only included the HNV areas classified as agricultural areas in the land-cover map. This means that 49% of HNV areas were taken into account. The remainder is not classified as agricultural land, but are mainly semi-natural areas (in Scotland and Spain), forests, heath lands and moorlands. Of all agricultural land included in the analysis 18% is HNV farmland (Fig. 6; green and red colours). The average relative species richness in these HNV areas is higher (58.5%) than on the remaining agricultural land (45.4%). For 33% of the HNV areas the predicted relative species richness is greater than or equals 75% (green colour). For 15% of the total agricultural area the relative species richness is greater than or equals 75% (green and blue colour). In total, 39% of areas with a relative species richness greater than or equal to 75% is situated in areas of High Nature Value. 4. Discussion and conclusions 4.1. Results 4.1.1. Explanation of results Based on the presented map of relative species richness, biodiversity in agricultural areas is lower in northwestern European countries and higher in those in southern and eastern Europe (Fig. 4). As stated before, semi-natural areas, such as those in United Kingdom, Greece, Spain, northern Scandinavia and many areas in mountainous regions, were not included in this analysis. The geographic pattern of the relative species richness resulted from the three pressures of land cover, land-use intensity and fragmentation, contributing 45%, 53% and 2%, respectively, to the exclusion of species from the original (50 km grid) species occurrence maps. These figures were derived by leaving out one driver at the time and recalculate the average relative species richness. Grassland has a higher relative species richness than arable land, and extensively managed land has a higher biodiversity than intensively managed land (as shown in Fig. 5). Although the southern and eastern European countries have a relatively larger share in arable land, the land-use intensity is medium to low for arable land and low for grassland. Therefore, the relative species richness is higher in these areas. The method does not include changes in the potential occurrence of species (Fig. 2). In case, for example, species would migrate northwards because of climate change this could not be included in the way the species richness indicator is set up now. Including new or modelled potential occurrence maps would solve this. The presented results should therefore be interpreted with care

and are valid for the coming years and not for analyses far into the future.

4.1.2. Validation and monitoring The relative species richness presented is constructed from available data and literature on pressures and ecological relations. Many of these sources are validated. However, reusing them in a new method does not necessarily have to have the same level of validation. Therefore, an independent validation of the results of this method would be a valuable addition. From the perspective of data availability a practical option is to do this separately for each species group, for parts of the covered area. Moreover, it is important that in case a pressure changes because of a policy change (and therefore a management change), and the relative species richness indicator changes, that this actually happens in reality. A monitoring system that is included in a policy would be a good option to do this. This information would also give a good indication of the validity of the method.

4.1.3. Comparison with High Nature Value farmland The map of High Nature Value farmland and the areas in the relative species index map that have a value of 75% or higher both divided the area into parts with high and low levels of biodiversity. The map in Fig. 6 is an overlay of the two maps. Postulated the comparison is not meant to be a validation – the maps have completely different objectives – it is interesting to explain the differences, since they both try to shed light on what areas are important for biodiversity. The areas classified as HNV without having a high level of biodiversity may be explained in several ways. First, in addition to areas with high biodiversity, the HNV map also includes areas with one or a few species of particular conservation importance. Second, the areas may have a high biodiversity in absolute terms (as modelled in Fig. 3), and thus be identified as HNV, however, this overlay was constructed with the relative species index. Third, areas without a high level of biodiversity may have been classified HNV due to an error in the HNV map or other reasons. Fourth, the biodiversity indicator may contain errors due to misclassification, for example, because modelling of land-use intensities was based on extrapolation of statistical relations, with its limitations in explaining the observed values. Despite these considerations, the map does show interesting differences between countries. In general, eastern European countries have considerable areas with high levels of biodiversity (>75%) that overlap with HNV areas. However, a large proportion of HNV areas were not classified as having high levels of biodiversity. An explanation could be that, especially in eastern Europe, certain cells of the HNV map only contain a small percentage of High Nature Value farmland. An exception in eastern Europe is Latvia. In Latvia large areas are modelled as having high levels of biodiversity without them being HNV areas. Therefore, for Latvia, either we overestimated the biodiversity levels, or the HNV classification was too conservative. In the United Kingdom the dominant category contains areas with high biodiversity levels without them being considered areas of High Nature Value. For the United Kingdom, HNV areas are mainly semi-natural areas not included in this study. The areas with modelled high levels of biodiversity without them being of High Nature Value, are areas adjacent to these semi-natural (and HNV classified) areas. In the Netherlands, most HNV areas do not have high relative species richness. In these HNV areas a particular species group is of interest: meadow birds. Southern Europe (especially Spain and southern France) presents a very diverse picture. The general conclusion is that the relative species richness map and the HNV map were constructed in very different ways and show

K.P. Overmars et al. / Ecological Indicators 37 (2014) 186–198

different results. This is partly due to uncertainties and to different objectives related to these indicators. 4.2. Methods 4.2.1. Selection of species The presented method for constructing a combined indicator for biodiversity in agricultural areas includes 132 species (plants and vertebrates). The first question here is: how representative are the selected species? Comparing our list of species with those considered in other studies on farmland biodiversity (Donald et al., 2006; Stoate et al., 2009; Fischer et al., 2011; Scholefield et al., 2011), we conclude that our list includes similar species of farmland biodiversity. The analysis included only five species groups and combined both flora and fauna, which remains a limited sample of biodiversity. Nevertheless, the number of species groups included is considered sufficient to construct an indicator that is representative of all biodiversity, for two reasons. First, the species in the BIOSCORE database are focal species, thus, their response to changes in pressures is representative of all species living within the same landscape (Delbaere et al., 2009; Louette et al., 2010). Second, the species included are those that are high up in the food chain. The presence of predators is a good indication of total biodiversity (Stoate et al., 2001). We believe that when in future more and better data on species occurrence will become available, the indicator could be improved by including this new information. In this respect this analysis should be seen as a methodological experiment that could be improved in the future. Nevertheless, the results do give a first impression. 4.2.2. Spatial distribution of species included in the analysis For the method to work properly, it was preferable that, for all regions, in all 50 km × 50 km blocks, a similar percentage of actually occurring species would be included in the study. For the modelled species number of Fig. 3 this was even a prerequisite. In Fig. 2, the potential number of species – the sum from all species occurrence maps – is presented. However, in the areas on the edges of the map, numbers of species are clearly lower (northern Scandinavia, Ireland, southern parts of the Mediterranean countries). This may have been caused by actual differences in total numbers of farmland species in these areas. Another explanation could be that many of the species included in the study occur in the central parts of Europe, with not all of them also occurring in the areas on the map edges. The question is whether this is being compensated, sufficiently, with species that occur specifically in these edge areas. To overcome the spatial differences in the percentages of actual species numbers included in the study, we constructed the relative species number (Fig. 4). With this indicator two locations may be compared for their biodiversity performance relative to the potential, whereas a comparison of actual species richness is not possible. In scenario analyses, however, this method may lead to biased results. When a location would have a low number of species, a change in one species could cause a significant change in relative species richness, when calculated according to this method. 4.2.3. Semi-natural areas As said in the methods section, semi-natural areas were not included in this study, although these areas are used for extensive agriculture and therefore contribute significantly to biodiversity in agricultural areas. Semi-natural areas include Natural grasslands, Scelophyllous vegetation (woody plants with small leathery evergreen leaves, mainly Mediterranean), Transitional woodland shrub and Land principally occupied by agriculture, with significant areas of natural vegetation.

195

In specifying the HNV farmland areas, semi-natural areas are one of the most important land-use types (Paracchini et al., 2008), and 51% of HNV farmlands are actually situated in these seminatural areas. In Europe, semi-natural areas are found mainly in the United Kingdom, Greece, Spain, northern Scandinavia and mountainous regions. To include semi-natural areas in the biodiversity indicator may alter the results. The difficulty with including semi-natural land in this analysis was that semi-natural areas are a mixture of land with and without farming activities: Some of the land classified in CORINE as seminatural area is not counted as utilised agricultural area (UAA) (EEA, 2009). We excluded the semi-natural areas because in many economic models (e.g. CAPRI) areas outside UAAs are not included. FSS data indicate that farms with High Nature Value land are more dependent on CAP support for their incomes, and that a larger part of their support comes from measures in the second pillar of the CAP (Osterburg et al., 2007). In theory, these lands may be abandoned or their use intensified (conversion to grassland or arable land) when the CAP changes. The area included in this study contains the major part of the agricultural area that generates most of the production and receives most CAP funding, especially in first-pillar funds. In a CAP scenario analysis, from the perspective of CAP expenditure and the effects of changes in subsidy distribution, this method gives a good indication of the effects on biodiversity. 4.2.4. Scale issues This study has focused on the large scale, both in extent (EU27) and spatial resolution (1 km grid). Landscape heterogeneity and landscape elements were not taken into account. The resolution of these landscape characteristics was beyond the level of detail of this study. This is mainly driven by data availability and practical reasons. It is difficult/not possible to include all landscape features relevant to agro-biodiversity in a European-wide assessment. However, landscape and field structures largely determine biodiversity. From a policy perspective, these levels are of importance since many policies are implemented on this level, for example, the agri-environment schemes. The indicator used in this study, however, was aimed at assessing the impact on biodiversity from three general pressures associated with large-scale policy-making (e.g. the CAP). Such indicators are important to improve environmental management and policy-making, as these rely on effective and simple indicators on a large spatial scale (Billeter et al., 2008). The presented indicator is meant to be used for large-scale analysis (EU level). Detailed studies may provide complementary information to cover the level from field to landscape. EU wide coverage is important since in the targeted policies (CAP) measures in one place may have an effect on another location. Within the EU this is especially important and therefore we included the whole of the EU. Actually the polices have also an influence outside the EU which should be taken into account. 4.2.5. Species richness vs abundance The main reason for using species richness in this indicator is data availability. However, biodiversity is more than species richness alone, for example, it also relates to abundance. Such data on abundance on EU scale is not easily available. Nevertheless, we considered that species richness, by itself, also would provide valuable information for assessing biodiversity. 4.2.6. Sensitivity to classes In the method at several points in the method classes were made. Land use classes from CORINE were merged and the corresponding suitability for specific species was therefore also aggregated. Furthermore, intensification was classified in classes where in reality this is a continuum. Here we used the best

196

K.P. Overmars et al. / Ecological Indicators 37 (2014) 186–198

information available. Not many data exist to map land use intensity as a continuous variable. Moreover, detailed information on the effects of intensity on species is also scarce. In general, we do know that many species do not strive in intensively managed agricultural landscapes. A reliable regression analysis, for example, between species abundance and Nitrogen input is often not possible to carry out. The approach taken in this study suits best with the available data. 4.2.7. Alternative presentation of the indicator The information from this analysis may be presented in alternative ways, depending on the objective for using the indicator. Presenting it as a relative species richness map is mainly targeted to assessing the impact of large-scale land-use policies, such as the CAP. Alternatively, more detail on species, species groups, or conservation status could be presented, although in such cases the indicator carries a certain risk of devaluated representativeness. In this study all species included received the same weight. Alternatively we could have given all taxa the same weight, for example. However, each approach has its advantages and disadvantages. In this study equal weighing at the species level was chosen because it is straightforward, the flora and fauna are already (almost) equally represented and the fauna consists of four groups. Since the full relative species richness includes 132 species, the indicator is considered to be representative of total biodiversity. Another interesting indicator would be to visualise the locational importance to European biodiversity, for example, by giving a location a higher value in case an occurring species is rare, or when its conservation status is high. Such an indicator could be useful in policy-making to identify priority areas, or the most cost-effective policy option. 4.3. Use in policy-making The objective of presented study is to provide a methodology for policy analysis with respect to farmland biodiversity at EU27 level. The indicator is based on potential occurrence of species and their relation with current pressures (land cover, land-use intensity, fragmentation). These pressures can be modelled in ex-ante policy evaluations (e.g. see Verburg et al., 2008), for example modelling the economic effects of policies on agricultural production and the relation land use and land use intensity effects. Coupling these modelled pressures to the potential species map will result in ex-ante assessments of biodiversity in agricultural areas. In that case the presented map may serve as the reference point with which the scenario alternatives for future situations are compared. This paper presents a first map of agricultural biodiversity with the EU as extent at a relatively high spatial resolution (1 km), using data on potential species occurrence and the pressures that limit this occurrence. There are many options to improve this map, especially by including more and detailed species occurrence data. This method and the resulting map may be an important resource for ex-ante assessments of future policies and for future studies of the dynamics of agricultural biodiversity in Europe. Acknowledgements This study was conducted as part of the EURURALIS 3.0 project commissioned by the Dutch Ministry of Agriculture, Nature and Food Quality (currently part of the Ministry of Economic Affairs, Agriculture and Innovation). For the work on land-use intensity some additional resources were provided by the EU FP7 project VOLANTE. We would like to thank all people involved for their contributions to this research. We would like to thank Annemieke Righart for editing the English.

Appendix A. List of species included in the indicator Birds (34) Alauda arvensis Alectoris chukar Athene noctua Carduelis cannabina Carduelis carduelis Ciconia ciconia Circus cyaneus Circus macrourus Coturnix coturnix Crex crex Cygnus olor Emberiza citrinella Falco tinnunculus Falco vespertinus Gallinago gallinago Glareola nordmanni Glareola pratincola Haematopus ostralegus Hirundo rustica Limosa limosa Miliaria calandra Milvus milvus Motacilla flava Otis tarda Passer montanus Perdix perdix Pluvialis apricaria Pterocles alchata Saxicola rubetra Tringa totanus Turdus viscivorus Tyto alba Upupa epops Vanellus vanellus Mammals (12) Meles meles Erinaceus europaeus Erinaceus roumanicus Lepus europaeus Apodemus agrarius Capreolus capreolus Cricetus cricetus Micromys minutus Microtus arvalis Microtus multiplex Crocidura russula Crocidura suaveolens Reptiles (12) Pseudopodus apodus Timon lepidus Vipera ammodytes Vipera latastei Algyroides nigropunctatus Chalcides striatus Elaphe quatuorlineata Elaphe situla Euleptes europaea Lacerta bilineata Lacerta trilineata Lacerta viridis

K.P. Overmars et al. / Ecological Indicators 37 (2014) 186–198

Amphibians (13) Alytes obstetricans Bombina bombina Bufo bufo Hyla arborea Pelobates fuscus Pelodytes punctatus Rana dalmatina Rana lessonae Rana ridibunda Triturus alpestris Triturus cristatus Triturus helveticus Triturus vulgaris

197

Herniaria glabra Lotus corniculatus Polygonum amphibium Ranunculus muricatus Ranunculus sardous Reynoutria sachalinensis Rumex acetosella Rumex thyrsiflorus Sagina procumbens Silene vulgaris Stellaria neglecta Trifolium repens

References Vascular plants (61) Arenaria biflora Cerastium cerastoides Clematis integrifolia Leucanthemum vulgare Pulsatilla montana Ranunculus acris Ranunculus bulbosus Ranunculus illyricus Ranunculus paludosus Ranunculus psilostachys Rumex salicifolius Thalictrum flavum Trollius europaeus Atriplex micrantha Atriplex oblongifolia Atriplex tatarica Cerastium tomentosum Chenopodium ambrosioides Chenopodium botrys Chenopodium giganteum Chenopodium opulifolium Consolida pubescens Dianthus tripunctatus Fumaria densiflora Gypsophila scorzonerifolia Oxybaphus nyctagineus Papaver argemone Papaver dubium Papaver hybridum Papaver rhoeas Papaver somniferum Phytolacca americana Polycarpon tetraphyllum Polycnemum majus Polygonum graminifolium Polygonum maritimum Silene dichotoma Silene fuscata Silene gallica Silene muscipula Silene noctiflora Spergula arvensis Arenaria serpyllifolia Bromus hordeaceus Cerastium glomeratum Chenopodium bonus-henricus Daucus carota Elymus repens Fallopia dumetorum

Alkemade, R., van Oorschot, M., Miles, L., Nellemann, C., Bakkenes, M., ten Brink, B., 2009. GLOBIO3: a framework to investigate options for reducing global terrestrial biodiversity loss. Ecosystems 12, 374–390. Berendse, F., Chamberlain, D., Kleijn, D., Schekkerman, H., 2004. Declining biodiversity in agricultural lanscapes and the effectiveness of agri-environment schemes. Ambio 33, 499–502. Bettio, M., Délince, J., Bruyas, P., Croi, W., Eiden, G., 2002. Area Frame Surveys: Aim, Principals and Operational Surveys. Building Agri-environmental Indicators, Focussing on the European Area frame Survey LUCAS. EC Report, pp. 12–27. Billeter, R., Liira, J., Bailey, D., Bugter, R., Arens, P., Augenstein, I., Aviron, S., Baudry, J., Bukacek, R., Burel, F., Cerny, M., De Blust, G., De Cock, R., Diekötter, T., Dietz, H., Dirksen, J., Dormann, C., Durka, W., Frenzel, M., Hamersky, R., Hendrickx, F., Herzog, F., Klotz, S., Koolstra, B., Lausch, A., Le Coeur, D., Maelfait, J.P., Opdam, P., Roubalova, M., Schermann, A., Schermann, N., Schmidt, T., Schweiger, O., Smulders, M.J.M., Speelmans, M., Simova, P., Verboom, J., van Wingerden, W.K.R.E., Zobel, M., Edwards, P.J., 2008. Indicators for biodiversity in agricultural landscapes: a pan-European study. J. Appl. Ecol. 45, 141–150. Bobbink, R., Hicks, K., Galloway, J., Spranger, T., Alkemade, R., Ashmore, M., Bustamante, M., Cinderby, S., Davidson, E., Dentener, F., Emmett, B., Erisman, J.W., Fenn, M., Gilliam, F., Nordin, A., Pardo, L., De Vries, W., 2010. Global assessment of nitrogen deposition effects on terrestrial plant diversity: a synthesis. Ecol. Appl. 20, 30–59. Bullock, J.M., Pywell, R.F., Walker, K.J., 2007. Long-term enhancement of agricultural production by restoration of biodiversity. J. Appl. Ecol. 44, 6–12. Butler, S.J., Boccaccio, L., Gregory, R.D., Vorisek, P., Norris, K., 2010. Quantifying the impact of land-use change to European farmland bird populations. Agric. Ecosyst. Environ. 137, 348–357. Delbaere, B., Nieto Serradilla, A., Snethlage, M.E., 2009. BioScore: A Tool to Assess the Impacts of European Community Policies on Europe’s Biodiversity. ECNC, Tilburg, the Netherlands. Donald, P.F., Green, R.E., Heath, M.F., 2001. Agricultural intensification and the collapse of Europe’s farmland bird populations. Proc. R. Soc. Lond. B Biol. Sci. 268, 25–29. Donald, P.F., Pisano, G., Rayment, M.D., Pain, D.J., 2002. The Common Agricultural Policy, EU enlargement and the conservation of Europe’s farmland birds. Agric. Ecosyst. Environ. 89, 167–182. Donald, P.F., Sanderson, F.J., Burfield, I.J., van Bommel, F.P.J., 2006. Further evidence of continent-wide impacts of agricultural intensification on European farmland birds, 1990–2000. Agric. Ecosyst. Environ. 116, 189–196. EC, 2010. The CAP towards 2020: meeting the food, natural resources and territorial challenges of the future – Communication from the Commission to the European Parliament and the Council. COM(2010) 672 final. European Commission, Brussels. EC, 2011. Our life insurance, our natural capital: an EU biodiversity strategy to 2020. In: EU (Ed.), COM(2011) 244 final, Brussels, p. 17. EEA, 2007. Halting the loss of biodiversity by 2010: proposal for a first set of indicators to monitor progress in Europe. Technical report 11/2007. EEA. EEA, 2009. Distribution and Targeting of the CAP Budget from a Biodiversity Perspective. European Environment Agency, Copenhagen. Eggers, J., Troltzsch, K., Falcucci, A., Maiorano, L., Verburg, P.H., Framstad, E., Louette, G., Maes, D., Nagy, S., Ozinga, W., Delbaere, B., 2009. Is biofuel policy harming biodiversity in Europe? GCB Bioenergy 1, 18–34. Ellenberg, H., Weber, H.E., Dull, R., 1992. Zeigerwerte van Pflanzen in Mitteleuropa. Scripta Geobotanica 18, 1–258. Firbank, L.G., Petit, S., Smart, S., Blain, A., Fuller, R.J., 2008. Assessing the impacts of agricultural intensification on biodiversity: a British perspective. Philos. Trans. R. Soc. Lond. B Biol. Sci. 363, 777–787. Fischer, C., Flohre, A., Clement, L.W., Batáry, P., Weisser, W.W., Tscharntke, T., Thies, C., 2011. Mixed effects of landscape structure and farming practice on bird diversity. Agric. Ecosyst. Environ. 141, 119–125. Gasc, J.P., Cabela, A., Crnobrnja-Isailovic, J., Dolmen, D., Grossenbacher, K., Haffner, P., Lescure, J., Martens, H., Martínez Rica, J.P., Maurin, H., Oliveira, M.E., Sofianidou, T.S., Veith, M., Zuiderwijk, A. (Eds.), 1997. Atlas of Amphibians and Reptiles in Europe. Societas Europaea Herpetologica, Muséum National d’Histoire Naturelle & Service du Petrimone Naturel, Paris.

198

K.P. Overmars et al. / Ecological Indicators 37 (2014) 186–198

Grashof-Bokdam, C.J., Van Langevelde, F., 2005. Green veining: landscape determinants of biodiversity in European agricultural landscapes. Landscape Ecol. 20, 417–439. Gregory, R.D., van Strien, A., Vorisek, P., Gmelig Meyling, A.W., Noble, D.G., Foppen, R.P.B., Gibbons, D.W., 2005. Developing indicators for European birds. Philos. Trans. R. Soc. Lond. B Biol. Sci. 360, 269–288. Hagemeijer, E.J.M., Blair, M.J. (Eds.), 1997. The EBCC Atlas of European Breeding Birds: Their Distribution and Abundance. T & A.D. Poyser, London. Herzog, F., Steiner, B., Bailey, D., Baudry, J., Billeter, R., Bukácek, R., De Blust, G., De Cock, R., Dirksen, J., Dormann, C.F., De Filippi, R., Frossard, E., Liira, J., Schmidt, T., Stöckli, R., Thenail, C., Van Wingerden, W., Bugter, R., 2006. Assessing the intensity of temperate European agriculture at the landscape scale. Eur. J. Agron. 24, 165–181. IUCN, Conservation International, NatureServe, 2006a. Global amphibian assessment. IUCN, Conservation International, NatureServe, 2006b. Global mammal assessment. IUCN, Conservation International, NatureServe, 2006c. Global reptiles assessment. Jacques, P., Gallego, F.J., 2005 The LUCAS project—the new methodology in the 2005–2006 surveys, Workshop on Integrating Agriculture and Environment: CAP Driven Land Use Scenarios Belgirate. Jalas, J., Suominen, J., 1972–1999. Atlas Florae Europaeae. Distribution of vascular plants in Europe 1-12. Committee for Mapping the Flora of Europe and Societas Biologica Fennica Vanamo, Helsinki. Kleijn, D., Kohler, F., Báldi, A., Batáry, P., Concepción, E.D., Clough, Y., Díaz, M., Gabriel, D., Holzschuh, A., Knop, E., Kovács, A., Marshall, E.J.P., Tscharntke, T., Verhulst, J., 2009. On the relationship between farmland biodiversity and land-use intensity in Europe. Proc. R. Soc. Lond. B Biol. Sci. 276, 903–909. Klein Goldewijk, K., Ramankutty, N., 2004. Land cover change over the last three centuries due to human activities: the availability of new global data sets. GeoJournal 61, 335–344. Laiolo, P., Dondero, F., Ciliento, E., Rolando, A., 2004. Consequences of pastoral abandonment for the structure and diversity of the alpine avifauna. J. Appl. Ecol. 41, 294–304. Lambin, E.F., Rounsevell, M.D.A., Geist, H.J., 2000. Are agricultural land-use models able to predict changes in land-use intensity? Agric. Ecosyst. Environ. 82, 321–331. Linnell, J., Salvatori, V., Boitani, L., 2007. Guidelines for population level management plans for large carnivores in Europe. A Large Carnivore Initiative for Europe report prepared for the European Commission. Final draft. Louette, G., Maes, D., Alkemade, J.R.M., Boitani, L., de Knegt, B., Eggers, J., Falcucci, A., Framstad, E., Hagemeijer, W., Hennekens, S.M., Maiorano, L., Nagy, S., Serradilla, A.N., Ozinga, W.A., Schaminée, J.H.J., Tsiaousi, V., van Tol, S., Delbaere, B., 2010. BioScore-Cost-effective assessment of policy impact on biodiversity using species sensitivity scores. J. Nat. Conserv. 18, 142–148. MacDonald, D., Crabtree, J.R., Wiesinger, G., Dax, T., Stamou, N., Fleury, P., Gutierrez Lazpita, J., Gibon, A., 2000. Agricultural abandonment in mountain areas of Europe: environmental consequences and policy response. J. Environ. Manage. 59, 47–69. Marshall, E.J.P., Moonen, A.C., 2002. Field margins in northern Europe: their functions and interactions with agriculture. Agric. Ecosyst. Environ. 89, 5–21. Mitchell-Jones, A.J., Reijnders, P., Amori, G., Bogdanowicz, W., Krystufek, B., 1999. The Atlas of European Mammals. Academic Press, London.

Neumann, K., Elbersen, B., Verburg, P., Staritsky, I., Pérez-Soba, M., de Vries, W., Rienks, W., 2009. Modelling the spatial distribution of livestock in Europe. Landscape Ecol. 24, 1207–1222. Osterburg, B., Nitsch, H., Laggner, A., Wagner, S., 2007. Analysis of policy measures for greenhouse gas abatement and compliance with the Convention on Biodiversity. Paracchini, M.L., Petersen, J.-E., Hoogeveen, Y., Bamps, C., Burfield, I., van Swaay, C., 2008. High Nature Value Farmland in Europe – An Estimate of the Distribution Patterns on the Basis of Land Cover and Biodiversity Data. Joint Research Centre, Institute for Environment and Sustainability, Ispra. PBL, 2011. http://geoservice.pbl.nl/website/GRIP/. PBL Netherlands Environmental Assessment Agency, Den Haag. Pocock, M.J.O., 2010. Can traits predict species’ vulnerability? A test with farmland passerines in two continents. Proc. R. Soc. Lond. B Biol. Sci., online. ´ K., Bejbek, V., 2008. Bird-habitat associaReif, J., Storch, D., Volínek, P., Kpastny, tions predict population trends in central European forest and farmland birds. Biodivers. Conserv. 17, 3307–3319. ´ K.V., 2000. Reliability of Ellenberg indicator values for moisSchaffers, A.P., Sykora, ture, nitrogen and soil reaction: a comparison with field measurements. J. Veg. Sci. 11, 225–244. Scholefield, P., Firbank, L., Butler, S., Norris, K., Jones, L.M., Petit, S., 2011. Modelling the European Farmland Bird Indicator in response to forecast land-use change in Europe. Ecol. Indic. 11, 46–51. Shriar, A.J., 2000. Agricultural intensity and its measurement in frontier regions. Agrofor. Syst. 49, 301–318. Sjödin, N.E., Bengtsson, J., Ekbom, B., 2008. The influence of grazing intensity and landscape composition on the diversity and abundance of flower-visiting insects. J. Appl. Ecol. 45, 763–772. Stoate, C., Báldi, A., Beja, P., Boatman, N.D., Herzon, I., van Doorn, A., de Snoo, G.R., Rakosy, L., Ramwell, C., 2009. Ecological impacts of early 21st century agricultural change in Europe – a review. J. Environ. Manage. 91, 22–46. Stoate, C., Boatman, N.D., Borralho, R.J., Carvalho, C.R., Snoo, G.R.d., Eden, P., 2001. Ecological impacts of arable intensification in Europe. J. Environ. Manage. 63, 337–365. Temme, A.J.A.M., Verburg, P.H., 2011. Mapping and modelling of changes in agricultural intensity in Europe. Agric. Ecosyst. Environ. 140, 46–56. Temple, H., Terry, A., 2007. The Status and Distribution of European Mammals. Office for Official Publications of the European Communities, Luxembourg. Tscharntke, T., Brandl, R., 2004. Plant–insect interactions in fragmented landscapes. Annu. Rev. Entomol. 49, 405–430. Van der Hoek, K.W., 1998. Nitrogen efficiency in global animal production. Environ. Pollut. 102, 127–132. Verboom, J., Alkemade, R., Klijn, J., Metzger, M.J., Reijnen, R., 2007. Combining biodiversity modeling with political and economic development scenarios for 25 EU countries. Ecol. Econ. 62, 267–276. Verburg, P.H., Eickhout, B., Meijl, H., 2008. A multi-scale multi-model approach for analyzing the future dynamics of European land use. Ann. Reg. Sci. 42, 57–77. Verburg, P.H., Neumann, K., Nol, L., 2011. Challenges in using land use and land cover data for global change studies. Global Change Biol. 17, 974–989. Verburg, P.H., Schulp, C.J.E., Witte, N., Veldkamp, A., 2006. Downscaling of land use change scenarios to assess the dynamics of European landscapes. Agric. Ecosyst. Environ. 114, 39–56.