Agroforestry creates carbon sinks whilst enhancing the environment in agricultural landscapes in Europe

Contents lists available at ScienceDirect Land Use Policy journal homepage: www.elsevier.com/locate/landusepol Agroforestry creates carbon sinks whilst enhancing the environment in agricultural landscapes in Europe Sonja Kay a,⁎ , Carlo Rega b , Gerardo Moreno c , Michael den Herder d , João H.N. Palma e,f , Robert Borek g , Josep Crous-Duran e , Dirk Freese h , Michail Giannitsopoulos i , Anil Graves i , Mareike Jäger j , Norbert Lamersdorf k , Daniyar Memedemin l , Rosa Mosquera-Losada m , Anastasia Pantera n , Maria Luisa Paracchini b , Pierluigi Paris o , José V. Roces-Díaz a,p , Victor Rolo c , Adolfo Rosati q , Mignon Sandor r , Jo Smith s , Erich Szerencsits a , Anna Varga t , Valérie Viaud u , Rafal Wawer g , Paul J. Burgess i , Felix Herzog a a Agroscope, Department of Agroecology and Environment, Zurich, Switzerland b European Commission, Joint Research Centre (JRC), Ispra, Italy c INDEHESA, Forestry School, University of Extremadura, Plasencia, Spain d European Forest Institute, Joensuu, Finland e Forest Research Centre, School of Agriculture, University of Lisbon, Lisbon, Portugal f MV Agroecology Research Centre, Mértola, Espírito Santo, Portugal g IUNG-PIB, Institute of Soil Science and Plant Cultivation – State Research Institute, Puławy, Poland h Brandenburg University of Technology (BTU), Cottbus-Senftenberg, Germany i Cranfield University, Cranfield, Bedfordshire, United Kingdom j AGRIDEA, Lindau, Switzerland k Büsgen-Institute, Georg-August-University Göttingen, Germany l Ovidius University of Constanta, Constanta, Romania m University of Santiago de Compostela, Lugo, Spain n Agricultural University of Athens, Karpenissi, Greece o C.N.R., Research Institute on Terrestrial Ecosystems (IRET), Porano, Italy p Swansea University, Department of Geography, Singleton Park, Swansea, United Kingdom q CREA Research Centre for Olive, Citrus and Tree Fruit, Spoleto, Italy r University of Agricultural Sciences and Veterinary Medicine Cluj-Napoca, Cluj-Napoca, Romania s Organic Research Centre, Newbury, Berkshire, United Kingdom t MTA Centre for Ecological Research, Vácrátót, Hungary u UMR SAS, INRA, Rennes, France ARTICLE INFO Keywords: Carbon storage Climate change mitigation Ecosystem services Farmland Resource protection Spatial deficit analysis ABSTRACT Agroforestry, relative to conventional agriculture, contributes significantly to carbon sequestration, increases a range of regulating ecosystem services, and enhances biodiversity. Using a transdisciplinary approach, we combined scientific and technical knowledge to evaluate nine environmental pressures in terms of ecosystem services in European farmland and assessed the carbon storage potential of suitable agroforestry systems, proposed by regional experts. First, regions with potential environmental pressures were identified with respect to soil health (soil erosion by water and wind, low soil organic carbon), water quality (water pollution by nitrates, salinization by irrigation), areas affected by climate change (rising temperature), and by underprovision in biodiversity (pollination and pest control pressures, loss of soil biodiversity). The maps were overlaid to identify areas where several pressures accumulate. In total, 94.4% of farmlands suffer from at least one environmental pressure, pastures being less affected than arable lands. Regional hotspots were located in north-western France, Denmark, Central Spain, north and south-western Italy, Greece, and eastern Romania. The 10% of the area with the highest number of accumulated pressures were defined as Priority Areas, where the implementation of agroforestry could be particularly effective. In a second step, European agroforestry experts were asked to propose agroforestry practices suitable for the Priority Areas they were familiar with, and identified 64 different systems covering a wide range of practices. These ranged from hedgerows on field boundaries to fast growing https://doi.org/10.1016/j.landusepol.2019.02.025 Received 19 July 2018; Received in revised form 10 January 2019; Accepted 18 February 2019 ⁎ Corresponding author at: Agroscope, Department of Agroecology and Environment, Research group Agricultural Landscape and Biodiversity, Reckenholzstrasse 191, 8046 Zurich, Switzerland. E-mail address: [email protected] (S. Kay). Land Use Policy 83 (2019) 581–593 Available online 06 March 2019 0264-8377/ © 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/BY/4.0/). T coppices or scattered single tree systems. Third, for each proposed system, the carbon storage potential was assessed based on data from the literature and the results were scaled-up to the Priority Areas. As expected, given the wide range of agroforestry practices identified, the carbon sequestration potentials ranged between 0.09 and 7.29 t C ha −1 a −1 . Implementing agroforestry on the Priority Areas could lead to a sequestration of 2.1 to 63.9 million t C a −1 (7.78 and 234.85 million t CO 2eq a −1 ) depending on the type of agroforestry. This corresponds to between 1.4 and 43.4% of European agricultural greenhouse gas (GHG) emissions. Moreover, promoting agroforestry in the Priority Areas would contribute to mitigate the environmental pressures identified there. We conclude that the strategic and spatially targeted establishment of agroforestry systems could provide an effective means of meeting EU policy objectives on GHG emissions whilst providing a range of other important benefits. 1. Introduction Increased market price volatility and the risks of changing climate are - according to the EU Agricultural Markets Briefs (September 2017) – the biggest challenges European farmers will face in near future (DG Agriculture and Rural Development, 2017). Facing the complex relationship between competitive farming and sustainable production, the current Common Agricultural Policy (CAP, the European framework for agricultural subsidies), supports farmers’ income, market measures and rural development (European Commission, 2016). In spite of crosscompliance mechanism and the recently introduced greening measure that links environmental standards to subsidies, the agricultural sector is still one of the prime causes of pressure on natural resources and the environment (EEA, 2017a). To address these environmental problems, the European Commission has issued policies such as the Nitrate Directive (91/676/CEE) in 1991, the Water Framework Directive (Directive 2000/60/EC) in 2000 and the Biodiversity Strategy in 2010 (COM(2011) 244). Nonetheless, major environmental problems persist and are still linked to or caused by intensive agricultural production on the one hand, and by land abandonment on the other (Plieninger et al., 2016). Most recently and in line with the COP21 Paris Agreement (UNFCCC, 2015) the Effort Sharing 2021–2030 (REGULATION (EU) 2018/842) includes agricultural practices, aiming to reduce greenhouse gas (GHG) emissions or balance with an equal amount of GHG sequestration. In this context, the future CAP for the next funding period after 2020 (CAP2020+) proposes three focal areas: a) “natural” farming, b) sustainable water management and use and c) dealing with climate change (European Commission, 2017a). This will require strategies to manage the above mentioned financial and environmental risks of production, ideas to expand the agricultural product range, and a focus on sustainable farming systems with climate adaptation and mitigation functions (Wezel et al., 2014). Agroforestry, the integrated management of woody elements on croplands or grasslands (European Commission, 2013a), may become part of those strategies because it provides multiple (annual and perennial) products while simultaneously moderating critical environmental emissions and impacts on soil, water, landscapes, and biodiversity (Torralba et al., 2016). In addition, it is highlighted as one of the agricultural practices with the greatest potential for climate change mitigation and adaptation (Aertsens et al., 2013;Hart et al., 2017). For example, agroforestry can enhance the sequestration of carbon in woody biomass and in the soil of cultivated fields (mitigation) (Kim et al., 2016), increase soil organic matter, improve water availability (adaptation to climate aridification) (Murphy, 2015), protect crops, pastures, and livestock from harsh-cli- mate events (adaptation to global warming and increasing wind speed) (Sánchez and McCollin, 2015). Against this background, our study aimed to evaluate the potential contribution of agroforestry towards achieving zero-GHG emissions agriculture in pursuit of the ambitious Paris Agreement COP21 and CAP targets. Using a transdisciplinary approach including scientific and practical knowledge, the study focused on three key questions: I. Where and to what extent is European agricultural land affected by (multiple) environmental pressures that could be reduced through agroforestry? II. Which regional types of agroforestry (combinations of various woody plants, crop / animal species and management practices) can be used to reduce these environmental pressures and provide multiple products? and – as an example of an ecosystem service that agroforestry can provide – III. What is the impact of the proposed systems on European climate change targets, in particular on carbon storage and GHG emissions? 2. Material and methods The study was conducted in three main phases: First, the agricultural areas most seriously affected by environmental pressures (“Pressure Areas”) were identified using various spatially explicit datasets on e.g. soil erosion, water pollution, and pollination pressures. In a second step, local agroforestry experts were consulted to propose suitable agroforestry practices for their regions suffering from environmental pressures. Finally, the annual carbon storage impact of the proposed systems was identified and evaluated in the light of European agricultural GHG emissions. In the next subsections, these three main phases are described more in detail, while we address advantages and limitations of the adopted approach, as well as possible improvements, in the Discussion section. 2.1. Identification of Priority Areas 2.1.1. Conceptual approach Bearing in mind that agroforestry is only one aspect of a diversified agriculture, our focus was on agricultural areas facing combined environmental pressures, in which agroforestry can mitigate several environmental pressures. Fig. 1 illustrates the conceptual background of the Priority Area approach. The analysis uses the Corine Land Cover 2012 (EEA, 2016) to identify the area of European arable and pasture land. From this farmland layer, the areas of high nature value such as Natura 2000 (EEA, 2015a), High Nature Value Farmland (EEA, 2015b;Paracchini et al., 2008), and the existing agroforestry areas (den Herder et al., 2017) were subtracted. The remaining “Focus Areas” (block II in Fig. 1) were the starting point for the pressure analysis. 2.1.2. Selection of indicators to assess environmental pressures The indicator selection passed three stages as visualised in Fig. 2. First, indicators characterizing benefits provided by agroforestry systems were chosen. Torralba et al. (2016) summarized them into i) timber, food, and biomass production, ii) soil fertility and nutrient cycling, iii) erosion control, iv) biodiversity provision and Hart et al. (2017) completed the list with v) climate change mitigation. Given that continental spatial datasets covering most of the European countries and addressing the indicators in a consistent way were limited, the selection focussed in a second step on the CAP 2014–2020 context indicators. The CAP monitoring and evaluation framework is composed on a set of socio-economic, sectorial, and environmental indicators to reflect the impact and provide (annual) information of the performance S. Kay, et al. Land Use Policy 83 (2019) 581–593 582 of the strategy (European Commission, 2017b,2014). Within this list four context environmental indicators were related to agroforestry benefits. These were i) water abstraction in agriculture (C.39) addressing pressures on available fresh water resources (indicator: e.g. irrigated area), ii) water quality (C.40) dealing with agricultural water pollution by nitrates and phosphates (indicator: e.g. Gross Nitrogen Balance), iii) soil organic matter (SOC) in arable land (C.41) as SOC influences soil structure, aggregate stability, nutrient availability, water retention and resilience (indicator: SOC content), and iv) soil erosion by water (C.42) the most widespread form of soil degradation (indicator: erosion) (European Commission, 2017b). Third, as indicators for climate change mitigation and biodiversity were not addressed within the CAP monitoring, we reviewed the literature and identified relevant datasets. In conclusion, only consistent spatial datasets, which were available with a wide European coverage, were included in the analysis. Accordingly, environmental pressures related to: i) soil health (soil erosion by wind and water, soil organic carbon), ii) water quality and abstraction (water pollution by nitrates, irrigation), iii) climate change (rising temperature), and iv) biodiversity (pollination and pest control pressures, reduced soil biodiversity) were identified. Individual pressure maps were spatially aggregated and combined into the “Pressure Areas” map showing all regions where one or several environmental pressures occur. To identify the “Priority Areas” for intervention, the sum of pressures per spatial unit (pixel size = 100 m x 100 m) was expressed as an accumulation map or a “heatmap of environmental pressures”. Fig. 1. Conceptual approach for the spatially explicit pressure analysis. European agricultural land: Arable and pasture land. Focus Areas: European agricultural land minus nature conservation areas, High Nature Value Farmland, and existing agroforestry land. Pressure Areas: Areas where at least one ecosystem service pressure was mapped. Priority Areas: Areas where environmental pressures accumulate (four out of eight in pasture and five out of nine in arable land). Fig. 2. Selection process of environmental pressure indicators. 1. Assessment of agroforestry benefits, 2. Evaluation of existing CAP context indicators, 3. Review of further environmental pressure indicators for climate change mitigation and biodiversity provision. (SOC: Soil organic carbon). S. Kay, et al. Land Use Policy 83 (2019) 581–593 583 2.2. Pressure Area analysis 2.2.1. Soil health pressures The European water erosion map (Panagos et al., 2015) and the Swiss soil erosion risk map (Prasuhn et al., 2013) together with the European wind erosion map (Borrelli et al., 2017) were used to locate areas with potentially critical loads of soil losses. According to Panagos et al. (2015) a critical threshold is reached if the soil loss is more than 5 t soil ha −1 a −1 . The analysis of potential wind erosion was limited to arable land, which is more affected than grassland. The Soil Organic Carbon (SOC) saturation capacity provided at European level by Lugato et al. (2014a),2014b) expresses the ratio between actual and potential SOC stocks. Regions with a ratio of less than 0.5 were identified as Pressure Areas, meaning that these soils contain less than half of their SOC storage potential. 2.2.2. Water quality pressures Irrigated fields regardless of whether they were pasture or arable land were included in the pressure analysis. Irrigation maps were provided by the JRC Water Portal (2017) and the Farm Structure Survey (FSS) (Eurostat, 2017a) and expressed the proportion of irrigated land on the total agricultural area. Regions with more than 25% of the agricultural area under irrigation were included as Pressure Area. The nitrogen surplus, which can lead to both high levels of nitrate leaching and denitrification to gaseous nitrous oxide, was assessed for the European Union using the CAPRI model by Leip et al. (2014). For Switzerland data were obtained from modelled accumulated nitrogen losses (BAFU, 2015). According to the German Ministry of Environment (BMUB, 2017), there is a critical load if the annual nitrogen surplus exceeds 70 kg N ha −1 a −1 and this threshold was used to identify areas with high nitrogen surplus. 2.2.3. Pressures related to changing climate Annual mean temperatures from the current climate (1970–2000 WorldClim; Hijmans et al., 2005) and the forecast for 2050 (HadGEM2- ES, Martin et al., 2011) were used to derive the predicted regional temperature increase up to 2050. According to Hart et al. (2012), agroforestry systems remain robust within an average temperature increase of up to 4 °C. Therefore, all areas with a predicted increase of temperature of more than 2 °C and less than 4 °C were qualified as Pressure Areas where agroforestry could potentially be beneficial. 2.2.4. Biodiversity pressures Soil fauna, microorganisms and biological functions derived from the spatial analysis by Orgiazzi et al. (2016) were used to assess soil biodiversity. The areas identified with “high” and “moderate-high” levels of risk were defined as Pressure Areas. The pollination assessment was based on the indicator of landscape suitability to support pollinators by Rega et al. (2017). The indicator is a dimensionless score; areas with “very low” and “low” suitability (corresponding to the first two quintiles of the values’ distribution) were defined as Pressure Areas. The pest control index (Rega et al., 2018) was used as input for the assessment of regions with potential pressures in natural pest control. Again, the indicator is a dimensionless score; areas with “very low” and “low” suitability to support natural pest control, corresponding to the first two quintiles of the values’ distribution, were combined and defined as Pressure Areas. 2.2.5. Selection of Priority Areas Using the thresholds previously mentioned (Table 1), the nine environmental pressures were spatially combined using GIS. In each spatial unit the number of pressures were added together by weighting each indicator equally. Implications, advantages and drawbacks of this methodological approach are addressed in the discussion section (4.1). In the resulting “heatmap”, the 10% of the area with the highest number of pressures were defined as the Priority Area for the implementation of agroforestry. Based on Mücher et al. (2010) the Priority Areas were clustered into seven biogeographical regions: Atlantic; Continental lowlands, Continental hills; Mediterranean lowlands, Mediterranean hills, Mediterranean mountains; and Steppic. The spatial analysis was performed in ArcGIS10.4 (ESRI, 2016). The Table 1 Spatial datasets with their respective characteristics and the threshold applied to define Pressure Areas (EU28: Austria, Belgium, Bulgaria, Croatia, Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Italy, Ireland, Latvia, Lithuania, Luxembourg, Malta, the Netherlands, Poland, Portugal, Romania, Spain, Slovakia, Slovenia, Sweden, and the United Kingdom; EU 27: is without Croatia; CH = value for Switzerland). Indicator Source Coverage Resolution Threshold Focus Area CORINE - Agricultural land EEA, 2016 all Europe 250 m Agroforestry area den Herder et al., 2017 EU 28, CH 100 m High Nature Value Farmland EEA, 2015b;Paracchini et al., 2008 all Europe (without Greece) 100 m Natura 2000, Ramsar areas EEA, 2015a EU 28, CH Soil Pressure Areas Soil erosion by water Panagos et al., 2015;Prasuhn et al., 2013 EU 28, CH 100 m > 5 t soil ha −1 a -1 (Panagos et al., 2015) Soil erosion by wind Borrelli et al., 2017 EU 28, CH 500 m > 5 t soil ha −1 a −1 (Panagos et al., 2015), limited to arable land Soil Organic Carbon (SOC) saturation capacity Lugato et al., 2014a,2014b EU 28 250 m < 0.5 Ratio between actual and potential SOC stock (Lugato et al., 2014a,2014b) Water related Pressure Areas Irrigation Eurostat, 2017a,2017b,2017c; JRC Water Portal, 2017 all Europe 100 m, 1000 m > 25% irrigated land Nitrogen surplus BAFU, 2015; Leip et al., 2014 EU 27, CH (without Cyprus) 1000 m (100 m CH) > 70 kg N ha −1 a −1 (BMUB, 2017). Climate Risk Areas Climate change / Temperature rise Hijmans et al., 2005 all Europe 2 - 4 °C between 1990 and 2050 (Hart et al., 2012) Biodiversity Pressure Areas Soil biodiversity Combination of soil fauna, soil microorganisms and soil biological function; Orgiazzi et al., 2016 EU 27 500 m Risk level “high”, “moderatehigh” (Orgiazzi et al., 2016) Landscape suitability to support pollination Rega et al., 2017 all Europe (without Cyprus) 100 m Classes “very low” and “low” (Rega et al., 2017) Pest control index Rega et al., 2018 all Europe (without Cyprus 100 m Classes “very low” and “low” (Rega et al., 2018) S. Kay, et al. Land Use Policy 83 (2019) 581–593 584 outcomes were processed in R (R Development Core Team, 2016) with packages plyr (Wickham, 2016), Hmisc (Harrell, 2018), and ggplot2 (Wickham et al., 2016). 2.3. Agroforestry recommendations Potential agroforestry practices, which are: 1) of interest to farmers and the most likely to be adopted by them, 2) the most adapted to mitigate the prominent environmental issues in the region, 3) the most developed in the region and 4) the most suitable to face climate change, were compiled by local experts and the authors for each Priority Area. A total of 20 experts, mainly national delegates of the European Agroforestry Federation (EURAF) or associated researchers, were asked for their contribution. We used a uniform emailing consisting of an explanation letter, maps of the Priority Area and a structured template. The template was divided into eight questions (see Table 3): i) type of agroforestry (e.g. silvopastoral, silvoarable; hedgerows, coppice, or single trees), ii) title and a short description of the system, iii) tree and hedgerow species, iv) number of trees per hectare or the percentage of woody cover per hectare, v) planting scheme (e.g. lines, scattered) and management system (e.g. year of harvesting / harvesting cycles), vi) crop species and products, vii) tree products, and viii) harvesting year. The outcomes were summarized by biogeographical region. 2.4. Assessment of carbon sequestration in biomass The total biomass production (aboveground wood and root biomass) of the woody elements and the carbon storage potential of the proposed agroforestry systems were assessed based on literature data (see Supplementary material) and from (regional) test sites [units: t biomass ha −1 a −1 ; t C ha −1 a −1 ]. Herein the values represented an average potential per year of tree life and did not consider any dynamics of tree growth over time, or other impact factors such as water and nutrient availability, temperature, tree density, etc. Potential minimum and maximum values of carbon storage in biomass (both above- and belowground) of each agroforestry practice for each biogeographic region were extracted separately for pasture and arable land. These values were used for upscaling the results to the “Priority Area”, assuming that in those regions, the total available farmland would be converted into agroforestry with one of the recommended agroforestry practices. 3. Results 3.1. Pressure assessment In EU (EU28 minus Cyprus and Croatia) and Switzerland, the total area of European agricultural land is 1,544,022 km 2 (CLC, EEA, 2016). Subtracting existing agroforestry and nature protection areas, the analysis was then restricted to 1,414,803 km 2 as Focus Area. This area consisted of 1,071,179 km 2 of arable land (≙92% of total European arable land) and 343,624 km 2 of pasture (88% of total European pasture). Fig. 3 gives an overview of the size of the individual “Pressure Areas” in relation to the Focus Area. Soil loss risks over 5 t soil ha −1 a −1 from water erosion were identified on 11.9% of the arable area and 9.5% of the pasture. Areas suffering from an annual loss greater than 5 t soil ha −1 a -1 by wind erosion were relatively small (1.5%), whereas a low SOC saturation capacity was present on 58.7% of arable lands and on 12.8% of pastures. In total, 8.4% of the arable areas and 1% of the pastures had irrigation levels greater than 25%. High nitrogen pollution risk was mapped on 20.6% of arable lands and on 34.5% of the pastures. Around 63.0% of arable lands and 53.6% of pastures were located in regions where temperature is expected to rise between 2 and 4 °C by 2050 according to the HadGEM2-ES forecast scenario. Pressures in biodiversity and resulting potential underprovision of ecosystem services are widely spread all over European agricultural land. In total, 66.4% of the arable lands and 36.8% of pastures in the Focus Area were predicted to have low or very low natural pest control potential, whilst 41.8% of the arable areas and 21.0% of pastures were predicted to be not suitable for supporting pollinators. Potential soil biodiversity pressures were mapped on 11.5% of arable lands and on 18.7% of pastures. By combining the nine individual pressure maps, we created a heatmap for environmental pressures (Fig. 4a). For the total Pressure Area, a lower proportion of pasture areas were identified than of arable lands. Only 4% of the arable lands in the Focus Fig. 3. The proportion of a) the arable land and b) the pasture affected by each of the environmental pressures across the selected Focus Areas. Wind erosion was only considered in arable areas. (SOC: Soil organic carbon). S. Kay, et al. Land Use Policy 83 (2019) 581–593 585 Fig. 4. a) Heatmap for the number of environmental pressures and b) Priority Areas (arable areas with more than five pressure indicators; and pasture areas with more than four pressure indicators). Table 2 Summary of the Priority Areas by country divided into biogeographical regions based on the landscape classification by Mücher et al. (2010). Biogeographical Region Country Arable land [km 2 ] Pasture [km 2 ] Total [km 2 ] Share of total agricultural land [%] Atlantic Total 29,611 29,088 58,698 9.74 Denmark 498 3,223 3,721 20.19 France 16,156 6,151 22,308 10.7 Germany 6,366 102 6,468 9.78 Ireland 6 7,133 7,139 17.12 Netherlands 2,624 3,030 5,654 32.96 UK 2,600 8,719 11,319 8.43 others 1,361 730 2,090 1.8 Continental Lowlands Total 7,644 1,259 8,903 6.24 Denmark 3,607 21 3,628 38.82 Germany 1,660 809 2,469 5.22 Poland 1,296 106 1,402 3.76 Others 1,081 322 1,403 2.88 Hills Total 13,906 4,360 18,265 4.11 Bulgaria 2,116 537 2,654 7.03 Germany 1,905 1,473 3,377 3.88 Poland 6,379 439 6,818 5.73 Romania 2,054 1,078 3,132 4.87 others 1,452 833 2,285 1.68 Mediterranean Lowlands Total 12,399 156 12,555 22.52 Greece 3,020 42 3,063 38.28 Italy 7,990 39 8,029 21.15 Spain 1,220 50 1,270 22.36 Others 169 25 193 4.7 Hills Total 20,226 650 20,876 15.53 Greece 2,340 117 2,457 22.04 Italy 6,985 83 7,069 15.64 Spain 9,676 227 9,903 25.02 Others 1,225 223 1,448 3.77 Mountains Total 12,858 628 13,486 10.96 Italy 1,071 78 1,149 10.66 Spain 11,176 429 11,606 12.34 Others 611 120 732 4.02 Steppic Total 2,948 1,026 3,974 11.54 Total 99,592 37,166 136,758 8.87 S. Kay, et al. Land Use Policy 83 (2019) 581–593 586 Table 3 Agroforestry practices for arable lands and pastures in the European biogeographical regions (an extract of the practices with the lowest, medium, and the highest carbon sequestration potential are shown. See Supplementary material for the complete list and references). (SRC: Short rotation coppice) Biogeographical region Agroforestry type Title Tree / hedgerow species Trees [trees ha −1 ], hedgerow [m ha −1 ] or wood cover [% ha −1 ] Management system Crop species and products Tree products Year of tree harvesting Carbon sequestration [t C ha −1 a −1 ] Atlantic pasture Silvopastoral, coppice Agroforestry for ruminants in France Pear (Pyrus spp), honey locust (Gleditsia triacanthos), service tree (Sorbus domestica), white mulberry (Morus alba), Italian alder (Alnus cordata), goat willow (Salix caprea), field elm (Ulmus minor), black locust (Robinia pseudoacacia), grey alder (Alnus incana) (single −2 m, double −6 m, triple −10 m), 4 m for trees, 1.3 m coppices x 20 m, (11% woody cover) Single, double, or triple lines Grazing, hay, silage Fodder-trees, woodchips 5–8 0.16-0.48 Atlantic pasture Silvopastoral, single trees Traditional orchard Fruit trees (apple – Malus domcestica, pear - Pyrus spp, plum -Prunus domestica) 80 trees ha −1 Lines Grazing, hay, silage Fruits (woodchips) 60 1.23 Atlantic pasture Silvopastoral, single trees High stem timber trees Poplar (Populus spp) 400 trees ha −1 , After 15-20 years: 120-150 trees ha −1 Lines Grazing, hay, silage Timber First cut: 15-20 harvest:25-30 2.78-6.35 Atlantic arable Silvoarable, hedgerows Productive boundary hedgerow Mixed hedgerow species: hawthorn (Crataegus spp), blackthorn (Prunus spinosa), field maple (Acer campestre), hazel (Corylus avellane) 0.03% ha −1 Boundary hedgerow Crop rotation with cereals (wheat, barley, oats), potatoes, squash, organic fertility building ley Woodchips Every 15 0.1-0.45 Atlantic arable Silvoarable, coppice Alley cropping – Short Rotation Coppice (SRC) Willow (Salix viminalis), hazel (Corylus avellana) 1000-1300 trees ha −1 (24% ha −1 ) Twin rows with 10- 15 m wide crop alley Cereals (wheat, barley, oats), potatoes, squash, organic fertility building ley Woodchips Every 2 for willow, every 5 for hazel 0.36-1.05 Atlantic arable Silvoarable, single trees High stem timber trees Walnuts (Juglans regia), maples (Acer spp), wild cherry (Prunus avium), checker tree (Sorbus torminalis), service tree (Sorbus domestica), apple (Malus domestica), pear (Pyrus spp). 28-110 trees ha −1 , (26- 50 m between rows) Lines Timber 60 Walnut: 0.32-2.75, cherry: 0.19-1.4 Continental hills pasture Silvopastoral, single trees Wooded grassland Fruit trees: cherry (Prunus avium), walnut (Juglans regia), apple (Malus domestica), etc. 60 trees ha −1 Lines Grazing, hay, silage Fruits 70-90 Cherry: 0.41-0.76, apple: 0.93-1.43, walnut: 0.86-1.16 Continental lowlands pasture Silvopastoral, coppice Agroforestry for freerange pig production Poplar (Populus spp), willow (Salix spp), various fruit trees 10-40% ha −1 (2.5 × 3.5 m) SRC lines Grazing, hay, silage Woodchips, fodder-trees 5-8 Poplar: 0.44-1.41 Continental hills pasture Silvopastoral, single trees High nature and cultural value wood pastures and wooded grasslands Sessile oak (Quercus petraea), beech (Fagus sylvatica), hornbeam (Carpinus betulus), wild fruit trees, mixed poplar (Populus spp.), willow (Salix spp.) 50-300 trees ha −1 (10- 50% ha −1 ) Scattered Grazing, hay, silage Acorns, fruits, timber, (foddertrees) Trees not harvested Oak: 0.71-2.83, beech: 0.59-2.34, hornbeam: 0.38- 1.55 Continental lowlands arable Silvoarable, coppice Alley cropping Poplar (Populus spp); Mixed hedgerow species: willow (Salix spp), hornbeam (Carpinus betulus), common ash (Fraxinus excelsior), common birch (Betula pendula), black locust (Robinia pseudoacacia) Rows A, B, and C: 10,000 trees ha −1 , Rows D, E, F, and G: 2222 trees ha −1 , (10% ha −1 ). Single and twin rows with 48, 96, and 144 m wide crop alleys. Crop rotation (wheat, maize, oilseed rape, barley) Woodchips Rows A, B, and C: every 3-5. Rows D, E, F, and G: every 8–10 0.15 - 0.44 Continental hills arable Silvoarable, single trees Orchard with vegetables or fruits (strawberries) Fruit trees: cherry (Prunus avium), walnut (Juglans regia), apple (Malus domestica), etc 60 trees ha −1 Lines Vegetable, berries (strawberries) Fruits, timber 70-90 Cherry: 0.41-0.76, apple: 0.93-1.43, walnut: 0.86 -1.16 Continental hills arable Silvoarable, single trees Non-native, energy tree with Alfalfa Pauwlonia (Paulownia tomentosa) 126 trees ha −1 (18 m x 5 m) Lines Triticale, alfalfa Timber 10-12 3.77 (continued on next page) S. Kay, et al. Land Use Policy 83 (2019) 581–593 587 Table 3 (continued) Biogeographical region Agroforestry type Title Tree / hedgerow species Trees [trees ha −1 ], hedgerow [m ha −1 ] or wood cover [% ha −1 ] Management system Crop species and products Tree products Year of tree harvesting Carbon sequestration [t C ha −1 a −1 ] Mediterranean hills pasture Silvopastoral, single trees Dehesa Holm oak (Quercus ilex) 25-50 trees ha −1 Scattered Grazing Acorns, foddertrees Trees not harvested 0.09–0.16 Mediterranean hills pasture Silvopastoral, single trees Grazed cork oak plantation Cork oak (Quercus suber) 113 trees ha −1 , after 20 years: 50 trees ha −1 Lines Grazing Cork, timber 80 0.34-1.29 Mediterranean hills/ mountains pasture Silvopastoral, single trees Grazed fruit plantations Olive (Olea europaea), almond (Prunus dulcis) 250 trees ha −1 Lines Grazing, legume rich mix (annual self seeding species) Fruits, oil, nuts Annual prunings, trees not harvested Olive: 1.97, almond:1.36 Mediterranean lowlands arable Silvoarable, single trees High stem timber trees Pedunculate oak (Quercus robur) 57 trees ha −1 Lines Cereals Timber 35 0.11-0.26 Mediterranean hills arable Silvoarable, single trees Fruit tree alley Olive (Olea europaea) 200-400 trees ha −1 Lines or scattered Wild asparagus Oil, forage Annual prunings, trees not harvested 1.57-3.14 Mediterranean mountains arable Silvoarable, single trees High stem timber trees Poplar (Populus spp) 200 trees ha −1 Lines Crop rotation wheat, oilseed rape, chickpeas Timber 15 5.76 - 7.29 Steppic arable Silvoarable, single trees High stem forest trees Poplar (Populus spp), willow (Salix spp.), black locust (Robinia pseudoacacia), pedunculate oak (Quercus robur), plain common and black walnut (Juglans nigra), common ash (Fraxinus excelsior), red oak (Quercur subra),), lime (Tilia sp.), 60–70 trees ha −1 Lines Vegetables Timber 70-90 Poplar: 1.72-2.85, oak: 0.32-1.2, walnut: 1.31 Steppic, arable Silvoarable, single trees Mixed timber and wild fruit species plantation Grayish oak (Quercus pedunculiflora), field maple (Acer campestre), lime (Tilia sp.), hawthorn (Crataegus sp), Rosa sp, blackthorn (Prunus spinosa) 100 trees ha −1 Lines Vegetables Fruits, fodder trees, timber Harvesting depends on species estimated from 25-120. Oak : 1.59, tilia:1.32 Steppic, arable Silvoarable, single trees Poplar plantation Poplar (Populus spp) 100 trees ha −1 Lines Sunflower, cabbage, corn, pepper and eggplant, water-melon and squash, cauliflower; wheat, beans Timber 35 2.88-4.76 S. Kay, et al. Land Use Policy 83 (2019) 581–593 588 Areas had no pressures, while in pasture it was around 12%. More than half of the pasture areas had less than three pressures, while 35% of arable area were affected by more than four pressures, and 9% had more than five pressures. Whilst we defined the Priority Areas as arable lands with more than five pressures, we set the threshold to only four pressures for pasture, as we evaluated only eight pasture pressure indicators (excluding soil erosion by wind). Together, they represent the worst 10% of the Pressure Area (Fig. 4b). These combined Priority Areas for arable and pasture land amounted to 136,758 km 2 , which corresponds to about 8.9% of the total European agricultural land. Table 2 gives an overview of the Priority Areas according to country and biogeographical region. 3.2. Potential agroforestry practices In total, 64 agroforestry practices were proposed by the authors and local experts. They cover a wide range of practices from hedgerow systems on field boundaries to fast growing coppices or scattered single tree systems. Table 3 lists, for each biogeographical region, the proposed system with the lowest, medium, and highest carbon sequestration potential. In line with the largest Pressure Areas, the highest number of agroforestry practices was proposed for Atlantic regions (14 silvopastoral and 9 silvoarable practices) followed by Mediterranean arable lands. The complete list can be found in Supplementary material. 3.3. Carbon storage potential For each system the annual carbon storage potential of the woody elements (including roots) was identified using data from the literature and in each geographical region, the minimum and maximum storage potential were determined. The wide range of practices selected corresponded to a wide range of carbon storage potentials, between 0.09 and 7.29 t C ha −1 a −1 . In Table 4 these data were upscaled to the entire Priority Area of each biogeographical region. Overall, implementing the proposed agroforestry practices in the Priority Areas could mitigate between 2.1 and 63.9 million t C a −1 depending on the systems chosen, which is between 7.7 and 234.8 million t CO 2eq a −1 . In 2015, the 28 members of the European Union (EU28) together with Switzerland emitted 4,504.9 million t of greenhouse gases (million t CO 2eq ), with agriculture contributing 12% (˜540 million t CO 2eq ; Eurostat, 2017b). Converting the conventionally used farmland in the Priority Area (which was about 8.9% of total agricultural land) to agroforestry could therefore capture between 1.4 and 43.4% of the European agricultural GHG emissions. 4. Discussion This research investigated three questions: I) Where and to what extent is European agricultural land affected by (multiple) environmental pressures? II) Which regional types of agroforestry can be used to reduce environmental pressures? and III) What is the potential contribution of the proposed systems to the European zeroemission agriculture climate targets? 4.1. European environmental Pressure Areas In response to the first question, several environmental pressures that can be mitigated by establishing agroforestry practices were selected. According to Alam et al. (2014) and Torralba et al. (2016) these include soil conservation, the improvement of water quality, nutrient retention, climate regulation, and enhanced biodiversity. We investigated nine environmental pressures and mapped their occurrence in European agricultural land, based on existing spatially explicit databases at a continental European scale. The best available data were used, although it should be noted that differences in scales (100–1000 m pixel size), time periods (2006–2017) and models (e.g. modelled soil losses in EU vs. soil erosion risk map in Switzerland) existed that might result in spatial inaccuracies (Schulp et al., 2014). However, as other authors have pointed out, this is an intrinsic limit of all pan-European, spatially explicit studies: “as fully harmonized data on the different aspects are not available, the possible bias from inconsistencies between the different data layers is unavoidable” (Malek and Verburg, 2017). All the datasets used, required some degree of modelling and the maps therefore show predicted rather than measured environmental pressures. Moreover, not all the existing environmental problems in agricultural areas could be addressed. Methane emissions, ammonia emissions, and zoonoses contamination, for example, were not included in the analysis presented here. In addition, biodiversity aspects in terms of quality and diversity (Zhang et al., 2007), the amenity value of the landscape, and natural hazards, such as avalanches, floods, droughts, and landslides (EEA, 2017b) were not considered. Recommendations from the literature were used to define the thresholds for delimiting the Pressure Areas. The definition of thresholds is always arbitrary to some extent: different thresholds exist and modifying these or using different models would affect the size and spatial location of the Pressure Areas. For erosion, we used 5 t soil ha −1 a −1 as a threshold for erosion caused by water and erosion caused by wind, whereas for example, adopting a “tolerable” soil erosion rate of 0.3 to 1.4 t soil ha −1 a −1 as recommended by Verheijen et al. (2009) would strongly have increased the Pressure Area. The 5 t soil ha −1 a −1 threshold was uniformly used for the whole of Europe. However, soil erosion threshold values could also be defined by the nature of the soils in a particular area, depending for example, on soil quality and depth, with lower quality and shallower soils given lower thresholds to reflect their already precarious state and the relative importance of conserving what remains. Surplus regions for nitrogen have also been defined in different Table 4 Potential carbon sequestration in the whole Priority Area using minimum and maximum carbon storage potential of agroforestry practices proposed for each biogeographical region. Biogeographical region Minimum carbon storage potential Maximum carbon storage potential [t C km −2 a -1 ] Priority Area [t C a −1 ] [t C km −2 a -1 ] Priority Area [t C a −1 ] Arable land Pasture Arable land Pasture Total Arable land Pasture Arable land Pasture Total Atlantic 10 16 296,109 465,401 761,510 275 635 8,142,998 18,470,618 26,613,616 Continental lowlands 15 44 114,660 55,396 170,056 159 141 1,215,401 177,518 1,392,919 Continental hills 27 38 375,461 165,661 541,122 377 283 5,242,545 1,233,741 6,476,286 Mediterranean lowlands 11 9 136,390 1,400 137,790 600 197 7,439,447 30,654 7,470,101 Mediterranean hills 11 9 222,488 5,850 228,338 530 197 10,719,872 128,053 10,847,925 Mediterranean mountains 11 9 141,441 5,650 147,092 729 197 9,373,711 123,676 9,497,387 Steppic hills 32 38 94,322 39,003 133,325 476 283 1,403,039 290,467 1,693,506 Total 1,380,871 738,362 2,119,233 43,537,013 20,454,727 63,991,740 S. Kay, et al. Land Use Policy 83 (2019) 581–593 589