Towards an enhanced indication of provisioning ecosystem services in agro-ecosystems

Bethwell, Claudia; Burkhard, Benjamin; Daedlow, Katrin; Sattler, Claudia; Reckling, Moritz; Zander, Peter

doi:10.1007/s10661-020-08816-y

Open Access
Published: 14 May 2021

Towards an enhanced indication of provisioning ecosystem services in agro-ecosystems

Environmental Monitoring and Assessment volume 193, Article number: 269 (2021) Cite this article

2528 Accesses
8 Citations
Metrics details

Abstract

Provisioning ecosystem services play a vital role in sustaining human well-being. Agro-ecosystems contribute a significant share of these services, besides food and fodder and also fuel and fibre as well as regulating and cultural ecosystem services. Until now, the indication of provisioning ecosystem services of agro-ecosystems has been based almost only on yield numbers of agricultural products. Such an indication is problematic due to several reasons which include a disregard of the role of significant anthropogenic contributions to ecosystem service co-generation, external environmental effects and strong dependence on site conditions. We argue for an enhanced indication of provisioning ecosystem services that considers multiple aspects of their delivery. The conceptual base for such an indication has been made by prior publications which have been reviewed. Relevant points were taken up in this article and condensed into a conceptual model in order to develop a more holistic and expanded set of indictors, which was then exemplarily applied and tested in three case studies in Germany. The case studies represent different natural conditions, and the indicator set application showed that ecosystem services (ES) flow—in terms of output alone—does not characterise agro-ecosystems sufficiently. The proposed aspects of provisioning ecosystem services can give a fuller picture, for example, by input-output relationships, as it is possible by just using single indicators. Uncertainties as well as pros and cons of such an approach are elaborated. Finally, recommendations for an enhanced indication of provisioning ecosystem services in agro-ecosystems that can help to integrate agricultural principles with ideas of sustainability and site-specific land use are derived.

Introduction

General background and objectives

The ecosystem services (ES) concept has gained great scientific importance especially during the last decade (Potschin et al. 2016), and the engagement of policy has also increased (Maes et al. 2012). For example, within its Biodiversity Strategy 2020^{Footnote 1} (EC 2011), the European Union focuses very much on ES and recognises that biodiversity and functioning ecosystems are the base for ES supply. To create a knowledge base on ES supply, all member states were asked to map and assess the state of ecosystems and their services in their national territories until 2020 based on the Strategy’s Target 2 Action 5 (Maes et al. 2016). In 2016, the USA released a memorandum^{Footnote 2} that directs agencies to incorporate ES into federal planning and decision-making. ES also find consideration in global sustainability policies such as the UN’s Sustainable Development Goals (SDGs; e.g. goal 15^{Footnote 3}) (Geijzendorffer et al. 2017).

ES are broadly defined as the benefits that nature provides to humans which are essential to their existence and well-being (Costanza et al. 1997; MEA 2005^{Footnote 4}). ES include provisioning ES such as food, materials or energy and regulating and maintenance ES like climate, water and erosion regulation along with cultural ES such as recreational services (TEEB 2010^{Footnote 5}; CICES^{Footnote 6}, see Haines-Young and Potschin 2013). A particular focus has been given to the generation of provisioning ES mostly in managed, to a lesser degree also in unmanaged, ecosystems (Plieninger et al. 2016), because these ES are at the core of direct human interest and activity, i.e. ensuring nutrition and material supply. There are different approaches to categorize provisioning ES (cf. MEA 2005; TEEB 2010; CICES, see Haines-Young and Potschin 2013). At the centre of most classification schemes is the biomass production from cultivated plants and animals, such as food and fodder biomass and raw materials. In addition, these ES provide genetic resources, medical and ornamental resources for humans and freshwater (TEEB 2010).

A fact that sets provisioning services apart from the other ES is that provisioning services often take the form of ecosystem ‘goods’, which actually can be more or less directly consumed or traded in markets, while for most of the other ES, such markets do not exist or function only poorly (Boyd and Banzhaf 2007). Moreover, in order to consider ecosystems that are managed based on anthropogenic system inputs, a newer definition of ES as ‘the contributions of ecosystem structure and function—in combination with other inputs—to human well-being (Burkhard et al. 2012a) has been proposed.

The MEA (2005, p. V) recognised that people are integral parts of ecosystems, and also the European Landscape Convention^{Footnote 7} sees landscape as ‘an area, as perceived by people, whose character is the result of the action and interaction of natural and/or human factors’ (European Council 2000). This particularly applies to intensively used landscapes, including agro-ecosystems^{Footnote 8}. Agro-ecosystems and their provisioning ES strongly rely on the modification of natural ecosystems, with input-dependent energy, matter and information flows. Consequently, it does not seem realistic to consider only purely nature-derived goods as ES (Pérez-Soba et al. 2012). Therefore, we strongly support the idea to include co-generated outcomes (commodity products) of managed agro-ecosystems as ES and to indicate them jointly with positive and negative externalities of their supply.

Agro-ecosystems are managed ecosystems (Zhang et al. 2007; Power 2010). They are highly subjected to anthropogenic system inputs (‘agro-ecosystem services’, Burkhard et al. 2014), can affect multiple other ES (Zhang et al. 2007) and lead to positive or negative environmental impacts (Petz and van Oudenhoven 2012). Thus, both natural and human-derived capitals are needed for the co-generation of agro-ecosystem services (Jones et al. 2016). The supply of provisioning services, like the production of food market products, forage, fuel, pharmaceuticals or energy crops, belongs to the main objectives of agro-ecosystems (Power 2010; Kandziora et al. 2013a). But agro-ecosystems also provide regulation and maintenance ES such as climate and water regulation (e.g. Balmford et al. 2011) and cultural ES such as landscape aesthetics or knowledge systems (Huang et al. 2015). However, agro-ecosystems can also be the source of ecosystem dis-services (e.g. soil erosion (Steinhoff-Knopp and Burkhard 2018) and nitrate leaching (cf. Fridman and Kissinger 2018; Zhang et al. 2007)).

Also, as agricultural production strongly depends on other ES, it has been repeatedly mentioned as a user of (especially regulating) ES (Power 2016, 2010; Zhang et al. 2007). Pollination, pest and disease control, regulation of erosion, water, local climate and nutrients are the most common examples (Plieninger et al. 2016; Guerra and Pinto-Correia 2016; Guerra et al. 2016; Dungait et al. 2012). Societal goals are aimed at enhancing regulating ES and human well-being-supporting services, while strengthening the competitiveness of the agricultural sector. At landscape levels, this requires an enhancement of landscape multi-functionality (Rossing et al. 2009; Renting et al. 2009; Huang et al. 2015). Appropriate agricultural management strategies and decisions could enhance positive effects of agro-ecosystem services (e.g. Stein-Bachinger et al. 2015; Duru et al. 2015) and reduce negative environmental impacts (e.g. Zhang et al. 2007; Garbach et al. 2017; Kanter et al. 2018).

Looking at these complex interactions, emerging ES synergies and trade-offs, an indication of agro-ecosystem services should be able to answer questions like: ‘How can provisioning ES of agriculture be indicated?’, ‘How can the different forms of system inputs be distinguished and measured effectively to assess the above mentioned anthropogenic inputs and their resulting actual ES flows, and the positive and negative effects of their supply?’ and ‘How can this enhanced indication be applied on a regional scale?’.

An enhanced integrative indication of provisioning ES supply in agro- (and other managed) ecosystems improves the prevailing estimation of their (1) ES potentials, (2) anthropogenic inputs, (3) actual ES flows, (4) environmental externalities and (5) ES demands and preferences under consideration of the (6) mapping of provisioning ES. However, up to now, the third aspect has almost exclusively been widely discussed and actually used to measure provisioning ES (cf. Maes et al. 2016). Some promising conceptual attempts to overcome this research gap have been recently published. Jones et al. (2016) distinguished between stocks and flows of natural and human-derived capital, but remained on a rather conceptual level. Qualitative interview data have been used by Fischer and Eastwood (2016) to analyse ES co-production (and disservices) as human-nature interactions. Human inputs such as use of fertilisers, energy, irrigation, tillage or management knowledge have been considered relevant for agricultural ES supply by Albert et al. (2015) and Burkhard et al. (2014). Schröter et al. (2014) included the densities of cabins and of hiking paths, both can be considered anthropogenic inputs, to account for recreational cultural ES. Related approaches such as HANPP (human appropriation of net primary production, Haberl et al. 2012) calculate the human impact on ecosystem functionality and related ES supply and have been applied frequently. However, further efforts are required to close the research gap related to the inappropriate indication of provisioning ES.

With this study, we aim to fill this gap and explore and discuss additional indicators that cover the broad range of aspects listed above. Therefore, the overall objective is twofold: (i) to develop an enhanced indicator set for provisioning ES with an integration of the aspects: ES potentials, anthropogenic inputs, actual ES flows, environmental externalities, ES demands and preferences, spatial modelling and mapping of provisioning ES and (ii) to apply and test our enhanced indicator set in three different case study regions. We proceed as follows:

1.
Elaborate the needs for new provisioning ES indicators in European and international policy and decision-making contexts.
2.
Review and discuss existing indicators that relate to the different aspects with respect to their current applications, their merits and drawbacks.
3.
Based on this discussion, develop and suggest an enhanced and more holistic/integrative indicator set which covers multiple aspects of provisioning ES.
4.
Test this indicator set in three concrete case studies in Germany by applying a bio-economic farm model approach.
5.
Discuss the new indicator approach and its case study application results, and look at uncertainties and pros and cons of its practical application.
6.
Derive final recommendations for using such an approach in a wider European/international context in science, policy and practice, and indicate how far this approach contributes to understanding interrelations between agriculture, ecosystems and landscapes.

Hence, we will address the following four research questions:

RQ1: Do we need a new rationale to describe provisioning ES supply realised as the combined outcome of the natural ecosystem potential and anthropogenic inputs under management?
RQ2: Does a more holistic/integrative set of indicators, which includes all discussed aspects of provisioning ES, allow for a better indication?
RQ3: Is this indicator set feasible for practical application at case study level (e.g. data needs and availability, measurability, comparability, revealing interdependencies, allowing valuation of environmental externalities, expendability by additional indicators)?
RQ4: What are the implications for a wider application of the indicator set in a European/international context (e.g. transferability of data availability and of measurability)?

Our key assumption is that outputs from agro-ecosystems are not ‘pure’ ES per se. Instead, they are highly influenced by anthropogenic system inputs, depend on other ES, result in environmental impacts and are bound to demands and preferences of markets and society. An integrative set of indictors as suggested in this study increases the quality of information for policy and planning of land use and agro-ecosystem management. For example, integrative indicators can help to enhance site-specific management of complex agricultural landscapes and related governance mechanisms. An overview of the main parts of the manuscript, the considered aspects of provisioning ES and the aim of the approach is shown in Fig. 1.

European and international policy and decision-making contexts

ES have found a prominent place in current policy and decision-making, certainly at European Union (EU) and global levels. In the following, three examples are given in which an application of an enhanced indicator set of provisioning ES could be beneficial.

1.
EU Biodiversity Strategy 2020

The EU Biodiversity Strategy 2020 is aimed at achieving more sustainable agriculture and forestry in the EU (target 3) and, at the same time, to halt the loss of global biodiversity and ES by 2020^{Footnote 9}. This includes the maintenance and restoration of ecosystems (target 2) and the improvement of knowledge of ecosystems and ES by Mapping and Assessing the state of Ecosystems and their Services (MAES^{Footnote 10}) within EU member states (action 5). Furthermore, the economic value of ES should be assessed and integrated into accounting and reporting systems at EU and national level by 2020 (EC 2011).

For provisioning ES from agriculture, Maes et al. (2014) suggested to use indicators such as yields of food, feed, fibre or energy crops (in t/ha); crop area (in ha); or produced amounts of biofuel, biodiesel or bioethanol (in kToe). They were aware that the proposed indicators did not consider that ES from agro-ecosystems are co-produced based on ecosystem and human management inputs. They furthermore asked for an agreed approach to discount human inputs (such as labour, machinery, irrigation, fertilisation, pest control) and to identify the contributions of ecosystems to production (Maes et al. 2014). Thus, the enhanced indication of provisioning ES from agro-ecosystems including environmental externalities as suggested in our study can help to bridge this gap and to go beyond commonly used agricultural production numbers as proxies. It remains to be tested how far data availability and quality hamper the enhanced indicator set’s implementation on EU and national levels.

2.
European Agricultural Policy (CAP)

All current components of the Common European Agricultural Policy (CAP) contribute to compensate farmers for the costs of providing ES through agricultural activities (Baur and Schläpfer 2018). Recommendations for using an enhanced indicator set in the context of the CAP are given against the backdrop of scientific suggestions to improve the ecological effectiveness of environmental components (cross compliance, single and group contracted agri-environmental and climate measures (AECM), greening measures, e.g. ecological focus areas) of previous and current CAP periods (e.g. Batáry et al. 2015). On the other hand, such an indicator set could help to improve the implementation of the recently proposed programme for the next CAP period 2021–2027 (EU 2018; EC 2019^{Footnote 11}; Jongeneel 2018). An enhanced indicator set can contribute science-based information and data to the current discussion of the CAP improvements, which can be summarised into the following five suggestions.

The first suggestion concerns the integrated consideration of socio-economic and ecological aspects as agro-ecosystem services. Agro-ecosystems link natural and human systems and the goods and services they generate (FAO 2014). The different aspects of our enhanced indicator set (see below) reflect this view. The integrated view on socio-economic and ecological aspects of agro-ecosystems (cf. Mouysset 2017) is essential for development, adaptation and valuation of AECM. An integrated view can ensure that these measures are economically feasible for farmers, as a main driver for farm decision-making processes (Wolters et al. 2014) and that sustainable management practices can be adapted to site conditions, mitigate environmental impacts and fulfil rising societal demands for environmentally friendly produced agricultural goods.

The second suggestion argues for a better targeting of measures. Environmental objectives (cf. Meyer et al. 2015) and spatial targeting (cf. Reed et al. 2014; Ekroos et al. 2014) mean to set priorities for ES and biodiversity targets in particular areas for applying corresponding measures. Environmental objectives are targeted in the case of expected promising results and effective application of expenditures to maintain or restore these results. The spatial targeting can consider the variability of biophysical conditions, management costs, potentials to deliver expected results from environmental measures and the appropriate scale and boundary of the targeted environmental objectives (Reed et al. 2014). An enhanced indicator set can be used to identify objectives and appropriate areas, due to the capacity of indicators for inter- and intra-regional comparisons.

The third suggestion, the integration of all measures at the landscape level, is seen by many authors as a crucial step to reach ES and biodiversity objectives (cf. Leventon et al. 2017; Lefebvre et al. 2014; Prager et al. 2012). This requires the collaboration between different stakeholders (e.g. Prager and Freese 2009; Prager 2015) and the application of systemic approaches (e.g. Lescourret et al. 2015). This has already been implemented in several European countries (e.g. Westerink et al. 2017a; de Krom 2017; Franks and Emery 2013). Local knowledge on farm, landscape or regional levels (cf. Zasada et al. 2017) can help to achieve superior ES and biodiversity objectives (e.g. McKenzie et al. 2013) or to contribute to a green infrastructure (e.g. Maes et al. 2015; Schmidt and Hauck 2018). Interactions between on- and off-field areas are complex and depend in detail on landscape structures and the targeted species (e.g. Concepcion et al. 2012), supporting the idea of integration. Significant positive effects on regulating (e.g. Westerink et al. 2017b) or cultural (e.g. van Berkel and Verburg 2014) ES can be achieved by collaboration in a landscape with a clear common vision and a local facilitator (e.g. Prager 2015). Especially biodiversity and climate change targets are suitable to derive multifunctional effects by coordination (Galler et al. 2015), but there are also limits in providing multiple ES (Maskell et al. 2013). That means, regional aims should be carefully defined by experts and local stakeholders, e.g. by hybrid governance approaches (e.g. Velten et al. 2018).

The fourth point bundles suggestions which require continuous information sampling for evidence-based decision-making on different levels and which stimulate ongoing learning processes amongst stakeholders and decision-makers. These suggestions include (i) introducing more result-based measures (cf. Plieninger et al. 2012; Burton and Schwarz 2013; Herzon et al. 2018; Reed et al. 2014; Russi et al. 2016), (ii) accompanying measures by monitoring (e.g. Prager 2015), (iii) allowing flexible application of measures and introducing adaptive management (cf. Meyer et al. 2015; Hodbod et al. 2016), (iv) fostering farm advisory and knowledge-exchange (e.g. Meyer et al. 2015; Schomers et al. 2015) and (v) developing and applying knowledge and innovation systems (e.g. Bommarco et al. 2018). Based on these points, an appropriate information stream can enable farmers to apply measures in a flexible way and to adapt to changing frame conditions within viable agricultural production systems. The enhanced indicator set provides a framework that accompanies such information-based processes.

The fifth suggestion is to find tailor-made solutions for the EU member states at the regional and national levels, because agri-structural characteristics and natural conditions differ between the member states and cause various conditions for an implementation of an overall programme in the countries (cf. Öhlund et al. 2015) and regions (cf. Kirchner et al. 2016). The application of an enhanced indicator set can help to identify specific tailor-made solutions at different organisational levels and thereby support the aim to sustain the diversity of European agricultural landscapes at various scales (c.f. Lefebvre et al. 2014).

3.
Ecosystem/natural capital accounting

Ecosystem/natural capital accounting is aimed at integrated assessments of human-environmental interrelations by measuring ecosystems, their condition and ES flows from ecosystems into economic and other human activities (SEEA EEA^{Footnote 12}; Science for Environment Policy 2017^{Footnote 13}). The System of Environmental Economic Accounts (SEEA) is connected to the Systems of National Accounts (SNA) and part of the statistical systems in many countries of the world. Steps of ecosystem accounting include assessing ecosystem extent (e.g. agricultural land area), ecosystem condition, ES supply and use and monetary ES assessments. SEEA also recommends to consider disservices that emerge from agricultural land use. The overall aim is to understand the dependence of economic activities on ecosystems and their condition. Besides the SEEA framework, detailed SEEA guidelines for practical applications were developed on EU level (Science for Environment Policy 2017) and globally (United Nations 2014). SEEA and SNA recognise that cultivated systems are managed systems with high human inputs. To properly define and ensure consistency with the (agricultural) production boundary, (natural) ecosystem contributions must be distinguished from (cultivated) anthropogenic production inputs. The existing SEEA guidelines are currently under revision, and the suggested enhanced indication of provisioning ES in agro-ecosystems can help to develop applicable indicators.

Existing approaches of indicating provisioning agro-ecosystem services

Agro-ecosystem services, their provision and potential indicators for their quantification have been studied intensively. However, to understand the relations between ecosystems and agriculture, we should not only look into the production process with the potential of agro-ecosystems to deliver provisioning ES (the related anthropogenic inputs and resulting actual provision, e.g. yields) but also analyse the positive and negative externalities of agricultural land use on ecosystems. The spatio-temporal phenomena of these aspects are important, and they can be assessed by spatial modelling and mapping of provisioning ES. As the valuation of these impacts depends on the perception of consumers, regional stakeholders and society at large, the relation between agricultural production and the ES preferences and demands also plays a role. Therefore, we distinguish between six different aspects for the indication of provisioning ES and suggest specific approaches for their assessment:

ES potential of agro-ecosystems

The natural potential to deliver provisioning ES varies substantially between different agro-ecosystems. Site conditions, like radiation and temperature, the site-specific long-term soil quality in terms of productivity (depending on soil types and characteristics) and crop features are all factors that define the potential of an agro-ecosystem to generate crop biomass (according to Duru et al. 2015, completed by site-specific soil quality). Limiting abiotic factors (e.g. water, nutrients) and reducing biotic factors (e.g. weeds, pests, diseases) can be compensated by anthropogenic inputs. However, the assessment of the input shares of the ecosystem-based ES potential and anthropogenic inputs is challenging. Therefore, assessments are often aimed at the overall potential for agricultural provisioning ES, by assigning site-specific, mostly natural, relative differences in soil and climate characteristics, for example, by soil quality ratings (cf. Bünemann et al. 2018). One example is the soil quality rating index (Müller et al. 2007: M-SQR-Index), used also as part of national natural capital accountancies (Albert et al. 2016). Further indicators are used that report on the sensitivity and actual condition of agricultural ecosystems which can increase or reduce soil productivity, e.g. in respect to soil biota (Barrios 2007; Griffiths et al. 2018), pest control (Chaplin-Kramer et al. 2013) or levels of soil erosion (e.g. Meyerson et al. 2005).

Anthropogenic inputs

The provisioning services of agro-ecosystems are the result of anthropogenic inputs in combination with natural ecosystem conditions. The actual ES flow can be assigned to various shares of natural and anthropogenic inputs (Duru et al. 2015). Anthropogenic inputs from an ecosystem perspective (Kandziora et al. 2013b) result in changes of the balances of energy, water and matter and also in structural variables. They are embedded in a specific management system, dependent on the aims of the farming activities and the site conditions, different cultivated crops, crop rotations and locally adapted management practices (e.g. tillage/non-tillage operations, fertilisation practices, pest and disease management). Thus, the whole management system can be seen as an external, anthropogenic input that heavily influences the characteristics of an agro-ecosystem (Pérez-Soba et al. 2012). Furthermore, legacy effects can occur, when previous input management strategies still have an effect over time (cf. Rutgers et al. 2012). Often, more unspecific terms like low vs. high intensity (e.g. Tamburini et al. 2016), intensity gradients (e.g. Syswerda and Robertson 2014) or ecological vs. conventional farming (e.g. Williams and Hedlund 2013; Batáry et al. 2012) are used to indicate anthropogenic inputs. However, a more detailed focus on the effects of specific management practices on the provision of ES would help to develop sustainable management strategies, as concluded by Williams and Hedlund (2013). Indicators are available which give more detailed insights into management practices, for example, for pesticide application, as used by Sattler et al. (2007: standardised treatment index). The development of indicator sets, which allow (i) to assess the consumption of energy, water and other resources (nitrogen, pesticides, machinery) per produced unit, i.e. to assess them by intensity indicators (cf. Ruiz-Martinez et al. 2015), (ii) to assess balances (e.g. GHG-balances) and (iii) to compare different production systems regarding their inputs and related ES, i.e. efficiency indicators (e.g. energy and water use efficiency), should accompany management strategies for innovative, resource-efficient solutions, which are embedded in the agricultural system (cf. Wolters et al. 2014).

Actual ES flow from agro-ecosystems

In order to indicate realised ES flows, a strategy of three steps was suggested by Meyerson et al. (2005): assess (i) the extent of an ecosystem, (ii) the condition of an ecosystem and (iii) quantities of some flows of ecosystem-oriented goods. Following this logic, land use/cover data has been used as a first proxy or a capacity estimation for actual ES flows (e.g. Burkhard et al. 2009, 2012b, 2014). However, this proxy exposes uncertainties (Hou et al. 2013) and should be completed (Van der Biest et al. 2015) by more detailed assessments (cf. Meyerson et al. 2005). Often, crop and grassland cultivated areas or crop yields (tons/energy per year and unit land) and livestock data (numbers or livestock units per unit land, tons/energy per year and region) (e.g. Fridman and Kissinger 2018; Kandziora et al. 2013a; Balbi et al. 2015) are primarily used for indication (Maes et al. 2016). Only a few studies use aggregated indicators like grain equivalent units (Koschke et al. 2013), which allow to compare agricultural production between regions. However, it is arguable whether such an indication, solely through used area and achieved yields, is adequate without taking anthropogenic inputs and external impacts into account, because it may not comply with the basic idea of the ES concept, which is to safeguard natural capital while maintaining sustainable flows of ES from nature to society (Burkhard et al. 2012a). Furthermore, the net primary production (cf. Haberl et al. 2012: NPP; e.g. Kandziora et al. 2013a) was used. A few studies include farm economic indicators, like sales and farm income (e.g. Crossman and Bryan 2009; Koschke et al. 2013; Kirchner et al. 2015; Firbank et al. 2018), but the factor costs, although necessarily assessed, because the farm income is based on a difference between sales and costs, are not explicitly described in most studies. In case the factor costs were made explicit, they belonged to the aspect anthropogenic input. An energy balance approach for the assessment of the ES flow of agro-ecosystems was suggested by Pérez-Soba et al. (2012). Indicators representing quality aspects of provisioning ES were applied for forage production (e.g. Van Vooren et al. 2018) and orchards (e.g. Demestihas et al. 2017). The quality aspects of crop production (cf. Wang et al. 2008), although important for the establishment of product prices, have not yet been represented by provisioning ES indicators. Other indicators include the actual amounts of marketed products for consumption (e.g. harvested biomass actually sold, cf. Geijzendorffer et al. 2017). Sometimes, these figures were corrected by the amounts of products that are spoiled and disposed of before consumption (e.g. Rasmussen et al. 2016).

Environmental externalities of provisioning ES

An ‘informed management’ emphasizes the mitigation of negative environmental impacts and the enhancement of regulating and cultural ES (Pérez-Soba et al. 2012). To assess the environmental impacts of provisioning ES, either positive (win-win/synergy; e.g. Bareille and Letort 2018; Daryanto et al. 2018; Everwand et al. 2017) or negative (trade-off; e.g. Gissi et al. 2018) externalities of production are considered (e.g. Bennett et al. 2009; Howe et al. 2014). For an assessment of environmental externalities, the identification and operationalization of indicators are two important steps (Kanter et al. 2018). Indicators cover a wide variety of possible externalities on different spatial scales (e.g. Williams and Hedlund 2013; Balbi et al. 2015; Schulte et al. 2014). For regulating ES, mainly information of soil-related ES is available (e.g. soil erosion control, nitrogen fixation), followed by indicators for pollination (e.g. pollinator abundance, pollination potential), whereby for cultural ES (e.g. rural tourism), only few indicators are available (Maes et al. 2016). Often, the externalities are assigned to broad categories of agricultural production, but their assessment should focus on specific management practices (Williams and Hedlund 2013), e.g. Garbach et al. (2017) and Techen and Helming (2017). As the scope of possible externalities is large, analysis must focus on a number of well-defined externalities that depend on societal perception (Rodríguez et al. 2006; Kroeger 2013). Generally, most assessments are based on agricultural intensity indicators (cf. Firbank et al. 2018), whereby site conditions, management practice, ES flow and site-specific sensitivity are all important determinants for environmental impacts (e.g. Albert et al. 2016; Sattler et al. 2010; Tsonkova et al. 2015).

For an assessment of externalities of provisioning ES for biodiversity aspects and regulating and cultural services within agricultural landscapes, the landscape structure in general (e.g. Kleijn et al. 2011), specific landscape elements (e.g. Firbank et al. 2018), interrelations between them and management practices (e.g. Tamburini et al. 2016) and their relations to site-specific sensitivities should be considered. For biodiversity aspects, these facets should be related to the habitat requirements of taxonomical and functional species groups (e.g. Liere et al. 2017; Birkhofer et al. 2018). Thus, the development of indicators is challenging and assessments often use proxies (e.g. Andersen et al. 2013) and/or scorings (e.g. Firbank et al. 2018; Tzilivakis et al. 2016; Overmars et al. 2014). Regulating ES that are used by farmers (e.g. natural pest control, pollination, soil erosion control) require accurate specific indicators (e.g. Rusch et al. 2012; Steinhoff-Knopp and Burkhard 2018) and are ideally related to landscape patterns (e.g. Duarte et al. 2018). Broader assessments of environmental quality are part of farm performance evaluations (e.g. Firbank et al. 2018) or describe developments of agricultural landscapes (e.g. Björklund et al. 1999).

ES demand and preferences

That ES supply is driven by demand holds especially true for the actual ES flow of provisioning ES from agro-ecosystems, which are mainly marketed products. Thus, this demand is related to consumer interests. Positive externalities of agriculture are usually driven by regional stakeholder or societal interests, e.g. provisioning of habitats for or of cultural ES of agricultural landscapes. Also, there is a demand to align the agricultural production in order to mitigate their negative externalities. Consumer, regional stakeholder and societal demand and preference assessments that regard provisioning as well as regulating and cultural ES from agro-ecosystems, their functions and structures help stakeholders (amongst them farmers) and politicians to decide on suitable agricultural management strategies to support the ES in high demand. Such integrated management strategies consider that agricultural products and non-marketed ES are jointly produced (Kragt and Robertson 2014; Tsonkova et al. 2015; Huang et al. 2015; Klapwijk et al. 2014).

Demand and preferences for consumer interests can be indicated through analyses of consumption patterns (i.e. consumption rates related to population density, cf. Villamagna et al. 2013). Additionally, the type or quality of marketed products is often used for indication at the individual level (e.g. share of organic and regional products, cf. Rödiger and Hamm 2015; Feldmann and Hamm 2015; Hempel and Hamm 2016). Willingness-to-pay (WTP) or willingness-to-accept (WTA) studies, employing stated or revealed preference analyses, can also be used for assessing regional stakeholder interests. These studies have to deal with the difficulty that people do not always recognise the capacity of ecosystems to provide ES (Martín-López et al. 2012). Demand and preferences are also influenced by availability of appropriate substitutes (cf. Rasmussen et al. 2016). Scarcity is another issue which is relevant in this context, as scarce resources are typically in higher demand (cf. Meyerson et al. 2005). Scarcity is often reflected in market prices driven up by high demand (e.g. Geijzendorffer et al. 2017).

Farmers and other regional stakeholders can benefit in manifold ways from specific locally provided ES, and these provided ES can, for example, influence the farmers’ perception of ES (Smith and Sullivan 2014; Teixeira et al. 2018). Generally, an individual demand of regional stakeholders can be quantified by methods like food diaries, group interviews, participant observations or surveys (Rasmussen et al. 2016) in specific studies or panels. The demand to mitigate negative externalities can be assessed by the amount of needed regulation to meet a desired environmental quality (cf. Villamagna et al. 2013). From an ecosystem perspective, the ecological work which is needed to achieve the socially defined environmental quality under a given ecological pressure determines the needed regulating service flow and depends on the regulating capacity of the ecosystem (Villamagna et al. 2013).

Apart from expressed demand at the consumer and regional stakeholder level, demand can also be expressed at the societal level through existing agri-environmental policies (cf. Schulte et al. 2014) or policies aiming at avoiding environmental risks (Wolff et al. 2015, 2017). These policies define indicators specific to their aimed and regularly reported environmental goods (e.g. Water Directive 2000 (EC 2000)^{Footnote 14}, fauna flora habitat directive (EC 1992)^{Footnote 15}).

Relations of ES demands to ES supply could be expressed as an ES footprint, defined as the area needed to generate particular ecosystem goods and services demanded by humans in a certain area at a certain time (Burkhard et al. 2012b). There are also studies which show the spatial variation of ES demand through mapping approaches (cf. Wolff et al. 2017). ES demand is, compared to ES supply, still insufficiently researched (cf. Geijzendorffer et al. 2017).

Spatial modelling and mapping of provisioning ES

As a cross-cutting issue, all indicators discussed above (aspects 1–5) show spatial and temporal variations. This can be best visualised by spatial modelling and mapping of provisioning ES. To show the spatial variation, single thematic ES maps can be used. To highlight temporal variations, a whole series of maps displaying changes over time is more appropriate (e.g. to show seasonal or annual variability). In dependency on the ES in question, different resolutions might be adequate.

For spatial modelling and mapping of provisioning ES supply, different data sources and methods can be used (Burkhard and Maes 2017). This includes look-up tables, based, for instance, on aggregated statistics and spatial interpolation, which are often used to display actual ES flow mapping (e.g. maize equivalent yield). Furthermore, causal relationships and environmental regression approaches, which are often used for regulating and cultural ES, can be used to assess environmental externalities (e.g. sequestered carbon) (Schröter et al. 2015). In a few studies, spatial ES supply concentration was indicated by assessing hot or cold spots (e.g. Früh-Müller et al. 2016) or temporally high ES supply can be indicated as a hot moment (Burkhard et al. 2014).

In principle, spatial modelling and mapping of ES can be done for all indicators described for the different aspects discussed above in sections of aspects 1–5.

An example for a mapped ES potential can be found in Albert et al. (2016), who suggest to use the yield potential of arable soils (BGR 2013) based on Müller et al. (2007) for national accountancies. Anthropogenic inputs are mapped mainly by intensity indicators, for example, the management intensity of grasslands (Estel et al. 2018) or agricultural land (van der Zanden et al. 2016) at a European scale. Mapped crop rotations (Koschke et al. 2013) depict the time and organisational processes of farms (Grunewald et al. 2013) and are important for the extent of environmental externalities, e.g. soil erosion (Guerra and Pinto-Correia 2016), habitat quality (Glemnitz et al. 2015). Actual ES flows in terms of crop yield and yield uncertainties are mapped on a regional scale (e.g. Balbi et al. 2015). A mapping of environmental externalities of provisioning ES considers trade-offs to regulating (e.g. Balbi et al. 2015), habitat (e.g. Willemen et al. 2012) or cultural ES (e.g. van Zanten et al. 2016). On an agricultural landscape scale, trade-offs between provisioning ES and habitat or cultural ES can be analysed by using landscape metric indicators, i.e. fragmentation, diversity, habitat connectivity, habitat richness and landscape heterogeneity (e.g. Frank et al. 2012; Koschke et al. 2013; Duarte et al. 2018; Kay et al. 2018). Geostatistical indicators can be used to explore patterns of ES (Ungaro et al. 2014, 2017). ES demand is regarded only by a few studies that for instance indicate a mismatch between ES supply and demand by matrix approaches, considering also provisioning ES (e.g. Burkhard et al. 2012b, 2014). Mostly, ES maps refer to the spatial variability of ES with only a few approaches that consider the temporal variability of ES supply (cf. Burkhard et al. 2014).

Overall indicators

In addition to the above listed indicators, also combined approaches which cover several aspects have been suggested. For instance, the EBI (Ecosystem service Bundle Index) was introduced by Van der Biest et al. (2014) to combine biophysical (ES supply-oriented) and socio-economic (ES demand-oriented) aspects.

When using indicators of all aspects, the following must be kept in mind: to avoid misinterpretation of indicators, it is important to explicitly describe the metadata of the analysis, e.g. objective, system-boundaries; characteristics of used data (data preparation, data resolution, data assessment aim); methods; uncertainties (e.g. Grêt-Regamey et al. 2014; Schulp et al. 2014) and scale (Hein et al. 2006; Haines-Young et al. 2012). For example, system boundaries play a key role in several aspects of provisioning ES. An example for a spatial boundary is that the results of GHG-emission of a production process differ between a calculation which only takes into consideration the emissions which occur during management on a field and a calculation which additionally considers the emissions in the upstream chain (e.g. mineral fertiliser production). The temporal boundary is related to the fact that crops are affected not only by management strategies and climate during the growing period but also by long-term soil processes related to the production of previous crops (Angus et al. 2015; Preissel et al. 2015). Thus, an assessment at the crop rotation scale (i.e. the rotation is the temporal system boundary) allows to quantify crop yield (which belongs to actual ES flow) while considering also long-term processes (Reckling et al. 2016).

Development of the enhanced indicator set for provisioning ES of agro-ecosystems

It is obvious that the generation of provisioning ES in human-modified agricultural land use systems is strongly dependent on natural and human-derived inputs (Jones et al. 2016; Power 2016). Respective indicator systems should therefore reflect these aspects by distinguishing between natural and anthropogenic contributions as well as by informing about environmental impacts. They should furthermore regard interests of consumers, regional stakeholders and the whole society, as well as describe spatio-temporal phenomena.

Conceptual model

The conceptual model in Fig. 2 gives an overview of ES flows from agro-ecosystems to society (after Burkhard et al. 2014; Bastian et al. 2013; and the ES ‘cascade’ model by Potschin and Haines-Young 2016). The role of anthropogenic inputs and ES potentials is illustrated in the left part of the ‘ES supply box’ in the centre. Thus, the ES potential is enhanced (as in the case of provisioning ES in agro-ecosystems) by anthropogenic inputs, activating a flow of usable ES that eventually benefit the human population and economic activities, which demand certain ES. The harnessing of ES and especially the optimisation of the supply of selected provisioning ES in agro-ecosystems based on anthropogenic inputs lead to trade-offs between ES, degradation of natural capital and environmental impacts (Rodríguez et al. 2006). The identification, quantification and assessment of trade-offs (and synergies) between ES and environmental impacts of human land use systems are one of the key strengths and a major application potential of the ES concept (Foley et al. 2005).

Suitable indicators are needed to quantify and to communicate the relevance and effects of the different components in the conceptual model. In this article, we will focus on (1) ES potentials, (2) anthropogenic inputs, (3) actual ES flows, (4) environmental externalities of provisioning ES, (5) ES demands and preferences and (6) spatial modelling and mapping of provisioning ES in agro-ecosystems by using the example of food provision. In most of the currently available provisioning ES indicator sets, the aforementioned components are summarised in one indicator, which normally is ‘yield in tons/area and time’ (see ‘Introduction’, part three, aspect three).

The distinction between ES potential, anthropogenic inputs and actual ES flow, altogether the central part of our conceptual model, is shown in Figure 3 (according to Duru et al. 2015, see also ‘Introduction’ part three, aspects one to three). In the conceptual model, we refer to an overall ES potential, defined by the site conditions (such as climate, long-term soil quality) and crop features. In a long-term perspective, the ES potential can be enhanced, for instance based on technological developments. The actual ES flow is defined by the ES potential, which can be limited by abiotic (e.g. water, nutrients) and biotic (e.g. weeds, pests, diseases) factors and which depends on the choice of the crop genotype and management as anthropogenic inputs. Furthermore, the choice of crop and land management strategies are important factors which determine whether and to which degree the agro-ecosystem can be considered self- or anthropogenic-regulated and how large the environmental impacts are.

Requirements to indicators of the enhanced indicator set

The following requirements are needed for indicators of the introduced aspects of provisioning ES (see ‘Introduction’, part three). Different levels of indicator integration are necessary to fulfil the requirements of those indicators, which refer to the ES supply (aspect one to four) and demand (aspect five). The levels of integration of the indicators range from detailed biophysical and socio-economic data on the site condition and farming activities up to complex and highly integrated data on environmental externalities (Fig. 4). Most of those indicators can be expressed by biophysical, non-monetary values to represent the ecological side of ES generation and to quantify the flow of matter and energy (see left side, Fig. 4) and by monetary values to represent the economic side and to quantify costs, sales and income (right side, Fig. 4).

First, to achieve transferability of the indicators, the basic datasets should ensure that the indicators are based mainly on quantitative, harmonised, available and if possible, spatially explicit datasets which meet the requirements of the ‘high-quality data label’ of ES indicators according to Maes et al. (2016). In the following sections, such data are referred to as ‘data of high-quality requirements’. That means regional analyses of agricultural provisioning ES start with (i) datasets that describe site conditions to define the ES potential, based on widely accessible geodata, and (ii) the crop cultivation and livestock data which are, for example, available from agricultural census or from the Integrated Administration and Control System (IACS)^{Footnote 16}.

Second, a detailed regional assessment of the management practices can be achieved by quantitative, non-aggregated indicators which are used to describe the anthropogenic inputs and the actual ES flow. The indicators are based on regional cultivated crops, livestock and management practices and can be expressed as biophysical values to assess the input (e.g. fertiliser products, amount and times of fertilisation, pesticide products, amount and times of pesticide application, use of fuels, machinery, applied working time) and the actual ES flow (e.g. crop yields), or as monetary values to assess the input (e.g. costs of inputs) and the actual ES flow (e.g. crops sales). For the anthropogenic input, not only the mentioned direct inputs but also indirect inputs (e.g. educational level of farmers) should be considered.

Third, to achieve comparability between different regions, quantitative, aggregated indicators are needed. Aggregated indicators can assess the anthropogenic input and the actual ES flow (as output), also in the case of regional different crop patterns and animal husbandry. The input and output can be made comparable through aggregated budgeting of provisioning ES. That means, for example, in terms of biophysical values to calculate aggregated numbers of fertilisation, grain equivalent unit yields and in terms of monetary values to calculate total factor costs and total sales.

Fourth, to relate input and output of agricultural production, quantitative, relational indicators can be used. They relate output to input, i.e. actual ES flow to anthropogenic input or express balances, i.e. N balances. Indicators, for example, expressed in biophysical values, are energy and water use efficiency and, in monetary values, farm income. Indicators of this category can mainly be used to assess resource efficiency and show ways for reducing environmental impacts and simultaneously maintain or improve actual ES flows.

Fifth, highly integrated and index-coded indicators can be used to assess environmental impacts. Ideally, these indicators are based on applied management practices to ensure that environmental externalities of management practices can be compared and implemented and that environmentally friendly practices can be monitored by the indicators.

Sixth, indicators for ES demand and preferences can reflect consumer interests and local and regional stakeholder interests—in this case, they are often derived from stakeholder processes or have a normative character to fulfil societal demands.

Enhanced indicator set

The comprehensive indicator set for assessing provisioning ES from agro-ecosystems captures the aspects that are described in the following. The respective indicators related to these aspects are presented in Table 1 and follow the requirements defined in the previous section. Major parts of the indicator set were utilised in the case study applications (Table 1 in bold).

Table 1 Indicators for provisioning ES, example for cultivated crops for human nutrition (indicators in bold were quantified in the case studies)

Full size table