The classification of ecosystem services has experienced a remarkable evolution from the Millennium Assessment 2005 to the latest versions of the Common International Classification of Ecosystem Services (Haines-Young and Potschin 2018), the Intergovernmental Platform on Biodiversity and Ecosystem Services (Brondizio et al. 2019), the National Ecosystems Classification System (US Environmental Protection Agency 2020) and The Economics of Ecosystems and Biodiversity (TEEB 2010). In INCA, we refer to the CICES classification, also mentioned in the SEEA EA (United Nations 2021). When assessing ES from an accounting perspective, we found that ecosystems can behave in various ways in delivering services for human needs. We look at two major groups (accounting clusters):

• Cluster 1 (sustainable thresholds) - for some ES, management practices could exceed sustainability thresholds: this is the case of provisioning services when the regeneration rate of natural resources is exceeded and resources cannot be regenerated anymore (e.g. the unsustainable extraction of wood). It also happens for services, such as water purification, where the exceeding absorption rate prevents ecosystems from an effective removal of pollutants, which in turn leads to ecosystem degradation. The former belongs to CICES section “provisioning ecosystem services”, the latter belongs to CICES section “regulation and maintenance services”.
• Cluster 2 (presence/absence of ecosystems)- for other ES, what matters is the presence of suitable ecosystems to generate the ES where they are needed. For example, the presence of: suitable habitats to host species (e.g. crop pollination and habitat and species maintenance); vegetation able to reduce the magnitude of matter and/or energy, contributing to flood control and avoiding soil erosion; natural attractive and accessible areas that offer the possibility for residents to enjoy outdoor activities (nature-based recreation). All these ES belong to CICES section “regulation and maintenance services”, while nature-based recreation is a “cultural” service.

Fig. 1 summarises the two ES accounting clusters:

Figure 1.  

Ecosystem services clustered by: (1) the existence of sustainability thresholds in management practices and (2) the presence or absence of suitable ecosystems (adapted from La Notte et al. (2019b).

1. the ES cluster including source-provision (regeneration rate) and sink (absorption rate) where the ES that is actually used can be higher than its sustainable use because resource extraction or pollutants emission exceed sustainable thresholds;
2. the ES cluster including source-suitability (e.g. existence of suitable habitats), buffer (e.g. protective role of vegetation in reducing threats) and information (e.g. opportunity to enjoy nature) where there can only be an actual use of ES.

Clustering of ES has strong implications when accounting for ES capacity in monetary terms. ES capacity can be defined as the ability to keep on generating ES in the future. As suggested in SEEA EA (United Nations 2021, United Nations 2019), the Net Present Value (NPV) is calculated using the ES actual flow. This procedure is correct in case that what matters is presence/absence of ecosystems (cluster 2); on the other hand, it can be misleading when the ES can be overused (i.e. for the first accounting cluster) as decribed in La Notte et al. (2019a). For provisioning (e.g. timber provision) and sink (water purification) ES, calculating the NPV of the actual flow (when exceeding sustainability thresholds) implies that a higher capacity is computed due to unsustainable management practices, eventually leading to ecosystem degradation. If capacity is computed from exceeding ES use, the message to policy-makers would be misleading in the following ways:

• there is no advantage for adopting sustainable practices (sustainable ES flow would lead to lower capacity);
• there is no alert mechanism for unsustainable policies: once the ES actual flow collapses (due to ecosystem degradation), it will be too late to act.

First lesson learned: For accounting purposes and, especially, when it comes to ES capacity, ES should be clustered correctly, according to the existence of a sustainability threshold in management practices (cluster 1) or the presence of ecosystems (cluster 2).

When using ecological models to assess ES, the resulting ES flow recorded in the supply and use table is, in turn, the result of the interaction between the ecological potential supply and the effective socio-economic need for each service. The match between the ES potential (ecological side of supply) and the ES demand (socio-economic side of demand) generates the ES actual flow (Fig. 2).

Figure 2.  

Conceptual scheme of the interaction between ecological (i.e. supply) and the socio-economic (i.e. demand) frames underpinning the quantification and accounting of ecosystem services (actual flow).

The interlinkage of ecological and socio-economic frames in the generation of ES actual flows plays an important role when interpreting the trends of ES flow over time. Fig. 3 reports the overall difference of crop pollination between 2000 and 2012. The pollination ecosystem service increased from 2000 to 2012 by 12.6%. At first glance, this seems a positive change. However, we found an increased demand for pollinator-dependent crops in 2012, which was not mirrored by an increase in suitable pollinator habitats (which actually decreased by 1.4%). This means there is a smaller habitat extent suitable to host wild pollinators and the ones that exist are under a growing pressure. This is a strong message for policy- and decision-makers for agricultural land management and ecosystem restoration strategies.

Figure 3.  

Changes in the crop pollination service in Europe (2000 and 2012).

Another example is provided by nature-based recreation (Fig. 4): a 20.4% increase from 2000 to 2012 is remarkable. However, almost half of such change (9.4%) can be explained by population increase in the EU (and, therefore, more people requiring nature-based recreation), while the other half (11%) is the real increase of the service (i.e. natural areas nearby human settlements).

Figure 4.  

Changes in nature-based recreation service in Europe (2000 -2012).

Second lesson learned: Two components interact when assessing ES actual flow, i.e ecological components assessed through the ES potential and the socio-economic components assessed through the ES demand. A correct interpretation of changes on ES use over time needs to be analysed considering the role of each side.

The socio-economic frame simplified in Fig. 2 is multifaceted. It can represent economic sectors such as agriculture (for pollination and soil retention) or households (for nature-based recreation). In both cases, the biophysical mapping of ES demand coincides with where economic assets and human settlements are located. Users of these ES are labelled "domestic", becasue they are within their national boundaries. The accounting of ES stops at its first users (e.g crop provision flows from the cropland ecosystem type to the agricultural sector). The transformation and trading of the products eventually generated is accounted in the conventional system of national accounts.

However, there are ES that contribute to addressing overarching environmental targets, such as climate change mitigation (by, for example, carbon sequestration) or halting biodiversity loss (habitat and species maintenance). For these ES, the global society can benefit from the services provided, regardless of where the service is generated (Fisher et al. 2009, Syrbe and Walz 2012Burkhard and Maes 2017). The concept of national boundaries does not apply to this type of ES, hence their beneficiaries/users are referred to as 'global'.

Fig. 5 confronts the two cases:

Figure 5.  

Ranking of domestic (a) and global (b) ecosystem services (€/Km²/year), in 2012 at European level.

• map (a) ranks the ES monetary value (€/km²/year) of a few domestic ES – crop and timber provision, pollination and nature-based recreation per EU NUTS2 regions with respect to a reference value, that is the EU average;
• map (b) ranks the ES monetary value (€/km²/year) of carbon sequestration per EU NUTS2 regions with respect to a reference value, that is the EU average.

Though some regions (see, for example, Finland and Sweden) seem relatively poor in terms of domestic ES, they may be rich in terms of global ES. For instance, also the presence of forests in (for example) Eastern Europe and some regions in France and Spain completely change the economic value distribution of ES across the EU. This outcome becomes key to address international debates on the role of countries or continents to address overarching environmental issues that will inevitably affect the planet for the future generations. ES for global society represent a sort of ecological public good and needs to be treated differently from other ES in accounting.

Third lesson learned: For most ES, the demand is spatially defined and can be allocated to specific users; however, for overarching environmental targets (such as climate change and biodiversity loss), the ES demand is represented by the global society, that cannot be spatially located to a specific place and exclude other places.

As shown in Fig. 2, when ES potential, which corresponds to the ecological supply and ES demand match, the ES actual flow is generated. However, there might be cases of unused ES due to lack of demand or oversupply and cases of unmet demand caused by a lack of ES providing the potential supply needed. While the existence of unused ES potential generates no harm, ES demand not covered by ES provision generates vulnerabilities, which can be explained as ecosystem deficit or lack of ecosystem contribution to human needs.

For example, the case of flood control (Fig. 6) shows that there are areas able to provide the flood control service. When they interact with areas which need protection from flooding, the ES actual flow can be recorded (ES match); when this interaction does not take place (ES mismatch), ES demand is unmet and, therefor,e the areas remain vulnerable.

Figure 6.  

Flood control service: match and mis-match between supply and demand in Europe (2012).

ES unmet demand can occur for ES that belong to the second accounting cluster (ref. Fig. 1) whose users can be spatially located (i.e. domestic users).

Fourth lesson learned: The mis-match between ES potential and ES demand can generate ES unmet demand.

A mis-match between the ecological and the socio-economic side may occur also for the ES which belong to the first accounting cluster (Fig. 1). What matters for the quantification of these ES is the sustainability threshold. The ES demand will be met at the expense of the ecological side and, thus, the actual flow will cover the whole ES demand, even when it exceeds the natural regeneration or absorption rate. However, it is possible to set a threshold to determine which part of the ES actual flow is sustainable and which part of the ES actual flow represents an overuse.

An example is provided for water purification (Fig. 7). In this case, the sustainability threshold is based on literature as the eutrophication level of 1 mg/l (Camargo and Alonso 2006).

Figure 7.  

Water purification service: match and mis-match in Europe (2012).

Fig. 7 shows that almost all water purification actual flow in 2012 is unsustainable. ES overuse represents another vulnerability because it will eventually lead to freshwater ecosystem degradation and all its consequences for human health and well-being.

The fifth lesson learned: The mis-match between ES potential and ES demand can generate ES overuse for source provision and sink services with serious consequences for sustainability targets.

A third mis-match that may occur between ES potential and ES demand concerns ES that belong to the second accounting cluster (Fig. 1) having the global society as users. These ES are not directed to a specific geographic area or to economic assets or inhabitants of a country. Scientific knowledge shows how the decrease in ES targeting climate change mitigation (e.g. carbon sequestration) and halting biodiversity loss (e.g. habitat and species maintenance) will harm the whole planet with all its inhabitants (Dasgupta 2021, Pachauri and Meyer 2014, Brondizio et al. 2019).

In this case, the mis-match is due to ES that are lost due to inappropriate human practices, for instance, in managing the territory. Fig. 8 proposes the example of carbon sequestration. The total potential removal of carbon in 2012 is 444 mln tonnes of CO₂. However, due to ecosystem emissions, the net removal of CO₂ is 306 mln tonnes. Global society is missing out on 138 mln tonnes of CO₂ that could have been sequestered, due to the unsustainable management of ecosystem types such as cropland, grassland and wetlands, which generate CO₂ emissions exceeding the (CO₂) removal capacity from the atmosphere (Fig. 8).

Figure 8.  

Carbon sequestration service: match and mismatch in Europe (2012).

It is necessary to specify that what is assessed as ES missed flows refers to the current context and not to optimal scenarios, i.e. the ES flow missed with respect to what is currently happening and not with respect to what should ideally be happening. The assessment of optimal ES provision would require additional processing in terms of concept development and modelling, which is not in the current domain of INCA.

The sixth lesson learned: The mismatch between ES potential and ES demand can generate ES missed flows.

In the SEEA EA, condition accounts and ES accounts are connected in a sequential logic chain; however, the SEEA EA framework does not provide guidance on how to establish a practical linkage between these two sets of accounts, which run in parallel.

INCA applications show that the linkage of ES accounts with variables used to calculate indicators of condition accounts is possible, feasible and desirable. Habitat and species maintenance clearly show how to establish this linkage (Fig. 9).

Figure 9.  

Habitat and species maintenance in Europe: from ecosystem condition to service (2012).

Habitat suitability is calculated to assess the presence of habitats in favourable conditions. Together with the indicator of bird species hotspots, habitat suitability is used to locate areas able to provide the ES habitat and species maintenance, for which people are willing to pay for its non-use value (Fig. 9). The combination of habitat condition with the presence of key species allows mapping four different zones to which different values will be attributed. Any change in the condition indicator will affect the assessment of the ES areas (distribution and quantities of the four zones) that, in turn, will affect the ES economic value. Condition accounts and ES supply and use table are, in this case, not running in parallel, but are linked by a cause-and-effect relationship, as long as the condition variables overlap with input of ES biophysical modelling.

The seventh lesson learned: There can be a linkage between ecosystem condition accounts and ES supply and use tables, when the variables chosen to compute the former are input variables for the assessment of the latter.

The SEEA EA complements the System of National Accounts (SNA) and the SEEA Central Framework (SEEA CF, United Nations et al. 2014b). In the SEEA CF, some of the resource accounts are reported as asset accounts. For example, timber accounts report both tables on forest land and on the volume of timber.

The work on wood provision in INCA highlighted that understanding whether and how this ES differs from or overlaps with what is proposed to be reported by the CF is a critical issue. After a few iterations, the latest version of this ES shows that the ecosystem contribution can be assessed as the net annual increment of wood biomass in cultivated forests/forest available for wood supply already reported in the timber account of the CF. This flow should not be confused with felling*4 or removals, because (as graphically simplified in Fig. 10) it takes place between different economic units sequentially linked.

Figure 10.  

From the timber service provision (EA) to forest environmental accounts (CF) to conventional economic accounts (SNA).

Another important example is the ES crop provision. The SNA reports crop production that includes both ecological and human inputs. A high crop production does not necessarily imply a high delivery of the service (from ecological inputs). In fact, high yields are often achieved with the use of artificial inputs such as fertilisers, plant protection products and machines (Perez-Soba et al. 2019). Disentangling ecosystem contribution from gross crop harvesting can: (i) clearly show the added value of introducing ES accounting (otherwise useless for provisioning ES, since agro-forestry harvests are already part of the SNA) and (ii) provide valuable information for policies directed to support sustainable agricultural practices.

Fig. 11 shows that areas with high yield growth: (a) often overlap with areas of low ecosystem contribution; (b) ecosystem contribution is calculated by using an Emergy-based ratio (Perez-Soba et al. 2019) that is multiplied by the crop yield to estimate the ES flow*5.

Figure 11.  

Total yield versus ecosystem contribution in Europe (2008).

Integrated accounting systems need to be coherent and consistent. This is the intrinsic feature that marks the difference between accounting and reporting. When composing and filling the accounts in INCA, each module is combined with economic and natural resource accounts in a way that does not create inconsistencies or discrepancies.

The eighth lesson learned: Ecosystem accounts can complement in a consistent and harmonised way economic and natural resources accounts, by making the whole accounting mechanism fully coherent.

INCA builds on the MAES initiative. The purpose of MAES was to map and assess ecosystems and their services without any specific accounting purpose. When using some of MAES outcomes in INCA, it becomes clear how the availability of a biophysical model is only the starting point and not the final outcome.

For example, in the case of nature-based recreation, ESTIMAP (Paracchini et al. 2014) provided the recreation opportunity spectrum as the key source of information to map ES potential. However, the accounting procedure required further processing. In fact, ES assessment required: (i) to identify the epicentres of nature-based recreation (rather than having a score allocated to each pixel) and (ii) to set buffers around these epicentres. Fig. 12 shows how the assessment of nature-based recreation starts from the ESTIMAP outcome: (a) where only some areas are considered for the assessment of the ES potential; (b) to finally calculate the ES actual flow; (c) where the ES potential interacts with ES demand, which needs to be correctly mapped.

Figure 12.  

Nature-based recreation in Europe (2012): from the recreation opportunity spectrum (a) to the ES potential (b).

The users identified for nature-based recreation in INCA are local residents and what matters in ES accounting is the distance from the nature-based hotspots they can reach on a daily basis.

Ninth lesson learned: When using already existing biophysical models for ES accounting, they will not directly provide the actual flow. Three steps are likely to be needed to adapt existing models: (i) building the ES potential as appropriate; (ii) identifying ES demand; (iii) combining ES potential and ES demand to assess the actual flow.

Previous classifications of ES used to define some of them as intermediate in order to avoid double counting of the same flows in the sum of ecosystem service flow for an accounting area. The “intermediate” tag often applies to “regulation and maintenance” services especially when their contribution is considered embedded in a final benefit used as a proxy for a given provisioning ES.

INCA applications demonstrate how the default definition of an ES as “intermediate” should not be used. In fact, treating an ES as intermediate or final depends on the methods (and the purpose) practitioners use to assess ES.

The intermediate flow may unfold “vertically” when many ecological flows merge into a single ES flow that embeds them all. This is the case, for example, for crop provision. Fig. 13 visually simplifies two cases: when (a) there is an agricultural production function that integrates all ecological inputs (such as water supply, pollination, soil retention); and when (b) each ES is modelled separately from the others and, thus, directly enter as final ES and is allocated to the agricultural sector.

Figure 13.  

Crop provision: how to single out different services from one flow.

The intermediate flow may unfold “horizontally” when the same ES flow is provided in sequence by different ecosystem types. This is the case of inter-ecosystem flows. Fig. 14 visually simplifies the two cases that can apply to water purification: in case (a), the outcome of the integrated model includes both the pollutants retained in the soil of the basin and the pollutant removed by rivers and lakes; in case (b), there are two separate intermediate flows: basin retention and rivers and lakes retention.

Figure 14.  

Water purification: how to single out one flow from different ecosystem types.

Tenth lesson learned: Any ecosystem service cannot be defined by default as “intermediate” because the intermediate or final role ES play depends on the assessment technique that is applied.

In SEEA EEA (United Nations et al. 2014a), inter-ecosystem flow is one of the possible service flows that may occur within and between ecosystems. The second service flow is intra-ecosystem flow: it applies when a service is provided within the same ecosystem type which is at the same place and likely contributes to maintain ecosystem condition.

For example, the ES on site soil retention is provided by almost all terrestrial ecosystem types. Only when provided on cropland is it accounted as ES and allocated to the agricultural sector; when provided by other ecosystem types, it is an intra-ecosystem flow and not accounted as final ES (although it may contribute to many other human activities in many differnt ways). Fig. 15 shows soil retention in physical terms by all ecosystem types: (a) soil retention in physical terms is then only mapped for cropland (b) and then translated in monetary terms (c).

Figure 15.  

On site soil retention in Europe (2012): intra-ES flow (a) and final ES flow (a, b).

Eleventh lesson learned: While some of the ES flows can be considered as final, others remain within the ecosystem and are not counted as final nor allocated to economic units.

The general way to account for ES in INCA (compliant with SEEA EA) is: first to assess the service in physical terms, then to translate the service in monetary terms and, finally, to fill in the supply and use table. For some ES, raw data used as proxy of the service are already available and can be used “as is”, a couple of examples are timber provision (where INCA uses timber accounts) and carbon sequestration (where INCA uses Land Use Land Use Changes and Forestry (LULUCF) dataset). These examples do not require any modelling in physical or monetary terms, because the quantity (already available datasets) is eventually multiplied by a price, stumpage price for timber and carbon rates (provided by OECD) for carbon sequestration.

This is what can be defined as a “fast-track” approach. For ES accounts based on the fast-track approach (only), the methodology proposed in SEEA EA to calculate enhancement and degradation within an ecosystem asset applies perfectly (ref. section 10.3 in UN et al. 2021).

For all the other ES in INCA, spatial modelling techniques are applied. Some services required modelling the ecosystem contribution that is eventually applied to already existing datasets, like crop provision and crop pollination. Other services are required to entirely model raw spatial data because no data exist or can be collected from current statistics.

There is a degree of complexity also in the typology of biophysical models used. Less complex models are those that map and combine ecological features without a spatial dependency of one area from the other, such as on-site soil retention, habitat and species maintenance and nature-based recreation. For other ES, conversely, modelling requires to explicitly consider the spatial configuration and interdependencies of the system at a fine granularity. Pest control and pollination services, for example, depend on the presence, structure and spatial arrangement of landscape features interspersed in the agrarian landscape (Zulian et al. 2013, Rega et al. 2018). In other cases, for example, water purification and flood control, spatial flows from other locations need to be considered.

Models can be applied not only to assess ES in physical terms, but also to value ES in monetary terms. In fact, complex valuation techniques are applied to estimate nature-based recreation, water purification and flood control and habitat and species maintenance.

Table 1 summarises the different degrees of complexity in assessing and valuing the nine ES that were part of INCA and that are able to cover a good variety of the possible cases that may occur.

Twelfth lesson learned: There can be many combinations of simple and complex techniques in biophysical and monetary assessment for each ES. In fact, the assessment of some ES can be completed throughout a fast-track approach, while others might need more demanding modelling procedures.