The European Union (EU) 7th Environment Action Programme (Commission 2014) and the EU Biodiversity Strategy of 2030 (Commission 2020) included objectives to develop Natural Capital Accounting (NCA) in the EU, with a focus on ecosystems and their services. The Knowledge Innovation Project on an Integrated System of Natural Capital and Ecosystem Services Accounting (KIP-INCA) carried out in the years 2016 – 2020 produced pilot ecosystem accounts for the EU that are largely based on models available at the time the project was conducted and using public datasets from the Statistical Office of the European Commision (EUROSTAT), explicit spatial data and Earth Observation (EO) products (Commission et al. 2018, Commission et al. 2019, Commission et al. 2021). Complementary to this KIP-INCA initiative, the European Commission (EC) supported the development of ecosystem accounting in Member States. Furthermore in 2021, the UN Statistical Commission (UNSC) adopted the System of Environmental-Economic Accounts – Ecosystem Accounts (SEEA EA) as an official standard (UNSD 2021). The SEEA EA constitutes an integrated and comprehensive global statistical framework for organising data about habitats and landscapes, measuring the ecosystem services, tracking changes in ecosystem assets and linking this information to economic and other human activity.

Currently, the EC developed an amendment to Regulation 691/2011 on European environmental economic accounts to include ecosystem accounts compliant with the SEEA EA (European Commission 2022). Therefore, in 2021 Eurostat awarded a grant to revise the methodolgoies of the KIP-INCA service accounting models (Vallecillo et al. 2019) and results to support regulary ecosystem accounting (European Commission - Eurostat 2020). The objectives of this project are the increased harmonisation of accounting methods within the EU, providing a tool - called INCA-tool - to produce ecosystem accounts at national scale, extending the time-series of EU wall-to-wall KIP-INCA accounts and facilitating NCA results in regular reporting (Commission et al. 2021). To assist further development and integration into existing tools, the INCA-tool needs to be open-sourced and follow the FAIR (Findability, Accessibility, Interoperability and Reusability) principles for produced data as described by Wilkinson et al. (2019), as well as for the research software itself as described by Lamprecht et al. (2020). The FAIR principles define a technical standard and, therefore, do not provide any quality control of the software or data itself. The in-depth discussion of the FAIR principles is described in a later section of this article. Nevertheless, since these principles are not a binary concept (Lamprecht et al. 2020), they define the scope of FAIRness of the tool and its output.

The main objective of this article is to introduce the INCA-tool as a reference tool for ecosystem accounting in the EU following the amendment to Regulation 691/2011. This includes the basic concept of the tool, the nine currently integrated and harmonised ecosystem services, as well as its usability. Moreover, we evaluate the FAIRness of the INCA-tool following the 15 principles as described in Lamprecht et al. (2020), as well as of the tool output following the 15 principles as decribed in Wilkinson et al. (2019). In order to showcase the modular build-up of the INCA-tool, one ecosystem service - soil retention - is presented.

To facilitate Member States in the implementation of Regulation 691/2011 (European Commission 2022) and assist EUROSTAT in the validation of the produced national ecosystem accounts, a reference tool was requested to generate SEEA EA compliant European accounts (European Commission - Eurostat 2020). Therefore, the software package must fulfil different tools, as well as user requirements.

A user requirements analysis was performed to identify all possible stakeholders (e.g. EU Institutions and Member States) and to determine their requirements and considerations through interviews. We identified three main users for the INCA-tools: EUROSTAT, Joint Research Centre (JRC) and EU Member States. Furthermore, the analysis showed a large variation of experience in the usage of NCA within the EU Member States. Some EU Member States have little to no experience in integrating NCA in their reporting, where others have expressed no needs for additional tools. Therefore, the INCA-tool has to support the following needs: (1) consultation and use of the results at national level, (2) integration of national data sources in existing KIP-INCA accounting models, (3) using the models as a starting point to develop methods more tailored to regional characteristics. The analysis unveiled that, to support the three main users, the INCA-tool needs to support three different levels of expertise (see Table 1).

The tool requirements are based on an analysis of existing modelling platforms and the requirements from the European Commission. Table 2 highlights the main features of existing ecosystem accounting tools, their potential use in SEEA EA and their compliance with EU Regulation 691/2011. This analysis shows that the ESTIMAP (Zulian et al. 2014) tool, developed by the JRC, was a potential candidate, but is tightly linked to the GRASS database concept (Neteler et al. 2012) and, hence, limited the adaptation, based on user requirements. Currently, it only supports four out of nine mandatory service accounts by the EU Regulation and does not include monetary valuation. Additional requirements were gathered through analysing technical descriptions of IT environments, user requirements gathered during other processes, such as the System of Environmental-Economic Accounts – Experimental Ecosystem Accounts (SEEA-EEA) revision process (UN et al. 2014), an independent expert review report of KIP-INCA ecosystem service accounts and the Mapping and Assessment for Integrated ecosystems (MAIA) project (Hein 2019) complements this information.

Therefore, we decided to build on the existing knowledge of the KIP-INCA models and decided to redesign the tool according to the following tool requirements:

• harmonisation of the existing models (Commission et al. 2018, Commission et al. 2019, Commission et al. 2021) in both processing and data formats;
• validation and enhancement of the quality of the existing models and introduction of compliance with the SEEA EA EU guidelines;
• open-sourced and following the FAIR (Findability, Accessibility, Interoperability and Reusability) principle for research software (Lamprecht et al. 2020);
• output of the tool (maps and tables) should be compliant with the FAIR principle for data (Wilkinson et al. 2019).

Taking into account the user requirements (Table 1) and tool requirements, the INCA tool is designed in a modular way to allow automated processing, as well as human induced processing through a user-friendly graphical interface. The modularity of the INCA-tool is given by two main components:

• a core library, which can run independently via a command line interface and contains all processing routines for the different accounting modules and
• a front-end library, currently a QGIS (QGIS Development Team 2022) plug-in, to enable users to easily set up processing runs and inspect the results.

Fig. 1 shows both components, implemented in the open source repository (GIT) and licensed under European Union Public Licence (EUPL).

Figure 1.  

Flow chart diagram showing the architecture of the INCA-tool.

The command line interface provided by the core library, allows advanced users to set up batch runs or scripted sensitivity analyses. The core library supports integration into larger processing frameworks, either using the command line tool or through python Application Programming Interface (API). In this sense, the QGIS plug-in interface is just one example of such an integration.

The core INCA-tool library was set up with extensibility and flexibility in mind (see Fig. 2). The library was developed, based on the KIP-INCA prototypes (Vallecillo et al. 2019) for five of the current implemented services, amongst which some of them are derived from ESTIMAP (Zulian et al. 2014). All models were harmonised and validated against references in order to ensure the functional correctness of the software. The harmonisation also included, next to the map generation, the statistical tabular outputs and interoperability with other software. We want to stress that, therefore, the FAIR principles were not only applied on the software itself, but also on its output data. The different ecosystem services are implemented as independent modules, which are then included in a processing framework that deals with basic tasks, such as command line parsing, configuration files, input data and log files. The package also contains a few shared general utility modules containing routines for common processing steps related to Graphical Information Systems (GIS) data or other frequently used input data formats and exporting tables in a standardised way. This modular structure makes it convenient to include other ecosystem services in the future.

Figure 2.  

Flow chart diagram showing the modular setup of the INCA core library.

The current implemented front-end (QGIS plug-in) takes away the complexity of having to know how to operate a programming language for the end user, by restricting its input to specifiying the necessary parameters and input data to run a calculation procedure. These can be selected via convenient drop-down menus. After completion of the automatic calculations, the users can inspect the tabular data and maps in the QGIS desktop. The integration of the plug-in in the official QGIS repository is planned.

The KIP-INCA accounts have been developed over the past years through the implementation of spatial models and integrating a variety of geospatial and other data not originally designed for statistical purposes (Vallecillo et al. 2019). Even though the results were published in several reports (Commission et al. 2018, Commission et al. 2019, Commission et al. 2021), researchers or statisticians were not able to replicate these results and were obliged to search through these reports to extract important information (e.g. account results or model parameters). This fact limits the usage of the KIP-INCA methodologies to reproduce regular ecosystem accounts.

To facilitate harmonisation of the services and usability of the tool, we adopted the FAIR (Findability, Accessibility, Interoperability and Reusability) principles for both the research software and the output data. The FAIRness of the software and the data is not limited to the fulfilment of all fifteen principles (Wilkinson et al. 2019, Lamprecht et al. 2020).

FAIRness assessment of the INCA-tool

The FAIRness of the INCA-tool was assessed by following the guideline by Lamprecht et al. (2020) as shown in Table 3. The assessment shows that the current version of the INCA-tool fully fulfils eight out of fifteen principles. It is planned to complement the five partially fulfilled and two unfulfilled principles by assigning a Digital Object Identifier for the software and each version, further align the metadata vocabulary to ARIES (Balbi et al. 2022) and advocate it as community vocabulary (Balbi and Bagstad 2021) and register the software in a searchable software registry.

Application of the FAIR principles for output data

Since the output from the INCA-tool is considered as a product, it is of equal importance to assess the application of the FAIR principles for this output data. The level of FAIRness of the output data was assessed by following the guidelines by Wilkinson et al. (2019) and is shown in Table 4. The assessment reveals that the current version of the INCA-tool provides output data that fully fulfils six out of fifteen principles. It is planned to complement the five partially fulfilled and three unfulfilled principles by aligning the metadata with a standardised vocabulary (from SEEA EA and GEO EO4EA) as soon as it has been provided. Furthermore, it is foreseen that the protocols, as implemented on the website for the European continental accounts, are further improved to facilitate uptake.

Currently, we see that 'Interoperability' is probably the most difficult principle to achieve and requires data compatibility, metadata compatibility and common APIs. The INCA-tool produces cloud-optimised geotiff (COG) raster images which are commonly recognised as an interoperable format that is supported by many platforms (Anonymous 2022). Tabular data are written in machine-accessible formats (like Comma-Separated Values - CSV format) where standard API can be defined (e.g. OpenSearch) to retrieve this information. Another important aspect of interoperability in the accounts is to harmonise the reporting across all EU Member States according to the SEEA EA standard. The standard, however, leaves room for interpretation and, therefore, the European Commission (EUROSTAT) has established a taskforce to have joint discussions for this harmonisation through developing EU guidelines to implement the SEEA EA standard. These EU guidelines are applied in the updates of the INCA-tool, making the tool more interoperable.

The assessment for the FAIRness of the software and the output data showed that our current implementation of the INCA-tool provides a certain degree of FAIRness. Nevertheless, the FAIR principles only describe the technical standard and do not provide any information on the functional correctness of the software itself. To ensure this correctness, we implemented a detailed evaluation and validation scheme for the models (unit tests), as well as the output data (cross-checks between table and map data, plausibility checks).

To demonstrate the integration of modular ecosystem services into the INCA-tool, the soil retention model was chosen. Soil retention, also known as sediment retention, requires a biophysical model and is an ecosystem service frequently included in ecosystem accounting. The service accounts for the value of the ecosystem to minimise soil erosion and, hence, contribute to the maintenance of soil quality and, therefore, of ecological processes. The Revised Universal Soil Loss Equation (RUSLE) model (Renard et al. 1991), as implemented in the KIP-INCA (Commission et al. 2021) is used for the biophysical calculations. RUSLE requires several spatial data inputs, such as the digital elevation model (DEM), land use - land cover, soil information, rainfall erosivity data and several coefficients. The INCA on-site soil retention account calculates the amount of soil retained by the ecosystem (use) as an interaction between the potential of ecosystems to reduce soil erosion by rain (the ecological side) and the demand (or need) for soil retention by ecosystems (the socio-economic side). The amount of soil erosion taking place at a higher rate than the soil formation rate (net losses) are provided in a complementary mismatch dataset. The monetary value is only calculated for cropland and expressed in 'real' values and 'nominal' values deflated to the reference year 2000. Cropland is considered a socio-economic flow contribution to the agricultural sector, while other ecosystem types are considered intra-ecosystem flows and, hence, not valuated. The soil retention module can be used to generate ecosystem accounts at different reporting levels (EU level, national level, regional level).

Fig. 3 depicts the workflow diagram for the soil retention model which consists of four specific modules, following the INCA architecture structure: the potential, the demand, the biophysical flow and the monetary flow for cropland. The fifth module is a generic module to calculate statistics and generate the tabular output. The INCA core soil retention model provides a python3 compliant API to configure the 13 input datasets and five configuration settings, next to some generic features as logging file, start run etc. Each module is further broken down and programmed into separate python sub-modules to ease integration and reuse with other accounts (e.g. mapping of Ecosystem Types to land-cover map).

Figure 3.  

Flow chart diagram showing the modular setup of the soil retention ecosystem service as implemented in the INCA-tool.

The 13 input datasets and five configuration files necessary to generate the soil retention account are represented in the QGIS graphical interface. A set of default input datasets at EU level were prepared for the accounting years 2000, 2006, 2012 and 2018 and can be used by users to reproduce the results. Nevertheless, each of these datasets can be replaced by MS using the INCA tool interface to create their optimised national accounts. Fig. 4 shows the graphical interface of the QGIS tool for the Soil Retention account for 2018 in Austria. The left window depicts the selection of the input datasets. The bottom left window shows the actual execution (run) button. The geospatial maps for the ecosystem flow are automatically ingested into the QGIS project as shown in the right window. Users can further add other geospatial data to their QGIS accounting project for analysis.

Figure 4.  

Screenshot of the QGIS front-end of the INCA-tool for the soil retention ecosystem service account - example shows the Graphical User Interface and ecosystem flow result for Austria in 2018.

The statistical reporting module is a generic module that not only calculates zonal statistics (provided as CSV files), but also automatically formats the tabular output in Supply-Use ecosystem accounting tables (provided as EXCEL files). Fig. 5 shows an example of this tabular output for the soil retention of 2018 over Europe (Austria is indicated with AT).

Figure 5.  

Supply and Use tables (SUT) for the soil retention ecosystem service for the year 2018. Note: the unit is in 1000 metric tonnes carbon.

Currently, nine ecosystem services, based on the KIP-INCA methodology (Commission et al. 2021), are integrated in a toolkit, based on the python programming language and the open QGIS tool. This new toolkit, named INCA-tool, was used to update the EU wall-2-wall KIP-INCA service accounts - published by JRC on the INCA website (Joint Research Centre 2021) - and is ready to generate yearly accounts. The tool is currently also in beta test by several European Member States to generate national accounts and is planned to be publicly released by the end of year 2022 as free and open source. The INCA-tool provides baseline methods for ecosystem accounts, but Member States are not limited to these methods. Nevertheless, the implemented methods will be the baseline for EU validation. The tool will be further extended in the coming years with more service accounts to support the amendment to Regulation 691/2011 on European environmental economic accounts. This tool facilitates the generation of SEEA-EA accounts compliant with the EU guidelines on ecosystem accounting for European Statistical Offices.

In a step forward for open science, we decided to implement the FAIR principles for software, as well as output data in the INCA-tool. The FAIRness concept is a relatively new topic to the ecosystem accounting community, well received, but requires more standardisation and integration. Despite the fact that we achieved a high level of FAIRness and plan to further raise this level, currently, the INCA-tool cannot be fully compliant with all principles until the community has decided on a standard vocabulary and registry.

Nevertheless, due to the modular design of the INCA-tool, it can be integrated into other tools, if the platforms support and can bind with python3. For that, we plan to further improve the tool to achieve semantic interoperability aligning with the SEEA interoperability strategy (Balbi and Bagstad 2021). This prepares for seamless ingestion in modelling approaches centred on interoperability like the Artificial Intelligence for Environment & Sustainability (ARIES) platform (Villa et al. 2014), while still being available for more commodity practices, such as using Geographic Information System tools (QGIS or ArcGIS).

The new INCA-toolkit is the next step in harmonising ecosystem accounting within the European Union. By Regulation 691/2011, the INCA-tool will be the reference for ecosystem accounting in Europe. Thanks to its modular design, its appliance of FAIR principles and its free and open-source licence, expert users in the community have the ability to improve existing services or add new services to the toolkit.

We would like to thank the JRC team of the KIP-INCA project for their support during the refactoring of the accounts. Special thanks to Joachim Maes and Ioanna Grammatikopoulou for organising the knowledge transfers and to Alessandra La Notte, Sara Vallecilo and Mayra Zurbaran and Chiara Polce for their technical support. We also would like to thank the EU Member States that provided feedback on the first beta version. The content of this publication is based on the work done within the INCA project for Eurostat; however, it does not reflect the official opinion of the European Union.

INCA 2021-2023

Service contract 2020.031

Conceptualisation (B.S.; S.K.; M.B.), Methodology (M.B., T.D., M.vL.), Software (T.D., M.B., M.vL., W.P.), Formal analysis (M.B., T.D., M.vL.), Investigation (M.B., T.D., M.vL., B.S., S.K.), Data Curation (M.B., T.D., M.vL.), Writing – original draft preparation (M.B., B.S., S.K., T.D., M.vL.), Writing – review and editing (M.B., B.S., S.K., W.P., M.vL), Visualisation (M.B., T.D., W.P.), Supervision (B.S., S.K.), Project administration (B.S., S.K.), Funding acquisition (B.S., S.K.)

No conflict of interests.