{"paper_id":"26c153d3-ecc3-4154-a805-e5e28b533d1a","body_text":"Project Report\nAuthor-formatted document posted on 06/06/2024\nPublished in a RIO article collection by decision of the collection editors.\nDOI: https://doi.org/10.3897/arphapreprints.e128926\nFuture projections of modelled soil EBVs for ecosystem\nrestoration and climate scenarios\n Irene Calderon Sanou, Carlos Guerra, Graciela Rusch, Sergei Põlme, Joachim Maes, \nNéstor Fernández\n\n \n \n \n \n \nFuture projections of modelled soil EBVs for ecosystem restoration and climate scenarios \n \n31/10/2023 \n \nLead beneficiary: \nMartin Luther University \n \nRef. Ares(2023)7447830 - 02/11/2023\nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 2 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nThis project receives funding from the European Commission’s Horizon 2020 research and innovation \nprogramme, under Grant Agreement n.101003553 \nPrepared under contract from the European Commission \nGrant agreement No. 101003553 \nEU Horizon 2020 Coordination and Support Action \n \nProject acronym: EuropaBON \nProject full title: EUROPA BIODIVERSITY OBSERVATION NETWORK: INTEGRATING DATA \nSTREAMS TO SUPPORT POLICY \nStart of the project: 01.12.2020 \nDuration: 36 months \nProject coordinator: Prof. Henrique Pereira \n Martin-Luther Universität Halle-Wittenberg (MLU) \n www.europabon.org \n \nType: Coordination and Support Action \nCall: The Sc5-33-2020 Call: “Monitoring ecosystems through innovation and \ntechnology” \n \n \n \n \n \n \nThe content of this deliverable does not necessarily reflect the official opinions of the European \nCommission or other institutions of the European Union. \n \n \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 3 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nProject ref. \nno. \n101003553 \nProject title EUROPA BIODIVERSITY OBSERVATION NETWORK: INTEGRATING DATA STREAMS TO \nSUPPORT POLICY \n \n \n \nDeliverable title Future projections of modelled soil EBVs for \necosystem restoration and climate scenarios \nDeliverable number D5.4 \nContractual date of delivery 31.10.2023 \nActual date of delivery 02.11.2023 \nType of deliverable Report \nDissemination level Public \nWork package number WP5 \nInstitution leading work package Martin Luther University \nTask number T5.4 \nInstitution leading task Martin Luther University \nAuthor(s) Irene Calderon-Sanou, Carlos Guerra, Graciela \nRusch, Sergei Põlme, Joachim Maes, Nestor \nFernandez \nEC project officer Laura Palomo-Rios \n \n \n \nDeliverable description This Deliverable describes the modular \nprocedure to estimate four essential \nbiodiversity variables important to monitor \nsoils and support their ecological restoration. \nThese four essential biodiversity variables \ninclude taxonomic diversity, functional \ndiversity, soil basal respiration and microbial \nbiomass. At the end, we also propose a \nsynthesis indicator that can be used to support \nthe identification of the ecological status of \nEuropean soils. \nKeywords soil restoration, taxonomic diversity of soils, \nfunctional diversity, soil basal respiration, \nmicrobial biomass, soil ecological status \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 4 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \n \n \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 5 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nContents \n \nExecutive summary 6 \n1. Introduction 7 \n2. Showcase participatory design 11 \n3. Policy targets 12 \n4. Essential Biodiversity Variables design 14 \n5. Discussion 30 \n6. References 30 \n \n \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 6 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nExecutive summary \nThe Deliverable discusses the critical role of soil in supporting terrestrial ecosystems, agriculture, and \nglobal climate regulation. It highlights that a significant portion of European soils are currently \nunhealthy, which has far -reaching consequences, including risks to human health, the environment, \nand the economy. Soil degradation affects food security, water quality, climate change, and \nbiodiversity. It also emphasizes the importance of soil monitoring in Europe to ensure sustainable land \nmanagement, preserve biodiversity, and mitigate environmental challenges. Healthy soils are essential \nfor agriculture, food security, and climate resilience. The European Union (EU) recognizes the \nimportance of soil health in addressing the food needs of a growing global population increasing the \nneed for improved soil monitoring. It also discusses the role of advanced remote sensing technologies \nin soil monitoring and mentions initiatives and projects focused on soil biodiversity and ecosystem \nfunctioning. \nThis Deliverable highlights the showcase in soil restoration and climate change mitigation that is aimed \nat developing soil essential biodiversity variables to test the capacity of current datasets and \nmonitoring frameworks to provide relevant information on the distribution and future trends of key \nsoil biodiversity variables. The development process also stakeholder engagement in the context of \nselecting essential biodiversity variables (EBVs). Two main approaches were used: a working group \ninvolving researchers and targeted meetings with institutional stakeholders. Key meetings were held \nwith stakeholders from organizations such as the European Environmental Agency, the Joint Research \nCenter, and German soil and biodiversity monitoring agencies. These mee tings aimed to gather input \nand insights on the selection of EBVs. The stakeholder engagement process revealed three key points: \ni) the importance of using readily available and open data to ensure data continuity and promote \ntransparency and accessibility (his includes data from sources like the European Soil Data Centre \n(ESDAC) and GBIF, as well as point data from the LUCAS sampling framework); ii) the need for diversity \nin modeling approaches, incorporating multiple data sources and levels of expertise ( the goal is to \nshowcase the potential use of existing information and allow various stakeholders to identify their \nroles in producing soil -based EBVs for different functions or ecosystem services); and iii) the \nimportance of including climate and land use prediction data in modeling frameworks when possible. \nOverall, the stakeholder engagement process emphasized the significance of data accessibility, diverse \nmodeling approaches, and the incorporation of relevant future data in the selection and development \nof essential biodiversity variables. \nFurthermore, the Deliverable discusses the existing gap in the EU's legal framework concerning soil \nmanagement and highlights the proposed Soil Monitoring Law as a comprehensive framework to \naddress this gap. The outcomes of this Deliverable align with sev eral EU environmental policy \ninitiatives, such as the EU Biodiversity Strategy for 2030, the Zero Pollution Action Plan, the Circular \nEconomy Action Plan, and the Chemicals Strategy for Sustainability. Finally, it proceeds with the \nselection of essential b iodiversity variables (EBVs) for soil -related monitoring within the EuropaBON \ninitiative. Three primary soil-related EBVs have been chosen: 'Functional Composition of soil biota' and \n'Community Biomass of soil microbes' and ‘Taxonomic diversity’. These EBVs belong to the 'Community \nComposition' class within the 'Terrestrial' realm category. \n \n \n \n \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 7 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \n1. Introduction \n \n1.1. State of the art in soil biodiversity and ecosystem function monitoring in Europe \n \nSoil plays a crucial role in supporting terrestrial ecosystems, agriculture, and even global climate \nregulation (Bardgett and van der Putten 2014) . Soils provide a habitat for countless organisms and \nserve as the primary medium for plant growth (Beck et al. 2005). Healthy soils support food production, \nclimate resilience, and overall well-being (H. R. P. Phillips et al. 2019; de Graaff et al. 2015) . However, \nin the European Union (EU), scientific evidence suggests that 60 -70% of soils are currently unhealthy, \nleading to degradation issues across member states. Soil degradation has far-reaching consequences, \nincluding risks to human health, the envir onment, climate, economy, and society. These risks \nencompass food security, water quality, flooding, drought impacts, biomass production, carbon \nemissions, and biodiversity loss (H. Phillips et al. 2023) . In Europe, where diverse landscapes and \necosystems abound, soil monitoring is of paramount importance to ensure sustainable land \nmanagement, preserve biodiversity, and mitigate environmental challenges such as climate change \nand land degradation. This is particularly relevant for the development of four key axes of current \nEuropean policy, namely assessing the outcomes of sustainable agriculture , ensuring that climate \nmitigation measures are successful, informing land-use planning and regulation , and conserving and \nrestoring biodiversity. As an example, agriculture is a cornerstone of European economies, and soil \nhealth directly influences crop productivity and food security (Foley et al. 2011) with multiple large \nfood distribution companies investing in their own soil monitoring systems to control and monitor the \ncapacity of producers to meet their market goals. Healthy soils are strategically crucial for securing \naccess to sufficient, nutritio us, and affordable food. As the EU plays a significant role in global food \nmarkets, it recognizes the importance of soils in good ecological condition in addressing the food needs \nof a growing global population, expected to reach 9 -10 billion by 2050. Since 95% of food is linked to \nsoil, soil degradation directly impacts food security and cross -border food markets. Globally, the \npressure on soil and land is increasing, with 4.2% of EU territory already affected by land take and soil \nsealing, primarily at th e expense of agricultural land. Additionally, soil degradation poses a threat to \nthe long-term fertility of agricultural soils, with erosion, organic carbon loss, nutrient exceedances, \ncompaction, and secondary salinisation affecting 61-73% of EU agricultural soils. Without sustainable \nsoil management and regeneration efforts, deteriorating soil health could become a central factor in \nfuture food security crises. \nHealthy soils are crucial for agriculture and the overall agro -ecosystem (Alberto Orgiazzi et al. 2022) . \nThey contribute to stable or increased crop yields, biomass production for non-food sectors, and long-\nterm production security for farmers. Healthy soils also play a vital role in the transition to a sustainable \nbioeconomy, preserving land value, and red ucing operational costs for farmers. Financial support for \nsoil-enhancing practices is available under the Common Agricultural Policy (CAP) (Pe’er et al. 2019) and \nproposed EU carbon removal certification framework and the recent proposal of a European Soil \nMonitoring Law. Soil degradation has adverse effects on human health, including respiratory and \ncardiovascular diseases due to airborne particulate matter from wind erosion. Sealed soils exacerbate \nheatwaves and reduce pollutant absorption capacity. Contaminated soils can compromise food safety. \nHealthy soils support environmental and recreational value, benefiting both rural and urban areas by \nsupporting the establishment of green spaces, improving air quality, and reducing heat islands. \nImproving soil health enhances the EU's resilience to climate change by increasing organic matter, \nwater retention, filtering capacity, and erosion resistance. Carbon farming practices help mitigate \nclimate change by storing CO2 in the soil (Freibauer et al. 2004). Enhanced soil water retention reduces \nflood intensity and mitigates drought effects. Certain soil bacteria and symbiotic fungi assist crop \nplants in tolerating drought. As climate-related hazards, including wildfires, intensify in Europe, healthy \nsoils with water retention capacity support resilient forest ecosystems. However, wildfires can lead to \nsoil degradation and increased erosion, landslides, and floods. Strengthening soil knowledge aids \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 8 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \ndisaster risk assessments, recognizing the diverse roles soils play in disaster mitigation. While EU and \nnational policies have made positive strides in improving soil health, significant gaps remain in \naddressing soil degradation drivers. Soils form slowly, but their health can be maintained or improved \nwith the right measures and practices. \nThe past decade has witnessed significant progress in soil monitoring sensors, technologies, and \napplications allowing for more precise and efficient data collection. In Europe, the adoption of \nadvanced remote sensing techniques has revolutionized soil mon itoring. These technologies enable \nresearchers to gather real-time data on soil moisture (Ruosteenoja et al. 2018; Almendra-Martín et al. \n2022), temperature (e.g., (Haesen et al. 2021) ; https://www.soiltempproject.com), pH (Ballabio, \nPanagos, and Monatanarella 2016), nutrient content, and microbial activity (Smith et al. 2021) across \nvarious spatial scales. A prime example is the deployment of satellite -based remote sensing, which \nprovides comprehensive coverage of soil properties at regional and continental scales. The European \nSpace Agency's (ESA) Soil Moisture and Ocean Sali nity (SMOS) mission and the European Space \nAgency's Copernicus program have played pivotal roles in this regard. These platforms allow for \nmonitoring soil moisture dynamics, aiding in drought prediction, flood risk assessment, and crop yield \noptimization. \nProjects such as the Soil Biodiversity and Ecosystem Functioning (SFB) and the European Soil \nBiodiversity Observation Network (EU-SOILS) aim to assess the impact of land use changes and climate \non soil organisms. These initiatives play a crucial role in gu iding sustainable land management \npractices. Together with these and more recent EU projects (e.g., work is still underway in projects like \nBENCHMARKS or SoilGUARD), over the last two decades the European Union has developed, either \nnationally or at Europe an level, several monitoring systems to track the condition of European soils. \nWhile originally developed to monitor changes in land use across Europe, the LUCAS (Land Use and \nCover Area Frame Survey) program includes since 2009 a module on soils. This mod ule initially \ndeveloped to track the amount of carbon sequestration in European soils led to the establishment of \nwhat can become a long -term soil biodiversity monitoring network in the European territory (A. \nOrgiazzi et al. 2018). \nThe European Green Deal aims to transform the EU into a fair, prosperous, and sustainable society. \nPart of this initiative includes the adoption of various strategies and plans, such as the EU Biodiversity \nStrategy for 2030, a Zero Pollution Action Plan, a n EU Climate Adaptation Strategy, and an EU Soil \nStrategy for 2030. The EU Biodiversity Strategy for 2030 emphasizes the need to protect soil fertility, \nreduce erosion, and increase soil organic matter through sustainable soil management practices. It \nalso addresses contaminated sites and aims to update the 2006 Soil Thematic Strategy to combat soil \ndegradation. The EU Soil Strategy for 2030 envisions achieving healthy soils by 2050 through voluntary \nand legislative actions. It proposes a Soil Monitoring La w supported by an impact assessment, \naddressing soil health indicators, monitoring provisions, and sustainable soil use. The 8th Environment \nAction Programme sets the objective of living well within planetary boundaries, highlighting the \nimportance of addressing soil degradation and ensuring soil protection through legislative proposals. \nInstitutional stakeholders, including the European Parliament, the Council of the EU, and others, have \ncalled for the development of an EU legal framework for soil. Global commitments related to soil, such \nas those under the Rio Conventions and UN Sustainable Development Goals, further emphasize the \nimportance of soil health. Currently, there is a lack of comprehensive and harmonized soil health data \nacross the EU. This legi slative proposal ( https://eur-lex.europa.eu/legal-\ncontent/EN/TXT/?uri=CELEX:52023PC0416) seeks to establish a coherent soil monitoring framework \nbased on common definitions of healthy soil to support sustainable soil management and stimulate \nresearch and innovation. Multidimensional soil health data can improve drought and disaster \nmanagement, climate change mitigation, and adaptation efforts, as well as support human health and \nbiodiversity. Sustainable soil management practices are expected to yield economic and \nenvironmental benefits. The proposal also addresses soil contamination, aiming to reduce risks to \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 9 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nhuman health and the environment by adopting a risk -based approach and promoting the polluter -\npays principle. The legislation allows Member States a gradual and proportionate approach to \nimplement soil governance, monitoring systems, soil health assessment , and sustainable soil \nmanagement measures. \nEffective management and conservation of biodiversity requires standardized and measurable \nindicators to assess its status and trends. To this respect, Essential Biodiversity Variables (EBVs) have \nemerged as a fundamental framework to address this need. As biodiversity faces unprecedented \nthreats due to habitat loss, climate change, and invasive species, EBVs represent a pivotal approach to \naddress these challenges by providing a structured framework for biodiversity monitoring and \nreporting. EBVs are quant ifiable, essential components of biodiversity that encapsulate key \ndimensions of the natural world. They serve as fundamental metrics to gauge the state and trends of \nbiodiversity across spatial and temporal scales (Pereira et al. 2013) . The conceptualization of EBVs \nharmonizes biodiversity monitoring and facilitates data comparability, enabling robust assessments of \necosystem health and informing conservation strategies (Proença et al. 2017) . Recently, soil specific \nessential biodiversity variables were defined (Guerra et al. 2021) (Figure 1) providing an holistic \noverview of how soil ecosystems should be monitored and studied, where soil organisms are \nintertwined with relevant soil chemical, physical, and functional properties, contributing to overall \nsocietal well-being. These soil EBVs include information on soil ecological functional aspects (e.g., soil \nmicrobial biomass and respiration), taxonomic and functional diversity, population abundance, and \nhabitat traits. \n \n \nFigure 1 Links between global soil essential biodiversity variables (EBVs) (outer ring) that are prioritized \nby the Soil Biodiversity Observation Network (SoilBON) and policy sectors (center) through the use of \nsoil ecological indicators (inner ring) (Guerra et al. 2021). \nIn this context, the Statistical Office of the European Union, known as EUROSTAT, conducts a regular \nsurvey called the Land Use and Coverage Area Frame Survey (LUCAS (Eiden, Vidal, and Georgieva 2002; \nBorrelli et al. 2022)) to assess land use, land cover, and changes in the European Union (EU) based on \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 10 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nEU Parliament and Council decisions. This survey, initiated in 2008, occurs every three years and \ninvolves over a million georeferenced sampling locations covering the EU territory. Surveyors visit \naround 270,000 of these points in the field to verify clas sifications and collect additional data. LUCAS \nincludes soil assessment as of 2009, initially in 25 EU member states. About 20,000 topsoil samples \nwere collected, and various physical and chemical properties were analyzed. The goal was to create a \nstandardized dataset of topsoil properties for the entire EU. The survey expanded to include Bulgaria \nand Romania in 2012 and locations at altitudes above 1000 meters in 2015. Electrical conductivity \nmeasurements were added to assess soil salinity issues. The third LUCAS Soil survey in 2018 (see Figure \n2) was already conducted with around 26,000 sampling locations, adding properties like bulk density, \nsoil biodiversity, visual soil erosion assessment, and organic horizon thickness measurement. LUCAS \nSoil is a subs tantial and freely accessible continental -scale soil database, including all land use types \nand benefiting various fields of research and informing policymaking. The ongoing scientific analyses \nof LUCAS Soil 2015 and 2018 samples and the current data prepa ration for the 2021 campaign offer \nopportunities for reviewing past data accessibility and discussing the future potential of LUCAS Soil as \na tool for analysing soil-related processes. \n \nFigure 2 Structure of LUCAS Soil survey over the sampling years (A. Orgiazzi et al. 2018). \nFollowing on this point, this monitoring approach requires a holistic approach that encompasses \nvarious European soil -related Essential Biodiversity Variables (EBVs) and employs standardized \nmonitoring systems at the European level to track the status and dynamics of soil biodiversity and \necosystem functioning over time. These EBVs should cover four key dimensions of soil systems, \nincluding soil physics, soil chemistry, soil biodiversity, and soil ecosystem functions, all of which are \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 11 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nlinked to specific ecological indicators (see Figure 1). The European soil research community has been \nmobilised to tackle these issues, and participate and lead several national and European efforts to \nmonitor soils and their biodiversity. However, these initiatives often rely on static, fragmented soil \nbiodiversity data without temporal resolution or coordination. While a European soil monitoring \nnetwork may not have the granularity to distinguish specific management practices, it can highlight \nsuccessful examples of conservation efforts focused on European soils. This can serve as a valuable \nreference point for comparisons among European regions and countries, aiding in the development of \nmore effective soil conservation policies. As mentioned before, eff ective soil monitoring is essential \nfor addressing various environmental challenges in Europe, including nature conservation, land \ndegradation, climate mitigation and adaptation, forestry, and food security. This monitoring initiative, \nalso as part of the current soil monitoring law, requires collaboration with local partners across diverse \necosystems and environmental conditions. It also entails capacity -building and knowledge -sharing \nmechanisms, as well as the creation of an open global repository of soil biodiversity resources. \nCurrently, soil biodiversity monitoring in Europe can be seen as a multi -tiered approach that involves \nEuropean (LUCAS Soils) coordinated sampling, cross -laboratory standardization, data aggregation, \nvalidation, and reporting, lead ing to a comparable mosaic of soil biodiversity and functional \nassessments across Europe. The goal is to place soils and their biodiversity at the forefront of European \nsustainability discussions. This effort is already underway in Europe, where several re search projects \ntogether with European institutions are providing essential biodiversity data to inform European \npolicies, including the European Biodiversity Strategy for 2030. While still not fully in place, a soil \nmonitoring program in Europe based on EBVs and derived indicators, can provide crucial tools to gauge \nprogress towards conservation and economic targets in the coming decades. It can also act as an early \nwarning system, signaling the success or failure of current policies in preserving and enha ncing soil \nquality and health across European landscapes. \n1.2. Showcase goals \n \nThis showcase aimed to develop soil essential biodiversity variables to test the capacity of current \ndatasets and monitoring frameworks to provide relevant information on the distribution and future \ntrends of key soil biodiversity variables. Here we are testing two data streams, one from a standardized \nsampling scheme, the LUCAS soil survey, and one from a combination of multiple expert and citizen \nscience data (e.g., using data from EdaphoBASE and GBIF). This showcase also uses several modeling \nframeworks to assess the distribution and condition of specific soil EBVs and, where possible, allow \nprediction of these variables for future conditions. \n \n \n \n2. Showcase participatory design \n \n2.1. Stakeholders’ engagement process \n \nStakeholder engagement is an important step to validate and specify the essential biodiversity \nvariables selected. In the case of this showcase, we developed a dual approach. On one hand we \ndeveloped a working group targeting researchers, and in parallel t argeted meetings with institutional \nstakeholders were held to have more direct input in the decision making process. Regarding the latter, \nwe had three key meetings with stakeholders at the European Environmental Agency, the Joint \nResearch Center, and from German soil and biodiversity monitoring agencies, including the Federal \nAgency for Nature Conservation (Bundesamt für Naturschutz), both at the begginign and at the end of \nthe process. This targeted consultation with stakeholders that have interests at di fferent scales \nallowed us to have a better grasp on the needs and limitations related to the selected EBVs. At the \nsame time, these stakeholders represent at the same time users (e.g., EEA) and data providers (e.g., in \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 12 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nthe case of the JRC and the BfN) with deep knowledge on the available information on soils and soil \nbiodiversity. Regarding the workshop with researchers, this was held in the beginning of the process \nto allow for an early discussion and integration of cri tical comments. This workshop included input \nfrom researchers with expertise in soil ecology, geomorfology, microbiology, soil macroecology, \nremote sensing, ecological modelling, and statistics. While we tried to engage a diverse group of \nstakeholders, at this stage we did not include policy makers nor were we able to involve expertise in \nsocial sciences that would be relevant to assert the capacity of the selected EBVs or indicators to \nsupport the assessment of soil ecosystem services. \n \n2.2. Key inputs from stakeholders \n \nFrom the meetings held, three points were raised: \ni) first, stakeholders stressed the need to use currently readily available and open data. This would not \nonly ensure the continuity of such dataset production (e.g., like in the case of the datasets curated by \nthe European Soil Data Centre (ESDAC) or the data streams related to GBIF), but also the promotion of \nopen access standards that facilitate transparency and data accessibility. From our consultation, these \ndatasets include point data from the LUCAS sampling framework, modelled data for specific soil \nproperties available at the ESDAC website, point data for soil species distribution available both in the \nEdaphoBASE and GBIF data portals, among others; \nii) the use of models should be diverse to include multiple data sources and levels of expertise. The \naim is to produce a showcase that highlights the potential use of current available information rather \nthan producing a single EBV. This will allow multiple stakeholders to identify their specific roles in the \nproduction of such soil based EBVs and clarify how different EBVs can be used to model different \nfunctions or ecosystem services; \niii) when possible, modelling frameworks should include and be prepared to ingest climate and land \nuse prediction data. When optimising models to describe current conditions, it is often the case that \nwe would have to use the best available datasets to represent those conditions. These datasets often \nare not the same as the ones used to estimate future conditions and/or create policy relevant \ninformation. The feedback was to use, when possible, a combination of both state of the art datasets \nto represent current conditions and equivalent datasets to represent future conditions; \nAll these point were fully integrated in the process of identifying and calculating the different EBVs \nthat were selected for this Showcase. \n \n \n3. Policy targets \n \n3.1. Reporting needs and data gaps \n \nThe European Union has been actively implementing environmental measures over the past three \ndecades with the goal of enhancing the quality of the environment for its citizens and ensuring a high \nquality of life. While EU law contains various provisions related to environmental protection, there is \na significant gap in the existing legal framework when it comes to soil. To address this gap, a proposal \non soil health has been introduced to establish a comprehensive and coherent EU-level framework for \nsoil management, currently addressed in the context of the Soil Monitoring Law. This proposal serves \nto complement and reinforce existing environmental legislation in several key areas. Firstly, it \naddresses soil contamination, including historical contamination, aligning with directives such as the \nIndustrial Emissions Directive, the Waste Framework and Landfill Directives, the Environmental \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 13 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nLiability Directive, and the Environmental Crime Directive. By doing so, it contributes significantly to \nthe protection of human health, which is a central objective of EU environmental policy. \nHealthy soils possess the natural ability to absorb, store, and filter water, and as a result, the proposal \nis expected to support the objectives of several other directives, including the Water Framework \nDirective, the Groundwater Directive, the Nitrates Directive, and the Environmental Quality Standards \nDirective. It addresses soil contamination and erosion while enhancing soil water retention. \nAdditionally, healthy soils play a crucial role in flood prevention, aligning with the objectives of the \nFloods Directive. The proposal contains provisions on sustainable soil management also complement \nexisting EU legislation related to nature, such as the Habitats and Birds Directives. By improving \nbiodiversity and preventing soil erosion, it supports the habitats of wild pollinators that nest in soils \nand contributes to reducing air pollution. Healthy soils serve as the foundation for life and biodiversity, \nincluding habitats, species, and genes. Furthermore, the knowledge, information, and data collected \nunder the monitoring requirements outlined in the proposal for a Soil Monitoring Law are expected to \nenhance the assessment of environmental impacts related to projects, plans, and programs conducted \nunder the Environmental Impact Assessment Directive and the Stra tegic Environmental Assessment \nDirective. In the framework of the Soil Monitoring Law, only soil respiration is taken as a viable EBV \n(although the proposal refers to it as an indicator) but this Deliverable aims to go beyond that to \nprovide a more comprehensive overview of the available ready to use information. \nIn specific terms, the proposal for a soil monitoring law aligns with several other EU environmental \npolicy initiatives, including: i) the EU Biodiversity Strategy for 2030, which aims to protect nature within \nthe EU and includes a proposal for a regulation on nature restoration in which soils play an important \nrole from an ecological point of view. The proposed regulation seeks to restore 20% of the EU's land \nand sea by 2030 and all ecosystems in need of restoration by 2050, with synergies between this \nproposal and the soil health proposal; ii) the Zero Pollution Action Plan, which envisions reducing air, \nwater, and soil pollution to non-harmful levels by 2050. This proposal aligns with efforts to revise and \nstrengthen existing EU legislation in the air and water sectors and legislation related to industrial \nactivities; iii) the Circular Economy Action Plan, which outlines measures to reduce microplastics and \nevaluates the Sewage Sludge Directive governing the quality of sludge used in agriculture; and the \nChemicals Strategy for Sustainability, which aims to protect citizens and the environment from \npotentially hazardous properties of chemicals while recognizing their importance to modern society. \n \n3.2. Cross-policy contributions \n \nDue to its encompassing nature, the European monitoring of soil biodiversity aligns seamlessly with \nseveral key EU policies and initiatives across various sectors, namely: \n \nEuropean Green Deal and EU Policy Objectives: soil biodiversity monitoring serves as a vital \ncomponent of the European Green Deal and contributes to EU policy objectives such as achieving \nclimate neutrality, promoting resilient nature and biodiversity, reducing pollution, establishing \nsustainable food systems, and enhancing human health and well-being. \n \nEuropean Climate Law: The objectives of implementing soil biodiversity monitoring across Europe are \nharmonious and synergistic with the European Climate Law. They support EU climate change \nadaptation goals by bolstering resilience and align with the ambition of achieving a cli mate-neutral \nEurope by 2050. Carbon sequestration in soils is a crucial aspect of this endeavour, balancing \ngreenhouse gas emissions even after a robust decarbonization process. \n \nLand Use, Land Use Change and Forestry (LULUCF) Regulation: soil biodiversity monitoring \ncomplements the revised LULUCF Regulation, which aims to reduce net emissions by 55% by 2030. The \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 14 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nLULUCF Regulation necessitates increased climate ambition in land use policies among EU Member \nStates. The monitoring of soil carbon stocks and nature-based climate mitigation in soils, as advocated \nby the proposal, aligns seamlessly with the objectives of the LULUCF Regulation. \n \nCertification Framework for Carbon Removal: soil biodiversity monitoring can support the proposed \ncertification framework for carbon removal aligns with the initiative to deploy high -quality carbon \nremovals through a voluntary EU certification framework. This framework is instrumental in enhancing \nsoil's capacity to absorb and store carbon. Regenerating healthy soil, in turn, increases its carbon \nsequestration potential and the generation of carbon removal credits. The creation of soil districts and \nrelated data and knowledge, as proposed, facilitat es the implementation of carbon removal \ncertification. \n \nValue of Healthy Soil: Certification of healthy soil is expected to elevate the value of carbon removal \ncertificates, garnering greater social and market recognition for sustainable soil management and \nassociated food and non -food products. This is likely to stimulate private f inancing, with businesses \nsupporting ecosystem services and sustainable practices related to soil health. Additionally, certified \nhealthy soil may increase the overall land value for purposes like collateral, sale, or succession. \n \nFarm to Fork Strategy: soil biodiversity monitoring also aligns with the Farm to Fork Strategy's \nobjective to reduce nutrient losses while maintaining soil fertility. It contributes to creating a more \nresilient EU food system. \n \nCommon Agricultural Policy (CAP): importantly, soil biodiversity monitoring supports CAP's efforts to \nenhance the environmental performance of the agricultural sector. It incorporates mandatory \nenvironmental and climate conditions under the CAP, including soil management practices. By \nstrengthening the CAP's innovation dimension, it encourages initiatives that address soil health issues. \n \nFarm Sustainability Data Network (FSDN): soil biodiversity monitoring is consistent with the \ntransformation of the Farm Accountancy Data Network into a Farm Sustainability Data Network as \noutlined in the Farm to Fork Strategy. The FSDN will collect farm-level sustainability data, including soil \nmanagement practices, aiding in benchmarking and advisory services for farmers. \n \nBeyond these policies, soil biodiversity monitoring plays a significant role in addressing the goals of the \nEuropean Habitats Directive by expanding taxonomic monitoring to an integral component of \necosystems that plays a crucial role in supporting various plant and animal species. The Directive also \nemphasises the restoration and maintenance of habitats in a favourable conservation status, \ntherefore, soil biodiversity monitoring provides essential data to evaluate the success of habitat \nrestoration efforts. \n \n \n4. Essential Biodiversity Variables design \n \n4.1. EBV design characteristics \n \nFrom the soil specific EBVs listed in Section 1, two soil-related EBVs were selected for the EuropaBON \nlist (https://github.com/EuropaBON/EBV-Descriptions/wiki). These are the ‘Functional Composition of \nsoil biota’ and the ‘Community Biomass of soil microbes’, representing soil biodiversity and functioning \nrespectively. Both soil -related EBVs are part of the EBV class ‘Community composition’ from the \n‘Terrestrial’ realm category. Beyond these, we will also consider two other EBVs corresponding to the \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 15 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \n‘Taxonomic Diversity of Soil Biota’ (Community Composition) and ‘Species distributions’ (Species \nPopulations). \n \n● ‘Functional Composition of soil biota ’ (EBV Class: Community Composition) describes the \nfunctional composition and diversity of soil biota based on morphological, physiological, \nphenological and behavioural traits or functional/taxonomic groups. It can be measured by \ndiversity metrics such as the functional group diversity across the whole soil community (e.g. \nrichness of soil functional groups present in a community), the species or OTU diversity within \nfunctional groups (e.g., diversity of Ectomycorrhizal fungi), or other functional diversity indices \nthat can include functional traits when available (e.g., functional richness or functional \ndivergence). \n \n● ‘Community Biomass of soil microbes ’ (EBV Class: Community Composition) refers to the \nestimated biomass of the living component of soil organic matter (bacteria, fungi and \nprotozoa) within contiguous spatial units (grid cells) across the EU over time. It can be \nmeasured by metrics such as the mass of soil microbes (i.e., bacteria, fungi and protozoa) per \nmass of dry soil, the mass of soil microbes per area. This EBV is also related to microbial activity \nsuch as microbial basal respiration, potential basal respiration, or the respiratory quotient \n(qO2; the ratio of basal respiration to microbial biomass). These last three variables refer to \nthe EBV class ecosystem function. \n \n● ‘Taxonomic Diversity of soil biota’ (EBV Class: Community Composition) refers to the diversity \nof key soil biota such as earthworms, collembola, fungi, bacteria, etc. It can be measured by \ndiversity metrics such as richness or Shannon diversity of species or OTUs for a given \ntaxonomic group (e.g. Annelida, Rotifera, Fungi). It can also be estimated from the stacked \ndistributions of species in a given taxon ( Species Distributions ; EBV Class: Species \nPopulations). \n \nAll EBVs were calculated using the 1km2 scale. In terms of the temporal resolution, it should vary from \n3 to 6 years, which is the temporal resolution of the LUCAS sampling survey (see section 4.2). Yet, for \nthe examples shown here, we are using a space for time substitution to be able to repre sent all \nvariables in both dimensions since there is no current data readily available to make temporal \nestimations. \n \n4.2. Input biodiversity data \n \nThe LUCAS (Land Use/Land Cover Area frame Survey) survey represents the main initiative aimed at \novercoming the soil monitoring gaps that have prevailed in the past at the European scale. It uses \nadvanced sampling techniques to gather comprehensive data on various aspects of soil, including its \nproperties, biodiversity, and ecological functions (Alberto Orgiazzi et al. 2022; Smith et al. 2021) . The \nLUCAS survey covers the two soil EBVs, which are measured every 3 years (starting in 2018) from 885 \nsites encompassing all the terrestrial land covers in Europe (Table 1). 'Community Biomass of soil \nmicrobes' is measured through standardized protocol s in the centralised lab in Leipzig. After a \ncoordinated sampling across European countries, soil samples are stored on ice and transported to \nIspra, Italy. From there, they are transported to Leipzig, Germany, for the measurement of potential \nbasal respiration, microbial biomass by substrate -induced respiration, and respiratory quotient. A \nrandom subset of samples is transported from Ispra to the Centro de Edafología y Biología Aplicada \ndel Segura -Consejo Superior de Investigaciones Científicas (CEBAS -CSIC, Murcia, Spain) for the \nmeasurement of ester -linked fatty acid methyl esters (FAMEs) as indicators of bacterial and fungal \nbiomass. A detailed description of the methods used to do these measurements can be found in (Smith \net al. 2021). Measuring the ‘Taxonomic Diversity of soil biota’ or the 'Functional Composition of soil \nbiota' requires a more comprehensive approach to sample the whole soil biota, and their functional \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 16 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nattributes for the latter. For this, the LUCAS survey employs eDNA metabarcoding, which offers a \nstandardized and cost -effective way to cover all kingdoms of life in soil samples. Details on the \nmethods used to extract and process eDNA from soil samples ca n be found in (Alberto Orgiazzi et al. \n2022). From eDNA data, the diversity of several taxa from the soil can be assessed to assess the \n‘Taxonomic diversity of soil biota’ (e.g., (Labouyrie et al. 2023; Köninger et al. 2023) ). The functional \naspect can be added by using existing functional databases allowing to assign different soil taxa to \nfunctional groups (e.g. FungalTraits for Fungi (Põlme et al. 2020) ). Data collected from the LUCAS \nsurvey is available at the European Soil Data Center (ESDAC). Yet, it's important to note that eDNA \nmetabarcoding may have some biases, favouring the detection and taxonomic resolution for microbes \nand microfauna, while potentially underrepresenting or providing less resolution for macrofauna, due \nto lower DNA presence in the soil. To achieve a more complete soil biota assessment of macrofauna \ntaxa, other existing non-commercial data infrastructures can be used, such as Eudaphobase (Table 1), \nwhich combines data from heterogeneous sources on soil animals, their distribution and habitat \nparameters of their sites of occurrence and makes these data available to the public. Other databases, \nfrom global to local, may exist for spe cific soil taxa, but will not be described here (e.g., section 4.3.2 \nfor the case of earthworms). \n \nTable 1 Raw data availability and access \nDATASET \nTITLE \nLUCAS Soils EUdaphobase GBIF \nRaw data \ncollection \ndesign \nOver ~20000 soil sampling \npoints across all EU member \nstates revisited every 3 years \n(starting in 2009) \nCurrently a small fraction \n(885 plots [2018], 1500 plots \n[2022]) of samples are used \nto measure microbial biomass \nand activity metrics and \naddress taxonomic \ncomposition of living \ncomponents of topsoil. \nStarted in 2018 across all EU-\nMS, sampled every 3 years \nMinimum sampling unit likely \nadequate for 1 x 1 km spatial \nresolution \n(https://esdac.jrc.ec.europa.e\nu/public_path/shared_folder/\ndoc_pub/JRC105923_LUCAS2\n018_JRCTechnicalReport.pdf). \nNon-commercial data \ninfrastructure developed by the \nSenckenberg Museum of Natural \nHistory Görlitz in Germany. \nCombines data from \nheterogeneous sources on soil \nanimals, their distribution and \nhabitat parameters of their sites \nof occurrence and makes these \ndata available to the public \n(open access). \nThe data originate from the \nscientific literature, unpublished \nresults of field studies (theses, \nreports), collections of museums \nand research institutions as well \nas raw data from research \nstudies and well-founded \nobservations. \nNon-commercial data \ninfrastructure. \nCombines data from heterogeneous \nsources on soil animals and \nmicrobes related to their \noccurrence, making these data \navailable to the public (open \naccess). \nThe data originates from the \nscientific literature, monitoring \nprograms, citizen science, \nunpublished results of field studies \n(theses, reports), collections of \nmuseums and research institutions. \nMonitoring \nprograms \nThe LUCAS Soil module is \ncoordinated by the European \nCommission’s Joint Research \nCentre (JRC). This survey \nrepresents the first attempt \nto construct a pan-European \ntopsoil database, which can \nserve as a baseline for EU-\nwide soil monitoring (A. \nOrgiazzi et al. 2018). It has a \nstandardized sampling \nprocedure & central \nlaboratory for measurements \nand samples processing. \nN/A N/A \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 17 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nTypes of data \naccess \nOpen access. \nData is available and \ndownloadable after \nregistration. \nOpen access. \nThe data warehouse is publically \navailable (open access) via a \nweb-based browser portal. \nOpen access. \nThe data warehouse is publically \navailable (open access) via a web-\nbased browser portal. \nData \nrepositories \nDatabase creation on \nEuropean Soil Data Center \n(ESDAC). \nhttps://esdac.jrc.ec.europa.e\nu/projects/lucas \nEUdaphobase will adapt the soil-\nzoological data platform \n“Edaphobase” to a pan-\nEuropean data warehouse for \nsoil biodiversity. \nData available through the GBIF \nrepository at www.gbif.org \nPersistent \nidentifier(s) \nN/A N/A N/A \nMetadata \ndescription \nMicrobial biomass is \nmeasured from topsoil \nsamples with substrate-\ninduced respiration; units [µg \nCmic g soil dw-1]. Other \nrelated measurements: \nRespiratory quotient, basal \nrespiration. \nMetagenomics with DNA \nmetabarcoding, including \nprimers for Bacteria and \nArchaea (16S rDNA), Fungi \n(ITS) and Eukaryotes (18S \nrDNA), which includes \nmicrofauna (e.g., nematodes), \nmesofauna (e.g., arthropods), \nand macrofauna (e.g., \nearthworms). \nCurrently includes data on \nNematoda, Collembola, \nOribatida, Gamasina, Chilopoda, \nDiplopoda, Isopoda, \nEnchytraeidae, and Lumbricidae, \ntheir distribution and habitat \nparameters of their sites of \noccurrence. \nData types comprise modern \ntaxonomic nomenclatures and \nsynonyms, geographical \nreferences, quantities of \ncollected organisms, soil \nparameters, vegetation, \nmeteorological data, sampling \nand extraction methods, \nidentification methods, \npreparation techniques and \nbehavioural data. \nOccurrence data of selected \nearthworm species using the Darwin \nCore standard \nQuality \ncontrol \nLUCAS general survey’s \nresults undergo a rigorous \nquality assurance process. \nInitially, an automated check \nensures data completeness \nand consistency, either during \ndata compilation or when \nuploaded to a central \nrepository. Subsequently, \nregional or central offices \nconduct visual inspections of \nall surveyed points. Data \npoints needing correction or \nclarification can be returned \nto field contractors for \nrectification. The same \napplies to the analytical \ninformation. \n \n \n4.3. The EBV model \n \nDifferent models were used to create the different soil-related EBV outputs. Here we present in detail \nthe methods used to create each EBV output map, including which data was used and how it was \nprocessed to obtain the EBV value, a description of the mode l used and the selected parameters, the \nperformance of the model and how uncertainty of the model was assessed. In some cases, e.g. for the \ncommunity biomass of soil microbes, different methods were used, as they were developed for \ndifferent aims. \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 18 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \n \n \n4.3.1 ‘Functional Composition of soil biota’ (EBV 64) \n \nInput data \nSoil biodiversity data was obtained from the LUCAS survey from 2018 (Origiazzi et al 2018). Raw eDNA \ndata for eukaryotes (18S) and prokaryotes (16S) was obtained from the ESDAC platform and cleaned \nfollowing a standard bioinformatic pipeline using DADA2 (Callahan et al. 2016) to obtain amplicon \nsequence variants (ASV) with taxonomic annotations. Eukaryote data was separated into Animals and \nProtists, and the rest discarded for this task. Protists were classified into functional groups using the \nfunctional information available in (Adl et al. 2019) . Due to the low taxonomic resolution of the 18S \nmarker for animals, these were classified into the key soil taxa, and only nematodes were functionally \nannotated using the NINJA platform (Sieriebriennikov, Ferris, and de Goede 2014). Bacteria ASVs were \nfunctionally annotated using FAPROTAX (Louca, Parfrey, and Doebeli 2016). For fungi, OTU tables with \ntaxonomic and functional annotations, derived from the ITS marker, were obtained from (Labouyrie \net al. 2023). For each broad taxonomic group, i.e., Bacteria, Protists, Fungi and Animals, the diversity \nof the functional groups identified was estimated using the Shannon diversity, i.e., the exponential of \nthe Shannon entropy. \nCovariates related to land cover and soil physical and chemical properties, including percentage of \ncoarse, sand, clay and silt, bulk density, pH, total nitrogen, potassium and phosphorus were also \nobtained from the LUCAS survey. For projections across Europe we used the values from the available \nmaps in the ESDAC. Climatic data was extracted from CHELSA v1.2 in the period from 1990 to 2020. \nSoil degradative processes such as soil erosion, soil pollution (i.e., heavy metals content) and soil \ncompaction were also accessed through the ESDAC platform or obtained directly from the EEA. \n \n \nModel description \nFor each response variable, i.e., the Shannon diversity of the selected 28 functional groups, the dataset \nwas first preprocessed by handling missing values, feature scaling, and encoding categorical variables, \nwhere applicable. A Random Forest regression model was employed to capture the effect of complex \ninteractions between the previous environmental variables on the diversity of the different functional \ngroups. The mean squared error was set as the optimization criteria and optimised key \nhyperparameters of the random forest algorithm (maximum tree depth, maximum features per tree) \non a hold out validation set. The model was implemented using Python's scikit-learn library. To assess \nthe ability of the model to generalise to new settings, each observation w as assigned according to its \nlocation to a grid cell (~100km). Afterwards, a random partition of the grid cells into 10 subsets for \nspatial block cross validation was generated. For each fold in turn, a regression model was trained on \nobservations of the r emaining folds and evaluated on its observations. The model’s predictive \nperformances were assessed using Mean Squared Error (MSE), the R -squared (R2) and the coefficient \nof determination (Spearman's rho2). Finally, the cross validation process resulted in 10 fitted models. \nAn ensemble of these models was used to predict the average and variance of the diversity of each \ntrophic group and the overall functional composition of soil biota (Figure 3, Figure 4). \n \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 19 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \n \nFigure 3 Workflow representation for modelling the ‘Functional Composition of soil biota’. The EBV model was \ntrained on data for measured samples, consisting of the diversity of 28 functional groups of soil biota, the \nenvironmental dataset and soil degradative processes. Environmental projections across Europe were used to \nbuild the EBV maps for each functional group and then for the overall functional composition of the soil biota by \nmaking a PCA with all the modelled functional groups. \n \nFigure 4 ‘Functional Composition of soil biota’. Here represented by the first two axes of a PCA created from \nthe diversity of 28 functional groups across Europe are presented in the figure using a RGB colour space. Different \ncolours represent a different functional composition that can be characterized from the right panels. While we \nuse here this representetaion to provide an overview of the variability in functional conditions, the EBV should \nbe represented with an Alpha diversity metric. \n \nModel performance \nThe ensemble models for across functional groups had an average±error R2 of 0.47±0.006, a MSE of \n2083.2±392 and a coefficient of determination of 0.78±0.002 . \n \nUncertainty assessment \nThe uncertainty of the model was estimated using the coefficient of variation (CV), which assesses the \nconsistency or variability of predictions across the individual models in the ensemble. The coefficient \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 20 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nof variation is a relative measure of variation and is defined as the ratio of the standard deviation to \nthe mean. It indicates how much the predictions of individual models deviate from the ensemble's \nmean prediction. The CV of all the ensemble models for the different functional groups was summed, \nand the areas having a probability above 0.9 of the total distribution of the summed CV were \nconsidered as uncertain. \n \n4.3.2 ‘Taxonomic diversity of soil biota’ \n \nInput data \nAvailable occurrence data of earthworm species from the Global Biodiversity Information Facility \n(GBIF; (GBIF.org 2022)), Edaphobase (Edaphobase.org 2021), and other existing large datasets were \ncollected (Figure 5). After data collection, records collecte d before 1970 were removed to estimate \ncurrent rather than past species occurrence patterns. The R package CoordinateCleaner was used to \nexclude data with common spatial and temporal errors. It was manually checked that record exclusion \nwas appropriate. Records removed included those from GBIF with coordinate uncertainty of more than \n1 km, non-observational records (i.e. keeping only living specimen, human observation, or preserved \nspecimen), records that were based on less than one observation, and those d escribing not species \nbut higher taxonomic levels. The combined dataset contained 98,732 occurrence records of 142 \nunique earthworm species collected across 45 European countries. The occurrence records were \nspatially thinned to adjust for sampling bias ca used by varying sampling density and resolution. To \navoid overfitting and increase robustness of the models, only the 41 species that were observed in at \nleast 10 grid cells were kept. Background data for the 41 focal species (i.e., number of records ≥10) \nwas generated with the R package biomod2 by randomly sampling 10,000 grid cells within the spatial \nextent of the environment for each species. Four groups of environmental descriptors were included \nin the model, namely climatic, land -use, topographic and edaphic factors. The CORINE Land Cover \ndataset from 2012 was used as a land-cover baseline. All environmental data were reprojected into a \n2 km2 and 5 km² grid system by either up- or downscaling, and standardized by dividing mean-centered \nvalues by their respective standard deviations. The variables were reprojected into WGS84 using \nArcGIS v10.7.1, if necessary, but merged them into one table in R to avoid conflicts with missing data. \nAreas with one or more missing values in environmental variables were excluded. The R package usdm \nwas used to check for correlation and collinearity between variables. Variables with r>0.8 (correlation \ncoefficient of Pearson) and variable inflation factor VIF≥10 were excluded from further analysis. To \navoid overfitting, the t en most important predictors for the focal earthworm species of the 24 least \ncorrelated predictors were identified. Future predictions of mean annual temperature (T) and \nseasonality of precipitation (P) were downloaded from CHELSA v1.2 for all available IP CC scenarios \n(n=15) and the time period from 2041 to 2070. The 3 Shared Socio-economic Pathways are SSP1 (with \nthe Representative Concentration Pathway [RCP] 2.6): Sustainability, additional radiative forcing of 2.6 \nwatt/m² by 2100; SSP3 (RCP 7.0): Regional Rivalry, 7 watt/m², middle to upper range of the bandwidth \nof all scenarios; and SSP5 (RCP 8.5): Fossil-fueled Development, 8.5 watts/m², upper edge in the range \nof scenarios described in the literature (SSPs; (Birch 2014)). The 5 Earth system models are GFDL -\nESM4, IPSL-CM6A-LR, MPI-ESM1-2-HR, MRI-ESM2-0, and UKESM1-0-LL. \n \nModel description \nThe maxent function from dismo in R (https://cran.r-project.org/web/packages/dismo/dismo.pdf) was \nused to perform Species Distribution Models (SDMs) for each of the 41 species present in more than \nor exactly ten grid cells (~2 km²) (Figure 5). Models were allowed to be tuned individually and the same \npseudo-absence dataset as for model fitting us ed. To identify the top ten variables for determining \nearthworm distribution, permutation importance was extracted, by permuting the values of each \npredictor and comparing the resulting reduction in training Area Under Curve (AUC) values. A large \nreduction in AUC (i.e., high permutation importance) indicates that the model is strongly influenced \nby that predictor. For species with n≥100 occurrences, ten variables with the highest median \npermutation importance across the 19 species were selected. Accordingly, the ten predictor variables \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 21 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nused for Species Distribution Modeling were annual mean temperature, precipitation seasonality, \ndistance to coast, proportion of area covered by agriculture, soil pH, phosphorus content, cation \nexchange capacity, elevation, clay+silt content, and human population density (Zeiss et al. 2023). The \n19 earthworm species with ≥100 records were modelled to avoid overfitting, and the ten algorithms \navailable in biomod2 (Thuiller et al. 2016) with adapted parameter settings were used. Committee \naveraging scores of the predictions, ten -fold cross-validation (80:20%), and True Skill Statistic (TSS) \nwere used to improve model performance during ensemble building. The committee averaging score \nis the average of the binary predictions of the individual models, giving both a prediction and a \nmeasure of uncertainty. If the prediction is close to 0 or 1, all models agree to predict 0 and 1, \nrespectively, while a prediction around 0.5 means that half of the models predict 1 and the other half \n0. During individual model building, less weight was given to older species observations as they do not \nnecessarily correspond to current species’ occurrences. For model evaluation, the Cohen’s Kappa, area \nunder the receiver operating characteristic curve (AUCROC), and TSS from BioMod output were used. \n \n \nFigure 5 Workflow representation for modelling the Taxonomic Diversity of Earthworms from (Zeiss et al. 2023). \nBuilding the species occurrence (left side) and environmental dataset (predictors, gray panel). n - Number of \noccurrence records before and after cleaning or spatial thinning. Earthworm icon was designed by Iconjam \n(flaticon.com). \n \nThe distribution of the 19 earthworm species was predicted and mapped with the best -performing \nmodels (i.e., the one with highest TSS) in R . The discrete presence-absence maps, which were derived \nfrom probability maps and model -specific probability threshold giving highest TSS, were summed to \nget the number of species per grid cell (i.e., Taxonomic Diversity, Figure 6). All maps were cropped to \nthe area in which prediction uncertainty, averaged across the 19 species distribution models, was \nlower than 0.1 (= mean and median uncertainty). The resulting investigated area spanned 864,140 km² \n(i.e., 172,828 grid cells à 5 km²). The future distribution and species richness of earthworms was \nprojected using current land cover, topographic and soil variables toget her with future climate \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 22 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nvariables. Only future climate variables were used, as we were interested in the potential climate effect \nrather than interactive effects of land cover and soil variables under future scenarios. Future \nearthworm distributions were predicted with 3 environmental datasets: both future climate variables \n(T and P); future T and current P; and current T and future P, resulting in 45 future projections per \nspecies. One ANOVA (lm function) including the ten environmental predictors was performed to \ncompare the 3 future projections of each scenario (factorial; T, P, TP) and to evaluate potential climate \neffects on the predicted species richness averaged across the 15 climate scenarios. \n \nFigure 6 Output map for the EBV ‘Taxonomic diversity of soil biota’ represented by earthworm species rchiness. \n(A) Spatial distribution of the current predicted overall species richness calculated as the number of species (max. \n19 species) being present with a probability higher than the species-specific threshold with maximum TSS value. \nDark gray areas indicate species richness values of 0. (B) Agreement across the 3 different Shared Socioeconomic \nPathway (SSP) scenarios. Gain represents areas in which gain of species richness is predicted in 3/2/1 scenarios, \nwhile remaining scenarios predict no change; loss represents areas in which loss is predicted, and mixed \nrepresents areas in which different scenarios predict gain and loss. Light gray areas were not predicted based on \nthe extent of the environmental dataset. Modified from (Zeiss et al. 2023). \n \nModel performance \nSDMs for the potential current and future distributions of the 19 highly recorded species of \nearthworms showed generally good performance (mean: Kappa=0.53 [SD 0.12], AUC=0.93 [SD 0.04], \nTSS=0.70 [SD 0.13]). \n \nUncertainty assessment \nUncertainty was estimated as the coefficient of variation of each SDMs under current climatic \nconditions and varied between species. \n \n4.3.3 ‘Community Biomass of soil microbes’ (EBV 61) \n \n4.3.3.1 Current distribution \n \nInput data \nData was obtained from the LUCAS survey from 2018 (A. Orgiazzi et al. 2018) . Variables included to \nrepresent the EBV were: potential basal respiration, microbial biomass and respiratory quotient. \nSamples from 185 grasslands, 289 forests , 347 croplands and 64 samples from other land-cover types \nincluding shrublands, bare land, and urban areas were included. Covariates included soil, climatic and \ngeographical variables, i.e., soil organic carbon content, soil water content, sand content, pH, annual \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 23 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nprecipitation, annual temperature, mean temperature of the 30 days preceding sampling, total \nprecipitation of the 30 days preceding sampling, latitude and elevation. Soil properties were taken \nfrom the LUCAS 2018 survey, and climate data were obtained from online databases. \n \nModel description \nStructural equation models (SEMs) were constructed using lavaan version 0.6.5. A general model was \nbuilt to unravel drivers of soil microbial respiration and biomass without any regard to land cover; so, \nfor this model, all available samples from the full range of land -cover types of the LUCAS Soil \nBiodiversity survey were included. To investigate whether and how different land -cover types affect \nsoil properties in different ways, multigroup analysis were used to consider the effects of three broad \nland-cover types (forest, cropland and grassland) individually, creating an SEM for each land -cover \ntype, to explore potential differences in the relationships between soil, climate, geographical and \nmicrobial parameters. This enables more precise predictions of s oil microbial properties and carbon, \nas models are parameterized to the conditions of the different land -cover types. The underlying \nstructural equations of the SEMs were used to create predictive maps of respiratory quotient, potential \nbasal respiration and microbial biomass across the European Union with a monthly step for 2018 and \nthen aggregated to obtain an average for the year (Figure 7). This was done because predictions are \nspecific to a 30 -day time period, due to inclusion of the mean temperature a nd total precipitation in \nthe 30 days before sampling. Predictions were made for the three respective land -cover types and \nthen aggregated to a single spatial representation across Europe. Sand content, pH, and carbon maps \ncame from studies published by th e European Soil Data Center (ESDAC), based on LUCAS 2009/2012 \nsoil property data. The predictive maps do not include any urban areas or areas above 1,000 m a.s.l., \nas these were not included in the modelled organic carbon map we used as input and were not (well-\n)represented in the LUCAS data used to create the models. \n \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 24 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \n \nFigure 7 Output maps for the EBV ‘Community Biomass of soil microbes’ taken from (Smith et al. 2021) . \nPredictive maps of mean (a) respiratory quotient, (b) microbial biomass and (c) potential basal respiration at 20 \nºC, and in 2018 across the European Union, excluding altitudes over 1,000 m, created by averaging the predictive \nmaps created for each month of 2018. \n \nModel performance \nThe R2 of the general model was 0.487. The R 2 of the models for each land cover type were: 0.359 for \ngrasslands, 0.404 for croplands and 0.655 for forests. \n \nUncertainty assessment \nThe environmental coverage of the current sampling design was estimated to evaluate the spatial \nuncertainty of the predictions using the Mahalanobis distance, which estimates a multidimensional \ndistance, and defined outliers as the 97.5% quantile of the ch i-squared distribution with n degrees of \nfreedom. This algorithm permitted identifying regions where the predictive maps were more or less \nreliable (Figure 8). \n \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 25 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \n \nFigure 8. Extrapolation of uncertainties associated with the survey used in this study. This map represents those \nlocations for which environmental conditions are better covered by our samples. Overall, our dataset covers a \nwide range of the terrestrial environmental conditions found in Europe, the rest being considered as outliers in \nrespect to the conditions covered by the sample sites used in this study (for this purposes an outlier is a location \nthat has a value above the 97.5% quantile of the Chi-squared distribution; in red). \n \n4.3.3.2 Future scenarios \n \nInput data \nThe variable response used was the microbial biomass from the LUCAS 2018 survey. Predictors \nincluded variables of soil chemical and physical properties, topography, climate, and land cover. \nBecause soil chemical variables may depend on the climate, their v alues were predicted first for the \ndifferent SSP, using the same model. These predictions were then used to predict the value of the \nmicrobial biomass in the future. Future predictions of climatic variables were downloaded from \nCHELSA v1.2 (Karger et al. 2017) for all available IPCC scenarios (n=15) and the time period from 2041 \nto 2070. The 3 Shared Socio -economic Pathways are SSP1 (with the Representative Concentration \nPathway [RCP] 2.6): Sustainability, additional radiative forcing of 2.6 watt/m² by 2100; SSP3 (RCP 7.0): \nRegional Rivalry, 7 watt/m², middle to upper range of the bandwidth of all scenarios; and SSP5 (RCP \n8.5): Fossil-fueled Development, 8.5 watts/m², upper edge in the range of scenarios described in the \nliterature (SSPs; (Pörtner & Roberts 2022)). The 5 Earth system models are GFDL-ESM4, IPSL-CM6A-LR, \nMPI-ESM1-2-HR, MRI-ESM2-0, and UKESM1-0-LL. \n \nModel description \nTo predict the future scenarios of soil chemical properties and further ‘ Community Biomass of soil \nmicrobes’, the same model described in 4.3.1 was used. The results are given in Figure 9. \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 26 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \n \nFigure 9 Output map for the future projections of the EBV ‘Community Biomass of soil microbes’. Differences in \nmicrobial biomass compared to the current scenario for the 3 different Shared Socioeconomic Pathway (SSP) \nscenarios: SSP1 in the left, SSP3 in the center and SSP5 in the right. Gain represents areas in which microbial \nbiomass is predicted to increase (red), while loss represents areas in which loss in microbial biomass is predicted \n(blue). \n \nModel performance \nThe ensemble model had an R2 of 0.85, a MSE of 24038.72 and a coefficient of determination of 0.87. \n \nUncertainty assessment \nThe uncertainty of the model was estimated using the coefficient of variation (CV), which assesses the \nconsistency or variability of predictions across the individual models in the ensemble. \n \n \n4.4. EBV-derived policy indicators \n \nThe use of soil -related EBVs can significantly improve the evaluation of ecological soil conditions, \nthereby facilitating a more comprehensive assessment of ecosystem health. This approach holds \nparticular relevance within the context of nature restoration legislation targets of evaluating the state \nof habitats and ecosystems, and in the context of the proposal for a Soil Monitoring Law, which \nreinforces the pressing need to prioritise soil monitoring for protection and restoration efforts in \nEurope. For this aim, valuable insights drawn f rom previous experiences with other environmental \ncompartments and directives can provide valuable guidance (Beck et al. 2005) . For instance, the \nconcept of ecological status, as defined in the European Water Directive Framework (WDF, European \nCommission, 2000), presents an integrated assessment that takes into account the various ecological \nfacets of aquatic ecosystems. This framework employs a range of ecological indicators, encompassing \nboth biological and physico-chemical parameters, to evaluate the condition of habitats. It sheds light \non the anthropogenic influences impacting ecosystem well -being and contributes to the formu lation \nof sustainable management strategies. In this framework, when there are slight but not significant \ndeviations of indicators from their natural or undisturbed states, the ecosystem can be designated as \nhaving a \"high\" ecological status, which is to b ecome a mandatory target for management (Birk et al. \n2012). In the context of soils, the undisturbed state refers to the natural condition of the soil system \npreceding any human intervention, but it can also denote the ecologically favorable state within a \nspecific habitat type. \nThe assessment of soil ecological status forms the base upon which complementary policies can be \ndeveloped, as depicted in Figure 1. One such policy initiative involves devising concrete strategies for \nthe restoration or enhancement of soils categorized as degraded or in a moderate state. Another \nsignificant policy priority centers on identifying strategic areas for the conservation of soil taxa and the \necosystem services they provide. When evaluating the attainment of good ecological status for soils, \nthe identification of restoration potential or soil nature conservation areas necessitates the \nconsideration and protection of the diverse ecological dimensions supported by soils. This includes \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 27 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \ntheir biodiversity, community composition, uniqueness, and the spectrum of ecosystem services they \ncontribute to (Guerra et al. 2022). Preserving these areas requires tailored approaches for restoration \nand conservation to ensure their preservation (Zeiss et al. 2022) . Here, we present the framework \ndeveloped to assess the ecological status with the use of soil related-EBVs produced as detailed in 4.3, \nto further identify areas of high restoration potential. \n \nFigure 10 Schematic representation and classification of the ecological status of soils. In green, a \ngradient of healthy soils is shown from “high ecological status”, referring to soils with high biological and \nfunctional activity and structure, to “moderate ecological status”, referring to still non-degraded soils with \nmoderate levels of biological and functional activity, both estimated when comparing to the same \nreference system (e.g., a combination of biogeographical region and land-use type). In dark orange and \nred, the soil degradation gradient is shown from “critically degraded soils”, referring to soils with high \nlevels of degradation and cumulative impacts resulting from severe or continued disturbances, to “ \ndegraded soils”, referring to soils with severe loss of function and/or biological activity but with less \ncumulative impacts. Black dashed arrows indicate different strategies for the improvement of the soil's \necological status and the potential for healthy soils to be sustainably used (e.g., fo r agricultural \nproduction) and to be spared for nature conservation of soil organisms. \n \n4.4.1. Methods for producing EBV-derived state indicators and trends \n \nThe ecological status of the European soils can be categorized into the following categories: critical \ndegraded soils, degraded soils, moderate status, good status, high status, as exemplified in Figures 10, \n11 and 12. The European commission has already e stablished a procedure to classify degraded soils. \nIn their approach, soils with at least one soil degradation process being above the defined threshold \nby the EEA, are considered as degraded. For simplification, soils with more than 1 degradation process \nabove the threshold can be classified as critically degraded. The rest of the soils can then be classified \ninto moderate, good or high status. For this, a ‘reference undisturbed state’ is needed to serve as a \nbaseline to compare the current EBVs state. A s olution is to use the predictions of the model used to \ncreate the EBVs (see section 4.3), but setting to 0 or the minimum all the soil degradation processes. \nFor this, soil degradation processes need to be included from the beginning in the model predictor s \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 28 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \n(e.g., see model described in 4.3.1). This ‘reference undisturbed state’ indicates the expected value of \nEBVs in a scenario where no soil degradation processes exist. \n \n \nFigure 11 Worflow for the classification of European soils based on their ecological status and to implement \nindicators for restoration and conservation. \n \nNext, the percentage of difference between the current EBV state and the ‘reference undisturbed \nstate’ can be calculated to define the ecological status of soil. If the difference between states is \nstatistically high, it means that the site is very distan t from the modelled reference conditions, so it \ncan be considered as having a moderate state for a given EBV or EBV metric. If the difference is \nintermediate, the site can be classified as having a good status. If the difference is low or insignificant, \ni.e., the value of the indicator is similar to the value expected from the reference conditions, then it \ncan be considered as having a high status. The overall ecological status classification for a soil area \n(pixel 1km) is determined by the element, i.e., EBV or EBV metric, with the worst status out of all the \necological indicators following the ‘one out, all out’ principle used in the WFD (European Communities, \n2005). It is advisable to conduct a separate evaluation based on habitat or land cover types an d \nbiogeographical regions. This is crucial due to the unique characteristics exhibited by certain soils, \nwhich may be atypical by nature and constitute rare habitats for biodiversity or unique landscapes. \nAlternatively, some soils may have undergone significant human modifications, such as those found in \nagricultural landscapes. These distinctive features should be duly considered when defining soil \nconditions and establishing the prerequisites for maintaining a healthy soil environment. \nIn order to effectively enhance degraded soils through soil improvement strategies, it is imperative to \nassess the potential for their restoration. One approach to evaluating this potential is by considering \nthe percentage of difference between the current EBV state and the 'reference undisturbed state' used \nto assess the ecological status of soils (Figure 12). A high difference between the current state and the \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 29 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nreference state signifies a significant potential for improvement for a given EBV. This potential can be \nrealised through the implementation of effective management strategies aimed at mitigating soil \ndegradation processes. \n \nFigure 12 Representation of the soil ecological status assessments in Europe. Degraded soils refers to soils one \n(degraded) or more (critically degraded) soil degradation process being above the critical limit (defined by the \nEEA). The rest of the soils were classified into moderate, good or high based on the deviation of soil-related EBV \nmetrics from an undisturbed reference state. \n \n4.4.2. Relevance of the new EBV product for policy reporting as compared to customarily used data \nflows. \n \nCurrent soil health assessments predominantly focus on categorizing soil conditions based on \nindicators related to soil degradation, such as erosion and heavy metal concentrations. These \nassessments designate soils as degraded when specific indicators exceed predefined thresholds, often \nreferred to as critical limits, as mentioned above. Typically, these assessments rely on physical and \nchemical measurements, including soil pH, carbon content, nutrient availability, chemical levels, and \nsoil texture. While these indicators provide valuable insights into soil health, they may not capture the \nfull complexity of soil biodiversity and functioning (Römbke et al. 2005) . Integrating EBVs related to \nsoil biodiversity and functions into these assessments, as shown in the previous section, may help to \naddress this limitation. These EBVs are rooted in the functional roles that soils play in supporting vital \necosystem process es such as carbon sequestration and rely on biological measurements, \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 30 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nencompassing soil biodiversity, activity levels, and biomass. The incorporation of soil biotic and \nfunctional EBVs offers a more comprehensive perspective on soil health and quality, enhancing our \nability to refine the classification of soil ecological status. By integrating these novel indicators into soil \nhealth assessments, we can attain a more holistic understanding of soil conditions. This, in turn, \ncontributes significantly to enhancing the accuracy of soil ecological status classifications, making ou r \npolicy reporting more robust and informed (Figures 10 and 12). \n \n \n5. Discussion \n \n5.1. Advantages and caveats of the EBV result \n \nThe main caveat of the current approach is its limitation in terms of taxonomic resolution. The current \ndatasets and the current knowledge about the taxonomic identity of many soil organisms is still limited \nboth in Europe and in the world. This is also wh ere the EBV approach has particular advantages by \nallowing a better representation of the distribution of particular species and the study of the potential \neffects of climate and land -use change. By using a limited pool of data for a particular species or \nfunction and transforming it into continuous representations of its distribution, EBVs provide a good \nbaseline for decision making but also to improve the quality of monitoring by identifying critical areas \nthat require further attention. \n \n5.2. Breakthroughs and lessons learned: \n \nBeing able to represent the taxonomic and functional diversity of soil organisms for Europe is an \nimportant step to show the potential that the current datasets (including LUCAS Soils) already have. \nThis also led to the classification of European soils into an ecological indicator depicting their ecological \nstatus. In both cases, particularly in the framework of the current soil monitoring law that is being \ndiscussed, these are valuable information to the identification and followup of the health of European \nsoils. Both goals (identifying healthy soils and tracking their progress), would benefit from more \nspatially, and more importantly, temporally extended datasets so that the analyses benefit from more \nup to date information. \n \n5.3. Outstanding challenges and proposed solutions \n \nOverall, no outstanding challenges were identified. That said, there is still a significant lack of validation \nof the soil organisms observations found in GBIF. This is mainly driven by the important amount of \ncitizen science data coming from iNaturalist d irectly to GBIF. While this information maybe relevant \nfor other taxonomic groups, in the case of soil organisms, the identification and classification is often \nflawed (e.g., a great number of miss identifications of Lumbricus terrestris since this is one of the most \ncommon earthworms present in Europe). Therefore, there is the need for a great deal of data curation \nprior to the use of such datasets in modeling approaches. At the same time, the LUCAS dataset, while \nnow the coverage is imp roving with the 2022 sampling, still lacks representation of more extreme \nconditions (e.g., holgy polluted sites, higher elevations), particularly for the subset regarding soil \nbiodiversity and function. That said, we were also able to show (Figure 8) that this is overall a minor \nissue when considering the European terrestrial systems. \n \n6. References \nAdl, Sina M., David Bass, Christopher E. Lane, Julius Lukeš, Conrad L. Schoch, Alexey Smirnov, Sabine \nAgatha, et al. 2019. “Revisions to the Classification, Nomenclature, and Diversity of Eukaryotes.” \nThe Journal of Eukaryotic Microbiology 66 (1): 4–119. \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 31 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nAlmendra-Martín, Laura, José Martínez-Fernández, María Piles, Ángel González-Zamora, Pilar Benito-\nVerdugo, and Jaime Gaona. 2022. “Influence of Atmospheric Patterns on Soil Moisture \nDynamics in Europe.” The Science of the Total Environment 846 (November): 157537. \nBallabio, Cristiano, Panos Panagos, and Luca Monatanarella. 2016. “Mapping Topsoil Physical \nProperties at European Scale Using the LUCAS Database.” Geoderma 261 (January): 110–23. \nBardgett, Richard D., and Wim H. van der Putten. 2014. “Belowground Biodiversity and Ecosystem \nFunctioning.” Nature 515 (7528): 505–11. \nBeck, Ludwig, Jörg Römbke, Anton M. Breure, and Christian Mulder. 2005. “Considerations for the \nUse of Soil Ecological Classification and Assessment Concepts in Soil Protection.” Ecotoxicology \nand Environmental Safety 62 (2): 189–200. \nBirch, Eugenie L. 2014. “A Review of ‘climate Change 2014: Impacts, Adaptation, and Vulnerability’ \nand ‘climate Change 2014: Mitigation of Climate Change.’” Journal of the American Planning \nAssociation. American Planning Association 80 (2): 184–85. \nBirk, Sebastian, Wendy Bonne, Angel Borja, Sandra Brucet, Anne Courrat, Sandra Poikane, Angelo \nSolimini, Wouter van de Bund, Nikolaos Zampoukas, and Daniel Hering. 2012. “Three Hundred \nWays to Assess Europe’s Surface Waters: An Almost Complete Overview of Biological Methods \nto Implement the Water Framework Directive.” Ecological Indicators 18 (July): 31–41. \nBorrelli, Pasquale, Jean Poesen, Matthias Vanmaercke, Cristiano Ballabio, Javier Hervás, Michael \nMaerker, Simone Scarpa, and Panos Panagos. 2022. “Monitoring Gully Erosion in the European \nUnion: A Novel Approach Based on the Land Use/Cover Area Frame Survey (LUCAS).” \nInternational Soil and Water Conservation Research 10 (1): 17–28. \nCallahan, Benjamin J., Paul J. McMurdie, Michael J. Rosen, Andrew W. Han, Amy Jo A. Johnson, and \nSusan P. Holmes. 2016. “DADA2: High-Resolution Sample Inference from Illumina Amplicon \nData.” Nature Methods 13 (7): 581–83. \nEiden, G., C. Vidal, and N. Georgieva. 2002. “Land Cover/Land Use Change Detection Using Point Area \nFrame Survey Data. Application of TERUTI, BANCIK and LUCAS Data.” Building Agro-\nEnviromnental. https://www.researchgate.net/profile/Claude-\nVidal/publication/267985758_Land_usecover_change_with_point_surveys_Land_CoverLand_U\nse_change_detection_using_point_area_frame_survey_data_Application_of_TERUTI_BANCIK_a\nnd_LUCAS_Data/links/54b8e6b00cf28faced625b1b/Land-use-cover-change-with-point-surveys-\nLand-Cover-Land-Use-change-detection-using-point-area-frame-survey-data-Application-of-\nTERUTI-BANCIK-and-LUCAS-Data.pdf. \nFoley, Jonathan A., Navin Ramankutty, Kate A. Brauman, Emily S. Cassidy, James S. Gerber, Matt \nJohnston, Nathaniel D. Mueller, et al. 2011. “Solutions for a Cultivated Planet.” Nature 478 \n(7369): 337–42. \nFreibauer, Annette, Mark D. A. Rounsevell, Pete Smith, and Jan Verhagen. 2004. “Carbon \nSequestration in the Agricultural Soils of Europe.” Geoderma 122 (1): 1–23. \nGraaff, M-A de, J. Adkins, P. Kardol, and H. L. Throop. 2015. “A Meta-Analysis of Soil Biodiversity \nImpacts on the Carbon Cycle.” Soil 1 (1): 257–71. \nGuerra, Carlos A., Richard D. Bardgett, Lucrezia Caon, Thomas W. Crowther, Manuel Delgado-\nBaquerizo, Luca Montanarella, Laetitia M. Navarro, et al. 2021. “Tracking, Targeting, and \nConserving Soil Biodiversity.” Science 371 (6526): 239–41. \nGuerra, Carlos A., Miguel Berdugo, David J. Eldridge, Nico Eisenhauer, Brajesh K. Singh, Haiying Cui, \nSebastian Abades, et al. 2022. “Global Hotspots for Soil Nature Conservation.” Nature 610 \n(7933): 693–98. \nHaesen, Stef, Jonas J. Lembrechts, Pieter De Frenne, Jonathan Lenoir, Juha Aalto, Michael B. \nAshcroft, Martin Kopecký, et al. 2021. “ForestTemp - Sub-Canopy Microclimate Temperatures of \nEuropean Forests.” Global Change Biology 27 (23): 6307–19. \nKöninger, Julia, Cristiano Ballabio, Panos Panagos, Arwyn Jones, Marc W. Schmid, Alberto Orgiazzi, \nand Maria J. I. Briones. 2023. “Ecosystem Type Drives Soil Eukaryotic Diversity and Composition \nin Europe.” Global Change Biology 29 (19): 5706–19. \nLabouyrie, Maëva, Cristiano Ballabio, Ferran Romero, Panos Panagos, Arwyn Jones, Marc W. Schmid, \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 32 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nVladimir Mikryukov, et al. 2023. “Patterns in Soil Microbial Diversity across Europe.” Nature \nCommunications 14 (1): 3311. \nLouca, Stilianos, Laura Wegener Parfrey, and Michael Doebeli. 2016. “Decoupling Function and \nTaxonomy in the Global Ocean Microbiome.” Science 353 (6305): 1272–77. \nOrgiazzi, A., C. Ballabio, P. Panagos, A. Jones, and O. Fernández-Ugalde. 2018. “LUCAS Soil, the \nLargest Expandable Soil Dataset for Europe: A Review.” European Journal of Soil Science 69 (1): \n140–53. \nOrgiazzi, Alberto, Panos Panagos, Oihane Fernández-Ugalde, Piotr Wojda, Maëva Labouyrie, Cristiano \nBallabio, Antonio Franco, Alberto Pistocchi, Luca Montanarella, and Arwyn Jones. 2022. “LUCAS \nSoil Biodiversity and LUCAS Soil Pesticides, New Tools for Research and Policy Development.” \nEuropean Journal of Soil Science 73 (5). https://doi.org/10.1111/ejss.13299. \nPe’er, Guy, Yves Zinngrebe, Francisco Moreira, Clélia Sirami, Stefan Schindler, Robert Müller, \nVasileios Bontzorlos, et al. 2019. “A Greener Path for the EU Common Agricultural Policy.” \nScience 365 (6452): 449–51. \nPereira, H. M., S. Ferrier, M. Walters, N. L. Geller, R. H. G. Jongman, J. Scholes, M. W. Bruford, et al. \n2013. “Essential Biodiversity Variables.” Science 339 (January): 277–78. \nPhillips, Helen, Eric K. Cameron, Nico Eisenhauer, Victoria Burton, Olga Ferlian, Yiming Jin, Sahana \nKanabar, et al. 2023. “Global Change and Their Environmental Stressors Have a Significant \nImpact on Soil Biodiversity -- a Meta-Analysis.” Authorea Preprints, February. \nhttps://doi.org/10.22541/au.167655684.49855023/v1. \nPhillips, Helen R. P., Léa Beaumelle, Katharine Tyndall, Victoria J. Burton, Erin K. Cameron, Nico \nEisenhauer, and Olga Ferlian. 2019. “The Effects of Global Change on Soil Faunal Communities: A \nMeta-Analytic Approach.” Riogrande Odontologico 5 (July): e36427. \nPõlme, Sergei, Kessy Abarenkov, R. Henrik Nilsson, Björn D. Lindahl, Karina Engelbrecht Clemmensen, \nHavard Kauserud, Nhu Nguyen, et al. 2020. “FungalTraits: A User-Friendly Traits Database of \nFungi and Fungus-like Stramenopiles.” Fungal Diversity 105 (1): 1–16. \nProença, Vânia, Laura Jane Martin, Henrique Miguel Pereira, Miguel Fernandez, Louise McRae, Jayne \nBelnap, Monika Böhm, et al. 2017. “Global Biodiversity Monitoring: From Data Sources to \nEssential Biodiversity Variables.” Biological Conservation 213 (September): 256–63. \nRömbke, Jörg, Anton M. Breure, Christian Mulder, and Michiel Rutgers. 2005. “Legislation and \nEcological Quality Assessment of Soil: Implementation of Ecological Indication Systems in \nEurope.” Ecotoxicology and Environmental Safety 62 (2): 201–10. \nRuosteenoja, Kimmo, Tiina Markkanen, Ari Venäläinen, Petri Räisänen, and Heli Peltola. 2018. \n“Seasonal Soil Moisture and Drought Occurrence in Europe in CMIP5 Projections for the 21st \nCentury.” Climate Dynamics 50 (3): 1177–92. \nSieriebriennikov, Bogdan, Howard Ferris, and Ron G. M. de Goede. 2014. “NINJA: An Automated \nCalculation System for Nematode-Based Biological Monitoring.” European Journal of Soil Biology \n61 (March): 90–93. \nSmith, Linnea C., Alberto Orgiazzi, Nico Eisenhauer, Simone Cesarz, Alfred Lochner, Arwyn Jones, \nFelipe Bastida, et al. 2021. “Large‐scale Drivers of Relationships between Soil Microbial \nProperties and Organic Carbon across Europe.” Global Ecology and Biogeography: A Journal of \nMacroecology 30 (10): 2070–83. \nThuiller, Wilfried, Damien Georges, Robin Engler, Frank Breiner, Maintainer Damien Georges, and \nContact Wilfried Thuiller. 2016. “Package ‘biomod2.’” Species Distribution Modeling within an \nEnsemble Forecasting Framework. \nftp://137.208.57.37/pub/R/web/packages/biomod2/biomod2.pdf. \nZeiss, Romy, Maria J. I. Briones, Jérome Mathieu, Angela Lomba, Jessica Dahlke, Laura-Fiona Heptner, \nGabriel Salako, Nico Eisenhauer, and Carlos A. Guerra. 2023. “Climate Effects on the Distribution \nand Conservation of Commonly Observed European Earthworms.” Conservation Biology: The \nJournal of the Society for Conservation Biology, September. https://doi.org/10.1111/cobi.14187. \nZeiss, Romy, Nico Eisenhauer, Alberto Orgiazzi, Matthias Rillig, François Buscot, Arwyn Jones, Anika \nLehmann, Thomas Reitz, Linnea Smith, and Carlos A. Guerra. 2022. “Challenges of and \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926\n\n europabon.org 33 | Page Dx.y Name \n This project receives funding from the European Union’s Horizon \n 2020 research and innovation programme under grant agreement \n No 101003553. \nOpportunities for Protecting European Soil Biodiversity.” Conservation Biology: The Journal of \nthe Society for Conservation Biology, May, e13930. \n \nAuthor-formatted document posted on 06/06/2024. DOI:  https://doi.org/10.3897/arphapreprints.e128926","source_license":"CC-BY-4.0","license_restricted":false}