An open-source metadataset of running European mid-and long-term agricultural field experiments

Mid-term (MTEs, 5–20 years) and long-term (LTEs, 20 + years) field experiments are key sources of information to design future climate-smart agriculture. Within the European Joint Program SOIL (EJP SOIL), we built the EJP SOIL-MTE/ LTE metadataset that contains metadata from 240 MTEs/LTEs across Europe. Metadata collected included precise descriptions of the treatments (combination of factors such as tillage, crop type/rotation, amendments/fertilizers, grazing and pest/weed management), soil and crop measurements and pedo-climatic information. Using different figures and dashboards, an overview of those MTEs/LTEs is presented and specific research themes (tillage systems, residue management, amendment type and cover crops) are further analysed within their pedo-climatic context. An interactive web portal developed in collaboration with the BonaRes project (https:// lte. bonar es. de), enables users to explore the metadataset and find relevant MTEs/LTEs for specific combinations of practices (e.g. all MTEs/LTEs that investigate cover crops on a Cambisol in no-tillage system). Finally, a SWOT (Strengths, Weaknesses, Opportunities, Threats) analysis of the metadataset was carried out to highlight the potential contribution of MTEs/LTEs to a harmonized European soil observation and monitoring approach. We propose that the metadataset could be elaborated with metadata from other existing MTEs/LTEs in Europe or even worldwide.

provision, prevention of soil degradation, restoration of land and maintenance of biodiversity more effectively.As Rasmussen et al. (1998) explained, sustainable agroecosystems are not a luxury.Within this context, long-term field experiments are crucial sources of knowledge on agricultural soil management (Cochran, 1939;Haddaway et al., 2015;Sandén et al., 2018).They are vitally important in monitoring, understanding and modelling the changes in soil properties and crop production occurring as a result of different agricultural management practices under the pressure of population and economic growth, and climate change (Montanarella et al., 2016;Smith, 2013).Long-term field experiments are also essential for future research and cannot be replaced by new analytical techniques or models (e.g.Donmez et al., 2023;Johnston & Poulton, 2018).On the contrary, they are an indispensable basis for the calibration and validation of these techniques.They also enable the monitoring of slowly changing parameters such as soil organic carbon (Baveye et al., 2016;Kätterer & Andrén, 1999).However, because of longevity, it is costly to maintain them.Therefore, efforts aiming at more widespread and cooperative use of long-term field experiments and their corresponding (meta)data should be encouraged.National initiatives, such as the German BonaRes project (https:// bonar es.de) (Grosse et al., 2020) or international networks such as the GLTEN (https:// glten.org), clearly demonstrate the value of an open access to long-term field experiments information.
Previous studies already resulted in valuable opensource data sets of long-term field experiments such as the qualitative overview map of Haddaway et al. (2015) or the quantitative approach of Sandén et al. (2018) that also incorporated pedo-climatic variables.Building on these studies, this initiative aimed at collecting metadata of running European field experiments both long-term (20+ years, LTE) and mid-term (5-20 years, MTE) at the factorial level using controlled vocabulary.At the time of writing, no experimental data were stored in the metadataset (however, some MTEs/LTEs provide data through the BonaRes portal, Specka et al., 2019) as, often, data embargoes are encountered and requiring data for submission can discourage contributors.The novelty of this work lies in the combination of the detailed metadata provided at the factorial level and the controlled vocabulary; both enabled an automated approach to check the entries and analyse the results.This ensured the reproducibility of the entire workflow and its future up-scaling.Overall, the findability accessibility, interoperability and reusability (FAIR) approach (Wilkinson et al., 2016) was followed.
While MTE/LTE information is often available at the national level such as with the BonaRes project for Germany (Grosse et al., 2021) or AnaEE (analysis and experimentation on ecosystems) for France (Mougin et al., 2015) or via international networks such as GLTEN (global long-term experiment network) and ILTER (international long-term ecological research, Mirtl et al., 2018), the information is scattered, heterogeneous in format, structure, used standards and (meta)data organization which makes comparisons between MTEs/LTEs difficult.A user-friendly web portal, using modern search queries and categories is a good example of an interactive interface that helps users to quickly find relevant MTEs/LTEs for specific management practices and pedo-climatic conditions, demonstrating the 'Findable' of the FAIR acronym.The open-access of the metadata (CC-BY for the published metadataset in the repository; CC0 for all metadata in the interactive map) collected is also essential as it enables users to reuse the collected data for future applications, as specified by the 'R' of 'Reusable/Reproducible' of the FAIR acronym (Wilkinson et al., 2016).The objectives of this research are: 1. To construct a living metadataset of European running agricultural field experiments that can be interactively explored.2. To provide an overview of the agricultural management practices investigated by the MTEs/LTEs within different pedo-climatic zones.3. To identify knowledge gaps within the current MTE/ LTE research.

| Selection criteria
The EJP SOIL programme defined several expected impacts such as 'fostering understanding of soil management and its influence on climate mitigation and adaptation, sustainable agricultural production and environment' and 'Understanding how soil carbon sequestration can contribute to climate change mitigation'.Most European running MTEs/LTEs that could provide information on these expected impacts were collected by the partners within the EJP SOIL project which includes 24 European countries.In order to be included in our metadataset, a set of criteria for inclusion of field experiments was provided.These quality criteria were checked and discussed by the partners of EJP SOIL and are described below.
Criteria: 1. Soil management (including grazing): the field experiment should focus on soil management.At least one soil management practice (i.e.fertilizer application, tillage, crop succession, etc.) is implemented and monitored over time.
2. Duration: the experiment should still be running and started at least 5 years ago.This is in contrast to the 20 years limit set by other initiatives such as GLTEN or BonaRes.This limit was decreased to encourage the owners of recently started experiments to share their metadata, hence increasing potential collaboration with other interested parties.In this work, both long-term (20+ years) field experiments and mid-term (5-20 years) field experiments (MTE) are included.Another reason to include MTEs along LTEs is that these MTEs (potential future LTEs) enable us to better observe the recent research trends.While this limit was reduced to 5 years, the long-term commitment of maintaining the experiments was assessed by the EJP SOIL partners.
3. Statistical soundness: the set-up of the MTE/LTE should allow a reliable statistical analysis of the data.For example, treatments have a number of replications (min 3), adequate plot size, adapted to the practice investigated.The filtering was done by the National Coordinators and we do not possess information about how many experiments were excluded by these criteria.
4. Recorded management history: information on the agricultural practices such as fertilizer application (e.g.type, application dose), tillage (e.g.type, depth), crop rotation, etc. that are studied within the LTE should be recorded over time.
5. Soil sampling at the start of the experiments/presence of a control treatment: the soil is thoroughly sampled at the start of the experiment or a control treatment is present.That way, a change in soil parameters can be studied over time.
6. Regular monitoring of crop/soil parameters (e.g.yield after each season): in a next step, the contact details of the owners of the MTEs/LTEs that complied with the selection criteria were listed after which, a template was sent to the owners to collect metadata of their MTEs/ LTEs.

| Template design
To collect metadata of the selected MTEs/LTEs, an excel template was developed.This template was built on the experience acquired from previous projects (e.g.Catch-C, Sandén et al., 2018).The terms used in the template are based upon the keywords tree of the Knowledge Library of the BonaRes project (https:// klibr ary.bonar es.de), controlled vocabulary from FAO AGROVOC thesaurus (http:// www.fao.org/ agrov oc/ ) and Corine Land Cover.As much as possible, a controlled vocabulary was enforced by the use of drop-down menus while keeping the flexibility for the user to add new terms if needed.Data validation was also used when numerical information was required.
These measures ensured a more homogeneous data set when all templates were merged together.Similar to the final data set, the template was divided into multiple tables (equivalent to sheets or tabs).The use of indices enabled the link between the tables to be defined, transforming the template into a relational metadataset.The file also contained instructions on how to fill the template ('README' table) and definitions of the vocabulary used ('description' table ).The 'experiment' table contained information about the experiment (location, exposition) as well as contact details of the MTE/LTE owner and/or operators.In this table, one row corresponds to one experiment.The 'soil-type' describes the inherent soil properties at the experiment site with information about one soil layer per row.This includes WRB soil group and USDA soil textural classes.The 'treatment' table describes the different treatments, taken as a unique combination of factor levels, one treatment per row.The following tables contain information relative to different categories of management practices: crops, amendment, irrigation, grazing and pest/weed management.Finally, the last table 'soil-crop-measurement' contains metadata about the soil and crop properties that were measured during the course of the experiment.The structure, the vocabulary and the definitions used were reviewed several times by the national representatives before finalizing and producing a released version of the template.Figure 1 illustrates the structure of the template.
Compared to templates with an all-encompassing table, the relational structure enables one-to-many relationships to be easily collected in a structured way, without creating excessive replications.For instance, if one experiment investigates the use of intercropping (e.g.growing peas and maize at the same time) versus monocropping (growing a single crop for a season), the intercropping treatment will be linked to several crops in the 'crop' table while the monocropping treatment will only be linked to one.In an all-encompassing table, this information will typically be coded on duplicated rows (Figure 2). Figure 2 illustrates the different options described above.

| Metadata collection
The template was implemented in a Google Sheet as this option enabled autocompletion for the dropdown menus.
Copies of the template were automatically made using Google Drive API and individual template links were distributed among the partners of EJP SOIL (a total of 177 templates with information of 260 experiments were collected); several experiments could be entered in one template.In a final step, each national representative delivered the template link to the corresponding MTE/LTE owners in her/his country and kept track of the response rate.Once the templates were completed, a checking script downloaded the templates and performed the automatic checks.These included examination of the type of the values (if a value was inside the drop-down list or if it was a new choice added by the user), and consistency as well as uniqueness of the index across the multiple tables.All templates that passed the automatic check (172 templates) were merged together to form the final EJP SOIL-MTE/LTE metadataset.In total, metadata from 240 different MTEs/LTEs (80 MTEs, 160 LTEs) from 21 countries across Europe were collected, resulting in 1654 unique factor level combinations (referred to as 'treatments' in the rest of the article).
Furthermore, based on the geographical coordinates of each MTE/LTE, climatic information was added from the BioClim dataset with resolution of ca. 1 km 2 (Fick  & Hijmans, 2017) and the 'Global Aridity Index and Potential Evapotranspiration Climate Database v2' dataset (Trabucco & Zomer, 2018).Each MTE/LTE was also assigned to one of the different European biogeographical regions (Habitat Directive 92/43/EEC).

| Metadataset exploration
The first objective of this study was to provide an interactive access to the EJP SOIL-MTE/LTE metadataset and to allow users to search and find MTEs/LTEs based on their metadata.To fulfil this objective, collaboration with the BonaRes project (https:// www.bonar es.de/ ) was established.Indeed, the BonaRes project already conducted such an interactive portal for their collection of German LTEs (https:// lte.bonar es.de/ ).For this purpose, the EJP SOIL-MTE/LTE metadataset was exported to a specific format provided by the BonaRes infrastructure.More information on the metadataset from BonaRes can be found in Donmez et al. (2022).
To provide an overview of the current MTEs/LTEs across Europe and to identify potential knowledge gaps that could be addressed by future field experiments (Objectives 2 and 3), specific research themes (e.g.tillage, amendment) were identified automatically using an algorithm, based on the information provided in the metadataset.This automatic extraction of research themes is only possible because of detailed information and the structure of the EJP SOIL-MTE/LTE metadataset.

| Overview
Figure 3 shows the distribution of MTEs/LTEs across Europe with a high density of collected experiments in Central Europe.
The main pedo-climatic variables and management practices are summarized in Figure 4.Note that a MTE/ LTE can be counted several times as generally different levels of a management practice category are present within the same MTE/LTE (e.g. an MTE/LTE will be included in both 'zero tillage' and 'conventional tillage' if the MTE/ LTE compares those two management practices).For each subplot, a small circular graph indicates the percentage of MTEs/LTEs that reported information on this topic.Nine research themes (amendment [mineral, organic and liming], tillage system, crop rotation, residue incorporation, cropping system, cover crops, farming system, pest control method and grazing) and the MTEs/LTEs that investigated one or more of these themes were identified.Figure 6 shows the histogram of the starting years of the MTEs/LTEs that were included in the metadataset per research theme as well as the number of MTEs/LTEs for each theme.
It is not uncommon that several research themes are investigated within one MTE/LTE; Figure 7 shows the number of research themes that were investigated within a single MTE/LTE with respect to its starting year.Before 1985, most LTEs investigated one or two research themes while more recent MTEs/LTEs investigated a combination of agricultural practices (up to six research themes for some).For instance, tillage is often investigated together with residue incorporation (nine MTEs/LTEs), different amendments (four MTEs/ LTEs) or crop rotations (three MTEs/LTEs).

| Specific research themes
LTEs were designed to answer specific research questions about the implementation of agricultural management practices.In the following figures, information about specific agricultural management practices is presented in a dashboard including the main practices investigated by MTEs/LTEs for specific research themes: tillage system (Figure 8), type of amendment (Figure 9), residue management (Figure 10), cropping system (Figure 11), cover crops (Figure 12) and crop rotation (Figure 13).The left part of the dashboard consists of an intersection figure, a so-called UpSet plot (Lex et al., 2014), that enables us to visualize the interactions between the different levels of the factor.The frequency of the levels is shown in the horizontal bar plot while the frequency of the intersections is shown in the vertical bar plot above.Additional bar charts are depicted in each dashboard to provide an overview of the pedo-climatic context of those MTEs/LTEs and to make a comparison with the entire F I G U R E 4 Overview of the selected categorical and numerical parameters and their distribution in terms of number of long-term experiments.To increase readability, levels with a small number of entries were grouped into the category 'Others'.The circle graph on the right-hand side of each graph denotes the data availability for this category.For instance, '64%' indicates that 64% of the MTEs/LTEs provided information on this topic.Orange plots are only relevant for arable land, green for pastures and black for both.EJP SOIL-MTE/LTE metadataset.Despite our efforts, not all metadata were available for all MTEs/LTEs and it is good to keep this in mind when interpreting the following dashboards.

| Amendments
This research theme includes all MTEs/LTEs that compare different types of amendments: mineral fertilizers, organic fertilizers, liming materials, plant biostimulants or no amendment.The dashboard of MTEs/LTEs investigating amendments is shown in Figure 9.This topic has been extensively researched in the past with a large number of experiments well spread over Europe and covering a wide range of climatic and pedological conditions.Most MTEs/LTEs compared the use of mineral, organic and no-fertilizer amendment while fewer investigated the effect of liming materials.

| Residue removal
Figure 10 shows the characteristics of the MTEs/LTEs investigating the effect of residue removal in Europe.It can be observed that most MTEs/LTEs compare full versus no residue removal.The distributions of pedoclimatic variables usually followed the distributions of the entire data set except for soil organic carbon (SOC).

| Cropping systems
Figure 11 shows that the MTEs/LTEs investigating different cropping systems such as monocropping versus intercropping or monocropping versus pastures are mainly located in the Northern part of Europe.Given the lower number of MTEs/LTEs investigating this topic, variance in pedo-climatic information is sparse.This lower number can also be explained by the fact that MTEs/LTEs usually tend to focus on one cropping system (e.g.only on pasture, monocropping or intercropping) and that fewer MTEs/LTEs compare the entire systems.This can also be observed in the relatively small number of MTEs/LTEs comparing different farming systems (Figure 6-evolution of research themes).

| Cover crops
According to the information collected in this study, the use of cover crops is a relatively new research theme investigated in MTEs/LTEs (see Figure 6 evolution).MTEs/ LTEs shown in the dashboard investigate the use of cover crops versus no cover crops (Figure 12).Radish, rye, legumes and ryegrass are among the most commonly used cover crops.MTEs/LTEs on cover crops are mainly concentrated in central Europe with no MTEs/LTEs in the Mediterranean or boreal regions.

| Crop rotation
Information about MTEs/LTEs investigating crop rotation versus no rotation are shown in Figure 13.As observed in Figure 6, the LTEs that were initiated before 1950 mostly focused on amendments.Since 1985, LTEs also included treatments on tillage, crop rotation and residue management.After 2000, an increase in MTEs/LTEs that investigated the use of cover crops and compared different cropping systems was noticed (e.g.Peigné et al., 2016).This really shows a shift in the research paradigm to more integrated agricultural methods.In addition, Figure 7 shows that there is a clear trend at investigating not only a specific practice but rather a combination of practices and the possible interactions.This more holistic approach tends to confirm the shift in paradigm towards a more ecological management of our agroecosystems (e.g.Gargano et al., 2021).
These trends are also observed in the dashboards where tillage systems (Figure 8) and amendments (Figure 9) research themes have a large number of experiments on diverse pedo-climatic conditions.On the contrary, MTEs/ LTEs investigating residues management (Figure 10) are in the lower range of the SOC distribution compared to other research themes, potentially explaining the motivation to increase it by keeping crop residues on the field (Lehtinen et al., 2014).Similarly, the distribution of research themes investigating cover crops (Figure 12) is also constrained by pedo-climatic variables.Indeed, in the Mediterranean, cover crops tend to compete for water with main crops (Gómez, 2017); which is less of an issue in the continental climatic zone.From Figure 13, the distribution of the MTEs/LTEs across climatic and soil properties demonstrates that crop rotation or the absence of rotation is of interest independently of the pedo-climatic context.
Other research themes were also reported in the EJP SOIL-MTE/LTE metadataset: Four MTEs/LTEs were reported to investigate the effect of grazing and three MTEs/ LTEs the use of different types of pest/weed management.Other research themes such as the use of agroforestry or the investigation of soil compaction were recorded with difficulty in our metadataset as the initial structure of the template and the choices of the drop-down menus did not allow characterization of the treatments of these MTEs/ LTEs using a controlled vocabulary.However, based on the verbose description of the treatments by the MTE/ LTE owners, these experiments can still be identified manually.Additionally, research themes investigating different occurrences of a practice (e.g. one main crop vs multiple cropping in the same growing year or yearly tillage vs occasional tillage) could not be visualized in these dashboards while well detected in the analysis.
In the next sections, a SWOT analysis aims to identify the Strengths, Weaknesses, Opportunities and Threats of the EJP SOIL-MTE/LTE metadataset.

| Strengths
A certain strength of the EJP SOIL-MTE/LTE metadataset is to provide a European overview of the metadata of MTEs/LTEs in an open-access data set.User-friendly access to the information using the BonaRes web portal enables users to easily find relevant MTEs/LTEs and additional related detailed information.The overview of different research themes through the dashboards (Figures 8-13) also enables us to identify knowledge gaps where few MTEs/LTEs were available (e.g.MTEs/LTEs investigating grazing) but also pedo-climatic zones less investigated (e.g. for coarser soil texture).The EJP SOIL-MTE/LTE metadataset is also general enough to include different practices and provide a granular level of details to describe each combination of factor levels using a controlled vocabulary.The automatic quality check ensures new templates are filled out properly.From its opensource nature and its flexibility, it provides a reusable template to describe metadata of agricultural experiments.Its relational structure also enables customized queries spanning several tables to be formulated (e.g.all MTEs/LTEs that investigate different tillage systems with cereals and organic amendments).The integration of the EJP SOIL-MTE/LTE metadataset to the existing infrastructure of BonaRes ensures the long-term accessibility and visibility of these MTEs/LTEs and demonstrates the interoperability inherent to our metadataset structure.Finally, the fact that a diverse network of partners and institutions across Europe characterize MTEs/LTEs, guarantees a multidisciplinary data acquisition, analysis and integration and a cost-effective soil monitoring.

| Weaknesses
It is important to note that while every effort has been made to ensure the completeness of this metadataset, it is not an exhaustive list of all possible running MTEs/ LTEs across Europe.This research has largely relied on the contributions of the EJP SOIL partners to complete the questionnaire.Not all questionnaires were completed and while the literature was reviewed to add additional sites, there may be data gaps within the results.The information collected for MTEs/LTEs is heterogeneous.Most MTEs/LTEs are fully described with the full crop rotation and management practices details included while other MTEs/LTEs only contain location and brief definitions of the treatments.These missing data on pedo-climatic variables (e.g.soil types) impede the representativity of the distribution of observations made for each research theme.Variables to be measured, methodologies, technological development and sampling schemes are not homogeneous among sites.While the use of controlled vocabulary certainly ensures a more homogeneous metadataset, it can also be a limitation for missing metadata fields specific to certain research themes (e.g.agroforestry with distance from the field to tree lines, or soil compaction).

| Opportunities
Electing a subset of MTEs/LTEs based on a specific treatment applied (e.g.zero-tillage) and obtaining the related data would enable us to derive pedo-climatic-specific conclusions on the effect of specific management practices and possibly translate these into EU-wide contextspecific recommendations.In addition, the long time span of MTEs/LTEs enables the identification of European trends (warming, drought) and local pressure (nutrient leaching, erosion risk) across pedo-climatic conditions.This information can later be used in space for time modelling where a crop model from, for instance, the North of France is fitted with data from the South of France to simulate future climatic effects.Thanks to the EJP SOIL-MTE/LTE metadataset format, one could incorporate it into other existing networks of long-term experiments such as AnaEE (Mougin et al., 2015), GLTEN or ILTER (Mirtl et al., 2018).The addition of terminated LTEs in the database, while not an objective of the current project, would be an additional important source of knowledge.As most MTEs/LTEs information can be extracted from scientific publications, a systematic mapping (e.g.Haddaway et al., 2015) could be envisaged, potentially aided by natural language processing (Blanchy et al., 2023;Padarian et al., 2020) that can help automatically fill some metadata of the template or at least identify new potential MTEs/ LTEs.In addition, MTEs/LTEs of the metadataset could serve as a knowledge hub and may well be integrated into Lighthouse farms to share the knowledge gained over decades.As EU strategic targets encompass a suite of soil and land management targets, this research can leverage the collaborative potential of existing research and the targeting of future research to support a transition towards meeting targets, such as those set out under the European Green Deal or the EU Soil Strategy.

| Threats
Maintenance of such a data set at the European level is a major threat and one of the reasons the collaboration with BonaRes was established, to ensure the longevity of the project beyond the funding period of EJP SOIL.It is planned that future contributors will be able to add new MTEs/LTEs to the data set beyond the end of the project by filling up a dedicated template or using a web form.Nevertheless, maintaining the accuracy of the information of already entered entries will remain challenging as MTEs/LTEs that are running now are never guaranteed to run 10 years from now.Regular update requests to MTE/ LTE owners can be a solution but requires manpower and responsiveness is not guaranteed.Figure 14 summarizes the main point discussed in the SWOT analysis.

| CONCLUSION
In this work, we describe the construction and analysis of the EJP SOIL-MTE/LTE metadataset, an open-source European wide relational metadataset containing information about 240 MTEs/LTEs.Data quality was ensured using a well-defined format, making use of controlled vocabulary and automatic checking scripts.Despite variable data availability within the metadataset, frequent management practices and pedo-climatic conditions were identified.In addition, selected MTEs/LTEs targeting specific research questions about the use of different tillage systems, cover crops, mineral/organic amendments or residue management were explored in detail using dashboards.These dashboards enabled us to detect which treatments were mostly compared against each other and in which pedo-climatic environment these MTEs/LTEs mainly took place.It is envisaged that MTE/LTE-sourced information around a management practice could lead to pedo-climatic specific recommendations at the EU level.While the maintenance of such a data set remains challenging, our collaboration with the BonaRes project aims to ensure its longevity and access beyond the funding period of EJP SOIL.The use of natural language processing techniques could also support the metadata extraction process from scientific articles and hence enhance maintainability.The open-source nature of the EJP SOIL-MTE/ LTE metadataset and usage of the FAIR principles makes it applicable and expandable for future climate smart agricultural projects.

F
Structure of the template used for metadata collection.The template consists of multiple tables (sheets) are connected by using ID's such as 'Experiment ID' and 'Treatment ID' (in bold).

F
I G U R E 2 Collecting metadata of an agricultural field experiment using (a) an all-encompassing table with multiple crops per row, (b) an all-encompassing table, with duplicated rows for T2 or (c) multiple tables linked by indices.

From
Figure 4, we notice that most of the MTEs/LTEs are located on Cambisol (n = 94) and Luvisol (n = 65) (IUSS Working Group, 2015) with a silt loam or sandy loam texture (USDA) under continental or Atlantic climatic conditions.MTEs/LTEs on arable land were largely dominant with only a few MTEs/LTEs on pastures.Conventional farming systems, inversion tillage, monocropping and inorganic fertilizer are also reported predominantly.The main types of measurements are displayed in Figure 4p and are mainly related to chemical soil properties.MTEs/LTEs often investigate multiple factors.Figure 5 visualizes trends of treatment combinations in a Sankey graph.For instance, one can now observe the flow of treatments with 'conventional farming', 'inversion tillage', 'monocropping', 'inorganic fertilizer' and 'no irrigation' values.

F
Distribution of the collected mid-and long-term field experiments across Europe with European biogeographical regions.

F
Sankey graph of the major combination of agricultural practices in the EJP SOIL-MTE/LTE metadataset.When information was missing, 'Unknown X' was used.F I G U R E 6 Evolution of newly started MTEs/LTEs with respect to their research themes.Note that one MTE/LTE can have several research themes (e.g. it can investigate both tillage and cover crops) and hence can be counted multiple times.The research theme about 'amendments' includes experiments that investigate mineral/organic or no fertilizer (67 MTEs/LTEs) and 10 MTEs/LTEs investigating liming.F I G U R E 7 Number of research themes investigated per MTE/LTE sorted by starting year.

F
I G U R E 8 Dashboard of mid-and long-term field experiments investigating different tillage practices.Subplot (a) shows the map of the experiments.The horizontal bar chart (c) shows how many experiments investigated this particular practice while the vertical bar chart (b) shows how many experiments investigated the comparison of different tillage practices (d).The subplots (e)-(j) show the distribution of metadata for the selected experiments (black) and for the entire data set (grey).The numbers in parentheses show for how many MTEs/ LTEs the information was available.F I G U R E 9 Dashboard of mid-and long-term field experiments investigating different types of amendments.Subplot (a) shows the map of the experiments.The horizontal bar chart (c) shows how many experiments investigated this type of amendment while the vertical bar chart (b) shows how many experiments investigated the comparison of different amendments (d).The subplots (e)-(j) show the distribution of metadata for the selected experiments (black) and for the entire data set (grey).The numbers in parentheses show for how many MTEs/ LTEs the information was available.MTEs/LTEs mainly compare inversion versus non-inversion tillage and zero tillage is often added to these two.Tillage systems have been well investigated across Europe and within different pedo-climatic zones.The distribution of soil properties (bulk density, texture, soil organic carbon or soil types) of these MTEs/LTEs is often similar to the general distributions observed in the entire metadataset.

F
I G U R E 1 0 Dashboard of mid-and long-term field experiments investigating different types of residue management.Subplot (a) shows the map of the experiments.The horizontal bar chart (c) shows how many experiments investigated this type of residue management while the vertical bar chart (b) shows how many experiments compared different types of residue management (d).The subplots (e)-(j) show the distribution of metadata for the selected experiments (black) and for the entire data set (grey).The numbers in parentheses show for how many MTEs/LTEs the information was available.F I G U R E 1 1 Dashboard of mid-and long-term field experiments investigating different cropping systems.Subplot (a) shows the map of the experiments.The horizontal bar chart (c) shows how many experiments investigated this type of farming system while the vertical bar chart (b) shows how many experiments compared different farming systems (d).The subplots (e)-(j) show the distribution of metadata for the selected experiments (black) and for the entire data set (grey).The numbers in parentheses show for how many MTEs/LTEs the information was available.

F
I G U R E 1 2 Dashboard of mid-and long-term field experiment investigating the presence or absence of cover crops.Subplot (a) shows the map of the experiments.The horizontal bar chart (b) shows the number of MTEs/LTEs that investigate specific types of cover crops.The subplots (c)-(h) show the distribution of metadata for the selected experiments (black) and for the entire data set (grey).The numbers in parentheses show for how many MTEs/LTEs the information was available.F I G U R E 1 3 Dashboard of mid-and long-term field experiments investigating the effect of crop rotation.Subplot (a) shows the map of the experiments.The most common crops for no rotation versus crop rotation experiments are shown in subplot (b).The subplots (c)-(h) show the distribution of metadata for the selected experiments (black) and for the entire data set (grey).The numbers in parentheses show for how many MTEs/LTEs the information was available.

F
Summary of the SWOT (Strengths, Weaknesses, Opportunities, Threats) analysis.