The LANDSUPPORT geospatial decision support system (S‐DSS) vision: Operational tools to implement sustainability policies in land planning and management

Nowadays, there is contrasting evidence between the ongoing continuing and widespread environmental degradation and the many means to implement environmental sustainability actions starting from good policies (e.g. EU New Green Deal, CAP), powerful technologies (e.g. new satellites, drones, IoT sensors), large databases and large stakeholder engagement (e.g. EIP‐AGRI, living labs). Here, we argue that to tackle the above contrasting issues dealing with land degradation, it is very much required to develop and use friendly and freely available web‐based operational tools to support both the implementation of environmental and agriculture policies and enable to take positive environmental sustainability actions by all stakeholders. Our solution is the S‐DSS LANDSUPPORT platform, consisting of a free web‐based smart Geospatial CyberInfrastructure containing 15 macro‐tools (and more than 100 elementary tools), co‐designed with different types of stakeholders and their different needs, dealing with sustainability in agriculture, forestry and spatial planning. LANDSUPPORT condenses many features into one system, the main ones of which were (i) Web‐GIS facilities, connection with (ii) satellite data, (iii) Earth Critical Zone data and (iv) climate datasets including climate change and weather forecast data, (v) data cube technology enabling us to read/write when dealing with very large datasets (e.g. daily climatic data obtained in real time for any region in Europe), (vi) a large set of static and dynamic modelling engines (e.g. crop growth, water balance, rural integrity, etc.) allowing uncertainty analysis and what if modelling and (vii) HPC (both CPU and GPU) to run simulation modelling ‘on‐the‐fly’ in real time. Two case studies (a third case is reported in the Supplementary materials ), with their results and stats, covering different regions and spatial extents and using three distinct operational tools all connected to lower land degradation processes (Crop growth, Machine Learning Forest Simulator and GeOC), are featured in this paper to highlight the platform's functioning. Landsupport is used by a large community of stakeholders and will remain operational, open and free long after the project ends. This position is rooted in the evidence showing that we need to leave these tools as open as possible and engage as much as possible with a large community of users to protect soils and land.

Geospatial CyberInfrastructure containing 15 macro-tools (and more than 100 elementary tools), co-designed with different types of stakeholders and their different needs, dealing with sustainability in agriculture, forestry and spatial planning.LAND-SUPPORT condenses many features into one system, the main ones of which were (i) Web-GIS facilities, connection with (ii) satellite data, (iii) Earth Critical Zone data and (iv) climate datasets including climate change and weather forecast data, (v) data cube technology enabling us to read/write when dealing with very large datasets (e.g.daily climatic data obtained in real time for any region in Europe), (vi) a large set of static and dynamic modelling engines (e.g.crop growth, water balance, rural integrity, etc.) allowing uncertainty analysis and what if modelling and (vii) HPC (both CPU and GPU) to run simulation modelling 'on-the-fly' in real time.Two case studies (a third case is reported in the Supplementary materials), with their results and stats, covering different regions and spatial extents and using three distinct operational tools all connected to lower land degradation processes (Crop growth, Machine Learning Forest Simulator and GeOC), are featured in this paper to highlight the platform's functioning.Landsupport is used by a large community of stakeholders and will remain operational, open and free long after the project ends.This position is rooted in the evidence showing that we need to leave these tools as open as possible and engage as much as possible with a large community of users to protect soils and land.Most importantly, there is an increasing number of legislation and policy frameworks aimed at achieving a better environment and agriculture that preserves natural resources and adapts to climate change (e.g.7th EAP, FAO Agenda, 17 Sustainable Development Goal [SDGs] of 2030 UN Agenda, EU directives).Some of this legislation, along with its implementation actions, such as the Nitrate Directive (year 1990) and Water Framework Directive (year 2000), has been in place for at least 2 decades.
It is clear that policies, data and stakeholder engagement are in place, but it is also dramatically clear that land degradation is increasing and our natural resources are under increasing pressure (Gowdy, 2020, UN, 2022-SDG Report) from climate change and land degradation processes.
It is also well-known that these degradation processes in turn induce biotic stresses, including augmenting the population of insects/pests and disease, increasing weed growth, threatening pollinators and increasing drought, waterlogging, salinity/alkalinity and abrupt rainfall patterns affecting agriculture in a series of ways (Shahzad et al., 2021).So how do we tackle all the contrasting issues mentioned above?
We argue that all the measures in place are simply not enough to produce a change.There is still something missing, something that would turn the above three elements (good policy, effective data and engaged communities) into operational toolseasy to use and freely available to everyonethat would support both (i) the many good environmental and agriculture policies that still face huge problems in their full implementation (e.g.reports on policies implementation: EU COM2015/120; EU COM2013/683) and (ii) positive environmental sustainability actions on the part of stakeholders and local communities.

| Aim
Here, we aim to demonstrate that a free web-based, smart, geospatial decision support system (S-DSS), based on GeoSpatial CyberInfrastructure (GCI), could make the difference in connecting data, policy and communities engagement.This is in line with this LDD special issue which requires the 'implementation of S-DSS to address the various sustainable land uses in different sectors such as in agriculture and forestry'.
In the LANDSUPPORT S-DSS (www.landsupport.eu),more than 100 operational tools are ready to support the implementation of sustainability policies and to assist a wide range of end-users for more sustainable land planning and management in agriculture, forestry, spatial planning, environmental protection, biodiversity and ecotourism.This S-DSS brings together diverse sources of data (e.g. the EU's Copernicus satellite and not satellite data, climate change data, soil maps, etc.), a large set of models simulating reality (e.g.crop growth and pesticide leaching) and a user-friendly graphical user interface, with the aim of enabling end-users to access and work on the platform easily.
We do not aim to provide a literature review of geoSpatial DSSs here, but we do wish to highlight the fact that, in recent years, there has been great progress towards the development of operational S-DSS tools using similar GCI infrastructure applied to for example: 1. Agriculture and forestry applications: including irrigation (Bonfante et al., 2019;Zhao et al., 2022), olive growing (Manna et al., 2020), viticulture (Terribile et al., 2017), forest planning and management (Marano et al., 2019;Povak et al., 2020) and crop planning and management (Kim & Kisekka, 2021).

3.
A large set of other multidisciplinary studies (S.Wang et al., 2019).
The specific development illustrated in this paper was carried out as a part of the EU-funded LANDSUPPORT project, which brought together 19 partners from 10 different countries across Europe, the Middle East and Asia.
Following similar approaches of many DSS-based papers, we have avoided the usual separation in this specific contribution's 'Materials & Methods, Results, Discussion' sections.This choice aims to enable easy reading sincedue to the complexity and interconnection of platform implementation, modelling development and user requirementseach section contains some elements overlapping (these may be methods and results) of a specific step which are a prerequisite to the next step.Thussequentiallywe treated: 1. Implementing the aim: key issues 2. The Landsupport platform.This includes (1) the requirements (including conceptual needs, required content of each tool and need to optimise tools), (2) the architecture of the system (IT infrastructure, dashboard, data and model) to implement the tools and (3) test of tools by end users.
3. Three case studies (one is given in the Supplementary materials), with their results and stats, highlight the platform's functioning.

| Implementing the aim: Challenging key issues
To produce operational tools which successfully support sustainability in agriculture, environmental protection and their connected policies, a necessary prerequisite is to acknowledge, analyse and solve the following key issues, often ignored, five specific issues: 1.The embedded, often overlooked, high physical and socioeconomic multifaceted complexity of the landscape This requires having a system that, for any type of landscape, addresses the following: • The physical, socio-economic, cultural variables and, thus, the trans-disciplinarity context of the landscape.
• Spatial variability and spatial uncertainty of influential geospatial variables (e.g.soil and hydrology).
• Multi-functionalities: the system must deliver useful operational results by capturing the deeply diverse multifunctionality of any landscape.For instance, this requires having, within one system, outputs for agriculture, environment, spatial planning, biodiversity, cultural heritage and environmental awareness.Basically, it is fundamental to capture as many dimensions of the very same landscape as possible.
• Site-specific dynamic nature of selected key geospatial variables change continuously in time and space.For instance, if the system seeks to address dynamic processes such as crop growth and nitrate leaching, then it is self-evident that monthly climate data are of little use since rainfall varies continuously over time and space.Therefore, it is necessary to capture at least the spatial and temporal 'dailybased' climate (e.g.rainfall) variations.This also applies to many other agriculture and environmental processes.Unfortunately, many operational tools dealing with agriculture and environment issues still employ highly aggregated datasets (e.g.Himics et al., 2020).For instance, Zwetsloot et al. (2021) in identifying synergies and tradeoffs when choosing different land uses employ climate data obtained from rather coarse scale European climatic zones, and in addition they have failed to use soil data (they highlight the coarse nature of currently available soil mapping).
2. The lack of a truly integrated physical approach to many agricultural/environmental problems.Here, there are at least two issues to be considered: The lack of a true integrated Earth Critical Zone (ECZ) vision.
Even when claiming interdisciplinarity, most operational approaches give priority to just one or two of the aspects (climate, plant, soil and bedrock) instead of developing a truly integrated ECZ approach.This is very unfortunate when dealing with environmental issues since many processes (e.g.nitrate leaching) affect the entire ECZ.In addition, soil information is often acknowledged in a very simplified way (e.g.no layering, great oversimplification of processes, etc.).
The need for process-based modelling approaches: in the last decades there have been many efforts at modelling to address agriculture and environmental issues.These may be schematically separated into empirical versus process-based approaches with pros and cons.When dealing with operational tools (e.g.DSS) for agriculture and environment, empirical models are by far the most widely employed.This is the case when using simple empirical multicriteria models or overlapping (as in a typical GIS system) data layers/knowledge about environmental variables (vegetation, soil, climate, etc.) to address complex agriculture and environmental issues depending on the 'Earth Critical Zone' (e.g.primary productivity or groundwater vulnerability).Here, we claim that an adaptive approach is required since empirical models are indeed essential in many cases (e.g.erosion at landscape scale), but process-based models are much more appropriate and powerful when dealing with interlinked biophysical processes and ecosystem services, at different temporal and spatial scales.In addition, one of the many drawbacks when using empirical models of this type of process (e.g.crop growth, water balance) is the very high cost of calibration and validation when transferring these models to new areas (Manna et al., 2009).To this end, process-based models, being based on superior general physical rules, are more replicable to new areas, decreasing the effort of new calibrations and validations and, thus, giving much better value for money.
3. The lack of factual scientific support (science-based solutions) to both farmers and regional governments for achieving both a realistic and performant sustainable management of agriculture, forest lands and many other environmental issues.Indeed, after many years in which both agriculture and environmental issues have benefited from the large availability of data, sensors and supposed tools, most farmers and regional governments have acknowledged a lack of support for sustainable management and planning activities.For instance, Lundström and Lindblom (2018) highlight the fact that most DSS applied to agriculture 'have not been used appropriately in practice' (Aubert et al., 2012;Eastwood et al., 2017;Rossi et al., 2014).Important reasons for this include the fact that developers 'normally consider only technology while the farmer must consider the technology in the whole complex situation of practice'.Such a lack of factual sound scientific support also applies to the guidance Research and Innovation (R&I) offers to policy makers in the designing or increasing the effectiveness of good land management practices (e.g.best practices, restoring carbon stocks, etc.) 4. The fragmentation of current approaches, models and DSS tools.A large variety of models and DSSs has emerged over recent decades (Geertmanand & Stillwell, 2009;Amelung et al. 2020;Manna et al., 2020) but typically these have been developed to address specific problems for specific end-user groups.Thus, they are of little use for delivering an integrated approach.Figure 1 aims to show this critical issue.Here, a number of cars are depicted, each representing an example of the many currently available specialised model/DSS systems designed to achieve a specific goal (e.g.climate change impact assessment, state of land degradation, crop productivity, etc.).
It should be extremely important that each of these models/DSSs be related to the others, but, unfortunately, this is not the case (the grey roads are not interconnected).Indeed, the current scenario is very fragmented and the many available models/DSSs systems do not interact with each other.This is not surprising considering the great fragmentation of disciplines around landscape analysis.Moreover, the majority of already existing models and DSSs is typically limited within a specific scale.5.The difficulties in the implementation of the many good policies to improve both the environment and agriculture so that they better preserve natural resources and adapt to climate change (e.g.7th EAP, FAO Agenda, 17 SDG of 2030 UN Agenda, EU directives).
The evidence of difficulties in the full implementation of these policies is reported in many official documents (e.g.reports about the implementation of a water framework Dir.COM2015/120 & Nitrate Dir.COM2013/683, UN, 2022).
It is believed that these difficulties arise from the fact that the implementation of much environmental and agriculture legislation requires, as a 'must', answers which vary in space (over the landscape) and time (dynamic).Here, it is necessary to underline the fact that the cause of this complexity is often the soils, whose properties vary in space and time (i.e. after soil tillage).In Table 1, some requirements, often overlooked, are reported that apply when implementing prominent agri-environmental legislation in the EU and beyond.
Thus, this hidden embedded complexity makes things rather difficult when considering points 1, 2, 3 and 4 above.The general problem is that what is required is a full implementation of many environmental policies which often require positive actions at a very detailed local scale and over very large areas (regions, countries, EU).From our understanding, current approaches are not challenging this complexity when addressing policy implementation as they offer either a simplistic aggregated view of the problem (e.g.NUT aggregation in CAPRI model, Himics et al., 2020) or address the complexity through the plethora of Agent-Based Modelling (WoS reports over 80 papers per year in only field of agriculture), which indeed produces interesting results (e.g.Bestmap https://cordis.europa.eu/project/id/817501,Agricore https://cordis.europa.eu/project/id/816078projects), but, as yet, not enough quantitative, farm-level rooted evidence (Shang et al., 2021).In our view, their view of soil-land complexity is very simplistic (Brown et al., 2022;Ziv et al., 2020), especially with respect to policy implementation which often requires quantitative, spatiallyexplicit soil-based approaches (e.g.EU Nitrate and Water Framework Directives, new CAP).

| The need for a novel concept
Here, an attempt is made to deal with the above issues, 1, 2, 3, 4 and 5, by overcoming the fragmented approach reported in Figure 1 and developing the S-DSS approach (LANDSUPPORT GCI), described in Figure 2 as a powerful 4-wheel car which, thanks to its unique engines, is able to address (black road) various objectives simultaneously, thus, overcoming the current fragmentation of tools and land policy implementation.In this way, with very limited investment and using the same infrastructure, it is feasible to reach new important additional objectives (e.g.socio-economy evaluation, spatial planning, what if modelling, water use efficiency, etc.) on a number of spatial scales.All the above is indeed feasible if intrinsic optimisations are achieved where the marginal cost of developing each new single engine is low because each engine is used for multiple purposes.
For instance, the high cost of developing the 'soil-plantatmosphere (SPA) agro-hydrologic simulation modelling working on high-performance computing (HPC) parallel processing' is counterbalanced by the fact that this model is used for many different issues such as evaluating (i) ecosystem services, (ii) farm management, (iii) soil compaction, (iv, v) nitrate and pollutant leaching, (vi) food security, (vii) impact of climate change and (viii) spatial planning.The same may apply to other modelling engines and, thus, the outcome of this approach might represent extraordinarily good value for money.Of course, data, model and HPC resources must work in accordance with the geographic scales, enabling actions to be tackled on the local scale where the largest multi-beneficial agriculture and 2030 SDGs deliveries will be produced while still delivering on very large spatial scales.
More specifically, our system should enable us to retain the following aspects, as depicted in Figure 3: (i) standard Web-GIS features, such as easy data updates and a user-friendly graphical user interface (GUI) and (ii) new additional features to be added to standard Web-GIS to enable a GUI to deliver multiscale, multi-stakeholder and multi-functionality outputs and permit upload of thematic maps by end-users and stakeholders.The latter is the case when the user is entering a new data layer (e.g.his/her own Region Of Interest-ROI as the focus of the analysis).Eventually, the system must include (iii) a third set of features well beyond Web-GIS architectures and philosophy.These refer to the use of dynamic databases (e.g.daily update of geospatial satellite or climate data) and the demand for 'on-the-fly' simulation modelling, as required by most dynamic environmental and/or agriculture applications (e.g.primary productivity, water balance, pollutant leaching, etc.).
Moreover, for specific applications (on the scale of the farm), the system must enable the uploading of soil analysis data produced by the farmer (e.g.soil textures) to ameliorate the performance of crop growth simulation modelling.In addition, the system requires computing codes to be easily updated; this is a crucial issue to ensure system flexibility and modularity.Here, we argue that current scientific and technological advances, the excellent availability of databases and, most importantly, the vision of the responsibility of scientists in leading sustainability (Bouma, 2015(Bouma, , 2020) make it possible to move down a new road.
These advances are made possible by (i) the large availability of geospatial data (maps, satellite, drone, etc.); (ii) progress in environmental sciences and adjacent sciences, including digital soil mapping (e.g.Chen et al., 2022;Huang et al., 2022;Piikki et al., 2021) and simulation modelling of the soil-plant-atmosphere system (Coppola et al., 2019;Penuelas & Sardans, 2021; see https://soil-modeling.org/ for a deeper insight); (iii) advances in open-source Web-GIS (Tavra & Škara, 2020); (iv) high performance computing and, especially, CPU and GPU processing (Badia et al., 2022;Goodman et al., 2019);(v) recent developments in building Geospatial-DSS for agriculture and the environment built on web-based 'geospatial cyberinfrastructure', thus enabling the 'acquisition, storage, management, and integration of both advanced and dynamic data (e.g.pedological, daily climatic, and land use), data mining and data visualization, and computer "on-the-fly" applications in order to perform simulation modelling (e.g.soil-water balance and crop growth), all potentially accessible via the Web'; (vi) new understanding of the key issues of transdisciplinarity (Bouma, 2020;Daliakopoulos & Keesstra, 2020;Terribile et al., 2015) and (vii) the increasing role of T A B L E 1 Some important European Union (EU) regulations concerning the management of agricultural/forestry and environmental issues (a more detailed version of this table is provided in the Supplementary material).

SDG and EU regulation/ directive
Key sustainability topics addressed by the selected legislation Required answer to implement the policy

| The need for operational tools
Here, we seek operational tools to better implement policies and actions dealing with sustainability in agriculture, forestry and spatial planning.In Table 2, we reported the main policies we aim to intercept by use of the LANDSUPPORT GCI system and also (last column) the list of 15 macro-tools we developed from the letter 'a' to 'o' (named macro-tools because each one is composed of many elementary tools).The expected achievements and tangible products produced after the use of the cited tools to fulfil both the implementation of selected policies and actions on various scales are reported in Table 3.
From the analysis of Tables 1-3, it is self-evident that we do not seek generic tools on aggregated scales, but rather tools that enable users to act at the specific scale of implementation of each policy.
For instance, the EU Nitrate directive (Dir.91/676) is indeed pan-European, but its implementation is on a regional scale and not  More specifically, each tool was co-designed with specific communities of users, which of course varied according to the particular tool.This ensured that user requirements were taken on board.
The community of users involved in the development process were the following: 1. Policy makers at local, regional and national levels (e.g.regional administrations, ministries, etc.).
4. Researchers (e.g. in the fields of soil, agriculture, environment, urban and landscape planning, sustainable development, etc.).
5. Sectoral associations and bodies (e.g.food security agency, Chamber of Agriculture).
8. Consultants in several LANDSUPPORT-related topicsfor example, soil, agriculture, environment, urban and landscape planning, sustainable development, and so on.
To engage stakeholders and end-users, a knowledge transfer chain, depicted in Figure 4, was developed.More specifically, knowledge transfer from researchers to end-users was ensured using the LANDSUPPORT platform (depicted as mode 1 in Figure 4).While a large set of general meetings and face-to-face meetings were organised with many users to ensure knowledge transfer from end-users to researchers (mode 2 in Figure 4).This required complementary types of knowledge (scientific and practical) and needs (challenges, incentives) to be taken into account so as to transform our system into concrete opportunities for end-users.

| The need to optimise tools
To satisfy all very high expectations reported above (Figures 2-4 and Tables 2 and 3), it is required to equip our system (depicted in Figure 5 as an engine to be placed into the LANDSUPPORT 4-wheel car) with a series of special features, the main ones of which were the following: The IT architecture designed to fulfil the above needs is reported below in the IT Landsupport description.
2.2 | The architecture of the system

| The IT infrastructure
Here, we move from a conceptual basis towards the actual implementation of LANDSUPPORT taking on-board all requirements reported in Section 1 and Section 2.1.The system is built on a GCI platform supporting 'acquiring, storage capacity, management and integration of both static (e.g.hydrogeology, soil) and dynamic data (e.g.rainfall, temperature, land use) and "on-the-fly" data elaboration and visualisation' (Terribile et al., 2015;Yang et al., 2010).
The database includes both vector and raster data (Table 4).
Vector data include three main types of geometries: points, lines and The data can be processed either (or both) by static and dynamic models (see Table 5) which can generate pdf reporting, interactive maps and tables and informative Html popups as outputs.User-tool interaction takes place within the Graphical User Interface.

| The dashboard
The GUI is given in Figure 7, and it includes a variety of graphical devices and processes aiming to combine geospatial data (analysis and visualisation), production of outputs as tables and maps, and easyto-use navigation.The dashboard includes three sections: the 'data viewer', the 'map' (central box) and the 'analysis tools' (left-hand box).
The 'data viewer' section (left-hand box, Figure 7) includes the 'Layers' sheet (Figure 7a), which enables the user to navigate through the different thematic maps, the 'Legend' sheet viewer (Figure 7b) and the 'Preferences' sheet (Figure 7c), where the user can modify the type of visualisation and obtain metadata.The 'Maps' box (Central box, Figure 7d) displays the maps which have been selected from the 'layers' sheet.In the Analysis Box (right-hand, Figure 7), there are two sections: the 'toolbox' (Figure 7e), which enables the end-user to browse through the many LANDSUPPORT tools, and the 'results' section (Figure 7f), where the user can visualise his own results, which have been produced and stored on the platform (including the applied model, scale, processing status, etc.).Additionally, at the very top of the dashboard, there are some devices (measure distances/areas, point locator and draw polygon) that enable the user to draw the ROI he/she is interested in and save it within a public (or personal) storage space for use it whenever he/she requires.The ROI can also be modified or deleted at a later stage.Finally, on the lower right-hand side of the GUI, the user will find the specific tool of interest (Figure 7g).

| Database
The database is connected with the 15 macro-tools and typically consists of geo-referenced data and metadata from different sources (Table 4).Up to now, LANDSUPPORT stored more than 450 layers, equally subdivided between raster and vector data.The data are stored in different folders according to the scale of application and themes.All stored data are reported in the Supplementary material and are available through the project catalogue at www.landsupport.eu.Some of these data, such as those found in official repositories of a public institution (e.g.ESDAC-JRC, Imperviousness-ISPRA), were already available, some were also interoperable (e.g.EU Copernicus); some others were held within the local repositories of public institutions (e.g.soil type and database of the Marchfeld region, Austria), a few required further processing (e.g.population data from Eurostat) and many new data had still to be created (soil database integration for the Campania region, Italy).Data are organised on the basis of the following: (i) theme area (soil, land use, geology, etc.), (ii) data type (vector data such as Corine CLC or raster data as DEM), (iii) spatial extension/resolution (European, National, Regional or local) and (iv) data details (source and year).
Before being integrated into the LANDSUPPORT database, raster and vector data were tested for quality and, where required, subjected to up-scaling procedures (e.g.coarser resolution data are better for some applications) or checked for spatial coordinates, missing data, outliers, etc. and then uploaded.
Table 4 represents an excerpt from the data layers stored in the database, either already available from existing data repositories or created during the project.They are include data on soil, geology, land use, morphology, meteorology, biodiversity, and so on.
To produce some of the tools, it was necessary to generate and harmonise a large set of these data.

| Models
All tools perform a variety of types and numbers of processing operations on the ROI selected by the user.This processing generates a variety of outputs, ranging from the visualisation of standard maps, through the production of tables and graphs, to geospatial functions that enable the complex processing required by some of our models.
Most of the tools involved writing specific codes aiming to produce new models or recompiling and adapting existing programme code to fit into our LANDSUPPORT system.
Table 5 gives a short description of the implemented models, which are at the basis of the 15 tools.Each of these models refers to a specific theme, a specific functionality and operating mode (e.g.'onthe-fly' or offline mode), the required input and, finally, the outputs produced.
Models can be organised according to their operating mode in the following two types: 1. Static models employ offline processing: one example is related to climatic data on a regional scale.This operation allows us to evaluate in advance the parameters of linear relationships between temperatures (average, minimum and maximum) and altitude, thus   F I G U R E 4 Knowledge transfer in Landsupport.
F I G U R E 5 Design of the engine to optimise geospatial decision support system tools for sustainability.CMCC, Centro euro Mediterraneo per il Cambio Climatico; FMIS, farm management information systems; GPU, graphic processing unit; HPC, high performance computing.
statistics of the output parameters.This data cannot be pre-calculated as it depends on the specific ROI designed by the user at the time of the query.

| Co-development and testing of the LANDSUPPORT platform
Successful S-DSS tools indeed require a continuous process of codevelopment and testing of tools by end-users both during and after the release of the different tools.
This activity has covered all phases of the DSS development process, namely: (i) the preparatory phase with needs assessment; (ii) the testing phase and (iii) the technical dissemination phase, which was the final step of the process and aimed at making the LANDSUPPORT platform known to its potential users and enabling them to use the platform through targeted hands-on sessions and training.
In our case, co-development included different forms of feedback to the developers, depending on the issues raised, such as (i) semistructured interviews, including requests and remarks from experts; (ii) e-mails describing problems in using tools and (iii) direct interaction with the developers.The feedback activities contributed to the codevelopment and co-creation of the S-DSS tools by identifying the main concerns of stakeholders concerning the tools.They also helped to establish a direct link between stakeholders and developers.
Assessment was made of users' judgments that emerged during the interviews regarding performance indicators, including interoperability, reliability, relevance for policy needs and overall satisfaction with the functionalities (usability and operational capacities) of the LANDSUPPORT DSS tools.
Here, we report some results: at the country level, the total number of institutions involved in the testing process in the three pilot countries (Italy, Austria and Hungary) is 55, and 367 people tested the tools.By category, the figures are as follows: 32 public institutions were involved and 127 people tested the tools.Regarding the actors of agriculture, environment and spatial planning, the numbers are 9, 5 and 9 for the institutions and 85, 106 and 49 for the people.
Throughout the duration of the project, 25 workshops were organised, involving 877 potential users and other stakeholders at European, national and regional levels and on tools belonging to different application areas.Throughout the duration of the project, the LANDSUPPORT platform received approximately 910 registrations and more than 4500 log-ins.
To highlight the functioning of the platform, we here report 2 case studies (in addition a third case on Sustainable Land Management [SLM] practices is reported in the Supplementary materials) being put to use aiming towards sustainable practices and lowering land degradation.These have been chosen to cover the different spatial extents:  Here, we report an example.In a predefined ROI, chosen in the Marchfeld region in Austria, two different land use management strategies were performed in the same 7 years of simulation (2012)(2013)(2014)(2015)(2016)(2017)(2018).
Case 1 (organic approach -Table 1) In Table 6, we report the main results after using the crop growth too.
From the above simulations, the user can obtain a clear quantification of the following ecosystem services: (i) crop production, (ii) soil filtering (Nitrate leaching) and (iii) C sequestration (SOC variation) capabilities, all performed under alternative crop rotations (with vs. without cover crop), fertilisation (mineral vs. organic), tillage (ploughing vs. minimum tillage) and crop residue management (retained vs. removed).
In this specific case, we observe that for the same area of 8.4 ha, in the case of raw soil, maize yield is 8.1 tons ha À1 year À1 in case 1 versus 10.4 tons ha À1 year À1 in case 2; but this increase of 2 tons ha À1 year À1 in annual production yield has a trade-off in terms of both (i) N leaching, which increases from 55.9 kg NO 3 -N ha À1 year À1 (case 1) to 129.2 kg NO 3 -N ha À1 year À1 (case 2) and (ii) annual SOC, which shows an increase of 0.53% for case 1 and a decrease of -0.14% for case 2. ) and harvested volume (m 3 ha À1 year À1 ).
The user using the 'MLFS tool' will be able, to select between the following parameters to be employed by the simulation process: Finally, MLFS supports a wide range of plot-designs and data types and can be applied to various forest types, from monocultures to forests with rich species compositions.
The tool produces two tables as outputs.The first table (given here as Table 8) reports for each forest plot captured by the ROI, the three parameters discussed above, namely: groeing stock, basal area and harvested volume.
The tool will then produce a second table (not reported) with the basic statistics (count, mean, min, max, standard deviation, 25% percentile, 50% percentile and 75% percentile) about growing stock, basal area and harvested volume for the entire selected ROI.
To highlight the importance of local pedoclimatic settings over management practices, we performed a sensitivity analysis using the forestry tool.
In Figure 8, a boxplot is reported where the numbers identifying 30 different ROIs (forest plot) of the same size (about 0.1 ha) and circular shape are given on the x-axis while the mean of harvested volume (m 3 ha À1 year À1 ) estimated for the year 2030, considering all possible combinations of four variables (yield harvesting, thinning, mortality and RCP scenario), is reported on the y-axis.The sensitivity analysis emphasises the existence of sites

| CONCLUSION
This work starts by acknowledging that the vast availability of data (e.g. from satellites, sensors and drones), policies (e.g.SDGs, EU environmental policies) and stakeholder engagements (e.g.EIP-AGRI) are not enough to have a positive impact on sustainable land management, especially considering that land degradation is worsening to a tipping point at which it will be difficult or impossible to reverse it.We assert that agriculture, forestry and environmental protection require something additional to the information at present accessible to address some key shortcomings due to: (i) the great physical and socio-economic multifaceted complexity of the landscape and land degradation processes, (ii) the lack of a truly integrated physical approach to many agriculture/environmental problems, (iii) the lack of factual scientific support for both farmers and regional governments so as to support the adoption of sustainable agriculture practices, (iv) the fragmentation of current approaches, models and DSS tools dealing with agriculture and environmental protection and The platform has been demonstrated here through three cases dealing with sustainable management practices in agriculture and forestry to lower land degradation processes.In the three case studies (one is reported in Supplementary materials), this is achieved by producing results and stats on sustainable crop growth, forestry and sustainable land management.
In terms of the future and legacy, the LANDSUPPORT platform will remain operational, open and free long after the project ends.This position is rooted in the evidence showing that we need to leave these tools as open as possible and engage as much as possible with a large community of users if we are to protect soils and land.Currently,

K
E Y W O R D S land degradation, land management, soil, spatial decision support system, sustainability 1 | INTRODUCTION 1.1 | Background There are numerous means available nowadays to challenge land degradation and implement environmental sustainability, both in the European Union (EU) and elsewhere.Amongst these are as follows 1. Everyday satellites (e.g.USGS-NASA, ESA Copernicus), sensors on board drones and in-field, which produce hundreds of terabytes of data that can support monitoring and management systems and models for sustainable land use.2. A large availability of precision farming techniques, robotics, omics, biopesticides and nanoparticles which support sustainable farming.3. Stakeholders engaged in research and innovation that lead to better-orientated land planning and management.A classic example is the EU's investment of great resources in implementing European Innovation Partnerships (EIP) for agriculture and sustainability and, more recently, in Soil Mission Living Labs and Lighthouses.

F
I G U R E 1 The figure depicts current fragmentation of models and decision support systems (DSSs) approaches.CC, climate change; ES, ecosystem services; LULC, land use and land change.
2.1.3| Co-design the tools Development, testing and engagement involved different types of users and their different needs due to the multisectoral and multiscale nature of the project.
Optimisation of the Landsupport geospatial cyber-infrastructure (CGI) approach.CC, climate change; ES, ecosystem services; LULUCF, land use and land change forestry; S-DSS, geospatial decision support system.
polygons, and are stored in PostgreSQL (open-source license database) and managed using the PostGIS extension.Raster data are typically composed of pixel arrays (with continuous or discrete values)and are managed through a data cube technology, in our case the rasdaman database(Baumann et al., 2021), allowing management, storage and recovery of huge multidimensional arrays.
2. Dynamic models employ 'on-the-fly' calculations: this is the most commonly used type of model for biophysical processes such as crop growth.Using daily climatic data, these models evaluate the climate-soil plant dynamics for each day and produce output phenology, yield and many other outputs.The latter can be used, for example, to populate PDF reports and graphs describing the basic T A B L E 3 (Continued)

3. 1 |
Crop growth case study3.1.1 | BackgroundThe new fair, green and performance-based EU Common Agricultural Policy (2023-2027) will be a key item in securing a more sustainable future for agriculture and forestry, as well as achieving the objectives of the European Green Deal (e.g.combat land degradation).This Deal requires each EU country to design a national strategic CAP plan which, in addition to many other aspects, has to put into practice enhanced conditionality, eco-schemes and farm advisory services, as well as agri-environmental and climate measures and investments, to address the Green Deal targets (e.g.Farm to Fork Strategy, Biodiversity Strategy 2030).For the best implementation of the strategic plans, the planned interventions (e.g.GAEC, SMR) should be designed 'considering À1 ha À1 ) Note: The models used in the 3 case studies reported below are reported with a grey background.Abbreviation: SOC, soil organic carbon.F I G U R E 7 Scheme of the graphical user interface and its different panels.[Colour figure can be viewed at wileyonlinelibrary.com] characteristics of the areas concerned, including soil and climatic conditions, existing farming system, land use, crop rotation, farming practices and farm structures' (territorial scope).In this specific scenario, this example case addresses the following specific issue: in designing Strategic Plan in line with the new Common Agricultural Policy (2023-2027), a national administrative authority aims to test, in a specific region, how different land use management scenarios have an impact on three ecosystem services, two of which are critical in terms of lowering land degradation: production, filtering capability and carbon sequestration.3.1.2| LANDSUPPORT crop growth tool solution 'Crop Growth' is a Landsupport tool that support the design of the Strategic Plan in line with the new Common Agricultural Policy.The 'Crop Growth' tool performs crop growth modelling and produces results regarding crop yield, so enabling simulation of management scenarios.The tool runs, 'on the fly', through a model request that allows the user to carry out the simulation.The heart of the tool is ARMOSA (Analysis of cRopping systems for Management Optimisation and Sustainable Agriculture) process-based crop model(Perego et al., 2013;Valkama et al., 2020), which simulates the high level of complexity found in those agroecosystem processes which vary in response to agricultural management (i.e.crop rotation, intercropping, crop residues management, fertilisation, irrigation and tillage) and pedoclimatic conditions.The user has to select the 'Crop Growth' tool and then he must choose between the following parameters: (i) the ROI; (ii) the date (start and end) of simulation; (iii) the crop rotations (on the basis of the local site); (iv) conventional agriculture versus organic farming (this choice is also associated with type of fertiliser); (v) occurrence of cover crop; (vi) irrigation management (either restoring 100% or 80% of crop requirement) and (vii) tillage (two options: ploughing at 30 cm depth or tillage limited at 15 cm in topsoil with no mixing of soil layers).The user can select also the type of output: (i) production (tons ha À1 year À1 ) of each crop included in the rotation (for each simulated year), (ii) mean annual nitrate leaching (NO 3 -N kg ha À1 year À1 ) at the base of the soil profile and (iii) mean annual change of soil organic carbon stock in the 30 cm topsoil layer (% of changes per year).
: Crop rotation = Sunflower + Maize; Irrigation = 100%; Fertilisation = Inorganic; Tillage = Conventional; Residue = No (resutls of this simulation are shown in Table (i) ROI (freely drawn or selected from administrative NUTS4 -NUTS3 areas); (ii) year of the simulation; (iii) harvesting yield (which represents the simulation of the intensity of forest cutting as a percentage); (iv) RCP scenarios (RCP: 4.5 and 8.5 on the basis of GHG emissions until 2100); (v) mortality rate (which is made on the basis of local conditions) and (vi) thinning rate, affecting future growing dynamics.
(e.g.sites18, 24 26, 28, 29 and 30)  where changes in yield harvesting, thinning rate, mortality rate and RCP scenario have a huge impact on harvest, while there are other sites (e.g. 1, 2, 4, 7, 9 and 22) where the impact of the three management variables and RCP scenarios have a minimum impact on the growing stock.This tool supports different users such as (i) forestry authorities to carry out scenario analysis and thus better fine-tune management practices on the basis of specific local settings, (ii) scientists interested in the long-term effects of environmental factors and climate on forest development and finally (iii) forest professionals and engineers, who can use the tool as a decision-making support in their forest planning process.
(v) difficulties in the implementation of the many good policies aimed at improving the environment and agriculture.Our vision is that robust operational S-DSS are the way forward to translate data availability into positive actions towards sustainability thus challenging land degradation.This was the starting point of the EU-funded LANDSUPPORT project, which brought together 19 partners from 10 different countries across Europe, the Middle East and Asia and developed 15 macro-tools (and more than 100 elementary tools) by establishing a free web-based smart geospatial decision support system.This system is based on Geospatial Cyber-Infrastructure IT architecture applied to the better implementation of a large set of EU, national and regional policies in the field of agriculture, forestry and the environment.The system brings together diverse sources of data including Sentinel satellites (from the EU's Copernicus programme), advanced climate and ECZ modelling, various technologies (e.g.datacube and parallel processing) and stakeholder engagement to develop and test the tools.All this complexity is 'hidden' behind an easy-to-use graphical user interface that has been created to enable end-users to easily access and work on the LANDSUPPORT platform.Here, we have shown that it is possible to overcome the current fragmentation of data and models by combining the following features into a single system: (i) a user-friendly GUI where all complexity is hidden; (ii) implementation of the concept of the operational multifunctionality of land and soil; (iii) adaptability to many different needs of end-users; (iv) implementation of 'what-if' modelling, so empowering the choices of the end-users (In this sense, we want to emphasise here that the system does not aim to provide the 'solution', but rather provide a set of 'options' for the user to choose from); (v) low cost of transferability of the approach to new areas and (vi) incorporation of bottom-up contributions from users (e.g.uploading of ROI or local soil data).
space Art.12) EU Member States shall define minimum standards and good practices (GAEC and SMR, ANNEX III) considering characteristics of the areas concerned, including soil and climatic conditions, existing farming system, land use, crop rotation, farming practices and farm structures.Schemes for the climate and the environment (Art.28)Elements common to several interventions (Art.98).
StaticLand takes limitation by circular use of land (sect.3.2.2) DynamicClosing the nutrient and carbon cycles (sect.3.2.3)DynamicSoil for healthy water resources (sect.3.4)StaticMakingsustainable soil management (sect.4.1)DynamicRestoring degraded soils and contaminated sites (sect. Policies addressed by the Landsupport System and list of 15 macro-tools (a more detailed version of this table is provided in the Supplementary material).
Desired features of the Landsupport decision support system system to address current agriculture and environmental sustainability challenges.[Colour figure can be viewed at wileyonlinelibrary.com]T A B L E 2 Achievements and tangible products produced after the Landsupport tools (a more detailed version of this table is provided in the Supplementary material).
R (Continues)allowing easier spatial interpolation of temperature (using Postgres/PostGIS table).
Some of the data layers stored in the Landsupport database (a more detailed version of this table is provided in the Supplementary material).Selection of models employed in Landsupport (a more detailed version of this table is provided in the Supplementary material).
region, country and Europe.Additional information on all case studies concerning both the environment (e.g.soil, geology, etc.) and local communities (e.g.NUTS4 population trends in the last decade) are available in other Landsupport tools (e.g.environmental report tool not explained in this paper).FI G U R E 6 Architecture of the Landsupport geospatial cyberinfrastructure. [Colour figure can be viewed at wileyonlinelibrary.com]TA B L E 4 a Note: ISPRA: https://www.isprambiente.gov.it/en.aData are created during the project.T A B L E 5 Results after the use of the crop growth tool in the first land use management case in Marchfeld region (Austria).
Results after the use of the crop growth tool in the second land use management case in Marchfeld region (Austria).
The horizontal line within the box marks is the median value.The box contains the middle 50% of the data points (IQR) and is a measure of data variability while the vertical bar for each ROI represents the upper and lower fences of the harvested volume based on all combinations of the variables.TA B L E 7 E 8 Sensitivity analysis of the harvest in the different region of interests (ROIs) (forest plot).[Colour figure can be viewed at wileyonlinelibrary.com] there is great interest from such stakeholders as public administrators, cooperatives, farmer associations, spatial planner associations, natural parks and metropolitan areas.Actually, members of the Italian Parliament proposed at the Italian Senate a new soil law on the basis of LANDSUPPORT (soil legislative bill ddl 2614).In addition, the project has received two awards (EU Success Story and Falling Walls global winner), which makes it hopeful that a widening of use, adoption and scope is welcome and possible.AFFILIATIONS 1 CRISP Research Center, Department of Agriculture, University of Napoli Federico II, Naples, Italy 2 Department of Agricultural and Environmental Sciences, University of Milan, Milan, Italy 3 Institute for Mediterranean Agricultural and Forestry Systems (ISAFOM), National Research Council (CNR), Naples, Italy 4 Crops For the Future UK, NIAB, Cambridge, UK 5 rasdaman GmbH, Bremen, Germany 6 Umweltbundesamt -Environment Agency Austria (EAA), Wien, Austria 7 EMM SRL, Napoli, Italy 8 ARIESPACE SRL (ARIES), Established in Centro Direzionale IS.A3,