Bridges not Walls: How can we build interoperability between diverse quality cultures without forcing limiting standards?

How can we build interoperability between diverse quality cultures without forcing limiting standards? Markus M¨oller on behalf of FAIRagro Julius K¨uhn Institute ·Department of Digitalisation and Artificial Intelligence ·Kleinmachnow, Germany · · [email protected] The Soil Mapping communication gap What is the best way to communicate DSM products? Annual conference of the German Soil Society (DBG) on 8 September 2023 Markus M¨oller Bridges not Walls 22. September 2025 2 / 16 The Soil Mapping communication gap German soil taxonomy Altermann, M. & Freund, K.L., 2015. Bodensch¨atzung in Deutschland: R¨uckblick – W¨urdigung – Ausblick. Deutsche Bodenkundliche Gesellschaft, Halle (Saale) K¨uhn, D., 2000. Zum Sinn und Zweck der bodenkundlichen Substratsystematik. In: Boden: Objekt menschlicher Nutzung und Forschung. Herrenh¨auser Forschungsbeitr¨age zur Bodenkunde, Band 3, S. 9–24. Markus M¨oller Bridges not Walls 22. September 2025 3 / 16 The Soil Mapping communication gap Digital Soil Mapping Definition . . . the creation and population of spatial soil information systems by the use of field and laboratory observational methods coupled with spatial and non-spatial soil inference systems. Also due to quality assurance concerns: DSM approaches have not yet been established as standard practice in German state geological services for soil information generation. Minasny, B., McBratney, Alex.B., 2016. Digital soil mapping: A brief history and some lessons. Geoderma 264, 301–311. https://doi.org/10.1016/j.geoderma.2015.07.017 Markus M¨oller Bridges not Walls 22. September 2025 4 / 16 The Soil Mapping communication gap What is the best way to communicate DSM products? The Digital Soil Mapping (DSM) literature has identified two levers: 1Uncertainty representations [as second map] that allow users to integrate uncertainty in decision making (but just one-third of publictions provide estimated uncertainty maps), and 2Scale-specific representation of soil information and associated uncertainties Courteille, L., Tardieu, L., Boukhelifa, N., Lutton, E., Lagacherie, P., 2025. What is the best way to communicate the uncertainty of a digital soil mapping product? Some lessons from an end-users survey. Geoderma 459, 117302. https://doi.org/10.1016/j.geoderma.2025.117302 Piikki, K., Wetterlind, J., S¨oderstr¨om, M., Stenberg, B., 2021. Perspectives on validation in digital soil mapping of continuous attributes—A review. Soil Use and Management 37, 7–21. https://doi.org/10.1111/sum.12694 M¨oller, M., Volk, M., 2015. Effective map scales for soil transport processes and related process domains — Statistical and spatial characterization of their scale-specific inaccuracies. Geoderma 247–248, 151–160. https://doi.org/10.1016/j.geoderma.2015.02.003 Schmidinger, J., Heuvelink, G.B.M., 2023. Validation of uncertainty predictions in digital soil mapping. Geoderma 437, 116585. https://doi.org/10.1016/j.geoderma.2023.116585 Markus M¨oller Bridges not Walls 22. September 2025 5 / 16 The Soil Mapping communication gap SOC content prediction in Bavaria PARAMETRIZATION FEATURE ENGINEERING MODELLING Soil Samples SCMaP-SRC Terrain Attributes RU Importance Testing Set Data Split Sampling Design Machine Learning Training Set RFE Prediction Uncertainty External Validation Accuracy Metrics M¨oller, M., Zepp, S., Wiesmeier, M., Gerighausen, H., Heiden, U., 2022. Scale-Specific Prediction of Topsoil Organic Carbon Contents Using Terrain Attributes and SCMaP Soil Reflectance Composites. Remote Sensing 14, 2295. https://doi.org/10.3390/rs14102295 Markus M¨oller Bridges not Walls 22. September 2025 6 / 16 The Soil Mapping communication gap Data quality communication obstacles Standardized data quality frameworks and metrics like ISO 19157-1:2023 serve to assess data quality from the perspective of producers, but may not be sufficient to communicate local and regional inaccuracies. SOC prediction for a Bavarian test site 5 10 15 5 10 15 Observation (test data) Prediction R2=0.838 RMSE =1.991 SLOPE =0.63 Prediction of soil organic carbon content utilizing multisource satellite data: assessing the influence of sensors, temporality, modeling resolution, and land use June 12, 2024 Table 1: Add caption Rank SUBSET 1 SUBSET 2 TA SCR TA SCR 1 TPI31 RED TPI419 SWIR2 2 TPI48 GREEN TPI647 GREEN 3 TPI74 NIR TPI74 RED 4 TPI647 SWIR2 TPI271 NIR 5 TPI419 BLUE TPI114 BLUE R20.86 0.62 RM SE 2.45 2.65 General comments The most accurate possible SOC prediction is an important topic in order to better assess the condition of agricultural soils. Against this background, the study is part of a large number of studies in which attempts are being made to increase the quality of the prediction by using satellite imagery of different sources. The focus of this study is on analysing multisource satellite data regarding the influence of sensors, temporality, modeling reolution and 1 Subset-specific validation results TA – terrain attribute, SCR – soil reflectance composite https://www.openagrar.de/receive/openagrar_mods_00097720 Markus M¨oller Bridges not Walls 22. September 2025 7 / 16 The Soil Mapping communication gap Enhanced Data Quality description ⇒Large Language Models (LLM) Metric Scale Source Data Quality Taxonomy Levels 1 2 3 R2global ISO 19157-1 Thematic Quality Quantitative Attribute Accuracy x RMSE global ISO 19157-1 Thematic Quality Quantitative Attribute Accuracy x PI local 10.1016/j.geoderma.2023.116585 Spatial Uncertainty Numerical Spatial Uncertainty x PICP global 10.1016/j.geoderma.2023.116585 Spatial Uncertainty Numerical Spatial Uncertainty Rel. 5 10 15 5 10 15 50% Prediction Interval Observation (test data) Prediction (Median) R² = 0.748 RMSE = 2.144 PICP50 = 0.529 CRPS = 0.967 PREDICTION INTERVAL COVERAGE (Reliability): 50% PICP: 0.554 (Target: 0.50) Markus M¨oller Bridges not Walls 22. September 2025 8 / 16 Bridges not Walls Claims 1There are “crucial” Data-Fitness-For-Purpose (DFFP) categories and metrics supporting intra-sectional interoperability! 2There are existing (conceptual) cross-border interoperability frameworks that enable the connection of different quality cultures! Markus M¨oller Bridges not Walls 22. September 2025 9 / 16 Conclusion The combination of Data-Fitness-For-Purpose categories and conceptual spatial structuring approaches can support cross-domain and multi-scale interoperability! Domain-specific data quality frameworks with accepted “crucial” quality metrics should be as complex as necessary! Simplifying and structuring complexity supports cross-domain interoperability (but we are also working on ontology-based interoperability implementations ⇒FAIRagro plenary). Expert-guided LLMs ⇒Dynamic enhancement of existing standards Markus M¨oller Bridges not Walls 22. September 2025 15 / 16 Conclusion The combination of Data-Fitness-For-Purpose categories and conceptual spatial structuring approaches can support cross-domain and multi-scale interoperability! Domain-specific data quality frameworks with accepted “crucial” quality metrics should be as complex as necessary! Simplifying and structuring complexity supports cross-domain interoperability (but we are also working on ontology-based interoperability implementations ⇒FAIRagro plenary). Expert-guided LLMs ⇒Dynamic enhancement of existing standards Markus M¨oller Bridges not Walls 22. September 2025 15 / 16 How can we build interoperability between diverse quality cultures without forcing limiting standards? World Cafe topics 1Are there “crucial” quality categories in your domain that are recognized as suitable to promote interoperability? 2Are there conceptual frameworks in your research domain that could enable cross-border interoperability?