Monitoring the implementation of regenerative practices in a carbon farming programme – CUP4SOIL showcase documentation

CUP4SOIL Showcase documentation Doc. ID C4S-SCS-002 Issue 1.2 Date 14.11.2025 Page 0 of 8 CUP4SOIL Monitoring the implementation of regenerative practices in a carbon farming programme – CUP4SOIL showcase documentation Jelle van Wesemael1, Celine Bassine1 1 SoilCapital (Belgium) Doc. ID SCS-002 Issue 1.2 Date 14.11.2025 CUP4SOIL Doc. ID C4S-SCS-002 Showcase documentation Issue 1.2 Date 14.11.2025 Page 1 of 8 Document based on EOC-Document-Template, Issue 1.0 CONTENTS 1. Purpose and structure of the document .............................................................................................. 2 2. CUP4SOIL project overview ................................................................................................................ 2 3. Showcase – Monitoring the implementation of regenerative practices in a carbon farming programme (Soil Capital) .................................................................................................................................................. 3 3.1 Partner description .............................................................................................................................. 3 3.2 Involved data ....................................................................................................................................... 3 3.3 Task for this showcase ........................................................................................................................ 4 3.4 Summary / evaluation of the CUP4SOIL data used ............................................................................ 8 LIST OF FIGURES Figure 1: Scatter plot comparing bare soil frequencies from SC and Soil Suite datasets............................. 5 Figure 2: Box plot of differences in bare soil frequency between SC and Soil Suite datasets. .................... 5 Figure 3: Box plot comparing the median frequency of bare soil between the two groups of agricultural practices. 6 Figure 4: Histogram comparing the distribution of bare soil frequency between two groups of farmers with different tillage practice; Group 1: Farmers who shifted to no-till before 2015, Group 2: Farmers with no change in practice. ......................................................................................... 7 Figure 5: Two farmers with different tillage practices (regenerative vs. conventional) superimposed on Soil Suite data (frequency of bare soil). ........................................................................................... 7 LIST OF TABLES Table 3-1: CUP4SOIL data list ...................................................................................................................... 3 CUP4SOIL Doc. ID C4S-SCS-002 Showcase documentation Issue 1.2 Date 14.11.2025 Page 2 of 8 Document based on EOC-Document-Template, Issue 1.0 1. Purpose and structure of the document One of the objectives of the CUP4SOIL project is to prepare and develop a user community that tests and validates data products suitable for soil health/quality assessment. One of the pillars of this objective is the development of showcases. This document describes the conducted showcase by the aim, partner, methodology and results. It is expected to gain important and user-based information for the usefulness of the generated EO-based information products as well as a more detailed view into the purpose, for which the information products might be used in the future. 2. CUP4SOIL project overview The CUP4SOIL project aims to contribute to a downstream EO-based service to support national and European agencies in reporting on soil health/quality and thus, contribute to Land Degradation Monitoring (LDN) and Sustainable Development Goal (SDG 15.3.1) reporting. The products can potentially also be used for other than national uses, such as for more local and/or commercial applications. It underpins the pre-operational Soil Monitoring System currently being developed within the ESA WorldSoils project with the potential to serve as a component of the Copernicus Land Monitoring Service. The synergies between the ESA WorldSoils project and the CUP4SOIL activities investigating soil health mapping will streamline the activities and boost user uptake. The 2-year CUP4SOIL project comprises the following objectives: • Suggest soil data products as new products for the Copernicus Land Monitoring Service to support national and European agencies for reporting on soil health/quality. • Generate European-wide example data products characterising soil health/quality • Develop a user community that tests and validates data products for soil health/quality information • Ensure close cooperation with the ESA WorldSoils project activities and other related projects/initiatives Initiatives and literature: In the first step, CUP4SOIL explores the different literature and project resources to get an update about the current discussion of essential soil health indicators. This is collected in a first version of a User Requirement Document (URD) that CUP4SOIL presents, discussed and adapted with a larger community during the first User Requirement virtual meeting. For this purpose, a specific online user survey is developed based on the framework of the Copernicus user requirements. The survey is planned to be repeated at the end of the project. Soil data products: In the next step, CUP4SOIL generates European-wide soil information products based on Sentinel-1 and Sentinel-2 data. For this purpose, DLR and ISRIC join their large-scale processing expertise and facilities. DLR is creating several soil-related input products such as soil reflectance composites, information about the cover frequency of soils and the vegetation dynamic on a high spatial resolution (20 m). These data products flow into the high-performance computing environment of ISRIC that generates information about soil organic carbon content, texture pH values, etc. using digital soil mapping approaches. The resulting output data will be made publicly available. Selected key users are involved to evaluate the usability of the proposed soil information products. It is also planned to compare the CUP4SOIL products with other existing European-wide soil products from running projects and initiatives. The results will be presented and discussed at the final workshop. Showcases: Based on this user exchange and feedback, the Copernicus URD is updated and required changes are integrated into the soil information processors of DLR and ISRIC to improve the soil product portfolio. In this phase, showcases are developed together with local municipalities, national institutes and European agencies. The showcases use the suggested Copernicus soil input data to derive higher-value information about soil quality and health for specific regions to demonstrate the usability and applicability of the soil product portfolio for Copernicus users. The results and evaluation of these showcase developments will be presented and discussed for a larger user community at the final CUP4SOIL workshop. CUP4SOIL Doc. ID C4S-SCS-002 Showcase documentation Issue 1.2 Date 14.11.2025 Page 3 of 8 Document based on EOC-Document-Template, Issue 1.0 3. Showcase – Monitoring the implementation of regenerative practices in a carbon farming programme (Soil Capital) This showcase focuses on the monitoring of regenerative agriculture using CUP4SOIL data. The main objective is to test the potential of this data for MRV (Monitoring, Reporting, and Verification) purposes, agronomic insights for farmers, and quality control. This dataset could be particularly valuable in assessing long-term changes in areas with high bare soil frequency, supporting the validation of continued regenerative agriculture practices post-program implementation. 3.1 Partner description Organisation: Soil Capital Main responsibilities: Soil Capital specializes in agricultural sustainability and carbon sequestration. The key focus in this showcase is to evaluate the potential of EO-derived soil products to enhance soil monitoring methodologies, especially for regenerative agriculture. Experience with image data: Soil Capital has experience using commercial service providers to monitor farming practices, including tillage, crop cover, and crop type detection. 3.2 Involved data Data (acronym) Data name Area Time period Used (checkbox) SCMaP Soil Suite SRC-MEAN Bare Surface Reflectance Composite (Mean) SRC-STD Bare Surface Reflectance Composite (Standard deviation) SRC-CI95 Bare Surface Reflectance Composite (95% Confidence) MASK Bare Soil Mask France, Belgium 2017-2022 x BSF Bare Soil Frequency France, Belgium 2017-2022 x BSC Bare Soil Count VPC Valid Pixel Count MREF-MEAN Reflectance Composite (Mean) MREF-STD Reflectance Composite (Standard deviation) Table 3-1: CUP4SOIL data list Other data involved: • Historical farming practices and monitoring data from a commercial supplier and • A sample of our farmer database at Soil Capital CUP4SOIL Doc. ID C4S-SCS-002 Showcase documentation Issue 1.2 Date 14.11.2025 Page 4 of 8 Document based on EOC-Document-Template, Issue 1.0 3.3 Task for this showcase Objective Test how CUP4SOIL data can be used for agricultural monitoring, particularly in the context of regenerative agriculture, by comparing it with other data sources (one commercial and one farmer-based). 1. Comparison with data from a commercial service provider Data: data from a commercial service provider used to monitor farmers' practices in the Soil Capital program (referred to as SC data in the rest of this comparison). The data provides crop types, tillage and cover crop detection for 1709 plots in France and Belgium, excluding permanent pastures, with dates ranging from 2017 to 2022. Method: • Calculation of the number of “bare soil days” between 2017 and 2022 for each plot in the SC data. Bare soil days" are days when the soil is not covered by a crop, by a cover crop or is tilled. • Soil Suite data are extracted for each plot as the median of the bare soil frequency values on the plot. Comparison and observations • Observation: systematic overestimation by Soil Suite compare to the SC data (however, underestimation by the commercial provider cannot be ruled out): o Scatter plot (Figure 1): Shows strong dispersion around the diagonal. Most points fall below the line y = x, suggesting that Soil Suite estimates higher bare soil percentages than the SC data. o Box plot (Figure 2): The distribution of differences (% bare soil SC - % bare soil SoilSuite) is centered around negative values, reinforcing the trend of overestimation by Soil Suite. Some extreme outliers (~ -40%) further highlight this. • Note: 33 out of the 1709 parcels (~2%) are classified as permanent in the Soil Suite mask but not by the provider, with bare soil frequencies up to 12% (Figure 1). • R² ≈ 0.36 indicates a moderate correlation between the datasets, suggesting they capture similar trends but with significant variability. CUP4SOIL Doc. ID C4S-SCS-002 Showcase documentation Issue 1.2 Date 14.11.2025 Page 5 of 8 Document based on EOC-Document-Template, Issue 1.0 Figure 1: Scatter plot comparing bare soil frequencies from SC and Soil Suite datasets. Figure 2: Box plot of differences in bare soil frequency between SC and Soil Suite datasets. 2. Comparison with SC farmer's dataset Data: based on Soil Capital farmer’s program (manual encoding of their data) with historical tillage practice records (France only). Hypothesis: Farmers practicing no-till for a long period should show lower bare soil frequencies than those using conventional tillage. CUP4SOIL Doc. ID C4S-SCS-002 Showcase documentation Issue 1.2 Date 14.11.2025 Page 6 of 8 Document based on EOC-Document-Template, Issue 1.0 Method: • Select farmers with known tillage histories: o Group 1: Farmers who shifted to no-till before 2015 (119 parcels) o Group 2: Farmers with no change in practice (889 parcels) • Soil Suite data are extracted for each plot as the median of the bare soil frequency values on the plot. Comparison and observations • The comparison of bare soil frequency (median %) between farmers practicing long-term no-till (group 1) and those using conventional tillage (group 2) reveals a clear and consistent trend. • Box plot insights (see Figure 3): o No-till group shows a significantly lower median bare soil percentage compared to till group. o Outliers in the tillage group suggest some parcels have exceptionally high bare soil exposure, which is not observed in the no-till group. • Histogram insights (see Figure 4): o The distribution for Group 1 is skewed toward low bare soil percentages, with a strong peak near 0%. o In contrast, Group 2 displays a wider spread, with a modal value around 15–20%, and a long tail extending up to 50%. o This supports the hypothesis that conventional tillage is associated with more frequent and extensive bare soil presence, whereas no-till practices maintain greater ground cover. Figure 3: Box plot comparing the median frequency of bare soil between the two groups of agricultural practices. CUP4SOIL Doc. ID C4S-SCS-002 Showcase documentation Issue 1.2 Date 14.11.2025 Page 7 of 8 Document based on EOC-Document-Template, Issue 1.0 Figure 4: Histogram comparing the distribution of bare soil frequency between two groups of farmers with different tillage practice; Group 1: Farmers who shifted to no-till before 2015, Group 2: Farmers with no change in practice. These results confirm the hypothesis that long-term adoption of no till practices results in a decrease in the frequency of bare soil, as shown by Soil Suite EO data. This not only validates Soil Suite's potential for differentiating tillage practices, but also reinforces its use in monitoring regenerative agriculture programs. The analysis confirms that historical farming practices can be meaningfully linked to EO-derived indicators to assess conservation practices at scale. Here is an illustration of our program parcels from two farmers with different tillage practices (regenerative vs. conventional) superimposed on Soil Suite data (frequency of bare soil). Figure 5: Two farmers with different tillage practices (regenerative vs. conventional) superimposed on Soil Suite data (frequency of bare soil). CUP4SOIL Doc. ID C4S-SCS-002 Showcase documentation Issue 1.2 Date 14.11.2025 Page 8 of 8 Document based on EOC-Document-Template, Issue 1.0 3.4 Summary / evaluation of the CUP4SOIL data used Usefulness • The CUP4SOIL data shows strong potential for assessing the adoption of regenerative agriculture and distinguishing between permanent and cultivated fields. Its spatial resolution is well-suited for both regional and field-scale agricultural analysis. • The "bare surface frequency" indicator and the 3-class mask (bare surface, permanent vegetation, other) are highly relevant for our use cases. • While Soil Suite may not fully replace other EO tools due to specific functionalities, it offers valuable complementarity through increased granularity. In particular, it is effective in differentiating between permanent and cultivated fields. • Its standardized data products can also serve as useful benchmarks for validating other soil-related datasets. • In the long term, the dataset is valuable for monitoring persistent changes in areas with high bare soil frequency—particularly in the context of regenerative agriculture—to assess whether regenerative practices are maintained after program implementation. Data handling The data formats are manageable, but when scaled to a national level (e.g., France), the volume of data becomes large and may result in slow downloads and processing times. Accuracy The Soil Suite appears to systematically overestimate bare soil frequency when compared to commercial datasets. This discrepancy could stem from underreporting by commercial providers or methodological differences. Suggestions for improvement • Temporal resolution: The current 5-year composite smooths out interannual variability and fails to capture short-term agricultural changes, such as annual tillage practices. Enhancing temporal granularity would significantly improve usability. • System integration: For better integration into automated monitoring systems, more current and temporally resolved datasets are essential.