Land Lying Fallow
The Land Lying Fallow use case demonstrates an approach to monitoring fallow land using satellite data to ensure compliance with agricultural eco-schemes. The approach involves leveraging Sentinel-1 and Sentinel-2 images to extract time series data, focusing on NDVI (Normalized Difference Vegetation Index) and radar coherence. These metrics help detect key land management activities by identifying abrupt changes in vegetation and soil cover.
The workflow includes visual interpretation and marker detection to track compliance with country-specific fallow land management rules. By using satellite data to automate the detection of activities like mowing or the presence of bare soil, the system flags parcels based on whether they meet eco-scheme requirements. This scalable approach improves the efficiency of compliance checks across different regions and countries.
The results demonstrate the effectiveness of using NDVI and radar coherence data, though challenges like parcel size and data quality require further refinement.
Czechia: The goal is to verify whether fallow crops are grown and if the required resting period from June 1 to August 15 is respected. The results show that 35% of the analyzed parcels were too small for monitoring. Among the remaining parcels, 9% were found to have no detected maintenance activity, and 46% were non-compliant due to grassland management activity occurring within the resting period. Only 7% of parcels fully complied with eco-scheme regulations.
Sweden: The objective is to distinguish between green and black fallow, as well as production grassland, ensuring that the mandatory resting period is respected and at least one maintenance activity occurs annually. The results indicated that 44% of the analyzed parcels met the compliance requirements. However, 27% had no detected maintenance activity, and 15% had grassland management activities taking place within the resting period, leading to non-compliance.
Spain: In Spain, the eco-scheme requires that fallow land have at least one maintenance activity between January and September. The analysis found that 49% of parcels were too small or lacked enough cloud-free satellite data for analysis. Among the remaining parcels, 46% were in compliance, while 5% did not meet the required maintenance activity.
Netherlands: In the Netherlands, the challenge is monitoring small strips of fallow land bordering larger agricultural parcels. Due to the narrow size of these strips, high-resolution satellite data were required. While the strips were generally identifiable, factors such as shadows from trees, farm buildings, and machinery made automated detection difficult. Visual interpretation remains the most reliable method for compliance checks.
Example of NDVI data analysis and flagging the analysed parcels in terms of their compliance with the rules applied on fallow land.