Advancing Productivity Monitoring in Alpine Grasslands
Enhanced satellite monitoring improves drought assessment, enabling more reliable grassland insurance and supporting climate resilience in mountain agriculture.
The Challenge
Alpine grasslands are essential for livestock farming, tourism and biodiversity. In South Tyrol (Italy), for example, they cover around 70% of the vegetated area while providing valuable ecosystem services such as climate regulation and preservation of biodiversity. As climate change increases the frequency and severity of droughts, grassland productivity declines, threatening forage production and the economic stability of mountain farms.
Traditional mitigation strategies and thus drought insurance schemes rely on field inspections to assess damage. For grasslands, however, these inspections are often too costly compared to the value of the yield. Earth Observation offers a scalable alternative through index-based insurance, where satellite-derived indicators are used to estimate drought impacts quickly, consistently and at relatively low-costs, across large areas.
Solutions: South Tyrolean study case
Since 2024, Eurac Research has provided a Grassland Productivity Index (GPI) to a farmers' consortium in South Tyrol. The GPI supports an index-based insurance scheme by estimating forage losses from satellite observations rather than field inspections. It is based on the strong relationship between measured grassland biomass and the Leaf Area Index (LAI), a key indicator of vegetation growth derived from Sentinel-2 satellite imagery.
Within the ScaleAgData Grassland R&I Lab, researchers focused on improving LAI estimation in Alpine environments. Frequent cloud cover often limits the availability of Sentinel-2 optical imagery, reducing the continuity of monitoring. To address this challenge, the team worked on two complementary approaches:
- improving Sentinel-2 preprocessing by refining terrain correction and cloud detection in mountainous areas/complex terrain
- and integrating Sentinel-1 Synthetic Aperture Radar (SAR) observations with topographic and meteorological data to enhance LAI estimation when optical data are unavailable.
Several data fusion strategies, including Machine Learning and Deep Learning models, were evaluated to combine optical, radar and ancillary datasets. To ensure a robust assessment of the different approaches, Eurac developed a dedicated field protocol and collected ground measurements across eight Alpine grassland parcels over two growing seasons. This unique dataset provided a valuable benchmark for evaluating satellite-derived LAI under the complex conditions of mountain agriculture.
Impact and future outlook
The improved Sentinel-2 LAI product showed excellent agreement with in situ measurements, accurately capturing the seasonal evolution of Alpine grasslands and management. In particular, it was able to detect key management practices such as mowing events (Fig. 1), demonstrating its suitability for operational grassland monitoring.
When incorporated into the Grassland Productivity Index (GPI), it successfully detected year-to-year variations in grassland productivity at the regional scale, including the severe agricultural drought that affected large parts of Europe in 2022. (Fig. 2).
These results demonstrate the potential of satellite-based monitoring to support operational drought assessment and more efficient index-based insurance schemes.
The project also developed a promising Deep Learning model [2] to predict Sentinel-2 LAI by integrating Sentinel-1 radar observations, digital elevation model information and remote sensing based soil moisture data. The model is openly available through the project's GitHub repository [3]. Validation against Sentinel-2 LAI over Alpine grasslands in South Tyrol during 2023 showed promising results [2] (Fig. 3). However, comparisons with independent in situ measurements highlighted that the current approach has limited ability to reproduce LAI dynamics outside the training period and calibration sites. These results underline the challenges of estimating complex biophysical variables such as LAI from SAR observations alone in heterogeneous mountain environments. Nevertheless, the analysis confirmed the sensitivity of Sentinel-1 radar features to grassland canopy structure and biomass, demonstrating their potential to complement optical observations and support more continuous grassland monitoring under cloudy conditions.
