AI Innovation to Predict Peatland Moisture and Strengthen Sustainable Supply Chains

A research team from Vilnius University’s Faculty of Mathematics and Informatics (VU MIF), in collaboration with Klasmann-Deilmann Ežerėlis, has secured a European Space Agency (ESA) grant to develop an AI-powered solution capable of predicting peatland moisture levels in real time.

By combining satellite radar, optical, and hyperspectral data with on-site field measurements, the project aims to deliver an interactive decision-support map that helps monitor peatland conditions and guide sustainable resource management.

“It’s remarkable how far both artificial intelligence and the space industry have advanced – together, they now enable solutions that were once beyond imagination,” says Dr. Valentas Gružauskas, the project’s lead researcher.

Aligning Industrial Operations with Environmental Goals

To achieve reliable forecasting, the team uses synthetic aperture radar (SAR) satellite data, which can penetrate cloud cover and observe surface conditions in any weather. Different radar wavelengths provide unique insights – shorter waves reflect off the surface, while longer ones capture vegetation canopy structure. When these radar data are combined with optical imagery and hyperspectral measurements, researchers can accurately estimate surface moisture levels.

“Because satellites don’t directly measure moisture, we conduct fieldwork in the Ežerėlis peatlands together with the Klasmann-Deilmann Ežerėlis team. The field data are integrated with satellite imagery and weather forecasts, allowing AI models to generate weekly peat moisture forecasts,” explains Dr. Gružauskas.

VU researcher Linas Beveinis highlights the broader environmental significance: “Many wetlands in Lithuania have been drained, and monitoring surface moisture is vital for maintaining ecosystem stability and managing emissions. This project bridges the gap between scientific research and practical resource management, providing a reliable tool for aligning industrial activity with environmental protection.”

Smarter Planning, Smaller Footprint

Unfavorable weather conditions have significantly affected peat production this year. Due to high moisture levels, only about 72% of the usual peat volume was extracted in Lithuania. The situation in Latvia was even more severe, with a state of emergency declared after output dropped to just 28% of the average level. Such volatility creates risks for the entire peat supply chain, especially as the Baltic States collectively supply around 40% of the world’s peat used in horticultural and forestry substrates.

“We haven’t seen conditions like this summer’s in about fifteen years. Some companies met their targets, while others reached less than half,” says Giedrius Kavaliauskas, Executive Director of the Growing Media Producers Association (ATGA).

He notes that while AI forecasting won’t change the weather, it can transform how the industry plans and responds:

“Peatlands provide a unique opportunity to study ecosystem dynamics and their link to climate change. This project helps us understand how peatland conditions evolve with the seasons and under increasingly extreme weather. Using remote sensing data, we can create methods that apply to other types of land use – improving efficiency, sustainability, and reducing environmental impact.”

Klasmann-Deilmann project manager Marijus Sandaras adds that the solution will enable real-time monitoring of peatland moisture and faster responses to environmental changes.

“This information is invaluable for planning extraction activities. It helps us assess conditions, react in time, and use our equipment more efficiently. As the climate continues to change, technologies like this are becoming an essential part of sustainable business operations,” he says.

Next Steps: Broader Environmental Integration

The project is currently in its early development phase, with initial results showing about 10% forecast deviation.

“If we can work with the VU team to reduce that to 5%, it will make a real difference for the entire industry,” says the ATGA director.

In the long term, the team plans to go beyond moisture prediction, incorporating additional indicators such as carbon cycle analysis and peatland restoration planning – contributing to smarter, data-driven management of natural resources across Europe.