FrameTest Webinar
- Taskscape Associates
- Sep 2
- 4 min read
Updated: Sep 11
Exploring a New Support Tool for Agricultural Decision Makers...
On August 28, 2025, the project hosted an interactive workshop on FrameTest - an innovative decision support tool designed to help policymakers and stakeholders compare agricultural management options. The webinar was part of an event series surrounding FRAMEwork's knowledge platform, Recodo.
Led by Dr Claudio Petucco and Dr Gustavo Martin Larrea Gallegos from LIST (Luxembourg Institute of Science and Technology), with support from Taskscape, the 90-minute session provided both theoretical grounding and practical demonstration of this new resource. Read on for a full deep dive, or if you're short on time find the key takeaways summarised on Recodo's LinkedIn page.
Navigating Complexity in Agro-ecological Systems
The workshop opened with context about Recodo and the H2020 FRAMEwork project's broader mission to support farmer clusters across Europe in implementing biodiversity-sensitive agricultural practices. This is especially important as the EU's latest Common Agricultural Policy (CAP) has increased the level of organic farming and environmentally sensitive measures required to combat significant biodiversity loss and food system resilience challenges.
As Dr Petucco explained, assessing these systems requires understanding multiple interconnected layers: ecological interactions (how biodiversity responds to or works within farming practices), social networks (how farmers exchange information), and economic factors (supply chains and input changes). Dr Petucco highlighted the lack of explicit models and data to predict outcomes when implementing such complex systems and outlined how:
"The motivation was really to try to combine and understand all these sociological, economic, and ecological interactions"
Interested in these issues? Explore more project webinars!
Making the Complex Accessible
FrameTest emerged from the need to bridge the gap between sophisticated agent-based modeling (ABM) and practical policy decision-making. The tool operates as a "surrogate model" or "meta-model" that provides rapid predictions based on extensive simulations from the underlying Frame-ABM system.
Dr Gustavo Martin Larrea described the technical foundation:
"Frame-ABM is a highly customizable model but it requires lots of parameters. For this reason, Frame-Test is built upon the results of the agent-based model using machine learning techniques to make the tool more accessible.”
The result is what Gustavo likened to "a mixing console in which you tune each one of the knobs and you get a prediction of a particular KPI or key performance indicator" – all without requiring coding skills or advanced computer literacy.
Real-World Applications
The team presented three compelling use cases that guided FrameTest's development:
Comparing Approaches: Testing whether farmer cluster approaches outperform farm-by-farm actions for specific goals like improving landscape connectivity (e.g, habitats)
Integrated Pest Management: Analyzing trade-offs between natural predation and pesticide use to reduce chemical inputs without unacceptable yield risks
Cost-Effectiveness Analysis: Determining which ecosystem interventions (flower strips, hedgerows, or combinations) deliver the most biodiversity per euro invested
Live Demonstration
During the live demo, participants saw FrameTest's intuitive web interface in action. The tool guides users through setting up scenarios with plain-language questions about:
Farming landscape characteristics (field sizes, aggregation patterns)
Farmer behaviors (tolerance to losses, response to dissimilarity)
Economic incentives (subsidy amounts for different conservation measures)
Ecosystem service efficiency (pest control effectiveness, pollination impacts)
The Explorer feature then allows real-time visualization of how changing any parameter affects multiple outcomes – from biomass production to life cycle impacts and economic performance.
Engagement and Future Development
The workshop attracted diverse participants, with Simone Martino from The James Hutton Institute and Iris Bohnet from the Czech University of Life Sciences contributing thoughtful questions about data requirements, calibration processes, and adaptability to different regional contexts. Theo Simmons, Taskscape, also raised interesting possibilities for private sector applications, based on questions fielded while representing Recodo at conferences - such as businesses looking to support groups like Farmer Clusters through corporate offsetting, with a need to assess and report on such efforts.
Overall, Q+A discussions centered on the tool's flexibility. While currently calibrated for Luxembourg's arable farming systems, the underlying framework can be adapted to other contexts, as Dr Petucco explained:
"The customization of Frame-ABM... has an initial time cost, but then it's very easy to adapt to local conditions."
Availability and Next Steps
The team announced plans to release both Frame-ABM and FrameTest as open-source packages through the Recodo platform. This will enable researchers and practitioners to:
Train their own FrameTest models for specific agricultural systems
Customize the underlying agent-based model for regional contexts
Expand the tool to include additional crops and farming systems
Resources and Further Information
For those interested in exploring FrameTest further or adapting it to their context, the team welcomes inquiries. For more information about FrameTest, visit Recodo where you can access the tool and contact the development team directly. As they emphasised, while the tool provides sophisticated modeling capabilities, its current value lies in making complex agricultural system analysis accessible to policymakers and stakeholders who need evidence-based insights for decision-making.
The workshop materials, including slides and supporting documentation will be published to YouTube and made available to webinar registrants and online visitors via the project's legacy pack in Autumn 2025.
