Skip to main content
Go to the home page of the European Commission (opens in new window)
English English
CORDIS - EU research results
CORDIS

BeeOmetrics: an AI-powered predictive platform based on wild bee activity for targeted environmental management.

Periodic Reporting for period 1 - BeeOmetrics (BeeOmetrics: an AI-powered predictive platform based on wild bee activity for targeted environmental management.)

Reporting period: 2024-03-01 to 2025-02-28

CONTEXT

There is a growing demand for accessible environmental monitoring tools, at farm or site level. That demand comes from organisations across all sectors of activity, may they be private companies (e.g. farmers, land developers) or public administrations (e.g. municipal councils) that oversee small parcels of land (up to 100ha) and have the power to make decisions that affect the natural environment within these parcels, i.e. "environmental managers".

These tools need to be cost-effective while delivering dependable data for accurate analysis of specific smaller areas. Remarkably, such a specialised tool is not available in the market of environmental monitoring solutions (20B€ in 2022, +6.30% CAGR 23-28).

Environmental monitoring solutions are observational techniques and tools (e.g. sensors, wireless communications) that detect, observe, and measure environmental conditions at a specific site or location. The current state-of-the-art solutions focus on measuring a few key indicators (e.g. presence of specific pollutants, name of most abundant plant species) measured directly by sensors or biochemical analysis in only one medium (i.e. air, soil or water)

Several of these solutions, such as satellite imaging, are not adapted to delivering environmental data that is relevant enough for a small (≈40ha) parcel. Current solutions that can deliver meaningful information on small land parcels collect data from only one type of observational tool in one medium.

Since each ecosystem is formed by several mediums where conditions are interrelated to each other, environmental managers would need several of these solutions to fully comply with their monitoring and reporting obligations. This means incurring significant capital investment (e.g. purchase of detecting device, training personnel in manual measurements) and/or ongoing operating costs (e.g. maintenance of the data source devices, manual on-site measurement). Moreover, the lack of integrated monitoring and analysis of different key indicators across all mediums of an ecosystem does not allow environmental managers to reliably assess the origin of environmental issues within their area of control or to evaluate the true impact of their decisions to improve the state of the ecosystem.

OVERALL OBJECTIVES

The BeeOmetrics project develops an ecosystem health monitoring system that supports end-users in environmental decision-making within the areas they control (down to 40ha). BeeOmetrics will overcome the limitations of current state-of-the-art approaches by delivering reliable KPIs of overall ecosystem health and providing predictive recommendations for improvement. BeeOdiversity seeks to advance the BeeOmetrics system from TRL4 to TRL6 within three years. This advancement involves the transformation of BeeOmetrics into a reliable predictive platform through rigorous testing at selected validation sites. Throughout the BeeOmetrics Transition project, we will accumulate pollution and environmental DNA (eDNA) data across an entire trophic chain (including soil, plants, pollen, and pollinators) from two different mediums (i.e. soil and air) and from multiple locations to significantly expand our databases and train our Machine Learning (ML) models. Our ultimate objective is to develop and validate a predictive platform tailored to different types of small-scale environmental managers, establishing BeeOdiversity as the benchmark in the environmental monitoring sector. Practically speaking, BeeOmetrics will offer a a predictive ecosystem health monitoring system that provides a reliable assessment of ecosystem health in small parcels (down to 40ha), with the sole need of installing on-site a cost-effective nature-based detecting device, the BeeÔtel (see uploaded BeeÔtel image).


PATHWAY TO IMPACT
Indicators and KPIs are the cornerstone of any decision-making process concerning biodiversity. BeeOmetrics will make environmental monitoring accessible for land managers and policy makers,
enabling many professionals, who previously lacked the means to monitor biodiversity and pollution (heavy metals, PFAS, dioxins, etc.) on their land/sites, to take decisions accordingly. Our solution
is low-tech, low-cost (compared to e.g. naturalistic surveys) and green (sheltering local wildlife). It provides the data and information needed by land managers. This could not only be a game changer in the day-to-day environmental management, but also help companies set credible nature positive roadmaps.

Additionally, BeeOmetrics helps EU market participants comply with global and EU green regulations, such as the EU Taxonomy and CSRD, enhancing their competitive edge. By streamlining monitoring processes and supporting regulatory compliance, BeeOmetrics creates measurable economic value while promoting sustainability and long-term resilience.

Finally, BeeOmetrics will support the achievement of several interrelated goals and targets of the European Green Deal:

(i) "Preserving and restoring ecosystems and biodiversity".
(ii) "From 'Farm to Fork': a fair, healthy, and environmentally friendly food system".

The BeeOmetrics platform will also support the EU Pollinator Initiative and its recent revision, the EC's "A new deal for pollinators" (24/1/2023)
During the 12 first months of the project (periodic report #1), the work performed can be summarized as follow:
• Prospecting and onboarding more than 25 participants
• Selecting 49 different sites for BeeÔtel installation and data collection.
• Ordering BeeÔtel from local production partners,
• Following up on delivery to site and manage installation with local partners
• Collect and first interpretation of data from the BeeÔtels installed (pollen samples & pictures of nest orifices).
• Development of an image recognition ML algorithm to extract information (i.e. occupancy rate, species of solitary bees present) from the pictures of the BeeÔtel.
• Data engineering to optimise datasets for model training: ingest data (clean, format, ensure
BeeÔtel
BeeÔtel
My booklet 0 0