Satellite, Drone, or Field Data? Designing Scalable Nature Monitoring Systems
Session Description
This session ran a hands-on design exercise built around Nabat.ai's mangrove restoration and monitoring work in the UAE. Attendees split into groups covering five measurement dimensions, seed and sapling success, vegetation health, carbon and biomass, biodiversity, and co-benefits such as water quality and community impact, and worked through which data sources they would choose to measure project outcomes, first optimising purely for accuracy and then under a second, resource-constrained scenario simulating a larger project in an unfamiliar location with limited field access.
Speakers
Taha Ghaznavi, Chief Product Officer, Nabat.ai
Mehdi Ajana, Head of Strategy, Nabat.ai
Vanessa Randon, Tech Ecology Manager, Nabat.ai
Watch the Session Recording
Key Takeaways
Groups generally converged on hybrid approaches combining satellite or drone data with field-based verification, rather than relying on a single data source, particularly for vegetation health and carbon and biomass measurement.
Biodiversity was consistently identified as the hardest dimension to measure remotely, with groups discussing proxies such as mangrove cover and habitat fragmentation, or working through local community members with knowledge of the landscape rather than direct field surveys.
Seed and sapling monitoring was seen as heavily reliant on drone-based sub-centimetre imagery, with groups noting that access to physical quadrats and plots would be preferable but not always feasible.
For co-benefits like water quality and community impact, groups agreed that satellite and drone data could not substitute for direct engagement, citing the need for surveys, local authority data, and in-person conversations with communities.
Under the resource-constrained scenario, groups shifted toward stratified spot-checking rather than comprehensive coverage, greater reliance on secondary or citizen-science data, and partnering with local groups already present in the landscape.
Several groups noted that ground-truthing requirements are often set by funders or reporting obligations rather than the project team's own preferences, particularly where restoration outcomes are tied to credit programmes that require an audit trail.
Seed monitoring was flagged as the hardest area to substitute for when drone access was removed, with machine-learning-based extrapolation from comparable sites offered as the least imperfect alternative.
