C-DRONE GUIDE · 7 SEPTEMBER 2026
Mapping and monitoring a mangrove by drone: Guiana, Mayotte, the Antilles, New Caledonia
A mangrove is not visited, it is endured. Soft mud you sink into up to the thigh, red-mangrove stilt roots that make a normal stride impossible, a tide that closes the way out, mosquitoes, caimans in French Guiana: a ground survey advances a few hundred metres a day, along a transect that will never describe the whole stand. That is exactly the kind of environment where flying changes the nature of the work. France manages 87,905 hectares of mangrove, spread across fifteen islands and one continental territory, French Guiana; Guiana and New Caledonia hold more than 90% of it, while Mayotte, with 667 hectares along 29% of its coastline, carries some of the tightest erosion and restoration stakes in the country. Here is what a drone actually measures there, for whom, and where the measurement stops.
Published on 7 September 2026, reviewed on 7 September 2026 — regulations in force as of September 2026.
Where France's mangroves are, and under what status
A mangrove is an intertidal forest of mangrove trees growing on mud between land and sea. It delivers three services few ecosystems combine: it damps swell and holds the shoreline — in Mayotte, the marine nature park recorded that reefs and mangroves reduced the effects of the swell during Cyclone Chido; it acts as a fish nursery, a breeding and growth zone for many harvested species; and it stores an exceptional volume of carbon, so-called blue carbon.
According to the SDES publication on key sea and coastal figures (2024 edition, data from the tropical wetlands resource centre, 2022), French mangroves cover 87,905 hectares, across fifteen islands and one continental territory. Contrary to a widespread assumption, Mayotte and Guiana are not the dominant pair: according to the ministry in charge of the Sea, French Guiana and New Caledonia alone hold more than 90% of French mangroves. Guiana has roughly 700 km² of mangrove covering close to 80% of its coastline, or 68% of France's overseas mangroves (Guiana water and biodiversity committee). Mayotte, by contrast, has only 667 hectares — 1.8% of the island's area — but spread along 76 km of coast, 29% of the Mahoran shoreline, with seven mangrove tree species present (Mayotte coastal observatory). Condition varies sharply: good in Guiana, on Europa, Juan de Nova, Wallis and across much of New Caledonia; unfavourable in Mayotte, Saint-Martin and Saint-Barthélemy; increasingly degraded in Martinique, Guadeloupe and around Nouméa.
On legal status, most mangroves fall within the public maritime domain, which governs both intervention permits and land control. In 2022, 58% of national mangroves were covered by a protection measure, 54,712 hectares excluding French Polynesia: entirely in the Indian Ocean, around 60% in both the Pacific and the Atlantic (SDES). The instruments overlap — Conservatoire du littoral sites, national nature reserves (Amana in Guiana, 14,400 hectares, created in 1998), biotope protection orders, issued by the prefect on land and river domain and by the minister for the sea on the public maritime domain — and Ramsar sites: Guadeloupe's Grand Cul-de-Sac Marin, listed in 1993, covers around 15,000 hectares including over 6,000 of mangrove, the widest belt in the Lesser Antilles; Mayotte's Vasière des Badamiers has been listed since 2011; the Amana reserve sits within the Basse-Mana site, around 59,000 hectares. Each of these statuses creates an obligation to know and to monitor — and that is where the drone comes in.
The real argument: terrain nobody can walk
The case for drones in mangroves is not modernity, it is physical impassability. The substrate is soft, waterlogged mud; red mangroves put out stilt roots that form a hip-high tangle; access depends on the tide, and the way out can close within tens of minutes. Add mosquito pressure, heat, and in Guiana the presence of caimans. An ecologist in waders moves slowly, takes a real risk, and brings back at best a few plots and one transect. Never a map.
A flight changes the scale: in a single morning at low tide you cover an entire mangrove unit and come away with a georeferenced orthophoto, a digital surface model and homogeneous coverage — the same data everywhere, including where nobody has ever set foot. In practice managers get:
- the exact extent and its change from one campaign to the next: advance or retreat of the front, gap opening, colonisation of a new mudflat — a major issue in Guiana, where Amazonian mud banks migrate westward and alternate accretion and erosion on a cycle estimated at around thirty years;
- a canopy height model from photogrammetry, separating low pioneer mangrove from mature stands;
- identification of dieback zones, grey or thinned canopy patches, often linked to a salinity shift, a blocked water exchange or pollution;
- monitoring of restoration and replanting works, with a baseline and comparable annual passes;
- quantification of cyclone damage: after Chido, on 14 December 2024, with gusts above 220 km/h, close to 80% of Mayotte's mangrove trees were defoliated and many broken — quantifying that from the ground, at that scale, was out of reach;
- detection of waste dumping and illegal infill along the fringe, a recurring dispute on the public maritime domain;
- production of defensible baseline surveys for an impact assessment or an offset file.
The principle is the one already described for coastal erosion monitoring by photogrammetry and for Posidonia seagrass monitoring in the Mediterranean, transposed to an environment where the obstacle is not water but vegetation and mud. Overseas, logistics and the local airspace framework add their own constraints, detailed in our guide to drone missions in France's overseas departments.
What the measurement really gives — and its three blind spots
On canopy height and above-ground biomass, the literature converges. A team led by Ali Karimi published in July 2025, in the journal Forests, a method for estimating above-ground biomass and blue carbon in Avicennia marina stands by combining canopy heights extracted from a drone survey with field-measured tree heights, explicitly framed as a low-cost tool for conservation and management (see the study on Google Scholar). The drone does not measure carbon: it measures a height, which allometric equations calibrated on field data convert into biomass and then carbon. Without that field calibration, there is no usable result.
On telling mangrove species apart, a team led by Yuanzheng Yang published in 2024, in the journal Remote Sensing, a fine-scale classification from drone multispectral and hyperspectral imagery: with hyperspectral data, overall accuracy exceeds 91% across four machine-learning models and reaches 97.15% with LightGBM (see the study on Google Scholar). The practical message cuts both ways: it is achievable, but it requires a suitable sensor and ground-truth surveys to train the model. With a plain RGB camera you get an extent and height map, not a species map.
Hence the three blind spots to state before signing:
- Soil carbon is entirely beyond the drone. Yet it dominates: work by Daniel C. Donato and colleagues, published in 2011 in Nature Geoscience, shows that soil organic matter accounts for 50% to more than 90% of a mangrove's total ecosystem carbon, biomass included (see the study on Google Scholar). A credible blue-carbon assessment requires soil cores and laboratory analysis. Root biomass is in the same category.
- Species are not readable by default. Without a multispectral or hyperspectral sensor and reference plots, species-level classification is not reliable — and it stays difficult under closed canopy or in mixed stands.
- The drone does not replace monitoring plots. Trunk diameter, natural regeneration, interstitial salinity, benthic fauna, actual seedling survival: none of that is visible from above. The flight densifies and generalises what plots measure; it does not stand in for them.
It is the same architecture as for any ecological drone monitoring programme: a continuous drone layer, calibrated on a small network of ground-measured points.
Who commissions this kind of mission, method and 2026 prices
Buyers are primarily public or semi-public: the Conservatoire du littoral, owner or assignee of many overseas mangroves; state environment services policing the public maritime domain and processing applications; marine nature parks and reserve managers; local authorities and inter-municipal bodies facing erosion at the foot of urban areas; ecological consultancies producing baselines and monitoring; and restoration or offset project owners, whose follow-up falls under the logic described in our guide to ecological compensation monitoring.
On method: fly at low tide, use a double-grid flight plan for photogrammetry, place GNSS control points on firm fringes since mud rules out setting any inside the stand, and above all repeat at comparable dates and tides — a change map is only worth something if both campaigns were acquired under equivalent conditions. In protected areas, the manager's agreement and respect for sensitive bird-breeding periods constrain the flight window as much as the weather does.
Ranges observed in 2026, excl. VAT, for an overseas intervention:
| Service | Range (excl. VAT, 2026) |
| Mapping a mangrove unit, orthophoto + extent (up to 50 ha) | €1,200 to €3,000 |
| Full photogrammetry with canopy height model | €2,000 to €4,500 per site |
| Annual monitoring of a restoration project (baseline + 2 passes) | €2,500 to €6,000 per year |
| Post-cyclone damage assessment, emergency mobilisation | quote-based, by extent and lead time |
These amounts cover flight preparation, acquisition, photogrammetric processing and delivery of GIS layers. They exclude soil cores, laboratory analysis, field plots, and inter-territory travel when the operator is not locally based — three line items to budget separately, and which can exceed the cost of the flight itself. On inland wetlands, the approach matches the one set out in our guide to wetland mapping for a water-law file. For a managed site, an impact assessment or a replanting programme, request a quote stating the territory, the area, the site's protection status and the monitoring frequency you need.