C‑DRONE
Night aerial view of an illuminated urban area

C-DRONE GUIDE · 5 SEPTEMBER 2026

Restoring river continuity at a weir or mill dam: what a drone survey brings to the project, price

An eighty-centimetre mill weir does not look like a regulated structure. It often has no remaining use, no clearly identified owner and no archives — and yet, if it sits on a watercourse classified in "list 2" under article L. 214-17 of the French environmental code, it carries a legal duty: to be managed, maintained and equipped so that migratory fish and sediment can pass. The river authorities that handle these cases almost always hit the same starting problem: nobody knows exactly what the structure is. What is the real head drop? How long is the impoundment? How much sediment is trapped behind it? What condition is the masonry in? A drone survey does not answer all of those questions, but it answers several of them in half a day on site, with defensible figures and a dated baseline. Here is which ones, how, and at what price.

Published on 5 September 2026, reviewed on 11 September 2026 — regulations in force as of September 2026.

Lists 1 and 2: where the duty comes from, and which structures it hits

Article L. 214-17 of the French environmental code requires the administration to draw up, basin by basin, two lists of watercourses — with different consequences depending on the list.

The timetable is what puts pressure on river managers: list 2 obligations apply to lawfully installed existing structures after a five-year period from publication of the lists. Where the works could not be completed within that period but a file proposing modification of the structure or a change in its operating arrangements has been lodged with the water police services, the owner or operator gets a further five years. In other words, producing a properly evidenced file on time is not an administrative formality: it is what opens the second window.

The European backdrop is Water Framework Directive 2000/60/EC, whose Annex V places river continuity among the hydromorphological quality elements supporting the biological elements of ecological status: high status is defined by continuity undisturbed by human activity, allowing uninterrupted migration of aquatic organisms and sediment transport. Restoring continuity is therefore not a standalone policy: it is one of the direct levers for reaching good status in water bodies.

The scale of the issue has been much better mapped since the European inventory carried out under the AMBER project. A study by Barbara Belletti, Carlos Garcia de Leaniz, Joshua Jones, Simone Bizzi and co-authors, published in 2020 in Nature, counts at least 1.2 million instream barriers across 36 European countries, a mean density of 0.74 barriers per kilometre of river; above all, the authors note that 68% of those structures are less than two metres high and are, for that reason, largely missing from national inventories (see the study on Google Scholar). Those are precisely the weirs now landing on French river authorities' desks.

What a flight documents at a weir, its impoundment and its longitudinal profile

A transverse river structure is awkward to survey from the ground: the crest is often unreachable without entering the water, the downstream face is buried in vegetation, and the impoundment stretches several hundred metres between wooded banks. Two to three hours of flying, by contrast, produce a complete geometric dataset.

Two field precautions govern the quality of the result. The first is season: in full leaf, the riparian canopy hides the bank and the toe of the structure; a winter or late-autumn flight at low flow yields a far more complete model. The second is vertical control: without surveyed ground control and independent check points, a photogrammetric model stays internally consistent but vertically adrift — useless for comparing two campaigns or for tying in a water surface profile. What each deliverable is worth is set out in our guide to the deliverables of a drone photogrammetry mission.

Trapped sediment: what the drone quantifies, what it cannot see

The question that usually decides the chosen scenario — full removal, partial lowering, gate opening, fish pass — is the sediment stock trapped in the impoundment: its volume, its grain size, and any contamination inherited from upstream industry. The drone answers only part of that question, and a regulatory file should say so plainly.

What it quantifies. Once deposits are exposed — impoundment drawn down, partial emptying, severe low flow, or after works — photogrammetry measures the volume of the sediment bars above a reference surface, with exactly the same mechanics as a stockpile volume calculation. On a deposit of a few thousand cubic metres, the relative uncertainty of a properly controlled model stays around a few per cent — ample for sizing a removal contract or for choosing between dredging and progressive remobilisation. It also documents the area and position of the deposits, which drives the choice of sampling points for quality analysis.

What it cannot see. Underwater, nothing. Aerial photogrammetry does not reliably reconstruct a submerged bed: refraction, turbidity and specular reflection degrade the model to the point of uselessness. While the impoundment holds water, the trapped volume is a bathymetric survey job — single-beam echo sounder, surface vessel or, in shallow clear water, the methods described in our guide to drone bathymetry of lakes and reservoirs. Nor does the drone say anything about grain size at depth, deposit cohesion or contamination: that remains coring and laboratory work. Finally, it replaces neither a hydraulic study — water surface modelling, the effect of drawdown on intakes and riverside foundations, sediment transport after remobilisation — nor a structural diagnosis of whatever part of the barrier is kept.

The sequence that works in practice is a combined one: bathymetry of the full impoundment for total volume, then a drone survey after drawdown to check what was actually remobilised.

Before and after works: evidencing the ecological gain

A continuity restoration project is ultimately judged on one question: did it work? Funders ask it, and so do residents who opposed the scheme. The answer rests on comparable baselines — same extent, same method, same control — before and after works.

That is exactly what a study by Alexandra D. Evans, Kevin H. Gardner, Scott Greenwood and Brett Still, published in 2022 in the journal Drones, put to the test: on the Bellamy River reservoir in New Hampshire, the authors compared drone and SfM photogrammetry monitoring against conventional field surveys, before and after a dam removal, and mapped vegetation recolonising the exposed sediment using machine learning, with overall accuracies of 83% to 100% depending on class. Their practical conclusion speaks directly to a river authority: the drone's continuous spatial coverage reveals geomorphic and vegetation changes that point-based ground surveys miss, at a cost compatible with repeated monitoring (see the study on Google Scholar).

The biological side of the gain is not measured by drone — it is measured with nets, traps and tagging. The order of magnitude a removal can deliver is nonetheless worth recalling in a file: on the Villestrup river in Denmark, a twelve-year study by K. Birnie-Gauvin, M. M. Candee, H. Baktoft, M. H. Larsen, A. Koed and K. Aarestrup, published in 2018 in River Research and Applications, recorded a brown trout smolt run rising from 1,660 individuals before any removal to 19,105 after six weirs had been taken out of the same system (see the study on Google Scholar). The drone does not establish that result: it establishes the physical state of the site that allows the result to be attributed to the works rather than to hydrological luck.

In practice a monitoring protocol comes down to three identical campaigns: one before works (baseline), one immediately after (as-built state, checking design levels), one after a year or two (vegetation recovery, bed stabilisation, residual sediment bars). The comparison only holds if all three share the same physically marked control points and the same processing chain — a subject we cover in our guide to riverbank and watercourse monitoring under the GEMAPI mandate, where the same repeated-campaign logic applies along the channel.

Feeding the national barrier register, building the file, and 2026 costs

The national register of obstacles to flow (ROE) is the participatory scheme coordinated by the French Office for Biodiversity that centralises the inventory of structures obstructing surface water flow; the data is published through the national water data service. It now exceeds 100,000 recorded obstacles across French territory, overseas regions included. Three types make up 98% of them: instream weirs, bridge-related obstacles and dams — and among those, overflow-type weirs are the most common, at around 30% of described obstacles.

Its gaps are documented, and they are exactly what a survey fills: head drop is recorded for only 62% of obstacles, and the proportion fitted with a fish passage device is particularly low — around 2% to 4% — for structures whose recorded head drop is under one metre. A drone round over a dozen structures in a single catchment produces, for each one, a dated geolocated photograph, a crest elevation, a measured head drop and an impoundment extent: enough to update a register entry with verifiable rather than estimated attributes, and enough to rank priority structures before commissioning detailed studies.

On funding, the French water agencies support both scenario-definition studies and removal or lowering works on obstacles, public or private, under their multi-year intervention programmes. Rates and eligibility conditions vary from basin to basin and from programme to programme: what is fundable must be checked with the competent agency, and topographic survey usually sits within the scope of supported studies. Where a local authority is the contracting party, our guide to public procurement of drone services sets out what to require in the specification — ground control, processing report, contractual tolerances, deliverable formats.

Orders of magnitude observed in 2026 (excl. VAT), derived from comparable photogrammetry and aquatic-environment monitoring work:

These ranges assume an accessible site, a workable weather window and no unusual airspace constraint; a structure sitting under restricted airspace, inside a national park core area or in a dense urban setting prices differently. For a specific project — one weir, a string of structures on a tributary, a pre-works campaign — request a quote stating the location, the watercourse classification, the estimated head drop and the planned emptying or drawdown date.

Frequently asked questions

Can a drone measure the head drop at a weir?

Yes, provided the crest and the downstream toe are visible above water at the time of the flight. A photogrammetric model tied to GNSS-surveyed ground control points reconstructs the drop between the crest and the downstream water surface to within a few centimetres. That matters: nationally, the French barrier register records a head drop for only 62% of listed structures. The drone measures a surface elevation, however, not a bed elevation: the depth of the plunge pool at the toe requires sounding or in-water reconnaissance.

Can a drone estimate the volume of sediment trapped in an impoundment?

Partly, and on one condition: the impoundment must be drawn down or emptied at the time of the flight so the deposits are exposed. Photogrammetry then measures the emerged volume above a reference surface, exactly as for a stockpile survey. While the water level is up, the water hides the deposit: aerial photogrammetry does not reliably see through the water column and a bathymetric survey is needed. Many projects combine both: a survey of the full impoundment before works, a drone survey after drawdown.

Does a drone survey replace a river-engineering consultancy study?

No, and presenting it that way in a regulatory file would be a mistake. The drone produces geometric data — orthophoto, digital terrain model, longitudinal profile, volumes, close-range masonry imagery. Hydraulic interpretation (water surface profile, fish passability, sediment transport, the effect of a drawdown on riverside structures and water intakes) and structural diagnosis belong respectively to a river-engineering consultancy and a structural engineer. The survey feeds their work; it does not stand in for it.

Request a free quote

Put it into practice

Also worth reading