C-DRONE GUIDE · 2 SEPTEMBER 2026

Drone water sampling: equipment, laboratory acceptance and pricing

On a lined lagoon, a flooded quarry pit, a retention basin with unstable banks or a dam reservoir, reaching the sampling point often costs more than the analysis itself: launching a boat, hiring a rope-access technician, rigging a lifeline, two operators and a whole morning for one litre of water. A drone fitted with a suspended sampler brings back that litre in minutes, with nobody standing at the water's edge. The question is not whether the aircraft can do it — the scientific literature has validated that, including at great depth — but whether the sample it brings back is admissible: a sample intended for regulatory analysis is worth what its traceability, containers, preservation and the qualification of the sampler are worth, not how it travelled. Here is the equipment actually carried, what laboratories accept, the three cases where they will refuse, the flight framework applicable in September 2026, and observed prices.

Published on 2 September 2026, reviewed on 2 September 2026 — regulations in force as of September 2026.

When the boat costs more than the analysis

Drone water sampling is not justified at a discharge point reachable from a walkway: there, an operator with a telescopic pole does the job in three minutes. It is justified when physical access becomes a project in itself. Four families of situations recur:

• The lagoon or lined basin of an industrial (ICPE) site. Walking on a geomembrane risks puncturing it; putting a boat on it means mobilising two people and a formal work-near-water procedure. The representative sampling point is often in the centre, beyond the reach of a pole.
• The flooded quarry pit or gravel pit. The banks are working faces: unstable walls, scree, no launching ramp. This is precisely the context of mine pit lakes, where the aerial method has been scientifically validated.
• The mudflat, marsh or wetland. The ground carries neither a person nor a quad bike, and a boat runs aground. The distance to cover is only a few dozen metres, which makes deploying a vessel absurd.
• The dam reservoir, bathing lake or stormwater basin. Here the problem is surface area: several sampling points far apart, half a day of navigation for one set of samples.

The trade-off is economic before it is technical. A two-operator waterborne intervention with a boat routinely amounts to half a day or a full day of billing, on top of which come the prevention plan, flotation equipment and, on an industrial site, a work-near-water permit. A flight is counted in minutes per point. The gain grows with the number of points: that is the method's main argument.

The drone does not remove the need to know where to sample. A preliminary reconnaissance flight — the one described in our guide to cyanobacteria and algal bloom monitoring on lakes — locates the densest scum, an incoming discharge or the downwind accumulation zone, and avoids the classic false negative of a sample taken from clear water 200 metres from the anomaly.

Four families of airborne samplers, and what each can do

There is no such thing as a "sampling drone" — there is a carrier aircraft and a payload. Four set-ups cover most needs, with very different levels of admissibility.

1. The weighted bottle on a line. The simplest arrangement: a laboratory-supplied bottle, weighted, lowered beneath the aircraft on a cord or winch, immersed for a few seconds at the surface, then raised. Advantage: the analysis bottle is the only container in contact with the water, which removes any question of cross-contamination through tubing. Limitation: surface sampling only (the top few decimetres), volume capped by payload, and incomplete filling if immersion is too brief. This is the set-up Horricks and co-authors used for marine bacteriology.

2. The valve or messenger-triggered sampler, a miniature Niskin. A bottle open at both ends is lowered and closed at the chosen depth by a remotely triggered mechanism or a sliding weight. This is the only set-up allowing sampling at a selected depth, an essential condition whenever the target parameter is stratified — dissolved oxygen, dissolved iron and manganese in a hypolimnion, ammonium at the bottom of a reservoir. Castendyk, Voorhis and Kucera validated this approach down to 92 m in a mine pit lake.

3. The onboard peristaltic pump. A flexible tube is lowered from the hovering aircraft and a pump fills one or more bottles. Advantages: larger volumes, several containers filled at the same point, and the option of flushing the circuit. Major drawback for admissibility: the water travels through tubing, and that tubing becomes a critical point — material compatible with the target parameters, flushing between points, and a field blank to be analysed to prove the absence of contamination. A serious laboratory will ask for it.

4. The suspended multiparameter probe. This is not a sampler: the probe measures temperature, pH, dissolved oxygen and conductivity in situ, bringing nothing back. It is nonetheless essential, because those four parameters degrade in a bottle and must be read on site — the French AQUAREF lake sampling guide explicitly classes them as field measurements. It also profiles the water column before deciding at what depth to trigger the sampler.

In practice a serious campaign combines set-ups: probe for profiling and labile parameters, valve sampler or weighted bottle for the analysis sample, and a pump only when volume demands it and the parameters tolerate the circuit.

What the payload really allows: mass, volume, number of points

One litre of water weighs one kilogram. That obvious fact governs the whole sizing of a campaign, and it is the point on which commercial promises most often trip up.

A common professional multirotor in France — the DJI Matrice 350 RTK, for instance — is rated at a maximum payload of around 2.7 kg for an empty mass of about 6.5 kg with its two batteries. Out of those 2.7 kg, the sampling mechanism, winch, cable and weight routinely consume 0.7 to 1.2 kg. In practice that leaves 1.5 to 2 litres of water per flight at the very most, and often less if a safety margin for wind is to be kept. The study by Graham and co-authors, published in 2022 in Science of The Total Environment, illustrates the threshold well: to bring back 2 litres of water per flight across six lakes in western Ireland, the authors had to use a carrier in the Matrice 600 Pro class, whose maximum payload reaches 6 kg — an aircraft markedly heavier and bulkier than those used in earlier studies.

The operational consequence is direct. A reasonably complete analysis report calls for several containers: a plastic bottle for general parameters, an amber glass bottle for organics, an acidified bottle for metals, a sterile bottle for bacteriology. The total frequently exceeds 2 litres. One sampling point then requires two flights, sometimes three. A provider promising ten complete points in one morning with a single aircraft and a single battery is not telling the truth: with a carrier of this class, expect rather one complete point every fifteen to twenty minutes, including ground handling, labelling, cool-box storage and battery changes.

Two often-forgotten corollaries. First, the flight is flown with a variable load: the aircraft takes off light and comes back heavy, which changes its behaviour and consumption — a point to include in the mass and power budget, as for any cargo drone lifting operation. Second, the load hangs on a line: it swings, it can snag, and the aircraft must stay high enough for the pilot to retain control if the sampler jams in a stump or a net.

What the laboratory looks at: the chain, not the aircraft

This is the point most sales presentations gloss over. No French text either provides for or prohibits sampling by aircraft: the question is never framed that way. What is regulated is the sampling chain, and it is regulated strictly.

The methodological foundation is the NF EN ISO 5667 series, "Water quality — Sampling", of which four parts directly concern an aerial campaign:

• Part 1 — Guidance on the design of sampling programmes and sampling techniques (ISO 5667-1:2020): this is what requires the objective to be defined before the equipment, and the representativeness of the chosen point to be justified.
• Part 3 — Preservation and handling of water samples (2024 edition): container type, preservative, temperature, darkness, maximum time between sampling and analysis, parameter by parameter. This is the most demanding part for a field operation.
• Part 4 — Guidance on sampling from lakes, natural and man-made (ISO 5667-4:2016): the reference for a lake, reservoir or flooded pit.
• Part 6 — Guidance on sampling of rivers and streams (ISO 5667-6:2014), for a discharge into running water.

On the laboratory side, regulatory water analyses fall under approval granted by the French minister for the environment, the terms of which are set by the order of 26 June 2023 on the approval of laboratories carrying out analyses in the field of water and aquatic environments under the Environmental Code. Approval is granted for a maximum of five years, parameter-matrix pair by parameter-matrix pair, and the list of approved laboratories is published on the ministry's LAB'EAU portal. Checking that your laboratory is approved for the specific parameter and matrix, and not merely "approved" in general, is part of the preparatory work.

On an ICPE (regulated industrial) site, the text goes further still. Article 58 of the order of 2 February 1998 — the so-called "integrated" order — provides that where the operator uses external bodies to monitor its discharges, it must ensure that every party in the sampling and analysis chain is approved or accredited by COFRAC or by a signatory to the European multilateral agreement. In other words: within that scope it is not only the laboratory that must be qualified, it is also the sampler. A non-accredited remote pilot cannot on their own perform an enforceable self-monitoring sample; they act as a means of implementation, under the responsibility and in the presence of the accredited sampler who conditions, labels and signs the sampling record. That arrangement should be spelled out in the quotation and validated with the inspectorate before the campaign.

Finally, field traceability is exactly the same as for sampling on foot: containers supplied by the laboratory, rinsing performed or deliberately omitted depending on the parameter, preservation cold and in the dark, a cool box with a temperature logger, and a time-stamped sampling record naming the operator, the point coordinates, the depth, weather conditions and in-situ measurements. The French AQUAREF lake sampling guide (version 2, 2022) sets out these requirements for lakes and reservoirs: an integrated euphotic-zone sample taken with a standard bottle, laboratory-supplied containers, cold chain, and systematic field measurements.

Three cases where drone sampling is not admissible

An honest provider will tell you where their method stops. Three situations fall outside its scope, and it is better to know them before ordering the campaign than in front of the inspectorate.

1. Volatile parameters. Volatile halogenated organic compounds, BTEX, light hydrocarbons, sulphides, and more generally any parameter requiring a bottle filled to the brim with no air bubble: part 3 of the ISO 5667 series requires filling with no headspace, which a bottle immersed for a few seconds on a line, hauled up vibrating beneath an aircraft and set down on the ground, cannot guarantee. Degassing begins during the ascent. For these parameters, manual sampling with immediate conditioning remains the reference; an aerial sample has value only as a screening indicator, never as an enforceable measurement.

2. Regulatory bacteriology. The picture is more nuanced than commonly assumed. Horricks and co-authors showed in 2022, in Environmental Monitoring and Assessment, that bacterial counts from drone-collected samples were not significantly different from those obtained by vessel, using sterile 250 mL bottles supplied by an accredited laboratory and immersion at 20 cm. Technically, therefore, the method holds. But two obstacles remain in France. First, ISO 5667-4 on lakes explicitly excludes guidance on sampling for microbiological examination from its scope: one must then rely on other parts of the series and on the laboratory's own requirements. Second, and above all, the sanitary control of bathing waters follows a protocol and uses a sampler designated by the health authority: a drone sample is not intended to replace it. Its legitimate place is operator self-monitoring, upstream of or alongside official control.

3. Standardised depth sampling, and the 24-hour composite sample. Two distinct limits combine here. The first is physical: lowering a sampler to a controlled depth requires a winch, cable length and hover stability that few configurations achieve — the 92 m reached by Castendyk and co-authors on the Thompson Creek pit lake in Idaho is a research result obtained with equipment and a method specifically validated for mine pit lakes, not a commercial standard. The second is regulatory and often decisive: for the aqueous effluents of an ICPE site, article 60 of the order of 2 February 1998 provides that, unless otherwise specified, limit values apply to samples, measurements or analyses averaged over twenty-four hours. No grab sample, aerial or otherwise, satisfies that requirement: it calls for a refrigerated automatic sampler stationed at the discharge point. The drone then remains relevant for characterising a lake, lagoon or basin — not for replacing the 24-hour composite sampler on a piped discharge.

Conversely, the method is entirely at home on general, stable physico-chemical parameters: suspended solids, COD, total nitrogen and phosphorus, chlorides, sulphates, metals in an acidified bottle, chlorophyll-a, and of course probe-based field measurements. That already covers the bulk of a lake diagnosis, a pre-works baseline survey or the monitoring of a retention basin.

The flight framework: carrying a load is not dropping it

Regulations in force in September 2026. A point of vocabulary immediately clears up the main confusion. Article 4(1)(f) of Implementing Regulation (EU) 2019/947 provides that in the open category, "during flight, the unmanned aircraft does not carry dangerous goods and does not drop any material". A sampler lowered on a line, immersed and hauled back up is not a drop: nothing is released, and the load stays attached to the aircraft from take-off to landing. Carrying a load, as such, is not prohibited in the open category. Two limits from the same article do apply without exception, however: a maximum take-off mass below 25 kg — including payload and collected water, which means reasoning on end-of-mission mass rather than take-off mass — and a height kept below 120 m from the closest point of the surface.

In practice, the mass of a carrier capable of sampling almost always places the operation in subcategory A3 of the open category: no overflight of any person, and a minimum distance of 150 m from residential, commercial, industrial and recreational areas. On an isolated lake or a quarry pit, that condition is often met naturally. On a lagoon in the middle of an operating industrial site, it is not: one must then either clear and control a ground area, or move to the specific category. Note that since 1 January 2026 the former French national scenarios have given way to the European standard scenarios STS-01 and STS-02, subject to prior declaration, or to an operational authorisation based on a risk assessment. The choice between the two regimes is detailed in our guide to choosing between the open and specific categories, the safety reasoning specific to carrying a load beneath the aircraft having been covered above.

Three further precautions specific to this type of mission:

• The exclusion zone beneath the aircraft. A suspended load requires keeping every third party out from directly underneath, along the whole outbound and return path, not just at the sampling point. On an operating site, that means signage and a written co-activity arrangement with the operator.
• Load loss and recovery. The sampler can snag on a stump, a net or riprap. The emergency cable-release procedure must be planned, recorded and compatible with the environment — releasing hardware into a drinking-water reservoir is not a neutral act.
• Local restrictions. A lake is often also a protected area, a nature reserve, a Natura 2000 site or the footprint of a sensitive structure. Check UAS restrictions on the official map and local orders before fixing the campaign date, and obtain the agreement of the water body manager, who is not always the landowner.

The provider's aviation liability insurance must explicitly cover the carriage and handling of a load: not all policies do so by default, and a generic certificate is not enough in a tender file.

Observed 2026 prices for a drone sampling campaign

The ranges below are orders of magnitude observed on the French market in 2026, excluding VAT and excluding laboratory analyses, which are billed as a separate item (see below). They assume a site reachable by vehicle and a water body without particular airspace restrictions; a control zone, a protected area or heavy industrial co-activity shifts the whole table upwards.

Analyses are a separate item, and they are billed by the laboratory. They cannot be quoted as a lump sum: the price depends entirely on the chosen parameter list, and the gap is considerable between a basic physico-chemical assessment (a few general parameters on one bottle) and a micropollutant screen or hazardous-substance search, which can amount to several times the cost of the flight. Ask the approved laboratory for a parameter-by-parameter quotation before ordering the campaign, not afterwards: that list determines the number of bottles, hence the volume to bring back, hence the number of flights and the price of the aerial service. A drone provider quoting an "analyses included" price without knowing the parameter list has not done that work.

The useful benchmark is not the price of the flight alone but the full cost of the alternative: launching a boat, two operators, flotation equipment, a prevention plan, a work-near-water permit, and navigation time between points. For a single, easily reachable point, the boat often remains cheaper. From three or four distant points onwards, or as soon as access requires a rope-access technician or a lift platform, the balance tips clearly in favour of flying.

Setting up a campaign: what to send us

A serious quotation is built from five elements, in this order: the list of parameters to be analysed and the chosen approved laboratory (it drives the volume, hence the number of flights); the sampling point plan with the target depth for each; the purpose — industrial self-monitoring, pre-works baseline, lake diagnosis, response to an inspectorate request — because it determines how demanding the sampling chain must be; the site environment (access, third parties present, airspace restrictions, water body manager); and the desired frequency.

Sampling combines naturally with other aerial surveys on the same trip: a bathymetric survey of a lake or reservoir to establish volume and siltation, a thermographic inspection of wastewater treatment basins, or a check of drinking-water catchment protection zones. Pooling mobilisation is the simplest saving on this type of campaign.

Describe the site, the points and the parameter list: we will come back with a written method, the appropriate sampler configuration, the applicable flight framework and an itemised quotation with analyses shown separately. Request a quote for a drone water sampling campaign.

Request a free quote

Also worth reading

• FEDER: Funding a Drone Project with EU Regional Funds (Local Authorities, SMEs)
• Protected species on facades: surveying nests by drone before renovation or demolition, price
• Measuring river discharge by drone: image velocimetry (LSPIV), method, accuracy, price