C-DRONE GUIDE · 4 SEPTEMBER 2026

Crop insurance and climate losses: what a drone survey adds to the claim file, price

Since 1 January 2023, compensation for weather-related crop losses in France has rested on a three-tier scheme: routine losses stay with the farmer, mid-range losses fall to subsidised multi-risk climate insurance, and exceptional losses trigger national solidarity beyond a threshold — 50% for arable crops, industrial crops, vegetables and viticulture, 30% for orchards, soft fruit and grassland. Everything therefore hinges on one measurement: the loss rate, assessed by an expert appointed by the authorised insurer. Yet a hailstorm or a frost pocket almost never hits a field evenly, and the visual signature of the damage fades within days. A dated aerial survey flown in the window right after the event does not set the statutory loss rate — the joint on-site assessment alone does that — but it documents what ground sampling captures poorly: the exact outline of the affected zones and their area. Here is what that record adds to the file, its limits, and the prices charged in 2026.

Published on 4 September 2026, reviewed on 7 September 2026 — regulations in force as of September 2026.

The three-tier scheme, and what applies to the 2026 season

The reform that took effect on 1 January 2023, stemming from Law n° 2022-298 of 2 March 2022 on wider uptake of crop insurance in agriculture, replaced the old agricultural-disaster regime for crop losses with a three-level architecture:

• Tier one — routine losses: borne by the farm itself, absorbed by the policy deductible and, where applicable, by tools such as the precautionary savings allowance.
• Tier two — multi-risk climate insurance (MRC): a subsidised contract covering the recognised climate hazards, including hail, frost, drought and waterlogging. Public support for eligible premiums is set at 70%, against roughly 62% on average in 2022, and the subsidisable deductible was lowered to 20% of average annual production, down from 25% before the reform.
• Tier three — national solidarity: beyond a loss threshold of 50% for arable crops, industrial crops, vegetables and viticulture, and 30% for orchards, soft fruit, grassland and speciality crops, the national solidarity indemnity (ISN) takes over.

For an insured farmer, losses above the threshold are compensated at 100%, of which 90% comes from national solidarity and 10% from the insurer. For an uninsured farmer, Decree n° 2025-1175 of 5 December 2025, which sets the parameters for the 2026 to 2028 harvests, organises a planned decline: for arable crops the rate falls from 35% for the 2025 harvest to 28% for 2026, then 21% in 2027 and 14% in 2028; the reduction is slower for orchards, soft fruit and grassland, at 31.5% for 2026. Exactly which crop group a given production belongs to is set out in the decree and should be checked with the authorised insurer before any calculation. The same text includes a safeguard clause: if public spending on the scheme exceeds €680 million a year, an interministerial order may temporarily set derogatory rates and thresholds until 31 December 2028.

One organisational point is decisive for the file: the authorised insurer (interlocuteur agréé) acts as a single window. It records the loss, calculates the damage and the ISN, and pays out both the insurance indemnity and the national solidarity payment. Since 2024, insurers have also handled national solidarity payments on grassland, including for uninsured farms.

Everything hinges on one measurement — and that is where files get weak

Deductible, trigger threshold, compensation rate: every tier of the scheme revolves around the same quantity, the crop loss rate, measured against a historical reference yield. A file is won or lost on a few percentage points: at 48% loss on a crop subject to the 50% threshold, national solidarity does not trigger; at 52%, it triggers on the excess. The financial stake of a precise assessment is therefore direct, for the farmer and the insurer alike.

Yet the hazards concerned are all marked by strong within-field variability. Hail moves in corridors a few dozen metres wide. Spring frost settles in low ground where cold air pools, sparing neighbouring slopes: our guide to spring frost risk mapping in vineyards and orchards covers that mechanism. Drought and waterlogging follow soil texture and topography, with contrasts documented in our guide to water stress monitoring using thermal and multispectral imagery. In every case, the assessment rests on sampling: a handful of plots, a handful of transects, from which a field-scale rate is extrapolated. The uncertainty of that extrapolation is the leading source of disagreement between farmer and expert.

The problem is structural enough to have driven entirely different insurance products elsewhere. The review by Willemijn Vroege, Tobias Dalhaus and Robert Finger published in 2019 in Agricultural Systems surveys index insurance for grassland across Europe and North America: because forage yield is almost impossible to measure directly on a grazed plot, these contracts pay out on an index correlated with yield — at the cost of basis risk, the gap between what the index says and the loss actually suffered (see the study on Google Scholar). That is precisely the trade-off the French scheme makes on grassland, compensated through a satellite grass-growth index. For crops still assessed by conventional field expertise, the question remains open: how do you document a field's real variability before it fades?

What a dated aerial survey evidences — and what it does not establish

A flight carried out soon after the event produces a georeferenced, time-stamped orthophoto of the whole field. On that basis, three things become measurable rather than estimated: the outline of the damaged zones and their area, expressed in hectares or as a share of the field; the within-field variability — the distribution of severe, moderate and undamaged zones, which shows where ground sampling should concentrate; and, where an earlier flight exists, the gap against the crop's condition at previous growth stages.

The scientific literature documents this capability. A study by Jianfeng Zhou, Mark Pavek, Seth Shelton, Zachary Holden and Sindhuja Sankaran, published in 2016 in Computers and Electronics in Agriculture, assessed simulated hail damage on potato using aerial multispectral imaging: the authors conclude that such imaging can assess the damage and that vegetation indices show potential for predicting yield losses (see the study on Google Scholar). More recently, a study run at the University of Padua by Jacopo Furlanetto, Nicola Dal Ferro, Francesco Morari and co-authors, published in 2023 in Precision Agriculture, compared drone and Sentinel-2 satellite imagery for estimating the leaf area index of maize defoliated by hail: the drone-based NDVI estimate reached an R² of 0.86 with a root-mean-square error of 0.71, against an R² of 0.65 and an error of 1.37 for Sentinel-2, at a ground resolution of 4 centimetres per pixel versus 10 metres for the satellite (see the study on Google Scholar). That resolution gap is why a thirty-metre hail band reads clearly on drone imagery and dissolves into a single satellite pixel.

Sensor choice follows the hazard. High-resolution RGB is enough to outline a flattened crop, a hail band across vines or a defoliated orchard. Multispectral becomes useful when the damage shows up first as a loss of vigour or leaf area — frost, drought, waterlogging — and above all when an earlier baseline flight allows a comparison; our guide comparing multispectral and RGB cameras in agriculture covers that choice.

The limits must be stated plainly. The drone does not establish the statutory loss rate. That rate is calculated against a historical reference yield and falls to the expert appointed by the authorised insurer, whose joint on-site assessment remains the only binding reference. An aerial image shows an affected area, not a yield loss: between 40% of foliage destroyed and 40% of the harvest lost, the relationship depends on the species, the growth stage and the plant's capacity to compensate. Finally, an aerial survey says nothing about grain or fruit quality, often decisive in the actual damage suffered. The aerial record is supporting evidence for the file, to be submitted on the same footing as dated ground photographs — and it is good practice to tell the authorised insurer it exists.

Who commissions the flight, and when

Four profiles commission this kind of mission, with distinct expectations.

• The farmer, who wants a dated record in hand before the expert's visit, especially on scattered fields where several days may pass before each site is inspected.
• The cooperative or grain merchant, which needs a quick read on the scale of an event across its collection area: flying a sample of member fields gives a territorial picture within days, useful for harvest forecasts as much as for supporting members.
• The agricultural expert, who uses the orthophoto ahead of the visit to target sampling plots rather than walking the field at random — a clear time saving on a large holding, and a more representative sample.
• The broker or insurer, documenting the extent of an event across a portfolio and scheduling expert visits according to the areas actually hit.

The main constraint is time. Policies set a short declaration deadline once the loss is noted, and the visual signature of the damage degrades fast: the plant compensates, rain or wind alters the look of the canopy, and any mowing, clean-up or harvesting erases the evidence on the ground. The useful window is measured in days, not weeks. Our guide to drone assessment of agricultural damage after hail or storms covers the operational side of that flight: altitude, flight plan, regulatory framework and deliverables. Where the same event also damaged barns, livestock housing or the farmhouse, surveying the buildings falls under a different procedure and a different policy, described in our guide to mass storm and hail damage surveys on buildings: the two files are built separately, even if a single site visit can feed both.

One practice genuinely changes the value of the file: having a baseline flight from before the loss. A farm already monitored during the season — vigour mapping, emergence counts, irrigation monitoring, as described in our guide to drones in precision agriculture — holds an independent record of the crop's condition before the event. That is what turns an "after" image into a measurement of change.

Method and 2026 prices

A typical mission runs in four steps: booking a slot as soon as the event is reported, a grid flight over the fields concerned, photogrammetric processing producing a georeferenced orthophoto, then a report organised field by field — areas by apparent damage level, illustration plates and time-stamped metadata usable as supporting evidence. For hazards with gradual expression (frost, drought, waterlogging), a multispectral pass is added to RGB; for a flattened or hail-struck crop, RGB is usually enough.

Ranges observed in France in 2026, excl. VAT:

These amounts cover the flight, image processing and the report; they cover neither the statutory agricultural assessment nor any counter-assessment. Our guide to how much a drone service costs places these prices within the wider French market. The regulatory information on this page reflects the rules in force in September 2026; the thresholds, rates and parameters of the current season should always be checked with the authorised insurer and the departmental land authority. For a farmer, a cooperative, an expert or a broker, request a quote stating the crop, the area, the date of the event and the response time needed.

Request a free quote

Also worth reading

• Race circuits and karting tracks: drone track surveys, and the rule that changes everything once there's a crowd
• Sargassum strandings in the French Caribbean: what a drone actually measures on the foreshore
• Sugarcane seen from a drone: vigour, ratoons and biomass in Réunion and the French Antilles