French study finds crowdsourced vineyard data can track water stress across wine regions

A seven-year French research project suggests that crowdsourced field observations can help track vine water stress across large wine regions, offering a practical tool as vineyards face hotter and drier growing seasons.

The study, published July 1 in OENO One, examined the Apex-Vigne project, which asks winegrowers and advisers to record weekly observations of vine shoot growth through a mobile app. Researchers said the project collected 34,233 observations across metropolitan France between 2019 and 2025 from more than 771 contributors working on 11,481 fields.

The paper was written by Don Ced Ogoumond, Bruno Tisseyre and Leo Pichon. It analyzes whether participatory data collection can produce reliable regional information on vine water status, a key measure for understanding how much water stress vines are under during the growing season.

That question matters well beyond academic research. Water stress affects shoot growth, yields and grape quality, and better regional monitoring could help wineries, cooperatives and growers make decisions on irrigation timing, harvest logistics and longer-term investment. In a wine sector under pressure from climate change, a system that gathers field data at scale could become a useful support tool if its limits are managed carefully.

The researchers focused on France because it combines many wine regions, climates and growing practices, and because it is the only country where Apex-Vigne has had a structured rollout aimed at wine professionals. They also looked closely at a 49,500 km2 area in southern France that includes vineyards in Languedoc, Provence and the Côtes du Rhône. That zone was chosen because Mediterranean conditions make water stress more common and because adoption of the app was higher there.

The method behind the project is simple. Contributors observe 50 vine shoot tips, known as apexes, and classify them into three stages: full growth, moderate growth or stopped growth. The app then calculates an indicator called iG-Apex using weighted values of 1, 0.5 and 0. As vines come under stronger water restriction during the season, shoot growth slows and the indicator tends to fall from near 1 toward 0.

Researchers said this makes iG-Apex a practical proxy for vine water status when water availability is the main factor limiting vegetative growth. In most cases, contributors monitor fields weekly from around full bloom to veraison, when grapes begin to ripen.

The app was first released in French for Android in June 2019. It was rebuilt and relaunched in April 2024 on both Android and iOS in five languages: French, English, Spanish, Portuguese and Italian. It records time stamps and geographic coordinates for each observation and uploads data to a central database when a network connection is available.

The study found that the volume of observations was enough to map vine water status at regional scale and to show how conditions changed over time. That is one of the paper’s central findings: crowdsourcing can generate useful spatial information for viticulture decision-making when enough participants contribute regularly.

But the researchers also found that participation did not follow a single pattern. Instead, observations reflected five different contribution behaviors tied to contributors’ own interests and work needs. The app was used for within-field experimentation on farms, routine monitoring at farm level and reference-field monitoring at regional level.

That diversity helped build a large database, but it also introduced bias. The paper says crowdsourced agricultural data are shaped by where contributors choose to observe and when they decide it is worth doing so. Some places or periods may be heavily sampled while others remain thinly covered. The authors argue that this unevenness is one of the main scientific challenges in using participatory data for regional vineyard monitoring.

The project’s participation strategy relied on direct benefit for users rather than volunteer enthusiasm alone. The app was promoted as a simple decision-support tool that could help growers diagnose vine water status quickly in their own fields. Communication campaigns ran through technical trade media and wine industry fairs, while the French Institute for Vine and Wine and local Chambers of Agriculture helped spread the tool among professionals.

That professional focus appears to be important. Unlike many citizen science projects built around hobbyists, agricultural crowdsourcing often depends on workers contributing as part of their daily activity. In this case, winegrowers and advisers were more likely to participate because the observations could also serve their own operational needs.

The paper places the work in a broader climate context. Regional vineyard monitoring is increasingly important for tracking water stress, frost events and pest development, all of which can affect yields and grape quality. The authors note that vine water status shows strong variation across space and time and is among the variables most exposed to climate change.

For beverage producers, especially in wine, that makes regional monitoring more than a technical exercise. If growers can identify which areas repeatedly face water constraints and when those constraints intensify during the season, they may be better positioned to adapt irrigation plans where allowed, organize grape intake at wineries and target vineyard investments more efficiently. The findings may also be relevant to other drink sectors that depend on agricultural raw materials under rising climate pressure, though the study itself focused on vineyards.

At the same time, the authors caution that scientific work remains to be done before crowdsourcing can be treated as a fully robust monitoring system. They say social science research is needed to better understand why contributors take part and what keeps them engaged over time. They also call for new data science methods that could automatically detect atypical observations.

The analysis covered all observations collected during the study period without pre-filtering outliers. The researchers said they assumed unusual observations represented only a small share of the total dataset. They used statistical methods including clustering analysis to identify user behavior patterns and geostatistical tools such as semi-variograms and kriging to assess spatial structure and produce interpolated maps.

The article was published as original research in cooperation with the 16th International Terroir Congress and the 3rd ClimWine Symposium, scheduled for July 5 through July 9 in Angers, France.

While the study does not present crowdsourcing as a complete answer to vineyard monitoring, it offers one of the clearest long-term tests so far of whether growers’ shared field observations can produce usable regional intelligence. After seven seasons of data collection, the answer from Apex-Vigne appears to be yes, but with clear caveats about sampling bias, contributor behavior and data quality control.