Researchers working on the case study. Credit: University of Cordoba

A monitoring system devised by the University of Cordoba ascertains the flowering stages of each hive, with high precision, exploiting data on bees’ behavior.

Beekeeping has existed for millennia, as evidenced by a painting in the Cueva de la Araña (Valencia), more than 8,000 years old, depicting a human figure collecting honey from a hole in a rock, with bees hovering around. This relationship between humans and bees is essential not only to obtain honey, pollen and wax, but also to conserve the honeybee (Apis mellifera L.), responsible for the pollination of thousands of crops.

The success of hives and beekeeping depends on flowering periods, which are irregular because they depend greatly on the season, rainfall, and temperatures. Reducing uncertainty and knowing the precise duration of flowering periods was the goal of a team from the Department of Electronics and Computer Engineering and the Department of Zoology at the University of Cordoba, which has been working for years on hive monitoring systems providing beekeepers with accurate data in real time.

The system designed allows them to ascertain the exact flowering time for each hive, from the beginning of the day until it ends, thanks to a weight sensor. “By observing how the weight of the hive varies throughout the day, and analyzing the curve that results from the weight measurements every 5 minutes, we obtain information about the current flowering stage,” explained Andrés Gersnoviez, lead author of the work.

Contrary to what one might think, it is not only the weight gain of the bees, loaded with nectar, that reveals the moment of flowering, but also when the bees begin to leave to forage (collect), the number of bees that leave and enter the hives, and how much time they spend outside them, among other things.

“If you know when the hive’s minimum weight occurs, which is when they go out to look for food, and when the maximum occurs, you know that they have already returned. In addition to this, knowing if that minimum and/or maximum have a peak or valley shape, as well as the difference in weight from the beginning of the day until it ends, all together tells us how long it has taken them to return, and the success they have had in their searches, which allows us to pinpoint the flowering phase,” the researcher explained.

With the data obtained from the hive sensors, the team designed a classifier using Artificial Intelligence algorithms with factors that describe the weight curve and relate that data to flowering. Once they had met their objective, they went further.

“We saw that, within the flowering period, we could distinguish between an earlier stage and an end one,” Gersnoviez said.

In this way, a system is obtained that is capable not only of determining whether flowering is impending, occurring, or finished, but also of distinguishing between an initial and a final flowering stage.

Technology to facilitate beekeeping

Beekeepers usually work with several apiaries featuring 40 or 50 hives. These apiaries are miles apart, sometimes hundreds, so knowing exact flowering times without having to visit them saves beekeepers time and allows them to carry out this age-old activity more efficiently.

According to José Manuel Flores, a researcher who has participated in the work on the Zoology side, this helps to enhance harvests. “The system tells you that the flowering will end in a few days. This allows beekeepers to plan where to go to in order to collect the honey. If you arrive too early and the flowering hasn’t finished, you’re losing the part of the harvest that the hive can still generate. Conversely, if you arrive late and the flowering finished days before, the bees are already feeding on that honey, and part of the harvest is also lost.”

It is especially useful for those who produce monofloral honey (of a single variety, such as orange blossom, chestnut, eucalyptus, or avocado, among many others), since, if it is not collected once the flowering is finished, the bees can take nectar from other flowerings of more distant crops, in which case the honey is no longer monofloral, and loses added value.

The monitoring system created through these teams’ collaboration provides much more information. Thanks to these sensors—which measure hives’ weight, humidity and temperature data every 5 minutes, and that can be consulted remotely via a computer—beekeepers can also track the changes that occur and verify whether there are any health problems, predators nearby, or any kind of interference, without having to actually visit the apiary.

The findings are published in the journal Computers and Electronics in Agriculture.

More information: Andrés Gersnoviez et al, Determination of flowering stage based on artificial intelligence and the daily weight of bee hives, Computers and Electronics in Agriculture (2025). DOI: 10.1016/j.compag.2025.110508

Citation: AI in bee hives: Monitoring system remotely predicts flowering periods (2025, December 2) retrieved 2 December 2025 from https://phys.org/news/2025-12-ai-bee-hives-remotely-periods.html

This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.