Reflexions, the University of Liège website that makes knowledge accessible


Fields under the watchful eye of observation satellites

2/17/17

Farmers will soon be able to accurately predict harvestable crop yields. In the context of the BELCAM project (Belgian Collaborative Agriculture Monitoring), the Department of Environmental Sciences and Management of the University of Liege is interested in developing tools capable of accurately predicting harvestable crop yields.

Financed by Belspo and piloted by the Catholic University of Louvain, BELCAM was started in 2014. The purpose of BELCAM is to apply new Earth observation systems to agriculture. Several Belgian and French partners are currently collaborating on the project including the Department of Environmental Sciences and Management of the University of Liege.

“Everyone has a specific role, at ULg, we are interested in predicting agricultural crop yields. In order to do this, we combine satellite images with crop growth-simulation software”, says Joost Wellens, a graduate in agricultural sciences who is coordinating the project for ULg.

Plateforme BELCAM

A sentinel in the sky

Every ten days, Dr. Wellens receives satellite images of agricultural parcels captured by the European satellite “Sentinel-2”. Its twin satellite will be placed in orbit alongside it in 2017 which means that the researcher will receive images every 5 days. Launched in 2015, Sentinel-2 supplies free images to anyone who requests it. It is a huge asset to scientists: “Before the launch of Sentinel-2, these images were very expensive. It costs up to between 2,000 and 3,000 Euros to obtain a detailed satellite image of a rather small zone/region. Free images existed, but their precision was quite mediocre and was not sufficient for monitoring at parcel level”, explains Dr. Wellens.

The use of satellite images nonetheless presents a major disadvantage. “If the sky is overcast, nothing can be seen and therefore no predictions can be made”. This is where simulation software comes into play. “We can solve this problem by combining usable satellite images with simulation software; that will simulate/model the plant behavior in between two images. If we see that there are some errors as soon as an image becomes available, we can correct theprogram on the fly”, explains the scientist. Currently, scientists are using the AquaCrop software-model, made freely accessible by the United Nations Food and Agriculture Organization (FAO). AquaCrop is capable of calculating the evolution of the canopy, that is to say, the speed of canopy growth and decline, and phenological stages. By means of this information, the model predicts whether or not a harvest will be profitable. 

A project combining researchers and farmers 

This will improve the production methods of farmers who are in direct collaboration with the researchers on the project. “We have created an internet platform where the farmer can register and supply data about his production. He can mark out the boundaries of his field on a satellite image background, state the crop type he has planted (corn, winter wheat or potatoes, for example), and indicate the planting date as well as the doses of applied fertilizer or any other information relative to the crop development”, explains Joost Wellens. This cooperation is essential because, according to the researcher “the more data we have, the more reliable our prediction tools will be and the better the advice we will be able to offer farmers”. 

The latter are informed about the profitability of their yields, for example, as compared with the region’s average produdctions. If necessary, recommendations could be given enabling them to increase their productivity. 

Foliose Cover- Aquacrop
Graphic simulating the foliose cover - Aquacorp software

Tools available to everyone 

In terms of results, the tool appears to be efficient. The project researchers found a correlation between the crop yield observed on the ground and that simulated by the software with an error of 1.6 tons per hectare for corn (in dry matter), and 0.9 tons per hectare (dry matter) for winter wheat and potatoes. “Between now and the end of the project, in 2019, we hope to be able to reduce this error. And in order to do this, we need to convince more farmers to share their data online”.

The other part of the project that needs to be further improved in the future is the “detection of anomalies”. “Our prediction models are effective when we find ourselves in what can be called “normal” situations. But what interests the farmer most of all, is what will happen to his yield in case of problems such as drought or heavy rain, for example. More research on this subject needs to be conducted”, admits Joost Wellens. In the long run, the team hopes that these tools will be used by all the Belgian pilot centers that currently offer advice to farmers. “If possible, we will try to create towards the end of the project a free application that can be used by the entire agricultural sector”, concludes the scientist.


© Universit� de Li�ge - https://www.reflexions.uliege.be/cms/c_438842/en/fields-under-the-watchful-eye-of-observation-satellites?printView=true - March 29, 2024