Design and co-construction of decision support tool for pasture management in Wallonia

Developing new technologies to improve the resources management, or our ability to predict and best confront changes, is a subject on track for this past decade in the field of grass and pasture. So, from an online inquiry including 156 respondents, 64.7 %, mainly young farmers (84% between 18 and 30 years old), expressed massively their wish to have a decision tool helping them to manage more efficiently their grass resources. But they were not inclined to spent more than 10 minutes on this tool per day suggesting a high automation. Furthermore, the typology established from extended interviews of 15 breeders was taken in account. So, from highlighted farmer’s needs, a decision tool to manage pastures was developed and installed on the first numerical platform dedicated to the agriculture in the Southern Region of Belgium, called WalleSmart. Even if any prediction model for the grass biomass or height can be used, we focussed on the use of compressed sward height (CSH) estimated from meteorological and satellite data with a prediction error around 20 mm. The structure of dataset used to feed the tool contained those predictions aggregated on a pixel basis having a 10m x 10m resolution in order to ease the biomass calculation and visualization. This represents 39 989 584 pixels. To optimize the computational resources, a dataset was created separately for each prediction date. So, the tool in R coding language and integrated in a docker on the WalleSmart server allows to observe the biomass, estimated from CSH values, at the last day of prediction by showing all the pastures belonging to the farmer. In order to take in account the possibility to divide a parcel to work with smaller area, the tool was also designed to allow the farmer to draw his parcels himself. Then, by clicking on the parcel, the tool shows the detail of the biomass by pixel and pasture. Finally, a retrospective of the grass growth is available. A second version of the tool is planned in order to take in account both the introduction of new elements such as the farmers animal and the opinion of the breeders concerning this first application obtained from a farmer’s focus group.