Understanding and modelling our environment

Despite 350,000 species of vascular plants having been identified worldwide, only 8% of them have protected status, which gives us an indication of how they are doing at a global level. To rectify this, Alexis Joly, a researcher with the Zenith project team at the Inria Sophia Antipolis research centre, working alongside colleagues from the Cirad (the French Agricultural Research Centre for International Development), the Leca (the Alpine Ecology Laboratory), the LIRMM (the Montpellier Laboratory of Computer Science, Robotics, and Microelectronics) and the University of Montpellier, had an idea: to use artificial intelligence and machine learning. The decision was then taken for an exploratory action to be set up, with Cactus launched in 2020. The aim of this exploratory action is to teach computer models to draw on meteorological and climate data, satellite images and data on occurrences in order to predict whether or not a given species would be found in a particular environment and to determine its conservation status. Long-term, the tool could even be used to predict the conservation status of habitats themselves and to compile the world’s very first biodiversity map.

Croco - king of the ocean?

Standing for Coastal and regional ocean community model, Croco is software for modelling ocean circulation. Its strength lies in the fact that it was developed by several different organisations.  Inria, the Ifremer (France’s National Institute for Ocean Science), the Shom (France’s Naval Hydrographic and Oceanographic Service), the IRD (the Research Institute for Development), the CNRS and the University of Toulouse all pooled their expertise in order to improve existing models and to create software capable of generating 3D simulations of ocean currents, changes in sea level and fluctuations in temperature and salinity based on external databases.

Croco also has the capacity to integrate meteorological data, in addition to data on sedimentology, biochemistry and data relating to the ecosystem. It has a whole host of applications, from monitoring marine pollution and managing fish stocks to issues linked to planning and defence/security. And there could be more on the horizon: the researchers are currently working on integrating data on sound waves so as to enable the software to trigger an early warning in the event of a tsunami. Clearly, Croco is still hungry for data...

On one hand, there was the Lacodam project team at the Inria Rennes - Bretagne Atlantique research centre, specialists in data mining and analysis. On the other were specialists in precision livestock farming from the Inrae in Rennes, who employ the use of sensors in order to understand the needs and behaviour of livestock animals and who collect a lot of data. When the two came together within the convergence lab #DigitAg, it was only a matter of time before research topics began to emerge.

Two joint PhDs have just been completed, while two more are set to follow. Here is a run-down of what their work has involved:  using sensors to detect when cows are in heat; to understand how to feed nursing sows; to study why certain cows are better at coping with heatwaves than others; and to identify environmental factors (e.g. sound, temperature) which impact the well-being of pigs in their sties. The objective is the same across all four of these research topics: to assist agriculture through digital while ensuring human values and animal wellbeing are respected.