In agriculture, it is important to understand how certain species — plant, animal, viral or bacterial — evolve over time and space. Many questions follow from this. For example, will these species be suited to future climatic conditions or farming practices? How can we control the arrival, adaptation and impact of crop-damaging pests that migrate from one continent to another? To do this, we need to be able to collect relevant data on the subject and mathematically model these changes in plant and animal populations in order to make forecasts useful to agriculture. The AgroStat project (for "Statistics for the Evolution and Dynamics of Populations and Species of Agronomic Interest") studies these dynamics through the complementary expertise of more than fifty scientists from different fields: biology, genetics, agronomy, mathematics and statistics.
Biological challenges, mathematical tools -------------------------------------------
One of the project's key points is to make it possible to anticipate the risks linked to pests — living organisms that damage cultivated plants or harvests. These include crop-damaging insects, viruses, bacteria and fungi… This requires a better understanding and modelling of the many factors that influence these risks. First, techniques must be put in place to collect reliable and relevant data on crops and pests in numerous locations. Next, we need to understand the routes taken by new invasive pest species. We also need to track the genetic mutations in plants that allow them to better adapt to new climatic or agricultural conditions. Mathematical models can also be used to describe and explain the spread of viral diseases in a given environment.
Combining all this knowledge should help predict future agricultural performance. This calls for mathematical tools capable of handling large amounts of heterogeneous data, while accounting for the uncertainty in measurements and models and the complexity of interactions between species and with the environment.
Collecting and analysing data -----------------------------------