A transdisciplinary strategy involving machine learning, genomics and ecological modelling to improve the resilience of beech forests to climate change

The main objective of FAGRESCUE is to quantify and characterize the standing adaptive of key adaptive traits in unmanaged beech forests to constitute gene pools that could eventually be exploited for future forest management programmes.

To do this, we will 1) develop a drone-based procedure and machine-learning based programmes for image analyses to yield fast and reliable phenotypic characterization of beech populations for both phenology and drought resistance; 2) screen two unmanaged forests for such phenotypic diversity using this method and conduct population genomics survey in order to identify the genetic factors involved in these traits using GWAS and 3) use predictive ecological and population-genetic modeling to predict the fate of both the phenotypes and the genotypes thus identified in various climate projections for the next 50-100 years.

To reach this goal, FAGRESCUE proposes to quantify and characterize the standing phenotypic and genetic variation in the bud break date and stress tolerance of unmanaged old-growth beech forests and develop predictive models of genotypes performance in future climate conditions. More precisely, we will 1) Develop a drone-based procedure and AI-based strategy for image processing to yield fast and reliable phenological characterization of beech populations at regional scale; 2) Identify the genomic bases of bud break date and drought tolerance through GWAS; (3) Use ecological and evolutionary modelling to predict the fate of both the phenotypes and the genotypes thus identified in various climate projections for the next 50-100 years.