Abstract ID: 107
Predicting the leaf area index in a dynamical S2S forecast system
Lead Author: Constantin Ardilouze
CNRM, Université de Toulouse, Météo France, CNRS, France
Keywords: Interactive vegetation, LA, ESM, predictability
Abstract: Processes governing land-atmosphere interactions have a flawed representation in dynamical subseasonal and seasonal forecast systems. In particular, the evolution of vegetation is generally prescribed, following an annual climatology for example. However, vegetation is a key player in the climate mean state and interannual variability, due to the impact on the surface water and energy budgets through the soil evaporation, the canopy transpiration and albedo. Furthermore, the time scale of the evolution of vegetation is slower than atmospheric processes, which could be an asset for atmospheric predictability because of the persistence of anomalous leaf area index initial states.
While recent efforts on land sources of subseasonal-to-seasonal predictability have mostly focused on soil moisture and its initialization, the impact of a better representation of the vegetation evolution on S2S forecast skill is yet to be explored.
In the framework of the European project H2020-CONFESS, we evaluate an ESM-based dynamical subseasonal-to-seasonal reforecast run with an interactive vegetation scheme.
We primarily focus on the extent to which the leaf area index (LAI) is predictable against a reference observational dataset, delivered during the CONFESS project. To our knowledge, inline predictions of LAI anomalies at the global scale have never been explored, and may provide avenues for innovative climate services such as crop management, or the anticipation of wildfire risk.
The potential of LAI as a source of atmospheric S2S predictability is also discussed by comparing the skill of screen-level temperature reforecasts run with or without the interactive vegetation scheme.
Co-authors:
Dayon, G. (CNRM)
Batté, L. (CNRM)
Decharme, B. (CNRM)
