Cerfacs Enter the world of high performance ...

Impacts of climate change on the continental hydrological cycle in Europe: mechanisms, uncertainties, diversity of tools and consistency of messages?

SDU2E (Sciences de l’Univers, de l’Environnement et de l’Espace) – [Subject to defense authorization]

https://youtube.com/live/T7pXRW3kz9M?feature=share

The water cycle is expected to be undergo significant modifications with climate change.
This will greatly exacerbate regional disparities in water resource availability. In Europe particularly, existing contrasts are projected to intensify.
Hydrological projections can be made by forcing global climate models with future anthropogenic forcing scenarios and are subject to several sources of uncertainty. The first stems from the scenario itself, the second is related to limitations in scientific knowledge and model imperfections, and the third is linked to the climate’s internal variability, which can be significant even at multi-decadal scales. Hydrological projections can also be generated using hydrological models, which represent continental hydrology more finely than global climate models. These models are forced by downscaled and bias-corrected climate data.
These downscaling and correction steps introduce additional uncertainties beyond those related to climate modelling and internal variability.
The objective of this thesis is to expand our knowledge of the uncertainties and mechanisms associated with the future evolution of the hydrological cycle in Europe and France. The results are based on projections from climate models forced by a high-end emission scenario, but also on projections from single model initial condition large ensembles, used to quantify the uncertainty arising from internal variability, and finally, on hydrological changes over France projected by different modelling tools in response to a high-end emission scenario and to different global warming levels. The results indicate that changes in atmospheric circulation and the physiological effect of CO2 partly explain the discrepancies
between the hydrological projections of climate models over Europe. For the coming decades, decadal internal variability constitutes a significant, and in some European regions, even a predominant source of uncertainty. The evaluation of model multi-decadal variability suggests that this uncertainty might be underestimated for the Mediterranean region. The multi-decadal variability in runoff projections originates from the multi-decadal variability of precipitation, which is itself driven by the multi-decadal variability of atmospheric circulation over the North Atlantic and European region. Finally, for the specific case of France, the different tools and models project a decrease in summer precipitation for all four studied watersheds (Seine, Loire, Rhône, and Garonne) and an increase in winter for all basins except the Garonne. The Garonne basin is also characterized by a robust decrease in annual mean precipitation, streamflow and runoff. For other hydrological changes, significant uncertainty remains due to uncertainties linked to climate modelling and internal variability, as well as disagreements between modelling tools. These disagreements are explained by three main factors: an alteration of the climate change signal induced by bias correction and downscaling methods; the lack of accounting for the physiological effect of CO2 in most regional and hydrological models; and the choice of the hydrological model, particularly influential for summer streamflow projections.

Jury

Agnès Ducharne	CNRS UMR METIS	Reviewer
Sophie Bastin	CNRS LATMOS	Reviewer
Florence Habets	CNRS ENS-PSL	Examiner
Lionel Jarlan	IRD CESBIO	Examiner
Jeanne Colin	Météo-France CNRM	Examiner
Julien Boé	CNRS UMR CECI	Thesis supervisor