Skip to Main content Skip to Navigation
Theses

Précipitations en Antarctique : de leur observation par télédétection à leur représentation dans un modèle de climat global.

Abstract : Antarctica is an immense continent made of ice. This region remains to this day one of the most unknown regions of our planet Earth. It contains nearly 90% of the world's fresh water, and in the current situation of global warming, this frozen reservoir is under serious threat. In the absence of reliable observations or efficient climate models, it is so far difficult to verify what is happening in Antarctica, and how it may evolve in the coming decades. Indeed, some studies predict significant ice melt in the west, other studies predict snow accumulation at the surface of its eastern region, mostly by precipitation events. It is, however, difficult to know if one of these two processes takes priority over the other, and which one. The CloudSat satellite is the only precipitation survey tool available on the continental scale. However, there are significant uncertainties about its measurements and the confidence that can be given to it is inevitably very low. Using precisely calibrated radar instruments deployed on the Dumont d'Urville coast station and the Princess Elisabeth continental station, we compared observations of precipitation from both surface and space. Comparison of these observations between the two datasets allowed to validate the satellite's precipitation measurements by re-evaluating an uncertainty on the measurement -- initially between 150 and 250% -- to less than 24%. This result gives more confidence to the measurements of this satellite, and from four years of continuous precipitation observation, we have developed the first three-dimensional and model-independent climatology of snowfall in Antarctica. Comparison of this dataset with theoretical precipitation rates calculated from the forced uplift of an air mass along a topographic slope shows that snowfall appears to be controlled at the first order by large-scale advection and forced uplift of wet flows against topography. We compared this dataset to the global climate model LMDz to assess its capability to represent precipitation over this polar region. Using several simulation configurations, we identified dynamic, warm and moist biases as well as biases in the moisture advections of the model. We have explored ways to reduce these biases, such as the use of atmospheric forcings (winds, temperature and specific humidity) or the sensitivity of microphysical precipitation parameters. In addition, a combination of forcing and numerical adjustment reducing the dissipation of the LMDz model on diabetic perturbations allows precipitation to be simulated in accordance with the available observations.
Complete list of metadata

https://hal-insu.archives-ouvertes.fr/tel-03625070
Contributor : Florentin Lemonnier Connect in order to contact the contributor
Submitted on : Wednesday, March 30, 2022 - 4:17:08 PM
Last modification on : Sunday, April 3, 2022 - 3:28:22 AM

File

Thèse de doctorat.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-03625070, version 1

Citation

Florentin Lemonnier. Précipitations en Antarctique : de leur observation par télédétection à leur représentation dans un modèle de climat global.. Climatologie. Sorbonne Université, 2019. Français. ⟨tel-03625070⟩

Share

Metrics

Record views

58

Files downloads

5