Florentin Lemonnier - Remote sensing observations and climate modeling of snowfall in Antarctica.

Event type: 
Seminar
Date: 
7 March 2018
Time: 
2.00 - 3.00pm
Location: 

Climate Change Research Centre, Seminar Room, Mathews Building 4th floor, UNSW, Sydney

Presenter: 
Florentin Lemonnier
Laboratoire de Météorologie Dynamique, Paris, France
Host: 
Climate Change Research Centre, UNSW, Australia

The Antarctic continent is a vast desert, the coldest and the most unknown area on Earth and contains its largest fresh water reservoir. Current global warming could threaten this ice sheet, leading to sea level rise. The main goal of the French-Swiss APRES3 project (Antarctic Precipitation, Remote Sensing from Surface and Space) is to document and understand current precipitation over the south polar cap and improve the representation of snowfall in climate models. Currently, climate models do not agree on the amount of snowfall over the ice sheet [Palerme et al., 2014] and improving their ability to predict climate change in the Antarctic region appears as a priority.

In the context of our research project, we deployed an array of instruments at the French Dumont d’Urville station in order to observe local precipitation in details. Using a dual polarization radar, we discovered that precipitation in coastal Antarctica is strongly influenced by dry katabatic winds flowing from the ice cap’s interior [Grazioli et al., 2017] which sublimate snowfall in the first kilometer above the surface. We compare these measurements to satellite observations made by the Cloud Profiling Radar (CPR) onboard CloudSat. We thereby reassess the accuracy of the CloudSat retrievals and find that the correlation between ground radar measurement and satellite retrievals is near-perfect. Using different CloudSat and ground radar vertical levels, we obtained an uncertainty on CloudSat retrieved snowfall rates of [-21,20%, +25,44%]. We apply this uncertainty to the whole Antarctic region and build a 3D dataset which gives the 3D structure of precipitation over the entire continent over multiple years.

This dataset is compared to 3D simulations performed using the IPSL Climate Model. We obtain a mean precipitation rate of the 172 mm/yr over the whole continent in the model, whereas the CloudSat radar gives a snowfall rate of 153 mm/yr. Locally, the IPSL Climate Model tends to underestimate snowfall rates but to overestimate the occurrence of precipitation events. The model thus predicts a realistic annual accumulation rate due to compensating errors, and the seasonal variability is also misrepresented. We explore the origin of this discrepancy, and propose ways to improve climate models for better climate predictions.

 

Brief biography: Florentin is a PhD student at the Dynamical Meteorology Laboratory (LMD), Sorbonne Université, and his thesis subject is divided in two parts: a first part dealing with satellite observations of snowfall using the CloudSat satellite and; a second part using ground observations from the French Dumont d'Urville station. He is also involved in the development of the IPSL Climate Model with a particular focus on the different parameterizations controlling snowfall over Antarctica, such as cloud to snow conversion processes, snowfall sublimation, and the growth of snow flakes. Simulations are compared with the satellite observations and the Dumont d'Urville field campaign observations to validate the model, all with the aim of making reliable climate predictions in this region.