Numerical modelling of the atmospheric boundary layer : from katabatic flow to pollutant transport

Gabriele Arduini, LEGI, Nov. 2013 - Nov. 2016

Phd student : Gabriele Arduini, LEGI, Nov. 2013 - Nov. 2016
Supervision : Chantal Staquet (LEGI, Grenoble) ; Charles Chemel (University of Hertfordshire, UK)
Funding : 50 % LabEx, 50 % University of Hertfordshire
Doctoral School : Terre Univers Environnement, Grenoble

Thesis subject

The atmospheric boundary layer (ABL) is the part of the atmosphere controlled by the heating and cooling of the ground surface. The height of the ABL highly varies, from a few tens of meters in the polar regions to about ten kms at the equator. When stable conditions prevail (namely the potential temperature gradient is positive), as this occurs at night or in winter, the ABL displays well-known properties : vertical motions are inhibited so that, for an urbanized site, the pollution gets trapped at the ground level, leading to acute health and environmental conditions. The extreme pollution episode which occurred in Beijing in January 2013, when stable conditions prevailed, dramatically illustrates this effect.
For complex terrain, such as an Alpine valley, the ABL dynamics under stable conditions are mainly driven by slope winds due to the cooling of the ground at sunset (down-slope or katabatic wind) and the heating of the ground at sunrise (up-slope or anabatic wind). In winter, when synoptic stable conditions prevail, the ABL dynamics become decoupled from the large scale atmospheric circulation and are controlled by slope winds (Whiteman 2000). Recent numerical simulations of the Grenoble valley (Largeron, 2010, PhD thesis) showed that katabatic winds may flow for 20 hours or so, ceasing for a few hours only about noon. Katabatic winds advect cold air down to the valley bottom, thereby creating a pool of cold air which promotes the formation of fog, for appropriate humidity conditions, and pollutant trapping.
Current numerical weather prediction (NWP) models do not represent Alpine valleys in an appropriate manner, the most part of the valleys pertaining to sub-grid scales. It follows that these valleys, more precisely the effects of these valleys on the resolved scales of a NWP model, have to be represented in a simplified manner in this model, through a sub-grid scale (SGS) parameterization. These effects are mixing and transport induced by the strong coupling between all the processes within the stable ABL of the valley. A deep understanding of these processes is needed before any parametrization can be derived.
The main purpose of the PhD subject is to study these processes within the context of an Alpine valley, along with their interaction and impact upon the development of a stable ABL. An idealized valley will first be considered, so as to get well-controlled conditions allowing for the characterization of the processes, before addressing the realistic case of the Passy valley in the Alpes. The estimate of the validity of realistic numerical simulations in the latter context will be assessed from in situ measurements, through field experiments in the Passy valley to which the student will participate.

Main steps of the thesis work

The PhD work is carried out in co-tutelle with the University of Hertfordshire (UK). The first part of the work (18 months) is being performed at LEGI, the second part (18 – 36 months) will be performed at the University of Hertfordshire.

Work already performed in LEGI (October 2013-May 2014) :

• Extensive bibliography on the PhD topic ; getting familiar with the WRF (Weather Research and Forecasting) model used for the numerical simulations.
• Set up of different numerical simulations of fully 3D idealized alpine valleys. The interaction between the valley wind which develops in such a valley and the stable ABL of the valley have been studied in detail, following the work of Burns and Chemel (Boundary Layer Meteorology, 2014a,b). Comparison with 2D numerical simulations in a vertical plane, which are commonly reported in the literature, has been performed to assess the validity of a 2D approach. This work is currently being written as a paper to be submitted to the journal Boundary Layer Meteorology. It was also presented at the 21st Symposium on Boundary Layers and Turbulence of the American Meteorology Society (AMS) held in June 2014 in Leeds (UK).

Outline of future work
Before performing the numerical modelling of the Passy valley, an intermediate study will be conducted, where the main features of this valley will be reproduced in an idealized context (September to December 2013). G. Arduini will then take part to the field campaign (January-February 2015). Data analysis, realistic numerical modelling and comparison between data and numerical results will form the remaining part of the PhD work.

Organization of the co-tutelle
Periodic meetings (every 3 months) and weekly skype meetings which involve the student and the two supervisors.
The meetings were held on :

• November 4 - 9 : visit of Gabriele Arduini and Chantal Staquet to the University of
  Hertfordshire.
• March 3 – 7 : visit of Gabriele Arduini to the University of Hertfordshire.
• June 2 – 5 : visit of Charles Chemel to LEGI.

Participation to a meeting
June 09 - 13, 2014 : Participation to the 21st Symposium on Boundary Layers and Turbulence of the American Meteorology Society (AMS) (poster).