Changes between Version 1 and Version 2 of imuk/projects/1998_01

Timestamp:

Sep 15, 2010 11:49:57 AM (15 years ago)

Author:

fricke

Comment:

--

imuk/projects/1998_01

@@ -2 +2 @@
 Responsible: Michael Schröter \\
 Project type: individual [http://www.dfg.de DFG] project \\
-Duration: ''15/06/1998 - 14/06/2001'' \\\\
+Duration: 15/06/1998 - 14/06/2001 \\\\
 During so called cold-air outbreaks cold and stably stratified air-masses are advected across relatively warm surfaces. The heating from below causes the development of a convective boundary layer in which a typical sequence of convection pattern can be observed: As soon as the cold air hits the warm surface, roll-like secondary flow pattern develop, appearing in satellite images as so called cloud-streets. Further downstream the roll pattern gradually changes to a hexagonal cellular convection pattern. While in classical laboratory experiments of Rayleigh-Benard convection the aspect ratios of the convection patterns are about 1:3, values up to 1:30 are typically observed during cold-air outbreaks. The aspect ratio is defined as the ratio between the wavelength or diameter of the pattern to their height. With boundary layer heights of 2 km, diameters of the convection pattern of about 60 km and more could be observed. \\\\
 Within the last two decades the phenomenon of cold-air outbreaks was increasingly studied using three-dimensional numerical models. These simulations suggested that adiabatic heat sources like latent heat release due to condensation are a condition precedent to observe large aspect ratios. Unfortunately, these simulations could not be carried out satisfactorily so far due to insufficient computer resources. To study the evolution of convective structures during cold-air outbreak situations the computational domains have to be large enough to contain the large cellular pattern, and they must have a spatial resolution fine enough in order to resolve smaller scales (single up- and downdrafts). Hence, the results of previous studies are afflicted with different uncertainties: Either the model domain was too small in order to allow a undisturbed cell growth or a too coarse resolution (up to 2 km) had to be used in order to be able to cover larger areas. Consequently it was unpossible to decide whether interactions between smaller and larger scales may affect the cell structure. \\\\