Performance of Different Sub-Grid-Scale Surface Flux Parameterizations for Urban and Rural Areas - Meteorological and Air Quality Models for Urban Areas - page 17

Chapter 3
Performance of Different Sub-Grid-Scale
Surface Flux Parameterizations for Urban
and Rural Areas

Sylvia Bohnenstengel and Heinke Schlünzen

Abstract With increasing computing power resolutions of mesoscale atmospheric
models increase. However, this does not necessarily improve model performance
as evaluations show. Especially, in very heterogeneous areas like urban areas, where
densely built areas with sealed surfaces and irrigated gardens and parks occur within
the same grid box the aggregation effect impacts model performance. To ensure a
good forecast of pollutant concentrations an accurate prediction of the flow field
and the vertical mixing is essential. Two schemes to parameterize sub-grid-scale
surface fluxes are applied to the area around Berlin for different meteorological
situations and grid resolutions. Parameter averaging and a flux aggregation method
with blending height concept are validated for grid resolutions 4, 8 and 16 km.
Model performance is determined by calculating hit rates from the German Weather
Service (DWD) routine data.

3.1 Introduction

The parameterization of surface fluxes with inclusion of sub-grid-scale land-use
becomes substantially important in very heterogeneous areas like urban ones. The
so-called “aggregation effect” (Giorgi and Avissar, 1997) affects the representa-
tiveness of the grid box averaged fluxes, since the system acts highly non-linear.
In urban areas different land-use types, which are highly varying in their surface
characteristics, occur within one grid box. A good forecast of concentration does
not only depend on emissions and chemical reactions, but also on the flow field.
The latter is investigated for the area of Berlin. Like other urban areas the urban
geometry results in lower wind speeds leading to a less intense vertical mixing in
contrast to suburban or rural areas. In the densely built areas the latent heat flux
is lower than in suburban and rural areas resulting in high Bowen ratios. Thus,