Environmental Engineering Reference
In-Depth Information
4.10 The Ability of Mesoscale Models to Predict
Vertical Profiles
1
2
E. Batchvarova and S.-E. Gryning
1
NIMH, Bulgaria Academy of Sciences, Sofia, Bulgaria
2
Risoe DTU, Roskilde, Denmark
Abstract
In the manuscript it is argued that examination of model performance
should be based on profiles and that the traditional evaluation based on surface
measurements at one level easily can be misleading. A model can perform very
well near the surface, suggesting good performance, but poorly at 100 meters.
When comparing model results and measurements, it is important to emphasize
that excellent model performance does not require an excellent match with
measurements. The representativeness of the measurements should be taken into
account. These points are discussed and argued in the manuscript with examples
drawn from a recent model evaluation effort.
Keywords
Atmospheric boundary layer, Models' evaluation methods, Vertical
profiles, Tall meteorological masts, Flux measurements, Representativeness of
measurements
1. Introduction
As part of the COST 728 action (Enhancing Mesoscale Meteorological Modelling
Capabilities for Air Pollution and Dispersion Applications) a major model
comparison and evaluation exercise is carried out.
One of the cases covers Central and Northern Europe in February and March
2003, when several PM10 episodes were observed. The predictions of several
models were shown to differ widely by Stern at al. (2008). Detailed analysis of
meteorological conditions in the same paper pointed out large differences in the
boundary layer height used in the chemical models.
Although the boundary-layer height plays a central role for the PM10 predictions
by chemical transport models and is often an output parameter from meso-scale
meteorological models, the way it is obtained is not transparent. Therefore when
modeled and measured heights of the boundary layer are compared it is often not
clear if the values are based on the same definitions.
This problem can be examined by evaluating the vertical profiles of meteoro-
logical parameters, such as wind speed and direction, humidity, temperature. In
