Notes
Outline
Some Results of Verification of LM
against Observations
Hellenic National Meteorological Service
George Galanis
E-mail: galan@hnms.gr
Abstract.
LM in Greece
Stations under study :
Athens’ Temperature (06 UTC) June – August
Athens’ Temperature (12 UTC) June – August
Thessaloniki’s Temperature (06 UTC) June – August
Heraklion Temperature (18 UTC) June – August
Kozani’s Temperature (12 UTC) June – August
Time Period: June – August 2001
We encountered significant systematic errors, especially at near coast locations during the evening hours.
Main reason: Some of the grid points used in the relevant interpolation schemes have totally different characteristics from the area in study (e.g. lay over the sea).
The Kalman filtering used gives excellent results towards the improvement of the final forecast, since it :
-  eliminates the systematic error
- decreases the standard deviation of the absolute error
Athens’ Temperature (18 UTC) February – April
Athens’ Temperature (12 UTC) February – April
Thessaloniki’s Temperature (06 UTC) March – April
Heraklion Temperature (18 UTC) February – April
Time Period: February – April 2001
The systematic error in the time series of this period is seriously decreased compared to the relevant of summer months
The proposed Kalman filter keeps giving serious help towards the elimination of this error.
Thessaloniki’s Wind’s Speed (06 UTC) June – August
Athens’ Wind Speed (12 UTC) June – August
Santorini’s Wind Speed (12 UTC) June – August
Heraklion Wind’s Speed (18 UTC) June – August
Verification of Wind’s speed
Time Period: June – August 2001
The direct outputs of LM do not encounter serious systematic errors.
In those cases where, even slightly, errors emerge, Kalman filter improves the final results.