In this note we present the
results of the verification of LM in Greece concerning 2m Temperature and 10m
Wind’s Speed for the
time periods February - March and June - August 2001.
The direct outputs encounter serious systematic
errors due, mainly, to the inability of the model to handle successfully
sub-grid scale phenomena as well as to the complexity of the morphology of
orography.
In order to
eliminate these errors we use a polynomial (multidimensional) Kalman filter
which is based on the values of only one parameter and on a backlog of 7 days.
The main advantages of this method are the easy adaptation to any alteration
of the form of data as well as its simplicity, since it can be applied at any
simple PC needing minor technical support.