Modeling a heavy rainfall case in North-East Estonia, August 2003. Influence of model resolution.

Modeling a heavy rainfall case in North-East Estonia, August 2003.

Influence of model resolution.

Andres LuhamaaUniversity of Tartu

2005

Overview of the event

● Precipitation amount: in one day 1.5 times more than monthly mean.

● High precipitation rates measured only in one (Estonian station)

Radar observations, aug 05, 17.00 – aug06, 06.00

Forecasts

● EMHI – general storm warning for Gulf of Finland, nothing more.

● FMI 22km forecast estimates high precipitation at the northern coast of Estonia.

Modelling experiment

● Influence of model resolution – 22km, 11km, 5km, 3km

● HH and NHH– Reference HIRLAM and non-hydrostatic

● Different versions of STRACO cloud scheme

– HIRLAM 6.1.0 - STR1– 6.1.2 (last major updates) - STR2– Research version for high resolutions (<5km)

called straco031 - STR3

FMI 22km forecast 11km STR-1 simulation

HH 11km simulation NHH 11km simulation

STR-2

HH 5km simulation NHH 5km simulation

STR-1

HH STR-2 5km simulation

NHH STR-2 5km simulation

NHH STR-3 5km simulation

NHH STR-1

NHH STR-2

3-km simulations

NHH STR-3

Conclusions:

● Nothing better than 22km HH forecast could be achieved ;(

● Significant difference between NHH and HH precipitation amounts already at 11km resolution!

– Parameterization tuning for HH->NHH modeling?

– As difference is only in amounts, is NHH necessary at coarser resolutions?