Overview of domain experiments SMHI (RCA3) GKSS (CLM) KNMI (RACMO2) MPI (REMO)

Overview of domain experiments

SMHI (RCA3)

GKSS (CLM)

KNMI (RACMO2)

MPI (REMO)

Overview

• SMHI: Mar-Nov 1997, 0.44/0.4° resol

• GKSS: Mar-Nov 1997, 0.44/0.4° resol, spectral nudging on/off

• KNMI: 1995, 0.44° resol, various rel.zones

• MPI: Mar-Nov 1979

SMHI: bias MSLP

• July

ENS PRUD PRUD + SPRUD + S+N

SMHI: bias MSLP

• August

ENS PRUD PRUD + SPRUD + S+N

SMHI: bias MSLP

• September

ENS PRUD PRUD + SPRUD + S+N

SMHI: conclusion

• Especially in July the ENS domain gives rise to rather large bias in MSLP. The bias is reduced later in the season.

• Weaker gradients give larger bias

• Initial conditions are not important.

GKSS MSLP results

• May 1997

GKSS MSLP results

• July 1997

GKSS MSLP results

• Sep 1997

Effect of spectral nudging

Large domainSmall domain

GKSS conclusions

• Bias seems smaller than SMHI (but is domain averaged)

• Patterns differ from SMHI

• Spectral nudging and smaller domain keep bias lower, but still quite large bias in autumn

small domain large domain

KNMI domain; mean over E-W transect

Small domain PXS, longitude-time section G500 anomaly

stronger BC forcing weaker forcing

tim

e

west east

Small domain PXS, longitude-time section G500 anomaly

stronger BC forcing weaker forcing

tim

e

west east

stronger BC forcing weaker forcing

Large domain PXL, longitude-time section G500 anomaly

tim

e

west east

Conclusions KNMI

• Bias in June in centre of domain

• Strength of relaxation AND domain size play a role:– Stronger relaxation in large domain Weaker

relaxation in small domain

• Explanation: with weak background forcing Rossby wave able to remain stationary

MPI 1979 exp: MSLP bias

MPI conclusion

• Development of strong dipole structure until Sep

• High bias in June/July

• After Sep: reset and new structure pattern

Overall conclusion

• All models produce high pressure bias in early summer: this seems a systematic result of RCM modelling

• Effect is increased likelihood of ridge with blocking circulation and potentially enhanced summer drying!

• Jens has a possible explanation for this (needs systematic exploration)

Arguments pro/con large domain

• Pro: – more variability in ensemble due to RCM-differences

– better development of low pressure systems (?), higher wind speeds

– potentially more impact assessment applications served

• Con: – systematic high pressure bias in summer

– difficult to compare with obs, even driven with ERA40

– more computer resources needed