Abstract
We conducted dynamical downscaling simulations for 20 years over the Asian CORDEX domain to validate uncertainty of RSM-ROMS, a regional atmosphere and ocean coupled climate model. We compared some surface quantities simulated by RSM-ROMS with observation data sets or simulated result by RSM, a regional atmosphere climate model. The results are summarized as follows.
1) The coupled model had some systematic negative bias for observed sea surface temperature (SST), but the correlation coefficient was over 0.99 and the model reasonably captured the interannual and seasonal variations. In addition, power spectrum suggested a detailed SST structure.
2) The ocean coupling increased a tendency to overestimate precipitation over the mountainous area. On the other hand, the coupled model more accurately reproduced precipitation over the oceans.
3) Although there was cool bias of sea water temperature (SWT) more than 4 degrees Celsius at 100m depth, the coupled model captured interannual validations of SWT as similar as those of SST. This SWT bias is larger than seasonal variations, and it indicates there is uncertainty in the coupled model.
4) Net heat flux was improved in the coupled model, especially in subtropical and tropical areas but got worse in middle latitude areas.
