Papers in Meteorology and Geophysics

Show abstractHide abstract

   Ten-year integrations of the JMA global model at T106 and T42 horizontal resolutions were compared to examine the role of model resolution in simulating the Indian monsoon and its variability. It has been found that both T106 and T42 models simulate mean monsoon climatology reasonably well in terms of the large-scale monsoon flow. The T106 model simulates rainfall distribution in western and northwestern parts of India better than the T42 model.
   Correlation of interannual variability (IAV) of simulated rainfall with observed IAV is poor. The simulated IAVs of the Indian monsoon rainfall for both models do not have a strong correlation with the SST, either. The T42 model simulated the rainfall variability for 1987/88 better than the T106 model. Monsoon rainfall variability may be largely due to internal dynamics in both models. Internally generated variability may be larger in T106 model than T42 model simulations.
   Rainfall anomaly patterns obtained from T42 model simulations are better than those of T106 model simulations for the two pairs of good and bad monsoon events (1988/87 and 1983/82). Responses of both the coarser and finer resolution models to the same imposed surface forcing differ over the Indian monsoon region.
   Both models simulate the synoptic evolution of the monsoon quite well. Monsoon activity in the T106 model is more intense than in the T42 model. Rainfall distribution is better obtained from T106 model simulations than T42 model simulations. This suggests that further work is necessary in relating intraseasonal variations with interannual monsoon variability for understanding the nature of monsoon variability.

View full abstract

Download PDF (4174K)

   Tide and storm surges in the Seto Inland Sea were numerically simulated in a σ-coordinate system using the Princeton Ocean Model (POM). For tide simulation, we obtained good results compared to the astronomical tide, particularly overall amplitude and phase features. For storm surge simulation, results from the model show sea surface elevation time series similar to that observed at each station. Several differences also arise, however. The discrepancy is attributable to the sea surface wind field over a wide area and the lack of sufficient resolution. More important ingredients are involved in reproducing the coastal region storm surge than nonlinearity or 3-dimensionality, however.
   Maximum sea surface elevation values for each location calculated using a 3-dimensional, stratified model were found to be up to about 10% larger than those calculated with a 2-dimensional, barotropic model. A physical explanation is proposed here.
   Tidal motion and storm surge are found to be almost independent in our simulations; that is, results are obtained merely by superposing the simulation driven by each independent forcing.

View full abstract

Register with J-STAGE for free!