Although an interannual variation with a period of about 5 years is dominant in the time series of annual number of the model TCs over the North Atlantic (Fig. 7), there is also an upward trend in annual number of the model TCs (Fig. 7a). Since no any trend is found in time series of model Nino3 SST anomaly (Fig. 5), other forcing might induce the change. In the present CGCM, an increase of any anthropogenic forcing is not considered. Therefore, one may consider that the SST change induced through slow oceanic adjustment from the initial state is related to the upward trend of model TC frequency. Thus, we computed the correlation between the SST and model TC frequency after applying a 7-year low-pass filter to both variables to remove ENSO signal.
In contrast to the correlation of the non-filtered SST and Atlantic TC frequency (Fig. 8a), there is the significant positive correlation in the region between 10° and 20°N from 30°W to Central America (Fig. 10a). The positive correlation is found also in the mid-latitude North Atlantic. The similar pattern is found in the trend and the 2nd EOF mode of SST anomalies (not shown). This also seems to be somewhat similar to the spatial pattern of SST anomalies associated with the Atlantic Multi-decadal Oscillation (AMO) (e.g., Goldenberg et al., 2001). It is noted that there is no significant correlation in the filtered vertical wind shear (Fig 10b). These imply that warmer SSTs in the region between 10° and 20°N over the North Atlantic may have potentially impacts on an increase in Atlantic TC frequency.
Was this article helpful?