Figure 7a shows the time series of annual number of the model TCs over the North Atlantic. There is an interannual variation with a period of about 5 years in the time series (Fig. 7b). Although the variation on decadal time scales as in the actual TC frequency observed over the North Atlantic is not found in this model, there is an upward trend in annual number of the model TCs (0.5/decade). This increase will be discussed later.

It has been accepted that SST is one of the key environmental variables affecting the interannual variability of TC frequency. Thus, we first present the correlation maps of SST anomalies with model annual TC frequency in Fig. 8a. The correlation of SST anomalies is quite similar to the model ENSO SST anomaly pattern (c.f. Fig. 6), suggesting that the interannual variability of model TC frequency is related to ENSO. In the Atlantic, there is no significant correlation in the region between 10° and 20°N from the west coast of Africa to Central America and along the eastern coast of North America, where most model TCs are formed (c.f., Figs. 3b and 8a). Moreover, the negative correlation is found off Dakar. This implies that TCs cool SST through an increase in ocean mixing and evaporation. Therefore, it is suggested that the interannual variability of TC frequency in the Atlantic simulated by this CGCM is constrained by the large-scale atmospheric circulation rather than by the local SSTs.

30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Year

Fig. 7 (a) Time series of annual number of model TCs (solid line) and least squares best-fit linear trend is denoted by dashed line. (b) Power spectra of annual number of model TCs (solid line) and dashed curve shows the power spectra of red noise

30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Year

—1 | |||

—— |

(per year)

(per year)

Fig. 7 (a) Time series of annual number of model TCs (solid line) and least squares best-fit linear trend is denoted by dashed line. (b) Power spectra of annual number of model TCs (solid line) and dashed curve shows the power spectra of red noise

Previous studies (e.g. Gray 1984; Goldenberg and Shapiro 1996; Vitart and Anderson 2001) have shown that an inverse relationship exists between North Atlantic TC activity and vertical wind shear in the main TC developing region between 10° and 20°N from the west coast of Africa to Central America (MDR) on seasonal and longer time scales. Thus, we next examine the relationship between the vertical wind shear and TC frequency in the model. There is the region with negative correlation between 10°N-20°N over the western part of North Atlantic (Fig. 8b). The similar spatial pattern is found in the correlation map between vertical wind shear and Nino3 SST anomaly (Fig. 8c). Therefore, the change in the vertical wind shear through atmospheric bridge associated with ENSO would cause the variation in annual TC frequency simulated by this CGCM. The results are in good agreement with the relationship between ENSO and Atlantic TC frequency in the observations (e.g., Goldenberg and Shapiro 1996).

SST with TC Num.

SST with TC Num.

Fig. 8 Maps of correlation coefficients of SST (a) and vertical wind shear (b) averaged during the months from August to October against the annual number of model TCs. Contour interval is 0.2 and values exceeding the 99 % confidence level are shaded. (c) Same as in (b) but for correlation coefficients of vertical wind shear against Nino3 SST anomaly averaged during the months from August to October. Note that sign is reversed

Fig. 8 Maps of correlation coefficients of SST (a) and vertical wind shear (b) averaged during the months from August to October against the annual number of model TCs. Contour interval is 0.2 and values exceeding the 99 % confidence level are shaded. (c) Same as in (b) but for correlation coefficients of vertical wind shear against Nino3 SST anomaly averaged during the months from August to October. Note that sign is reversed r=-0.6

22 20 1B 16

Fig. 9 Scatter diagrams for annual TC number (a) and annual hurricane number (b) against Nino3 SST anomaly during August-October. Correlation coefficient is denoted in the top of each panel. Linear regression lines are denoted by solid lines

Figure 9 shows the scatter diagrams for annual TC number and annual Hurricane number against Nino3 SST anomaly during August-October. Here, the model tropical storms (TSs) and Hurricanes are defined as model TCs reaching maximum surface winds more than 17 and 33 m/s, respectively. It is noted that there are no TCs with maximum surface winds more than 43 m/s belonging to the category 2 on the Saffir-Simpson scale because the horizontal resolution of the model used in the present study does not have the ability to simulate such strong TCs (c.f. Oouchi et al. 2006).

The annual number of model Atlantic TCs tends to increases in model La Nina years while it seems to be below the average in model El Nino years (Fig. 9a). There is also the similar relation in the annual number of model Hurricanes (Fig. 9b). The regression coefficients of model annual TC number and annual Hurricane number against Nino3 SST anomaly are —0.94/°C and —0.45/°C, respectively. The result suggests that in model El Nino years there are fewer intense TCs than in model La Nina years, which is generally consistent with the observation (e.g., Landsea et al. 1999).

Was this article helpful?

This is a collection of 3 guides all about survival. Within this collection you find the following titles: Outdoor Survival Skills, Survival Basics and The Wilderness Survival Guide.

## Post a comment