The potential contribution of TCs to the ISV was estimated by calculating the difference between the ISV was the original and the TC-removed vorticity at 850 hPa. As demonstrated in Fig. 2 and Hsu et al. (2008), the Kurihara scheme cleanly separates TCs from the background large-scale circulation. The variance contributed by TCs and large-scale circulation are also well separated. The original 32-76-day variance of 850 hPa vorticity presented in Fig. 8(a) reveals two major ISV regions along the TC tracks: one elongated region in the Philippine Sea and the other south of Japan. The former corresponds to the northwestward TC tracks, while the latter corresponds to the recurving TC tracks toward Japan. After the removal of TCs, the variance in these two regions reduced dramatically, as shown in Fig. 8(b), while the variance outside these two regions

remained almost unchanged. The variance difference shown in Fig. 8(c) indicates that the amount of the variance contributed by the clustered TCs can be as large as 50-80%.

To illustrate how the TCs contribute to the ISV in the statistical sense, the time series of the original and TC-removed 850 hPa vorticity was averaged over 125°E-140°E, 15°N-22.5°N, where the maximum variance is observed, and the difference between the two is shown in Fig. 9(a). Positive vorticity peaks occurred in groups in June, August, and October, indicating the clustering of TCs, while weak negative vorticity was observed during the TC-inactive period (July and September). Removal of TCs apparently results in a significantly reduced amplitude of the positive vorticity peaks, but it has no effect on negative vorticity for an obvious reason. This leads to the overall enlargement of positive vorticity during the TC-active periods, which recurred on the intraseasonal time scale, and therefore the increased ISV for the whole JJASO season. For example, the ISV in the above region drops from 5.9 x 10-11 to 2.2 x 10~n s~1 after the TC removal. This significant reduction can also be seen in the spectra of the area- averaged vorticity presented in Fig. 9(b). While the original and the TC-removed spectral density exhibit qualitatively

Figure 9. (a) Time series of the area-averaged 850 hPa vorticity over 125°E-140° E, 15° N-22.5° N from June to October 2004. The dotted and dashed lines represent the original and TC-removed vorticity fields, respectively, while the thick solid line represents the difference. (b) Spectra density of the original (dark solid line) and the TC-removed (light solid line) time series shown in (a). Dashed lines represent the 95% confidence limits. Fluctuations with periods longer than 120 days were removed before calculating the spectral density.

Figure 9. (a) Time series of the area-averaged 850 hPa vorticity over 125°E-140° E, 15° N-22.5° N from June to October 2004. The dotted and dashed lines represent the original and TC-removed vorticity fields, respectively, while the thick solid line represents the difference. (b) Spectra density of the original (dark solid line) and the TC-removed (light solid line) time series shown in (a). Dashed lines represent the 95% confidence limits. Fluctuations with periods longer than 120 days were removed before calculating the spectral density.

similar distribution in frequency, the intrasea-sonal peak (statistically significant at the 0.05 level) drops by about 50% in the TC-removed case. In addition, removal of TCs also results in a reduction of the seasonal mean vorticity. For example, the seasonal mean vorticity in the region chosen for Fig. 4(a) is 2.1 x 10~5 in the original vorticity but reduces to almost zero in the TC-removed vorticity. These results indicate that the presence of TCs in the tropical western North Pacific not only enlarges the ISV

but also increases the seasonal mean vorticity along the TC tracks. In a recent study, Hsu et al. (2008) demonstrated that the presence of TCs enhances not only the ISV but also the inter-annual variance.

In view of the large reduction in the ISV after the removal of TCs, one would wonder whether the propagation of the ISO is affected. Figure 10 presents the latitude-time Hovmiiller diagram of the 850 hPa vorticity averaged between 120°E and 150°E, where the ISO exhibits obvious northward propagation. The original ISO exhibited two cycles of oscillation, with the largest amplitudes near 5°N and 15°N, and a node near 10°N [Fig. 10(a)]. In between the occurrence of maximum amplitudes, northward propagation from 5°N to 15°N was evident. After the removal of TCs, the amplitude of the vor-ticity weakens significantly between 10°N and 20°N, where TC tracks were located, while the pattern and amplitude at other latitudes are nearly unchanged [Fig. 10(b)]. The northward propagation from 5°N to 15°N, which is the major characteristic of the ISO in this region during the boreal summer, is still evident.

Since the TCs' effect on the ISO mainly occurred in the 10°N-20°N latitudinal band, the longitude-time Hovmuiller diagram of the 850 hPa vorticity averaged between 10° N and 20°N was examined. In Fig. 10(c), two cycles of the ISO with the maximum amplitude at various longitudes are evident. There are signs of westward propagation in the early half of summer, especially over 120°E-140°E, and eastward propagation in the latter half of summer. After the removal of TCs, the amplitude of the ISO reduces significantly, while the pattern remains similar [Fig. 10(d)], although there are indications of weakened eastward propagation between 140°E and 155°E in the latter half of summer. Overall, the most significant effect of TCs on the ISO is the large reduction in the amplitude. In comparison, the TCs' effect on the phase and propagation of the ISO is minimal.

Figure 10. Hovmtiller diagrams for the (upper left) original and (upper right) the TC-removed 32—76-day filtered 850 hPa vorticity averaged over 120° E—150°E, and for the (lower left) original and the (lower right) TC-removed 32— 76-day filtered 850 hPa vorticity averaged over 10°N-20°N. The contour intervals are 3 X 10~6 and 2 X 10~6 for the upper panel and the lower panel, respectively.

Figure 10. Hovmtiller diagrams for the (upper left) original and (upper right) the TC-removed 32—76-day filtered 850 hPa vorticity averaged over 120° E—150°E, and for the (lower left) original and the (lower right) TC-removed 32— 76-day filtered 850 hPa vorticity averaged over 10°N-20°N. The contour intervals are 3 X 10~6 and 2 X 10~6 for the upper panel and the lower panel, respectively.

Was this article helpful?

## Post a comment