Figure 1 shows the composite distributions of pentad mean STEA after the dates when the selected 144 TCs entered the North Pacific from October to December in the period 1979–2019. It can be seen that the strongest STEA is located at the upper levels, and it decreases down to the lower levels. The maximum STEA at 300 hPa is located around 45°N, 170°W, with a peak value above 20 m–2 s–2 (Fig. 1a). The STEA at 500 hPa exhibits a zonal distribution east of (45°N, 150°E) (Fig. 1b), and the maximum intensity of STEA at 850 hPa exhibits a southwest–northeast oriented pattern, with a positive center maximum value of about 8 m–2 s–2 near 45°N, 165°E over the KOE region (Fig. 1c). It is noteworthy that the spatial distribution of STEA at 850 hPa is apparently different from the spatial distributions at 500 hPa and 300 hPa. The distribution of STEA associated with TCs at 850 hPa is roughly located over the mean path of most of the TCs which take northeastward tracks after recurving and entering the KOE region. We speculate that the STEA in the lower troposphere might be related to the momentum transfer by the TC vortices. At the mid-to-upper levels, the amplification of eddy perturbations is controlled by a preexisting Rossby wave train that disperses downstream and modifies the large-scale flow patten associated with the TC–extratropical flow interactions (Archambault et al., 2015). Additionally, the spatial patterns of the climatological STEA from October to December are quite similar to those connected with the TCs that enter the KOE region at all levels of the troposphere (not shown).
To explore the relationship between the STEA and TC intensity, Fig. 2 shows the scatter diagram of the TC EKE with the pentad mean STEA for the selected 144 TC cases. On the whole, the positive linear correlation is significant in October and November, but significance is not evident in December due to the small sample size (Figs. 2c, f, and i). A linear relationship of TC EKE with the corresponding STEA intensity can be seen in October and November from the regression polynomial (Figs. 2a, b, d, e, g, and h). STEA intensity is significantly correlated with TC EKE, with a correlation coefficient of 0.39/0.32/0.41 (0.49/0.64/0.68), exceeding the 95% confidence level, at 300 hPa/500 hPa/850 hPa in October (November) (Table 1). Due to the very small number of TC cases in December, there is almost no linear relationship between TC EKE and STEA. In general, TC intensity is highly indicative of the subsequent STEA, with a correlation coefficient of 0.37/0.33/0.45, exceeding the 99% confidence level, at 300 hPa/500 hPa/850 hPa from October to December for the period 1979–2019 (Table 1). Meanwhile, STEA reaches its most significant correlation with TC EKE in the lower troposphere from October to November. This is possible because TC activity, accompanied by abundant kinetic energy, is transported into the target region at the lower levels, favoring more transient eddy development over the North Pacific and thus combining the lower-level STEA with TC EKE. Most STEA cases in November are located more closely around the linear regression polynomial, indicating that the intensity of STEA associated with TC activity in November is stronger than that in October and December. This is partially attributed to the peak strength of STEA over the North Pacific occurring in November, as well as simultaneous intense TCs entering into this region when the climatological STEA values are already high (not shown).
Figure 2. Scatter diagrams of TC eddy kinetic energy (EKE; m2 s−2) with pentad mean STEA (m2 s−2) at 300 hPa [(a), (b), (c)], 500 hPa [(d), (e), (f)], and 850 hPa [(g), (h), (i)] for the selected 113 TC cases in October (left), 24 TC cases in November (middle), and 7 TC cases in December (right). Red lines are linear regression curves, and the correlation coefficients between the TC EKE and STEA are listed in Table 1.
October November December October to December 300 hPa 0.39** 0.49* 0.23 0.37** 500 hPa 0.32** 0.64** –0.01 0.33** 850 hPa 0.41** 0.68** –0.10 0.45** * Significant at the 95% confidence level.
** Significant at the 99% confidence level.
Table 1. Correlation coefficients matrix between the TC EKE and corresponding STEA from October to December in the period 1979–2019.
To further examine the influence of TC activity on the spatial variation of STEA intensity, results of the regressions performed on the TC EKE series of the 144 TC cases for pentad mean STEA in the troposphere are shown in Fig. 3. It can be seen that although the linear contribution of TC activity to STEA exhibits a spatial discrepancy between different atmospheric levels, the positive impacts of TCs on STEA are consistent throughout the entire troposphere, and this is highlighted by the robust positive regression coefficients. As shown in Fig. 3a (3b), the most significant regression coefficients in association with strong STEA are characterized by a tripolar pattern, with three maximum centers from west to east, located around (45°N, 140°E), (45°N, 175°E), and (45°N, 170°W) [(42°N, 160°E), (40°N, 170°E), and (40°N, 170°W)] at 300 hPa (500 hPa), respectively. Yet, the evident positive regression relationship between the TC EKE and STEA at 850 hPa is largely concentrated west of 165°E, exhibiting a northeast–southwest-oriented pattern from south of Japan to the northwestern WNP. These results suggest that TC activity can exert great influences on a broad area at the mid-to-upper levels, migrating eastward to the Northeast Pacific. Meanwhile, in the lower troposphere, the impact of TCs on STEA is primarily localized over the KOE region. Additionally, based on the distribution of standard regression coefficients, the significant positive contribution of TCs to STEA at the upper levels is located over the central North Pacific (Fig. 3a), which is more extensive than the contribution at the mid-to-lower levels (Figs. 3b and c). The results of monthly regression show that the contribution of TCs to STEA in October is largest in late autumn and early winter, and the contribution of TCs in November and December is less based on the significant region (Figs. 3g–3l), as there are fewer TC samples in these two months.
Figure 3. Simultaneous regression upon normalized TC EKE for pentad mean STEA (shading; m2 s−2) at 300 hPa [(a), (d), (g), (j)], 500 hPa [(b), (e), (h), (k)], and 850 hPa [(c), (f), (i), (l)] for the selected 144 TC cases during October–December (left), 113 TC cases in October (left-center), 24 TC cases in November (right-center), and 7 TC cases in December (right). Anomalies enclosed by gray contours denote the coefficients are significant above the 95% confidence level by the F test.
We also examine the interannual change of the relationship between the TC EKE and STEA by calculating the interannual means of TC EKE and pentad mean STEA cases from 1979 to 2019. Figure 4a shows the scatter diagrams of annual-averaged TC EKE and STEA for 41 years at different atmospheric levels during 1979–2019, respectively. It is noted that the TC EKE index is significantly correlated with the subsequent STEA at 300 hPa/500 hPa/850 hPa/all level over the North Pacific at the interannual time scale, with a correlation coefficient of 0.52/0.44/0.49/0.47, exceeding the 99% confidence level. The strongest correlation on the interannual time scale is located in the upper troposphere, which reflects a long time-scale variability and is different from the results shown in Table 1. Based on the slope of the linear regression curve, STEA in the upper troposphere (300 hPa) exhibits a more significant linear growth relationship with TC EKE than that in the mid-to-lower levels during the period after cyclones enter the KOE region (Fig. 4a). This suggests that the impact of TCs on STEA has an effect that gradually increases with height, which is closely related to the interaction between TCs and the westerly jet in the upper troposphere (Archambault et al., 2015; Chen et al., 2017). Moreover, the results of simultaneous regression exhibit a strong spatial variation of STEA at 300 hPa connected with TC activity broadly located over the western and central North Pacific (35°–45°N, west of around 180°; Fig. 4b), which is similar to the significant regions at 500 hPa and 850 hPa (Figs. 4c and 4d). This agreement throughout the entire troposphere suggests a consistent feature at the interannual time scale of TC influence on STEA changes over the North Pacific. This interannual influence is probably connected with the development of downstream wave packets associated with the TCs in the extratropical baroclinic storm-track region in a climatological sense (Archambault et al., 2015; Keller et al., 2019).
Figure 4. (a) Scatter diagrams of TC EKE (m2 s−2) with pentad mean STEA (m2 s−2) from October to December for 41 years in the period 1979–2019. Lines are linear regression curves at the different levels. Simultaneous regression upon annual-mean normalized TC EKE for pentad mean STEA (shading; m2 s−2) at 300 hPa (b), 500 hPa (c), and 850 hPa (d) from 1979 to 2019. Anomalies enclosed by gray contours in (b)–(d) denote the coefficients are significant above the 95% confidence level by the F test.
Figures 5a, e, and f show the temporal strength distributions of the pentad mean STEA overlapped with the TCs when they enter the KOE region from October to December in the period 1979–2019. During the annual cycle, there is a minimum in the strength of the climatological STEA over the North Pacific in midwinter (Nakamura, 1992). Before the midwinter suppression, STEA reaches its strongest intensity during autumn to early winter (October through December). In particular, the strongest STEA is located at the upper levels and decreases with reduced height. The total STEA over the North Pacific from October to December displays interannual (Figs. 5b and f) and interdecadal variations (Fig. 5j). Figures 5c, g, and k show the interannual series of STEA associated with the entered TCs at 300 hPa, 500 hPa, and 850 hPa, respectively. There is a high linear correlation between their changes, indicating that the effect of TCs on STEA is consistent throughout the whole troposphere on the interannual time scale. In order to further study the quantitative contribution of TCs to STEA, the ratios of STEA associated with the entered TCs to the total STEA at the different levels are calculated and shown in Figs. 5d, h, and l. From the October–December climatic average, the ratio in the North Pacific is 4.24%/4.48%/6.18% at 300 hPa/500 hPa/850 hPa. That is, the contribution of TCs to STEA in late autumn and early winter reaches about 4%–6%. In addition, the contribution of TCs to the STEA in autumn and early winter experienced an interdecadal change in the early 2000s, which can be detected at each level of the troposphere. Compared to the higher ratios before the 2000s, there is a marked decrease from the early 2000s through the early 2010s (Figs. 5d, h, and l). In order to clarify the spatial characteristics of the interannual variation of TC influence on STEA, the anomalies of STEA in the composited TC high-impact and low-impact years in the period 1979–2019 are presented in Fig. 6. In general, the anomaly distribution of STEA presents the opposite phase in the TC high-impact and low-impact years, and the positive anomalies of STEA have a wider range during the high-impact years (Figs. 6a, c, and e). In addition, the significant area affected by the anomaly is different between east and west and between upper levels and lower levels.
Figure 5. Distributions of pentad mean STEA intensity (shading; m2 s−2) averaged over the North Pacific (30°–60°N, 130°E–160°W) at 300 hPa (a), 500 hPa (e), and 850 hPa (i) overlapped with the red TC symbols when a TC entered the Kuroshio/Oyashio Extensions. Interannual variations of the total STEA [m2 s−2; (b), (f), (j)], the STEA associated with the entered TCs [m2 s−2; (c), (g), (k)], and the ratio of STEA associated with the entered TCs to the total STEA [%; (d), (h), (l)] at 300 hPa (upper-right), 500 hPa (middle-right), and 850 hPa (lower-right) from October to December. The red lines and percentages denote the averaged ratio of STEA associated with the entered TCs to the total STEA in the period 1979–2019.
Figure 6. Anomalies of pentad mean STEA (shading; m2 s−2) in the composited TC high-impact years [(a), (c), (e)] and low-impact years [(b), (d), (f)] in the period 1979–2019 at 300 hPa (top), 500 hPa (middle), and 850 hPa (bottom) based on the thresholds with one standard deviation of normalized ratios of the STEA associated with TCs to the total STEA in Figs. 5d, 5h, and 5l. The high-impact (low-impact) years are 1992, 1993, 1994, 1996, 2002, and 2013 (1982, 1983, 1985, 1991, 2005, 2006, 2011, and 2017). Differences enclosed by solid contours denote that the anomalies are significant above the 95% confidence level by the two tailed Student’s t test.
In this study, to establish the quantitative relationship between TCs and STEA, we calculate the TC EKE before a cyclone crosses through the boundaries (130°E or 30°N) of the North Pacific region, and we calculate the pentad mean STEA beginning on the date when the selected TC enters the KOE zone. This is a traditional method to parameterize the transient eddy activity in synoptic disturbances, including TCs (Ren et al., 2010; Ha et al., 2013). In fact, considering the definition of STEA, which is calculated by the variance of meridional velocity averaged for a certain period of several days, we also detect the lagging relationship between TC EKE and STEA defined by the different period-means. Figure 7 shows the correlation coefficients between TC EKE and STEA calculated by the n day-mean; please note that the results presented in this paper correspond to the condition of n = 5 (pentad mean). It can be seen that STEA defined by the mean variance within a period of 12 days-mean/10 days-mean/13 days-mean all have significant linear relationships with the TC EKE at 300 hPa/500 hPa/850 hPa, which can exceed the 95% confidence level (Fig. 7). This suggests that from a climatology perspective, the impact of TC activity on STEA over the North Pacific can last as long as two pentads after a TC enters the mid-latitudinal region. Meanwhile, the effective temporal lengths of TC influence on the STEA at the lower and upper levels are slightly longer than that at the midlevel (12 days/11 days vs. 9 days). In addition, we also calculated the relationship between TCs and STEA defined by the synoptic transient eddy EKE at different levels, and similar conclusions can be drawn (not shown). This suggests that the relationship revealed in this study between TC activity and the strength of STEA over the North Pacific is independent of the definition of STEA.
|October||November||December||October to December|
|* Significant at the 95% confidence level.|
** Significant at the 99% confidence level.