1. Introduction

In the boreal summer, the Asian jet is climatologically enhanced in the upper troposphere along the northern fringe of the enhanced Tibetan high associated with the activated summer Asian monsoon. The enhanced Asian jet works as an effective Rossby waveguide (Hoskins and Ambrizzi 1993), contributing to frequent Rossby wave propagation. Lu et al. (2002) and Enomoto et al. (2003) showed the existence of a wave train associated with the Rossby wave propagation along the Asian jet, and Enomoto et al. (2003) referred to it as the Silk Road pattern. Kosaka et al. (2009) showed using empirical orthogonal function (EOF) analysis that the Silk Road pattern is extracted as principal components of the upper-tropospheric atmospheric variability over southern Eurasia based on monthly dataset of the Japanese 25-Year Reanalysis (JRA-25; Onogi et al. 2007), indicating that the wave propagation along the Asian jet is one of the essential factors contributing to East Asian summer climate (Enomoto et al. 2009).

It is well known that the Rossby wave propagation from Eurasia along the Asian jet frequently causes Rossby wave breaking (RWB) near the Asian jet exit region (e.g., Postel and Hitchman 1999, 2001; Enomoto 2004; Abatzoglou and Magnusdottir 2006; Hitchman and Huesmann 2007), accompanied by mixing of potential vorticity between the tropics and extratropics (Homeyer and Bowman 2013). Bowley et al. (2019) showed a climatological maximum of the anticyclonic RWB frequency over the North Pacific midlatitudes east of Japan during summer. The RWB is associated with an amplified blocking anticyclone near Japan, with its equivalent barotropic structure, referred to as the Bonin high (Enomoto et al. 2003). The enhanced Bonin high influences anomalous hot summer conditions over Japan (e.g., Ogasawara and Kawamura 2007; Enomoto et al. 2009). The interannual variability of the RWB frequency near Japan is thus an important factor to predict seasonal climate during the boreal summer over Japan. Previous studies showed influences of the frequent anticyclonic RWB over the North Pacific on its farther downstream anomalous circulation such as the North Atlantic Oscillation (NAO) (Strong and Magnusdottir 2008), and sea surface temperature (SST) over the North Pacific and the resulting phase modulation of the Pacific decadal oscillation (PDO; Mantua et al. 1997) (Strong and Magnusdottir 2009). However, the dynamical mechanism that may cause the interannual variability of the RWB frequency itself from near Japan to the east has not been clarified as yet. Although Bowley et al. (2019) indicated that the anticyclonic RWB frequency over the North Pacific is significantly correlated with El Niño–Southern Oscillation (ENSO) indices, the related process has not been revealed.

This study investigates the large-scale oceanographic and atmospheric circulation associated with the interannual variability of the RWB frequency through linear regression analysis, focusing on the influence of ENSO on the RWB frequency. This line of approach is important to further understand the ENSO impact on the summer Asian monsoon and the related large-scale atmospheric circulation, and climate conditions over East Asia during the boreal summer.

The reminder of the present paper is organized as follows. Section 2 describes the dataset and analytical methods. In section 3, the climatology and ENSO composite of the RWB frequency are described to show that the RWB is favorable and exhibits large interannual variation near Japan during the boreal summer. In section 4, results of the linear regression analysis for the RWB frequency in August are provided to show large-scale oceanographic, atmospheric circulation, and boreal summer climate conditions in Japan associated with the frequent RWB, through the influences of ENSO. Section 5 provides the main findings and discussion of the results in this study.

2. Data and methods

The data used in this study are those from daily and monthly mean datasets of the Japanese 55-Year Reanalysis (JRA-55) for July–August during the 61-yr period from 1958 to 2018, with a horizontal resolution of 1.25° and 37 pressure levels from 1000 to 1 hPa (Kobayashi et al. 2015). We also used the monthly mean dataset of COBE-SST (Ishii et al. 2005) for the same months during the 61-yr period, with a resolution of 1°, to analyze SST. Here, anomalies are defined as a departure from the climatology, which is obtained as the 30-yr monthly averages from 1981 to 2010. We applied a horizontal smoothing filter to relative vorticity fields using a triangular truncation retaining N = 24 wavenumbers (T24) to exclude the disturbances at a scale smaller than synoptic eddies, as with the method of Takemura and Mukougawa (2020, hereafter TM20). A similar pattern is obtained even if the wavenumbers are slightly altered (not shown).

The region of Japan has four major divisions¹: northern Japan, eastern Japan, western Japan, and Okinawa/Amami. Monthly mean temperature anomalies averaged over the four regions were obtained from the Japan Meteorological Agency (JMA) database, which is available on the JMA website.

The propagation of Rossby wave packet is analyzed using the wave activity flux defined by Takaya and Nakamura (2001). The horizontal component of wave activity flux is defined as follows:

where u is the zonal wind, υ is the meridional wind, and ψ is the geostrophic streamfunction at a reference latitude of ϕ₀ = 40°N, at a pressure level. The overbars (primes) denote the basic states (perturbations), defined as the climatology (anomalies). The term $\overline{\mathbf{U}}=\left(\overline{u},\overline{\upsilon }\right)$ is the climatological wind vector. The subscripts x and y denote the partial derivatives with respect to longitude and latitude, respectively. The wave activity flux is derived from the monthly average.

To assess the waveguide of the Rossby wave packets, the stationary Rossby wavenumber K_s (Hoskins and Ambrizzi 1993) is calculated from the zonal wind. The wavenumber K_s is defined as follows:

where β is the meridional gradient of the planetary vorticity, u_B is the zonal wind with zonal wavenumbers k < 3 representing the basic zonal flow, and ${\beta }^{*}\equiv \beta -{\partial }^2{u}_B/\partial y^2$ is the meridional gradient of the absolute vorticity referred to as effective β. To examine the diffluence and deceleration of basic flow near the Asian jet exit region, which is a preconditioning of the RWB (Colucci 2001), the anomalous planetary-scale stretching deformation $d_p^{\prime }$ was calculated from the horizontal wind as follows:

which is defined according to Mak and Cai (1989) and Bluestein (1992). Here, the subscript p denotes the planetary-scale component defined by zonal wavenumbers k < 3 in this study. This wavenumber decomposition can exclude flows due to the Rossby waves. The anomalous diffluent basic flow thus corresponds to anomalous negative $d_p^{\prime }$ (Colucci 2001).

We calculated advection term of Rossby wave source (Sardeshmukh and Hoskins 1988) to assess an influence of tropical large-scale divergence in the upper troposphere on the meridional shift of Asian jet. The anomalous absolute vorticity advection by a horizontal divergent wind is expressed as follows:

where ξ is the absolute vorticity, v_x is the horizontal divergent wind vector, and ζ is the relative vorticity. The overbars and primes are defined as in Eq. (1). The first and second terms of RHS in Eq. (4) indicate the contributions of anomalous and climatological divergent wind to the absolute vorticity tendency, respectively.

To calculate the RWB frequency near Japan, we used a daily dynamical blocking index (Pelly and Hoskins 2003) based on the meridional distribution of potential temperature θ on the dynamical tropopause defined by 2 potential vorticity units (PVU; 1 PVU = 10⁻⁶ K kg⁻¹ m² s⁻¹), which is conserved without diabatic processes. The blocking index B (K) is expressed as follows:

where λ is the longitude, ∆λ is the width of zonal average, ϕ is the latitude, ∆ϕ is the typical meridional scale of the wave breaking, and ϕ₀ is the central latitude. We defined ∆λ as 5° and ∆ϕ as 30°, respectively, according to Pelly and Hoskins (2003). A daily RWB is detected when B has a positive value: there is high θ to the north and low θ to the south. To illustrate an example of RWB event, θ on 2 PVU and 200-hPa absolute vorticity on 22 August 2016 are shown in Figs. 1a and 1b, respectively. An anticyclonic meridional overturning is clearly seen east of Japan in both of θ and the absolute vorticity fields, showing that the positive B corresponds to negative latitudinal gradient of the absolute vorticity in the upper troposphere. Although the blocking index can identify not only the RWB but also “Ω”-shaped blocking (Pelly and Hoskins 2003), it is expected to be a suitable approximation of the RWB frequency. The advantages to using the blocking index B are that the RWB events are identified as Lagrangian behavior (Pelly and Hoskins 2003) and that anomalous circulations theoretically attain their largest amplitude on the tropopause level (Bowley et al. 2019). We defined the monthly RWB frequency as the ratio of the number of days on which the RWB is detected in the month.

(a) Potential temperature on 2 PVU (K) and (b) 200-hPa absolute vorticity (10⁻⁵ s⁻¹) on 22 Aug 2016.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

View Full Size

(a) Potential temperature on 2 PVU (K) and (b) 200-hPa absolute vorticity (10⁻⁵ s⁻¹) on 22 Aug 2016.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

(a) Potential temperature on 2 PVU (K) and (b) 200-hPa absolute vorticity (10⁻⁵ s⁻¹) on 22 Aug 2016.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

El Niño and La Niña years to perform composite analysis are extracted according to a criteria of ENSO occurrence defined by the JMA²: the 5-month running mean SST deviation for Niño-3, which is defined as SST averaged over the area 5°S–5°N, 150°–90°W, satisfies above +0.5°C and below −0.5°C for at least six consecutive months, respectively. Here the SST deviation is defined as the deviation from the latest sliding 30-yr averages. Table 1 shows the El Niño and La Niña years for July and August. The 14–17 samples of the El Niño and La Niña years are extracted to construct the composite during the 61-yr period from 1958 to 2018 (Table 1).

Table 1.

El Niño and La Niña years extracted to composite in July and August during the 61-yr period from 1958 to 2018.

View Table

3. Climatology and ENSO composite of RWB frequency

Figures 2a and 2b show monthly climatological RWB frequency and its standard deviation averaged over 25°–45°N in July and August, respectively. There is a wide area of clear maximum of the RWB frequency over a region from Japan to the date line in midlatitude, indicating a frequent RWB near the Asian jet exit region during the boreal summer. The RWB frequency near Japan in August particularly shows a marked enhancement. The latitudinal band of 25°–45°N corresponds to the anticyclonically sheared basic state south of the jet core in the boreal summer. According to the algorithm defined by Bowley et al. (2019), the climatological RWB frequency shown in Figs. 2a and 2b can be divided into anticyclonic and cyclonic types as shown in Figs. 2c and 2d, respectively. A large difference in the RWB frequency between anticyclonic and cyclonic types is clearly seen from Japan to the date line, and more than 70% of the detected RWB near Japan is categorized as the anticyclonic type, consistent with the result of composite analysis in TM20. A large standard deviation is also seen over the region, indicating a large interannual variability of the RWB frequency. Although a wide area of another maximum for the RWB frequency is also seen over a region from eastern North America to the western North Atlantic, the frequency over the region in August is about half of that from Japan to the date line (Fig. 2b). A localized maximum of the RWB frequency is also seen from 50° to 60°E, consistent with the result of Samanta et al. (2016) that showed impacts of an occurrence of the RWB over the region on Indian summer monsoon activity through cold and dry air intrusion toward India.

Climatological RWB frequency (%) averaged over 25°–45°N in (a),(c) July and (b),(d) August, with total (black lines on the upper panels), anticyclonic (red lines on the lower panels), and cyclonic (blue lines on the lower panels) RWB. Gray, light red, and light blue shadings indicate the standard deviations of each RWB frequency during the climatological period. Green shading denotes longitudinal area near Japan from 130° to 160°E.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Climatological RWB frequency (%) averaged over 25°–45°N in (a),(c) July and (b),(d) August, with total (black lines on the upper panels), anticyclonic (red lines on the lower panels), and cyclonic (blue lines on the lower panels) RWB. Gray, light red, and light blue shadings indicate the standard deviations of each RWB frequency during the climatological period. Green shading denotes longitudinal area near Japan from 130° to 160°E.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

Climatological RWB frequency (%) averaged over 25°–45°N in (a),(c) July and (b),(d) August, with total (black lines on the upper panels), anticyclonic (red lines on the lower panels), and cyclonic (blue lines on the lower panels) RWB. Gray, light red, and light blue shadings indicate the standard deviations of each RWB frequency during the climatological period. Green shading denotes longitudinal area near Japan from 130° to 160°E.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

Figures 3a and 3b show composited monthly RWB frequency averaged over 25°–45°N in the El Niño (red line) and La Niña (blue line) years and their standard deviations. The monthly RWB frequency near Japan significantly increases in the La Niña years, particularly in August, indicating that the La Niña conditions are favorable for the RWB occurrence over the region. This is consistent with the result of Bowley et al. (2019), who indicated the statistically significant relationship between the frequent anticyclonic RWB and negative phase of ENSO (i.e., La Niña). We hereafter focus on circulation characteristics related to the RWB frequency in August; those in July show similar characteristics.

As in Fig. 2, but for the climatology (black line), composite of El Niño years (red dashed line), and La Niña years (blue dashed line). Solid color lines indicate that the differences in frequency between the composite and the climatology are significant with a confidence level of 95%. Light red and blue shadings indicate the standard deviations during the El Niño and La Niña years, respectively.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

View Full Size

As in Fig. 2, but for the climatology (black line), composite of El Niño years (red dashed line), and La Niña years (blue dashed line). Solid color lines indicate that the differences in frequency between the composite and the climatology are significant with a confidence level of 95%. Light red and blue shadings indicate the standard deviations during the El Niño and La Niña years, respectively.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

As in Fig. 2, but for the climatology (black line), composite of El Niño years (red dashed line), and La Niña years (blue dashed line). Solid color lines indicate that the differences in frequency between the composite and the climatology are significant with a confidence level of 95%. Light red and blue shadings indicate the standard deviations during the El Niño and La Niña years, respectively.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

4. Regression analysis for the RWB frequency

This section shows large-scale oceanographic and atmospheric circulation anomalies regressed onto areal averages of the RWB frequency within 25°–45°N, 130°–160°E in August. Figure 4a shows interannual time series of the areal-averaged RWB frequency over the region during the 61-yr period from 1958 to 2018. The frequency shows large interannual variability as shown in Fig. 2b. It also shows decadal-like variations as seen in the 11-yr running mean time series (blue line in Fig. 4a), suggesting the relationship to the PDO, corresponding with the result of Strong and Magnusdottir (2009). Although a small long-term decreasing trend of the RWB frequency with −0.05% yr⁻¹ is seen (red dashed line in Fig. 4a), the linear trend is not significant. The trend is not significant even if the PDO-related variability was removed using the monthly PDO index, which is available in JMA’s website,³ from the interannual variability of the RWB frequency, although the confidence level slightly increased (not shown), suggesting that a primary cause of the insignificance is the large interannual variability. The RWB frequency in August 2010 exceeded 40%, which is more than twice the climatology of 18.6%. This is consistent with an unprecedented heat wave observed over Japan in summer 2010 (e.g., Kosaka et al. 2012) and with the monthly mean temperature anomalies highest on record for August since 1946 (JMA 2011). Figure 4b shows the histogram of the RWB frequency in August. The monthly RWB frequency during the 61-yr period is nearly normally distributed, indicating the validity of the linear regression analysis for the frequency.

(a) Interannual time series of the areal averaged RWB frequency over 25°–45°N, 130°–160°E (bars; %) and its 11-yr running mean (blue line), and (b) histogram of the RWB frequency with a class width of 5% in August during the period from 1958 to 2018. Red and green dashed lines in (a) indicate linear regression lines of the frequency during the period from 1958 to 2018 and the period from 1980 to 2010, respectively.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

View Full Size

(a) Interannual time series of the areal averaged RWB frequency over 25°–45°N, 130°–160°E (bars; %) and its 11-yr running mean (blue line), and (b) histogram of the RWB frequency with a class width of 5% in August during the period from 1958 to 2018. Red and green dashed lines in (a) indicate linear regression lines of the frequency during the period from 1958 to 2018 and the period from 1980 to 2010, respectively.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

(a) Interannual time series of the areal averaged RWB frequency over 25°–45°N, 130°–160°E (bars; %) and its 11-yr running mean (blue line), and (b) histogram of the RWB frequency with a class width of 5% in August during the period from 1958 to 2018. Red and green dashed lines in (a) indicate linear regression lines of the frequency during the period from 1958 to 2018 and the period from 1980 to 2010, respectively.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

Figure 5a shows anomalous SST regressed onto the RWB frequency, and its statistical significance. Significant anomalous cold SST is dominant over the equatorial central to eastern Pacific, indicating that the La Niña conditions are favorable for the RWB near Japan as described in section 3. This result is consistent with the result of TM20, who performed a lag composite analysis of 44 RWB cases and indicated that the significant anomalous cold SST dominates over the equatorial central to eastern Pacific (see Fig. 6 in TM20). Anomalous warm SST, in contrast, is clearly seen near Japan associated with the increased RWB frequency. The regressed latent and sensible heat fluxes shown in Fig. 5b exhibit negative anomalies near Japan, indicating anomalous heat exchange from atmosphere to sea surface. This result suggests that anomalous hot conditions over the region in the lower troposphere associated with the frequent RWB shown in Fig. 5c is a factor contributing to the anomalous warm SST near Japan. The regressed anomalous surface wind speed (not shown) is not significant to the sea east of Japan, also indicating an essential contribution of the anomalous hot condition to the anomalous negative heat flux, compared to that of the weak wind speed. The anomalous warm SST east of Japan is similar to the negative phase of PDO (Fig. 5a), also indicating the relationship between the RWB frequency and the decadal-like SST variations over the North Pacific. Further investigation is needed to assess the detailed factor of the anomalous warm SST in terms not only of the contribution of atmosphere but also of the oceanographic factors including the PDO.

(a) SST, (b) latent and sensible upward heat flux, and (c) 850-hPa temperature regressed onto the areal averaged RWB frequency over 25°–45°N, 130°–160°E in August. Contour intervals are 0.1°C in (a), 3 W m⁻² in (b), and 0.1°C in (c). Solid and dashed contours denote positive and negative regressed anomalies, respectively. Negative (positive) anomalies in (b) indicate anomalous heat flux from the atmosphere (ocean) to the ocean (atmosphere). Dots indicate the significant anomalies with a confidence level of 95% based on Student’s t test.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

View Full Size

(a) SST, (b) latent and sensible upward heat flux, and (c) 850-hPa temperature regressed onto the areal averaged RWB frequency over 25°–45°N, 130°–160°E in August. Contour intervals are 0.1°C in (a), 3 W m⁻² in (b), and 0.1°C in (c). Solid and dashed contours denote positive and negative regressed anomalies, respectively. Negative (positive) anomalies in (b) indicate anomalous heat flux from the atmosphere (ocean) to the ocean (atmosphere). Dots indicate the significant anomalies with a confidence level of 95% based on Student’s t test.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

(a) SST, (b) latent and sensible upward heat flux, and (c) 850-hPa temperature regressed onto the areal averaged RWB frequency over 25°–45°N, 130°–160°E in August. Contour intervals are 0.1°C in (a), 3 W m⁻² in (b), and 0.1°C in (c). Solid and dashed contours denote positive and negative regressed anomalies, respectively. Negative (positive) anomalies in (b) indicate anomalous heat flux from the atmosphere (ocean) to the ocean (atmosphere). Dots indicate the significant anomalies with a confidence level of 95% based on Student’s t test.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

Figure 6a shows 200-hPa anomalous velocity potential regressed onto the RWB frequency. The anomalous large-scale convergence is centered over the central to eastern Pacific in the upper troposphere, corresponding to suppressed convection associated with the anomalous cold SST over the region. The anomalous convergence extends to the midlatitude in the North Pacific, suggesting its association with warm SST anomalies east of Japan (Fig. 5a). The upper-tropospheric anomalous divergence, in contrast, is seen around the Indian Ocean. These anomalous divergence patterns with zonal wavenumber 1 indicate the enhanced Walker circulation from the eastern Indian Ocean to the Pacific, compared to the climatology (not shown). Anomalous northward divergent winds are dominant over southern Eurasia (vectors in Fig. 6c) associated with the upper-tropospheric anomalous divergence around the Indian Ocean (Fig. 6a), indicating enhanced Asian summer monsoon. Figure 6b shows the climatological (gray shaded) and regressed anomalous (color contour) zonal wind at 200 hPa. Anomalous westerly wind is seen along and north of the climatological jet core from central Asia to East Asia, indicating enhanced and northward shifted tendencies of the Asian jet. The enhanced Asian jet can be explained by the anomalous northward divergent wind through Coriolis acceleration. To assess the influence of the anomalous northward divergent wind on the northward shifted Asian jet, the advection term of the 200-hPa Rossby wave source calculated from the regressed and climatological divergent wind and vorticity described in Eq. (4) is shown in Fig. 6c. A negative vorticity tendency is clearly seen over the latitudinal band over 40°–50°N, where the anomalous northward divergent winds cross the climatological Asian jet (green contour), contributing to the northward shifted Asian jet. These results indicate that the upper-tropospheric anomalous divergence around the Indian Ocean contributes to the enhanced and northward shifted Asian jet.

As in Fig. 5, but for 200-hPa (a) velocity potential, (b) zonal wind, and (c) absolute-vorticity advection term of Rossby wave source (shading; 10⁻¹¹ s⁻²). The absolute vorticity advection is calculated from the regressed anomalies and climatology. Shading in (b) indicates 200-hPa climatological zonal wind (m s⁻¹). Vectors and green contour in (c) indicate 200-hPa regressed anomalous divergent wind (m s⁻¹) and climatological zonal wind (m s⁻¹). The contour intervals are 1 × 10⁵ m² s⁻¹ in (a), 0.5 m s⁻¹ in (b), and 5 m s⁻¹ in (c) for values larger than 20 m s⁻¹.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

View Full Size

As in Fig. 5, but for 200-hPa (a) velocity potential, (b) zonal wind, and (c) absolute-vorticity advection term of Rossby wave source (shading; 10⁻¹¹ s⁻²). The absolute vorticity advection is calculated from the regressed anomalies and climatology. Shading in (b) indicates 200-hPa climatological zonal wind (m s⁻¹). Vectors and green contour in (c) indicate 200-hPa regressed anomalous divergent wind (m s⁻¹) and climatological zonal wind (m s⁻¹). The contour intervals are 1 × 10⁵ m² s⁻¹ in (a), 0.5 m s⁻¹ in (b), and 5 m s⁻¹ in (c) for values larger than 20 m s⁻¹.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

As in Fig. 5, but for 200-hPa (a) velocity potential, (b) zonal wind, and (c) absolute-vorticity advection term of Rossby wave source (shading; 10⁻¹¹ s⁻²). The absolute vorticity advection is calculated from the regressed anomalies and climatology. Shading in (b) indicates 200-hPa climatological zonal wind (m s⁻¹). Vectors and green contour in (c) indicate 200-hPa regressed anomalous divergent wind (m s⁻¹) and climatological zonal wind (m s⁻¹). The contour intervals are 1 × 10⁵ m² s⁻¹ in (a), 0.5 m s⁻¹ in (b), and 5 m s⁻¹ in (c) for values larger than 20 m s⁻¹.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

Figures 7a and 7b show the regressed 200-hPa anomalous meridional wind and its meridional averages, respectively, between 40° and 45°N. A wave train is clearly seen along the Asian jet with a phase of anomalous anticyclonic circulation over Japan, contributing to the frequent RWB over the region. The regressed meridional wind anomalies averaged over 40°–45°N (blue line in Fig. 7b) contribute to slight westward shift of the climatological wave phase along the Asian jet (black dashed line) and the wave amplification near the jet exit region associated with the RWB. Wave activity fluxes shown as vectors in Fig. 7a indicate the Rossby wave propagation along the Asian jet, and its contribution to the RWB. Furthermore, the equatorward propagating Rossby wave is seen from Japan to its east, corresponding to the frequent anticyclonic RWB as shown in Abatzoglou and Magnusdottir (2006). The wave train in the upper troposphere is consistent with the formation of the Silk Road pattern (Enomoto et al. 2003; Kosaka et al. 2009). The phase of the wave train might be regulated by the meridional displacement of the Asian jet through internal atmospheric variability (e.g., Simmons et al. 1983; Sato and Takahashi 2006; Hong and Lu 2016) and the large-scale orography such as the Tibetan Plateau (e.g., Ringler and Cook 1995, 1999). The regressed wave pattern corresponds to that shown in Fig. 2b of Kosaka et al. (2009), which shows 200-hPa anomalous vorticity regressed onto the first mode of the upper-tropospheric meridional wind over Eurasia, suggesting its close relationship to the Silk Road pattern.

(a) 200-hPa meridional wind regressed onto the areal averaged RWB frequency over 25°–45°N, 130°–160°E, and (b) its meridional averages between 40°–45°N in August. (c),(d) Latitudinal profiles for 200-hPa K_s and eddy kinetic energy (m² s⁻²) averaged over 60°–120°E, respectively. The contour interval in (a) is 0.2 m s⁻¹. Shading in (a) is 200-hPa climatological zonal wind. The vectors in (a) are wave activity flux (Takaya and Nakamura 2001; m² s⁻²) calculated from the 200-hPa regressed anomalies and climatology. In (b)–(d), black dashed and blue solid lines denote the climatology and that added with the regressed anomalies explained by one standard deviation of the RWB frequency, respectively. Red solid lines in (c) and (d) denote the regressed anomalies shown in the second (red colored) axis. In (c), the lines are not shown over easterly winds.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

View Full Size

(a) 200-hPa meridional wind regressed onto the areal averaged RWB frequency over 25°–45°N, 130°–160°E, and (b) its meridional averages between 40°–45°N in August. (c),(d) Latitudinal profiles for 200-hPa K_s and eddy kinetic energy (m² s⁻²) averaged over 60°–120°E, respectively. The contour interval in (a) is 0.2 m s⁻¹. Shading in (a) is 200-hPa climatological zonal wind. The vectors in (a) are wave activity flux (Takaya and Nakamura 2001; m² s⁻²) calculated from the 200-hPa regressed anomalies and climatology. In (b)–(d), black dashed and blue solid lines denote the climatology and that added with the regressed anomalies explained by one standard deviation of the RWB frequency, respectively. Red solid lines in (c) and (d) denote the regressed anomalies shown in the second (red colored) axis. In (c), the lines are not shown over easterly winds.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

(a) 200-hPa meridional wind regressed onto the areal averaged RWB frequency over 25°–45°N, 130°–160°E, and (b) its meridional averages between 40°–45°N in August. (c),(d) Latitudinal profiles for 200-hPa K_s and eddy kinetic energy (m² s⁻²) averaged over 60°–120°E, respectively. The contour interval in (a) is 0.2 m s⁻¹. Shading in (a) is 200-hPa climatological zonal wind. The vectors in (a) are wave activity flux (Takaya and Nakamura 2001; m² s⁻²) calculated from the 200-hPa regressed anomalies and climatology. In (b)–(d), black dashed and blue solid lines denote the climatology and that added with the regressed anomalies explained by one standard deviation of the RWB frequency, respectively. Red solid lines in (c) and (d) denote the regressed anomalies shown in the second (red colored) axis. In (c), the lines are not shown over easterly winds.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

Figure 7c shows a latitudinal profile of 200-hPa K_s averaged over 60°–120°E. Although the zonal wind anomalies (Fig. 6b) are quite smaller than the climatological wind along the Asian jet (Fig. 7a), the northward shifted Rossby waveguide at 200-hPa is seen associated with the increased RWB frequency (blue line in Fig. 7c), compared to the climatology (black dashed line in Fig. 7c). To highlight a change of the Rossby waveguide from the climatology, the anomalous K_s at 200 hPa is denoted by the red line in Fig. 7c. The anomalous K_s shows the increased wavenumber from the center to north of the Asian jet core, indicating that Rossby waves with more extensive wavenumber range than the climatology can propagate along the Asian jet. Figure 7d shows a latitudinal profile of eddy (i.e., zonal wavenumbers k ≧ 3) kinetic energy averaged over the same longitudinal range, according to the procedure of Enomoto (2004). The large eddy kinetic energy corresponds to the increased Rossby wave activity. The increased eddy kinetic energy is seen along the waveguide with the large K_s (blue line in Fig. 7d), compared to the climatology (black dashed line in Fig. 7d). The difference between the eddy kinetic energy and its climatology (red line in Fig. 7d) clearly shows the increased wave activity along the latitudinal band of 40°N, indicating the favorable condition for propagation of the Rossby wave with the extensive wavenumber range along the Asian jet as described in Fig. 7c. The active wave propagation along the Asian jet is consistent with the frequent RWB east of Japan.

Figures 8a and 8b show anomalous 200-hPa geopotential height and potential temperature at 2 PVU regressed onto the RWB frequency. The 200-hPa height in Fig. 8a shows not only the wave train shown in Fig. 7a but also zonally elongated positive anomalies, indicating the northward extension of the Tibetan high and the associated northward-shifted Asian jet over midlatitude Eurasia. Positive and negative height anomalies are clearly seen near and south of Japan, respectively, consistent with the increased RWB near Japan in the La Niña years. The negative height anomalies correspond to an enhanced and westward extended mid-Pacific trough compared to the climatology (gray shading). The regressed anomalous potential temperature at 2 PVU in Fig. 8b also shows positive and negative anomalies from Japan to its south. This dipole anomalies of potential temperature are associated with an enhancement of the climatological “inverse-S”-shaped overturning of potential temperature (i.e., the anticyclonic RWB), corresponding to the frequent RWB near Japan. The enhanced anticyclonic RWB exhibits westward intrusion of low potential temperature [i.e., high potential vorticity (PV) air mass] toward the subtropical western North Pacific. To assess an influence of the enhanced and northward shifted Asian jet (Figs. 6b and 8a) on the frequent RWB, the regressed 200-hPa anomalous planetary-scale (i.e., zonal wavenumbers k < 3) stretching deformation described in Eq. (3) is shown in Fig. 8c. An anomalous negative stretching deformation is seen from north of mainland Japan to its east (Fig. 8c), where is located over and downstream of the positive 200-hPa height anomalies near Japan (Fig. 8a). This result indicates that enhanced diffluence and deceleration of basic flow near the Asian jet exit region, which are associated with the northward extension of the Tibetan high and the enhanced and northward shifted Asian jet due to the La Niña conditions, contribute to the frequent RWB over the region. The relationship between the northward shifted Asian jet and the frequent anticyclonic RWB is also consistent with the result of Rivière (2009), who showed the influence of meridional shift of jet stream on the RWB type using quasigeostrophic and primitive equation models.

As in Fig. 5, but for (a) 200-hPa geopotential height, (b) potential temperature at 2 PVU, and (c) 200-hPa planetary-scale stretching deformation $d_p^{\prime }$. Contour intervals are 5 m in (a), 0.5 K in (b), and 1 × 10⁻⁷ s⁻¹ in (c). Shadings in (a) and (b) are the climatology and in (c) the climatological 200-hPa geopotential height with zonal wavenumbers k < 3.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

View Full Size

As in Fig. 5, but for (a) 200-hPa geopotential height, (b) potential temperature at 2 PVU, and (c) 200-hPa planetary-scale stretching deformation $d_p^{\prime }$. Contour intervals are 5 m in (a), 0.5 K in (b), and 1 × 10⁻⁷ s⁻¹ in (c). Shadings in (a) and (b) are the climatology and in (c) the climatological 200-hPa geopotential height with zonal wavenumbers k < 3.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

As in Fig. 5, but for (a) 200-hPa geopotential height, (b) potential temperature at 2 PVU, and (c) 200-hPa planetary-scale stretching deformation $d_p^{\prime }$. Contour intervals are 5 m in (a), 0.5 K in (b), and 1 × 10⁻⁷ s⁻¹ in (c). Shadings in (a) and (b) are the climatology and in (c) the climatological 200-hPa geopotential height with zonal wavenumbers k < 3.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

Recently, TM20 revealed a process that the Rossby wave propagation along the Asian jet can excite the Pacific–Japan (PJ) pattern (Nitta 1987), through the RWB east of Japan and the associated high-PV intrusion toward the subtropical western North Pacific and the consequent enhanced cumulus convection over the region. The PJ pattern is a lower-tropospheric summer teleconnection pattern characterized by a meridional seesaw between enhanced (suppressed) convection over the Philippines and anticyclonic (cyclonic) circulation anomalies over Japan, which is well known to cause anomalous summer climate near Japan. The results of composite analysis in TM20 and the statistical relationship shown in this study suggest that the frequent RWB associated with the La Niña conditions may contribute to the monthly mean anomalous circulation south of Japan due to the similar process. To show the interannual variability of the convection over the subtropical western North Pacific associated with the RWB frequency, anomalous 500-hPa vertical p velocity (ω) regressed onto the RWB frequency is shown in Fig. 9a. Anomalous ascent is seen around the northern Philippines, west of the intruded anomalous low potential temperature over the subtropical western North Pacific (Fig. 8b) associated with the westward extended mid-Pacific trough. This is consistent with the relationship between the high-PV intrusion and the enhanced convection shown by TM20. The anomalous ascent around the northern Philippines associated with the La Niña–related frequent RWB is also consistent with enhanced tropical cyclone (TC) activity around the region in the La Niña conditions, which is accompanied by zonal shift of the TC genesis position, as shown by previous studies (e.g., Chia and Ropelewski 2002; Camargo and Sobel 2005). Furthermore, anomalous 850-hPa relative vorticity regressed onto the RWB frequency shown in Fig. 9b exhibits the formation of the PJ pattern, with anomalous positive vorticity north of the Philippines and anomalous negative vorticity over Japan. Figure 10 shows a scatter diagram between the RWB frequency and PJ index in August. Here the PJ index is defined as difference in the monthly mean 850-hPa relative vorticity between 17.5°–27.5°N, 110°–130°E (green solid box in Fig. 9b) and 30°–40°N, 125°–155°E (green dashed box in Fig. 9b). The RWB frequency and the PJ index show a positive correlation (+0.50), indicating that the frequent RWB is closely related to the PJ pattern. The anomalous negative vorticity over Japan indicates an extension of the North Pacific subtropical high toward mainland Japan, corresponding to anomalous hot summer climate over the region as shown later. These regressed anomalies indicate that the frequent RWB can cause the frequent formation of the PJ pattern, through the mechanism shown in TM20. The position of the vorticity anomalies associated with the PJ pattern is different from the composite shown in TM20 (Fig. 3 of their paper), primarily because of the difference in the target regions of the RWB to perform the statistical analyses. An anomalous cyclonic circulation is also seen over the Okhotsk Sea, indicating the existence of an anomalous meridional tripolar pattern including the PJ pattern from the Philippines to the Okhotsk Sea shown by Hirota and Takahashi (2012).

As in Fig. 5, but for (a) 500-hPa vertical p velocity (ω) and (b) 850-hPa relative vorticity. Contour intervals are 0.4 × 10⁻² Pa s⁻¹ in (a) and 0.5 × 10⁻⁶ s⁻¹ in (b). The negative (green dashed line) and positive (orange line) anomalous ω denote the anomalous ascent and descent, respectively. Green solid and dashed boxes indicate the defined region of the PJ index. See text for the definition of the PJ index.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

View Full Size

As in Fig. 5, but for (a) 500-hPa vertical p velocity (ω) and (b) 850-hPa relative vorticity. Contour intervals are 0.4 × 10⁻² Pa s⁻¹ in (a) and 0.5 × 10⁻⁶ s⁻¹ in (b). The negative (green dashed line) and positive (orange line) anomalous ω denote the anomalous ascent and descent, respectively. Green solid and dashed boxes indicate the defined region of the PJ index. See text for the definition of the PJ index.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

As in Fig. 5, but for (a) 500-hPa vertical p velocity (ω) and (b) 850-hPa relative vorticity. Contour intervals are 0.4 × 10⁻² Pa s⁻¹ in (a) and 0.5 × 10⁻⁶ s⁻¹ in (b). The negative (green dashed line) and positive (orange line) anomalous ω denote the anomalous ascent and descent, respectively. Green solid and dashed boxes indicate the defined region of the PJ index. See text for the definition of the PJ index.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

Scatter diagram of the areal averaged RWB frequency over 25°–45°N, 130°–160°E (horizontal axis; %) and monthly mean PJ index (vertical axis; 10⁻⁶ s⁻¹) in August. The correlation coefficient is shown at the bottom right of the panel.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Scatter diagram of the areal averaged RWB frequency over 25°–45°N, 130°–160°E (horizontal axis; %) and monthly mean PJ index (vertical axis; 10⁻⁶ s⁻¹) in August. The correlation coefficient is shown at the bottom right of the panel.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

Scatter diagram of the areal averaged RWB frequency over 25°–45°N, 130°–160°E (horizontal axis; %) and monthly mean PJ index (vertical axis; 10⁻⁶ s⁻¹) in August. The correlation coefficient is shown at the bottom right of the panel.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

Furthermore, the relationship between the frequent RWB and the anomalous hot summer climate in Japan should also be described, according to the aforementioned result. Figure 11 shows scatter diagrams between the RWB frequency and regional averaged temperature anomalies in northern, eastern, and western Japan and Okinawa/Amami in August. The RWB frequency and the monthly-mean temperature anomalies in northern, eastern, and western Japan show positive correlations. A correlation coefficient between the RWB frequency and temperature anomalies averaged over the three regions from northern to western Japan also shows a positive correlation coefficient (R = +0.55). The results indicate that the frequent RWB associated with the La Niña conditions is closely related to hot summer conditions over Japan, through the upper-tropospheric anomalous warm anticyclone and the enhanced PJ pattern. These positive correlations imply that the decadal-like variability of the RWB frequency as described in Fig. 4a can contribute to that of the boreal summer climate in Japan, corresponding to the result of Urabe and Maeda (2014), who indicated a close relationship between the decadal variability and summer–autumn climate in Japan. The RWB frequency and the temperature anomalies in Okinawa/Amami, in contrast, show weak negative correlation in association with the anomalous cyclonic circulation north of the Philippines.

Scatter diagram of the areal averaged RWB frequency over 25°–45°N, 130°–160°E (horizontal axis; %) and monthly mean regional averaged temperature anomalies (vertical axis; °C) in (a) northern, (b) eastern, and (c) western Japan, and (d) Okinawa/Amami in August. The correlation coefficients are shown at the bottom right of the panels.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Scatter diagram of the areal averaged RWB frequency over 25°–45°N, 130°–160°E (horizontal axis; %) and monthly mean regional averaged temperature anomalies (vertical axis; °C) in (a) northern, (b) eastern, and (c) western Japan, and (d) Okinawa/Amami in August. The correlation coefficients are shown at the bottom right of the panels.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

Scatter diagram of the areal averaged RWB frequency over 25°–45°N, 130°–160°E (horizontal axis; %) and monthly mean regional averaged temperature anomalies (vertical axis; °C) in (a) northern, (b) eastern, and (c) western Japan, and (d) Okinawa/Amami in August. The correlation coefficients are shown at the bottom right of the panels.

Citation: Journal of Climate 33, 15; 10.1175/JCLI-D-19-0958.1

• Download Figure
• Download figure as PowerPoint slide

5. Conclusions and discussion

This study investigated the statistical relationship between the RWB frequency east of Japan and the large-scale oceanographic and atmospheric circulation during the boreal summer, focusing on the impacts of the La Niña conditions on the frequent RWB through large-scale anomalous divergence and the resultant modulation of the Asian jet.

The RWB frequency in the Northern Hemisphere midlatitude in July and August showed its maximum climatologically and large interannual variation near the Asian jet exit region near Japan. The large difference in the RWB frequency between anticyclonic and cyclonic types was clearly seen from Japan to the date line, and more than 70% of the RWB near Japan was categorized as the anticyclonic type. The total RWB frequency significantly increased during the La Niña years, compared to the climatology.

The SST regressed onto the RWB frequency in August shows significant cold anomalies over the equatorial central to eastern Pacific, corresponding to the favorable conditions for the RWB during the La Niña events. The regressed SST also showed the negative phase of PDO-like pattern with warm SST anomalies east of Japan, suggesting its relationship to the RWB frequency. The large-scale anomalous convergence and divergence in the upper troposphere was seen over the central to eastern Pacific and around the Indian Ocean, respectively. It was consistent with the enhanced Walker circulation from the eastern Indian Ocean to the Pacific and Asian summer monsoon, which is one of the major characteristics seen during the La Niña events. The Asian jet is enhanced and shifted northward compared to the climatology, associated with the upper-tropospheric divergence around the Indian Ocean and the enhanced Asian summer monsoon, contributing to the frequent anticyclonic RWB east of Japan through the enhanced diffluence and deceleration of basic flow near the jet exit region. The northward shifted Rossby waveguide with extensive wavenumber K_s range and the associated favorable condition of Rossby wave propagation is also seen over Eurasia along the Asian jet, contributing to the increased anticyclonic RWB frequency. These statistical results and the related physical interpretation indicate that the La Niña conditions can contribute to the frequent RWB near Japan, through the enhanced Asian monsoon and the resultant modulation of the Asian jet. Furthermore, the expanded Tibetan high, which can be regarded as a huge mass of low-PV air, may provide us another interpretation for the influence of the enhanced Asian monsoon on the frequent RWB. In the Tibetan high, the low-PV air is supplied due to the anomalous diabatic heating associated with the enhanced monsoon (Hoskins 2015), and is cut off near the Asian jet exit region (i.e., the RWB occurrence). This physical interpretation also supports the impact of ENSO-related enhanced Asian monsoon on the frequent RWB shown in this study.

The enhanced anticyclonic overturning of potential temperature gradient at 2 PVU in association with the frequent RWB indicated an amplified ridge near Japan and the westward extended mid-Pacific trough. The anomalous ascent was seen near the northern Philippines, west of the extended mid-Pacific trough. This relationship could be explained by the process shown by TM20, which indicated that the high-PV intrusion toward the subtropical western North Pacific associated with the RWB contributed to the enhanced convection over the region through dynamically induced ascent. Although TM20 indicated a primary role of the high-PV intrusion to the dynamically induced anomalous ascent compared with that of warm SST conditions over the region from their partial correlation analysis, thermodynamically induced ascent due to the anomalous cold air inflow related to the high PV may also contribute to the enhanced convection. The regressed anomalous relative vorticity in the lower troposphere showed the formation of the PJ pattern, which was also consistent with the results of TM20, although the position of the PJ pattern is different from the composite shown in TM20 primarily because of the difference in the target regions of the statistical analyses. Although the lower-tropospheric anomalous anticyclonic circulation over Japan shown in Fig. 9b is presumed to be partly caused by the upper-tropospheric amplified ridge over the region shown in Fig. 8a, the combined effects of the upper-tropospheric amplified anticyclone resulted from the Rossby wave propagation over Eurasia and the PJ pattern can contribute to an unprecedented hot summer conditions over Japan. The combined impacts on the anomalous summer climate are suggested by Ogasawara and Kawamura (2007) and Thompson et al. (2019), who showed the relationship between extreme hot summer climate over East Asia and the combined factors of the Silk Road pattern and the extended North Pacific subtropical high associated with the PJ pattern. The anomalous ascent seen near the Korean Peninsula (Fig. 9a), where is the western fringe of the extended North Pacific subtropical high (Fig. 9b), also can partly contribute to the upper-tropospheric positive height anomalies (Fig. 8a; i.e., the frequent RWB) through a diabatic generation of low-PV air due to the enhanced convection. The frequent RWB near Japan associated with La Niña conditions is closely related to hot summer climate over Japan, with positive correlations between the RWB frequency and monthly-mean regional averaged temperature anomalies over northern to western Japan in August. Further investigation for the relationship with summer climate in larger region such as East Asia will provide us the more detailed impact of the frequent RWB.

The long-term decreasing trend and decadal-like variability of the RWB frequency shown in Fig. 4a is also one of the interesting characteristics to be further examined. Kubota et al. (2016) showed a synchronization between the PDO phase and the correlation of ENSO and the PJ pattern, using a 117-yr historical index of the PJ pattern. Their results motivate us to examine decadal changes of the relationship between the RWB frequency and the PJ pattern with longer analysis period. Although the long-term trend of the RWB frequency is statistically insignificant primarily because of the large interannual variation, it might be affected by modulated atmospheric circulation associated with the climate change. Sugi et al. (2002) showed that an increasing tendency of static stability associated with the climate change contributes to weakened tropical atmospheric circulation and the resultant suppressed convection over the Asian monsoon region, indicating an El Niño–like atmospheric response. Some previous studies suggested that the Asian jet core would shift equatorward in future climate during summer (e.g., Hirahara et al. 2012; Horinouchi et al. 2019), corresponding to a weakened extension of the Tibetan high associated with a suppressed summer Asian monsoon circulation (Sooraj et al. 2015). It is expected that the weakened Tibetan high and the related southward shifted Asian jet will act to decrease the RWB frequency near Japan, associated with zonally symmetric flow prevailing from Eurasia to the North Pacific and the consequent weakened diffluence and deceleration of the basic zonal flow near the Asian jet exit region. This possible process corresponds to the El Niño–like atmospheric response, and is an opposite sense of the tendency described in section 4. Although Bowley et al. (2019) showed a long-term increasing trend of the RWB frequency over the western part of the North Pacific in midlatitude, they assessed the long-term trend during a shorter period from 1980 to 2010, compared to the 61-yr period in this study. The RWB frequency assessed during the period from 1980 to 2010 actually shows an increasing trend (green dashed line in Fig. 4a) with a confidence level of 90%, partly due to the decadal-like variability of the RWB frequency (blue line in Fig. 4a). Strong and Magnusdottir (2009) showed that variability of the RWB frequency caused a phase modulation of PDO through the influence of atmospheric variability on SST condition over the North Pacific. The decadal-like variability of the RWB frequency shown in Fig. 4a also suggests a relationship to the PDO. To assess the significance of these long-term variabilities, historical dataset for atmospheric circulation with longer period will be required. For example, a diagnosis of high-resolution experiments with large-ensemble historical and nonwarming simulations, called “Database for Policy Decision-Making for Future Climate Change” (d4PDF; Mizuta et al. 2017), will be useful to elucidate the long-term variabilities of the RWB frequency.