1. Introduction
In the boreal summer, Rossby wave propagation along the Asian jet (Lu et al. 2002), such as the Silk Road pattern (Enomoto et al. 2003; Enomoto 2004), frequently causes Rossby wave breaking (RWB) near the jet exit region (Postel and Hitchman 1999, 2001; Abatzoglou and Magnusdottir 2006; Hitchman and Huesmann 2007). RWB is a large-scale irreversible overturning of isentropic potential vorticity (PV) (McIntyre and Palmer 1985). Summertime RWB over the extratropical Northern Hemisphere is evidently displayed utilizing a 350- or 360-K isentropic surface, which approximately corresponds to the dynamical tropopause and plays an important role in stratosphere–troposphere airmass exchange. The RWB near Japan in summer is closely associated with the Bonin high, which often transports heat waves to the surrounding area (Enomoto et al. 2003; Enomoto 2004). The RWB is accompanied by an equatorward intrusion of a high PV air mass, which frequently promotes enhanced moisture inflow (de Vries 2021) and dynamically induced ascent (Takemura et al. 2017), causing enhanced convection and heavy precipitation. The RWB frequency interannual variability near Japan is thus a significant factor in predicting seasonal temperature and precipitation anomalies during the boreal summer. Using a reanalysis dataset, Takemura et al. (2020) elucidated a linkage between El Niño–Southern Oscillation (ENSO) and RWB frequency near Japan by modulating large-scale tropical convection and the Asian jet. Bowley et al. (2019) also indicated a statistical relationship between ENSO and RWB frequency in the extratropical North Pacific. Takemura et al. (2021) demonstrated the impact of interdecadal sea surface temperature (SST) cooling and the associated suppressed convection over the central part of the tropical North Pacific on the RWB frequency interdecadal variation near Japan. These findings suggest a relationship between the RWB frequency and ENSO, as well as the Pacific decadal oscillation (Mantua et al. 1997). Li and Lin (2015) analyzed multimodel ensemble forecasts initiating in May and indicated that the Asian jet meridional shift was statistically related to ENSO. Lu et al. (2006) also demonstrated an externally forced Asian jet variability owing to precipitation over the western tropical Pacific, indicating the seasonal predictability of the Asian jet derived from tropical convection. However, the impact of the seasonal predictability of ENSO-related tropical circulation on the RWB frequency forecast skills remains unclear.
The Japan Meteorological Agency (JMA) upgraded the seasonal ensemble prediction system for the operational seasonal forecast from the Japan Meteorological Agency/Meteorological Research Institute-Coupled Prediction System version 2 (JMA/MRI-CPS2; hereafter referred to as CPS2; Takaya et al. 2018) to version 3 (CPS3; Hirahara et al. 2023). Using monthly averaged hindcast data in CPS2, Takemura et al. (2022) investigated seasonal predictability for the Asian jet deceleration near Japan in the boreal summer, which is favorable for RWB occurrence (Naoe and Matsuda 2002; Barnes and Hartmann 2012), indicating moderate to high forecast skills of the jet deceleration in CPS2. However, RWB frequency seasonal predictable skill itself has not been examined in previous studies. The underestimation of RWB frequencies in numerical weather forecasting is due not only to their horizontal resolution, but also to the existence of biases in the large-scale climatological atmospheric fields and sea surfaces temperatures in the model (Scaife et al. 2010, 2011). Hence, in this study we elucidate the seasonal predictability of CPS2 and CPS3 RWB frequency. Takemura et al. (2022) also demonstrated a significant relationship between the forecast error of the jet deceleration near Japan and that of ENSO. Hirahara et al. (2023) indicated improved forecast skill of ENSO in CPS3 compared with that of CPS2. They attributed improvement of ENSO prediction to the eddy-permitting higher resolution of the model, improved initial oceanographic conditions and initial perturbations in CPS3. Hence, we expect improved RWB frequency seasonal predictability near Japan in CPS3. The reproducibility of the relationship between the RWB frequency annual variability near Japan and ENSO, which is elucidated by Takemura et al. (2020) based on a reanalysis dataset, should also be examined in advance to evaluate the RWB frequency seasonal predictability in CPS3. Based on the abovementioned perspective, this study evaluates the seasonal predictability of RWB frequency near Japan in July–August (JA), when climatological maximum temperatures are observed after the end of the rainy season in Japan, utilizing daily CPS3 hindcast data. The RWB frequency near Japan is closely related to the frequency of hot days in the country, an important factor in determining summer climate. This study is conducted to deepen our understanding the possible origins of seasonal predictability of summer climate near Japan. This line of approach is expected to provide seasonal forecasters with very significant backgrounds for interpreting the summer seasonal forecast obtained from the numerical forecast model.
2. Data and methods
To evaluate the RWB frequency seasonal predictability near Japan during JA, daily CPS2 and CPS3 hindcast data, which were derived from 6-hourly data, during a 30-yr period from 1991 to 2020, were utilized. The initial dates for each month from which the seasonal prediction starts are listed in Table 1. The ensemble size for each seasonal hindcast, initiated each month, was 10, which comprised five members for each of the two initial dates.
Initial dates for CPS2 and CPS3 hindcasts for each month from January to June.
Daily mean data of the Japanese 55-year Reanalysis (JRA-55; Kobayashi et al. 2015), which was derived from 6-hourly data in JA during the 30-yr period from 1991 to 2020, were also utilized to evaluate large-scale atmospheric circulation. JRA-55 has a horizontal grid interval of 1.25° and 37 pressure levels ranging from 1000 to 1 hPa. The climatology of the CPS hindcast and JRA-55 was defined as the average over the 30-yr period from 1991 to 2020.
To compute the RWB frequency, we utilized a daily blocking index (Pelly and Hoskins 2003) for each longitude, defined by the meridional difference in potential temperature on the dynamic tropopause, defined as the 2 PVU (PV unit; 1 PVU = 10−6 K kg−1 m2 s−1), between a region 15° to the north and south for each grid point. The adopted meridional scale was 15° based on Pelly and Hoskins (2003). Here, we adopted the surface of 2 PVU with the highest altitude when the 2-PVU surface is folded and multivalued in altitude on a certain grid point. RWB was detected when the blocking index had a positive value; there was a high (low) potential temperature to the north (south). Here, to detect RWBs including instantaneous events in summer Asian jet exit region, we did not apply any temporal constraints that take into account the persistence of RWBs in this study. The RWB frequency was defined as the ratio of the number of days on which RWB was detected in JA, according to the method of Takemura et al. (2020), who also indicated that over 70% of RWBs near Japan are categorized as anticyclonic, based on the algorithm of Bowley et al. (2019). This algorithm detects anticyclonic and cyclonic RWB types based on the morphological characteristics of the dynamic tropopause potential temperature contours. The algorithm detects anticyclonic (cyclonic) RWB if the westernmost point along the potential temperature contours within the RWB region is poleward (equatorward) of the easternmost point (Bowley et al. 2019). Therefore, in the detected anticyclonic (cyclonic) RWB, the high potential temperature streamer is located west (east) of the low potential temperature streamer. We also utilized the Bowleys’ algorithm to classify each RWB event into anticyclonic and cyclonic types. The RWB frequency near Japan was defined as the average value over the 25°–45°N, 130°–160°E region, according to the definition of Takemura et al. (2020).
3. Forecast skills of RWB frequency
To elucidate the RWB frequency seasonal predictability for CPS3, we first evaluated reproducibility of the climatology and interannual variability of the RWB frequency.
a. Reproducibility of climatological RWB frequency
Figure 1 depicts the RWB frequency distributions in JA during the 30-yr period from 1991 to 2020 averaged for all ensemble members for six initial months from January to June for CPS3 and the JRA-55 reanalysis counterpart. The RWB frequency in the midlatitude band from 30° to 50°N for JRA-55 (Fig. 1g) depicts two maxima over the North Pacific and from the southern United States to the North Atlantic. The predicted RWB frequency for CPS3 (contours in Figs. 1a–f) is approximately 20% over the midlatitude North Pacific, comparable with that of JRA-55 (contour in Fig. 1g). To grasp how the CPS3 hindcasts represent a zonal distribution of RWB frequencies characterized by two local maxima in the midlatitude North Pacific and North Atlantic, the RWB frequency normalized by its zonal mean is superimposed as shadings in Fig. 1. The RWB frequency normalized by its zonal mean for CPS3 has almost the same distribution as that of JRA-55 (shading in Fig. 1). Although the RWB frequency south of Japan for CPS3 is underestimated compared to that of JRA-55, the RWB frequency spatial distribution from the south to the seas east of Japan is similar to that of JRA-55 (contours in Fig. 1).
Rossby wave breaking (RWB) frequency distributions in July–August (JA) during the 30-yr period from 1991 to 2020 (contour interval: 5%) averaged over all ensemble members of CPS3 for each initial month of (a) January, (b) February, (c) March, (d) April, (e) May, and (f) June. (g) The corresponding distribution for the Japanese 55-year Reanalysis (JRA-55). Shadings indicate the RWB frequency normalized by its zonal mean.
Citation: Weather and Forecasting 38, 6; 10.1175/WAF-D-22-0226.1
b. Interannual variability of RWB frequency
Figures 2 and 3 depict the interannual time series of the predicted RWB frequency near Japan during the 30-yr period from 1991 to 2020 for the CPS3 and CPS2 hindcasts, respectively, initiating from January to June. Although the ensemble-mean RWB frequency of CPS3 is underestimated compared to that of JRA-55, the interannual variability has some similarities between them (black and blue lines in Fig. 2). ACCs between the ensemble mean and JRA-55 range from 0.4 to 0.6 for the initial months of February to June, indicating moderate or high forecast skills for the RWB frequency interannual variability near Japan. In a few years, the RWB frequency intermember maxima near Japan have values close to that of JRA-55 (blue error bars in Fig. 2). Moderate or high forecast skills of the RWB frequency near Japan in CPS3 with positive ACCs suggest high reproducibility of the relationship between the RWB frequency and ENSO, which is described in section 4. Further investigation will be needed in future works to clarify the reason that ACCs for forecasts starting from March to April are higher than those from May and June. Conversely, the RWB frequency in CPS2 (Fig. 3) is not only significantly underestimated compared with that of JRA-55, but also demonstrated much smaller interannual variability than CPS3 (Fig. 2) and JRA-55. Smaller ACCs between the ensemble mean and JRA-55 for CPS2, ranging from −0.2 to 0.3 for initial months from January to June, also indicate a low RWB frequency reproducibility.
RWB frequency interannual variation near Japan (%; see the definition in section 2) during the 30-yr period from 1991 to 2020 for JRA-55 (black lines) and ensemble mean in CPS3 (blue lines). Blue boxes depict the ensemble spread. Blue whiskers indicate the full range of the ensemble members. Red-colored numbers in the upper left of each panel depict the anomaly correlation coefficient (ACC).
Citation: Weather and Forecasting 38, 6; 10.1175/WAF-D-22-0226.1
As in Fig. 2, but for CPS2 depicted by green lines, boxes, and error bars.
Citation: Weather and Forecasting 38, 6; 10.1175/WAF-D-22-0226.1
c. Possible factors for improved RWB frequency forecast skill in CPS3
To investigate possible factors for the low RWB frequency reproducibility near Japan in CPS2, the biases of 200-hPa zonal wind averaged during summer from the JRA-55 climatology for CPS3 and CPS2 are compared in Fig. 4. The Asian jet in CPS3 shifts slightly northward compared with that of JRA-55, with positive and negative biases of small magnitudes north and south of the jet core over Eurasia near the longitudinal area from 60° to 120°E, respectively (Fig. 4a). The subtropical jet in CPS3 decelerates compared with that of JRA-55 from Japan to the east (Fig. 4a). Conversely, the subtropical jet in CPS2 significantly shifts southward compared with that of JRA-55, with positive biases of large magnitudes south of the jet core (Fig. 4b). CPS2 has a much larger bias along the subtropical jet than that of CPS3 (Figs. 4a,b), indicating a significantly improved subtropical jet forecast skill from Eurasia to the seas east of Japan. The significant southward-shifted bias in CPS2 is consistent with the low RWB frequency reproducibility as described in the next subsection (Fig. 3).
Bias of 200-hPa zonal wind (shading; m s−1) averaged during summer from the JRA-55 climatology for hindcasts initiating from 26 Apr during the 30-yr period from 1991 to 2020 for (a) CPS3 and (b) CPS2. Contours indicate the time average of 200-hPa zonal wind during the 30-yr period.
Citation: Weather and Forecasting 38, 6; 10.1175/WAF-D-22-0226.1
d. Reproducibility of anticyclonic and cyclonic RWB frequency types
The aforementioned Asian jet biases are expected to be further related to the reproducibility of the ratios of the anticyclonic and cyclonic types of RWB frequencies near Japan. To test this hypothesis, the horizontal distributions of the anticyclonic and cyclonic RWB frequencies for CPS3 and CPS2 are depicted in Fig. 5. Both RWB types are present near Japan (Takemura et al. 2020). The region with large anticyclonic RWB frequencies extending from the south of Japan to the extratropical North Pacific in CPS3 (Fig. 5a) has a similar shape to that of the large RWB frequencies in the model and JRA-55 (Fig. 1), indicating a high anticyclonic RWB frequency reproducibility. Conversely, the anticyclonic RWB frequency in CPS2 is notably exaggerated near the latitudinal band of 20°N over the subtropical western North Pacific (Fig. 5b), while the higher frequency at 20°N is not due to RWB near the Asian jet exit region. This is because in CPS2, potential vorticity can be expected to reverse meridionally along the southern periphery of eastward extended Tibetan high with the southward shifted Asian jet (Fig. 4b). Furthermore, the anticyclonic RWB frequency in CPS2 is evidently underestimated from the south of Japan to the extratropical North Pacific (Fig. 5b) compared with that of CPS3 and JRA-55 (Figs. 1 and 5a), demonstrating a low anticyclonic RWB frequency reproducibility in CPS2.
(a),(b) Anticyclonic and (c),(d) cyclonic RWB frequencies averaged over JA during the 30-yr period from 1991 to 2020 for hindcasts, initiating from January to June for (left) CPS3 and (right) CPS2.
Citation: Weather and Forecasting 38, 6; 10.1175/WAF-D-22-0226.1
The cyclonic RWB frequency near Japan in CPS3 is much smaller than the anticyclonic RWB frequency (Fig. 5c), which is also consistent with JRA-55 data (Takemura et al. 2020). The cyclonic RWB frequency in CPS2 is exaggerated near the latitudinal band of 20°N over the subtropical western North Pacific (Fig. 5d), suggesting a meridional reversal of potential vorticity in association with the southward-biased jet in CPS2 (Fig. 4b), similar to the anticyclonic RWBs (Fig. 5b).
Further, the anticyclonic and cyclonic RWB frequencies near Japan in CPS3 and CPS2 are quantitatively evaluated. Table 2 depicts the anticyclonic and cyclonic RWB frequencies and ratios averaged over the 30-yr period from 1991 to 2020 for all ensemble hindcasts, initiating from January to June. In CPS3, the anticyclonic RWB frequency (6.5%) is higher than the cyclonic RWB frequency (1.2%), with a high ratio of 72% for the anticyclonic RWB. The ratio of anticyclonic RWB frequency in CPS3 is almost the same as that in JRA-55 (Takemura et al. 2020), indicating the reproducibility of the RWB types. The anticyclonic RWB frequency in CPS2 (4.7%) is smaller than that in CPS3 and almost the same as the cyclonic RWB frequency (3.0%), which is larger than that of CPS3. As a result, the ratio of the cyclonic RWB frequency to the anticyclonic RWB frequency in CPS2 (0.64) is much larger than those of CPS3 (0.19) and JRA-55 (Takemura et al. 2020). The large RWB characteristics biases in CPS2 are presumably associated with the significant southward-shifted Asian jet bias (Fig. 4b).
Anticyclonic and cyclonic RWB frequencies and ratios averaged during the period from 1991 to 2020 for all ensemble hindcasts initiating from January to June in CPS3 and CPS2. The fifth row indicates the unclassifiable type (i.e., neither anticyclonic nor cyclonic types) of RWBs, according to the method of Bowley et al. (2019). The bottom row depicts the ratio of the cyclonic to anticyclonic RWB frequencies.
4. Reproducibility of the relationship between RWB frequency and ENSO
To investigate the high reproducibility of the relationship between the RWB frequency near Japan and ENSO-related atmospheric circulation anomalies (Takemura et al. 2020) in CPS3, this section describes a regression analysis of oceanographic and atmospheric circulation on the RWB frequency using all ensemble hindcasts initiating from January to June.
Using JRA-55, Takemura et al. (2020) elucidated that the RWB frequency near Japan increases (decreases) during La Niña (El Niño) years. Models with high RWB reproducibility are expected to accurately represent the relationship between the RWB frequency and ENSO. Figure 6 depicts the distributions of anomalous RWB frequencies in CPS3 composited during El Niño and La Niña years in the summer. Herein, the El Niño (La Niña) years in summer, which are enlisted in Table 3, are defined as years when the 5-month running mean SST anomaly averaged over the Niño-3 region (5°S–5°N, 150°–90°W) continues at +0.5°C or higher (−0.5°C or lower) for at least six consecutive months. The SST anomaly is defined as the deviation from the average over the latest sliding 30 years. The predicted RWB frequency during El Niño (La Niña) years is statistically smaller (larger) than the corresponding average during the 30-yr period from 1991 to 2020. This statistical result is consistent with that of Takemura et al. (2020), demonstrating the reproducibility of the relationship between RWB frequency and ENSO in CPS3.
Composite anomalies of the RWB frequency (shading; %) during (a) El Niño years, (b) La Niña years, and (c) the difference between La Niña and El Niño years. Dots denote the anomalies with a 99% confidence level. Contours indicate the averaged RWB frequency in JA during the 30-yr period from 1991 to 2020 for hindcasts starting from January to June (contour interval: 5%).
Citation: Weather and Forecasting 38, 6; 10.1175/WAF-D-22-0226.1
El Niño and La Niña years in summer detected using the definition of El Niño–Southern Oscillation (ENSO) events in JMA (see text for the detail).
Global atmospheric conditions related to the reproducibility of the relationship between the RWB frequency near Japan and ENSO (Takemura et al. 2020) are also investigated. Figure 7 depicts the oceanographic and atmospheric circulation anomalies regressed onto the RWB frequency near Japan for CPS3 and CPS2. In CPS3, the regressed SST demonstrated a La Niña–like pattern with negative anomalies over the central to eastern equatorial Pacific and positive anomalies from the South China Sea to the western tropical North Pacific (Fig. 7a). This anomalous SST pattern in the tropical Pacific is consistent with the reanalysis (Takemura et al. 2020). In the tropical southeastern Indian Ocean (i.e., west of Sumatra), there are significantly positive SST anomalies, which share a similar horizontal shape with the negative phase of the Indian Ocean dipole (IOD) mode (Saji et al. 1999). The regressed precipitation, which is characterized by suppressed convection over the equatorial Pacific and enhanced convection from the eastern tropical Indian Ocean to the Maritime Continent (Fig. 7b), resembles the characteristics of La Niña and negative IOD (Fig. 7a). Positive SST anomalies and enhanced convection in the southeastern part of the tropical Indian Ocean are not seen in the reanalysis (Takemura et al. 2020). This suggests a false relationship between the RWB frequency and IOD in CPS3.
(a),(f) Sea surface temperature (SST; °C); (b),(g) precipitation (mm day−1); (c),(h) 200-hPa zonal wind (m s−1); (d),(i) 200-hPa height (m); and (e),(j) 200-hPa meridional wind (m s−1) anomalies regressed onto the RWB frequency near Japan in JA using all ensemble hindcasts starting from January to June during the 30-yr period from 1991 to 2020 in (left) CPS3 and (right) CPS2. Dots indicate the regression with a 99% confidence level. Red dashed boxes denote the region near Japan (25°–45°N, 130°–160°E) to compute the area-averaged RWB frequency. Green contours in (c), (e), (h), and (j) depict 200-hPa zonal wind in each ensemble hindcast averaged during the period from 1991 to 2020.
Citation: Weather and Forecasting 38, 6; 10.1175/WAF-D-22-0226.1
To determine an explanation for this inconsistency between CPS3 and the reanalysis, we examined the predicted dipole mode index (DMI) interannual variation in CPS3. Here, the DMI is defined as the difference between the deviation of the monthly mean SST over the region spanning 10°S–10°N, 50°–70°E from the latest sliding 30-yr average and the corresponding deviation over the region covering 10°S–0°, 90°–110°E. A positive (negative) DMI indicates a positive (negative) IOD phase. Figure 8 depicts the interannual variation in the predicted DMI index for all hindcasts starting from May for CPS3, CPS2, and JRA-55. The predicted DMI interannual variability for CPS3 had a statistically significant relationship with the reanalysis (left panels of Figs. 8a,b), with correlation coefficients greater than 0.7. However, the DMI variability in CPS3 was much larger than that in JRA-55 (right panel of Figs. 8a,b). Figure 9 further depicts the root-mean-square error (RMSE) of the SST anomalies averaged during summer from the JRA-55 climatology for CPS3. The RMSE of SST demonstrates a maximum in the tropical southeastern Indian Ocean (Fig. 9), which is consistent with the exaggerated variability of the predicted DMI for CPS3 (right panel of Figs. 8a,b). Thus, the regressed SST pattern on the RWB frequency near Japan (Fig. 7a) in CPS3 with significant positive anomalies in the southeastern Indian Ocean could be due to the large SST anomaly biases in the tropical southeastern Indian Ocean in CPS3. Conversely, the predicted DMI interannual variation in CPS3 has larger positive correlations with the reanalysis (Figs. 8a,b) than did CPS2. Therefore, the DMI prediction skill in CPS3 is higher than that in CPS2, whereas SST in the tropical southeastern Indian Ocean still has large biases, which are presumed to cause forecast errors in the RWB frequency near Japan.
Interannual time series of SST indices (shown at left) and scatter diagrams of the observed SST indices obtained from JRA-55 (x axis) against the predicted ones (shown at right) for (a)–(d) CPS3 and (e)–(h) CPS2 hindcasts initiating from May. The SST indices are DMI for (a), (b), (e), and (f) and Niño-3 SST for (c), (d), (g), and (h), which are averaged in July in (a), (c), (e), and (g) and in August in (b), (d), (f), and (h) from 1991 to 2020. Black and blue lines in the time series panels denote the analysis and ensemble mean, respectively. Dark (light) blue open circles correspond to predictions of the ensemble mean (each ensemble member). “COR” at the top right of (a)–(d) and (e)–(h) indicates correlation coefficients between the ensemble mean and analysis. Blue and light-blue circles in the scatter diagrams indicate ensemble mean and members, respectively.
Citation: Weather and Forecasting 38, 6; 10.1175/WAF-D-22-0226.1
Root-mean-square error of SST anomalies (shadings; K) for CPS3 averaged during summer from the JRA-55 climatology for hindcasts starting from 26 Apr during the 30-yr period from 1991 to 2020.
Citation: Weather and Forecasting 38, 6; 10.1175/WAF-D-22-0226.1
The regressed 200-hPa zonal wind in the tropics demonstrates positive anomalies from the Maritime Continent to the central Pacific and negative anomalies from the eastern Pacific to the western Indian Ocean (Fig. 7c), indicating an enhanced and slightly westward-shifted Walker circulation. The upper-level zonal wind from Eurasia to the east of Japan demonstrates westerly and easterly wind anomalies north and south of 40°N, respectively, indicating a northward shift of the Asian jet. The regressed 200-hPa height demonstrates zonally elongated positive anomalies in the Northern Hemisphere midlatitudes (Fig. 7d), which correspond to the atmospheric response to La Niña conditions (e.g., Robinson 2002; Alexander et al. 2004; Seager et al. 2010). In the tropical North Pacific, the upper-level height demonstrates negative anomalies, corresponding to the enhanced mid-Pacific trough (Fig. 7d). The regressed 200-hPa meridional wind anomalies (Fig. 7e) depict a wavy pattern along the Asian jet (green contours in Fig. 7e), corresponding to the Silk Road pattern (Enomoto et al. 2003; Enomoto 2004). The regressed wavy pattern implies that the enhanced Silk Road pattern is favorable for increased and decreased RWB frequencies near Japan in La Niña and El Niño conditions, respectively. These regressed anomalous circulations in CPS3 are consistent with the reanalysis (Takemura et al. 2020) except for SST anomalies associated with the negative IOD, indicating reproducibility of the relationship between the RWB frequency near Japan and ENSO.
The regression analysis results for CPS2 are also depicted in Fig. 7. The signal associated with ENSO is absent in CPS2 as the RWB frequency in CPS2 is significantly underestimated compared with that in JRA-55 and has smaller interannual variability (Fig. 3). We do not observe any La Niña–like or negative IOD-like SST patterns in the equatorial Pacific (Fig. 7f), contrary to the regressed SST in CPS3 (Fig. 7a) and the reanalysis (Takemura et al. 2020). In association with the absence of ENSO and IOD signals, the regressed precipitation demonstrates no significant anomaly in the equatorial region (Fig. 7g). The regressed 200-hPa zonal wind anomalies in the tropics are positive in the western hemisphere (Fig. 7h), which contradicts the regression in CPS3 (Fig. 7c). The upper-level zonal wind anomalies from Eurasia to the east of Japan are positive and negative north and south of 40°N (Fig. 7h), respectively, indicating a northward shift of the Asian jet. Although the northward-shifted jet is consistent with the results of CPS3 (Fig. 7c), the SST anomalies do not have ENSO-related characteristics (Fig. 7f), and positive precipitation anomalies are evident near the Asian monsoon region from South Asia to the east of the Philippines (Fig. 7g). The regressed 200-hPa height demonstrates zonally elongated positive anomalies in the extratropical Northern Hemisphere (Fig. 7i) associated with the Asian jet northward shift (Fig. 7h). Conversely, the regressed height has no significant negative anomalies in the tropical Pacific consistent with an ENSO response in CPS3 (Fig. 7d). The regressed 200-hPa meridional wind anomalies (Fig. 7j) demonstrate a wavy pattern along the Asian jet (green contours in Fig. 7j), consistent with CPS3 values (Fig. 7e), corresponding to the Silk Road pattern (Enomoto et al. 2003; Enomoto 2004). These regressed anomalous extratropical circulations in CPS2 are consistent with CPS3, but the precipitation and tropical SST anomalies in CPS2 have diverse characteristics from those in CPS3. These results indicate that the relationship between the RWB frequency near Japan and ENSO in CPS3 has been significantly improved compared with CPS2 data.
Although the regression analysis assumes a symmetric response of SST and atmospheric circulation anomalies to the sign of the RWB frequency anomaly, these anomalies could also demonstrate an asymmetric response. To explore the antisymmetric relationship in CPS3 between ENSO and RWB frequency, scatterplots of SST, precipitation, and upper-tropospheric zonal wind north of the Asian jet axis against the RWB frequency are depicted in Fig. 10. For hindcasts with larger RWB frequency, SST over the Niño-3 region (5°S–5°N, 150°–90°W) is relatively low (Fig. 10a), convective activity over the equatorial Pacific is suppressed (Fig. 10b), and upper-level westerly wind is stronger (Fig. 10c). Additionally, these relationships have nonnegligible nonlinear characteristics (Fig. 10) that are not incorporated in the regression analysis. Thus, we conduct a supplemental composite analysis of SST and atmospheric circulation anomalies for the CPS3 hindcasts of the predicted RWB frequency with magnitudes larger than +1 standard deviation (σ) and lower than −1σ, respectively. A similar horizontal anomaly pattern with opposite polarity was detected in the composite of the enhanced and suppressed RWB frequencies (not depicted), supporting the regression analysis results (Fig. 7).
Scatterplots of RWB frequency (x axis; %) and (shown on the y axes) (a) SST averaged over Niño-3 region (5°S–5°N, 150°–90°W; °C), (b) precipitation averaged over 5°S–5°N, 150°–90°W (mm day−1) and (c) 200-hPa zonal wind averaged over a region spanning 40°–50°N, 90°E–180°, which is to the north of the climatological Asian jet axis (m s−1).
Citation: Weather and Forecasting 38, 6; 10.1175/WAF-D-22-0226.1
We conducted the same analyses for ensemble hindcasts for each month and obtained approximately the same results (not depicted). Hence, regression analyses verified the reproducibility of the relationship between RWB frequency and ENSO in CPS3.
5. Conclusions and discussion
This study investigated the RWB frequency seasonal predictability near Japan in JA using daily CPS3 hindcast data, which is a new operational seasonal prediction system for the JMA. Although the RWB frequency near Japan during JA in CPS3 was underestimated compared with that of the reanalysis (JRA-55), the RWB frequency interannual variability was generally well reproduced with a moderate or higher forecast skill for ensemble hindcasts, initiating from February to June. The RWB frequency forecast skill in CPS3 was much more enhanced than that of CPS2, which was the previous operational seasonal prediction system of the JMA, in association with the improvement of the model bias related to the Asian jet meridional position. The prediction for the ratio of anticyclonic RWB frequency to cyclonic frequency in CPS3 was also significantly improved compared with that of CPS2.
Regression analysis against the RWB frequency near Japan using all ensemble hindcasts in CPS3 indicated that the predicted RWB frequency increased during La Niña conditions, consistent with the conclusions obtained by Takemura et al. (2020) through the JRA-55 analysis. The regressed anomalies in CPS3 with larger RWB frequency demonstrated an SST pattern similar to that of La Niña and the negative phase of IOD with a consistent tropical convection anomaly. For the La Niña condition in CPS3, the regressed upper-level height anomalies were negative over the tropical Pacific and positive over the extratropical Northern Hemisphere, corresponding to the enhanced mid-Pacific trough and northward-shifted subtropical jet. The regressed meridional wind anomaly is characterized by a wavy pattern along the Asian jet over Eurasia, consistent with the relationship between the Silk Road pattern and the RWB near the Asian jet exit region (Enomoto et al. 2003; Enomoto 2004).
Regression analysis using all ensemble hindcasts for CPS3 detected significant positive SST anomalies associated with the negative IOD in the southeastern part of the tropical Indian Ocean, which is inconsistent with the regression analysis using JRA-55. The impacts of a negative IOD on the East Asian summer climate have not been completely examined in previous studies as thoroughly as positive IOD. The latter impact on anomalously hot summer conditions in East Asia was elucidated by Guan and Yamagata (2003) through quasi-stationary Rossby wave propagation enhancement along the Asian jet stream based on a case study of 1994. Takemura and Shimpo (2019) indicated that the northeastward extension of the Tibetan high toward East Asia during positive IOD events is closely associated with the frequent RWB occurrence near Japan. Alternatively, Qiu et al. (2014) focused on the impact of the negative IOD and indicated that increased moisture inflows toward East Asia during negative IOD events cause unusual weather conditions that give rise to natural disasters. This result suggests that RWB events near Japan are infrequent during negative IOD periods, which is inconsistent with Fig. 7a, indicating that negative IOD prefers frequent RWB events. Hence, the impact of a negative IOD on the RWB frequency near Japan should be quantitatively evaluated in future work.
The improved forecast skill for ENSO in CPS3, as demonstrated by Hirahara et al. (2023), could be associated with the improved RWB frequency reproducibility near Japan for CPS3, as depicted in Figs. 2 and 3. The significantly enhanced prediction skill of ENSO for CPS3 (CPS2) is also verified by Figs. 8c and 8d (Figs. 8g,h), indicating interannual variations in the predicted Niño-3 SST in July and August for hindcasts initiating in May. The prediction reproduced the observed Niño-3 SST interannual variability with high temporal correlation coefficients larger than 0.90. Moreover, the magnitude of the interannual variation in the predicted Niño-3 SST is similar to the observations. CPS3 has larger temporal correlation coefficients than CPS2; hence, RWB frequency reproducibility near Japan is significantly improved for CPS3 compared with CPS2.
The newly obtained insight into the improved interannual variability reproducibility of the RWB frequency near Japan for CPS3 provides forecasters with useful guidance to enhance forecast skills on the occurrence probability of anomalous weather conditions during summers in Japan. This is because the majority of these anomalous weather events, such as unprecedented heat waves and heavy rainfall events, are associated with an enhanced upper-tropospheric anomalous anticyclone or a cutoff cold low in the exit region of the summer Asian jet near Japan, which is closely related to the RWB (Enomoto et al. 2009). Thus, the RWB frequency interannual variability reproducibility near Japan should be enhanced to improve the prediction skill for the occurrence probability of anomalous weather in Japan during the summer season. Although this study focused on the RWB frequency near Japan to assess the frequency of hot days in the country, the predictability of RWB strength and duration should be examined in future works.
Acknowledgments.
The authors are very grateful to Dr. Ben Kirtman and three anonymous reviewers for their constructive and helpful comments. The Generic Mapping Tools (GMT) were used to create the graphics. This work was supported by JSPS KAKENHI Grant 22H04493.
Data availability statement.
The datasets analyzed in this study [the Japanese 55-year Reanalysis (JRA-55)] are available at https://jra.kishou.go.jp/JRA-55/index_en.html. The hindcast maps and datasets of JMA-MRI-CPS2 and CPS3 are available at https://www.data.jma.go.jp/tcc/tcc/products/model/index.html.
REFERENCES
Abatzoglou, J. T., and G. Magnusdottir, 2006: Planetary wave breaking and nonlinear reflection: Seasonal cycle and interannual variability. J. Climate, 19, 6139–6152, https://doi.org/10.1175/JCLI3968.1.
Alexander, M. A., N.-C. Lau, and J. D. Scott, 2004: Broadening the atmospheric bridge paradigm: ENSO teleconnections to the tropical West Pacific-Indian Oceans over the seasonal cycle and to the North Pacific in summer. Earth’s Climate: The Ocean-Atmosphere Interaction, Geophys. Monogr., Vol. 147, Amer. Geophys. Union, 85–103, https://doi.org/10.1029/147GM05.
Barnes, E. A., and D. L. Hartmann, 2012: Detection of Rossby wave breaking and its response to shifts of the midlatitude jet with climate change. J. Geophys. Res., 117, D09117, https://doi.org/10.1029/2012JD017469.
Bowley, K. A., J. R. Gyakum, and E. H. Atallah, 2019: A new perspective toward cataloging Northern Hemisphere Rossby wave breaking on the dynamical tropopause. Mon. Wea. Rev., 147, 409–431, https://doi.org/10.1175/MWR-D-18-0131.1.
de Vries, A. J., 2021: A global climatological perspective on the importance of Rossby wave breaking and intense moisture transport for extreme precipitation events. Wea. Climate Dyn., 2, 129–161, https://doi.org/10.5194/wcd-2-129-2021.
Enomoto, T., 2004: Interannual variability of the Bonin high associated with the propagation of Rossby waves along the Asian jet. J. Meteor. Soc. Japan, 82, 1019–1034, https://doi.org/10.2151/jmsj.2004.1019.
Enomoto, T., B. J. Hoskins, and Y. Matsuda, 2003: The formation mechanism of the Bonin high in August. Quart. J. Roy. Meteor. Soc., 129, 157–178, https://doi.org/10.1256/qj.01.211.
Enomoto, T., H. Endo, Y. Harada, and W. Ohfuchi, 2009: Relationship between high-impact weather events in Japan and propagation of Rossby waves along the Asian jet in July 2004. J. Meteor. Soc. Japan, 87, 139–156, https://doi.org/10.2151/jmsj.87.139.
Guan, Z., and T. Yamagata, 2003: The unusual summer of 1994 in East Asia: IOD teleconnections. Geophys. Res. Lett., 30, 1544, https://doi.org/10.1029/2002GL016831.
Hirahara, S., and Coauthors, 2023: Japan Meteorological Agency/Meteorological Research Institute Coupled Prediction System version 3 (JMA/MRI-CPS3). J. Meteor. Soc. Japan, 101, 149–169, https://doi.org/10.2151/jmsj.2023-009.
Hitchman, M. H., and A. S. Huesmann, 2007: A seasonal climatology of Rossby wave breaking in the 320–2000-K layer. J. Atmos. Sci., 64, 1922–1940, https://doi.org/10.1175/JAS3927.1.
Kobayashi, S., and Coauthors, 2015: The JRA-55 reanalysis: General specifications and basic characteristics. J. Meteor. Soc. Japan, 93, 5–48, https://doi.org/10.2151/jmsj.2015-001.
Li, C., and Z. Lin, 2015: Predictability of the summer East Asian upper-tropospheric westerly jet in ensembles multi-model forecasts. Adv. Atmos. Sci., 32, 1669–1682, https://doi.org/10.1007/s00376-015-5057-z.
Lu, R., J.-H. Oh, and B.-J. Kim, 2002: A teleconnection pattern in upper-level meridional wind over the North African and Eurasian continent in summer. Tellus, 54A, 44–55, https://doi.org/10.3402/tellusa.v54i1.12122.
Lu, R., Y. Li, and B. Dong, 2006: External and internal summer atmospheric variability in the western North Pacific and East Asia. J. Meteor. Soc. Japan, 84, 447–462, https://doi.org/10.2151/jmsj.84.447.
Mantua, N. J., S. R. Hare, Y. Zhang, J. M. Wallace, and R. C. Francis, 1997: A Pacific interdecadal climate oscillation with impacts on salmon production. Bull. Amer. Meteor. Soc., 78, 1069–1080, https://doi.org/10.1175/1520-0477(1997)078<1069:APICOW>2.0.CO;2.
McIntyre, M. E., and T. N. Palmer, 1985: A note on the general concept of wave breaking for Rossby and gravity waves. Pure Appl. Geophys., 123, 964–975, https://doi.org/10.1007/BF00876984.
Naoe, H., and Y. Matsuda, 2002: Rossby wave propagation and blocking formation in realistic basic flows. J. Meteor. Soc. Japan, 80, 717–731, https://doi.org/10.2151/jmsj.80.717.
Pelly, J. L., and B. J. Hoskins, 2003: A new perspective on blocking. J. Atmos. Sci., 60, 743–755, https://doi.org/10.1175/1520-0469(2003)060<0743:ANPOB>2.0.CO;2.
Postel, G. A., and M. H. Hitchman, 1999: A climatology of Rossby wave breaking along the subtropical tropopause. J. Atmos. Sci., 56, 359–373, https://doi.org/10.1175/1520-0469(1999)056<0359:ACORWB>2.0.CO;2.
Postel, G. A., and M. H. Hitchman, 2001: A case study of Rossby wave breaking along the subtropical tropopause. Mon. Wea. Rev., 129, 2555–2569, https://doi.org/10.1175/1520-0493(2001)129<2555:ACSORW>2.0.CO;2.
Qiu, Y., W. Cai, X. Guo, and B. Ng, 2014: The asymmetric influence of the positive and negative IOD events on China’s rainfall. Sci. Rep., 4, 4943, https://doi.org/10.1038/srep04943.
Robinson, W. A., 2002: On the midlatitude thermal response to tropical warmth. Geophys. Res. Lett., 29, 1190, https://doi.org/10.1029/2001GL014158.
Saji, N. H., B. N. Goswami, P. N. Vinayachandran, and T. Yamagata, 1999: A dipole mode in the tropical Indian Ocean. Nature, 401, 360–363, https://doi.org/10.1038/43854.
Scaife, A. A., T. Woollings, J. Knight, G. Martin, and T. Hinton, 2010: Atmospheric blocking and mean biases in climate models. J. Climate, 23, 6143–6152, https://doi.org/10.1175/2010JCLI3728.1.
Scaife, A. A., and Coauthors, 2011: Improved Atlantic winter blocking in a climate model. Geophys. Res. Lett., 38, L23703, https://doi.org/10.1029/2011GL049573.
Seager, R., N. Naik, W. Baethgen, A. Robertson, Y. Kushnir, J. Nakamura, and S. Jurburg, 2010: Tropical oceanic causes of interannual to multidecadal precipitation variability in southeast South America over the past century. J. Climate, 23, 5517–5539, https://doi.org/10.1175/2010JCLI3578.1.
Takaya, Y., and Coauthors, 2018: Japan Meteorological Agency/Meteorological Research Institute-Coupled Prediction System version 2 (JMA/MRI-CPS2): Atmosphere-land-ocean-sea ice coupled prediction system for operational seasonal forecasting. Climate Dyn., 50, 751–765, https://doi.org/10.1007/s00382-017-3638-5.
Takemura, K., and A. Shimpo, 2019: Influence of positive IOD events on the northeastward extension of the Tibetan High and East Asian climate condition in boreal summer to early autumn. SOLA, 15, 75–79, https://doi.org/10.2151/sola.2019-015.
Takemura, K., Y. Kubo, and S. Maeda, 2017: Relation between a Rossby wave-breaking event and enhanced convective activities in August 2016. SOLA, 13, 120–124, https://doi.org/10.2151/sola.2017-022.
Takemura, K., H. Mukougawa, and S. Maeda, 2020: Large-scale atmospheric circulation related to frequent Rossby wave breaking near Japan in boreal summer. J. Climate, 33, 6731–6744, https://doi.org/10.1175/JCLI-D-19-0958.1.
Takemura, K., H. Mukougawa, and S. Maeda, 2021: Interdecadal variability of Rossby wave breaking frequency near Japan in August. SOLA, 17, 125–129, https://doi.org/10.2151/sola.2021-021.
Takemura, K., H. Mukougawa, Y. Takaya, and S. Maeda, 2022: Seasonal predictability of summertime Asian jet deceleration near Japan in JMA/MRI-CPS2. SOLA, 18, 19–24, https://doi.org/10.2151/sola.2022-004.
