Kuroshio-Enhanced Convective Rainband Associated with an Extratropical Cyclone in the Cold Season

Abstract

This study examined the roles of heat fluxes from the Kuroshio Current in enhancing a frontal convective rainband associated with an extratropical cyclone that brought record-breaking heavy rainfall to Miyake Island, Japan, in January 2017. A simulation of the rainband (control experiment) and sensitivity experiments without the sensible and latent heat fluxes from the Kuroshio Current were conducted using a regional cloud-resolving model. The rainband developed along a non-classic front (outer front), which formed to the north of a warm front associated with the cyclone. The control experiment reproduced the intensity and migration of the rainband well. As the rainband developed, the heat fluxes from the Kuroshio Current became evident around the cold conveyor belt of the cyclone to the south of the rainband. The latent heat fluxes were ∼ 2.3 times larger than the sensible heat fluxes. Comparisons between the control and sensitivity experiments demonstrated that the heat fluxes, especially the latent heat fluxes, enhanced the rainband. The sensible heat fluxes slightly intensified convective instability in the lower troposphere, whereas the latent heat fluxes significantly increased the near-surface moisture content and the convective instability. A frontal updraft along the outer front released the increased convective instability, which intensified the moisture convergence, condensation, and updraft, enhancing the rainband. The findings show that the heat fluxes from the Kuroshio Current, especially the latent heat fluxes, enhanced the heavy rainfall-producing rainband by increasing the moisture content and the convective instability.

1. Introduction

Recent studies have shown that the midlatitude ocean, especially the Kuroshio Current, is critical to the occurrence of heavy rainfall events around Japan (Miyama et al. 2012; Manda et al. 2014; Iizuka and Nakamura 2019; Sekizawa et al. 2019; Yamamoto 2020). Miyama et al. (2012) examined a convective rainband that developed along the Kuroshio Current over the East China Sea in May 2010 and found that the high sea surface temperature (SST) of the Kuroshio Current intensified the rainband by decreasing static stability. Manda et al. (2014) reported that the magnitude of the SST over the East China Sea significantly affected heavy rainfall over the Kyushu region, Japan, during the Baiu/Meiyu season. Iizuka and Nakamura (2019) showed that the differences in SST distributions over the Sea of Japan altered the evolution of an intense convective precipitation system that formed over the Tohoku region, Japan, in August 2013. Sekizawa et al. (2019) demonstrated that moisture supply from the Kuroshio Current partly contributed to a torrential rainfall event over western Japan in July 2018. Yamamoto (2020) showed that warm SST anomalies in the Tsushima Strait influenced heavy rainfall in the region in August 2013.

All the studies focused on heavy rainfall events during the warm season. Furthermore, Hirata et al. (2019; hereafter H19) reported that the Kuroshio Current could affect heavy rainfall during the cold season. H19 specifically focused on a heavy rainfall event on Miyake Island (Fig. 1), Japan, on 8 January 2017. The Japan Meteorological Agency (JMA) observatory on the island, which opened in 1942, recorded this recordbreaking daily precipitation event (139.5 mm). Two rainbands that developed over a warm front and a non-classic front (outer front), which formed to the north of the warm front associated with an extratropical cyclone, caused heavy rainfall. Figure 1 shows the frontal structure of the cyclone and associated rainband at 06:00 UTC on 8 January 2017. A surface front extended from the center of the cyclone toward the southeast, corresponding to the warm front (Fig. 1a). Another front lay to the north of the warm front; this front was the outer front (H19). Then, a convective rainband (the outer frontal rainband) developed along this outer front (Fig. 1b). This rainband passed over Miyake Island at approximately 06:00 UTC on 8 January 2017, bringing heavy rainfall to the island. H19 showed that sensible heat fluxes from the Kuroshio Current increased markedly just before cyclogenesis compared with those fluxes from the colder SST region to the north of the current. They demonstrated that this meridional contrast in sensible heating led to the initial formation of the outer front that produced heavy rainfall and concluded that the Kuroshio Current contributed to triggering heavy rainfall by modifying the frontal structure of the cyclone.

In addition to the role of the Kuroshio Current in modifying the frontal structure, it could have played another crucial role in the occurrence of heavy rainfall. As shown by Hirata et al. (2015, 2016, 2018), when extratropical cyclones develop over the Kuroshio Current, strong sensible and latent heat fluxes occur from the warm current around the cyclone center, especially around the cold conveyor belt (CCB; e.g., Carlson 1980; Schultz 2001), enhancing latent heat release. Because latent heat release is critical for the rapid development of extratropical cyclones (e.g., Catto 2016), the above-cited and other studies (e.g., Nuss and Kamikawa 1990; Gyakum and Danielson 2000; Kuwano-Yoshida and Minobe 2017) primarily focused on the roles of surface heat fluxes and enhanced latent heating in cyclone intensification. Moreover, such enhanced latent heat release implies an increase in the rainfall. Thus, we consider that heat fluxes from the warm current contribute to heavy rainfall events associated with extratropical cyclones through rainfall intensification. However, this science issue has not yet been clarified sufficiently. H19 highlighted the formation of the outer front before cyclogenesis but did not address this issue.

Therefore, this study examined the roles of surface heat fluxes from the Kuroshio Current in enhancing rainfall associated with the extratropical cyclones during the cold season. This study examined the same event studied by H19. As mentioned above, the event was attributed to two rainbands along the warm and outer fronts. In particular, the convective rainband along the outer front (outer frontal rainband) grew over the Kuroshio Current, causing intense rainfall on Miyake Island over a short time (H19). Thus, this investigation sheds light on the development of the outer frontal rainband.resolution runs and associated CRE changes and implications. The conclusion and further discussion are provided in Section 5.

2. Model, experiments, and data

2.1 Model

The influence of the Kuroshio Current on the outer frontal rainband was examined by conducting cloudresolving numerical experiments of the rainband using the Cloud Resolving Storm Simulator (CReSS; Tsuboki and Sakaikibara 2002). The settings used for the cloud microphysics and for parameterization of sub-grid scale turbulent eddies were the same as those used in H19. The horizontal grid spacing was set at 0.02° in both longitude and latitude. The number of vertical layers was 57, with the lowest layer set to 100 m and the height of the topmost level at 22,800 m. The model domain spanned 115–170°E and 18–52°N. The experimental period was from 12:00 UTC on 7 January 2017 to 08:00 UTC on 8 January 2017, including the time taken for the outer frontal rainband to pass over Miyake Island. We used JMA Global Spectral Model (GSM) data (Japan Meteorological Agency 2013) as the initial and lateral boundary conditions for the simulations. We also used the Japan Coastal Ocean Predictability Experiment 2 (JCOPE2; Miyazawa et al. 2009) as the SST condition and assumed that the condition was fixed with time as the initial condition in the simulations (Fig. 2). The JCOPE2 data used in this study are the daily mean data.

Fig. 1.

Fig. 1. (a) Map of the magnitude of the horizontal gradient of potential temperature (shading; K km⁻¹), geopotential height (contours; m), and horizontal wind (arrows; m s⁻¹) at 975 hPa derived from the Japan Meteorological Agency (JMA) mesoscale objective analysis data (Japan Meteorological Agency 2013) at 06:00 UTC on 8 January 2017. The contour interval is 20 m. (b) Map of precipitation intensity (shading; mm h⁻¹) derived from JMA radars at 06:00 UTC on 8 January 2017.

Fig. 2.

(a) Map of sea surface temperature (SST) (shading and contours; °C) and sea surface currents (arrows; m s⁻¹) on 7 January 2017. The ocean data are derived from the Japan Coastal Ocean Predictability Experiment 2 (JCOPE2; Miyazawa et al. 2009). The SST data are daily mean data. Contour intervals measure 0.5°C. The open circle indicates Miyake Island. The surface heat fluxes from the ocean area enclosed by the green rectangle were removed in no sensible heat fluxes (NSH) and no latent heat fluxes (NLH) runs.

2.2 Experiments

We conducted a rainband simulation experiment [control (CNTL) run] and two sensitivity experiments without surface heat fluxes. We first conducted the CNTL run. Figures 3a–d (3e–h) show the horizontal distributions of the sensible (latent) heat fluxes in the CNTL run during the developmental stage of the outer front. The figures also show a two-dimensional (2D) frontogenesis function (e.g., Martin 2006) at 980 hPa, horizontal wind vectors at 980 hPa, and sea level pressure (SLP). During this stage, a frontogenesis zone extending from the southwest to the northeast, corresponding to the outer front, became evident to the northeast of the cyclone center. At the same time, the sensible and latent heat fluxes from the Kuroshio Current increased around the outer front. The magnitude of the latent heat fluxes (∼ 450 W m⁻²) was ∼ 2.3 times greater than that of the sensible heat fluxes (∼ 200 W m⁻²). Because the CCB and warm currents are conducive to increases in surface heat fluxes (e.g., Hirata et al. 2015, 2016, 2018), CCB could be critical for enhancing these fluxes, as discussed later in greater detail (Section 4.1). The effects of these fluxes on the outer frontal rainband were clarified by conducting two experiments with the following conditions: no sensible heat fluxes (NSH) and no latent heat fluxes (NLH) from the ocean region around the Kuroshio Current (136–142.5°E, 30–34.5°N, the region indicated by the green box in Figs. 2, 3), after 00:00 UTC on 8 January 2017 when these fluxes started to increase (Fig. 3).

Fig. 3.

Horizontal distributions of surface sensible heat fluxes (shading; W m⁻²), positive two-dimensional (2D) frontogenesis function at 980 hPa (blue contours; 10⁻⁴ K km⁻¹ s⁻¹), horizontal wind at 980 hPa (arrows; m s⁻¹), and SLP (black contours; hPa) in the CNTL run at (a) 21:00 UTC on 7 January, (b) 00:00 UTC on 8 January, (c) 03:00 UTC on 8 January, and (d) 06:00 UTC on 8 January 2017. The intervals of the blue and black contours are 1 × 10⁻⁴ K km⁻¹ s⁻¹ and 4 hPa, respectively. Winds < 13 m s⁻¹ are not shown. The surface heat fluxes from the ocean area enclosed by the green rectangle were removed in the NSH and NLH runs. (e)–(h) is the same as (a)–(d), except for the surface latent heat fluxes (shading; W m⁻²).

2.3 Data

To overview the frontal structure of the cyclone in Fig. 1a, we used JMA mesoscale objective analysis data with a spatial resolution of 0.125° in longitude and 0.1° in latitude (Japan Meteorological Agency 2013). Moreover, to show the observed features of the outer frontal rainband in Fig. 1b, we used precipitation intensity data derived from the operational JMA radars.

3. Mesoscale features of the outer frontal rainband

We compared the simulated outer frontal rainband among the three experiments. Figures 4a–c show the horizontal distribution of precipitation intensity at 04:00 UTC, 06:00 UTC, and 08:00 UTC on 8 January 2017 in the CNTL run. Figures 4d–f and 4g–i are the same as Figs. 4a–c but show the NSH and NLH runs, respectively. The CNTL run simulated the genesis of convective cells within the outer frontal rainband, their northward migration, and their passage around Miyake Island (Figs. 4a–c). These features of the rainband in the CNTL run resembled those in the observations (Fig. 1b; Fig. 4 in H19), and the CNTL run successfully reproduced the features of the outer frontal rainband. The NSH run simulated the convective cells along the outer front (Figs. 4d–f), and the basic features of the rainband in the NSH run resembled those in the CNTL run. However, the precipitation intensity was slightly lower in the NSH run than in the CNTL run. In the NLH run, the development of convective cells along the outer front was significantly suppressed (Figs. 4g–i). Consequently, the precipitation around Miyake Island in the NLH run was the weakest among all the experiments.

Fig. 4.

Horizontal distributions of precipitation intensity (shading; mm h⁻¹) at the surface at (a) 04:00 UTC, (b) 06:00 UTC, and (c) 08:00 UTC on 8 January 2017 in the CNTL run. The open blue circle with a 20-km radius shows Miyake Island. (d)–(f) is the same as (a)–(c), except for the NSH run. (g)–(i) is the same as (a)–(c), except for the NLH run. Maximum and area-averaged values of precipitation intensity within the blue circle with a 20-km radius in Fig. 4 at 06:00 UTC on 8 January 2017 are noted in panels (b), (e), and (h).

We quantitatively compared the precipitation intensities near Miyake Island among the three runs by estimating the maximum and area-averaged values of precipitation intensities within the blue circle (20-km radius) (Fig. 4) at 06:00 UTC on 8 January 2017. The obtained values are shown in Figs. 4b, 4e, and 4h. The maximum precipitation intensity values in the NSH and NLH runs were smaller than those in the CNTL run by 22.3 mm h⁻¹ and 35.8 mm h⁻¹, respectively. Furthermore, the average precipitation intensity values in the NSH and NLH runs were smaller than those of the CNTL run by 0.5 mm h⁻¹ and 4.0 mm h⁻¹, respectively. Thus, the contributions of the sensible and latent heat fluxes to the maximum precipitation intensity values are 42 % and 67 %, respectively, whereas the contributions to the average precipitation intensity values are 6 % and 48 %, respectively. These comparisons indicate that the heat fluxes, especially the latent heat fluxes, are critical for intensifying the outer frontal rainband.

Next, we examined the mesoscale structure of the outer frontal rainband (Fig. 5). Figures 5a–c show the spatial distribution of the moisture flux at 980 hPa and its horizontal convergence around Miyake Island at 06:00 UTC on 8 January 2017 in the CNTL, NSH, and NLH runs. In all the runs, moisture flux convergence occurred near the surface along the outer frontal rainband (Fig. 4). Among all the runs, the convergence around the south of Miyake Island was the strongest in the CNTL run. The convergence in the NSH run was slightly weaker than that in the CNTL run, and the convergence in the NLH run was considerably weaker than in others.

Fig. 5.

Horizontal distributions of moisture flux (arrows; kg m kg⁻¹ s⁻¹) and its horizontal convergence (shading; × 10⁻⁵ kg kg⁻¹ s⁻¹) at 980 hPa at 06:00 UTC on 8 January 2017 in the (a) CNTL, (b) NSH, and (c) NLH runs. (d)–(f) Latitude–height cross-sectional maps of the meridional–vertical moisture fluxes (arrows; kg m kg⁻¹ s⁻¹), the horizontal moisture flux convergence (blue contours; × 10⁻⁵ kg kg⁻¹ s⁻¹), and latent heating rate induced by condensation (shading; K h⁻¹) along the red lines in the panels (a–c). The interval of the blue contours is 0.5 × 10⁻⁵ kg kg⁻¹ s⁻¹. (g)–(i) is the same as (d)–(f), except for the total mixing ratio of cloud water and cloud ice (contour; g kg⁻¹) and the total mixing ratio of rain water, snow, and graupel (shading; g kg⁻¹).

Figures 5d–f show the vertical cross-section of the meridional-vertical moisture fluxes, horizontal moisture flux convergence, and latent heating rate related to condensation along the red lines shown in Figs. 5a–c. The lines cross the precipitation maxima nearest Miyake Island in each run (Figs. 4b, e, h). In the CNTL run, the strongest moisture flux convergence occurred below the 850-hPa level between 33.95°N and 34°N (Fig. 5d). The northward moisture fluxes were stronger around the convergence area than the southward fluxes, indicating that the moisture to the south of the front was the main source of rain. Over this convergence area, the moisture condensed, and latent heat was released. An updraft attending the latent heating was also observed. Because updrafts can enhance the near-surface moisture convergence and further condensation, such a mesoscale structure could intensify precipitation. Compared with the CNTL run, the moisture flux convergence, condensation heating, and attendant updraft were all suppressed in the NSH run (Fig. 5e). Among the runs, these three components were the weakest in the NLH run (Fig. 5f). The differences in the mesoscale features of the three runs (Fig. 5) correspond to the differences observed in the precipitation intensity of each run (Fig. 4). Therefore, it is conceivable that the heat fluxes, especially the latent heat fluxes, modified the mesoscale structures associated with the outer frontal rainband, enhancing rainfall.

We also investigated the vertical distribution of hydrometeors (Figs. 5g–i). Figures 5g–i are the same as Figs. 5d–f, except for the total mixing ratio of cloud water and cloud ice (cloud hydrometeors) and the total mixing ratio of rain water, snow, and graupel (precipitating hydrometeors). In the CNTL run, clouds formed around 34°N near the surface to 650 hPa (Fig. 5g). The condensation of vapor around the front (Fig. 5d) contributed to the formation of the clouds. Moreover, higher values of the mixing ratio of precipitating hydrometeors (≥ 1.6 g kg⁻¹) were found above and below the core of the cloud (Fig. 5g). Furthermore, the values of the mixing ratio of the precipitating hydrometeors were high near the surface (≥ 2.0 g kg⁻¹), corresponding to the heavy rainfall at the surface. The mesoscale structure around the front (Fig. 5d) facilitated the formation of these maxima in the precipitation hydrometeors. In the NSH run, the height of the clouds was lower (Fig. 5h) than that in the CNTL run (Fig. 5g). The formation of the precipitating hydrometeors was suppressed in the NSH run compared with that in the CNTL run. Among the three runs, the mixing ratio values of the cloud and precipitating hydrometeors were the lowest in the NLH run among the three runs (Fig. 5i). These differences in hydrometeors corroborate the findings that the heat fluxes, especially the latent heat fluxes, are critical for enhancing the outer frontal rainband.

4. Surface heat fluxes and their impact on environmental fields of the outer frontal rainband

Because of the importance of surface heat fluxes found above, we further examined the factors affecting the increase in surface heat fluxes around the Kuroshio Current and their impact on the environmental fields of the outer frontal rainband. First, we explored the roles of SST distribution around the Kuroshio Current and CCB in the intensification of surface heat fluxes in the CNTL run. An analysis of the environmental conditions in the CNTL run and comparisons among the CNTL, NSH, and NLH runs are presented in Sections 4.2 and 4.3, respectively.

4.1 Analysis of the increase in surface heat fluxes

Because the impact of latent heat fluxes on the outer frontal rainband was higher than that of the sensible heat fluxes (Section 3), we focus on the latent heat fluxes in this section. Latent heat fluxes, LHF, are calculated in the model using a bulk formula as follows LHF=ρ_aL_vC_h|V_a|[Qvs (SST) – Qv_a], where, |V_a| is the wind speed, Qvs (SST) is the saturation water vapor mixing ratio of SST, Qv_a is the water vapor mixing ratio, ρ_a is the density, L_v is the latent heat of water evaporation, and C_h is the bulk coefficient of water vapor. The subscript a indicates the bottom layer of the atmosphere in the CReSS model. To understand the factor of the increase in latent heat fluxes, Figs. 6a–f show the horizontal distributions of latent heat fluxes, SST, |V_a|, [Qvs (SST) – Qv_a], Qvs (SST), and Qv_a at 06:00 UTC 8 January 2017 in the CNTL, along with features of a frontal structure and circulation fields.

Fig. 6.

Horizontal distributions of (a) latent heat fluxes (shading; W m⁻²), (b) SST (shading; °C), (c) wind speed in the bottom layer |V_a|, (d) the differences between saturation water vapor mixing ratio of SST Qvs (SST) and water vapor mixing ratio in the bottom layer Qv_a (shading; g kg⁻¹), (e) Qvs (SST) (shading; g kg⁻¹), and (f) Qv_a (shading; g kg⁻¹) at 06:00 UTC on 8 January 2017 in the CNTL run. A 2D frontogenesis function at 980 hPa (blue contours; 10⁻⁴ K km⁻¹ s⁻¹), horizontal wind in the bottom layer (arrows; m s⁻¹), and SLP (black contours; hPa) are also shown in these figures. The intervals of the blue and black contours are 1 × 10⁻⁴ K km⁻¹ s⁻¹ and 4 hPa, respectively. The closed circle shows Miyake Island.

First, we examine the relationship between the latent heat fluxes (Fig. 6a) and SST (Fig. 6b). Latent heat fluxes exceeding 350 W m⁻² were observed over a wide area around the south of the outer front between 30°N and 34°N (Fig. 6a). In particular, maxima of latent heat fluxes exceeding 450 W m⁻² were found immediately to the east of the cyclone center, between 137°E and 140°E (Fig. 6a). Warm SST associated with the Kuroshio Current intruded zonally at around 33°N (Fig. 6b). Moreover, a warm SST area, extending from the southwest to the northeast existed to the east of the cyclone center, between 137°E and 141°E (Fig. 6b). With its axis around 33°N, the Kuroshio Current transported warm water broadly around the region, forming this warm SST area associated with oceanic mesoscale eddy structures (Fig. 2). The maxima of the latent heat fluxes appeared around this warm SST area (Figs. 6a, b).

Next, we investigated the location of the CCB. A zonally extended frontogenesis area was observed immediately to the north of the cyclone center around 32°N (Fig. 6), corresponding to the warm front. As mentioned in Sections 1–3, the outer front existed to the north of the warm front. Near-surface southeasterly winds prevailed between the warm and outer fronts (Fig. 6c). The Qv_a (Fig. 6f) around the southeasterly winds were lower than those to the south of the warm front, i.e., the warm sector of the cyclone. Because CCB is on the northern side of a warm front and is characterized by dry air (e.g., Carlson 1980; Schultz 2001), the atmospheric features around the southeasterly winds between the two fronts corresponded well to the features of CCB.

Therefore, we examined the influence of SST distribution and CCB on latent heat fluxes. The high SST around the Kuroshio Current created the high Qvs (SST) (Fig. 6e). The overlap of the high Qvs (SST) with the low Qv_a of CCB (Fig. 6f) increased the difference between these two variables, [Qvs (SST) – Qv_a] (Fig. 6d). Moreover, |V_a| around the CCB near the cyclone center was particularly strong (Fig. 6c). The increased [Qvs (SST) - Qv_a] (Fig. 6d) and strong |V_a| (Fig. 6c) produced the maxima of latent heat fluxes to the east of the cyclone center (Fig. 6a), indicating that the maxima of heat fluxes were caused by the Kuroshio Current and the CCB. The latent heat fluxes to the south of Miyake Island and the outer front also increased because of the combined influence of the Kuroshio Current (Figs. 6b, e) and the CCB (Figs. 6c, f). Figure 5 shows that the atmospheric conditions to the south of the front were critical for developing the outer frontal rainband. Thus, the heat fluxes from the Kuroshio Current to the south of the outer front are crucial for forming the environmental conditions that favored the intensification of the outer frontal rainband, which are discussed in detail in the following subsections.

4.2 Analysis of environmental fields of the rainband in the CNTL run

To examine the atmospheric conditions around the outer front in the CNTL run, Figs. 7a–c show the horizontal distributions of the water vapor mixing ratio (Qv) at 980 hPa, equivalent potential temperature (θ_e) at 980 hPa, and convective instability in the lower troposphere defined by the difference in θ_e between 850- and 980-hPa levels at the same time as that shown in Figs. 5 and 6 (06:00 UTC on 8 January 2017). These figures show a 2D frontogenesis function at 980 hPa, horizontal wind vectors at 980 hPa, and SLP. The Qv at 980-hPa was higher to the south of the outer front (> 8 g kg⁻¹) than to the north (Fig. 7a). The 980-hPa θ_e was also relatively high around the outer front (> 310 K; Fig. 7b), partially influenced by the Qv distribution. Corresponding to the relatively high θ_e area, a convectively unstable area was found in the lower troposphere (Fig. 7c). Around the south of the outer front, the southeasterly winds associated with CCB prevailed at 980 hPa. These findings indicate that CCB flow also influences transporting convectively unstable and moist air into the outer front. Moreover, the release of convective instability and associated condensation occurred over the front, enhancing the convective rainband.

Fig. 7.

(a) Horizontal distributions of water vapor mixing ratio Qv at 980 hPa (shading; g kg⁻¹), 2D frontogenesis function at 980 hPa (blue contours; 10⁻⁴ K km⁻¹ s⁻¹), horizontal wind at 980 hPa (arrows; m s⁻¹), and SLP (black contours; hPa) at 06:00 UTC on 8 January 2017 in the CNTL run. The intervals of the blue and black contours are 1 × 10⁻⁴ K km⁻¹ s⁻¹ and 4 hPa, respectively. The black line indicates the position of the vertical cross-section shown in Fig. 8. (b) The same as (a), except for the equivalent potential temperature θ_e at 980 hPa (shading; K). (c) Same as (a), except for the convective instability defined by the difference in θ_e between 850- and 980-hPa levels (shading; K).

To better understand the relationship between the environmental conditions and outer frontal rainband, Fig. 8 shows a vertical cross-section of the θ_e, 2D frontogenesis function, and the meridional–vertical wind vectors along the black line shown in Figs. 7a–c. The line crosses the grid point at the highest precipitation intensity in the convective cell to the south of Miyake Island (Fig. 4b). The red dots in this figure indicate the moist area where the Qv exceeds 8 g kg⁻¹. The 2D frontogenesis associated with the outer front was evident between 33.9°N and 34.0°N. To the south of the frontogenesis area, a convectively unstable layer between the 980- and 850-hPa levels and the moist area nearest to the surface was observed, consistent with Figs. 7a–c. Moreover, the southerly winds on the southern side of the outer front, which might correspond to CCB flow, convey the convectively unstable and moist layer to the outer front. Warmer air on the southern side of the outer front was lifted around the front. Thus, it is conceivable that the frontal updraft lifted the unstable and moist layer to the south of the front, releasing the instability and the attendant moisture condensation. Note that the relatively high near-surface Qv to the south of the outer front not only increased convective instability but it also became an important source of rain over the outer front, correlating with the results shown in Fig. 5d. These findings indicate that, in addition to the horizontal frontogenesis and the southerly flow of CCB, the convectively unstable and moist conditions to the south of the outer front are crucial environmental factors affecting the formation of the mesoscale structure (moisture convergence, condensation heating, updraft, and hydrometeors) around the outer front (Figs. 5a, d, g) and the associated intensification of the rainband (Fig. 4b).

Fig. 8.

Latitude–height cross-sectional map of θ_e (shading; K), positive 2D frontogenesis function (blue contours; 10⁻⁴ K km⁻¹ s⁻¹), and meridional and vertical winds (arrows; m s⁻¹) at 06:00 UTC on 8 January 2017 in the CNTL run along the black line in Fig. 7. The blue contour interval is 0.5 × 10⁻⁴ K km⁻¹ s⁻¹. The red dots indicate the moist area where the Qv exceeds 8 g kg⁻¹.

4.3 Comparisons of environmental fields of the rainband among the three runs

Because the surface sensible and latent heat fluxes from the Kuroshio Current increased significantly to the south of the outer front around the CCB (Fig. 3), they might contribute to the increases observed in the Qv and convective instability. To confirm this hypothesis, Figs. 9a–c show the differences in the 980-hPa Qv, the 980-hPa θ_e , and convective instability between the NSH and CNTL runs at the same time as that shown in Figs. 5–8. Figures 9d–f are the same as Figs. 9a–c, except for the differences between the NLH and CNTL runs. Horizontal wind vectors at 980 hPa and SLP in the NSH and NLH runs are shown in Figs. 9a–c and 9d–f, respectively.

Fig. 9.

Differences in (a) Qv (shading; g kg⁻¹) at 980 hPa, (b) θ_e (shading; K) at 980 hPa, and (c) convective instability defined by the difference in θ_e between 850- and 980-hPa levels (shading; K) between the NSH and CNTL runs (the former minus the latter) at 06:00 UTC on 8 January 2017. Vectors and contours are 980-hPa horizontal winds and SLP in the NSH run, respectively. The closed circle shows Miyake Island. (d)–(f) is the same as (a)–(c), except for the differences between the NLH and CNTL runs. Vectors and contours in these figures indicate 980-hPa horizontal winds and SLP in the NLH run. The red dots in (a) show where the relative humidity exceeds 99 % in the CNTL run.

The difference in the Qv between the NSH and CNTL runs was insignificant around the CCB (Fig. 9a). Hirata et al. (2018) proposed that the sensible heat supply from the ocean can increase Qv by increasing the saturation water vapor mixing ratio if the near-surface air is saturated. The red dots in Fig. 9a show where the relative humidity exceeds 99 % at 980 hPa in the CNTL run. The air around CCB was unsaturated. Thus, the lack of response of the Qv to the sensible heat fluxes is because of the unsaturated atmospheric conditions in this case. However, the Qv in the NLH run decreased significantly (∼ 1 g kg⁻¹) around the southern side of the outer front (Fig. 9d).

The 980-hPa θ_e around CCB was lower in the NSH run than in the CNTL run by ∼ 0.5 K (Fig. 9b). Further, differences in the 980-hPa θ_e between the NLH and CNTL runs exceeded 3.5 K (Fig. 9e). Although the convective instability around CCB also decreased in both sensitivity experiments, the instability was suppressed more in the NLH run compared with the NSH run (Figs. 9c, f), which might reflect the differences in the near-surface θ_e between the two runs (Figs. 9b, e).

We further examined the vertical profile of Qv (Fig. 10a) and θ_e (Fig. 10b) on the upstream side of Miyake Island, indicated by the red rectangle in Fig. 9. The Qv profiles were similar between the NSH and CNTL runs (Fig. 10a), consistent with the results shown in Fig. 9a. The difference in the Qv between the NLH and CNTL runs was the largest nearest the surface and it decreased with height (Fig. 10a). Above the 920-hPa level, the differences were hardly noticeable. The differences in the θ_e among the three runs were also most evident near the surface (Fig. 10b). The θ_e difference values corroborated those shown in Figs. 9b and 9e. Unlike the differences in the Qv (Fig. 10a), the differences in the θ_e between the CNTL and NLH runs were observed, even above the 920-hPa level, but these values were relatively small (∼ 1 K). This could be because of the differences in the condensation heating associated with convection (Figs. 5d, f). These θ_e profiles indicate that the θ_e differences near the surface (below ∼ 950 hPa) contributed most to the differences in convective instability among the three runs. The observation that the latent heat fluxes from the Kuroshio Current were larger than the sensible heat fluxes (Fig. 3) might explain why the response of the θ_e and convective instability to the latent heat fluxes is larger than to the sensible heat fluxes.

Fig. 10.

Vertical profiles of area-averaged values of (a) Qv (g kg⁻¹) and (b) θ_e (K) within the red rectangle depicted in Fig. 9. The red, green, and blue curves indicate the profiles in the CNTL, NSH, and NLH runs, respectively.

5. Summary

This study examined the role of the surface heat fluxes from the Kuroshio Current in enhancing the outer frontal rainband that brought heavy rainfall to Miyake Island, Japan, in January 2017 by using three cloud-resolving numerical experiments (CNTL, NSH, and NLH runs) with a horizontal resolution of 0.02°. As the rainband developed, the sensible and latent heat fluxes from the Kuroshio Current increased around CCB to the south of the outer front in the CNTL run (Figs. 2, 3, 6). The magnitude of the latent heat fluxes was ∼ 2.3 times larger than that of the sensible heat fluxes. The CNTL run successfully simulated the features of the outer frontal rainband (Figs. 1, 4). However, slight and significant weakening of the outer frontal rainband were observed in the NSH and NLH runs, respectively (Fig. 4). Further, we showed that the differences in the mesoscale features (horizontal moisture convergence, condensation, updraft, and hydrometeors) around the outer front in the three runs corresponded to the differences in the outer frontal rainband intensity (Fig. 5). Analysis of the CNTL run showed that, in addition to the horizontal frontogenesis and southwesterly flow of CCB, convective unstable and moist conditions near the surface enhanced the rainband (Figs. 7, 8). The frontal uplifting released convective instability, facilitating the formation of the mesoscale features and associated intensification of the rainband. The comparisons between the NSH and CNTL runs showed that the sensible heat fluxes slightly enhanced convective instability (Figs. 9a–c, 10). Moreover, the comparison between the NLH and CNTL runs demonstrated that the latent heat fluxes significantly increased the moisture content and convective instability (Figs. 9d–f, 10). The effects of heat fluxes can enhance the outer frontal rainband. Therefore, we conclude that the heat fluxes from the Kuroshio Current, especially the latent heat fluxes, contributed to heavy rainfall associated with the extratropical cyclone by enhancing the convective rainband and by modifying the frontal structure of the cyclone described in H19.

The findings of this study are significant for future studies. Previous studies have typically focused on the role of the warm conveyor belt during heavy precipitation associated with extratropical cyclones (e.g., Pfahl et al. 2014; Catto et al. 2015). However, the findings presented here show that CCB contributes to intensifying surface heat fluxes from the midlatitude ocean and to the transportation of modified air to precipitation areas (Figs. 3, 6–8 in this paper; Hirata et al. 2015, 2016, 2018). We emphasize the role of CCB and how studies on CCB could help us to better understand the relationship between the midlatitude ocean and heavy precipitation associated with extratropical cyclones. Moreover, this study demonstrated that the surface heat fluxes from the midlatitude ocean can enhance heavy rainfall associated with extratropical cyclones. The magnitude of SST values is a critical factor in determining the magnitude of surface heat fluxes (Fig. 6). A recent study demonstrated that the increase in SST values because of global warming is larger around Japan than in other ocean areas (Wu et al. 2012). Moreover, SST distributions around Japan vary significantly because of oceanic phenomena, such as the Kuroshio large meander (e.g., Kawabe 1995) and mesoscale warm eddies (Sugimoto and Hanawa 2011). Such variability of SST around Japan could markedly affect heavy precipitation associated with extratropical cyclones (Nakamura et al. 2012; Hayasaki et al. 2013) and further studies are required to gain insights into this research question.

Acknowledgments

The authors thank the anonymous reviewer, Dr. Atsuyoshi Manda and Dr. Masayuki Kawashima for their helpful comments. The authors thank Dr. Yasumasa Miyazawa for providing JCOPE2 data. The CReSS model was developed by the Institute for Space–Earth Environmental Research (ISEE), Nagoya University (https://www.isee.nagoya-u.ac.jp/en/co-re.html). The JMA radar, GSM, and MSM data are available from the website of the Research Institute for Sustainable Humanosphere, Kyoto University (http://database.rish.kyoto-u.ac.jp/index-e.html). JCOPE2 data were provided by the Japan Agency for Marine-Earth Science and Technology (JAMSTEC) (https://www.jamstec.go.jp/jcope/htdocs/e/distribution/index.html). This study was supported in part by JSPS KAKENHI Grant Numbers JP20H05170, JP19H05696, and JP19H05701, and the Ministry of Education, Culture, Sports, Science and Technology of Japan under the framework of the TOUGOU Program. The numerical experiments in this study were conducted on the Earth Simulator of JAMSTEC. The figures in this article were prepared using the Grid Analysis and Display System (http://cola.gmu.edu/grads/). We used the MG tool developed by Masaya Kato to perform some of the analyses in this study.

References

•  Carlson,  T. N., 1980: Airflow through midlatitude cyclones and the comma cloud pattern. Mon. Wea. Rev., 108, 1498-1509.
•  Catto,  J. L., 2016: Extratropical cyclone classification and its use in climate studies. Rev. Geophys., 54, 486-520.
•  Catto,  J. L.,  E.  Madonna,  H.  Joos,  I.  Rudeva, and  I.  Simmonds, 2015: Global relationship between fronts and warm conveyor belts and the impact on extreme precipitation. J. Climate, 28, 8411-8429.
•  Gyakum,  J. R., and  R. E.  Danielson, 2000: Analysis of meteorological precursors to ordinary and explosive cyclogenesis in the western North Pacific. Mon. Wea. Rev., 128, 851-863.
•  Hayasaki,  M.,  R.  Kawamura,  M.  Mori, and  M.  Watanabe, 2013: Response of extratropical cyclone activity to the Kuroshio large meander in northern winter. Geophys. Res. Lett., 40, 2851-2855.
•  Hirata,  H.,  R.  Kawamura,  M.  Kato, and  T.  Shinoda, 2015: Influential role of moisture supply from the Kuroshio/Kuroshio Extension in the rapid development of an extratropical cyclone. Mon. Wea. Rev., 143, 4126-4144.
•  Hirata,  H.,  R.  Kawamura,  M.  Kato, and  T.  Shinoda, 2016: Response of rapidly developing extratropical cyclones to sea surface temperature variations over the western Kuroshio–Oyashio confluence region. J. Geophys. Res.: Atmos., 121, 3843-3858.
•  Hirata,  H.,  R.  Kawamura,  M.  Kato, and  T.  Shinoda, 2018: A positive feedback process related to the rapid development of an extratropical cyclone over the Kuroshio/Kuroshio Extension. Mon. Wea. Rev., 146, 417-433.
•  Hirata,  H.,  R.  Kawamura,  M. K.  Yoshioka,  M.  Nonaka, and  K.  Tsuboki, 2019: Key role of the Kuroshio Current in the formation of frontal structure of an extratropical cyclone associated with heavy precipitation. J. Geophys. Res.: Atmos., 124, 6143-6156.
•  Iizuka,  S., and  H.  Nakamura, 2019: Sensitivity of midlatitude heavy precipitation to SST: A case study in the Sea of Japan area on 9 August 2013. J. Geophys. Res.: Atmos., 124, 4365-4381.
• Japan Meteorological Agency, 2013: Outline of the operational numerical weather prediction at the Japan Meteorological Agency. Appendix to WMO Technical Progress Report on the Global Data-processing and Forecasting System (GDPFS) and Numerical Weather Prediction (NWP) Research. 201 pp. [Available at: http://www.jma.go.jp/jma/jma-eng/jma-center/nwp/outline2013-nwp/pdf/outline2013_all.pdf.]
•  Kawabe,  M., 1995: Variation of current path, velocity, and volume transport of the Kuroshio in relation with the large meander. J. Phys. Oceanogr., 25, 3103-3117.
•  Kuwano-Yoshida,  A., and  S.  Minobe, 2017: Storm-track response to SST fronts in the Northwestern Pacific region in an AGCM. J. Climate, 30, 1081-1102.
•  Manda,  A.,  H.  Nakamura,  N.  Asano,  S.  Iizuka,  T.  Miyama,  Q.  Moteki,  M. K.  Yoshioka,  K.  Nishii, and  T.  Miyasaka, 2014: Impacts of a warming marginal sea on torrential rainfall organized under the Asian summer monsoon. Sci. Rep., 4, 5741, doi:10.1038/srep05741.
•  Martin,  J. E., 2006: Mid-latitude Atmospheric Dynamics: A First Course. John Wiley & Sons Ltd, 324 pp.
•  Miyama,  T.,  M.  Nonaka,  H.  Nakamura, and  A.  Kuwano-Yoshida, 2012: A striking early-summer event of a convective rainband persistent along the warm Kuroshio in the East China Sea. Tellus A, 64, 18962, doi: 10.3402/tellusa.v64i0.18962.
•  Miyazawa,  Y.,  R.  Zhang,  X.  Guo,  H.  Tamura,  D.  Ambe,  J.-S.  Lee,  A.  Okuno,  H.  Yoshinari,  T.  Setou, and  K.  Komatsu, 2009: Water mass variability in the western North Pacific detected in a 15-year eddy resolving ocean reanalysis. J. Oceanogr., 65, 737-756.
•  Nakamura,  H.,  A.  Nishina, and  S.  Minobe, 2012: Response of storm tracks to bimodal Kuroshio path states south of Japan. J. Climate, 25, 7772-7779.
•  Nuss,  W. A., and  S. I.  Kamikawa, 1990: Dynamics and boundary layer processes in two Asian cyclones. Mon. Wea. Rev., 118, 755-771.
•  Pfahl,  S.,  E.  Madonna,  M.  Boettcher,  H.  Joos, and  H.  Wernli, 2014: Warm conveyor belts in the ERA-Interim dataset (1979–2010). Part II: Moisture origin and relevance for precipitation. J. Climate, 27, 27-40.
•  Schultz,  D. M., 2001: Reexamining the cold conveyor belt. Mon. Wea. Rev., 129, 2205-2225.
•  Sekizawa,  S.,  T.  Miyasaka,  H.  Nakamura,  A.  Shimpo,  K.  Takemura, and  S.  Maeda, 2019: Anomalous moisture transport and oceanic evaporation during a torrential rainfall event over western Japan in early July 2018. SOLA, 15A, 25-30.
•  Sugimoto,  S., and  K.  Hanawa, 2011: Roles of SST anomalies on the wintertime turbulent heat fluxes in the Kuroshio-Oyashio confluence region: Influences of warm eddies detached from the Kuroshio Extension. J. Climate, 24, 6551-6561.
•  Tsuboki,  K., and  A.  Sakakibara, 2002: Large-scale parallel computing of Cloud Resolving Storm Simulator. High Performance Computing.  Zima,  H. P.,  K.  Joe,  M.  Sato,  Y.  Seo, and  M.  Shimasaki (eds.), Springer, 243-259.
•  Wu,  L.,  W.  Cai,  L.  Zhang,  H.  Nakamura,  A.  Timmermann,  T.  Joyce,  M. J.  McPhaden,  M.  Alexander,  B.  Qiu,  M.  Visbeck,  P.  Chang, and  B.  Giese, 2012: Enhanced warming over the global subtropical western boundary currents. Nat. Climate Change, 2, 161-166.
•  Yamamoto,  M., 2020: Ensemble simulations of the influence of regionally warm sea surface on moisture and rainfall in Tsushima Strait during August 2013. Atmos. Res., 238, 104876, doi:10.1016/j.atmosres.2020.104876.