Abstract

Mesoscale structures of the wintertime marine atmospheric boundary layer (MABL) as climatological imprints of oceanic fronts within the Kuroshio–Oyashio Extension (KOE) region east of Japan are investigated by taking advantage of high horizontal resolution of the ERA-Interim global atmospheric reanalysis data, for which the resolution of sea surface temperature (SST) data has been improved. These imprints, including locally enhanced sensible and latent heat fluxes and local maxima in cloudiness and precipitation in association with locally strengthened surface-wind convergence in the vicinities of SST fronts along the warm Kuroshio Extension and cool Oyashio to its north, are also identified in high-resolution satellite data. In addition to these mesoscale MABL features, meridionally confined near-surface baroclinic zones and zonally oriented sea level pressure (SLP) minima associated with the dual SST fronts are represented in ERA-Interim only in the period of high-resolution SST, but those imprints of the Oyashio front are missing in the low-resolution SST period. In the presence of the prevailing monsoonal northerlies, latitudinal displacements of the SLP trough, baroclinic zone, and the peak meridional gradient of the turbulent heat fluxes from each of the corresponding SST fronts are also found to be sensitive to the frontal width that depends on the SST resolution. The analysis herein suggests that the converging surface northerlies into the SLP minima can contribute positively to the formation of a surface baroclinic zone along the Kuroshio Extension, while a stronger baroclinic zone along the Oyashio front is maintained primarily through the pronounced cross-frontal contrast in sensible heat release from the ocean.

1. Introduction

In the Kuroshio–Oyashio Extension (KOE) region east of Japan, the warm Kuroshio Extension (KE) and the cool Oyashio are confluent to yield a pronounced meridional gradient in sea surface temperature (SST) climatologically (Yasuda 2003). The region is thus also called the subarctic frontal zone (SAFZ), whose meridional displacement causes prominent decadal SST anomalies (Nakamura and Kazmin 2003; Frankignoul et al. 2011; Taguchi et al. 2012). In the KOE region, turbulent sensible heat flux (SHF) and latent heat flux (LHF) from the ocean are enhanced in the cold season, because of large air–sea difference in temperature and humidity under dry, cold continental air advected by the prevailing monsoonal northerlies (e.g., Taguchi et al. 2009; Kwon et al. 2010). Despite a huge amount of heat release into the atmosphere, SST remains relatively high in winter along the KE owing to its advective effect.

It has been argued that surface baroclinicity is restored through sharp meridional contrast of SHF across oceanic fronts (Nakamura et al. 2008; Nonaka et al. 2009; Hotta and Nakamura 2011). Taguchi et al. (2009) showed that the restoration mechanism is indeed operative in the KOE region, contributing to the recurrent development of synoptic-scale disturbance and thereby the formation of the Pacific storm track (Nakamura et al. 2004).

Recent studies have revealed that locally enhanced turbulent heat fluxes along the western boundary currents can organize mesoscale structures in marine atmospheric boundary layer (MABL) that cannot form only through atmospheric processes (e.g., Small et al. 2008; Kelly et al. 2010). Investigating long-term ship-measured climatology, Tanimoto et al. (2011) found that enhanced SHF and LHF along the KE warm the overlying MABL locally, forming a trough of sea level pressure (SLP) in winter and thereby influencing the distribution of surface geostrophic wind through the “hydrostatic effect” (Lindzen and Nigam 1987). The formation of the SLP trough also acts to force frictional wind convergence near the surface and associated updraft at the MABL top. At the same time, static stability is also reduced within the overlying MABL, where the “vertical mixing effect” (Wallace et al. 1989; Hayes et al. 1989), thus enhanced, translates a larger amount of westerly momentum down from the free troposphere to modify the ageostrophic wind field. Experiments by Koseki and Watanabe (2010) with an atmospheric general circulation model (AGCM) suggest that these two mechanisms can be operative comparably over the KE in January (cf. Shimada and Minobe 2011). Samelson et al. (2006) argued that the positive correlations between SST and surface wind stress away from the immediate vicinity of oceanic fronts may be attributable to the deeper MABL over the warmer SST.

Recent high-resolution satellite observations and numerical experiments have suggested mesoscale influences of SST on clouds and precipitation systems. Tokinaga et al. (2009) found that surface wind convergence associated with the pressure trough can locally enhance cloud formation on the warmer flank of the KE front in winter based on satellite measurements. Similar mechanisms can also be operative over the Gulf Stream (Minobe et al. 2008). Furthermore, Iizuka (2010) showed numerically that high-resolution SST acts to augment interannual variance of precipitation around the KE.

The aforementioned mesoscale oceanic imprints on MABL over the KOE region are identified as climatological features. For investigation of their interannual variations, ship-observed data are not suited because of their sparseness in both time and space. For that purpose atmospheric reanalysis data are more suited, with their additional advantage of providing three-dimensional distribution of atmospheric variables. Unlike in the free troposphere, however, the state of MABL represented by atmospheric reanalysis data is not strongly constrained by observed atmospheric data, and it can therefore be sensitive to the prescribed SST, especially around the oceanic frontal zones where SST gradient is tight. Even if the horizontal resolution of the atmospheric model used for a reanalysis is sufficiently high, those oceanic imprints cannot be represented in the reanalysis unless fine structures are resolved in the SST field prescribed.

To investigate the sensitivity of MABL structure in atmospheric reanalysis to prescribed SST, we take advantage of characteristics of the ERA-Interim global atmospheric reanalysis (Dee et al. 2011) produced by the European Centre for Medium-Range Weather Forecasts (ECMWF), in which resolution of prescribed SST has been improved twice. Our investigation targets the KOE region, or the North Pacific SAFZ, which is not merely a single frontal zone but rather characterized by a pair of SST fronts (Yasuda 2003; Nonaka et al. 2006; Seo et al. 2014). In the KOE region, part of the Oyashio flows eastward to the north of the KE, forming the mixed water region in between, whereas the other part flows southward along the east coast of Japan to form extremely tight SST gradient on the northern flank of the KE. In comparison with high-resolution satellite observations, we investigate how the representation of the mesoscale imprints of these SST fronts on MABL in ERA-Interim is modified as the resolution of the SST field has been improved. Indeed, Chelton (2005) showed that mesoscale wind stress fields in the eastern tropical Pacific represented in the ECMWF model are sensitive to the SST data prescribed as the lower boundary condition. We focus on winter months (December–March), when turbulent heat fluxes are the strongest in the year.

The rest of this paper is organized as follows. The datasets used in the present study are introduced in section 2. In sections 3 and 4 we investigate distributions of SST, air temperature, SLP, and turbulent heat fluxes within the KOE region based on the ERA-Interim data. A comparison with satellite observations is made in section 5. In section 6, the possible influence of the SLP minimum on the formation of a baroclinic zone is discussed. A summary and discussion are given in section 7.

2. Data

a. ERA-Interim

In this study we mainly use the ERA-Interim global atmospheric reanalysis from January 1979 to March 2012, available on a 0.75° × 0.75° latitude–longitude grid (Dee et al. 2011). The horizontal resolution of the atmospheric model used for ERA-Interim is at a TL255 spectral truncation (equivalent to ~79-km grid intervals, corresponding to 0.75° grid intervals) with 60 hybrid vertical levels (14 levels below the 800-hPa level). We use monthly fields of three-dimensional atmospheric motion, sea level pressure, geopotential height, temperature, and humidity, available at the pressure levels at intervals of 25 hPa below the 750-hPa level. In addition to precipitation, cloud amount and column liquid/ice water, boundary layer height is also used, which is defined as the altitude at which the bulk Richardson number first exceeds 0.25 (ECMWF 2007). For surface air temperature (SAT) and surface wind, we use air temperature and wind velocities at the lowest model level, respectively, rather than 2-m temperature and 10-m wind components, which depend strongly on MABL parameterization schemes. Turbulent heat fluxes from the ocean as output from the forecasting system are also used.

Although the resolution of the atmospheric model is fixed throughout the data period, the resolution of SST data prescribed at the lower boundary of the model has been improved twice. Specifically, the resolution is 1.0° × 1.0° in January 1979 through December 2001, 0.5° × 0.5° in January 2002 through January 2009, and 0.05° × 0.05° since February 2009. We have confirmed that the latter improvement exerts virtually no impacts on the atmospheric fields of the reanalysis. This is because the SST resolution in the latest period is too fine to be assigned for the TL255 atmospheric model, and therefore much finer structures resolved in the original SST field since 2009 cannot be resolved in the atmospheric model. In contrast, the former improvement does exert substantial impacts especially on the MABL, as discussed in detail in the following sections, since the SST data whose original resolution is 1.0° cannot represent mesoscale features even if rearranged onto the grid system with 0.75° intervals. We thus refer to the period from January 1979 to December 2001 as the “low-resolution SST” period (LR period) and the period since January 2002 as the “high-resolution SST” period (HR period).

b. Satellite data

The SST fields assigned for the ERA-Interim data are compared with the Advanced Very High Resolution Radiometer (AVHRR) Pathfinder optimally interpolated SST (OISST) data produced by the National Oceanic and Atmospheric Administration (NOAA) (Reynolds et al. 2007). The monthly satellite SST data are available since November 1981 on a 0.25° × 0.25° grid. The resolution is enough to distinguish the Oyashio front from the KE front (Figs. 1a–c). We have confirmed that the climatological distribution of equatorward SST gradient based on the OISST data is very similar to that based on the Advanced Microwave Scanning Radiometer for the Earth Observing System (AMSR-E) data for their overlapping period from December 2002 to March 2011.

Fig. 1.

(a) Climatological-mean wintertime (December–March) distribution of the SCOW surface winds (m s−1; arrows) for the period from December 1999 to March 2009, superimposed on OISST (contoured for every 1°C) and J-OFURO2 surface sensible and latent heat fluxes combined (SHF + LHF; W m−2; shaded as indicated by the color bar at the bottom) for the period from January 1988 to December 2008. Also shown are the corresponding latitudinal profiles of (b) OISST [for 149.375°E; black line (bottom x axis)] and SHF + LHF [for 149.5°E; red line (top x axis)] and of (c) equatorward gradients of OISST [°C (100 km)−1; black line (bottom x axis)] and SHF + LHF [W m−2 (100 km)−1; red line (top x axis)]. (d)–(f) As in (a)–(c), but for ERA-Interim for the HR period, with latitudinal profiles at 149.25°E. (g)–(i) As in (d)–(f), respectively, but for ERA-Interim for the LR period.

Fig. 1.

(a) Climatological-mean wintertime (December–March) distribution of the SCOW surface winds (m s−1; arrows) for the period from December 1999 to March 2009, superimposed on OISST (contoured for every 1°C) and J-OFURO2 surface sensible and latent heat fluxes combined (SHF + LHF; W m−2; shaded as indicated by the color bar at the bottom) for the period from January 1988 to December 2008. Also shown are the corresponding latitudinal profiles of (b) OISST [for 149.375°E; black line (bottom x axis)] and SHF + LHF [for 149.5°E; red line (top x axis)] and of (c) equatorward gradients of OISST [°C (100 km)−1; black line (bottom x axis)] and SHF + LHF [W m−2 (100 km)−1; red line (top x axis)]. (d)–(f) As in (a)–(c), but for ERA-Interim for the HR period, with latitudinal profiles at 149.25°E. (g)–(i) As in (d)–(f), respectively, but for ERA-Interim for the LR period.

Since MABL structures in atmospheric reanalysis data may strongly depend on parameterization schemes as well as prescribed SST in data assimilation, the following multiple satellite datasets are used for validating the atmospheric reanalysis data. For validating the mean sea surface wind field, the Scatterometer Climatology of Ocean Winds (SCOW; http://cioss.coas.oregonstate.edu/scow/) is used, where monthly climatologies of surface wind vectors and their divergence/convergence are compiled on a 0.25° × 0.25° grid through QuikSCAT measurements over the 11-yr period from September 1999 to October 2009 (Risien and Chelton 2008). The Moderate Resolution Imaging Spectroradiometer (MODIS) collection 051 level 3 products are employed for validating monthly-mean cloud liquid water path, cloud ice water path, cloud-top pressure, and cloud fraction. The MODIS instrument has been on board National Aeronautics and Space Administration (NASA) satellites Terra and Aqua. The MODIS product on a 1.0° × 1.0° grid is available for the period from July 2002 to August 2012. Two kinds of cloud fraction, called “cloud mask cloud fraction” and “cloud optical properties cloud fraction,” are available as the MODIS products (Hubanks et al. 2008). The former cloud fraction is larger than the latter by approximately 15%, although the horizontal distributions of their wintertime climatology are mutually similar around the KOE region (not shown). In the present study, we focus only on the horizontal distribution of cloud fraction based on the cloud optical properties cloud fraction, whose values are closer to those of the corresponding variable based on ERA-Interim. We also employ the AMSR-E precipitation product. The data are available on a 0.25° × 0.25° grid from June 2002 to September 2011.

In addition, the Japanese Ocean Flux Datasets with Use of Remote Sensing Observations version 2 (J-OFURO2) (Kubota and Tomita 2007) are used for the validation of monthly fields of SHF and LHF. These fluxes have been evaluated by applying the bulk formulas derived by Fairall et al. (2003) to high-resolution SST and marine meteorological databased mainly on satellite and in situ measurements and in part on atmospheric reanalysis data. The particular flux data are available on a 1.0° × 1.0° grid from January 1988 to December 2008.

3. SST and surface baroclinicity within the KOE region

a. Wintertime climatologies of SST, SHF, LHF, and surface wind

Figure 1 shows the wintertime (December–March) climatological-mean fields of SST, surface turbulent heat fluxes (SHF and LHF combined), and surface wind, in addition to latitudinal profiles of SST, the heat fluxes, and their equatorward gradients, based on the satellite observations (Figs. 1a–c) and the ERA-Interim data in the HR period (Figs. 1d–f) and LR period (Figs. 1g–i). Both in the satellite observations and reanalysis, the monsoonal northwesterly winds prevail over the KOE region, inducing strong upward heat fluxes, especially along the warm KE. Both in the LR and HR periods, heat fluxes in ERA-Interim are slightly less than in the J-OFURO2 estimation, especially off the southeastern coast of Japan.

As highlighted in the meridional profile in Fig. 1c (black line), the OISST data indicate dual local maxima of equatorward SST gradient (dSSTdy), one corresponding to the Oyashio front (around 40°N) and the other to the KE front (around 36°N), and its local minimum in between that corresponds to the Kuroshio–Oyashio mixed water region. The J-OFURO2 heat fluxes are locally enhanced on the warmer flanks of these SST fronts (red line in Fig. 1b). Correspondingly, equatorward gradients of the heat fluxes also exhibit dual maxima that coincide with these SST fronts (red line in Fig. 1c). These features are insensitive to the period for averaging. Specifically, similar distributions can be obtained as the wintertime climatologies of the OISST and J-OFURO2 data for the period from December 1999 to December 2008 (i.e., virtually the same period as the averaging period for the SCOW data). The J-OFURO2 data are consistent with objectively analyzed turbulent heat fluxes based on the OAFlux data (Yu and Weller 2007) (not shown).

Despite some slight differences in averaging periods, ERA-Interim in the HR period (Figs. 1d–f) well reproduces the mesoscale structures in the field of surface heat fluxes associated with the dual SST fronts as captured by the satellite observations (Figs. 1a–c). In the LR period, by contrast, those mesoscale distributions in the surface heat fluxes are missing in ERA-Interim, owing to the smoothed SST distribution assigned (Figs. 1g–i). In this period, equatorward gradients of heat fluxes and SST exhibit a broad single maximum within the KOE region. Indeed, differences in the wintertime SST climatology between the HR and LR periods (HR minus LR) within the KOE region (Fig. 2) are characterized by warm anomalies on the southern flanks of the Oyashio and KE fronts around 39° and 34°N, respectively, and by cold anomalies to their north, consistent with dual peaks in dSSTdy in the HR period. The zonal band of cold SST anomalies centered at 37°N and slight local weakening in surface winds from the LR period into the HR period may contribute to the slight reduction in the heat fluxes into the HR period (Figs. 1d–f). Qualitatively the same distributions are obtained if the corresponding climatologies are constructed separately for SHF and LHF (not shown). Specifically, SHF exhibits more distinct dual-peak profiles in the satellite observations and HR period of ERA-Interim, whereas LHF is twice as large as SHF along the KE owing to higher SST. In summary, Fig. 1 demonstrates the presence of mesoscale features of the surface heat fluxes reflecting the dual SST fronts within the KOE region, showing unambiguous sensitivity of their representation in the reanalysis data to the resolution of SST data assigned.

Fig. 2.

(a) Difference of the wintertime climatology of SST fields based on ERA-Interim between its high- and low-resolution SST periods. (b) The corresponding latitudinal profile at 149.25°E. Gray horizontal lines in (b) indicate the peaks of dSSTdy in the high-resolution SST period.

Fig. 2.

(a) Difference of the wintertime climatology of SST fields based on ERA-Interim between its high- and low-resolution SST periods. (b) The corresponding latitudinal profile at 149.25°E. Gray horizontal lines in (b) indicate the peaks of dSSTdy in the high-resolution SST period.

b. Meridional profiles of SST, SAT, and SHF based on ERA-Interim

In this section, meridional profiles of equatorward gradients of SST (dSSTdy), SAT (dSATdy), and SHF (dSHFdy) are compared in detail between the LR and HR periods of ERA-Interim. We consider dSATdy as a measure of surface baroclinicity. We focus on a particular longitudinal sector (146.25°–150°E), in which the Oyashio and KE fronts are clearly separated (Figs. 3c,d). Since the Oyashio front is tilting from southwest to northeast, we avoided longitudinal averaging not to introduce any artificial smoothing to mesoscale structures.

Fig. 3.

Climatological-mean wintertime (December–March) distributions of (a) dSSTdy [°C (100 km)−1; shaded as indicated by color bar at the bottom], dSATdy [°C (100 km)−1; black contours at intervals of 0.1 from 1.3] and dSHFdy [W m−2 (100 km)−1; red contours at intervals of 4 from 0] based on the ERA-Interim LR period. (b) The corresponding latitudinal profiles at 149.25°E of dSSTdy (blue line at blue labels on bottom x axis), dSATdy (black line and black labels on bottom x axis), and dSHFdy (red line and red labels on top x axis). (c),(d) As in (a),(b), but for the HR period.

Fig. 3.

Climatological-mean wintertime (December–March) distributions of (a) dSSTdy [°C (100 km)−1; shaded as indicated by color bar at the bottom], dSATdy [°C (100 km)−1; black contours at intervals of 0.1 from 1.3] and dSHFdy [W m−2 (100 km)−1; red contours at intervals of 4 from 0] based on the ERA-Interim LR period. (b) The corresponding latitudinal profiles at 149.25°E of dSSTdy (blue line at blue labels on bottom x axis), dSATdy (black line and black labels on bottom x axis), and dSHFdy (red line and red labels on top x axis). (c),(d) As in (a),(b), but for the HR period.

Figure 3 compares the wintertime ERA-Interim climatologies of dSSTdy, dSATdy, and dSHFdy within the KOE region between the LR (Figs. 3a,b) and HR periods (Figs. 3c,d). Light low-pass filtering with three-point running mean in the meridional direction has been applied to dSATdy in order to remove two-grid noises (not shown). For the sake of consistency, the same meridional low-pass filtering has been applied also to dSSTdy and dSHFdy. Despite this meridional smoothing, both dSSTdy and dSHFdy in the HR period exhibit distinct dual maxima associated with the Oyashio and KE fronts (Figs. 3c,d). In the LR period (Figs. 3a,b), by contrast, dSSTdy and dSHFdy each exhibit only a single broad maximum, which is weaker than the primary maximum along the Oyashio front in the HR period. In both periods, dSATdy peaks are weaker than their corresponding dSSTdy peaks, which is presumably due to the relaxing effect by atmospheric disturbances on dSATdy. This leads to enhanced heat supply from the ocean surface to the overlying atmosphere on the warmer flanks of the SST fronts, which can locally modify the thermal structure of MABL. These relationships can be confirmed through a simple thermodynamic argument as in section 3d and  appendix C.

As evident in latitudinal–time sections of dSSTdy, dSATdy, and dSHFdy for 149.25°E (Figs. 4a–c), their latitudinal distributions in January based on the ERA-Interim data change suddenly in 2002. In most of the years since 2002 during the HR period, dSHFdy and dSSTdy each exhibit a secondary peak near the KE front in addition to the primary peak near the Oyashio front, although the peak along the KE front is less robust and even missing in some of the years. In the LR period before 2002, by contrast, the dSSTdy peak associated with the KE front is totally missing, and the KOE region is represented as a broad frontal zone whose axis resides near the Oyashio front. In the LR period, dSSTdy is weaker than in the HR period by approximately 30% (Figs. 4b and 4c). Correspondingly, dSATdy and dSHFdy each exhibit a single maximum, which tends to be weaker than in the HR period. The aforementioned contrasts between the HR and LR periods are evident not only in January (Figs. 4a–c) but also for other winter months (December, February, and March; not shown).

Fig. 4.

(a) Latitude–time (year) section for 149.25°E of dSSTdy [°C (100 km)−1; shaded as indicated by color bar below (a)], dSATdy [°C (100 km)−1; black contours at intervals of 0.3 from 1.2 and thickened for 1.2], and dSHFdy [W m−2 (100 km)−1; red contours at intervals of 12 and thickened for 0] based on ERA-Interim for January. Yellow arrow indicates the timing of the improvement of SST resolution in January 2002. (b) The corresponding meridional profiles of dSSTdy (blue lines and blue labels on bottom x axis), dSATdy (black lines and black labels on bottom x axis), and dSHFdy (red lines and red labels on top x axis) averaged over the LR period. (c) As in (b), but for the HR period. (d) As in (a), but for SHF + LHF [W m−2; shaded as indicated by color bar below (d)] and meridional hSLP (hPa; black contours at 0.1 intervals and thickened for 0; only nonpositive values are plotted). (e),(f) As in (b),(c), but for SHF + LHF (red lines and red labels on top x axis labels) and hSLP (black lines, bottom x axis). Error bars in (e) and (f) indicate ±1 standard errors of hSLP at each latitudinal grid. The standard errors are estimated as RMS errors divided by the square root of the degrees of freedom (N − 1), where N denotes the number of years for a given period.

Fig. 4.

(a) Latitude–time (year) section for 149.25°E of dSSTdy [°C (100 km)−1; shaded as indicated by color bar below (a)], dSATdy [°C (100 km)−1; black contours at intervals of 0.3 from 1.2 and thickened for 1.2], and dSHFdy [W m−2 (100 km)−1; red contours at intervals of 12 and thickened for 0] based on ERA-Interim for January. Yellow arrow indicates the timing of the improvement of SST resolution in January 2002. (b) The corresponding meridional profiles of dSSTdy (blue lines and blue labels on bottom x axis), dSATdy (black lines and black labels on bottom x axis), and dSHFdy (red lines and red labels on top x axis) averaged over the LR period. (c) As in (b), but for the HR period. (d) As in (a), but for SHF + LHF [W m−2; shaded as indicated by color bar below (d)] and meridional hSLP (hPa; black contours at 0.1 intervals and thickened for 0; only nonpositive values are plotted). (e),(f) As in (b),(c), but for SHF + LHF (red lines and red labels on top x axis labels) and hSLP (black lines, bottom x axis). Error bars in (e) and (f) indicate ±1 standard errors of hSLP at each latitudinal grid. The standard errors are estimated as RMS errors divided by the square root of the degrees of freedom (N − 1), where N denotes the number of years for a given period.

c. Latitudinal relationship between local maxima of dSSTdy, dSATdy, and dSHFdy

As evident in Fig. 4a, the peak latitudes of dSATdy and dSHFdy tend to wobble within the KOE region during the LR period, despite the fact that the corresponding latitudinal wobbling of the dSSTdy peak is much less. In the HR period, by contrast, both the dSATdy and dSHFdy peaks tend to be anchored in the immediate vicinity of the dSSTdy peak. The latitudinal displacement of the local maxima of monthly dSATdy and dSHFdy from the nearest dSSTdy peak is quantified in the histograms shown in Fig. 5, which have been constructed for 149.25°E in a manner as described in  appendix A.

Fig. 5.

(a) Probability density function of frequency (%; abscissa) for wintertime (December–March) monthly Δlat (intervals of 0.75°; ordinate; negative values for southward displacements) of a local maximum of dSATdy from the corresponding maximum of dSSTdy based on the ERA-Interim data at 149.25°E. Black and gray lines correspond to the LR and HR periods, respectively. The numbers of the local maxima sampled for the individual periods are indicated in the upper portion of the panel. (b) As in (a), but for monthly latitudinal displacements of a local peak of dSHFdy from the corresponding peak of dSSTdy.

Fig. 5.

(a) Probability density function of frequency (%; abscissa) for wintertime (December–March) monthly Δlat (intervals of 0.75°; ordinate; negative values for southward displacements) of a local maximum of dSATdy from the corresponding maximum of dSSTdy based on the ERA-Interim data at 149.25°E. Black and gray lines correspond to the LR and HR periods, respectively. The numbers of the local maxima sampled for the individual periods are indicated in the upper portion of the panel. (b) As in (a), but for monthly latitudinal displacements of a local peak of dSHFdy from the corresponding peak of dSSTdy.

As shown in Sampe et al. (2010) and Hotta and Nakamura (2011), SST fronts with strong dSSTdy, including the KE and Oyashio fronts, act to anchor surface baroclinic zones (identified as dSATdy peaks) through cross-frontal differential sensible heating (dSHFdy). Throughout the analysis period, dSHFdy indeed tends to peak at SST fronts that are identified as dSSTdy peaks (Fig. 5b). However, the prevailing monsoonal northerlies act to advect the baroclinic zones downstream. In fact, local dSATdy maxima tend to be located slightly southward (i.e., downwind) of the corresponding local maxima of dSSTdy and dSHFdy in the LR period (black line in Fig. 5a). These latitudinal relationships seem consistent with previous numerical studies on spatial relationships among mesoscale anomalies of SST, SHF, virtual potential temperature, and surface wind stress associated with tropical instability waves (Small et al. 2003) and meandering of the Agulhas Return Current (O’Neill et al. 2010). In the HR period, by contrast, local maxima of both dSATdy and dSHFdy tend to coincide with the corresponding maxima of dSSTdy at the SST fronts. Into the histograms in Fig. 5, the peaks around both the KE and Oyashio fronts are incorporated. The same tendency as seen in Fig. 5 can be captured in the same histograms but constructed separately for the KE and Oyashio fronts (not shown). The construction has been made under the assumption that the two fronts are separated at 38.5°N, where dSSTdy reaches its climatological minimum. These characteristics are consistent among all of the six meridians from 146.25° to 150°E at which the ERA-Interim data are available (not shown).

To highlight the differences in latitudinal relationship among the peaks between the LR and HR periods in a more quantitative manner, we statistically verified that the displacements tend to be smaller in the HR period than in the LR period through the chi-squared test with one degree of freedom. Specifically, cases in which the displacements are small (i.e., between −0.75° and 0.75°) are more often observed in the HR period than in the LR period at the 99.9% confidence level for all of the six meridians within the KOE region.

Interestingly, Fig. 5b indicates that there are more cases when dSHFdy peaks slightly north (i.e., upwind) of the corresponding dSSTdy peak than the opposing cases, in spite of the advective effect of the monsoonal northerlies. This tendency is obvious in the LR period, and it is hinted also in the HR period especially in the vicinity of the KE front just east of Japan (Fig. 3c). This counterintuitive displacement is discussed in the next subsection.

d. Interpretation of the relationship among the dSSTdy, dSATdy, and dSHFdy peaks

In this section, an attempt is made to quantify the advective effect of the monsoonal flow and elucidate the factors contributing to the difference in the latitudinal displacement of dSATdy peaks from the corresponding dSSTdy peaks in ERA-Interim between the LR and HR periods. In our attempt, we examine the solution of the thermodynamic equation under an idealized condition that mimics the wintertime KOE region. We assume that SHF uniformly warms the overlying MABL, in which air density ρ is assumed to be constant. The heating rate of MABL by SHF can thus be represented as SHF/(ρCpH), where Cp denotes specific heat of dry air at constant pressure and H the depth of the atmospheric mixed layer (Nonaka et al. 2009), which corresponds to the MABL depth. With the aerodynamic bulk formula for SHF, the heating rate can be written as

 
formula

where CH denotes the heat transfer coefficient and W = (u2 + υ2)1/2 is the surface wind speed, with u and υ representing the zonal and meridional wind velocities, respectively. We consider a situation where SHF is positive (i.e., upward) where SST > SAT, as typically observed in winter over the KOE region. For simplicity, both SST and SAT are assumed to be zonally uniform, and thermal advection is thus yielded only by the meridional wind velocity. Although idealized, this setting is nevertheless useful for investigating thermal conditions for the wintertime KOE region as a first-order approximation. In fact, heat budget analysis based on the ERA-Interim data shows that the advection of monthly-mean air temperature by monthly-mean meridional winds averaged over the MABL depth can offset about 52% of the SHF heating, if the MABL top is defined for a given month as the level at which the difference in monthly-mean potential temperature from the 1000-hPa level (denoted as Δθ) reaches 1.5 K. This definition is adopted in consideration of an apparent tendency for the boundary layer height determined on the basis of the bulk Richardson number to be deeper than mixed layer into which direct thermal influence of SHF reaches (Figs. 6a,b). The climatological MABL thus defined extends up to the 880-hPa level (thick gray lines in Figs. 6a,b), and its depth H is approximately 1200 m.

Fig. 6.

(a) Meridional cross section of the climatological-mean wintertime (December–March) potential temperature (thin gray contours at every 1 K) and the MABL top (thin black line) at 149.25°E, from ERA-Interim in its LR period. Thick black (gray) line indicates the MABL top defined as the level at which monthly potential temperature is higher than at the 1000-hPa level by 0.5 K (1.5 K). (b) As in (a), but for the HR period. (c) As in (a), but for latitudinal profile of L (km) calculated from monthly-mean surface winds and MABL top. Definitions of the MABL top for thick black line and thick gray line are the same as thick black thick gray line in (a). (d) As in (c), but for the HR period. Thin gray vertical lines in (c) and (d) indicate the SST fronts where dSSTdy maximizes.

Fig. 6.

(a) Meridional cross section of the climatological-mean wintertime (December–March) potential temperature (thin gray contours at every 1 K) and the MABL top (thin black line) at 149.25°E, from ERA-Interim in its LR period. Thick black (gray) line indicates the MABL top defined as the level at which monthly potential temperature is higher than at the 1000-hPa level by 0.5 K (1.5 K). (b) As in (a), but for the HR period. (c) As in (a), but for latitudinal profile of L (km) calculated from monthly-mean surface winds and MABL top. Definitions of the MABL top for thick black line and thick gray line are the same as thick black thick gray line in (a). (d) As in (c), but for the HR period. Thin gray vertical lines in (c) and (d) indicate the SST fronts where dSSTdy maximizes.

In such an idealized situation as discussed above, the thermodynamic equation for the MABL may be simplified by retaining the two dominant terms:

 
formula

where the y axis is defined as being positive for southward and positive υ thus corresponds to the northerlies. We further assume that SST gradient is sinusoidal on mesoscale and it is superposed on a constant gradient Ty. That is,

 
formula

or

 
formula

where T1 signifies half the mesoscale SST difference across an oceanic front centered at y0 in association with the sinusoidal SST gradient, T2 a parameter that can be specified by setting the upstream boundary condition of the domain (e.g., at 43.5°N), and D a measure of the frontal width (Fig. 7). If υ, W, and H are assumed to be positive and independent of y, the steady-state solution of (2) for SAT can be obtained as

 
formula

or

 
formula

with

 
formula

for which SHF can be expressed from (1) as

 
formula

or

 
formula

In (9), C0 represents an integral constant, and L = (υ/W)(H/CH), whose physical meaning and interpretation are discussed in  appendix C.

Fig. 7.

Schematic normalized latitudinal profiles of the sinusoidal components of analytical solutions of SAT [thin solid; from (5)] and SHF [dashed; from (8)] for a given sinusoidal profile of SST (thick solid). SST and SHF are normalized with the magnitudes of their sinusoidal components, while SAT is normalized with the magnitude of the sinusoidal component of SST.

Fig. 7.

Schematic normalized latitudinal profiles of the sinusoidal components of analytical solutions of SAT [thin solid; from (5)] and SHF [dashed; from (8)] for a given sinusoidal profile of SST (thick solid). SST and SHF are normalized with the magnitudes of their sinusoidal components, while SAT is normalized with the magnitude of the sinusoidal component of SST.

For interpreting the analytical solutions, their sinusoidal components are discussed first. Both dSATdy and dSHFdy have a cosine component if dSSTdy does so. The southward displacement (δy) of the maximum of the cosine component of dSATdy from the dSSTdy peak is given in (7) (positive δy for southward), and that of the maximum dSHFdy from the maximum dSSTdy is given by δy* = δy − D/2.

Equation (7) indicates that both δy and δy* depend on the width of the SST front D and the parameter L, which represents the atmospheric conditions as indicated above. In the following, we assess the sensitivity of δy and δy* to D and L to give an interpretation for the latitudinal displacements of the dSATdy and dSHFdy peaks from the corresponding dSSTdy peak as described in session 3c. In reality, L is not necessarily uniform over the wintertime KOE region (Figs. 6c,d). To obtain the analytical solutions of (2), however, L must be specified as a constant parameter. As described in  appendix C, typical values of L (i.e., L = 461 km for the LR period and L = 377 km for the HR period) are used for constructing Fig. 8, in which δy and δy* are plotted as functions of D. Validity and uncertainty of the assumption of a constant L value are discussed in the next subsection.

Fig. 8.

(a) Latitudinal displacements of the peaks of dSATdy (° lat; solid lines; δy, positive for southward) and dSHFdy (° lat; dashed lines; δy* = δyD/2) from the corresponding peak of dSSTdy based on the analytical solution given in (6) and (9), respectively, as functions of D of the SST front. Black and gray lines signify the evaluations with averaged L values under the MABL depth defined with Δθ = 1.5 K for the LR (i.e., L = 461 km) and HR (i.e., L = 377 km) periods for ERA-Interim. Gray arrows on the x axis indicate ranges of parameter D assumed for the LR and HR periods. (b) As in (a), but for the corresponding phase displacement.

Fig. 8.

(a) Latitudinal displacements of the peaks of dSATdy (° lat; solid lines; δy, positive for southward) and dSHFdy (° lat; dashed lines; δy* = δyD/2) from the corresponding peak of dSSTdy based on the analytical solution given in (6) and (9), respectively, as functions of D of the SST front. Black and gray lines signify the evaluations with averaged L values under the MABL depth defined with Δθ = 1.5 K for the LR (i.e., L = 461 km) and HR (i.e., L = 377 km) periods for ERA-Interim. Gray arrows on the x axis indicate ranges of parameter D assumed for the LR and HR periods. (b) As in (a), but for the corresponding phase displacement.

As evident in Fig. 8, the solution predicts positive (southward) δy and negative (northward) δy* under the northerlies (υ > 0). This analytical model prediction is consistent with the southward and northward displacements of the peaks of dSATdy and dSHFdy, respectively, from the corresponding dSSTdy peak, as revealed in the preceding section based on the ERA-Interim data (Fig. 5), although the prediction tends to overestimate those displacements. The prevailing advective effect results in the southward displacement of the dSATdy peak, maximizing the gradient of air–sea thermal contrast (SST − SAT) and thereby dSHFdy slightly north of the dSSTdy peak.

Although the phase difference between the dSATdy and dSSTdy peaks (i.e., πδy/D) decreases with D (solid lines in Fig. 8b), the corresponding latitudinal difference δy increases (solid lines in Fig. 8a). The northward displacement of the dSHFdy peak from the corresponding dSSTdy peak increases with D, if measured as difference either in phase or in latitude (−δy*). According to Fig. 8a, our analytical model predicts that δy should change from ~1.5° to ~0.9° in latitude and −δy* from ~0.4° to ~0° in latitude in responding to the reduction in a typical D value from ~3.75° in latitude for the LR period to ~1.88° in latitude for the HR period (Table 1), as discussed in  appendix B. In contrast, virtually no difference is introduced in the theoretical solution by the difference in L between the LR and HR periods through the corresponding differences in H and surface winds (gray versus black lines in Fig. 8). We therefore conclude that the reductions in the latitudinal displacements of the dSATdy and dSHFdy peaks from the corresponding dSSTdy peak that emerge around 2002 in the ERA-Interim data can be attributed mainly to the narrowing of the KE and Oyashio fronts resulting from the improved resolution of the prescribed SST data for the reanalysis rather than from the possible slight weakening of the monsoonal northerlies.

Table 1.

Assumed maximum and minimum values of the parameters in (3) as discussed in  appendix B.

Assumed maximum and minimum values of the parameters in (3) as discussed in appendix B.
Assumed maximum and minimum values of the parameters in (3) as discussed in appendix B.

e. Applicability of the analytical model derived under idealized conditions

In the preceding subsection, we focused on the sinusoidal component of the analytical solution of the simplified thermodynamic equation. However, peak positions of dSATdy and dSHFdy derived from the analytical solutions (5) and (8) depend also on the uniform background SST gradient Ty and the integral constant C0. Furthermore, such parameters as L, D, y0, T1, T2, and Ty cannot be determined uniquely from the observations. Applicability and uncertainty of our analytical model are thus discussed in this subsection.

The integral constant C0 can be specified in determining the upstream boundary condition for SAT, for example, based on the wintertime SAT climatology at 43.5°N, 149.25°E around the northern boundary of the KOE region. In the following we assess how the uncertainties in determining L, which should implicitly include some errors arising from the assumption of uniform H, υ, and W, could affect our analytical solutions. As discussed in the preceding subsection, influence exerted on δy by the uncertainties in L seems to be relatively small, but their overall influence on the solutions may be greater through the integral constant.

For the assessment, we use the minimum L for the MABL top defined under the assumption of Δθ = 0.5 K and the corresponding maximum L defined with Δθ = 1.5 K for both of the LR and HR periods, while other parameters are prescribed as their typical values along the 149.25°E meridian. With these parameters, the latitudinal displacement between the peaks of dSATdy and dSSTdy that arises only from the sinusoidal SST component (δy) is evaluated for both of the two periods. Those δy values are shown in Table 2 for comparison with the corresponding displacements (Δlat) derived from the full formula of the solution (6). For the LR period, δy can explain as much as 85% of Δlat if evaluated with the minimum L (256 km), while only 55% of Δlat can be explained by δy if evaluated with the maximum L (534 km). The contribution from C0 to Δlat thus increases from 15% to 45% as L becomes nearly doubled in acting to enhance the southward influence of the upstream boundary with increasing the e-folding meridional scale in the solution (5). Although substantially less pronounced, qualitatively the same tendency is found for the HR period. Table 3 summarizes the differences in Δlat and δy between the LR and HR periods based on Table 2. If evaluated with the minimum L values for the two periods, the difference in δy can explain 73% of the difference in Δlat. The contribution from the δy difference to the Δlat difference is reduced to 39%, however, if evaluated with the maximum L values. Considering the uncertainties arising from our idealized theoretical framework, the changes in the width of the SST fronts D from the LR period into the HR period should be regarded as one of the main mechanisms for the corresponding changes in the latitudinal displacements of the dSATdy and dSHFdy peaks from the corresponding dSSTdy peak revealed in the ERA-Interim data.

Table 2.

Latitudinal displacement of a peak of dSATdy from the specified dSSTdy peak derived from the full expression of the analytical solution (Δlat), the corresponding displacement derived only from the sinusoidal component (δy), and their ratio (δy/Δlat), evaluated on the basis of the minimum or maximum in the range of the parameter L based on ERA-Interim for its LR and HR periods.

Latitudinal displacement of a peak of dSATdy from the specified dSSTdy peak derived from the full expression of the analytical solution (Δlat), the corresponding displacement derived only from the sinusoidal component (δy), and their ratio (δy/Δlat), evaluated on the basis of the minimum or maximum in the range of the parameter L based on ERA-Interim for its LR and HR periods.
Latitudinal displacement of a peak of dSATdy from the specified dSSTdy peak derived from the full expression of the analytical solution (Δlat), the corresponding displacement derived only from the sinusoidal component (δy), and their ratio (δy/Δlat), evaluated on the basis of the minimum or maximum in the range of the parameter L based on ERA-Interim for its LR and HR periods.
Table 3.

Differences in Δlat and δy between the LR and HR periods and their ratio, based on data in Table 2.

Differences in Δlat and δy between the LR and HR periods and their ratio, based on data in Table 2.
Differences in Δlat and δy between the LR and HR periods and their ratio, based on data in Table 2.

4. SLP minimum and turbulent heat fluxes over the KOE region in winter

a. SLP and turbulent heat fluxes based on ERA-Interim

Figure 9 shows the climatological-mean wintertime distributions of the sensible and latent heat fluxes combined from the ocean (SHF + LHF; shaded) and of SLP (green lines) based on the ERA-Interim data separately for the LR and HR periods. For each meridian, local departures of the climatological-mean SLP from its meridional nine-point running-mean values are regarded as high-pass-filtered SLP (hSLP), whose distributions are superimposed on Fig. 9 with black lines. This procedure is equivalent to 6.75° latitudinal high-pass filtering in Tanimoto et al. (2011).

Fig. 9.

As in Fig. 3, but for SHF + LHF [W m−2; shaded in (a) and (c) and red lines in (b) and (d) (top x axis)], SLP [green contours at 1 hPa intervals in (a) and (c)], and meridional hSLP [black contours at 0.1-hPa intervals in (a) and (c) and black lines in (b) and (d) (bottom x axis); dashed lines in (a) and (c) for the negative values]. Gray horizontal lines in (b) and (d) indicate the SST fronts where dSSTdy maximizes.

Fig. 9.

As in Fig. 3, but for SHF + LHF [W m−2; shaded in (a) and (c) and red lines in (b) and (d) (top x axis)], SLP [green contours at 1 hPa intervals in (a) and (c)], and meridional hSLP [black contours at 0.1-hPa intervals in (a) and (c) and black lines in (b) and (d) (bottom x axis); dashed lines in (a) and (c) for the negative values]. Gray horizontal lines in (b) and (d) indicate the SST fronts where dSSTdy maximizes.

As evident in Fig. 9, SHF + LHF forms a zonal band of well-defined maxima around 35°N in both of the periods. Although somewhat less pronounced than in the ship-measured wintertime climatology by Tanimoto et al. (2011), a SLP minimum (trough) is reproduced in the ERA-Interim data along the band of SHF + LHF maxima slightly to its south in both of the periods. In addition, Fig. 10 indicates that high-pass-filtered virtual potential temperature (hθυ) exhibits a distinct positive local maximum throughout the MABL around 34°N. This maximum coincides with the distinct hSLP minimum, located slightly south of the local maxima of high-pass-filtered SHF + LHF (hHF) and SST (hSST) along the KE. These features in the ERA-Interim data are consistent with a regional atmospheric model experiment by Tanimoto et al. (2011). The correspondence suggests the primary importance of the hydrostatic effect or pressure adjustment effect in the formation of the pressure trough along the KE. In the corresponding meridional section in the HR period for 142.5°E (not shown), along which the KE front is the strongest (Fig. 3c), the aforementioned features are about twice as strong as those along 149.25°E (Fig. 10b).

Fig. 10.

(a) (middle) Climatological-mean wintertime (December–March) meridional section at 149.25°E of pressure vertical velocity (−ω) (Pa s−1; black contours at 0.01 intervals; dashed for descent and thickened for 0), meridional hθυ (K; shaded as indicated by the color bar at the bottom), and MABL top pressure (white line), from ERA-Interim in its LR period. (top) The corresponding latitudinal profiles of hθυ [red line (left y axis)] and −ω (pressure velocity) [black line (right y axis; positive upward)] at 950 hPa. (bottom) Meridional hSLP [hPa; black line (left y axis)], hHF [W m−2; red line (red right y axis)], and hSST [°C; blue line (blue right y axis)]. (b) As in (a), but for the HR period. Gray vertical lines in the bottom panels signify the SST fronts where dSSTdy maximizes.

Fig. 10.

(a) (middle) Climatological-mean wintertime (December–March) meridional section at 149.25°E of pressure vertical velocity (−ω) (Pa s−1; black contours at 0.01 intervals; dashed for descent and thickened for 0), meridional hθυ (K; shaded as indicated by the color bar at the bottom), and MABL top pressure (white line), from ERA-Interim in its LR period. (top) The corresponding latitudinal profiles of hθυ [red line (left y axis)] and −ω (pressure velocity) [black line (right y axis; positive upward)] at 950 hPa. (bottom) Meridional hSLP [hPa; black line (left y axis)], hHF [W m−2; red line (red right y axis)], and hSST [°C; blue line (blue right y axis)]. (b) As in (a), but for the HR period. Gray vertical lines in the bottom panels signify the SST fronts where dSSTdy maximizes.

Although the aforementioned gross features are common to the two periods, the secondary minimum and maximum of hSLP and hHF, respectively, are evident around 39°N only in the HR period (Figs. 9c,d and 10b). These secondary minimum and maximum in the wintertime climatology are noticeable also in the climatological-mean fields for individual winter months in the HR period. In fact, for January (Figs. 4d,f), the northern hSLP minimum at 39°N is more distinct than the southern minimum at 34°N. Although standard errors are relatively large, we have statistically confirmed that the secondary minima are observed more frequently in the HR period than in LR period in December, January (Figs. 4e,f), and February. As shown in Fig. 10b, hθυ also exhibits its shallow secondary maximum that is collocated with the northern hSLP minimum, and secondary maxima of hHF and hSST as well. These mesoscale structures can therefore be considered as imprints of the Oyashio front around 40°N, which are resolved only in the HR period. In this period, the corresponding dual-peak profiles are noticeable in the ERA-Interim boundary layer height (thick white lines in Fig. 10b), as a manifestation of the destabilizing effect on MABL due to locally enhanced heat release from the ocean on the warmer flanks of the Oyashio and KE fronts. In the LR period (Fig. 10a), by contrast, the boundary layer height exhibits a broad single peak that is collocated with the corresponding peaks in hHF and hSST around 37°N.

b. Latitudinal relationship between hSLP and turbulent heat fluxes

Tanimoto et al. (2011) argued that the climatological-mean SLP minimum is located slightly south of the local SHF + LHF maximum because of the advective effect of the prevailing northerlies. The same displacement is found in latitude–time sections of hSLP and SHF + LHF for 149.25°E based on ERA-Interim for both the LR and HR periods (Figs. 4e,f), in a manner consistent with previous numerical model studies (Small et al. 2003; O’Neill et al. 2010). In virtually the same manner as described in  appendix A, we identified pairs of negative hSLP minima and nearby positive SHF + LHF maxima based on the monthly ERA-Interim data (Fig. 4d). The latitudinal distance between the minimum and maximum that consist of each of the pairs was recorded for constructing a histogram for both of the LR and HR periods. When two or more pairs of these extremes were identified for a given month, we selected a particular pair whose hSLP minimum is located between 32.25° and 36°N around the KE. When multiple pairs were still identified, only the southernmost pair was retained for the histograms. Even if all the pairs identified between 30° and 42°N were retained for the histograms, the conclusion derived from the statistical analysis below would nevertheless be unchanged with statistical confidence slightly enhanced.

The histogram thus constructed in Fig. 11a shows an obvious tendency for a hSLP minimum to be located slightly to the south of the corresponding SHF + LHF maximum for both the LR and HR periods, although the latitudinal displacements tend to be smaller in the HR period. The histogram in Fig. 11a indicates that cases of the small displacements (between −0.75° and 0.75°) are more frequently observed in the HR period than in the LR period in excess of the 95% confidence level for all the six meridians within the KOE region.

Fig. 11.

(a) As in Fig. 5, but for the monthly latitudinal displacement of a local minimum of meridional hSLP relative to the corresponding local maximum of the sensible and latent heat fluxes combined, based on ERA-Interim for its LR (black line) and HR periods (gray line). (b) As in (a), but based only on 30 months for the weakest northerlies in the LR period.

Fig. 11.

(a) As in Fig. 5, but for the monthly latitudinal displacement of a local minimum of meridional hSLP relative to the corresponding local maximum of the sensible and latent heat fluxes combined, based on ERA-Interim for its LR (black line) and HR periods (gray line). (b) As in (a), but based only on 30 months for the weakest northerlies in the LR period.

To assess whether a weakening tendency observed in the northerlies into the HR period exerts any influence on the reducing tendency in the displacement of hSLP minima from the corresponding SHF + LHF maxima, we limit our sampling for the LR period to those 30 pairs for which the northerly wind velocity averaged over the KOE region (32°–40°N, 146.25°–150°E) is weakest. The number of the sampled pairs thus becomes the same as that for the HR period. The histogram thus constructed (Fig. 11b) for those 30 pairs in the LR period indicates that the displacements tend to be slightly reduced relative to the original histogram for the LR period (black line in Fig. 11a), and the statistical significances is also slightly reduced. Nevertheless, the difference in the displacements for the 30 pairs between the LR and HR periods is still statistically significant, at least, at the 90% confidence level. We thus conclude that the difference in the displacement is caused primarily by the difference of the prescribed SST for ERA-Interim but not by the weakening of the northerlies into the HR period.

5. Mesoscale atmospheric structures within the KOE region

From satellite and in situ observations in winter, Tokinaga et al. (2009) showed that surface wind convergence, turbulent heat fluxes, cloud liquid water content, and cloud-top altitude are climatologically enhanced locally around the KE. In this section we investigate finer horizontal distributions of these atmospheric quantities in the presence of the dual SST fronts within the KOE region. Figure 12 shows climatological-mean wintertime fields of these atmospheric quantities for the LR period (Figs. 12a–f) and the HR period (Figs. 12g–l) of the ERA-Interim data in addition to high-resolution satellite observations (Figs. 12m–r). In the ERA-Interim data during its HR period, the distinct hSLP minimum around 34°N (also in Figs. 9 and 10) accompanies surface wind convergence (Figs. 12g,h), ascent in MABL (black contours in Fig. 10b), and a zonal band of enhanced total precipitation (Figs. 12k,l). Furthermore, the secondary hSLP minimum appears to accompany weak surface wind convergence (or weakening of the divergence) and weak ascent around 39°N, 150°E near the Oyashio front. Satellite observations also capture this mesoscale surface wind convergence near the Oyashio front in addition to a distinct zonal band of convergence along the KE (Figs. 12m,n). A band of strong divergence located on the northern flank of the Oyashio front is also evident in ERA-Interim (Fig. 12g), as confirmed by the satellite observations (Fig. 12m).

Fig. 12.

(a) Climatological-mean wintertime (December–March) distributions of wind convergence at the lowest model level (10−5 s−1; shaded as indicated by the color bar on the far right) and meridional hSLP (black contours at 0.1-hPa intervals; dashed for negative values) based on ERA-Interim in its LR periods. (b) As in (a), but for the corresponding latitudinal profiles of wind convergence (red lines; top x axis) and hSLP (black line; bottom x axis) at 149.25°E. (c) As in (a), but for SST (black contours at 2°C intervals) and TCA (%; shaded as indicated by the color bar on the far right). (d) As in (b), but for SST (thick black line; black labels on bottom x axis) and TCA (thick green line; green labels on bottom x axis). The corresponding meridional high-pass-filtered profiles of SST and TCA are superimposed with thin black line (black labels on top x axis) and thin green line (green labels on top x axis), respectively. (e) As in (a), but for TP (mm day−1; shaded as indicated by the color bar on the far right ) and the fraction CP/TP (hatched where CP/TP exceeds 42.5%). (f) As in (b), but for TP (blue line on top x axis) and CP/TP (black line; bottom x axis). (g)–(l) As in (a)–(f), but for the HR period. (m) As in (a), but for surface wind convergence based on SCOW (shaded) and SST based on OISST (black contours at 2°C intervals) in place of hSLP for the period from December 1999 to March 2009. (n) As in (m), but for the corresponding latitudinal profiles of wind convergence (red lines; top x axis) and SST (black line; bottom x axis) at 149.375°E. (o) As in (m), but for TCA based on MODIS (%; shaded) in place of wind convergence for the period from December 2002 to March 2012. (p) As in (n), but TCA at 149.5°E in place of surface wind convergence. The corresponding profiles of meridionally high-pass-filtered SST and TCA are superimposed with thin black line (black labels on top x axis) and thin green line (green labels on top x axis), respectively. (q),(r) As in (m),(n), but for TP of AMSR-E smoothed with meridional three-point running mean (mm day−1; shaded) in place of wind convergence for the period from December 2002 to March 2011. Gray horizontal lines in (b),(d),(f),(h),(j),(l),(n),(p), and (r) indicate SST fronts, at which dSSTdy maximizes.

Fig. 12.

(a) Climatological-mean wintertime (December–March) distributions of wind convergence at the lowest model level (10−5 s−1; shaded as indicated by the color bar on the far right) and meridional hSLP (black contours at 0.1-hPa intervals; dashed for negative values) based on ERA-Interim in its LR periods. (b) As in (a), but for the corresponding latitudinal profiles of wind convergence (red lines; top x axis) and hSLP (black line; bottom x axis) at 149.25°E. (c) As in (a), but for SST (black contours at 2°C intervals) and TCA (%; shaded as indicated by the color bar on the far right). (d) As in (b), but for SST (thick black line; black labels on bottom x axis) and TCA (thick green line; green labels on bottom x axis). The corresponding meridional high-pass-filtered profiles of SST and TCA are superimposed with thin black line (black labels on top x axis) and thin green line (green labels on top x axis), respectively. (e) As in (a), but for TP (mm day−1; shaded as indicated by the color bar on the far right ) and the fraction CP/TP (hatched where CP/TP exceeds 42.5%). (f) As in (b), but for TP (blue line on top x axis) and CP/TP (black line; bottom x axis). (g)–(l) As in (a)–(f), but for the HR period. (m) As in (a), but for surface wind convergence based on SCOW (shaded) and SST based on OISST (black contours at 2°C intervals) in place of hSLP for the period from December 1999 to March 2009. (n) As in (m), but for the corresponding latitudinal profiles of wind convergence (red lines; top x axis) and SST (black line; bottom x axis) at 149.375°E. (o) As in (m), but for TCA based on MODIS (%; shaded) in place of wind convergence for the period from December 2002 to March 2012. (p) As in (n), but TCA at 149.5°E in place of surface wind convergence. The corresponding profiles of meridionally high-pass-filtered SST and TCA are superimposed with thin black line (black labels on top x axis) and thin green line (green labels on top x axis), respectively. (q),(r) As in (m),(n), but for TP of AMSR-E smoothed with meridional three-point running mean (mm day−1; shaded) in place of wind convergence for the period from December 2002 to March 2011. Gray horizontal lines in (b),(d),(f),(h),(j),(l),(n),(p), and (r) indicate SST fronts, at which dSSTdy maximizes.

In addition, mesoscale patches of surface wind convergence and divergence along the Oyashio front can be identified in the satellite observations (Fig. 12m) and, though less pronounced, also in ERA-Interim in its HR period (Fig. 12g). A close inspection reveals that those patches of convergence and divergence are roughly collocated with westward and eastward gradient, respectively, of satellite-observed SST (contours in Fig. 12m). These divergence/convergence patterns near the Oyashio front are suggestive of the importance of the vertical mixing effect associated with SST gradient (Chelton et al. 2004) under the monsoonal northwesterlies, while the hydrostatic effect seems more important in shaping the larger-scale patterns of surface wind convergence/divergence within the KOE region. Although distinct surface wind convergence along the KE is reproduced also in ERA-Interim during its LR period, the mesoscale patches of surface wind convergence are virtually absent due to the lack of mesoscale distribution in the prescribed SST field (Figs. 12a,b).

As shown in Figs. 12o,p, total cloud amount (TCA) based on the MODIS observations maximizes along the KE and also over the mixed water region south of the Oyashio front. This is a manifestation of local TCA maxima over local SST maxima where surface winds are convergent, as highlighted in the meridional high-pass filtered fields of TCA and SST (thin lines in Fig. 12p). This mesoscale TCA–SST correspondence around the Oyashio front is reproduced in ERA-Interim during its HR period (Figs. 12i,j), but it is missing in the LR period (Figs. 12c,d). The influence of the Oyashio front is particularly evident in low-level clouds and noticeable in midlevel clouds (Figs. 13b,c), whereas the influence of the KE front is more distinct in midlevel clouds than in low-level clouds. This particular difference between the two SST fronts arises probably from a greater amount of evaporation from the warmer, which can lead to stronger and deeper ascent. Apparently, the influence of the SST fronts does not reach high-level clouds (Fig. 13a).

Fig. 13.

Meridional sections of the wintertime (December–March) climatologies at 149.5°E of (a) high-, (b) middle-, and (c) low-level cloudiness (%; black lines; left y axis) based on MODIS measurements and their corresponding meridionally high-pass-filtered profiles (%; gray lines; right y axis). As indicated on top of the individual panels, the classification of high-, middle-, and low-level cloudiness is based on cloud-top pressure (Ptop).

Fig. 13.

Meridional sections of the wintertime (December–March) climatologies at 149.5°E of (a) high-, (b) middle-, and (c) low-level cloudiness (%; black lines; left y axis) based on MODIS measurements and their corresponding meridionally high-pass-filtered profiles (%; gray lines; right y axis). As indicated on top of the individual panels, the classification of high-, middle-, and low-level cloudiness is based on cloud-top pressure (Ptop).

Though less pronounced, similar dual-peak profiles are also observed in the total column cloud water (TCLW + TCIW) in the wintertime climatology of the MODIS measurements and ERA-Interim during its HR period (not shown). It should be pointed out that, compared to the MODIS measurements, ERA-Interim underestimates total column cloud ice water (TCIW) along the KE by as much as 80%. Likewise, the secondary peak of high-pass-filtered total column cloud liquid water (TCLW) near the Oyashio front is much less pronounced in the ERA-Interim data.

The total precipitation (TP) and the fraction of convective precipitation (CP) in TP both based on ERA-Interim in its HR period (Figs. 12k,l) are locally enhanced along the KE and near the Oyashio front in association with locally enhanced surface wind convergence. Correspondingly, AMSR-E precipitation also exhibits local enhancement near these oceanic fronts (Figs. 12q,r), despite a light meridional smoothing applied. A close inspection reveals a tendency for MODIS-measured cloud-top pressure to lower locally, as an indication of local elevation of cloud top where surface wind convergence is measured by SCOW, for example, around 39°N, 145°E, and vice versa (not shown). The local elevation of cloud top is evident particularly for low and middle level clouds (Fig. 13). These relationships suggest that surface wind convergence associated with mesoscale SST distribution can enhance development of convective clouds, as suggested by Tokinaga et al. (2009), but even on finer horizontal scales.

6. Possible role of the SLP minimum in forming a baroclinic zone

As discussed above, the SLP minimum acts to induce the frictional convergence of surface winds. In this section we discuss a possible role of the SLP minimum in forming a baroclinic zone by augmenting local gradient of near-surface temperature, which has been considered as essential for baroclinic growth of synoptic-scale disturbances (e.g., Hoskins et al. 1985; Nakamura et al. 2004). Many studies have investigated frontogenesis processes through a frontogenesis function F (e.g., Miller 1948; Hoskins 1982; Ogura and Portis 1982), which is usually expressed as a time tendency in horizontal gradient of potential temperature in terms of divergence and deformation wind fields. In the present study, however, we slightly modify the frontogenesis function by expressing it as a time tendency of near-surface air temperature gradient. The particular expression may be derived from the thermodynamic equation for MABL on a given pressure surface as

 
formula
 
formula
 
formula
 
formula
 
formula
 
formula

In (10), T denotes air temperature and R the gas constant of air. The first and second terms in each line of (10) act to strengthen zonal and meridional gradients of T. The first [(10a)], second [(10b)], and third [(10c)] lines on the RHS of (10) are called confluence, shear, and tilting terms, respectively. The fourth [(10d)] and fifth [(10e)] lines represent contributions from adiabatic volume change and diabatic heating due to SHF, respectively. We decompose each term of (10) into contributions from monthly-mean circulation and submonthly transients. For example, the former contribution to the zonal confluence term may be expressed as

 
formula

where the overbars signify monthly-mean quantities. The contribution from submonthly transients can be obtained, for example, as

 
formula

where the first term has been evaluated from 6-hourly data. Finally, the residual terms can be estimated by assuming that

 
formula

We have confirmed that the tendency term is indeed negligible compared to the major terms.

Figure 14 shows winter-mean climatological fields of the major terms of (10) evaluated at the 975-hPa level based on ERA-Interim in its HR period, with R = 287 J K−1 kg−1, Cp = 1004 J K−1 kg−1, H = 1200 m, and ρ = 1 kg m−3. For simplicity, H and ρ are assumed to be uniform in space. Among all the terms of (10), the monthly-mean contribution to the meridional SHF term (Figs. 14a,b) is prominent owing to the cross-frontal enhancement of dSHFdy (Figs. 3c,d). Specifically, this contribution is positive along the Oyashio front to reinforce near-surface baroclinicity, and so is the case along the KE front just off the east coast of Japan. Otherwise, the SHF contribution is generally destructive (i.e., frontolysis) along the KE, especially on its southern flank (around 34°N) due to greater SHF to its north (Figs. 1d–f). Counteracting this effect, the monthly-mean contribution from the meridional confluence term (Figs. 14c,d) is prominent and constructive along the KE, owing to the near-surface northerlies converging into the hSLP minimum around 34°N. This may be an indication of a possible role of the SLP minimum in maintaining near-surface baroclinicity along the KE front. The monthly-mean contributions to the meridional volume change term (Figs. 14e,f) and the meridional tilting term (Figs. 14g,h) are also strong along the KE, but they almost cancel out one another thus yielding no significant net forcing. Small-scale wavy patterns east of Japan are probably a manifestation of the spectral truncation in representing the topography in the global atmospheric model (Milliff and Morzel 2001). The transient eddy meridional advection (Figs. 14i,j) overall acts to weaken near-surface baroclinicity as poleward heat transport by baroclinically developing eddies. The residual term is also generally destructive (Figs. 14k,l), which includes such contributions as errors arising from our estimation of the SHF contribution, the spectral truncation errors discussed above and contributions from latent heat release and radiative heating.

Fig. 14.

Climatological-mean wintertime (December–March) distributions of the dominant contributions to the frontogenesis function [K (100 km)−1 day−1; shaded as indicated by the color bar at the bottom] in (10) from (a) monthly-mean meridional gradient of SHF, (c) monthly-mean meridional confluence, (e) monthly-mean meridional volume change, (g) monthly-mean meridional tilting, (i) eddy meridional advection, and (k) a residual term. All evaluated at 975 hPa from the ERA-Interim data for its HR period. The corresponding latitudinal profiles at 149.25°E of the contributions in (a),(c),(e),(g),(i), and (k) are given in (b),(d),(f),(h),(j), and (l), respectively. Gray horizontal lines indicate the axis of the KE and Oyashio fronts at which dSSTdy maximizes.

Fig. 14.

Climatological-mean wintertime (December–March) distributions of the dominant contributions to the frontogenesis function [K (100 km)−1 day−1; shaded as indicated by the color bar at the bottom] in (10) from (a) monthly-mean meridional gradient of SHF, (c) monthly-mean meridional confluence, (e) monthly-mean meridional volume change, (g) monthly-mean meridional tilting, (i) eddy meridional advection, and (k) a residual term. All evaluated at 975 hPa from the ERA-Interim data for its HR period. The corresponding latitudinal profiles at 149.25°E of the contributions in (a),(c),(e),(g),(i), and (k) are given in (b),(d),(f),(h),(j), and (l), respectively. Gray horizontal lines indicate the axis of the KE and Oyashio fronts at which dSSTdy maximizes.

The aforementioned analysis has revealed the primary importance of the cross-frontal SHF gradient in maintaining the near-surface baroclinicity along the Oyashio front, probably through a process that may be called “oceanic baroclinic adjustment” (Sampe et al. 2010; Hotta and Nakamura 2011). Along the KE, however, the contribution from cross-frontal SHF gradient (dSHFdy) is destructive, especially on its southern flank. Just off the eastern coast of Japan the width of the KE front is rather broad (Fig. 3c). According to our analytical model in section 3d, dSHFdy peaks to the north of the frontal axis (Fig. 3c), acting to enhance dSATdy at 36.5°N (Fig. 14a). Rather, the northerlies converging into the SLP minimum contribute positively to the reinforcement of dSATdy at 34°N through their southward-decreasing cold advection to the north of the SLP minimum. At the same time, the converging northerlies weaken the advective effect acting on the baroclinic zone along the KE and thus act to reduce the displacement of the baroclinic zone from the KE front.

7. Summary

In the present study we have performed a comprehensive investigation of mesoscale structures in the wintertime MABL in the presence of fine SST distribution associated with the KE and Oyashio fronts east of Japan, to extend the findings by Tokinaga et al. (2009) and Tanimoto et al. (2011). Through analysis of the ERA-Interim atmospheric reanalysis data and high-resolution satellite data, climatological imprints of mesoscale SST distribution have been identified on such mesoscale features as near-surface atmospheric baroclinic zones marked by local SAT maxima, local minima in a high-pass-filtered SLP field and associated surface wind convergence, ascent, and high-pass-filtered MABL depth, cloudiness, cloud water amount, and (convective) precipitation. These oceanic imprints on MABL have been verified through their high sensitivity to the resolution of SST data assigned for ERA-Interim that has been improved from its LR period into the HR period.

Specifically, the structure of near-surface atmospheric baroclinic zones in the ERA-Interim data is found to be sensitive to the resolution of the prescribed SST (Fig. 3). Although the prevailing monsoonal northerlies act to yield southward displacement of the baroclinic zones from the corresponding SST fronts across which SHF gradient maximizes, the latitudinal displacement analyzed in ERA-Interim is significantly smaller in its HR period than in its LR period. Our theoretical analysis of a simplified thermodynamic equation has demonstrated an unambiguous tendency for the latitudinal displacement to decrease with the narrowing of an oceanic front. It suggests that the reduced displacement in the HR period relative to the LR period may be due primarily to the ability of the SST data prescribed for ERA-Interim in its HR period to resolve narrow multiple oceanic fronts and associated fine SHF distribution within the KOE region, whereas the SST data for the LR period represent the KOE region as a broad single frontal zone. Nevertheless, even in the HR period, the KE front just east of Japan is broad enough to yield apparent northward and southward displacements of the peaks of the SHF and SAT gradients, respectively, from the frontal axis.

The improved SST resolution is also found to influence the SLP distribution analyzed in the ERA-Interim data. Even in the LR period, the climatological wintertime pressure trough is analyzed south of the KE front, as indicated in the corresponding climatology based on in situ observations (Tanimoto et al. 2011). In the period of higher-resolution SST, meridionally high-pass-filtered SLP (hSLP) in the ERA-Interim data exhibits another minimum just south of the Oyashio front, which is missing in the LR period. The southward displacement of a particular hSLP minimum from the corresponding peak of the turbulent heat fluxes decreases significantly into the HR period. We have verified that this decrease in the displacement is not due to the slight weakening tendency in the monsoonal northerlies but rather due to the improved resolution of the prescribed SST for the ERA-Interim data.

We argue that the prominent SLP minimum along the KE may exert certain influence on the reinforcement of the nearby near-surface baroclinic zone. Even in the presence of the prevailing monsoonal northerlies, the SHF gradient across the Oyashio front contributes primarily to the maintenance of the nearby baroclinic zone against the destructive effect by synoptic-scale disturbances, as in the spring situation (Taguchi et al. 2009). For the KE front just east of Japan, in contrast, its broad width acts to enlarge the southward displacement of the baroclinic zone from the peak SHF gradient, according to our analytical model. Owing to this displacement, the SHF gradient is destructive for maintaining the baroclinic zone to its south. Instead, the converging northerlies are frontogenetic around the baroclinic zone, while acting to weaken their advective effect around the SLP minimum. It remains for future study to elucidate how the SLP minimum maintains the baroclinic zone and thereby helps the recurrent development of synoptic-scale disturbances migrating into the Pacific storm track.

Although somewhat underestimated in magnitude, mesoscale features in MABL represented in the ERA-Interim HR period are found to be overall consistent with high-resolution satellite observations. Our findings indicate that state-of-the-art high-resolution reanalysis datasets of the global atmosphere, including ERA-Interim, can reproduce mesoscale imprints on MABL of fine SST distributions associated with oceanic jets, fronts, and mesoscale eddies as long as they are resolved in the SST field assigned for the reanalysis and the spatial resolution of the atmospheric model used for the reanalysis is sufficiently high. The present study has revealed that apparent long-term changes in the MABL structure represented in the ERA-Interim data are largely an artifact of the improved SST resolution. Nevertheless, the present study demonstrates the potential usefulness of high-resolution reanalysis data such as ERA-Interim for the latest decade for studying interannual variability of the mesoscale oceanic imprints in MABL, if combined with the corresponding satellite data. For this purpose, an additional product of the new Japanese reanalysis of the global atmosphere (JRA-55; Ebita et al. 2011), for which high-resolution (0.25° × 0.25°) SST fields since 1985 are prescribed, and a high-resolution coupled atmosphere–ocean reanalysis (i.e., CFSR; Saha et al. 2010) should be useful.

Acknowledgments

We thank Dr. N. Komori for the information of the SCOW dataset. We also thank Drs. A. Kuwano-Yoshida, B. Taguchi, M. Nonaka, T. Miyama, and K. Takaya for their variable comments. The four anonymous reviewers provided useful and constructive comments that have led our paper to its substantial improvement. This study is supported in part by Japanese Ministry of Education, Culture, Sports Science and Technology (MEXT) through Grants-in-Aid for Scientific Research in Innovative Areas 2205 and 2409, and by the Japanese Ministry of Environment through the Environment Research and Technology Department Fund A1201.

APPENDIX A

Construction of the Histograms in Fig. 5

The histograms in Fig. 5 in section 3c have been constructed from monthly fields in winter (December–March) of SST, SAT, and SHF based on the ERA-Interim data, to which meridional three-point running mean has been applied before evaluating their equatorward gradient. First, we identified local dSSTdy maxima along the 149.25°E meridian between 30° and 42°N whose magnitude is greater than 1.0°C (100 km)−1. If two maxima are only 1.5° latitude apart (with only a single grid point in between), the stronger one was retained. Typically, there is a single dSSTdy peak in the LR period but dual peaks in the HR period, and the histograms have been constructed separately for those two periods. For a given dSSTdy peak, we then identified the closest positive peak in both dSATdy and dSHFdy regardless of its position to the south or north of the dSSTdy maximum. In the case where two peaks of dSATdy (or dSHFdy) were identified at the same latitudinal distance from a given dSSTdy peak, the stronger peak was retained. The peak of dSATdy (or dSHFdy) thus identified was then paired with the corresponding dSSTdy peak, and the latitudinal distance between them was recorded for the histograms in Fig. 5. If the two peaks are more than 5° in latitude apart, they were not regarded as a pair and therefore not included in the histograms. Note that the sample number of these pairs can be more than that of total months, since dual dSSTdy peaks were extracted from a particular month in the HR period.

APPENDIX B

Determining the Parameter Ranges of D, T1, y0, Ty, and T2

In substituting a given SST profile into (2) in section 3d, several parameters have to be determined. Not defined uniquely, each of these parameters has certain uncertainties. At first, the uniform background SST gradient Ty and a constant T2 in the idealized SST profile (3) were determined through linear least-squared fitting of the wintertime climatological-mean SST profile at 149.25°E based on ERA-Interim for both of the LR and HR periods. In doing so, the northern boundary of the KOE region is assumed to be at 43.5°N, since the climatological-mean SST gradient obviously weakens north of this latitude (Fig. B1). Parameters for the sinusoidal component of SST (i.e., D, T1, and y0) were then estimated from local deviations of SST from the linear least-squared fitting, as shown in Fig. B2 for the latitudinal sector 35.25°–43.5°N, in which the Oyashio front is the major SST front. According to Figs. B2c,d, the axial position of the SST front (y0), at which the sinusoidal component is zero, can be identified around 39.75°–40.5°N for the LR period and around 40.5°–41.25°N for the HR period. Amplitude (T1) can be defined as half of the difference between the maximum and minimum of the sinusoidal component. Specifically, T1 for the HR period is 1.46°C, from the maximum (+1.61°C) at 39.75°N and the minimum (−1.31°C) at 42°N (Fig. B2d). Likewise, T1 for the LR period is 0.60°C from the maximum (+0.67°C) at 37.5°N and the minimum (−0.54°C) at 42°N (Fig. B2c). At the same time, the frontal width can be defined as the latitudinal distance between the maximum and minimum of the sinusoidal component. Specifically, D is 2.25° latitude for the HR period and 4.5° latitude for the LR period. The values of the parameters are found rather sensitive to the latitudinal domain to which the least-squared fitting is applied. We have repeated the fitting and specification of those parameters for the following latitudinal domains: 36.75°–43.5°N and 37.5°–43.5°N. From those values, we define the maximum, minimum, and average values of those parameters as listed in Table 1.

Fig. B1.

Climatological-mean wintertime (December–March) latitudinal profiles of SST (°C) at 149.25°E based on the ERA-Interim data for its (a) LR and (b) HR periods. Vertical lines indicate 43.5°N.

Fig. B1.

Climatological-mean wintertime (December–March) latitudinal profiles of SST (°C) at 149.25°E based on the ERA-Interim data for its (a) LR and (b) HR periods. Vertical lines indicate 43.5°N.

Fig. B2.

(a) Climatological-mean wintertime (December–March) latitudinal profiles at 149.25°E of SST (°C; black line), its linear least-squared fitting for 35.25°–43.5°N (°C; gray line) and (c) their difference based on the ERA-Interim data for its LR period. (b),(d) As in (a),(c), but for the HR period.

Fig. B2.

(a) Climatological-mean wintertime (December–March) latitudinal profiles at 149.25°E of SST (°C; black line), its linear least-squared fitting for 35.25°–43.5°N (°C; gray line) and (c) their difference based on the ERA-Interim data for its LR period. (b),(d) As in (a),(c), but for the HR period.

APPENDIX C

Physical Meaning and Specific Values of the Parameter L

Equation (7) in section 3d indicates that the latitudinal displacement δy of the dSATdy peak from the dSSTdy peak increases with the parameter L, which is defined as L = (υ/W) (H/CH). The parameter L depends on the MABL depth H and wind direction υ/W relative to the orientation of the SST gradient but not υ itself. Strengthening of the northerly component (i.e., greater positive υ) enhances the advective effect, acting to increase δy. This effect can be counteracted by an increase in scalar wind speed (i.e., greater W) that enhances dSHFdy. This enhancement effectively reduces the departure of dSATdy from dSSTdy (i.e., dSSTdy − dSATdy), thus acting to diminish the latitudinal displacement between these peaks in a steady state. Furthermore, a tendency for a larger H to lead to larger L means less effective heating of MABL by SHF, which is assumed to warm the MABL uniformly under the constant air density assumed. In other words, L represents a contribution of advective effect, which acts to increase δy, relative to the thermal forcing by SHF, which acts to diminish δy. As evident in the last term of the solution (5), L also represents the e-folding meridional scale of the boundary effect. A larger value of L means greater advective effect relative to the thermal forcing by SHF under a thermally equilibrium state.

In constructing Fig. 8a, L for both of the LR and HR periods is defined as the wintertime climatology of monthly-mean L values, which were calculated from monthly-mean H defined with Δθ = 1.5 K and monthly-mean u and υ with a typical value of the heat transfer coefficient CH =1.3 × 10−3 used commonly for the two periods. Then, a constant L has been obtained for each of the periods as the algebraic average of the maximum and minimum of its wintertime climatology at 149.25°E within the KOE region.

A typical value of /D increases from 3.5 in the LR period to 5.9 in the HR period with greater fractional reduction in D than in L, leading to an increase in phase offset [i.e., arctan(/D)] (Fig. 8b). Furthermore, if typical values of Ty, D, and L are substituted into (4) and (6) and the integral constant term is neglected for simplicity, our analytical model predicts that the peak of dSATdy is about 0.8 times stronger than that of dSSTdy for the LR period and 0.6 times stronger for the HR period, leading to positive SST − SAT and thus enhanced upward SHF on the warmer flank of the SST front. The result is roughly consistent with ERA-Interim (Fig. 3). In the LR period, the dSATdy peak is about 0.7 times stronger than the dSSTdy peak, while the dSATdy peak is about 0.5 and 0.8 times stronger near the Oyashio and KE fronts, respectively, in the HR period.

REFERENCES

REFERENCES
Chelton
,
D. B.
,
2005
:
The impact of SST specification on ECMWF surface wind stress fields in the eastern tropical Pacific
.
J. Climate
,
18
,
530
550
, doi:.
Chelton
,
D. B.
,
M. G.
Schlax
,
M. H.
Freilich
, and
R. F.
Milliff
,
2004
:
Satellite measurements reveal persistent small-scale features in ocean winds
.
Science
,
303
,
978
983
, doi:.
Dee
,
D. P.
, and Coauthors
,
2011
:
The ERA-Interim reanalysis: Configuration and performance of the data assimilation system
.
Quart. J. Roy. Meteor. Soc.
,
137
,
553
597
, doi:.
Ebita
,
A.
, and Coauthors
,
2011
:
The Japanese 55-year Reanalysis “JRA-55”: An interim report
.
SOLA
,
7
,
149
152
, doi:.
ECMWF
,
2007
: IFS documentation Cy31r1. Tech. Rep., European Centre for Medium-Range Weather Forecasts, 950 pp. [Available online at http://old.ecmwf.int/research/ifsdocs/CY31r1/.]
Fairall
,
C. W.
,
E. F.
Bradley
,
J. E.
Hare
,
A. A.
Grachev
, and
J. B.
Edson
,
2003
:
Bulk parameterization of air–sea fluxes: Updates and verification for the COARE algorithm
.
J. Climate
,
16
,
571
591
, doi:.
Frankignoul
,
C.
,
N.
Sennechael
,
Y.-O.
Kwon
, and
M. A.
Alexander
,
2011
:
Influence of the meridional shifts of the Kuroshio and the Oyashio Extensions on the atmospheric circulation
.
J. Climate
,
24
,
762
777
, doi:.
Hayes
,
S. P.
,
M. J.
McPhaden
, and
J. M.
Wallace
,
1989
:
The influence of sea surface temperature on surface wind in the eastern equatorial Pacific: Weekly to monthly variability
.
J. Climate
,
2
,
1500
1506
, doi:.
Hoskins
,
B. J.
,
1982
:
The mathematical theory of frontogenesis
.
Annu. Rev. Fluid Mech.
,
14
,
131
151
, doi:.
Hoskins
,
B. J.
,
M. E.
McIntyre
, and
A. W.
Robertson
,
1985
:
On the use and significance of isentropic potential vorticity maps
.
Quart. J. Roy. Meteor. Soc.
,
111
,
877
946
, doi:.
Hotta
,
D.
, and
H.
Nakamura
,
2011
:
On the significance of sensible heat supply from the ocean in the maintenance of mean baroclinicity along storm tracks
.
J. Climate
,
24
,
3377
3401
, doi:.
Hubanks
,
P. A.
,
M. D.
King
,
S. A.
Platnick
, and
R. A.
Pincus
,
2008
: MODIS Atmosphere L3 Gridded Product Algorithm Theoretical Basis Document. ATBD-MOD-30, 96 pp. [Available online at http://modis-atmos.gsfc.nasa.gov/_docs/L3_ATBD_2008_12_04.pdf.]
Iizuka
,
S.
,
2010
:
Simulations of wintertime precipitation in the vicinity of Japan: Sensitivity to fine-scale distributions of SST
.
J. Geophys. Res.
,
115
,
D10107
, doi:.
Kelly
,
K. A.
,
R. J.
Small
,
R. M.
Samelson
,
B.
Qiu
,
T. M.
Joyce
,
Y.-O.
Kwon
, and
M. F.
Cronin
,
2010
:
Western boundary currents and frontal air–sea interaction: Gulf Stream and Kuroshio Extension
.
J. Climate
,
23
,
5644
5667
, doi:.
Koseki
,
S.
, and
M.
Watanabe
,
2010
:
Atmospheric boundary layer response to the mesoscale SST anomalies in the Kuroshio Extension
.
J. Climate
,
23
,
2492
2507
, doi:.
Kubota
,
M.
, and
H.
Tomita
,
2007
:
Present state of the J-OFURO air–sea interaction data product
.
Flux News
,
4
,
13
15
.
Kwon
,
Y.-O.
,
M.
Alexander
,
N.
Bond
,
C.
Frankignoul
,
H.
Nakamura
,
B.
Qiu
, and
L.
Thompson
,
2010
:
Role of the Gulf Stream and Kuroshio–Oyashio systems in large-scale atmosphere–ocean interaction: A review
.
J. Climate
,
23
,
3249
3281
, doi:.
Lindzen
,
R. S.
, and
S.
Nigam
,
1987
:
On the role of sea surface temperature gradients in forcing low-level winds and convergence in the tropics
.
J. Atmos. Sci.
,
44
,
2418
2436
, doi:.
Miller
,
J. E.
,
1948
:
On the concept of frontogenesis
.
J. Meteor.
,
5
,
169
171
, doi:.
Milliff
,
R. F.
, and
J.
Morzel
,
2001
:
The global distribution of the time-average wind-stress curl from NSCAT
.
J. Atmos. Sci.
,
58
,
109
131
, doi:.
Minobe
,
S.
,
A.
Kuwano-Yoshida
,
N.
Komori
,
S.-P.
Xie
, and
R. J.
Small
,
2008
:
Influence of the Gulf Stream on the troposphere
.
Nature
,
452
,
206
209
, doi:.
Nakamura
,
H.
, and
A. S.
Kazmin
,
2003
:
Decadal changes in the North Pacific oceanic frontal zones as revealed in ship and satellite observations
.
J. Geophys. Res.
,
108
,
3078
, doi:.
Nakamura
,
H.
,
T.
Sampe
,
Y.
Tanimoto
, and
A.
Shimpo
,
2004
: Observed associations among storm tracks, jet streams and midlatitude oceanic fronts. Earth’s Climate: The Ocean–Atmosphere Interaction, Geophys. Monogr., Vol. 147, Amer. Geophys. Union, 329–345.
Nakamura
,
H.
,
T.
Sampe
,
A.
Goto
,
W.
Ohfuchi
, and
S.-P.
Xie
,
2008
:
On the importance of midlatitude oceanic frontal zones for the mean state and dominant variability in the tropospheric circulation
.
Geophys. Res. Lett.
,
35
,
L15709
, doi:.
Nonaka
,
M.
,
H.
Nakamura
,
Y.
Tanimoto
,
T.
Kagimoto
, and
H.
Sasaki
,
2006
:
Decadal variability in the Kuroshio–Oyashio Extension simulated in an eddy-resolving OGCM
.
J. Climate
,
19
,
1970
1989
, doi:.
Nonaka
,
M.
,
H.
Nakamura
,
B.
Taguchi
,
N.
Komori
,
A.
Kuwano-Yoshida
, and
K.
Takaya
,
2009
:
Air–sea heat exchanges characteristic of a prominent midlatitude oceanic front in the South Indian Ocean as simulated in a high-resolution coupled GCM
.
J. Climate
,
22
,
6515
6535
, doi:.
Ogura
,
Y.
, and
D.
Portis
,
1982
:
Structure of the cold front observed in SESAME-AVE III and its comparison with the Hoskins–Bretherton frontogenesis model
.
J. Atmos. Sci.
,
39
,
2773
2792
, doi:.
O’Neill
,
L. W.
,
S. K.
Esbensen
,
N.
Thum
,
R. M.
Samelson
, and
D. B.
Chelton
,
2010
:
Dynamical analysis of the boundary layer and surface wind responses to mesoscale SST perturbations
.
J. Climate
,
23
,
559
581
, doi:.
Reynolds
,
R. W.
,
T. M.
Smith
,
C.
Liu
,
D. B.
Chelton
,
K. S.
Casey
, and
M. G.
Schlax
,
2007
:
Daily high-resolution-blended analyses for sea surface temperature
.
J. Climate
,
20
,
5473
5496
, doi:.
Risien
,
C. M.
, and
D. B.
Chelton
,
2008
:
A global climatology of surface wind and wind stress fields from eight years of QuikSCAT scatterometer data
.
J. Phys. Oceanogr.
,
38
,
2379
2413
, doi:.
Saha
,
S.
, and Coauthors
,
2010
:
The NCEP Climate Forecast System reanalysis
.
Bull. Amer. Meteor. Soc.
,
91
,
1015
1057
, doi:.
Samelson
,
R. M.
,
E. D.
Skyllingstad
,
D. B.
Chelton
,
S. K.
Esbensen
,
L. W.
O’Neill
, and
N.
Thum
,
2006
:
On the coupling of wind stress and sea surface temperature
.
J. Climate
,
19
,
1557
1566
, doi:.
Sampe
,
T.
,
H.
Nakamura
,
A.
Goto
, and
W.
Ohfuchi
,
2010
:
Significance of a midlatitude SST frontal zone in the formation of a storm track and an eddy-driven westerly jet
.
J. Climate
,
23
,
1793
1814
, doi:.
Seo
,
Y.
,
S.
Sugimoto
, and
K.
Hanawa
,
2014
:
Long-term variations of the Kuroshio Extension path in winter: Meridional movement and path state change
.
J. Climate
,
27
,
5929
5940
, doi:.
Shimada
,
T.
, and
S.
Minobe
,
2011
:
Global analysis of the pressure adjustment mechanism over sea surface temperature fronts using AIRS/Aqua data
.
Geophys. Res. Lett.
,
38
,
L06704
, doi:.
Small
,
R. J.
,
S.-P.
Xie
, and
Y.
Wang
,
2003
:
Numerical simulation of atmospheric response to Pacific tropical instability waves
.
J. Climate
,
16
,
3723
3741
, doi:.
Small
,
R. J.
, and Coauthors
,
2008
:
Air–sea interaction over ocean fronts and eddies
.
Dyn. Atmos. Oceans
,
45
,
274
319
, doi:.
Taguchi
,
B.
,
H.
Nakamura
,
M.
Nonaka
, and
S.-P.
Xie
,
2009
:
Influences of the Kuroshio/Oyashio Extensions on air–sea heat exchanges and storm-track activity as revealed in regional atmospheric model simulations for the 2003/4 cold season
.
J. Climate
,
22
,
6536
6560
, doi:.
Taguchi
,
B.
,
H.
Nakamura
,
M.
Nonaka
,
N.
Komori
,
A.
Kuwano-Yoshida
,
K.
Takaya
, and
A.
Goto
,
2012
:
Seasonal evolutions of atmospheric response to decadal SST anomalies in the North Pacific subarctic frontal zone: Observations and a coupled model simulation
.
J. Climate
,
25
,
111
139
, doi:.
Tanimoto
,
Y.
,
T.
Kanenari
,
H.
Tokinaga
, and
S.-P.
Xie
,
2011
:
Sea level pressure minimum along the Kuroshio and its extension
.
J. Climate
,
24
,
4419
4433
, doi:.
Tokinaga
,
H.
,
Y.
Tanimoto
,
S.-P.
Xie
,
T.
Sampe
,
H.
Tomita
, and
H.
Ichikawa
,
2009
:
Ocean frontal effects on the vertical development of clouds over the western North Pacific: In situ and satellite observations
.
J. Climate
,
22
,
4241
4260
, doi:.
Wallace
,
J. M.
,
T. P.
Mitchell
, and
C.
Deser
,
1989
:
The influence of sea surface temperature on surface wind in the eastern equatorial Pacific: Seasonal and interannual variability
.
J. Climate
,
2
,
1492
1499
, doi:.
Yasuda
,
I.
,
2003
:
Hydrographic structure and variability in the Kuroshio–Oyashio transition area
.
J. Oceanogr.
,
59
,
389
402
, doi:.
Yu
,
L.
, and
R. A.
Weller
,
2007
:
Objectively analyzed air-sea heat fluxes for the global ice-free oceans (1981–2005)
.
Bull. Amer. Meteor. Soc.
,
88
,
527
539
, doi:.

Footnotes

This article is included in the Climate Implications of Frontal Scale Air–Sea Interaction Special Collection.