Abstract

To clarify the summertime evolution of decadal sea surface temperature (SST) anomalies and related physical processes in the midlatitudes of the North Pacific, numerical solutions of a three-dimensional bulk mixed layer model are analyzed, focusing on the contribution of the net shortwave radiative forcing at the sea surface. A quantitative heat budget analysis for the ocean mixed layer relating to late-1980s decadal SST change reveals that the decadal SST anomalies decay from late spring to early summer over the entire midlatitudes of the North Pacific. This quasi-seasonal decay of the decadal SST anomalies is controlled by an anomalous local thermal damping (i.e., anomalous surface heat fluxes). From midsummer to early autumn the anomalous net shortwave radiation flux associated with a meridional shift of the storm track acts to induce strong seasonal damping of the decadal SST anomaly in the northern Kuroshio–Oyashio Extension region. In contrast, in the north of the subtropical frontal region, the net shortwave radiation flux anomaly, which results from changes in low-level stratiform cloud cover, plays a major role in seasonally enhancing the decadal SST anomaly. Consequently, the SST anomalies formed by these radiative forcings cause significant variations in the local thermal damping rate at the sea surface over the period from late summer to early autumn.

1. Introduction

The Kuroshio–Oyashio Extension (KOE) region of the midlatitude North Pacific is central to our understanding of the Pacific decadal variation since the slow dynamic and fast thermal adjustments that take place in the upper ocean of the KOE region control the dominant decadal-scale sea surface temperature (SST) variation over the extratropical North Pacific (e.g., Miller and Schneider 2000). Using a coupled general circulation model, Latif and Barnett (1994, 1996) advocated that a slow dynamic adjustment in subtropical upper-ocean circulation, caused by the first baroclinic Rossby wave mode, controlled the phase reversal of decadal SST anomalies in the midlatitudes of the North Pacific. A number of numerical and observational studies have also detected decadal SST anomalies in the KOE region due to the delayed response of the ocean over time scales of a few years to wind stress curl anomalies (e.g., Miller et al. 1998; Deser et al. 1999; Venzke et al. 2000; Xie et al. 2000; Hanawa and Kamada 2001; Seager et al. 2001; Schneider et al. 2002; Vivier et al. 2002; Luo and Yamagata 2003). In addition, it has been demonstrated that the fast thermal adjustment in the ocean mixed layer caused by wintertime heat flux anomalies at the sea surface amplifies the decadal SST anomalies in the midlatitudes of the North Pacific (Barsugli and Battisti 1998; Saravanan 1998).

While a number of studies have revealed that these oceanic responses to the large-scale changes of the atmosphere (the slow dynamic and fast thermal adjustments of the upper ocean) play a central role in the dominant decadal climate variation over the North Pacific, the possible feedbacks to the atmosphere are still poorly understood (see Fig. 6 of Miller and Schneider 2000). Some studies have suggested that the KOE region is an area where the decadal SST anomalies can control the atmospheric condition (e.g., Pierce et al. 2001; Seager et al. 2001; Schneider et al. 2002). The horizontal ocean heat transport in the KOE region is so strong that any anomalous surface heat flux forcing due to changes in atmospheric conditions must be heavily suppressed. In this case, the resulting changes in decadal SST anomalies can contribute to determining the anomalous surface heat flux (see Nonaka and Xie 2003; Tanimoto et al. 2003; Xie 2004). Nevertheless, there is no hard evidence to indicate that the feedback in the KOE region is strong enough to destabilize the extratropical atmosphere. While the atmosphere–ocean coupling in the extratropics is generally thought to be strongest in winter (e.g., Deser and Blackmon 1993; Nakamura et al. 1997), the previous studies have often obtained unrealistic atmospheric responses to the wintertime decadal SST anomalies in the extratropics. For example, the wintertime response of atmospheric general circulation models to extratropical SST anomalies is too small in amplitude (e.g., Lau and Nath 1994; Bladé 1997). In addition, an atmospheric response to SST anomalies in winter has been demonstrated in the formation of a high surface air pressure anomaly to the east of the positive SST anomaly (e.g., Palmer and Sun 1985; Lau and Nath 1990; Wallace et al. 1990; Ferranti et al. 1994), while the decadal surface air pressure anomalies are observed to be shifted to the north of the decadal SST anomalies (e.g., Trenberth and Hurrell 1994; Graham 1994; Kachi and Nitta 1997; Nakamura et al. 1997). Peng et al. (1995, 1997) pointed out that the atmospheric response to an SST anomaly around the KOE region depends largely on monthly differences in the atmospheric conditions (i.e., background conditions). These results suggest that the investigation of the seasonal modulation of decadal SST anomaly is an important step toward better understanding the possible feedbacks to the atmosphere. Accurate modeling and detailed analyses are required for a specific season when the decadal SST anomaly is strong enough to control the atmospheric condition, rather than for winter. The seasonal-scale modulation of decadal SST anomalies and relating physical processes should be further examined in the midlatitudes of the North Pacific.

Some previous studies have reported a strong seasonal-scale modulation of decadal SST anomalies over the KOE region, where the decadal SST anomalies are observed to be much stronger in winter and weaker in summer (e.g., Nakamura and Yamagata 1999). The local reemergence process (Alexander and Deser 1995; Alexander et al. 1999; Alexander et al. 2001; Hanawa and Sugimoto 2004; Sugimoto and Hanawa 2005) may play a major role in the seasonal enhancement of the decadal SST anomaly in autumn (e.g., Schneider et al. 2002). The decadal SST anomaly formed in the previous winter remains beneath the ocean mixed layer during the summer and entrains into the ocean mixed layer in autumn and the following winter. Mochizuki and Kida (2006, hereafter MK06) have examined the seasonal modulation of the decadal SST anomalies in the KOE region (defined as 35°–45°N, 150°E–180°), and the related seasonal heat budget of the ocean mixed layer. They revealed that the seasonal-scale enhancement of the decadal SST anomaly was controlled by horizontal Ekman transport of heat in early winter and by vertical entrainment in autumn. Although they also noted that local thermal damping dominated in the summertime decay of decadal SST anomalies, the main physical process controlling the summertime evolution of the decadal SST anomaly still remained a controversial issue. As the main focus of MK06 was directed to the seasonal enhancement of decadal SST anomalies during autumn–winter, when the contribution of net shortwave radiation flux was negligible, they ignored the contribution from decadal variations in the net shortwave radiation flux (and the downward longwave radiation flux) at the sea surface. Recent studies have suggested that the interannual variations in the net shortwave radiation flux at the sea surface associated with stratiform cloud cover variations are correlated well with those in the SST over the midlatitude of the North Pacific (e.g., Frankignoul 1985; Klein and Hartmann 1993; Norris and Leovy 1994; Zhang et al. 1998; Alexander et al. 2004). Over the region where predominantly stratocumulus clouds are observed, for example, the suppressed downward solar radiation flux resulting from enhanced cloud cover can act to cool the upper ocean. Park and Leovy (2004) pointed out a close relationship between the SST anomaly and the anomalous low-level cloud cover, particularly associated with El Niño–Southern Oscillation. Norris et al. (1998) and Norris (2000) also reported the close relationship between cloud cover and SST on interannual to interdecadal time scales. Nevertheless, these studies did not focus on the quantitative impact of cloud radiative forcing on the seasonal heat budget of the ocean mixed layer, particularly with respect to decadal SST anomalies.

In line with these advances, to clarify the physical processes controlling the summertime evolution of the decadal SST anomalies in the midlatitudes of the North Pacific, we evaluate the relative importance of the vertical and horizontal temperature transports, the latent and sensible heat fluxes, and the cloud radiative forcing at the sea surface to the seasonal heat budget of the ocean mixed layer. To do this, we integrate a three-dimensional (3D) bulk mixed layer model developed in MK06 with atmospheric properties at the sea surface as boundary conditions and investigate the results of the numerical solutions. We focus particularly on the influence of variability in the net shortwave radiation flux on the summertime evolution of decadal SST anomalies in the midlatitudes of the North Pacific. It should be noted that our seasonal heat budget analyses clarify the physical processes responsible for the seasonal-scale modulation of the decadal SST anomaly in summer (e.g., amplification of decadal SST anomaly and/or possible feedbacks to the atmosphere) rather than that controlling the decadal-scale variation of SST (e.g., phase reversals of decadal SST anomaly).

The model and experimental designs used in the present paper are briefly described in section 2. In section 3, we present the seasonal modulation of decadal anomalies in the modeled SST field in relation to observations. In section 4, we examine the seasonal heat budgets of the ocean mixed layer relating to decadal SST anomalies, focusing on the contribution from the net shortwave radiation flux anomaly at the sea surface and resulting changes in the sea surface heat flux. Section 5 provides further discussion using long-term data of net shortwave radiation flux. A summary is presented in section 6.

2. Model and experimental design

a. Model description related to decadal variation

A modeling study is an effective means of investigating the seasonal heat budget of the ocean mixed layer in the midlatitudes of the North Pacific in which the observed data are insufficient to estimate the strength of vertical heat transport by entrainment. The simple ocean model employed here is the 3D bulk mixed layer model used in MK06. The model predicts two physical variables—the anomalous mixed layer temperature (T ′), which is a good proxy for SST anomaly, and the anomalous mixed layer depth (Hm). The governing equation for the anomalous heat budget of the ocean mixed layer is given as

 
formula

(e.g., Niiler and Kraus 1977; Frankignoul 1985). The right-hand terms represent the rates of anomalous heat transport by geostrophic advection ug, Ekman transport uE, vertical entrainment we, net heat flux at the sea surface Q′, and horizontal diffusion, respectively. Similarly, the governing equation for the anomalous mixed layer depth is written as

 
formula

The right-hand terms represent the entrainment velocity, the contributions of geostrophic advection and Ekman pumping, and, last, the horizontal diffusion. To investigate the seasonal modulation of the decadal temperature anomaly, the two components and (· · ·)′ are defined for each term as representing the climatological seasonal variation (i.e., the daily climatological values) at each grid point and the deviations from these daily mean climatologies, respectively. The terms in Eqs. (1) and (2) are then linearized in the integration, with the assumption that (· · ·)′ ≪ .

As documented in MK06, the model includes two key processes relating to decadal variation and its seasonal modulation, namely, the slow dynamic adjustment in upper-ocean circulation caused by the first baroclinic Rossby wave mode (Latif and Barnett 1994, 1996) and the local reemergence process of the SST anomalies (Alexander and Deser 1995; Alexander et al. 1999; Alexander et al. 2001; Hanawa and Sugimoto 2004; Sugimoto and Hanawa 2005). Considering the first baroclinic Rossby wave mode, the anomalous geostrophic current velocity ug is calculated using the anomalous geostrophic streamfunction predicted by the potential vorticity equation on a 1° grid. MK06 have explored some concerns resulting from this representation of the anomalous geostrophic velocity. They have reported that anomalous horizontal temperature transports by smaller eddy components (regarded as subgrid-scale disturbances in the 1° grid) make a slight contribution to the seasonal heat budget relating to the decadal SST anomaly in the KOE region, while the shorter time-scale deviations of the geostrophic velocity, which are not represented by the anomalous potential vorticity equation of our model, contribute negligibly to the seasonal modulation of decadal SST anomalies, particularly during summer–autumn. In line with our earlier work, our model also includes anomalous horizontal temperature transports by smaller eddy components that can be statistically estimated using a linear regression equation based on higher-resolved observational datasets (see appendix B of MK06). The horizontal heat transport by the anomalous geostrophic current velocity estimated here plays the major role in the phase reversals of the decadal SST anomaly in the KOE region (see Fig. 5 of MK06). The SST anomaly lags the anomalous temperature advection by a couple of years, consistent with previous findings of many studies (e.g., Miller et al. 1998; Deser et al. 1999; Xie et al. 2000; Yasuda and Kitamura 2003).

The temperature transport anomaly from vertical entrainment at the bottom of the mixed layer [weΔT]′, which is related to the local reemergence process, is calculated from the potential energy equation of the mixed layer (for details, see Qiu and Kelly 1993; Yasuda et al. 2000). The potential energy input due to the vertical entrainment at the bottom of the mixed layer, net sea surface heat flux, and the absorption of solar radiation is balanced with the conversion rate between turbulent kinetic energy and potential energy in the mixed layer, which can be parameterized by the sum of the effects of wind stirring and cooling at the sea surface. In this formulation, the rate of the anomalous heat transport by vertical entrainment during autumn and winter depends mainly on the strength of the turbulent kinetic energy input at the sea surface due to wind stirring and cooling. Note that the vertical entrainment process is calculated as an asymmetric process because the detrainment process has no thermodynamical effect on the mixed layer temperature evolution in the case of negative entrainment. To determine the local reemergence process, the temperature anomaly beneath the mixed layer (T ′diff) is controlled in our model, because this value affects the anomalous temperature difference between the mixed layer and water just below the mixed layer, ΔT ′(=T ′ − T ′diff) (for details, see Qiu and Kelly 1993). The value of T ′diff within the mixed layer is always set to be T ′, while the value beneath the mixed layer is governed by a simple vertical diffusive model (e.g., Kim 1976) to carry the memory of the mixed layer temperature anomaly formed when the mixed layer is deep. Note that the vertical diffusion is included not to vertically transport the temperature anomalies, but to reduce the unsmoothed fluctuations found in the vertical water temperature profile for numerical stability. The bulk mixed layer model and this vertical diffusive model are alternately integrated and the temperature anomaly beneath the ocean mixed layer, which has the memory of the mixed layer temperature anomaly, affects the heat budget of the mixed layer through changes in the entrainment velocity and the mixed layer depth. Although the often-used vertical diffusive models that ignore the contribution from the horizontal temperature advection beneath the mixed layer (e.g., Schneider et al. 2002; Sugimoto and Hanawa 2005) are one of the limitations of the present study, such simplified models are still able to keep the temperature anomalies beneath the mixed layer, and by and large represent a seasonally reemergence process of mixed layer temperature anomalies (MK06).

b. Experimental design

The model covers a region between 10° and 70°N in the North Pacific on a 1° latitude × 1° longitude grid. Daily atmospheric data (wind velocity, air temperature, and specific humidity) for near the sea surface from the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis dataset are utilized as boundary conditions. Sea surface flux values (sensible and latent heat fluxes and wind stress) are calculated from the bulk aerodynamic formulae. The climatological geostrophic current velocity is calculated using the monthly mean dynamic topography obtained from Ocean Topography Experiment (TOPEX)/Poseidon altimeter data. The anomalous geostrophic velocity is predicted by the potential vorticity equation from January 1950 to December 1999 with the anomalous geostrophic streamfunction set uniformly to zero as the initial condition. The bulk mixed layer model governed by Eqs. (1) and (2) is integrated from January 1960 to December 1999. Data on the climatologies of SST and mixed layer depth and the initial (January 1960) SST anomaly conditions are taken from the Simple Ocean Data Assimilation (SODA) dataset (Carton et al. 2000a, b). As we use different analyses and observational data for forcing our model, the modeling is not wholly consistent. Nevertheless, our simple approach is worthwhile for the identification of basic mechanisms involved in the seasonal modulation of decadal SST anomalies since it includes the key processes responsible for decadal variation and its seasonal modulation (MK06).

Two experiments (referred to as CTL and RAD) are performed to examine the seasonal modulation of decadal SST anomalies by focusing on the effect of anomalous net shortwave radiation fluxes at the sea surface (Table 1). Monthly mean data from the International Satellite Cloud Climatology Project (ISCCP) observation (Rossow and Schiffer 1999) is used to define the net shortwave radiation flux values at the sea surface and the monthly mean values are interpolated to the daily values by spline functions. In the RAD simulation the net shortwave radiation flux anomalies are defined as the deviations from the daily climatological values, while those in the CTL simulation are set to be zero. Compiled monthly mean data from 1983 and beyond are mainly used in the following investigations since the ISCCP observed data are available only from July 1983 onward. Thus, the present work should be regarded as a case study of late-1980s decadal SST change rather than a comprehensive analysis for the dominant Pacific decadal variation. Note that the late-1980s SST change in the midlatitudes of the North Pacific reflects a typical phase change of the dominant decadal variation observed in the North Pacific (e.g., White 1995; Yasuda and Hanawa 1997; Schneider and Miller 2001; Yasunaka and Hanawa 2002). For a fuller statistical discussion, we perform additional 40-yr simulations using the radiation flux from the NCEP–NCAR reanalysis data rather than the ISCCP-observed data and examine the statistical features of these numerical solutions in greater detail. Note that the reanalyzed radiation flux data strongly depend on the particular model used for the assimilation since the NCEP–NCAR reanalysis is based on cloud parameterizations in weather forecasts and is initialized without direct input of information derived from cloud observations.

Table 1.

List of experiments with various datasets for net shortwave radiation flux at the sea surface.

List of experiments with various datasets for net shortwave radiation flux at the sea surface.
List of experiments with various datasets for net shortwave radiation flux at the sea surface.

In the CTL and RAD experiments, the contribution of the downward longwave radiation flux at the sea surface is ignored since low-level cloud cover effectively controls the strength of the downward flux at the sea surface and very large errors are expected, even in the ISCCP-observed data. For example, the ISCCP low-level cloud cover takes into account only low-level clouds that are not obscured by higher-level clouds. Weare (2000) has reported that the dominant mode of the low-level cloud cover variations derived by ground-based observation is closer to that of the NCEP–NCAR reanalysis dataset than that of the ISCCP observation. Results of additional experiments using the reanalyzed downward longwave radiation flux dataset will be briefly discussed in section 6, while further studies are required to obtain a more reliable longwave radiation flux dataset that covers the entire ocean.

3. Seasonal modulation of decadal SST anomalies

The numerical solution of the RAD simulation displays a realistic interannual SST variation in the midlatitude North Pacific (Fig. 1) where strong interannual variations of cloud cover are observed in summer (e.g., Norris et al. 1998; Norris 2000). For all summers, the SST anomaly of the RAD case is approximately 20% larger in amplitude than that of the CTL case. This enhancement of the SST anomalies leads to a more favorable comparison with the optimum interpolation SST (OISST) observation (Reynolds and Smith 1994). To define the decadal anomalies as deviations from the climatological seasonal variations, we derived 5-yr running means for each month (i.e., averaging the values for the same month over 5 yr to all data). We did not remove a linear trend since the time period analyzed in the present study (from 1985 to 1997) is too short to separate the dominant decadal SST variation from the globally warming trend. While Levitus et al. (2005a) pointed out the presence of global warming signals in the upper ocean, the subsequent study (Levitus et al. 2005b) revealed that it was the Atlantic Ocean that had played the decisive role in the global increase of the upper-ocean heat content during the past 50 yr. The modeled decadal SST anomalies display a typical decadal variation of observed SST and its seasonal modulation (Fig. 2): A warm decadal SST anomaly is simulated in the early 1990s, whereas a cool SST anomaly is obtained in the mid-1980s (e.g., White 1995; Deser et al. 1996; Yasuda and Hanawa 1997; Schneider and Miller 2001). These SST anomalies in the western part of the midlatitude North Pacific (hereafter called the northwestern Pacific) are strongest in winter (Nakamura and Yamagata 1999; Schneider et al. 2002; Mochizuki and Kida 2003), while those in the eastern part of the midlatitude North Pacific (hereafter called the northeastern Pacific) are strongest in summer. The SST anomalies of both simulations are almost the same during the winter–spring interval (Fig. 2d). On the other hand, the decadal SST anomaly derived from the RAD data during summer–autumn is damped in the northwestern Pacific but is enhanced in the northeastern Pacific (Fig. 2d), when compared to the CTL data. These differences in the decadal SST anomaly between the RAD and CTL cases are definitely caused by changes in the net shortwave radiation flux (Fig. 3a). Note that both the difference in the decadal SST anomaly between the two simulations and the decadal anomaly of net shortwave radiation flux are less (but significantly) correlated with changes in cloud cover (Fig. 3b) since the strength of cloud radiative forcing is closely related to not only cloud cover but also the cloud property, which is quite different between deep convective clouds and shallow stratiform clouds. As shown by the following analyses, the types of the clouds that dominantly control the net shortwave radiation flux variations may depend on time and space. For instance, the low-level stratiform clouds seem to be mostly observed west of North America in summer, while the convective clouds may control net shortwave radiation flux variations over the wintertime central North Pacific where strong synoptic disturbances are observed.

Fig. 1.

Hovmöller diagrams of SST anomalies averaged over 38°–42°N from (a) the OISST observation, (b) the RAD simulation, (c) the CTL simulation, and (d) the differences between the RAD and CTL simulations. In (a), (b), and (c), the contour intervals are 1°C; the darker(lighter) shaded regions denote areas greater(less) than 0.5°C (−0.5°C). In (d), the contour intervals are 0.2°C; the darker (lighter) shaded regions denote areas greater (less) than 0.1°C (−0.1°C).

Fig. 1.

Hovmöller diagrams of SST anomalies averaged over 38°–42°N from (a) the OISST observation, (b) the RAD simulation, (c) the CTL simulation, and (d) the differences between the RAD and CTL simulations. In (a), (b), and (c), the contour intervals are 1°C; the darker(lighter) shaded regions denote areas greater(less) than 0.5°C (−0.5°C). In (d), the contour intervals are 0.2°C; the darker (lighter) shaded regions denote areas greater (less) than 0.1°C (−0.1°C).

Fig. 2.

As in Fig. 1 but for the SST anomalies to which 5-yr running means were applied for each month (averaging the values for the same month over 5 yr). In (a), (b), and (c), the contour intervals are 0.5°C; the darker (lighter) shaded regions denote areas greater (less) than 0.25°C (−0.25°C). In (d), the contour interval is 0.1°C; the darker (lighter) shaded regions denote areas greater (less) than 0.05°C (−0.05°C).

Fig. 2.

As in Fig. 1 but for the SST anomalies to which 5-yr running means were applied for each month (averaging the values for the same month over 5 yr). In (a), (b), and (c), the contour intervals are 0.5°C; the darker (lighter) shaded regions denote areas greater (less) than 0.25°C (−0.25°C). In (d), the contour interval is 0.1°C; the darker (lighter) shaded regions denote areas greater (less) than 0.05°C (−0.05°C).

Fig. 3.

Hovmöller diagrams of (a) anomalous net shortwave radiation flux at the sea surface and (b) anomalous cloud cover from the ISCCP observation. Values are averaged over 38°–42°N and 5-yr running means are applied for each month. In (a), the contour intervals are 5 W m−2; the darker (lighter) shaded regions denote areas greater (less) than 2.5 W m−2 (−2.5 W m−2). In (b), the contour intervals are 2%; the darker(lighter) shaded regions denote areas greater (less) than 1% (−1%). Values of R represent the coefficients of the pattern correlation with Fig. 2d.

Fig. 3.

Hovmöller diagrams of (a) anomalous net shortwave radiation flux at the sea surface and (b) anomalous cloud cover from the ISCCP observation. Values are averaged over 38°–42°N and 5-yr running means are applied for each month. In (a), the contour intervals are 5 W m−2; the darker (lighter) shaded regions denote areas greater (less) than 2.5 W m−2 (−2.5 W m−2). In (b), the contour intervals are 2%; the darker(lighter) shaded regions denote areas greater (less) than 1% (−1%). Values of R represent the coefficients of the pattern correlation with Fig. 2d.

The spatial patterns of the seasonal modulation of the decadal SST anomaly are also better simulated in the RAD case (Fig. 4) when focusing on the late-1980s decadal SST change (e.g., White 1995; Deser et al. 1996; Yasuda and Hanawa 1997; Schneider and Miller 2001; Yasunaka and Hanawa 2002, 2003). The significant decadal SST anomalies cover the entire midlatitude North Pacific in winter. These anomalies are also detected in spring, although not in the southeastern part of the KOE region (south of 39°N and east of 165°E). This absence possibly reflects the fact that time scales of local thermal damping at the sea surface are shorter in the subtropical regions than those in the high-latitude regions (e.g., Lau and Nath 1996; Watanabe and Kimoto 2000). The decadal SST anomalies in the entire KOE region decay to insignificant levels in summer but reemerge in the following autumn. On the other hand, the SST anomalies in the northeastern Pacific are weakly enhanced during summer and persist throughout the year.

Fig. 4.

Seasonal averages of the decadal SST anomalies during 1989–92 minus those during 1985–88 for the (left) RAD data and (right) OISST observed data. The significant areas are shaded at 95% confidence levels; solid lines represent positive values and dashed lines represent negative values.

Fig. 4.

Seasonal averages of the decadal SST anomalies during 1989–92 minus those during 1985–88 for the (left) RAD data and (right) OISST observed data. The significant areas are shaded at 95% confidence levels; solid lines represent positive values and dashed lines represent negative values.

Figure 5a shows averages of the observed net shortwave radiation flux at the sea surface during July–September for 1989–92 minus those for 1985–88. Positive values denote net shortwave radiation flux anomalies acting to warm the upper ocean. The net shortwave radiation flux around the KOE region is damped north of 40°N (and is enhanced south of 40°N) after the late 1980s. A similar spatial pattern is also obtained from the NCEP–NCAR reanalysis data (Fig. 5b). These changes of net shortwave radiation flux are strongly related to those of the observed total cloud amount (Fig. 6a), which is primarily controlled by changes in the amount of high-level cloud (Fig. 6b). For the following heat budget analyses, therefore, we focus on the northern part of the KOE region (hereafter the NKOE region), defined within the limits 39°–44°N, 150°E–180° (see Fig. 5) rather than the entire KOE region. The enhancement of the net shortwave radiation flux is also noticeable in the northeastern Pacific (Fig. 5) where a strong SST anomaly is observed throughout the year (see Fig. 4). Thus, we similarly examine the seasonal heat budget north of the subtropical frontal region (here after the NSTF region), defined within the limits 35°–40°N, 130°–160°W (see Fig. 5). In the NSTF region, changes in low-level cloud cover control the total cloud cover and the net shortwave radiation flux on decadal time scales (Fig. 6d). Note that a number of observational studies have reported that the low-level stratiform cloud cover is observed to be strongest in the summer seasons (e.g., Klein and Hartmann 1993; Yuter et al. 2000; Duynkerke and Teixeira 2001; Garreaud et al. 2001; Bretherton et al. 2004).

Fig. 5.

Averages of the net shortwave radiation flux at the sea surface during July–September 1989–92 minus those of 1985–88, for (a) the ISCCP observed data and (b) the NCEP–NCAR reanalysis data. The contour intervals are 10 W m−2. The significant areas are shaded at 95% confidence levels; solid lines represent positive values and dashed lines represent negative values. The two rectangles represent the NKOE region and the NSTF region defined in the present study (see text for details).

Fig. 5.

Averages of the net shortwave radiation flux at the sea surface during July–September 1989–92 minus those of 1985–88, for (a) the ISCCP observed data and (b) the NCEP–NCAR reanalysis data. The contour intervals are 10 W m−2. The significant areas are shaded at 95% confidence levels; solid lines represent positive values and dashed lines represent negative values. The two rectangles represent the NKOE region and the NSTF region defined in the present study (see text for details).

Fig. 6.

As in Fig. 5a except for (a) the total cloud amount, (b) the high-level cloud amount, (c) the midlevel cloud amount, and (d) the low-level cloud amount derived from the ISCCP observations. Contour intervals are 2%.

Fig. 6.

As in Fig. 5a except for (a) the total cloud amount, (b) the high-level cloud amount, (c) the midlevel cloud amount, and (d) the low-level cloud amount derived from the ISCCP observations. Contour intervals are 2%.

Norris et al. (1998) and Norris (2000) obtained similar spatial patterns as the leading EOF modes of the interannual to interdecadal variations of summertime convective clouds. Since they analyzed unfiltered datasets and the principal component (the leading EOF mode) seemed to be composed of interannual variations and a specific decadal variation, their results would give us an important insight into the definition of the above two key regions. Note that the physical relationships among high-level cloud cover, storm-track position, net shortwave radiation flux, and SST are still unclear with respect to decadal variations. In addition, as documented in section 1, they did not focus on the quantitative impact of changes in the net shortwave radiation flux on the ocean heat budget. In the following sections, therefore, we clarify the dominant physical process controlling the seasonal modulations of decadal SST anomalies in these key regions (i.e., the NKOE and NSTF regions). Note that the difference in the area size between the NKOE region in the present study and the KOE region in MK06 have no effect on the major conclusions of the present study.

4. Seasonal heat budget of the ocean mixed layer

Upper panels of Fig. 7 show the lagged regressed values of the decadal SST anomalies for each month (i.e., seasonally modulated decadal SST anomalies) to the annual averages of the decadal SST anomaly (i.e., decadal SST anomalies with no seasonal modulation) for the two key regions in the midlatitude North Pacific. These regressed values are defined as the product of one standard deviation of the decadal SST anomaly for each month and the correlation coefficient between the annual average of the decadal SST anomaly and the monthly values of the decadal SST anomaly [see Eq. (7) of MK06]. From the definition, the regressed value represents the statistical amplitude of the decadal SST anomaly in a specific month. To clarify the physical processes that control seasonal modulation in decadal SST anomalies, we have similarly calculated the lagged regressed values of the selected terms in Eq. (1) for each month relative to the annually averaged decadal SST anomaly (Fig. 7). Note that in the present study the surface heat flux term represents the sum of the effects of sensible and latent heat fluxes and the upward component of the longwave radiation flux at the sea surface. As documented in section 2, the net shortwave radiation flux anomaly is externally given and the downward component of the longwave radiation flux anomaly is ignored in the present study.

Fig. 7.

Lagged regressed values of the decadal time series for SST anomaly and selected terms in Eq. (1) for each month relative to the annually averaged decadal SST anomaly. Terms are SST anomaly (°C), Ekman temperature transport [−HmuE·T]′ (°C month−1), vertical entrainment –[weΔT]′ (°C month−1), net shortwave radiation flux –QNSWRS/coρpo (°C month−1), and sea surface heat flux –Qsfc/coρpo (°C month−1). The values of each plotted term have been divided by the climatological mixed layer depth Hm and averaged over (a) the NKOE region and (b) the NSTF region. The solid and broken lines represent the RAD case and the CTL case, respectively; vertical bars represent significant ranges at a 95% confidence level.

Fig. 7.

Lagged regressed values of the decadal time series for SST anomaly and selected terms in Eq. (1) for each month relative to the annually averaged decadal SST anomaly. Terms are SST anomaly (°C), Ekman temperature transport [−HmuE·T]′ (°C month−1), vertical entrainment –[weΔT]′ (°C month−1), net shortwave radiation flux –QNSWRS/coρpo (°C month−1), and sea surface heat flux –Qsfc/coρpo (°C month−1). The values of each plotted term have been divided by the climatological mixed layer depth Hm and averaged over (a) the NKOE region and (b) the NSTF region. The solid and broken lines represent the RAD case and the CTL case, respectively; vertical bars represent significant ranges at a 95% confidence level.

From late autumn to spring in the NKOE regions of the RAD data, the net shortwave radiation flux no longer contributes to the seasonal heat budget of the ocean mixed layer relating to the decadal SST anomaly (Fig. 7a). The seasonal enhancement of the decadal SST anomalies is due to enhanced heat transfer by vertical entrainment in autumn and Ekman transport in early winter, while the anomalous surface heat flux acts to damp the decadal SST anomalies in spring, in agreement with the results of MK06. Although the temperature advection by geostrophic currents dominantly contributes to SST change on decadal time scales (phase reversals of decadal SST anomaly, see Fig. 5 of MK06), it contributes less to the seasonal modulation of the decadal SST anomalies when compared to the Ekman transport and the vertical entrainment (Fig. 7a). Note that the slight contribution of geostrophic advection in autumn may reflect the effect of the mean circulation in transporting the seasonally enhanced temperature anomaly near Japan into the NKOE region.

On the other hand, the net shortwave radiation flux anomaly during August–October is small (Fig. 7a) but, nevertheless, significantly damps the decadal SST anomaly in the NKOE region (Fig. 8a), whereas MK06 found that the summertime shortwave radiation flux makes only a slight contribution when averaged over the entire KOE region. The largest difference in SST and surface heat flux anomalies between the CTL and RAD cases is detected during September–November (Fig. 8a), which is about one month later than the negative peak of the summertime shortwave radiative forcing observed during August–October (Fig. 8a). The net shortwave radiative forcing, which is more effective at damping the SST anomaly, acts to cool the upper ocean and rapidly damps the decadal SST anomaly on seasonal time scales. This in turn leads to an anomalous surface heat flux that acts as a strong local atmospheric forcing during the subsequent September–November. In October of the RAD data, for example, the decadal SST anomaly is roughly 20%–30% smaller and the anomalous surface heat flux is about twice larger when compared to the CTL data. Nevertheless, the surface heat flux asserts less control over the seasonal enhancement of the decadal SST anomaly than either the Ekman transport or the vertical entrainment.

Fig. 8.

Lagged regressed values of the decadal time series of differences between the RAD and CTL simulations for SST anomaly and selected terms in Eq. (1) for each month relative to the annually averaged decadal SST anomaly of the RAD simulation. Terms are SST anomaly (°C), net shortwave radiation flux –QNSWRS/coρpo (°C month−1), and sea surface heat flux −Qsfc/coρpo (°C month−1). The values of each plotted term have been divided by the climatological mixed layer depth Hm and averaged over (a) the NKOE region and (b) the NSTF region, respectively. The vertical bars represent significant ranges at a 95% confidence level.

Fig. 8.

Lagged regressed values of the decadal time series of differences between the RAD and CTL simulations for SST anomaly and selected terms in Eq. (1) for each month relative to the annually averaged decadal SST anomaly of the RAD simulation. Terms are SST anomaly (°C), net shortwave radiation flux –QNSWRS/coρpo (°C month−1), and sea surface heat flux −Qsfc/coρpo (°C month−1). The values of each plotted term have been divided by the climatological mixed layer depth Hm and averaged over (a) the NKOE region and (b) the NSTF region, respectively. The vertical bars represent significant ranges at a 95% confidence level.

In the early 1990s when the SST values of the NKOE region are above normal, the high-level cloud cover is enhanced (Figs. 6a,b) and the net shortwave radiation flux is suppressed at the sea surface (Fig. 5). This corresponds to the northward shift of the storm track, whose intensity is defined here using the standard deviation of upward wind velocity relative to the monthly mean value at a height of 500 hPa (Fig. 9a). This northward shift of the storm track would be closely related to the enhanced meridional SST gradient (Fig. 9b) and the negative sea level pressure anomaly north of 40°N (Fig. 9c). Note that the positive sea level pressure anomaly around 30°N, 170°E implies that the footprinting process (Vimont et al. 2001, 2002) may play a major role in the seasonal-scale enhancement of decadal SST anomalies: The atmospheric local thermal forcing (the anomalous surface heat flux generated by changes in the atmospheric conditions) may enhance the decadal SST anomaly during summer–autumn. An additional heat budget analysis (lagged regression analysis) reveals that the atmospheric thermal forcing, which is mainly due to an air temperature anomaly near the sea surface resulting from changes in the meridional wind speed, is the major contributor to the autumnal enhancement of decadal SST anomaly east of Japan (south of 40°N and west of 150°E, not shown), while such atmospheric local thermal forcing is very small in the NKOE region. As suggested by Vimont et al. (2001), the footprinting process seems to affect the summertime SST mainly in the subtropics (i.e., south of the NKOE and NSTF regions) and still makes only a secondary contribution over the NKOE and NSTF regions.

Fig. 9.

As in Fig. 5a but for (a) the standard deviation of upward wind velocity relative to the monthly mean value at a height of 500 hPa, (b) the meridional SST gradient in the RAD case, and (c) the sea level pressure. Values of wind velocity and sea level pressure are taken from the NCEP–NCAR reanalysis. Contour intervals in (a), (b), and (c) are 0.5 hPa s−1, 0.5°C (103 km)−1, and 0.2 hPa.

Fig. 9.

As in Fig. 5a but for (a) the standard deviation of upward wind velocity relative to the monthly mean value at a height of 500 hPa, (b) the meridional SST gradient in the RAD case, and (c) the sea level pressure. Values of wind velocity and sea level pressure are taken from the NCEP–NCAR reanalysis. Contour intervals in (a), (b), and (c) are 0.5 hPa s−1, 0.5°C (103 km)−1, and 0.2 hPa.

For the NSTF region, in contrast, the summertime decadal SST anomaly in the RAD case is larger than that in winter (Fig. 7b). Those changes in the net shortwave radiation flux related to the anomalous low-level cloud cover (Figs. 6a,d) act to enhance the decadal SST anomaly and have a major influence on the summertime heat budget of the ocean mixed layer (Fig. 7b). As the positive peak of the summertime shortwave radiative forcing is observed during June–August (Fig. 8b), the SST anomaly in the RAD case is much stronger than that in the CTL case during July–September (Fig. 8b). In addition, the positive SST anomaly enhanced by this summertime net shortwave radiative forcing induces strong local thermal damping (mainly through anomalous latent heat and sensible heat fluxes) and may act to destabilize the atmospheric boundary layer. The atmospheric conditions derived from the NCEP–NCAR reanalysis also imply this SST radiation positive feedback process (e.g., Norris et al. 1998; Norris 2000). In the early 1990s when the SST values of the NSTF region are above normal, the lower troposphere is significantly destabilized at a height of 800 hPa (Fig. 10a) and the atmospheric boundary layer is moistened due to enhanced evaporation at the sea surface (Fig. 10b). The low-level clouds that are topped by a strong stable layer must be suppressed in this atmospheric condition, especially since a number of observational studies have indicated that low-level marine stratiform cloud cover is associated with greater atmospheric stability (e.g., Klein and Hartmann 1993; Weaver and Ramanathan 1997; Norris et al. 1998; Yuter et al. 2000; Duynkerke and Teixeira 2001; Garreaud et al. 2001; Bretherton et al. 2004).

Fig. 10.

Vertical profiles of the composited values (a) for the vertical gradient of potential temperature and (b) for the specific humidity derived from the NCEP–NCAR reanalysis data. Plotted values are the averages over the NSTF region during July–September of 1989–92 minus those of 1985–88. The horizontal bars represent significant ranges at a 95% confidence level.

Fig. 10.

Vertical profiles of the composited values (a) for the vertical gradient of potential temperature and (b) for the specific humidity derived from the NCEP–NCAR reanalysis data. Plotted values are the averages over the NSTF region during July–September of 1989–92 minus those of 1985–88. The horizontal bars represent significant ranges at a 95% confidence level.

These lagged regression analyses reveal that the net shortwave radiation flux anomaly contributes to the summertime evolution of the decadal SST anomalies and consequently also controls the surface heat flux anomaly from late summer to early autumn in the NSTF region. Composite maps of decadal SST anomalies and related surface heat flux anomalies also support these results (Figs. 11, 12 and 13). Due to the anomalous net shortwave radiation flux at the sea surface, the decadal SST anomaly around the NSTF region is greatly enhanced during June–October, while that over the NKOE region is slightly damped during August–October (Figs. 13a,b,c). Corresponding to these differences in the modified SST anomaly, the resulting surface heat flux values are changed (Figs. 13d,e,f). From late spring to early summer, the sea surface heat flux acts as a local thermal damping mechanism (Figs. 11d and 12d) and the anomalous radiation fluxes have little effect on the decadal SST anomaly and the related surface heat flux anomaly in both the NKOE and NSTF regions (Figs. 13a and 13d). From midsummer to early autumn, on the other hand, the enhanced SST anomaly caused by the net shortwave radiation flux anomaly in the RAD case enhances the local thermal damping rate in the NSTF region (Figs. 13b,c,e,f). The surface heat flux anomaly in midsummer displayed in the RAD data is twice as large as that of the CTL data (Figs. 11e and 12e). In addition, the surface heat flux of the RAD data acts to damp the decadal SST anomaly even in early autumn, while that of the CTL data acts to enhance the SST anomaly (Figs. 11f and 12f). This surface heat flux anomaly in the RAD case generally warms the lower troposphere when SST values are above normal (Fig. 13), and again may imply a SST–radiation positive feedback process (see Fig. 10). As for the NKOE region, the surface heat flux forcing to the atmosphere seems to be only slightly controlled by the net shortwave radiation flux anomaly and the resulting SST anomaly. Nevertheless, the atmospheric thermal forcing registered in the RAD data is significantly larger than that in the CTL data in early autumn, particularly to the east of Japan (Figs. 13c and 13f), although this difference is small in magnitude when compared to the decadal anomaly of surface heat flux (Fig. 7a).

Fig. 11.

Averages of the (left) RAD-derived SST anomalies and (right) surface heat flux anomalies during (top) April–June, (middle) June–August, and (bottom) August–October of 1989–92 minus those of 1985–88. The significant areas are shaded at 95% confidence levels. Solid lines represent positive values and dashed lines represent negative values. The two rectangles represent the NKOE and NSTF region.

Fig. 11.

Averages of the (left) RAD-derived SST anomalies and (right) surface heat flux anomalies during (top) April–June, (middle) June–August, and (bottom) August–October of 1989–92 minus those of 1985–88. The significant areas are shaded at 95% confidence levels. Solid lines represent positive values and dashed lines represent negative values. The two rectangles represent the NKOE and NSTF region.

Fig. 12.

As in Fig. 11 but for the results of the CTL case.

Fig. 12.

As in Fig. 11 but for the results of the CTL case.

Fig. 13.

As in Fig. 11 but for the differences between the RAD and CTL case. Note that the contour intervals in the left panels are 0.1°C.

Fig. 13.

As in Fig. 11 but for the differences between the RAD and CTL case. Note that the contour intervals in the left panels are 0.1°C.

5. Discussion

The above analysis is limited to a case study of late-1980s decadal SST change in the midlatitude North Pacific because all of the analyses have been conducted for the period 1983–99. To make a more detailed examination of the statistical characteristics of the seasonal modulation of decadal SST variation, we perform two additional 40-yr integrations (the RADN and CTLN simulations) from 1960 to 1999 (Table 1). The anomalous net shortwave radiation flux at the sea surface is defined using the NCEP–NCAR reanalysis data rather than the ISCCP observations in the RADN case, while this flux is set to the climatological seasonal value derived from the NCEP–NCAR reanalysis in the CTLN case. Although the reanalyzed radiation flux data are generally dependent on the particular model used for the assimilation, the composited values of the net shortwave radiation flux taken from the NCEP–NCAR reanalysis are consistent with those from the ISCCP observations for decadal SST change in the late 1980s (Fig. 5). In addition, the seasonally modulated decadal variations of the reanalyzed net shortwave radiation flux are consistent with those of the ISCCP data in both the NKOE and NSTF regions (Figs. 14c,d). The modeled decadal SST anomaly and its seasonal modulation in the RADN case result in a similar behavior to that observed in the RAD case (Figs. 14a,b). Thus, both the RADN and CTLN datasets merit more detailed examination. Note that the decadal anomaly of the summertime net shortwave radiation flux taken from the NCEP–NCAR reanalysis (used in the RADN case) is smaller in magnitude than that from the ISCCP observation (used in the RAD case). The ratios of the standard deviation of the summertime net shortwave radiation flux in the NCEP–NCAR reanalysis relative to that of the ISCCP data are 0.49 over the NKOE region and 0.67 over the NSTF region.

Fig. 14.

(top) Temporal distributions of decadal SST anomalies (°C) and (bottom) contribution from net shortwave radiation flux to the seasonal heat budget (°C month−1). Plotted values are the averages over (left) the NKOE region and (right) the NSTF region. Solid and dotted lines represent the RAD simulation and the RADN simulation, respectively. Dash–dotted lines of the top panels represent the OISST observation, while those of the bottom panels represent the differences between the RAD case and the RADN case.

Fig. 14.

(top) Temporal distributions of decadal SST anomalies (°C) and (bottom) contribution from net shortwave radiation flux to the seasonal heat budget (°C month−1). Plotted values are the averages over (left) the NKOE region and (right) the NSTF region. Solid and dotted lines represent the RAD simulation and the RADN simulation, respectively. Dash–dotted lines of the top panels represent the OISST observation, while those of the bottom panels represent the differences between the RAD case and the RADN case.

For the dominant physical processes controlling the summertime evolution of the decadal SST anomaly, the RADN data during 1960–99 provide qualitatively similar results to those from the RAD data during 1983–99, while some differences are found in the amplitude of the contribution from sea surface fluxes. The net shortwave radiation flux anomaly acts to damp the summertime decadal SST anomaly in the NKOE region of the RADN data, while the amplitude is stronger and the seasonal peak of the contribution is detected about one month earlier when compared to the RAD case (Figs. 7 and 15). The seasonal peak of the resulting changes in sea surface heat flux is also about one month earlier in the RADN data (Figs. 8 and 16). These slight differences might reflect the specific characteristics of the late-1980s decadal SST change in contrast to the statistical features of the Pacific decadal variation. In addition, differences between the RAD case and the RADN case could result from differences in data processing between observational and reanalysis datasets. The atmospheric response to a SST anomaly seems to be more clearly represented in a reanalyzed dataset than in the observations, probably because the cloud amounts in the reanalysis data are not directly assimilated but are derived from an atmospheric model using observed SST data as a boundary condition. Arakawa and Kitoh (2004) have pointed out that reanalyzed precipitation data are overdependent on the lower boundary conditions (the observed SST values). We speculate that the reanalyzed cloud cover may adjust to the SST changes instantaneously. In the NSTF region, the net shortwave radiation flux anomaly in the RADN case is small in amplitude (half of that in the RAD case) but still significantly contributes to the summertime enhancement of decadal SST anomaly (Figs. 7b and 15b). The anomalous net shortwave radiation flux still represents a dominant contribution to the summertime heat budget, and the surface heat flux anomaly is sufficiently modified to enhance the SST anomaly (Figs. 8b and 16b). This slightly smaller flux contribution in the RADN case relative to the RAD case may be due to less reliable values of low-level stratiform cloud amounts in the reanalysis data. Nevertheless, results of the present study are clearly relevant to the comprehensive description of the dominant Pacific decadal variation, while the shortwave radiation flux values from the NCEP–NCAR reanalysis data strongly depended on less reliable cloud fields of the specific model than those of the ISCCP-observed data.

Fig. 15.

Lagged regressed values of the decadal time series for SST anomaly and selected terms in Eq. (1) for each month relative to the annually averaged decadal SST anomaly. Terms are SST anomaly (°C), net shortwave radiation flux –QNSWRS/coρpo (°C month−1), and sea surface heat flux −Qsfc/coρpo (°C month−1). Values of each plotted term have been divided by the climatological mixed layer depth Hm and averaged over (a) the NKOE region and (b) the NSTF region. The solid and broken lines represent the RADN case and the CTLN case (the results of 40-yr integrations using the NCEP–NCAR reanalyzed radiation data), respectively. The vertical bars represent significant ranges at a 95% confidence level.

Fig. 15.

Lagged regressed values of the decadal time series for SST anomaly and selected terms in Eq. (1) for each month relative to the annually averaged decadal SST anomaly. Terms are SST anomaly (°C), net shortwave radiation flux –QNSWRS/coρpo (°C month−1), and sea surface heat flux −Qsfc/coρpo (°C month−1). Values of each plotted term have been divided by the climatological mixed layer depth Hm and averaged over (a) the NKOE region and (b) the NSTF region. The solid and broken lines represent the RADN case and the CTLN case (the results of 40-yr integrations using the NCEP–NCAR reanalyzed radiation data), respectively. The vertical bars represent significant ranges at a 95% confidence level.

Fig. 16.

As in Fig. 8 but for lagged regressed values calculated using the results of the RADN and CTLN simulations rather than the RAD and CTL simulations.

Fig. 16.

As in Fig. 8 but for lagged regressed values calculated using the results of the RADN and CTLN simulations rather than the RAD and CTL simulations.

6. Summary

The present study has examined the seasonal heat budget of the ocean mixed layer in the context of late-1980s decadal SST change and has clarified the physical process controlling the summertime evolution of decadal SST anomalies in the midlatitudes of the North Pacific. In particular, we have focused on the quantitative contributions of both the net shortwave radiation flux and the heat flux at the sea surface. Note that the modeled decadal SST anomalies seem to be slightly smaller than observed SST anomalies only in a narrow area (along 40°N, 150°E–180°) (e.g., Figs. 2a,b, 4). Nevertheless, these differences in magnitude of the decadal SST anomalies between the RAD simulation and the observation should not change the major conclusions of the present study as the negative anomaly of the net shortwave radiation flux is shifted northward to this narrow area (Fig. 5) and the area-averaged SST values over the NKOE and NSTF regions are almost the same as the observed values (Figs. 14a,b).

The decadal SST anomaly in the NKOE region decays seasonally from late spring to early autumn, of which the quasi-seasonal decay from late spring to early summer is controlled by the anomalous local thermal damping, in agreement with the results of an earlier study (MK06). From midsummer to early autumn, on the other hand, an anomalous radiation flux affects the seasonal heat budget of the ocean mixed layer. When the summertime SST around the KOE region is above normal, the storm track is observed to be shifted northward due to changes in the meridional SST gradient. This northward-shifted storm track changes the high-level cloud cover around the KOE region and thus suppresses the net shortwave radiation influx at the sea surface in the NKOE region. This reduced radiation influx acts to damp the summertime decadal SST anomaly in the NKOE region. Corresponding to the change in net shortwave radiative forcing, the sea surface heat flux anomaly is slightly enhanced from late summer to early autumn, but this change is significant as it can obscure the rapid enhancement (reemergence) of the decadal SST anomaly in the following autumn in the NKOE region of the RAD data.

Over the NSTF region, in contrast, the summertime decadal SST anomaly is enhanced. The local thermal damping process rarely influences the seasonal heat budget from late spring to early summer. From midsummer to early autumn, the net shortwave radiation flux anomaly, which results from changes in low-level stratiform cloud cover, is the major contributor to the seasonal heat budget of the ocean mixed layer. When the summertime SST in the NSTF region is above normal, the surface heat flux anomaly warms the atmospheric boundary layer and consequently can suppress the stratiform cloud cover by destabilizing the lower troposphere over the period from midsummer to early autumn. This result implies that the net shortwave radiation flux anomaly and the accompanying heat flux anomaly at the sea surface (the thermal forcing of the local atmosphere in the NSTF region due to the decadal SST anomaly) may act as an SST–radiation positive feedback process. This feedback process would play a major role in amplifying the decadal SST anomalies in the NSTF region, though this possibility is not fully examined here.

The NCEP–NCAR reanalysis and the ISCCP observational datasets display similar spatiotemporal variations in net shortwave radiation flux at the sea surface (Figs. 5 and 14), which is primarily controlled by the changes in observed cloud cover (Fig. 6). However, the downward component of longwave radiation flux, of which the contribution is ignored in our experiments, is not controlled in this fashion (not shown). The ISCCP downward longwave radiation flux values before 1985 are much stronger than those after 1986 in the region where the climatological low-level cloud cover is observed to be very large. It is uncertain whether the ISCCP downward longwave radiation flux values during 1983–85, which act to overheat the upper ocean in the mid- and high latitudes of the North Pacific, either represent a source of long-term variability in globally averaged cloud cover or reflect an artificial and/or instrumental bias when obtaining the observed data. Although some recent studies have examined the long-term variability of cloud cover and its radiative forcing at the top of the atmosphere (e.g., Cess and Udelhofen 2003; Norris 2005), the radiation variability at the sea surface remains unclear. Thus, the ISCCP longwave radiation flux data seem to be unsuitable for the present analyses of late-1980s decadal change. For a brief discussion on the contribution from the downward longwave radiation flux, we have performed the additional experiments using the reanalyzed data. Results of these additional experiments suggest that the downward longwave radiation flux makes a negligible contribution to the seasonal heat budget associated with the decadal SST anomaly (not shown). Nevertheless, we are unable to conclude that the downward longwave radiation flux has little effect on the seasonal modulation of a decadal SST anomaly, since the reanalyzed low-level cloud amounts and radiation flux data should strongly depend on the particular model. Further work is required to obtain a more reliable longwave radiation flux dataset that covers the entire ocean.

In the present paper, the input datasets used as boundary conditions have been derived from different instrumental systems and modeling approaches, which may not be consistent. In addition, the model has been integrated by atmospheric forcing at the sea surface, while the atmospheric boundary conditions have been derived from the NCEP–NCAR reanalysis data, which reflect the spatiotemporal variations of observed SST values. Since the presence of a possible SST–radiation feedback process has been implied in these results, research using a general circulation model that better couples the atmosphere and ocean to accurately examine atmospheric responses and air–sea coupling processes is required. This should be performed with information on oceanic thermal conditions at the sea surface that is independent of both the ISCCP cloud data and the NCEP–NCAR reanalysis data used to provide the boundary conditions.

Acknowledgments

The authors wish to thank Professor S.-P. Xie and anonymous reviewers for their valuable comments and useful suggestions for the improvement of the manuscript. The authors gratefully acknowledge helpful comments with Professor H. Kida on developing the model. Thanks are also due to Professor J. P. Matthews for his careful reading and editing of this paper to improve the draft. This research formed part of the Innovative Program of Climate Change Projection for the 21st Century and the Research Revolution 2002 Project for Sustainable Coexistence of Human, Nature and the Earth. All of the figures were produced by the GFD–DENNOU Library.

REFERENCES

REFERENCES
Alexander
,
M. A.
, and
C.
Deser
,
1995
:
A mechanism for the recurrence of wintertime midlatitude SST anomalies.
J. Phys. Oceanogr.
,
25
,
122
137
.
Alexander
,
M. A.
,
C.
Deser
, and
M. S.
Timlin
,
1999
:
The reemergence of SST anomalies in the North Pacific Ocean.
J. Climate
,
12
,
2419
2433
.
Alexander
,
M. A.
,
M. S.
Timlin
, and
J. D.
Scott
,
2001
:
Winter-to-winter recurrence of sea surface temperature, salinity, and mixed layer depth anomalies.
Prog. Oceanogr.
,
49
,
41
61
.
Alexander
,
M. A.
,
N-C.
Lau
, and
J. D.
Scott
,
2004
:
Broadening the atmospheric bridge paradigm: ENSO teleconnections to the tropical west Pacific–Indian Oceans over the seasonal cycle and to the North Pacific.
Earth’s Climate: The Ocean–Atmosphere Interaction, Geophys. Monogr., No. 147, Amer. Geophys. Union, 85–103
.
Arakawa
,
O.
, and
A.
Kitoh
,
2004
:
Comparison of local precipitation–SST relationship between the observation and a reanalysis dataset.
Geophys. Res. Lett.
,
31
.
L12206, doi:10.1029/2004GL020283
.
Barsugli
,
J. J.
, and
D. S.
Battisti
,
1998
:
The basic effects of atmosphere–ocean thermal coupling on midlatitude variability.
J. Atmos. Sci.
,
55
,
477
493
.
Bladé
,
I.
,
1997
:
The influence of midlatitude ocean–atmosphere coupling on the low-frequency variability of a GCM. Part I: No tropical SST forcing.
J. Climate
,
10
,
2087
2106
.
Bretherton
,
C. S.
, and
Coauthors
,
2004
:
The EPIC 2001 stratocumulus study.
Bull. Amer. Meteor. Soc.
,
85
,
967
977
.
Carton
,
J. A.
,
G.
Chepurin
,
X.
Cao
, and
B.
Giese
,
2000a
:
A simple ocean data assimilation analysis of the global upper ocean 1950–95. Part I: Methodology.
J. Phys. Oceanogr.
,
30
,
294
309
.
Carton
,
J. A.
,
G.
Chepurin
, and
X.
Cao
,
2000b
:
A simple ocean data assimilation analysis of the global upper ocean 1950–95. Part II: Results.
J. Phys. Oceanogr.
,
30
,
311
326
.
Cess
,
R. D.
, and
P. M.
Udelhofen
,
2003
:
Climate change during 1985–1999: Cloud interactions determined from satellite measurements.
Geophys. Res. Lett.
,
30
.
1019, doi:10.1029/2002GL016128
.
Deser
,
C.
, and
M. L.
Blackmon
,
1993
:
Surface climate variations over the North Atlantic Ocean during winter: 1900–1989.
J. Climate
,
6
,
1743
1753
.
Deser
,
C.
,
M. A.
Alexander
, and
M. S.
Timlin
,
1996
:
Upper-ocean thermal variations in the North Pacific during 1970–1991.
J. Climate
,
9
,
1840
1855
.
Deser
,
C.
,
M. A.
Alexander
, and
M. S.
Timlin
,
1999
:
Evidence for a wind-driven intensification of the Kuroshio Current Extension from the 1970s to the 1980s.
J. Climate
,
12
,
1697
1706
.
Duynkerke
,
P. G.
, and
J.
Teixeira
,
2001
:
Comparison of the ECMWF reanalysis with FIRE I observations: Diurnal variation of marine stratocumulus.
J. Climate
,
14
,
1466
1478
.
Ferranti
,
L.
,
F.
Molteni
, and
T. N.
Palmer
,
1994
:
Impact of localized tropical and extratropical SST anomalies in ensembles of seasonal GCM integrations.
Quart. J. Roy. Meteor. Soc.
,
120
,
1613
1645
.
Frankignoul
,
C.
,
1985
:
Sea surface temperature anomalies, planetary waves, and air–sea feedback in the middle latitudes.
Rev. Geophys.
,
23
,
357
390
.
Garreaud
,
R. D.
,
J.
Rutllant
,
J.
Quintana
,
J.
Carrasco
, and
P.
Minnis
,
2001
:
CIMAR-5: A snapshot of the lower troposphere over the subtropical southeast Pacific.
Bull. Amer. Meteor. Soc.
,
82
,
2193
2207
.
Graham
,
N. E.
,
1994
:
Decadal-scale climate variability in the tropical and North Pacific during the 1970s and 1980s: Observations and model results.
Climate Dyn.
,
10
,
135
162
.
Hanawa
,
K.
, and
J.
Kamada
,
2001
:
Variability of core layer temperature (CLT) of the North Pacific subtropical mode water.
Geophys. Res. Lett.
,
28
,
2229
2232
.
Hanawa
,
K.
, and
S.
Sugimoto
,
2004
:
“Reemergence” areas of winter sea surface temperature anomalies in the world’s oceans.
Geophys. Res. Lett.
,
31
.
L10303, doi:10.1029/2004GL019904
.
Kachi
,
M.
, and
T.
Nitta
,
1997
:
Decadal variation of the global atmosphere–ocean system.
J. Meteor. Soc. Japan
,
75
,
657
675
.
Kim
,
J-W.
,
1976
:
A generalized bulk model of the oceanic mixed layer.
J. Phys. Oceanogr.
,
6
,
686
695
.
Klein
,
S. A.
, and
D. L.
Hartmann
,
1993
:
The seasonal cycle of low stratiform clouds.
J. Climate
,
6
,
1587
1606
.
Latif
,
M.
, and
T. P.
Barnett
,
1994
:
Causes of decadal climate variability over the North Pacific and North America.
Science
,
266
,
634
637
.
Latif
,
M.
, and
T. P.
Barnett
,
1996
:
Decadal climate variability over the North Pacific and North America: Dynamics and predictability.
J. Climate
,
9
,
2407
2423
.
Lau
,
N-C.
, and
M. J.
Nath
,
1990
:
A general circulation model study of the atmospheric response to extratropical SST anomalies observed in 1950–79.
J. Climate
,
3
,
965
989
.
Lau
,
N-C.
, and
M. J.
Nath
,
1994
:
A modeling study of the relative roles of tropical and extratropical SST anomalies in the variability of the global atmosphere–ocean system.
J. Climate
,
7
,
1184
1207
.
Lau
,
N-C.
, and
M. J.
Nath
,
1996
:
The role of the “atmospheric bridge” in linking tropical Pacific ENSO events to extratropical SST anomalies.
J. Climate
,
9
,
2036
2057
.
Levitus
,
S.
,
J. I.
Antonov
, and
T. P.
Boyer
,
2005a
:
Warming of the world ocean, 1955–2003.
Geophys. Res. Lett.
,
32
.
L02604, doi:10.1029/2004GL021592
.
Levitus
,
S.
,
J. I.
Antonov
,
T. P.
Boyer
,
H. E.
Garcia
, and
R. A.
Locarnini
,
2005b
:
EOF analysis of upper ocean heat content, 1956–2003.
Geophys. Res. Lett.
,
32
.
L18607, doi:10.1029/2005GL023606
.
Luo
,
J-J.
, and
T.
Yamagata
,
2003
:
A model study on the 1988–89 warming event in the northern North Pacific.
J. Phys. Oceanogr.
,
33
,
1815
1828
.
Miller
,
A. J.
, and
N.
Schneider
,
2000
:
Interdecadal climate regime dynamics in the North Pacific Ocean: Theories, observations, and ecosystem impacts.
Prog. Oceanogr.
,
47
,
355
379
.
Miller
,
A. J.
,
D. R.
Cayan
, and
W. B.
White
,
1998
:
A westward-intensified decadal change in the North Pacific thermocline and gyre-scale circulation.
J. Climate
,
11
,
3112
3127
.
Mochizuki
,
T.
, and
H.
Kida
,
2003
:
Maintenance of decadal SST anomalies in the midlatitude North Pacific.
J. Meteor. Soc. Japan
,
81
,
477
491
.
Mochizuki
,
T.
, and
H.
Kida
,
2006
:
Seasonality of decadal sea surface temperature anomalies in the northwestern Pacific.
J. Climate
,
19
,
2953
2968
.
Nakamura
,
H.
, and
T.
Yamagata
,
1999
:
Recent decadal SST variability in the northwestern Pacific and associated atmospheric anomalies.
Beyond El Niño: Decadal and Interdecadal Climate Variability, A. Navarra, Ed., Springer, 49–72
.
Nakamura
,
H.
,
G.
Lin
, and
T.
Yamagata
,
1997
:
Decadal climate variability in the North Pacific during the recent decades.
Bull. Amer. Meteor. Soc.
,
78
,
2215
2225
.
Niiler
,
P. P.
, and
E. B.
Kraus
,
1977
:
One-dimensional models of the upper ocean.
Modelling and Prediction of the Upper Layers of the Ocean, E. B. Kraus, Ed., Pergamon, 143–172
.
Nonaka
,
M.
, and
S-P.
Xie
,
2003
:
Covariations of sea surface temperature and wind over the Kuroshio and its extension: Evidence for ocean-to-atmosphere feedback.
J. Climate
,
16
,
1404
1413
.
Norris
,
J. R.
,
2000
:
Interannual and interdecadal variability in the storm track, cloudiness, and sea surface temperature over the summertime North Pacific.
J. Climate
,
13
,
422
430
.
Norris
,
J. R.
,
2005
:
Multidecadal changes in near-global cloud cover and estimated cloud cover radiative forcing.
J. Geophys. Res.
,
110
.
D08206, doi:10.1029/2004JD005600
.
Norris
,
J. R.
, and
C. B.
Leovy
,
1994
:
Interannual variability in stratiform cloudiness and sea surface temperature.
J. Climate
,
7
,
1915
1925
.
Norris
,
J. R.
,
Y.
Zhang
, and
J. M.
Wallace
,
1998
:
Role of low clouds in summertime atmosphere–ocean interactions over the North Pacific.
J. Climate
,
11
,
2482
2490
.
Palmer
,
T. N.
, and
Z.
Sun
,
1985
:
A modeling and observational study of the relationship between sea surface temperature in the northwest Atlantic and the atmospheric general circulation.
Quart. J. Roy. Meteor. Soc.
,
111
,
947
975
.
Park
,
S.
, and
C. B.
Leovy
,
2004
:
Marine low-cloud anomalies associated with ENSO.
J. Climate
,
17
,
3448
3469
.
Peng
,
S.
,
L. A.
Mysak
,
H.
Ritchie
,
J.
Derome
, and
B.
Dugas
,
1995
:
The differences between early and midwinter atmospheric responses to sea surface temperature anomalies in the Northwest Atlantic.
J. Climate
,
8
,
137
157
.
Peng
,
S.
,
W. A.
Robinson
, and
M. P.
Hoerling
,
1997
:
The modeled atmospheric response to midlatitude SST anomalies and its dependence on background circulation states.
J. Climate
,
10
,
971
987
.
Pierce
,
D. W.
,
T. P.
Barnett
,
N.
Schneider
,
R.
Saravanan
,
D.
Dommenget
, and
M.
Latif
,
2001
:
The role of ocean dynamics in producing decadal climate variability in the North Pacific.
Climate Dyn.
,
18
,
51
70
.
Qiu
,
B.
, and
K. A.
Kelly
,
1993
:
Upper-ocean heat balance in the Kuroshio Extension region.
J. Phys. Oceanogr.
,
23
,
2027
2041
.
Reynolds
,
R. W.
, and
T. M.
Smith
,
1994
:
Improved global sea surface temperature analyses using optimum interpolation.
J. Climate
,
7
,
929
948
.
Rossow
,
W. B.
, and
R. A.
Schiffer
,
1999
:
Advances in understanding clouds from ISCCP.
Bull. Amer. Meteor. Soc.
,
80
,
2261
2287
.
Saravanan
,
R.
,
1998
:
Atmospheric low-frequency variability and its relationship to midlatitude SST variability: Studies using the NCAR climate system model.
J. Climate
,
11
,
1386
1404
.
Schneider
,
N.
, and
A. J.
Miller
,
2001
:
Predicting western North Pacific Ocean climate.
J. Climate
,
14
,
3997
4002
.
Schneider
,
N.
,
A. J.
Miller
, and
D. W.
Pierce
,
2002
:
Anatomy of North Pacific decadal variability.
J. Climate
,
15
,
586
605
.
Seager
,
R.
,
Y.
Kushnir
,
N. H.
Naik
,
M. A.
Cane
, and
J.
Miller
,
2001
:
Wind-driven shifts in the latitude of the Kuroshio–Oyashio Extension and generation of SST anomalies on decadal timescales.
J. Climate
,
14
,
4249
4265
.
Sugimoto
,
S.
, and
K.
Hanawa
,
2005
:
Remote reemergence areas of winter sea surface temperature anomalies in the North Pacific.
Geophys. Res. Lett.
,
32
.
L01606, doi:10.1029/2004GL021410
.
Tanimoto
,
Y.
,
H.
Nakamura
,
T.
Kagimoto
, and
S.
Yamane
,
2003
:
An active role of extratropical sea surface temperature anomalies in determining anomalous turbulent heat flux.
J. Geophys. Res.
,
108
.
3304, doi:10.1029/2002JC001750
.
Trenberth
,
K. E.
, and
J. W.
Hurrell
,
1994
:
Decadal atmosphere–ocean variations in the Pacific.
Climate Dyn.
,
9
,
303
319
.
Venzke
,
S.
,
M.
Munnich
, and
M.
Latif
,
2000
:
On the predictability of decadal changes in the North Pacific.
Climate Dyn.
,
16
,
379
392
.
Vimont
,
D. J.
,
D. S.
Battisti
, and
A. C.
Hirst
,
2001
:
Footprinting: A seasonal connection between the tropics and mid-latitudes.
Geophys. Res. Lett.
,
28
,
3923
3926
.
Vimont
,
D. J.
,
D. S.
Battisti
, and
A. C.
Hirst
,
2002
:
Pacific interannual and interdecadal equatorial variability in a 1000-yr simulation of the CSIRO coupled general circulation model.
J. Climate
,
15
,
160
178
.
Vivier
,
F.
,
K. A.
Kelly
, and
L.
Thompson
,
2002
:
Heat budget in the Kuroshio Extension region: 1993–99.
J. Phys. Oceanogr.
,
32
,
3436
3454
.
Wallace
,
J. M.
,
C.
Smith
, and
Q.
Jiang
,
1990
:
Spatial patterns of atmosphere–ocean interaction in the northern winter.
J. Climate
,
3
,
990
998
.
Watanabe
,
M.
, and
M.
Kimoto
,
2000
:
Behavior of midlatitude decadal oscillations in a simple atmosphere–ocean system.
J. Meteor. Soc. Japan
,
78
,
441
460
.
Weare
,
B. C.
,
2000
:
Near-global observations of low clouds.
J. Climate
,
13
,
1255
1268
.
Weaver
,
C. P.
, and
V.
Ramanathan
,
1997
:
Relationships between large-scale vertical velocity, static stability, and cloud radiative forcing over Northern Hemisphere extratropical oceans.
J. Climate
,
10
,
2871
2887
.
White
,
W. B.
,
1995
:
Design of a global observing system for gyre-scale upper ocean temperature variability.
Prog. Oceanogr.
,
36
,
169
217
.
Xie
,
S-P.
,
2004
:
Satellite observations of cool ocean–atmosphere interaction.
Bull. Amer. Meteor. Soc.
,
85
,
195
208
.
Xie
,
S-P.
,
T.
Kunitani
,
A.
Kubokawa
,
M.
Nonaka
, and
S.
Hosoda
,
2000
:
Interdecadal thermocline variability in the North Pacific for 1958–1997: A GCM simulation.
J. Phys. Oceanogr.
,
30
,
2798
2813
.
Yasuda
,
I.
,
T.
Tozuka
,
M.
Noto
, and
S.
Kouketsu
,
2000
:
Heat balance and regime shifts of the mixed layer in the Kuroshio Extension.
Prog. Oceanogr.
,
47
,
257
278
.
Yasuda
,
T.
, and
K.
Hanawa
,
1997
:
Decadal changes in the mode waters in the midlatitude North Pacific.
J. Phys. Oceanogr.
,
27
,
858
870
.
Yasuda
,
T.
, and
Y.
Kitamura
,
2003
:
Long-term variability of North Pacific subtropical mode water in response to spin-up of the subtropical gyre.
J. Oceanogr.
,
59
,
279
290
.
Yasunaka
,
S.
, and
K.
Hanawa
,
2002
:
Regime shifts found in the Northern Hemisphere SST field.
J. Meteor. Soc. Japan
,
80
,
119
135
.
Yasunaka
,
S.
, and
K.
Hanawa
,
2003
:
Regime shifts in the Northern Hemisphere SST field: Revisited in relation to tropical variations.
J. Meteor. Soc. Japan
,
81
,
415
424
.
Yuter
,
S. E.
,
Y. L.
Serra
, and
R. A.
Houze
Jr.
,
2000
:
The 1997 Pan American climate studies tropical Eastern Pacific process study. Part II: Stratocumulus region.
Bull. Amer. Meteor. Soc.
,
81
,
483
490
.
Zhang
,
Y.
,
J. R.
Norris
, and
J. M.
Wallace
,
1998
:
Seasonality of large-scale atmosphere–ocean interaction over the North Pacific.
J. Climate
,
11
,
2473
2481
.

Footnotes

Corresponding author address: Takashi Mochizuki, Frontier Research Center for Global Change, Japan Agency for Marine–Earth Science and Technology, 3173-25 Showa-machi, Kanazawa-ku, Yokohama 236-0001, Japan. Email: motizuki@jamstec.go.jp