1. Introduction
The ocean plays an important role in climate variation due to its large thermal inertia. Pioneering works by Namias and Born (1970, 1974) first noted that sea surface temperature (SST) anomalies in the late winter tend to persist beyond the warming season in the midlatitude oceans. They hypothesized that temperature anomalies formed in the deep winter mixed layer are preserved under the shallow summer mixed layer. As the mixed layer deepens again in the next winter, thermal anomalies are entrained into the mixed layer. Then the anomalies recur at the sea surface in the following winter. Later, Alexander et al. (1999) called this mechanism “reemergence” of winter SST anomalies. Several authors have shown local reemergence (Alexander and Deser 1995; Alexander et al. 1999, 2001; Watanabe and Kimoto 2000; Timlin et al. 2002; Deser et al. 2003). Hanawa and Sugimoto (2004, hereafter HS04) detected seven “collocated” reemergence areas in the world’s oceans and showed that all the reemergence areas correspond to the formation areas of major mode waters. Sugimoto and Hanawa (2005a, 2006, manuscript submitted to J. Oceanogr., hereafter SUHA) investigated why reemergence does not occur in the North Pacific eastern subtropical mode water, and then they pointed out two major reasons: a vigorous mixing in the lower part of the mode water (Sugimoto and Hanawa 2005a; SUHA) and less heat input in the warming season (SUHA). In addition, Coëtlogon and Frankignoul (2003) showed that nonlocal reemergence also occurs in the area along the Gulf Stream path remote from the area where waters acquire SST anomalies in winter. Sugimoto and Hanawa (2005b, hereafter SH05) also pointed out that “remote” reemergence occurs as a result of the movement of the North Pacific subtropical mode water (NPSTMW) and showed that the remote reemergence area is situated at the sea area under the Aleutian low, that is, the central North Pacific. Here, the meaning of collocated reemergence area is that the reemergence area is situated at the same location of the formation area where SST anomalies are set in the previous winter, while the meaning of remote reemergence area is that the reemergence area is situated at a different location from the formation area.
The Pacific decadal oscillation (PDO) is a leading mode of decadal variability in SST anomalies in the North Pacific (Mantua et al. 1997; Minobe 1997, 2000; Mantua and Hare 2002). The PDO index is given by the time coefficient of the leading empirical orthogonal function (EOF) mode of the monthly SST in the Pacific north of 20°N as the representative mode of SST variability (Mantua et al. 1997). It is well known that this decadal variability greatly affects the marine ecosystem and fisheries (Ebbesmeyer et al. 1991; Mantua et al. 1997). The Kuroshio/Oyashio Extension area is a key region where negative feedback from ocean to atmosphere could occur through surface heat flux (Latif and Barnett 1994, 1996). Several authors have suggested that SST anomalies in the Kuroshio/Oyashio Extension region are formed as the delayed response to the change of the wind stress curl over the central Pacific (Miller and Schneider 2000; Seager et al. 2001; Schneider et al. 2002; Qiu 2003; Miller et al. 2004). Using a coupled model, Schneider et al. (2002) showed some evidence for a positive feedback between the Kuroshio Extension SST and the North Pacific Ekman pumping, while there was no negative feedback as suggested by Latif and Barnett (1994, 1996). However, feedback mechanisms to explain the PDO have not been clarified yet (see the review of Miller and Schneider 2000).
The NPSTMW is formed in the south of the Kuroshio Extension and is one of the major surface water masses in the subtropical gyre. To understand the decadal variability in the North Pacific, it is important to elucidate the long-term variability of the NPSTMW. Recently, Hanawa and Kamada (2001) and Abe (2004) have shown that a core-layer temperature (CLT) of NPSTMW has a decadal-scale variation, using various hydrographic datasets such as the World Ocean Database 2001 (Conkright et al. 2002). Here, CLT is defined as the temperature of the layer having a minimum temperature gradient, that is, the core layer of the mode water. Hanawa and Kamada have also proposed a working hypothesis of decadal-scale variation of CLT and the Aleutian low activity as follows: first, the strengthening of the Aleutian low causes an increase of the Kuroshio transport with a lag of 3 to 5 yr. This time lag is in line with the findings by Deser et al. (1999) and Yasuda and Kitamura (2003). Then, the increasing supply of warm water to the NPSTMW formation area causes CLT warming with a lag of about 2 yr. This oceanic warming in the northwestern part of the North Pacific may give some negative impact (feedback) on the Aleutian low activity, and then the Aleutian low begins to be weakened. A similar process with a sign reversal continues, and the oscillation is closed.
In the above discussion, the decadal-scale variations of the CLT and Aleutian low activity have been postulated to be in a close relationship with each other. However, it has not necessarily been clear how NPSTMW formed in the northwestern North Pacific impacts Aleutian low activity in the central North Pacific. In addition, the observational datasets used in the previous works are not long enough to discuss the decadal-scale variation. Therefore, as a first step to fill the gap of the linkage between variations of the CLT and Aleutian low activity, we try to elucidate how the remote reemergence of NPSTMW impacts on winter SST variation in the central North Pacific, using various long-term datasets.
The remainder of this paper is organized as follows: In section 2, the datasets used are outlined. In section 3, a period-dependent nature of occurrence of remote reemergence is described. In section 4, causes of a period-dependent nature of reemergence of NPSTMW are explored. In section 5, performing a multiple regression analysis, we quantitatively examine an impact of remote reemergence on SST variation in the central North Pacific. Section 6 gives our summary.
2. Datasets and method
In the present study, various kinds of long-term datasets such as SST, CLT, and an index representing atmospheric variation are used. The analysis period is 74 years from 1930 to 2003.
To analyze basinwide SST fields on a monthly basis, we use two SST datasets: the Extended Reconstruction Sea Surface Temperature (ERSST) dataset (Smith and Reynolds 2004) and the Global Sea Ice–Sea Surface Temperature (GISST) dataset, version 2.3b (Rayner et al. 1996).
To represent long-term variation of NPSTMW temperature, the data of CLT recently prepared by Abe (2004) are used, which is shown in Fig. 1a. These CLT data are produced based on hydrographic data archived in the World Ocean Database 2001 and the Japan Fisheries Oceanography Database, and those collected by the Japan Oceanographic Data Center. The yearly CLT data are those averaged in the region of 30°–35°N, 140°–150°E in space and from May to December in time (see Abe 2004 for details).
To observe a reemergence process, the upper-ocean thermal data of White (1995) are also used. Further, to examine upper-ocean conditions, we use the gridded climatologies of temperature and salinity profiles from the World Ocean Atlas 1998 (WOA98: Boyer et al. 1998). Since these data are given at standard depth levels, the interpolated data of 10 m by the Akima (1970) scheme are prepared.
To represent the winter atmospheric forcing field in the central North Pacific, we use the North Pacific index (NPI: Trenberth and Hurrell 1994), which is an average of sea level pressure (SLP) in the region of 30°–65°N, 160°E–140°W, where the region of maximum SLP variance in the North Pacific in winter is covered. The NPI is a measure of the strength of the Aleutian low during the Northern Hemisphere winter. In the present study, the NPI is averaged from December to February and is normalized by the standard deviations.
Ishi and Hanawa (2005) investigated large-scale variability of the wintertime wind stress curl field in the North Pacific and discussed its relation to atmospheric teleconnection patterns. They showed that the leading EOF mode of wintertime wind stress curl anomalies had a high correlation with the NPI. In addition, the simultaneous correlation coefficient between the leading mode and the Sverdrup transport representing westerly boundary currents of the wind-driven circulation was very high. Therefore, they concluded that the NPI is a good indicator of spinup or spindown in both the subtropical gyre and subpolar gyre. Based on Ishi and Hanawa’s result, in the present study, a 5-yr running mean of the NPI is used as the measure of spinup/spindown of the North Pacific subtropical gyre, and we call it the “spinup/spindown” index in the present study. Figure 1b shows time series of this index. Here the sign of the spinup/spindown index is reversed from that of the original NPI to easily interpret strengthening or weakening of the subtropical gyre. That is, the positive (negative) spinup/spindown index means spinup (spindown) of the subtropical gyre. Naturally, as noted by Ishi and Hanawa, the real ocean responds to atmospheric forcing with some time delay due to oceanic baroclinic adjustment, such as the propagation of baroclinic Rossby waves. Actual time lag of this delayed response will be discussed in a later section.
To calculate buoyancy flux, the data of net surface heat flux, evaporation minus precipitation (E − P) flux, and wind stress obtained from the 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-40: Simmons and Gibson 2000) are used.
In performing a correlation analysis, significance levels are determined using the degrees of freedom calculated by the Davis (1976) method. That is, the degrees of freedom can be obtained from the data length divided by the integral time scale in the combined autocorrelation functions of the compared time series.
3. Period-dependent nature of the occurrence of remote reemergence
Using the 21-yr-long SST dataset from 1982 to 2002, SH05 discovered the existence of the remote reemergence of NPSTMW. In this section, we investigate more detailed characteristics of remote reemergence using much longer datasets of 74 years from 1930 to 2003.
To detect collocated and remote reemergence areas, following HS04 and SH05, lag correlation analyses are done using the reference month of February, when the mixed layer in the NPSTMW formation area is deepest throughout a year (see HS04). Further, in the next subsection, the remote reemergence area is a priori set as the reemergence area (RA) region (29°–35°N, 175°E–175°W), following SH05. For neatness, we call SST averaged in the RA region in February “RA-SST.”
a. Running correlation analysis between RA-SST and CLT
To explore a nature of occurrence of remote reemergence, first, a “running” correlation analysis between RA-SST for a given winter and the CLT (Fig. 1a) in the previous winter is performed using a window of 15 yr. A correlation coefficient for a given year is calculated using the data of ±7 yr from the given year. That is, for instance, a correlation coefficient of the 1987 RA-SST is calculated using the data for 15 years from 1980 to 1994. Then, the reference year is shifted by one year and, finally, a correlation coefficient for every year is calculated in the same way. Based on these calculations, a time series of correlation coefficient is obtained from 1937 to 1996.
Figure 2 shows the time series of correlation coefficients from two SST datasets (ERSST and GISST). Both time series are quite similar, and a period-dependent nature of occurrence of remote reemergence is clearly seen. That is, there are two periods: an occurrence period showing a significant correlation and a nonoccurrence period not showing a significant correlation. In the time series using the ERSST dataset, there are three periods with high positive correlation coefficients exceeding the 90% significance level: 1944–48, 1969–72, and 1984–91. It is found that the period from 1982 to 2002 treated by SH05 corresponds almost to the third occurrence period. On the other hand, from 1974 to 1982, the correlation coefficient is approximately zero. That is, during this period reemergence signals never appear: RA-SST variation is independent from CLT variation. In addition, in the late 1950s, RA-SST and CLT are negatively correlated.
To examine whether low-frequency running correlations displayed in Fig. 2 are merely due to noise in the time series, we performed the statistical test proposed by Gershunov et al. (2001). As a result, we found that the obtained time series of the correlation coefficient is significant and meaningful. Therefore, the relationship between RA-SST and CLT showing a period-dependent nature characterized by low frequency is robust.
b. Further evidence of the period-dependent nature of remote reemergence
In the previous subsection, a correlation analysis was done between CLT and RA-SST anomalies in the reemergence area fixed a priori. To confirm this result, and to check the possibility whether or not the remote reemergence area moves, two types of lag correlation analyses are made. The first is a lag correlation analysis between averaged SST anomalies in the NPSTMW formation area and SST anomalies at each grid point in the North Pacific. The second is a lag correlation analysis between CLT and SST anomalies at each grid point.
Please note that, as shown in Hanawa and Kamada (2001) and Abe (2004), although CLT anomalies are correlated to some degree with winter SST anomalies in the NPSTMW formation area, CLTs are not necessarily consistent with winter SSTs in the formation area. This is because CLTs tend to be the lowest temperature (actually highest density) in winter, while SSTs are monthly or seasonal mean values. In addition, the CLT and SST datasets are almost independent of each other since most of the data sources are those based on different observations. Therefore, two correlation analyses made in this subsection will give independent evidence on the period-dependent nature of occurrence of remote reemergence.
To conduct these analyses, based on the result in the previous subsection, we divide the analysis period into five 15-yr-long periods as follows: period I (1940–54), period II (1950–64), period III (1960–74), period IV (1970–84), and period V (1980–94). Five 15-yr-long periods roughly correspond to each of the occurrence and nonoccurrence periods found in the previous subsection: the former is periods I, III, and V and the latter is periods II and IV.
1) Lag correlation analyses between SST anomalies in NPSTMW formation area and those at each grid point in the North Pacific
Based on the results of HS04 and SH05, and referring to the related previous studies (e.g., Hanawa and Hoshino 1988; Suga and Hanawa 1990; Hanawa and Yoritaka 2001), we set the NPSTMW formation area as the region of 28°–34°N, 140°–160°E. Figure 3 shows the regions where correlation coefficients between SST anomalies averaged within the NPSTMW formation area in February and SST anomalies at each grid point in the following February exceed the 90% significance level. It is seen that two SST datasets give quite similar results. In periods I, III, and V, reemergence areas are more or less detected around the central North Pacific and are situated at almost the same location as the RA region (29°–35°N, 175°E–175°W) found in SH05. On the other hand, in periods II and IV, significant reemergence signals are never detected anywhere in the North Pacific in both SST datasets. These results confirm that of the analysis made in the previous subsection.
2) Correlation analyses between CLT and SST anomalies at each grid point
To further confirm a period-dependent nature of occurrence of remote reemergence, we perform another lag correlation analysis between anomalies of CLT and SST at each grid point. Figure 4 shows distributions of correlation coefficients for periods I through V. Left panels show the areas where the correlation coefficients between anomalies of SST in February and CLT with no time lag exceed the 90% significance level. Although those areas are different in shape and size, in gross geographical view, the distributions of significant correlation coefficients correspond well to the formation area of NPSTMW at every period (e.g., Hanawa and Talley 2001). In periods II and IV, significant correlation areas are distributed in narrower region compared with those in the other periods. This may manifest the lower formation rate of NPSTMW in these two periods.
Right panels show results of the lag correlation analysis between anomalies of SST at each grid point in February of the next year (i.e., with the lead − lag of +1 yr for CLT) and CLT. In periods I, III, and V, the areas with high correlation coefficients are distributed around the central North Pacific and are situated at almost the same location as the RA region detected by SH05, and as those shown in Fig. 3. On the contrary, in periods II and IV, a reemergence signal is never detected. This result is also consistent with that of the analysis made in the previous subsection.
We further examined relationships between CLT and SST in the subsequent summer (with the lead − lag of +6 months for CLT), and SST in the previous winter (with the lag of −1 yr for CLT). In both cases, a significant region could not be detected (not shown here). That is, these facts mean that reemergence signals are surely set in the NPSTMW formation area in the previous winter (e.g., HS04; SH05) and the signals disappear at the sea surface during the subsequent summer; then the thermal anomalies preserved in the subsurface reemerge in the central North Pacific (see SH05). As a whole, it can be said that the analyses made in this subsection also confirm the period-dependent occurrence of remote reemergence.
4. Causes of period-dependent occurrence of remote reemergence
Based on the two examinations made in the previous subsection, we can say that the period-dependent occurrence of remote reemergence is very robust. In this section, we seek the causes that are responsible for this period-dependent occurrence of remote reemergence of NPSTMW.
a. Relationship between remote reemergence and spinup/spindown of the subtropical gyre
As mentioned in the introduction, Ishi and Hanawa (2005) showed that both the subtropical gyre and subpolar gyre were forced to spin up or spin down by Aleutian low activity, namely by the leading EOF mode of wintertime wind stress curl anomalies. Therefore, we regarded the 5-yr running mean of the NPI as an index of spinup/spindown of the subtropical gyre. Here we examine the relationship between the spinup/spindown index and the period-dependent occurrence of remote reemergence. Figure 5 shows the lag correlation coefficient between the spinup/spindown index shown in Fig. 1a and the time series of the running correlation coefficient shown in Fig. 2. The lag correlation coefficient takes maximum values at the lag of 6 to 8 yr (i.e., the spinup/spindown index leads 6 to 8 yr). Therefore, it is suggested that period-dependent occurrence of remote reemergence is closely related to the spinup/spindown of the subtropical gyre, eventually to the Aleutian low activity, with a delay time of 6 to 8 yr. That is, when the subtropical gyre begins to spin up (spin down) owing to the strengthening (weakening) of the Aleutian low in winter, remote reemergence begins (ceases) to function with a delay time of about 7 yr.
As mentioned in the introduction, Hanawa and Kamada (2001) and Abe (2004) showed that the CLT has a significant correlation with the NPI at lag 5 to 7 yr. Considering the delay time of 1 yr for remote reemergence process, it can be said that the delay time of 6 to 8 yr between occurrence of remote reemergence and NPI (actually spinup/spindown index) found in the present study is plausible. Therefore, we may roughly regard the period of positive (negative) phases of the spinup/spindown index with lag 7 yr as the occurrence (nonoccurrence) period. Henceforth, we call this period the R (nonR).
The specific period of the R and nonR in the study interval is as follows: Period R is 1932–37, 1945–53, 1967–70, 1981–95, and 2000–03 (ending year of the analysis period), while period nonR is 1930 (beginning year of the analysis period) to 1931, 1938–44, 1954–66, 1971–80, and 1996–99. Note that 1930 (2003) is the beginning (ending) year of this study period, and the actual beginning (ending) year would be different from that. It is interesting to note that the mean duration of periods R and nonR is 8.5 yr. Therefore, the time period needed to close one cycle from period R to period nonR is roughly estimated as 17 yr on an average. This time length of 17 yr is consistent with the previous findings on decadal-scale variation in the North Pacific (e.g., Mantua et al. 1997: Minobe 1997; White and Cayan 1998).
In the next subsection, in order to investigate why remote reemergence has a period-dependent nature, we will compare the upper-oceanic conditions in periods R and nonR by composite analyses.
b. Ocean initial stratification
First, in order to trace the signals of SST anomalies set in the surface mixed layer in winter, as made in Fig. 4 of SH05, a lag correlation analysis is performed using the upper-ocean thermal dataset of White (1995). Although the dataset of White covers from 1955, it is known that the source data are very limited before 1970. Therefore, only data after 1970 are used in the following analysis.
Figures 6a and 6b show cross sections of correlation coefficients along the latitudinal lines of 28°, 30°, 32°, and 34°N in periods R and nonR, respectively. Layers having significant correlation coefficients for the reference temperature averaged from the sea surface to 20 m within the NPSTMW formation area in February are shaded. Mixed layer depth (MLD) and isothermal lines averaged in each of the periods are superposed. Here, MLD is defined as the depth where the temperature is 1.0°C less than at the sea surface, following Hanawa and Hoshino (1988).
Figure 6a for period R shows that the high correlation layers move eastward as seasons go by. In the warming season (the lag of 6 months), these layers are isolated in the subsurface and then the layers outcrop to the sea surface around 180° in the next winter. This outcrop region is situated at almost the same location as the RA region. This feature is just the manifestation of occurrence of remote reemergence in period R. On the other hand, Fig. 6b for period nonR shows that the high correlation layers never recur at the sea surface in the next winter, although thin and narrow layers are isolated in the subsurface in warming season. That is, for period nonR, this lag correlation analysis indicates that signals formed in one winter are not preserved in the subsurface throughout a year.
In terms of oceanic initial stratification such as MLD, upper-ocean heat content (OHC), integrated buoyancy flux loss, and strength of mixing we investigate differences of background conditions in the subsurface between periods R and nonR.
Figure 7 shows distributions of mean vertical temperature gradients meridionally averaged from 30° to 34°N for (a) period R and (b) period nonR in winter and summer; MLDs are also displayed. Layers having a vertical temperature gradient less than 1.5 × 10−2 °C m−1 in summer cross sections correspond well to those with the high correlation layers shown in Fig. 6. That is, it can be said that high correlation layers represent layers of NPSTMW. In period R, it is found that the NPSTMW is thick (see Fig. 6a) and widely distributed in latitude and longitude (see also Fig. 4). Therefore, based on the result shown in Figs. 4, 6 and 7, one may speculate that, in period R, a large amount of NPSTMW is formed, the NPSTMW voluminously moves eastward in subsurface and then recurs at the sea surface one year later. This suggests that the difference in “volume” of a water mass is a key factor controlling occurrence of reemergence.
Here, it is worthwhile to note that at the lag 12 months in Fig. 6, we can observe water masses having a significant correlation different from the ones occupying the subsurface in the cross sections of 28° and 30°N in period R. This might be the manifestation that some NPSTMW enters deeper layers through subduction, as already pointed out by Suga and Hanawa (1995). Since this signal is not seen in period nonR, it is considered that the subduction rate in period R is greater than that in period nonR. That is, the subduction rate has decadal variability associated with Aleutian low activity, as Xie et al. (2000) pointed out for the North Pacific central mode water using an ocean general circulation model.
Figure 8 shows a map of winter mean OHC difference between periods R and nonR. The OHC is calculated from the upper-ocean thermal dataset of White (1995) and is defined as vertically averaged temperature from the sea surface to the depth 300 m as a proxy of OHC, following Hasegawa and Hanawa (2003). During period R, positive OHC anomalies appear in the subtropical region. In particular, large positive anomalies are situated in the central North Pacific. Here, positive OHC anomalies equivalently mean the existence of a deeper thermocline. Indeed, as shown in Fig. 7, the winter MLD in period R is from 20 to 30 m deeper than that in period nonR. Therefore, in period R, it is considered that a large amount of NPSTMW is formed due to the deeper winter mixed layer. This observational result of deeper winter MLD in the western North Pacific is consistent with the result of a coupled ocean–atmosphere model that the thermocline deepens due to propagation of a baroclinic Rossby wave associated with Aleutian low activity (e.g., Schneider et al. 2002). In addition, we evaluated the temperature change of the winter mixed layer due to net heat flux in the RA region. The evaluated temperature difference between period R and period nonR was −0.09°C. This value is much smaller than the OHC difference (0.50°C). This implies that deeper MLD and positive OHC anomalies in the central North Pacific are mainly due, not to atmospheric forcing, but to oceanic processes such as oceanic heat transport, as pointed out in the Kuroshio Extension region by Qiu (2000) and Vivier et al. (2002).
Sugimoto and Hanawa (2005a) have proposed that the buoyancy flux was one of the important factors for an occurrence of reemergence. In terms of the buoyancy flux integrated from the month of deepest MLD to 12 months later, they have pointed out that an occurrence of reemergence was classified into three types: “normal type” (reemerge around 1 yr later), “early type” (reemerge around 7 to 9 months later), and “nonoccurrence type” (no reemergence).
Next, in order to examine the strength of mixing, we calculate the Turner angle defined by arctan(α∂T/∂Z − β∂S/∂Z, α∂T/∂Z + β∂S/∂Z), where T is potential temperature, S is salinity, α is the thermal expansion coefficient, and β is the saline contraction coefficient. The Turner angle can be regarded as a measure of the potential for double diffusion within a water column (Turner 1973; Ruddick 1983). Figure 10 shows a cross section of the Turner angle meridionally averaged from 30° to 34°N, using summer-mean temperature and salinity profiles of WOA98. Around 200 m, the Turner angle is greater than 65° from 160°E to 170°W. The large Turner angle strongly suggests the presence of salt-finger-type convection during summer that tends to diffuse the temperature anomalies in the subsurface. As a climatological feature, the layer having large Turner angle is situated in the lower part of the layer with high correlation in summer (see Figs. 6a,b). Therefore, the following speculation on persistence of thermal anomalies in the subsurface would be expected to hold. Since a smaller amount of NPSTMW is formed during period nonR, the reemergence signals gradually disappear due to vigorous salt-finger-type convection in the lower portions of a water mass, and consequently reemergence cannot occur. On the other hand, during period R, since a large amount of NPSTMW is formed in subsurface, some part of the water mass is preserved from strong mixing and therefore reemergence can occur.
To summarize, it is found that the RA region in period R (period nonR) has more favorable (unfavorable) conditions for the occurrence of reemergence, in terms of background conditions such as formation rate of NPSTMW, winter MLD, winter OHC, and integrated buoyancy flux (see Table 1).
5. Impact of remote reemergence on winter SSTs in the central North Pacific
In this section, we quantitatively estimate how remote reemergence of NPSTMW impacts SST variability in the central North Pacific. It is known that the RA region is the region where the first EOF mode for SST anomalies in the North Pacific shows decadal variation, described as the Pacific decadal oscillation (e.g., Mantua et al. 1997). The Aleutian low is located over this area in winter, and its activity also shows decadal-scale variation (e.g., Trenberth and Hurrell 1994; Ishi and Hanawa 2005).
Previous studies (e.g., Nakamura et al. 1997) show that the SST variation in the central North Pacific including the RA region is related to the Aleutian low activity to some degree. Actually, the RA-SST of the ERSST dataset has a simultaneous correlation coefficient exceeding the 99% significance level with a single regression model obtained from NPI (0.58; see Fig. 11a). This means that surface thermal forcing due to the Aleutian low represented by NPI controls SST variation in winter to some degree. Therefore, in order to quantitatively elucidate the impact of remote reemergence on the RA-SST, a multiple regression analysis is performed using two explanatory variables of NPI and CLT, normalized by unit standard deviation. Here, since the correlation coefficient between NPI and CLT in the previous year is only −0.04, NPI and CLT can be regarded as independent variables.
In period R, since C1 and C2 obtained from the ERSST dataset (GISST dataset) are 0.32 (0.33) and 0.40 (0.27), respectively, the impact of CLT is almost the same degree as that of NPI. On the other hand, in period nonR, they are 0.01 (0.10) and 0.45 (0.40) using the ERSST dataset (GISST dataset), respectively. That is, C1 in period nonR is much smaller than that in period R. Therefore, it can be said that for the RA-SST, remote reemergence has an impact equivalent to that with NPI during period R, while there is no significant impact during period nonR.
Figures 11b and 11c show time series of observed RA-SSTs and modeled RA-SSTs for the ERSST and GISST datasets, respectively. Multiple correlation coefficients between modeled RA-SST and observed RA-SST are 0.78 (ERSST dataset) and 0.74 (GISST dataset), both of which well exceed the 99% significance level. Therefore, it can be said that this multiple regression model can well reproduce the observed RA-SSTs, using CLT and the surface thermal forcing represented by NPI. Furthermore, since the variance of the modeled SSTs accounts for about 60% of total variance of the observed one, it can be pointed out that CLT and NPI are major components controlling SST field in the central North Pacific.
6. Summary
In the present study, we investigated how remote reemergence of NPSTMW impacts winter SST variation in the central North Pacific. Since remote reemergence was also detected using the CLT dataset recently prepared by Abe (2004), the present result confirmed the finding on remote reemergence by SH05. It was found that occurrence of remote reemergence strongly depended on a specific time period, and the background conditions, such as formation rate of mode water, winter MLD, winter OHC, and integrated buoyancy flux, were very important factors in determining the period-dependent remote reemergence. During the occurrence periods, the reemergence area was situated at almost the same location as the remote reemergence region detected by SH05, and this remote reemergence had an impact equivalent with Aleutian low activity on winter SST variation in the central North Pacific. On the other hand, during nonoccurrence periods, there was no significant contribution of remote reemergence to the SST field in the central North Pacific. A multiple regression analysis could reproduce winter SST variation in the remote reemergence area well, using two explanatory variables of CLT and NPI. Modeled SST variance accounted for most of the observed SST variance in the remote reemergence area.
Acknowledgments
The authors wish to express their sincere thanks to members of Physical Oceanography Group at Tohoku University for their fruitful comments and heartfelt encouragements. Three anonymous reviewers gave useful comments. The first author (SS) was financially supported by the 21st Century Center-of-Excellence (COE) Program, “Advanced Science and Technology Center for the Dynamic Earth (E-ASTEC),” at Tohoku University. This study was done as part of the second author’s (KH) research program supported by Grants-in-Aid for Scientific Research by the Ministry of Education, Science, Sports and Culture.
REFERENCES
Abe, M., 2004: Temporal variation of the core layer temperature of North Pacific subtropical mode water using newly archived hydrographic dataset (in Japanese). M.S. thesis, Department of Geophysics, Tohoku University, 66 pp.
Akima, H., 1970: A new method of interpolation and smooth curve fitting based on local procedures. J. Assoc. Comput. Methods, 17 , 589–603.
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. Progress in Oceanography, Vol. 49, Pergamon, 41–61.
Boyer, T. P., S. Levitus, J. Antonov, M. Conkright, T. O’Brien, and C. Stephens, 1998: Salinity of the Pacific Ocean. Vol. 5, World Ocean Atlas 1998, NOAA Atlas NESDIS 31, 166 pp.
Coëtlogon, G., and C. Frankignoul, 2003: The persistence of winter sea surface temperature in the North Atlantic. J. Climate, 16 , 1364–1377.
Conkright, M. E., and Coauthors, 2002: Introduction. Vol. 1, World Ocean Database 2001, S. Levitus, Ed., NOAA Atlas NESDIS 42, 167 pp.
Davis, R., 1976: Predictability of sea surface temperature and sea level pressure anomalies over the North Pacific Ocean. J. Phys. Oceanogr., 6 , 249–266.
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.
Deser, C., M. A. Alexander, and M. S. Timlin, 2003: Understanding the persistence of sea surface temperature anomalies in midlatitudes. J. Climate, 16 , 57–72.
Ebbesmeyer, C. C., D. R. Cayan, D. R. McLain, F. H. Nichols, D. H. Peterson, and K. T. Redmond, 1991: 1976 step in the Pacific climate: Forty environmental changes between 1968–75 and 1977–84. Proc. Seventh Annual Pacific Climate Workshop, Interagency Ecology Studies Program Rep. 26, Pacific Grove, CA, California Department of Water Resources, 115–126.
Gershunov, A., N. Schneider, and T. Barnet, 2001: Low-frequency modulation of the ENSO–Indian monsoon rainfall relationship: Signal or noise? J. Climate, 14 , 2486–2492.
Hanawa, K., and I. Hoshino, 1988: Temperature structure and mixed layer in the Kuroshio region over the Izu Ridge. J. Mar. Res., 46 , 683–700.
Hanawa, K., and J. Kamada, 2001: Variability of core layer temperature (CLT) of North Pacific subtropical mode water. Geophys. Res. Lett., 28 , 2229–2232.
Hanawa, K., and L. D. Talley, 2001: Mode waters. Ocean Circulation and Climate, G. Siedler, J. Church, and J. Gould, Eds., Academic Press, 373–386.
Hanawa, K., and H. Yoritaka, 2001: North Pacific subtropical mode water observed in long cross section along 32.5°N line. J. Oceanogr., 57 , 679–692.
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.
Hasegawa, T., and K. Hanawa, 2003: Heat content variability related to ENSO events in the Pacific. J. Phys. Oceanogr., 33 , 407–421.
Ishi, Y., and K. Hanawa, 2005: Large-scale variability of wintertime wind stress curl field in the North Pacific and their relation to atmospheric teleconnection patterns. Geophys. Res. Lett., 32 .L10607, doi:10.1029/2004GL022330.
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.
Mantua, N. J., and S. R. Hare, 2002: The Pacific decadal oscillation. J. Oceanogr., 58 , 35–44.
Mantua, N. J., S. R. Hare, Y. Zhang, J. M. Wallace, and R. Francis, 1997: A Pacific interdecadal climate oscillation with impacts on salmon production. Bull. Amer. Meteor. Soc., 78 , 1069–1079.
Miller, A. J., and N. Schneider, 2000: Interdecadal climate regime dynamics in the North Pacific Ocean: Theories, observations, and ecosystem impacts. Progress in Oceanography, Vol. 27, Pergamon, 257–260.
Miller, A. J., J. F. Chai, S. Chiba, J. R. Moisan, and D. J. Neilson, 2004: Decadal-scale climate and ecosystem interactions in the North Pacific Ocean. J. Oceanogr., 60 , 163–188.
Minobe, S., 1997: A 50–70 year climatic oscillation over the North Pacific and North America. Geophys. Res. Lett., 24 , 683–686.
Minobe, S., 2000: Spatio-temporal structure of the pentadecadal variability over the North Pacific. Progress in Oceanography, Vol. 47, Pergamon, 381–408.
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.
Namias, J., and R. M. Born, 1970: Temporal coherence in North Pacific sea-surface temperature patterns. J. Geophys. Res., 75 , 5952–5955.
Namias, J., and R. M. Born, 1974: Further studies of temporal coherence in North Pacific sea surface temperature. J. Geophys. Res., 79 , 797–798.
Qiu, B., 2000: Interannual variability of the Kuroshio Extension system and its impact on the wintertime SST field. J. Phys. Oceanogr., 30 , 1486–1502.
Qiu, B., 2003: Kuroshio Extension variability and forcing of the Pacific decadal oscillation: Responses and potential feedback. J. Phys. Oceanogr., 33 , 2465–2482.
Rayner, N. A., E. B. Horton, D. E. Parker, C. K. Folland, and R. B. Hackett, 1996: Version 2.3b of the Global Sea-Ice and Sea Surface Temperature data set 1903–1994. Climate Research Tech. Note 74, 35 pp.
Ruddick, B. R., 1983: A practical indicator of the stability of the water column to double diffusive activity. Deep-Sea Res., 30A , 1105–1107.
Schneider, N., A. J. Miller, and D. W. Pierce, 2002: Anatomy of North Pacific decadal variability. J. Climate, 15 , 586–605.
Seager, R., Y. Kunshnir, N. Naik, M. A. Cane, and J. A. 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.
Simmons, A. J., and J. K. Gibson, 2000: The ERA-40 project plan. Tech. Rep. ERA-40, European Centre for Medium-Range Weather Forecasts, Reading, United Kingdom, 63 pp.
Smith, T. M., and R. W. Reynolds, 2004: Improved extended reconstruction of SST (1854–1997). J. Climate, 17 , 2466–2477.
Suga, T., and K. Hanawa, 1990: The mixed-layer climatology in the northwestern part of the North Pacific subtropical gyre and the formation area of Subtropical Mode Water. J. Mar. Res., 48 , 543–566.
Suga, T., and K. Hanawa, 1995: The subtropical mode water circulation in the North Pacific. J. Phys. Oceanogr., 25 , 958–970.
Sugimoto, S., and K. Hanawa, 2005a: Why does reemergence of winter sea surface temperature anomalies not occur in eastern mode water areas? Geophys. Res. Lett., 32 .L15608, doi:10.1029/2005GL022968.
Sugimoto, S., and K. Hanawa, 2005b: Remote reemergence areas of winter sea surface temperature anomalies in the North Pacific. Geophys. Res. Lett., 32 .L01606, doi:10.1029/2004GL021410.
Timlin, M. S., M. A. Alexander, and C. Deser, 2002: On the reemergence of North Atlantic SST anomalies. J. Climate, 15 , 2707–2712.
Trenberth, K. E., and J. W. Hurrell, 1994: Decadal atmosphere-ocean variations in the Pacific. Climate Dyn., 9 , 303–319.
Turner, J. S., 1973: Buoyancy Effects on Fluids. Cambridge University Press, 367 pp.
Vivier, F., K. Kelly, and L. Thompson, 2002: Heat budget in the Kuroshio Extension region: 1993–99. J. Phys. Oceanogr., 32 , 3436–3454.
Watanabe, M., and M. Kimoto, 2000: On the persistence of decadal SST anomalies in the North Atlantic. J. Climate, 13 , 3017–3028.
White, W. B., 1995: Design of a global observing system for gyre-scale upper ocean temperature variability. Progress in Oceanography, Vol. 36, Pergamon, 169–217.
White, W. B., and D. R. Cayan, 1998: Quasi-periodicity and global symmetries in interdecadal upper ocean temperature variability. J. Geophys. Res., 103 , 21335–21354.
Xie, S. P., M. Nonaka, and S. Hosoda, 2000: Interdecadal thermocline variability in the North Pacific for 1958–97: A GCM simulation. J. Phys. Oceanogr., 30 , 2798–2813.
Yasuda, T., and Y. Kitamura, 2003: Long-term variability of North Pacific subtropical mode water in response to spinup of the subtropical gyre. J. Oceanogr., 59 , 279–290.
Time series (a) of the CLT by Abe (2004) and (b) of the spinup/spindown index, defined as 5-yr running average of the NPI but sign reversed. See the text.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Time series (a) of the CLT by Abe (2004) and (b) of the spinup/spindown index, defined as 5-yr running average of the NPI but sign reversed. See the text.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Time series (a) of the CLT by Abe (2004) and (b) of the spinup/spindown index, defined as 5-yr running average of the NPI but sign reversed. See the text.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Time series obtained by a running correlation analysis between RA-SST and CLT in the previous year, using two SST datasets. For example, 1987 value shown by the cross represents a correlation coefficient between RA-SST from 1980 to 1994 and CLT from 1979 to 1993. Broken and dashed horizontal lines denote the 99% and 90% significance levels, respectively.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Time series obtained by a running correlation analysis between RA-SST and CLT in the previous year, using two SST datasets. For example, 1987 value shown by the cross represents a correlation coefficient between RA-SST from 1980 to 1994 and CLT from 1979 to 1993. Broken and dashed horizontal lines denote the 99% and 90% significance levels, respectively.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Time series obtained by a running correlation analysis between RA-SST and CLT in the previous year, using two SST datasets. For example, 1987 value shown by the cross represents a correlation coefficient between RA-SST from 1980 to 1994 and CLT from 1979 to 1993. Broken and dashed horizontal lines denote the 99% and 90% significance levels, respectively.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Distribution of lag correlation coefficients between SST anomalies averaged within the formation area of NPSTMW indicated by the box in February, and SST anomalies at each grid point in February of the next year, for five periods: (a) I (1940–54), (b) II (1950–64), (c) III (1960–74), (d) IV (1970–84), and (e) period V (1980–94). Shading denotes the results of ERSST, and dark (light) shading indicates the region where the correlation coefficient exceeds the 95% (90%) significance level. The 95% (90%) significance level is 0.55 (0.45). Contours denote the results of the GISST dataset, and thick (thin) contour indicates the region where the correlation coefficient exceeds the 95% (90%) significance level.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Distribution of lag correlation coefficients between SST anomalies averaged within the formation area of NPSTMW indicated by the box in February, and SST anomalies at each grid point in February of the next year, for five periods: (a) I (1940–54), (b) II (1950–64), (c) III (1960–74), (d) IV (1970–84), and (e) period V (1980–94). Shading denotes the results of ERSST, and dark (light) shading indicates the region where the correlation coefficient exceeds the 95% (90%) significance level. The 95% (90%) significance level is 0.55 (0.45). Contours denote the results of the GISST dataset, and thick (thin) contour indicates the region where the correlation coefficient exceeds the 95% (90%) significance level.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Distribution of lag correlation coefficients between SST anomalies averaged within the formation area of NPSTMW indicated by the box in February, and SST anomalies at each grid point in February of the next year, for five periods: (a) I (1940–54), (b) II (1950–64), (c) III (1960–74), (d) IV (1970–84), and (e) period V (1980–94). Shading denotes the results of ERSST, and dark (light) shading indicates the region where the correlation coefficient exceeds the 95% (90%) significance level. The 95% (90%) significance level is 0.55 (0.45). Contours denote the results of the GISST dataset, and thick (thin) contour indicates the region where the correlation coefficient exceeds the 95% (90%) significance level.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
As in Fig. 3 but for distributions of correlation coefficients between anomalies of CLT and (left) SST in February, and (right) SST in February of the next year, for five periods: (a) I, (b) II, (c) III, (d) IV, and (e) V.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
As in Fig. 3 but for distributions of correlation coefficients between anomalies of CLT and (left) SST in February, and (right) SST in February of the next year, for five periods: (a) I, (b) II, (c) III, (d) IV, and (e) V.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
As in Fig. 3 but for distributions of correlation coefficients between anomalies of CLT and (left) SST in February, and (right) SST in February of the next year, for five periods: (a) I, (b) II, (c) III, (d) IV, and (e) V.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Lag-correlation analysis between time series of a running correlation analysis shown in Fig. 2 and the spinup/spindown index shown in Fig. 1b. Three lines denote the 99%, 95%, and 90% significance levels.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Lag-correlation analysis between time series of a running correlation analysis shown in Fig. 2 and the spinup/spindown index shown in Fig. 1b. Three lines denote the 99%, 95%, and 90% significance levels.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Lag-correlation analysis between time series of a running correlation analysis shown in Fig. 2 and the spinup/spindown index shown in Fig. 1b. Three lines denote the 99%, 95%, and 90% significance levels.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Cross sections of correlation coefficients between the reference temperature averaged from the sea surface to 20 m within the NPSTMW formation area in February and the subsurface temperatures, at the lags 0, 6, and 12 months, along the latitudinal lines of (from left to right) 28°, 30°, 32°, and 34°N for (a) period R and (b) period nonR. The darkly and lightly shaded areas denote the layers having lag correlation coefficients exceeding the 95% and 90% significance levels, respectively. Here, the 95% (90%) significance level is 0.60 (0.52). Temperatures are superposed and the thick solid line denotes MLD.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Cross sections of correlation coefficients between the reference temperature averaged from the sea surface to 20 m within the NPSTMW formation area in February and the subsurface temperatures, at the lags 0, 6, and 12 months, along the latitudinal lines of (from left to right) 28°, 30°, 32°, and 34°N for (a) period R and (b) period nonR. The darkly and lightly shaded areas denote the layers having lag correlation coefficients exceeding the 95% and 90% significance levels, respectively. Here, the 95% (90%) significance level is 0.60 (0.52). Temperatures are superposed and the thick solid line denotes MLD.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Cross sections of correlation coefficients between the reference temperature averaged from the sea surface to 20 m within the NPSTMW formation area in February and the subsurface temperatures, at the lags 0, 6, and 12 months, along the latitudinal lines of (from left to right) 28°, 30°, 32°, and 34°N for (a) period R and (b) period nonR. The darkly and lightly shaded areas denote the layers having lag correlation coefficients exceeding the 95% and 90% significance levels, respectively. Here, the 95% (90%) significance level is 0.60 (0.52). Temperatures are superposed and the thick solid line denotes MLD.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Cross sections of (top) winter and (bottom) summer mean vertical temperature gradients [°C (100 m)−1] meridionally averaged from 30° to 34°N, for (a) period R and (b) period nonR. Temperatures are superposed. Shading indicates the layers where the mean vertical temperature gradient is less than 1.5 [°C (100 m)−1]. Contour indicates temperature, interval: 1.0°C. The thick solid and dashed lines denote MLD for periods R and nonR, respectively.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Cross sections of (top) winter and (bottom) summer mean vertical temperature gradients [°C (100 m)−1] meridionally averaged from 30° to 34°N, for (a) period R and (b) period nonR. Temperatures are superposed. Shading indicates the layers where the mean vertical temperature gradient is less than 1.5 [°C (100 m)−1]. Contour indicates temperature, interval: 1.0°C. The thick solid and dashed lines denote MLD for periods R and nonR, respectively.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Cross sections of (top) winter and (bottom) summer mean vertical temperature gradients [°C (100 m)−1] meridionally averaged from 30° to 34°N, for (a) period R and (b) period nonR. Temperatures are superposed. Shading indicates the layers where the mean vertical temperature gradient is less than 1.5 [°C (100 m)−1]. Contour indicates temperature, interval: 1.0°C. The thick solid and dashed lines denote MLD for periods R and nonR, respectively.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Winter (January through March) mean OHC (vertically averaged temperature from the sea surface to 300 m) difference between periods R and nonR. Dashed contours are negative, Contour interval: 0.1°C. Dark (light) shading indicates the area where the OHC difference exceeds 0.4°C (0.2°C).
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Winter (January through March) mean OHC (vertically averaged temperature from the sea surface to 300 m) difference between periods R and nonR. Dashed contours are negative, Contour interval: 0.1°C. Dark (light) shading indicates the area where the OHC difference exceeds 0.4°C (0.2°C).
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Winter (January through March) mean OHC (vertically averaged temperature from the sea surface to 300 m) difference between periods R and nonR. Dashed contours are negative, Contour interval: 0.1°C. Dark (light) shading indicates the area where the OHC difference exceeds 0.4°C (0.2°C).
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
As in Fig. 8 but for buoyancy flux integrated from the month of deepest MLD to 12 months later. Positive value indicates oceanic buoyancy gain, Contour interval: 1.0 × 10−6 m2 s−3. Dashed contours are negative (oceanic buoyancy loss). Dark (light) shading indicates the area where the integrated buoyancy fluxes are less than −0.4 × 10−6 m2 s−3 (−0.2 × 10−6 m2 s−3).
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
As in Fig. 8 but for buoyancy flux integrated from the month of deepest MLD to 12 months later. Positive value indicates oceanic buoyancy gain, Contour interval: 1.0 × 10−6 m2 s−3. Dashed contours are negative (oceanic buoyancy loss). Dark (light) shading indicates the area where the integrated buoyancy fluxes are less than −0.4 × 10−6 m2 s−3 (−0.2 × 10−6 m2 s−3).
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
As in Fig. 8 but for buoyancy flux integrated from the month of deepest MLD to 12 months later. Positive value indicates oceanic buoyancy gain, Contour interval: 1.0 × 10−6 m2 s−3. Dashed contours are negative (oceanic buoyancy loss). Dark (light) shading indicates the area where the integrated buoyancy fluxes are less than −0.4 × 10−6 m2 s−3 (−0.2 × 10−6 m2 s−3).
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Distribution of the Turner angle in summer (July through September) meridionally averaged from 30° to 34°N, Contour interval: 5.0°. Dark (light) shading shows the layers where the Turner angle exceeds 65.0° (45.0°).
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Distribution of the Turner angle in summer (July through September) meridionally averaged from 30° to 34°N, Contour interval: 5.0°. Dark (light) shading shows the layers where the Turner angle exceeds 65.0° (45.0°).
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Distribution of the Turner angle in summer (July through September) meridionally averaged from 30° to 34°N, Contour interval: 5.0°. Dark (light) shading shows the layers where the Turner angle exceeds 65.0° (45.0°).
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Time series of observed SST anomalies and modeled SST anomalies in the RA region in February using two SST datasets: (a) ERSST and a single regression model, (b) ERSST and a multiple regression model, and (c) GISST and a multiple regression model. Dashed line indicates observed SST anomalies in February, and solid line those of modeled SST.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Time series of observed SST anomalies and modeled SST anomalies in the RA region in February using two SST datasets: (a) ERSST and a single regression model, (b) ERSST and a multiple regression model, and (c) GISST and a multiple regression model. Dashed line indicates observed SST anomalies in February, and solid line those of modeled SST.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Time series of observed SST anomalies and modeled SST anomalies in the RA region in February using two SST datasets: (a) ERSST and a single regression model, (b) ERSST and a multiple regression model, and (c) GISST and a multiple regression model. Dashed line indicates observed SST anomalies in February, and solid line those of modeled SST.
Citation: Journal of Climate 20, 2; 10.1175/JCLI4004.1
Comparison of important factors for occurrence of reemergence between periods R and nonR.
