Rapid Warming in Global Sea Surface Temperature since around 2013

Transcription

1 SOLA, 2017, Vol. 13, 25 30, doi: /sola Rapid Warming in Global Sea Surface Temperature since around 2013 Yusuke Urabe 1, 2, Tamaki Yasuda 1, 2, and Shuhei Maeda 2 1 Climate Prediction Division, Japan Meteorological Agency, Tokyo, Japan 2 Climate Research Department, Meteorological Research Institute, Ibaraki, Japan Abstract Since around 2013, the globally averaged sea surface temperature has rapidly warmed up and reached its highest on record. During this time, there was an intensifying El Niño event that caused positive temperature anomalies in the tropical Pacific Ocean. Compared with the conditions observed in 1997/98, when the previous highest record was marked associated with strong El Niño event, there were notable differences detected in the recent conditions. In the tropical Pacific, remarkable warming near sea surface associated with strong El Niño event in 2015/16 started from significantly warmed conditions along with positive temperature anomaly redistributed from the western part since early 2014, resulting in positive anomalies in the central to eastern part remaining for more than two years, much longer than 1997/98 event. In addition, substantial warming was observed in the North Pacific around 2013 and contribution of the North Pacific region to the global averaged SST anomaly marked significantly large value and was comparable to that of the tropical Pacific. (Citation: Urabe, Y., T. Yasuda, and S. Maeda, 2017: Rapid warming in global sea surface temperature since around SOLA, 13, 25 30, doi: /sola ) 1. Introduction The globally averaged surface temperature exhibits centennial increases related to increasing greenhouse gases and subsequent net energy gain at the top of the atmosphere. The Fifth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC 2013) indicates that ocean warming dominates the global energy change inventory. Recently, the globally averaged sea surface temperature (GASST) is rapidly warming up since around 2013, considerably exceeding the highest record in 2014 and 2015, and the record-breaking anomalies are still observed in early 2016 (Fig. 1a). The increase is accompanied by a strong El Niño event, presenting, at least qualitatively, consistent condition with those indicated by previous studies such as Trenberth et al. (2002). However, the recent increase of GASST and observed anomalies are rather insistent and it is worth examining how much the El Niño event attributed the recent warming of GASST and whether the conditions are similar to or different from those observed when the previous highest record was marked along with the 1997/98 El Niño event. In addition to El Niño in the tropical Pacific, the warming in the North Pacific around 2013 is also remarkable and possibly has some impacts on GASST. In this study, oceanic conditions and related atmospheric variabilities associated with the observed significant increase in GASST are presented and compared with those observed along with the previous highest record of GASST. 2. Data and methods SST and subsurface temperature distributions are obtained via objective analysis (COBE-SST; Ishii et al. 2005) and ocean data assimilation (MOVE-G2; Toyoda et al. 2013) operated by the Japan Meteorological Agency (JMA). The Japanese 55-year Corresponding author: Yusuke Urabe, Japan Meteorological Agency, Otemachi, Chiyoda-Ku, Tokyo , Japan. y_urabe@met. kishou.go.jp. 2017, the Meteorological Society of Japan. Fig. 1. (a) Time series of globally averaged sea surface temperature (GASST) anomalies ( C). The gray, blue, and red lines represent annual (July June) mean GASST anomalies, their five-year running mean, and the long-term linear trend, respectively. The trend is C per century and significant at 95% confidence level according to Student s t-test. (b) Monthly time series of contributions to a GASST anomaly in the tropical Pacific (10 S 10 N) (red line), North Pacific (10 N 60 N) (dark blue line), South Pacific (60 S 10 S) (green line), Atlantic (yellow line), and Indian Ocean (light blue line). Contributions are determined from the areaintegrated SST anomaly in each basin, which is divided by the area of the global ocean domain. (c) Monthly time series of anomaly of NINO-3 SST. Reanalysis (JRA-55) dataset (Kobayashi et al. 2015) is used to investigate atmospheric circulation patterns. Monthly data until July 2016 are used for each dataset. In this study, the SST anomaly averaged in the NINO-3 region (5 S 5 N, 150 W 90 W) is referred to as NINO-3 SST, and its anomaly is used as an indicator of the El Niño-Southern Oscillation (ENSO). Although the time-filtered (for example, three or five months running mean) NINO-3 SST is commonly used as El Niño index, the strict definition of the occurrence of El Niño/La

2 26 Urabe et al., Rapid Warming in Global Sea Surface Temperature Niña events and/or obstruction of monthly-scale variability do not matter to results and discussions of this study. Therefore, simple monthly time-series of NINO-3 SST is used. Climatology is defined as average for the period from 1981 to Unless otherwise noted, anomaly refers to the deviation from the climatology. Considering typical lifecycles of El Niño and La Niña events, annual mean values are defined as average from July to June. Average from July 2014 to June 2015 is termed annual mean value for 2014/15, for example. 3. GASST variability and contribution of the basinaveraged SSTs The GASST exhibits centennial warming trend with decadal time scale fluctuations (Fig. 1a). While the GASST for the period has stayed below the previous highest record of 1997/98, the recent GASST shows the rapid warming after 2013 and continuously breaks the highest record in 2014/15 and 2015/16. Generally similar conditions are identified in other datasets such as HadSST3 (Kennedy et al. 2011a, b) and ERSSTv4 (Huang et al. 2014; Liu et al. 2014) (figure not shown). Figure 1b shows the time series of contributions of SST anomalies integrated in the tropical Pacific (10 S 10 N), North Pacific (10 N 60 N), South Pacific (60 S 10 S), Atlantic, and Indian Oceans to GASST anomalies. The meridional extent of tropical Pacific region is defined as above because SST variability associated with ENSO shows strong signal in 10 S 10 N (figure not shown). These contributions are determined from the area-integrated SST anomaly in each basin, which is divided by the area of the global ocean domain. It should be noted that summation of contribution in these five regions is not equal to GASST anomaly because the Arctic Sea and the Southern Ocean are included in GASST. Historically, the tropical Pacific (red line) has the greatest contribution, and interannual variabilities in the tropical Pacific and Indian Ocean (light blue line) corresponding to El Niño and La Niña events are evident, which are consistent with those reported by previous studies (e.g. Klein et al. 1999). Indeed, the previous highest record was marked in 1997/98 when a strong El Niño event occurred (Figs. 1a and 1c). Compared with 1997/98, the North Pacific (dark blue line), which started to get warmer around 2013, stands out and the South Pacific and Atlantic regions show relatively weak contributions in recent years. The tendencies of area-integrated contribution in each basin and GASST (defined as the difference of annual mean values between the current and previous years) are presented in Table 1. North Pacific region has ten times larger contribution than other basin from 2012/13 to 2013/14, and more than half of tropical Pacific from 2013/14 to 2014/15. The tropical Pacific got warmer associated with that in NINO-3 SST anomaly and the warming in the Indian Ocean is also almost in phase in 2015 (Fig. 1c). The amplitude of increase in the Tropical Pacific (0.090 C) and Indian Ocean (0.054 C) from 2013/14 to 2015/16 (summation of 2014 and 2015 in Table 1) are comparable to those observed from 1996/97 to 1997/98 (0.091 C and C, respectively), which is consistent with the fact that NINO-3 SST anomaly shows comparable amount of increase in each period (regression components to NINO-3 SST are C from 1996/97 to 1997/98 and C from 2013/14 to 2015/16, respectively). The important difference between the two cases is the duration of warming. It persisted for two years or more from 2013/14 to 2015/16, much longer than that from 1996/97 to 1997/98 (Figs. 1b and 1c). These time series indicate that the development of an El Niño condition in the tropical Pacific plays dominant roles in the remarkable increase in GASST from 2014 and precedent warming in the North Pacific contributes to the extremely large GASST anomaly which drastically exceeding that of 1997/98. It should be noted that deviations from the linear trend (difference between gray and red lines in Fig. 1a) in 2015/16 is 0.28 C, which is substantially larger than that in 1997/98 (0.20 C) and any other years in history. The GASST anomaly in 2015/16 is significantly higher Table 1. (a) The tendencies of area-integrated contribution in each basin to GASST anomaly (defined as the difference of annual mean values between the current and previous years; for example, the values in 2014 stand for differences between 2014/15 and 2013/14). Basin segmentation is same as Fig. 1b and TRPAC is tropical Pacific, NPAC is North Pacific, SPAC is South Pacific, IND is Indian Ocean, and ATL is Atlantic, respectively. Units are ( C). (b) Same as (a) for GASST and NINO-3-regression (NINO-3 REG) (see black and red lines in Fig. 4). (a) Annual difference of SST anomaly contribution to global average TRPAC NPAC SPAC IND ATL (b) Annual difference of SST Anomaly GASST NINO-3 REG than the previous records even if the effect of long-term global warming is considered. 4. Observed oceanic and atmospheric conditions Figure 2a shows the distribution of annual mean global SST anomaly in 2013/14. In the tropical Pacific region, positive SST anomalies in the western part and negative SST anomalies from the southern central to eastern part are observed, indicating that La Niña-like condition persists since around 2000 to 2013 as shown in previous studies (Urabe and Maeda 2014; and references therein). In 2014/15 (Fig. 2b), in contrast, SST anomalies turn to positive and increase in the central to eastern part of the tropical Pacific and the amplitude of which is significantly enhanced in 2015/16 (Figs. 2c and 2g). Significant warming tendencies are recognized in the North Pacific around the west coast of North America from 2012/13 to 2014/15 (Figs. 2e and 2f) resulting in remarkable positive anomalies in these region (Figs. 2a and 2b). In addition, continuous warming is found in the Indian Ocean. These spatial patterns of warming are quite consistent with the area-integrated anomalies presented in Fig. 1b. Figures 2i, 2j, 2k, and 2l are same as Figs. 2a, 2b, 2c, and 2d but for vertically averaged temperature (VAT) from the surface to 300-m depth. In the tropical Pacific, positive anomalies observed in the western part in 2013/14 (Fig. 2i), central and eastern parts in 2014/15 (Fig. 2j), and concentrated in the eastern part in 2015/16 (Fig. 2k). In the western Pacific, substantial cooling initially emerged off-equator from near Philippines to the date line, in which positive VAT anomalies observed before 2014 (Figs. 2i and 2j). These variabilities observed in VAT pattern indicates that drastic changes have occurred not only at the sea surface but also in the ocean subsurface, associated with the development of El Niño including eastward propagations of warm Kelvin waves along the equator and westward propagations of cold Rossby waves in the off-equatorial region (Kessler 1990; White et al. 2003; Hasegawa and Hanawa 2003). According to Urabe and Maeda (2014), the warm water in the western Pacific has continuously accumulated in the last decade. In 2015/16, zonally antisymmetric VAT anomaly pattern developed in the tropical Pacific with positive (negative) anomalies in the central to eastern (western) part. These characteristics observed in 2015/16 are similar to those in 1997/98 (Figs. 2d and 2l), note, however, that SST is significantly warmer in the North Pacific in the former (Fig. 2h).

3 SOLA, 2017, Vol. 13, 25 30, doi: /sola Fig. 2. (a) Annual mean SST anomaly in 2013/14 (averaged from July 2013 to June 2014). (b) 2014/15. (c) 2015/16. (d) 1997/98. (e) Difference of annual mean SST anomaly between 2013/ /13, (f) 2014/ /14, (g) 2015/ /15, (h) 2015/ /98. Dots indicate that the anomalies are statistically significant at 95% confidence level according to Student s t-test. (i l) Same as (a d) for vertically averaged temperature (VAT) from sea surface to 300-m depth. Units are ( C).

4 28 Urabe et al., Rapid Warming in Global Sea Surface Temperature Atmospheric forcing is supposed to play a dominant role in mid-latitude regions of the North Pacific, whereby substantial warming had proceeded before an increase in a NINO-3 SST anomaly. The forcing from atmospheric field contributes to increase or decrease of SST between start and end of the period. Therefore, we calculated 12 months running mean of atmospheric anomalies for each month from July 2012 (mean from July 2012 to June 2013; approximately corresponding to contribution to SST difference between July 2012 and July 2013) to June 2013 and averaged them in order to get atmospheric forcing that contributed to the difference between annual mean SST anomalies between 2012/13 and 2013/14. The calculated contribution of sea-level pressure (SLP), wind speed at 10-m height (WS10m) and downward latent heat flux at the sea surface (LHF) around the Pacific region are presented in Figs. 3a, 3b, and 3c. Positive (negative) LHF means ocean heat gain (loss). In the climatology of SLP, local maximum corresponding to the North Pacific High in the latitude band of 20 N 40 N and the minimum corresponding to the Aleutian Low in 40 N 60 N are observed (Fig. 3a). In SLP anomalies in this period, negative anomalies are observed around the central North Pacific, and positive anomalies are surrounding them from the Kamchatka Peninsula to the west coast of North America (Fig. 3a), indicating that SLP contrast between the Aleutian Low and the North Pacific High became weaker than normal. Consistently, negative WS10m anomaly is observed in the large area of the North Pacific region (Fig. 3b) indicating that atmospheric circulation near the surface also became weaker. In addition, there are meridionally antisymmetric patterns with positive (negative) SLP anomalies in the Northern (Southern) hemisphere around the eastern part of the tropical Pacific. Maeda et al. (2016) suggest that feedbacks among wind, evaporation, and SST and intensified convective activity in the intertropical convergence zone contributed to antisymmetric patterns associated with warming in the eastern tropical region in the North Pacific. As a whole, weaker wind anomalies suppress evaporation, and latent heat loss from the ocean is reduced around the extratropics and in the central to eastern tropical region of the North Pacific (Fig. 3c), the spatial pattern of which is qualitatively consistent with that of the region where SST increased and/or were kept warmer during this period (Fig. 2d). These relations among wind speed, latent heat flux, and SST variability are confirmed by statistical analysis. As shown in Fig. 3e, WS10m and LHF show significant negative correlation in most regions, indicating that the weakened surface wind is consistent with increased downward latent heat flux at sea surface. In addition, significant positive correlation between LHF and SST increase is recognized northeastern Pacific, which confirms the consistency between positive LHF anomaly contribution and SST increase in this region. By supposing that temperature variability in surface mixed layer is vertically uniform and equal to that of SST, we also estimated SST change caused by the LHF contribution (Fig. 3c) using the climatology of mixed layer depth based on de Boyer Montégut et al. (2004). Estimated SST increase caused by LHF averaged in the northeastern Pacific region (40 N 60 N, 150 W 120 W) is 0.76 C, which compensate for about 60% of observed SST increase (Fig. 3d) averaged in the same region (1.25 C). This result indicates that LHF is supposed to play an important role for SST variability in this region where rapid warming was observed from 2012/13 to 2013/14. Although the weakening of easterly trade wind and the North Pacific high are known as typical response to El Niño event, El Niño conditions were quite weak until the early 2015 and effect of El Niño event on atmospheric circulation fields were also unobvi- Fig. 3. (a) Spatial pattern of sea level pressure (SLP) (hpa). Contour indicates annual mean climatology and shading indicates anomaly contributing to SST variability from 2012/13 to 2013/14. (b) Same as shading in (a) for wind speed at 10-m height (WS10m) (m s 1 ). (c) Same as (b) for downward latent heat flux at sea surface (LHF) (W m 2 ). Positive (negative) LHF means ocean heat gain (loss). (d) SST anomaly difference between 2013/ /13. (e) Correlation coefficients between LHF and WS10m anomalies. (f) Correlation coefficients between SST variability and contribution from LHF anomaly. Analysis period is from 1958 to Dots indicate that correlation coefficients are significant at 95% confidence level according to Student s t-test.

5 SOLA, 2017, Vol. 13, 25 30, doi: /sola Fig. 4. (a) Lag correlation (black line) and regression (red line) coefficients between anomalies of GASST and NINO-3 SST. Positive lag means NINO- 3 SST leads GASST. (b) Time series of GASST anomalies separated into the component calculated from the one month lag regression to NINO-3 SST anomalies (red lines; NINO-3-regression) and the residual component (blue line). (c) Time series of residual component similar to blue line of (b) for recent period from 1995 to (d) Time series of NINO-3-regression (red line) and GASST anomalies (black line; equal to sum of blue line in (c) and red line) for recent period from 1995 to ous (Menkes et al. 2014; Maeda et al. 2016). Therefore, the results presented here suggest that some processes other than ENSO at mid- and/or high latitudes played important roles for the warming in the North Pacific in 2013/14/ Statistical estimation of El Niño contribution Trenberth et al. (2002) show that the global surface temperature increases accompanying El Niño events with lags of several months. According to the lag correlation and regression coefficients shown in Fig. 4a, the GASST exhibits a similar response with rather small lags and a maximum regression of C per 1 C anomaly of NINO-3 SST after a one-month lag. Figures 4b, 4c, and 4d show the time series of GASST anomalies separated into a component calculated from the one-month-lag regression to NINO-3 SST anomalies (hereafter, referred to as NINO-3- regression ) and the residual component. Year-by-year variability of NINO-3-regression is presented in Table 1 along with that of GASST. Although annual variabilities exist in the residual component as well as in the long-term trend (Fig. 4b), increase tendency became weaker after late 2014 compared with the rapid warming observed since late 2012 (Fig. 4c) and NINO-3-regression (0.094 C) compensates for more than 65% of GASST increase (0.147 C) from 2014/15 to 2015/16 (Table 1). Therefore, the remarkable increase in the GASST since the end of 2014 is mainly attributed to the development of a strong El Niño event (Fig. 4d). Similarly, the residual component shows no increase from 1997 to early 1998 and increase of GASST observed in this period is likely to be caused by El Niño event, too. However, there is no clear increase for some years before 1997 and the significant increase in GASST persisted only for one year, associated with a development and decay of the El Niño event. In the recent years, in contrast, increase in residual component emerged in around 2013 and 2014 and positive values persisted since then, which is supposed to correspond to the positive anomalies in the North Pacific indicated in Section. 4 and contribute to the recordbreaking values marked in 2015/16. Specifically, the increase in North Pacific region contributes to GASST by C between 2012/13 and 2014/15, which corresponds to 75% of GASST increase (0.099 C) in the same period (Table 1). Even if 2015/16 is included, the North Pacific region compensates for 30% of GASST increase since 2012/13 (0.246 C). 6. Summary and discussion In this study, oceanic conditions and related atmospheric variabilities associated with the rapid warming in the GASST since around 2013 are presented. Although this increase is generally attributed to the emergence and development of an El Niño event, significant differences were found compared with 1997/98 when the previous highest record was marked associated with a strong El Niño. In the tropical Pacific, warming near the surface had substantially progressed in 2014 before the El Niño condition developing into a strong level. As a result, the warmed ocean surface in the tropical Pacific persisted from early 2014 to early 2016 or further, much longer than that observed from early 1997 to mid Although it is beyond the scope of this study, more detailed investigations are necessary in order to estimate the influence of subsurface temperature anomaly on SST variability in the tropical Pacific. Warming in the North Pacific before early 2014 also contributed to extreme anomaly observed in 2015/16, substantially larger than that in 1997/98 even if the effect of longterm global warming is considered. This warming in the North Pacific is consistent with reduced latent heat loss associated with weakened atmospheric circulation and not likely to be affected by tropical variability. The remarkably large positive anomaly observed in the North Pacific in 2013 is referred to as blob and attributed to atmospheric forcing including advection and entrainment in addition to surface heat flux in the recent study by Bond et al. (2015). However, detailed mechanism that induced the anomalous atmospheric field is still unclear and further investigation is necessary. Recent studies indicate that La Niña-like conditions associated with decadal climate variability trigger ocean heat uptake into Pacific subsurface and play an important role in weakened increase of the global surface temperature (the so-called hiatus ; Easterling and Wehner 2009) in the last decade (England et al. 2014; Liu et al. 2016). Urabe and Maeda (2014) indicate that positive subsurface temperature anomaly in the western tropical Pacific had been continuously accumulating along with the recent La Niña-like condition in decadal timescale. The subsurface temperature anomaly redistribution observed in the tropical Pacific in recent years (Figs. 2i, 2j, and 2k) and warming in GASST are in stark contrast with the conditions continued during the hiatus, in other words, the recent warming event is possibly accompanied by a rebound from the hiatus. Although it is difficult to show whether global warming hiatus ended around , we should con-

6 30 Urabe et al., Rapid Warming in Global Sea Surface Temperature tinue to monitor the climate system carefully in order to assess the subsequent status of global warming. Acknowledgements The authors are grateful to anonymous reviewers for their detailed instructions and comments. The authors would also like to thank the members of the climate prediction division of JMA and the Meteorological Research Institute for interesting discussion and instructive comments. Edited by: M. Nonaka References Bond, N. A., M. F. Cronin, H. Freeland, and N. Mantua, 2015: Causes and impacts of the 2014 warm anomaly in the NE Pacific. Geophys. Res. Lett., 42, doi: /2015gl de Boyer Montégut, C., G. Madec, A. S. Fischer, A. Lazar, and D. Iudicone, 2004: Mixed layer depth over the global ocean: An examination of profile data and a profile-based climatology. J. Geophys. Res., 109, doi: /2004jc Easterling, D. R., and M. F. Wehner, 2009: Is the climate warming or cooling? Geophys. Res. Lett., 36, doi: /2009gl England, M. H., S. McGregor, P. Spence, G. A. Meehl, A. Timmermann, W. Cai, A. S. Gupta, M. J. McPhaden, A. Purich, and A. Santoso, 2014: Recent intensification of wind-driven circulation in the Pacific and the ongoing warming hiatus. Nat. Climate Change, 4, , doi: /nclimate Hasegawa, T., and K. Hanawa, 2003: Head content variability related to ENSO events in the Pacific. J. Phys. Oceanogr., 33, Huang, B., V. F. Banzon, E. Freeman, J. Lawrimore, W. Liu, T. C. Peterson, T. M. Smith, P. W. Thorne, S. D. Woodruff, and H.-M. Zhang, 2014: Extended reconstructed sea surface temperature version 4 (ERSST.v4). Part I: Upgrades and intercomparisons. J. Climate, 28, , doi: / JCLI-D IPCC, 2013: Working group I contribution to the fifth assessment report of the intergovernmental panel on climate change, Climate Change 2013: The Physical Science Basis, Cambridge University Press, Cambridge, UK and New York, USA. Ishii, M., A. Shouji, S. Sugimoto, and T. Matsumoto, 2005: Objective analyses of sea-surface temperature and marine meteorological variables for the 20th century using ICOADS and the Kobe Collection. Intl. J. Climatol., 25, Kennedy, J. J., N. A. Rayner, R. O. Smith, M. Saunby, and D. E. Parker, 2011a: Reassessing biases and other uncertainties in sea-surface temperature observations since Part 1: Measurement and sampling errors. J. Geophys. Res., 116, D14103, doi: /2010jd Kennedy, J. J., N. A. Rayner, R. O. Smith, M. Saunby, and D. E. Parker, 2011b: Reassessing biases and other uncertainties in sea-surface temperature observations since 1850, Part 2: Biases and homogenisation. J. Geophys. Res., 116, D14104, doi: /2010jd Kessler, W. S., 1990: Observations of long Rossby waves in the northern tropical Pacific. J. Geophys. Res., 95, , doi: /jc095ic04p Klein, S., B. J. Soden, and N.-C. Lau, 1999: Remote sea surface temperature variations during ENSO: Evidence for a tropical atmospheric bridge. J. Climate, 12, Kobayashi, S., Y. Ota, Y. Harada, A. Ebita, M. Moriya, H. Onoda, K. Onogi, H. Kamahori, C. Kobayashi, H. Endo, K. Miyaoka, and K. Takahashi, 2015: The JRA-55 reanalysis: General specifications and basic characteristics. J. Meteor. Soc. Japan, 93, 5 48, doi: /jmsj Liu, W., B. Huang, P. W. Thorne, V. F. Banzon, H.-M. Zhang, E. Freeman, J. Lawrimore, T. C. Peterson, T. M. Smith, and S. D. Woodruff, 2014: Extended reconstructed sea surface temperature version 4 (ERSST.v4). Part II: Parametric and structural uncertainty estimations. J. Climate, 28, , doi: /jcli-d Liu, W., S. P. Xie, and J. Lu, 2016: Tracking ocean heat uptake during the surface warming hiatus. Nature Commun., 7, 10926, doi: /ncomms Maeda, S., Y. Urabe, K. Takemura, T. Yasuda, and Y. Tanimoto, 2016: Active role of the ITCZ and WES feedback in hampering the growth of the expected full-fledged El Niño in SOLA, 12, 17 21, doi: /sola Menkes, C. E., M. Lengaigne, J. Vialard, M. Puy, P. Marchesiello, S. Cravatte, and G. Cambon, 2014: About the role of Westerly Wind Events in the possible evelopment of an El Niño in Geophys. Res. Lett., 41, , doi: / 2014GL Toyoda, T., Y. Fujii, T. Yasuda, N. Usui, T. Iwao, T. Kuragano, and M. Kamachi, 2013: Improved analysis of seasonalinterannual fields using a global ocean data assimilation system. Theor. Appl. Mech. Japan, 61, Trenberth, K. E., J. M. Caron, D. P. Stepaniak, and S. Worley, 2002: Evolution of El Niño-Southern Oscillation and global atmospheric surface temperatures. J. Geophys. Res., 107, doi: /2000jd Urabe, Y., and S. Maeda, 2014: The relationship between Japan s recent temperature and decadal variability. SOLA, 10, , doi: /sola White, W. B., Y. M. Tourre, M. Barlow, and M. Dettinger, 2003: A delayed action oscillator shared by biennial, interannual, and decadal signals in the Pacific Basin. J. Geophys. Res., 108, doi: /2002jc Manuscript received 24 June 2016, accepted 1 February 2017 SOLA: jstage. jst. go. jp/browse/sola/