Twentieth-century sea surface temperature trends

Science

Volumn 275, Issue 5302, 1997, Pages 957-960

  Cane, Mark A ^a   Clement, Amy C ^a   Kaplan, Alexey ^a   Kushnir, Yochanan ^a   Pozdnyakov, Dmitri ^a   Seager, Richard ^a   Zebiak, Stephen E ^a   Murtugudde, Ragu ^b  

^a Department of Earth and Environmental Sciences   (United States)

^b UNIVERSITIES SPACE RESEARCH ASSOCIATION   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

AIR TEMPERATURE; ARTICLE; ATMOSPHERE; CLIMATE; GREENHOUSE EFFECT; MODEL; PACIFIC OCEAN; PRIORITY JOURNAL; SEA; TROPICS;

GLOBAL WARMING; LONG-TERM TRENDS; SST;

EID: 0030621272     PISSN: 00368075     EISSN: None     Source Type: Journal    
DOI: 10.1126/science.275.5302.957     Document Type: Article

References (44)

1. J. T. Houghton et al., Eds., Climate Change 1995 -The Science of Climate Change (Cambridge Univ. Press, Cambridge, 1995).
2. J. F. B. Mitchell, T. C. Johns, J. M. Gregory, S. F. B. Tett, Nature 376, 501 (1995).
3. R. J. Charlson et al., Science 255, 423 (1992).
4. J. T. Kiehl and B. P. Briegleb [ibid. 260, 311 (1993)] and J. Hansen, M. Sato, and R. Ruedy [Atmos. Res. 37, 175 (1995)] demonstrated that the direct radiative influence of sulfate aerosols is far too small to correct the simulations; it would have to be indirect, through their ability to alter cloud properties. The amplitude of the latter is highly speculative; compare (1) and A. Jones, L. Roberts, A. Slingo, Nature 370, 450 (1994). Hansen, Sato, and Ruedy (in preparation) argue that the direct effect of the aerosols may be entirely canceled by a semidirect effect in which heating due to aerosol absorption reduces low cloud cover. Some inconsistencies of the aerosol cooling hypothesis with the observed spatial and temporal patterns of temperature change have been pointed out by T. R. Karl, R. W. Knight, G. Kukla, and J. Gavin [in Aerosol Forcing of Climate, R. J. Charlson and J. Heintzenberg, Eds. (Wiley, Chichester, UK, 1995), pp. 363-382].
5. J. T. Kiehl and B. P. Briegleb [ibid. 260, 311 (1993)] and J. Hansen, M. Sato, and R. Ruedy [Atmos. Res. 37, 175 (1995)] demonstrated that the direct radiative influence of sulfate aerosols is far too small to correct the simulations; it would have to be indirect, through their ability to alter cloud properties. The amplitude of the latter is highly speculative; compare (1) and A. Jones, L. Roberts, A. Slingo, Nature 370, 450 (1994). Hansen, Sato, and Ruedy (in preparation) argue that the direct effect of the aerosols may be entirely canceled by a semidirect effect in which heating due to aerosol absorption reduces low cloud cover. Some inconsistencies of the aerosol cooling hypothesis with the observed spatial and temporal patterns of temperature change have been pointed out by T. R. Karl, R. W. Knight, G. Kukla, and J. Gavin [in Aerosol Forcing of Climate, R. J. Charlson and J. Heintzenberg, Eds. (Wiley, Chichester, UK, 1995), pp. 363-382].
6. J. T. Kiehl and B. P. Briegleb [ibid. 260, 311 (1993)] and J. Hansen, M. Sato, and R. Ruedy [Atmos. Res. 37, 175 (1995)] demonstrated that the direct radiative influence of sulfate aerosols is far too small to correct the simulations; it would have to be indirect, through their ability to alter cloud properties. The amplitude of the latter is highly speculative; compare (1) and A. Jones, L. Roberts, A. Slingo, Nature 370, 450 (1994). Hansen, Sato, and Ruedy (in preparation) argue that the direct effect of the aerosols may be entirely canceled by a semidirect effect in which heating due to aerosol absorption reduces low cloud cover. Some inconsistencies of the aerosol cooling hypothesis with the observed spatial and temporal patterns of temperature change have been pointed out by T. R. Karl, R. W. Knight, G. Kukla, and J. Gavin [in Aerosol Forcing of Climate, R. J. Charlson and J. Heintzenberg, Eds. (Wiley, Chichester, UK, 1995), pp. 363-382].
7. J. T. Kiehl and B. P. Briegleb [ibid. 260, 311 (1993)] and J. Hansen, M. Sato, and R. Ruedy [Atmos. Res. 37, 175 (1995)] demonstrated that the direct radiative influence of sulfate aerosols is far too small to correct the simulations; it would have to be indirect, through their ability to alter cloud properties. The amplitude of the latter is highly speculative; compare (1) and A. Jones, L. Roberts, A. Slingo, Nature 370, 450 (1994). Hansen, Sato, and Ruedy (in preparation) argue that the direct effect of the aerosols may be entirely canceled by a semidirect effect in which heating due to aerosol absorption reduces low cloud cover. Some inconsistencies of the aerosol cooling hypothesis with the observed spatial and temporal patterns of temperature change have been pointed out by T. R. Karl, R. W. Knight, G. Kukla, and J. Gavin [in Aerosol Forcing of Climate, R. J. Charlson and J. Heintzenberg, Eds. (Wiley, Chichester, UK, 1995), pp. 363-382].
8. D.-Z. Sun and Z. Liu, Science 272, 1148 (1996).
9. J. D. Neelin and H. A. Dijkstra, J. Clim. 8, 1325 (1995); H. A. Dijkstra and J. D. Neelin, ibid., p. 1343.
10. J. D. Neelin and H. A. Dijkstra, J. Clim. 8, 1325 (1995); H. A. Dijkstra and J. D. Neelin, ibid., p. 1343.
11. A. C. Clement, R. Seager, M. A. Cane, S. E. Zebiak, ibid. 9, 2190 (1996).
12. M. Latif, R. Kleemon, C. Eckert, in preparation.
13. J. Bjerknes, Mon. Weather Rev. 97, 163 (1969).
14. M. A. Cane, in Teleconnections Linking Worldwide Climate Anomalies, M. Glantz, R. W. Katz, N. Nicholls, Eds. (Cambridge Univ. Press, Cambridge, 1991), pp. 345-369.
15. The thermocline is a region of sharp temperature change separating the warm surface layers from the cold abyssal waters.
16. S. E. Zebiak and M. A. Cane, Mon. Weather Rev. 115, 2262 (1987). Details of the calculations used to produce Fig. 1 are described in (7). The model domain is 29°S to 29°N, 124°E to 80°W.
17. J. McCreary and P. Lu, J. Phys. Oceanogr. 24, 466 (1994); Z. Liu, S. G. H. Philander, R. Pacanowski, ibid, p. 1606.
18. J. McCreary and P. Lu, J. Phys. Oceanogr. 24, 466 (1994); Z. Liu, S. G. H. Philander, R. Pacanowski, ibid, p. 1606.
19. J. R.. Toggweiler, K. Dixon, W. S. Broecker, J. Geophys. Res. 96, 20467 (1991).
20. R. Seager and R. Murtugudde, J. Clim., in press.
21. T. R. Knutson and S. Manabe, ibid. 8, 2181 (1995).
22. S. Tett, ibid., p. 1473.
23. G. A.. Meehl and W. M. Washington [Nature 382, 56 (1996)] provide an alternative mechanism for greater east Pacific warming. Following V. Ramanathan and W. D. Collins [ibid. 351, 27 (1991)], they assumed that the albedo of deep convective clouds increases with SST. As the climate warms, this larger albedo reduces the solar radiation absorbed in the west Pacific relative to that in the east. This cirrus cloud thermostat mechanism has been criticized by some [see, for example, R. Fu, A. D. DelGenio, W. B. Rossow, W. T. Liu, ibid. 358, 394 (1992); D. Hartmann and M. Michelson, J. Clim. 6, 2049 (1993); R. T. Pierrehumbert, J. Atmos. Sci. 52, 1784 (1995)].
24. G. A.. Meehl and W. M. Washington [Nature 382, 56 (1996)] provide an alternative mechanism for greater east Pacific warming. Following V. Ramanathan and W. D. Collins [ibid. 351, 27 (1991)], they assumed that the albedo of deep convective clouds increases with SST. As the climate warms, this larger albedo reduces the solar radiation absorbed in the west Pacific relative to that in the east. This cirrus cloud thermostat mechanism has been criticized by some [see, for example, R. Fu, A. D. DelGenio, W. B. Rossow, W. T. Liu, ibid. 358, 394 (1992); D. Hartmann and M. Michelson, J. Clim. 6, 2049 (1993); R. T. Pierrehumbert, J. Atmos. Sci. 52, 1784 (1995)].
25. G. A.. Meehl and W. M. Washington [Nature 382, 56 (1996)] provide an alternative mechanism for greater east Pacific warming. Following V. Ramanathan and W. D. Collins [ibid. 351, 27 (1991)], they assumed that the albedo of deep convective clouds increases with SST. As the climate warms, this larger albedo reduces the solar radiation absorbed in the west Pacific relative to that in the east. This cirrus cloud thermostat mechanism has been criticized by some [see, for example, R. Fu, A. D. DelGenio, W. B. Rossow, W. T. Liu, ibid. 358, 394 (1992); D. Hartmann and M. Michelson, J. Clim. 6, 2049 (1993); R. T. Pierrehumbert, J. Atmos. Sci. 52, 1784 (1995)].
26. G. A.. Meehl and W. M. Washington [Nature 382, 56 (1996)] provide an alternative mechanism for greater east Pacific warming. Following V. Ramanathan and W. D. Collins [ibid. 351, 27 (1991)], they assumed that the albedo of deep convective clouds increases with SST. As the climate warms, this larger albedo reduces the solar radiation absorbed in the west Pacific relative to that in the east. This cirrus cloud thermostat mechanism has been criticized by some [see, for example, R. Fu, A. D. DelGenio, W. B. Rossow, W. T. Liu, ibid. 358, 394 (1992); D. Hartmann and M. Michelson, J. Clim. 6, 2049 (1993); R. T. Pierrehumbert, J. Atmos. Sci. 52, 1784 (1995)].
27. G. A.. Meehl and W. M. Washington [Nature 382, 56 (1996)] provide an alternative mechanism for greater east Pacific warming. Following V. Ramanathan and W. D. Collins [ibid. 351, 27 (1991)], they assumed that the albedo of deep convective clouds increases with SST. As the climate warms, this larger albedo reduces the solar radiation absorbed in the west Pacific relative to that in the east. This cirrus cloud thermostat mechanism has been criticized by some [see, for example, R. Fu, A. D. DelGenio, W. B. Rossow, W. T. Liu, ibid. 358, 394 (1992); D. Hartmann and M. Michelson, J. Clim. 6, 2049 (1993); R. T. Pierrehumbert, J. Atmos. Sci. 52, 1784 (1995)].
28. "Coupling" is used in the ENSO literature to mean a feedback loop wherein a given change in SST causes a change in equatorial wind stress that results in a change in thermocline depth and finally SST. The ratio of the final to initial SST change would be a measure of "coupling" strength. Note that the oceanic changes are dynamical. Thermodynamics enters the loop only through its effect on wind stress.
29. N. C. Lau, S. G. H. Philander, M. J. Nath, J. Clim. 5, 284 (1992); S. Manabe and R. J. Stouffer, ibid. 9, 376 (1996); J. D. Neelin et al., Clim. Dyn. 7, 73 (1992).
30. N. C. Lau, S. G. H. Philander, M. J. Nath, J. Clim. 5, 284 (1992); S. Manabe and R. J. Stouffer, ibid. 9, 376 (1996); J. D. Neelin et al., Clim. Dyn. 7, 73 (1992).
31. N. C. Lau, S. G. H. Philander, M. J. Nath, J. Clim. 5, 284 (1992); S. Manabe and R. J. Stouffer, ibid. 9, 376 (1996); J. D. Neelin et al., Clim. Dyn. 7, 73 (1992).
32. A. Kaplan et al. (in preparation) present a global analysis and validation. The method [developed by A. Kaplan, M. A. Cane, Y. Kushnir, and B. Blumenthal (in preparation)] uses empirically derived time and space covariance estimates to obtain an optimally smoothed analysis of Atlantic SSTs from 1856 to 1991. The analysis procedure is similar to so-called "optimal interpolation" in that it uses estimates of spatial covariance to fill gaps in the data. In addition, it makes limited use of temporal covariance estimates in a Kalman Smoother algorithm; the SST data comes from (27).
33. Although the analysis extends from 1856 to 1991, we went back no further than 1900 because of suspicions that the 19th-century SSTs are systematically in error (1, 20).
34. The Lamont model has a well-known bias that places the maximum SST anomaly too far east; allowing for this would further increase the resemblance of Fig. 1 to the observed pattern.
35. K. Trenberth and T. Hoar, Geophys. Res. Lett. 23, 57 (1996).
36. B. Rajagopalan, U. Lall, M. A. Cane, in preparation.
37. The global pattern correlation between Fig. 3, A and B, is 0.97. We also computed the first empirical orthogonal function of the data with ENSO removed; its principal component is largely trend and correlates with the trend patterns at 0.95 or more.
38. M. Bottomley, C. K. Folland, J. Hsiung, R. E. Newell, D. E. Parker, Global Ocean Surface Temperature Atlas (GOSTA) (Her Majesty's Stationary Office, London, 1990). This atlas is based on the U.K. Meteorological Office Main Marine Data Bank and includes corrections for systematic measurement errors. The version of the GOSTA data set that was used in (21), MOHSST5 [D. E. Parker et al., in Natural Climate Variability on Decade to Century Timescales (National Academy Press, Washington, DC, 1996), pp. 244-250], incorporates COADS (28) values where GOSTA data are deficient and uses an improved scheme of corrections for instrumental biases [C. K. Folland and D. E. Parker, Q. J. R. Meteorol. Soc. 121, 319 (1995)].
39. M. Bottomley, C. K. Folland, J. Hsiung, R. E. Newell, D. E. Parker, Global Ocean Surface Temperature Atlas (GOSTA) (Her Majesty's Stationary Office, London, 1990). This atlas is based on the U.K. Meteorological Office Main Marine Data Bank and includes corrections for systematic measurement errors. The version of the GOSTA data set that was used in (21), MOHSST5 [D. E. Parker et al., in Natural Climate Variability on Decade to Century Timescales (National Academy Press, Washington, DC, 1996), pp. 244-250], incorporates COADS (28) values where GOSTA data are deficient and uses an improved scheme of corrections for instrumental biases [C. K. Folland and D. E. Parker, Q. J. R. Meteorol. Soc. 121, 319 (1995)].
40. M. Bottomley, C. K. Folland, J. Hsiung, R. E. Newell, D. E. Parker, Global Ocean Surface Temperature Atlas (GOSTA) (Her Majesty's Stationary Office, London, 1990). This atlas is based on the U.K. Meteorological Office Main Marine Data Bank and includes corrections for systematic measurement errors. The version of the GOSTA data set that was used in (21), MOHSST5 [D. E. Parker et al., in Natural Climate Variability on Decade to Century Timescales (National Academy Press, Washington, DC, 1996), pp. 244-250], incorporates COADS (28) values where GOSTA data are deficient and uses an improved scheme of corrections for instrumental biases [C. K. Folland and D. E. Parker, Q. J. R. Meteorol. Soc. 121, 319 (1995)].
41. S. D. Woodruff, R. J. Slutz, R. L. Jenue, P. N. Steurer, Bull. Am. Meteorol. Soc. 68, 521 (1987).
42. Trends and confidence limits are based on a median estimator, the Sens slope test [see, for example, R. O. Gilbert, Statistical Methods for Environmental Pollution Monitoring (Van Nostrand Reinhold, New York, 1987)]. The more commonly used least squares trend estimates are less robust. The least squares trend estimates are (in degrees Celsius per century, with 95% confidence limits) +0.72 ± 0.63 for the analysis, +0.22 ± 0.69 for GOSTA, and +0.35 ± 0.62 for COADS.
43. For a recent review with many references, see N. C. Lau, Bull. Am. Meteorol. Soc., in press.
44. E. K. Schneider, B. Kirtman, R. S. Lindzen, J. Clim., in press.