• Abarca del Rio, R., 1999: The influence of global warming in Earth rotation speed. Ann. Geophys., 17 , 806811.

  • Abarca del Rio, R., , D. Gambis, , and D. A. Salstein, 2000: Interannual signals in length of day and atmospheric angular momentum. Ann. Geophys., 18 , 347364.

    • Search Google Scholar
    • Export Citation
  • Alexander, M. A., , I. Blade, , M. Newman, , J. R. Lanzante, , N-C. Lau, , and J. D. Scott, 2002: The atmospheric bridge: The influence of ENSO teleconnections on air–sea interaction over the global oceans. J. Climate, 15 , 22052231.

    • Search Google Scholar
    • Export Citation
  • Anderson, J. R., , and R. D. Rosen, 1983: The latitude-height structure of 40–50 day variations in atmospheric angular momentum. J. Atmos. Sci., 40 , 15841591.

    • Search Google Scholar
    • Export Citation
  • Barsugli, J. J., , J. S. Whitaker, , A. F. Loughe, , P. D. Sardeshmukh, , and Z. Toth, 1999: The effect of the 1997/98 El Niño on individual large-scale weather events. Bull. Amer. Meteor. Soc., 80 , 13991411.

    • Search Google Scholar
    • Export Citation
  • Black, R. X., , D. A. Salstein, , and R. D. Rosen, 1996: Interannual modes of variability in atmospheric angular momentum. J. Climate, 9 , 28342849.

    • Search Google Scholar
    • Export Citation
  • Chang, E. K. M., 1998: Poleward-propagating angular momentum perturbations induced by zonally symmetric heat sources in the Tropics. J. Atmos. Sci., 55 , 22292248.

    • Search Google Scholar
    • Export Citation
  • Chao, B. F., 1989: Length-of-day variations caused by El Niño–Southern Oscillation and Quasi-Biennial Oscillation. Science, 243 , 923925.

    • Search Google Scholar
    • Export Citation
  • Cox, C. M., , and B. F. Chao, 2002: Detection of a large-scale mass redistribution in the terrestrial system since 1998. Science, 297 , 831833.

    • Search Google Scholar
    • Export Citation
  • de Viron, O., , S. L. Marcus, , and J. O. Dickey, 2001: Atmospheric torques during the winter of 1989: Impact of ENSO and NAO positive phases. Geophys. Res. Lett., 28 , 19851988.

    • Search Google Scholar
    • Export Citation
  • de Viron, O., , V. Dehant, , H. Goosse, , and M. Crucifix, and Participating CMIP Modeling Groups, 2002: Effect of global warming on the length-of-day. Geophys. Res. Lett., 29 .1146, doi:10.1029/2001GL013672.

    • Search Google Scholar
    • Export Citation
  • Dickey, J. O., , S. L. Marcus, , and R. Hide, 1992: Global propagation of interannual fluctuations in atmospheric angular momentum. Nature, 357 , 484488.

    • Search Google Scholar
    • Export Citation
  • Dickey, J. O., , S. L. Marcus, , R. Hide, , T. M. Eubanks, , and D. H. Boggs, 1994: Angular momentum exchange among the solid Earth, atmosphere, and ocean: A case study of the 1982–1983 El Niño. J. Geophys. Res., 99 , 2392123937.

    • Search Google Scholar
    • Export Citation
  • Gross, R. S., 2000: Combinations of Earth-orientation measurements: SPACE97, COMB97, and POLE97. J. Geod., 73 , 627637.

  • Hide, R., , and J. O. Dickey, 1991: Earth's variable rotation. Science, 253 , 629637.

  • Hines, K. M., , D. H. Bromwich, , and G. J. Marshall, 2000: Artificial surface pressure trend in the NCEP–NCAR reanalysis over the Southern Ocean and Antarctica. J. Climate, 13 , 39403952.

    • Search Google Scholar
    • Export Citation
  • Hoerling, M. P., , J. W. Hurrell, , and T. Xu, 2001a: Tropical origins for recent North Atlantic climate change. Science, 292 , 9092.

  • Hoerling, M. P., , J. S. Whitaker, , A. Kumar, , and W. Wang, 2001b: The midlatitude warming during 1998–2000. Geophys. Res. Lett., 28 , 755758.

    • Search Google Scholar
    • Export Citation
  • Houghton, J. T., , Y. Ding, , D. J. Griggs, , M. Noguer, , P. J. van der Linden, , and D. Xiaosu, Eds.,. 2001: Climate Change 2001: The Scientific Basis. Cambridge University Press, 881 pp.

    • Search Google Scholar
    • Export Citation
  • Huang, H-P., , P. D. Sardeshmukh, , and K. M. Weickmann, 1999: The balance of global angular momentum in a long-term atmospheric data set. J. Geophys. Res., 104 , 20312040.

    • Search Google Scholar
    • Export Citation
  • Huang, H-P., , K. M. Weickmann, , and C. J. Hsu, 2001: Trend in atmospheric angular momentum in a transient climate change simulation with greenhouse gas and aerosol forcing. J. Climate, 14 , 15251534.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and Coauthors. 1996: The NCEP/NCAR 40-year reanalysis project. Bull. Amer. Meteor. Soc., 77 , 437471.

  • Kang, I-S., , and K-M. Lau, 1994: Principal modes of atmospheric circulation anomalies associated with global angular momentum fluctuations. J. Atmos. Sci., 51 , 11941205.

    • Search Google Scholar
    • Export Citation
  • Kistler, R., and Coauthors. 2001: The NCEP–NCAR 50-year reanalysis: Monthly means CD-ROM and documentation. Bull. Amer. Meteor. Soc., 82 , 247267.

    • Search Google Scholar
    • Export Citation
  • Madden, R. A., 1987: Relationship between changes in the length of day and the 40- to 50-day oscillation in the Tropics. J. Geophys. Res., 92 , 83918399.

    • Search Google Scholar
    • Export Citation
  • Marshall, G. J., , and S. A. Harangozo, 2000: An appraisal of NCEP/NCAR reanalysis MSLP data viability for climate studies in the South Pacific. Geophys. Res. Lett., 27 , 30573060.

    • Search Google Scholar
    • Export Citation
  • Mo, K. C., , J. O. Dickey, , and S. L. Marcus, 1997: Interannual fluctuations in atmospheric angular momentum simulated by the National Centers for Environmental Prediction medium range forecast model. J. Geophys. Res., 102 , 67036713.

    • Search Google Scholar
    • Export Citation
  • Pan, Y. H., , and A. H. Oort, 1983: Global climate variations connected with sea surface temperature anomalies in the eastern equatorial Pacific Ocean for the 1958–73 period. Mon. Wea. Rev., 111 , 12441258.

    • Search Google Scholar
    • Export Citation
  • Peixoto, J. P., , and A. H. Oort, 1992: Physics of Climate. American Institute of Physics, 520 pp.

  • Ponte, R. M., , and R. D. Rosen, 1999: Torques responsible for evolution of atmospheric angular momentum during the 1982–83 El Niño. J. Atmos. Sci., 56 , 34573462.

    • Search Google Scholar
    • Export Citation
  • Ponte, R. M., , and D. Stammer, 2000: Global and regional axial ocean angular momentum signals and length-of-day variations (1985–1996). J. Geophys. Res., 105 , 1716117171.

    • Search Google Scholar
    • Export Citation
  • Räisänen, J., 2003: CO2-induced changes in atmospheric angular momentum in CMIP2 experiments. J. Climate, 16 , 132143.

  • Rosen, R. D., 1993: The axial momentum balance of Earth and its fluid envelope. Surv. Geophys., 14 , 129.

  • Rosen, R. D., , and D. A. Salstein, 1983: Variations in atmospheric angular momentum on global and regional scales and the length of day. J. Geophys. Res., 88 , 54515470.

    • Search Google Scholar
    • Export Citation
  • Rosen, R. D., , and D. A. Salstein, 1985: Contribution of stratospheric winds to annual and semiannual fluctuations in atmospheric angular momentum and the length of day. J. Geophys. Res., 90 , 80338041.

    • Search Google Scholar
    • Export Citation
  • Rosen, R. D., , and D. A. Salstein, 2000: Multidecadal signals in the interannual variability of atmospheric angular momentum. Climate Dyn., 16 , 693700.

    • Search Google Scholar
    • Export Citation
  • Rosen, R. D., , D. A. Salstein, , T. M. Eubanks, , J. O. Dickey, , and J. A. Steppe, 1984: An El Nino signal in atmospheric angular momentum and earth rotation. Science, 225 , 411414.

    • Search Google Scholar
    • Export Citation
  • Sardeshmukh, P. D., , G. P. Compo, , and C. Penland, 2000: Changes of probability associated with El Nino. J. Climate, 13 , 42684286.

  • Seager, R., , N. Harnik, , Y. Kushnir, , W. Robinson, , and J. Miller, 2003: Mechanisms of hemispherically symmetric climate variability. J. Climate, in press.

    • Search Google Scholar
    • Export Citation
  • Smith, T. M., , R. W. Reynolds, , R. E. Livezey, , and D. C. Stokes, 1996: Reconstruction of historical sea surface temperatures using empirical orthogonal functions. J. Climate, 9 , 14031420.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., 1990: Recent observed interdecadal climate changes in the Northern Hemisphere. Bull. Amer. Meteor. Soc., 71 , 988993.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., , and J. W. Hurrell, 1994: Decadal atmosphere–ocean variations in the Pacific. Climate Dyn., 9 , 303319.

  • Trenberth, K. E., , and D. P. Stepaniak, 2002: A pathological problem with NCEP reanalysis in the stratosphere. J. Climate, 15 , 690695.

    • Search Google Scholar
    • Export Citation
  • Weickmann, K. M., , G. N. Kiladis, , and P. D. Sardeshmukh, 1997: The dynamics of intraseasonal atmospheric angular momentum oscillations. J. Atmos. Sci., 54 , 14451461.

    • Search Google Scholar
    • Export Citation
  • Weickmann, K. M., , W. A. Robinson, , and C. Penland, 2000: Stochastic and oscillatory forcing of global atmospheric angular momentum. J. Geophys. Res., 105 , 1554315557.

    • Search Google Scholar
    • Export Citation
  • Yang, S., , K-M. Lau, , and K-M. Kim, 2002: Variations of the East Asian jet stream and Asian–Pacific–American winter climate anomalies. J. Climate, 15 , 306325.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    The monthly mean anomalies of global relative angular momentum (black) and Niño-3.4 SST index (red). The anomaly is defined as the departure from the mean seasonal cycle. Both time series are undetrended. Niño-3.4 index is in °C, angular momentum in 1025 kg m2 s−1 (1 AMU). The curves are correlated at 0.52

  • View in gallery

    Same as Fig. 1, but the black curve is the 12-member ensemble average of the global relative AAM anomaly from the NCAR CCM3 simulations. The curves are correlated at 0.61

  • View in gallery

    (a) The monthly mean total atmospheric angular momentum (black) and its decadal/interdecadal component (red) defined as the first six harmonics plus long-term mean. (b) The mean annual cycle of total AAM, defined as the sum of the annual, semiannual, and triannual harmonics. (c) The monthly anomaly of total AAM (black), relative AAM (red), and Niño-3.4 SST index (blue). The AAM curves are defined as the remainder of the monthly AAM time series after the decadal/interdecadal component and mean seasonal cycle are removed. Units are AMU

  • View in gallery

    Same as Fig. 3, but for LOD. The LOD in (c) has been converted to the equivalent AAM using Eq. (1). An arbitrary 5-yr segment of the mean annual cycle of AAM (from Fig. 3b), converted to equivalent LOD, is superimposed in red in (b). The monthly AAM anomaly from Fig. 3c is superimposed in red in (c). Units are ms for (a) and (b), and AMU for (c)

  • View in gallery

    The time–latitude diagram for the monthly anomaly of the vertically and zonally averaged angular velocity u cosθ, based on NCEP reanalysis. The time series of the monthly anomaly at each latitude is constructed in the same way as its globally integrated counterpart in Fig. 3. Regions with the magnitude of the anomaly exceeding 0.5 m s−1 are colored in red (positive) and blue (negative), with shading at an interval of every additional 0.5 m s−1. The four bold horizontal bars indicate the Sep–Mar periods for the 1965/66, 1972/73, 1982/83, and 1997/98 El Niños

  • View in gallery

    The Sep–Mar average of the monthly anomaly of the vertically and zonally averaged angular velocity, u cosθ, for the four El Niño events marked in Fig. 5: (a) 1972/73 (solid) and 1965/66 (dashed), (b) 1997/98 (solid) and 1982/83 (dashed). Units are m s−1

  • View in gallery

    The monthly anomalies of the relative AAM with (black) and without (red) the contribution from the stratosphere, defined as the uppermost eight sigma levels. The anomalies are defined in the same manner as in Figs. 3 and 4. Units are AMU

  • View in gallery

    Same as Fig. 5, but for only the stratospheric portion of the data (the uppermost eight sigma levels) used to construct the vertical/zonal average of u cosθ. Interval for shading is 3 m s−1

  • View in gallery

    Same as Fig. 5, but for the 12-member ensemble average of the u cosθ in the NCAR CCM3 simulations. Shading interval is 0.5 m s−1

  • View in gallery

    The Sep–Mar average of the monthly anomaly of the vertically and zonally averaged angular velocity, u cosθ, for the 1965/66 El Niño, simulated by the NCAR CCM3. (a) The thick black curve shows the 12-member ensemble average; the thin black curves the individual ensemble members. (b) Two outliers of the individual ensemble members with relatively small responses in the Northern Hemisphere are shown. Units are m s−1

  • View in gallery

    Same as Fig. 10, but for the 1972/73 El Niño event. The two selected individual ensemble members shown in (b) are one with a relatively small response in the Northern Hemisphere (solid), and another with the strongest negative response on the equator (dashed)

  • View in gallery

    The Nov 1972–Feb 1973 anomaly of (a) SST based on Reynolds's reconstructed data (Smith et al. 1996); (b) vector wind at 150 hPa (wind direction in arrows, wind speed shaded for anomalies greater than 3, 6, and 9 m s−1); (c) streamfunction at 150 hPa, contour interval 1.0 × 106 m2 s−1. Both (b) and (c) are from reanalysis, and anomalies are based on a 1968–96 climatology

  • View in gallery

    The 5-yr running means of the monthly mean time series in Figs. 1 and 2. The Niño-3.4 SST index is shown in red, global relative AAM of the NCEP reanalysis in blue, and the ensemble mean of the NCAR CCM3-simulated global relative AAM in black. The gray shading indicates one std dev among the 12 ensemble members in the CCM3 simulations. Units are °C for Niño-3.4 index, AMU for angular momentum

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 153 153 3
PDF Downloads 32 32 7

Unusual Behavior in Atmospheric Angular Momentum during the 1965 and 1972 El Niños

View More View Less
  • 1 NOAA–CIRES Climate Diagnostics Center, Boulder, Colorado
  • | 2 Atmospheric and Environmental Research, Inc., Lexington, Massachusetts
© Get Permissions
Full access

Abstract

The global atmospheric angular momentum (AAM) is known to increase with tropical eastern Pacific sea surface temperature (SST) anomalies during El Niño events. Using a reanalysis dataset, the ratio of the monthly AAM anomaly to El Niño SST anomaly (based on the Niño-3.4 index) is found to be approximately 1 angular momentum unit (=1025 kg m2 s−1) per degree Celsius for most post-1975 El Niños. This ratio is much smaller, however, during the 1965/66 and 1972/73 El Niños, raising the possibilities that either the early reanalysis data are in error due to sparse observations, or the atmospheric response to the two early El Niños was unusual. The possibility of a severe data problem in the reanalysis is ruled out by cross-validating the AAM time series with independent measurements of length of day. The latitudinal structures of the zonal wind anomalies in 1965/66 and 1972/73 are examined for both the reanalysis and a set of general circulation model (GCM) simulations. Multiple GCM runs with specified SST produce a more positive ensemble-mean AAM anomaly in 1965 than its counterpart in the reanalysis. The GCM-simulated ensemble-mean zonal wind anomaly resembles the canonical El Niño response with accelerations of subtropical zonal jets in both hemispheres, a pattern that is almost absent in the reanalysis. On the other hand, a large spread exists among the individual ensemble members in the 1965/66 GCM simulations. Although the majority of the individual ensemble members shows the canonical El Niño response, two outliers (out of 12 runs) exhibit very small zonal wind responses in the Northern Hemisphere similar to the reanalysis. Thus, the observed AAM anomaly during 1965/66 is interpreted as an outlier with atmospheric noise being strong enough to overwhelm the canonical El Niño response. The low AAM in the 1972/73 event is related in the reanalysis to a significantly negative zonal wind response on the equator. This signal is robustly reproduced, although with a slightly smaller amplitude, in the ensemble mean and all individual ensemble members in the GCM simulations. The small ensemble standard deviation and large ensemble-mean response on the equator indicate that the negative response is due to the lower-boundary forcing related to the SST anomaly. The fact that the AAM anomaly in 1972/73 is not well correlated with the Niño-3.4 index, then, indicates that SST anomalies outside the conventional El Niño region may be responsible for the low AAM. The uncharacteristically low values of global AAM in 1965/66 and 1972/73 contribute to a low mean for the decade before 1975, which, combined with high AAM in the post-1980 era, produces a significant upward trend in AAM in the second half of the twentieth century. If the weak AAM anomalies during the two pre-1975 El Niños are due to random noise or incidental non-El Niño influences, taking them at face value would result in an overestimate of about 15%–20% in the multidecadal trend of AAM due to boundary forcing alone. Notably, a multidecadal trend in AAM is also simulated in the ensemble mean of the multiple GCM runs, but its magnitude is smaller than the observed counterpart and more consistent with the multidecadal trend of the Niño-3.4 index. The implications of these findings for climate change detection are discussed.

Current affiliation: Lamont-Doherty Earth Observatory, Columbia University, Palisades, New York

Corresponding author address: Dr. Huei-Ping Huang, Lamont-Doherty Earth Observatory, Columbia University, P.O. Box 1000, Palisades, NY 10964-8000. Email: huei@ldeo.columbia.edu

Abstract

The global atmospheric angular momentum (AAM) is known to increase with tropical eastern Pacific sea surface temperature (SST) anomalies during El Niño events. Using a reanalysis dataset, the ratio of the monthly AAM anomaly to El Niño SST anomaly (based on the Niño-3.4 index) is found to be approximately 1 angular momentum unit (=1025 kg m2 s−1) per degree Celsius for most post-1975 El Niños. This ratio is much smaller, however, during the 1965/66 and 1972/73 El Niños, raising the possibilities that either the early reanalysis data are in error due to sparse observations, or the atmospheric response to the two early El Niños was unusual. The possibility of a severe data problem in the reanalysis is ruled out by cross-validating the AAM time series with independent measurements of length of day. The latitudinal structures of the zonal wind anomalies in 1965/66 and 1972/73 are examined for both the reanalysis and a set of general circulation model (GCM) simulations. Multiple GCM runs with specified SST produce a more positive ensemble-mean AAM anomaly in 1965 than its counterpart in the reanalysis. The GCM-simulated ensemble-mean zonal wind anomaly resembles the canonical El Niño response with accelerations of subtropical zonal jets in both hemispheres, a pattern that is almost absent in the reanalysis. On the other hand, a large spread exists among the individual ensemble members in the 1965/66 GCM simulations. Although the majority of the individual ensemble members shows the canonical El Niño response, two outliers (out of 12 runs) exhibit very small zonal wind responses in the Northern Hemisphere similar to the reanalysis. Thus, the observed AAM anomaly during 1965/66 is interpreted as an outlier with atmospheric noise being strong enough to overwhelm the canonical El Niño response. The low AAM in the 1972/73 event is related in the reanalysis to a significantly negative zonal wind response on the equator. This signal is robustly reproduced, although with a slightly smaller amplitude, in the ensemble mean and all individual ensemble members in the GCM simulations. The small ensemble standard deviation and large ensemble-mean response on the equator indicate that the negative response is due to the lower-boundary forcing related to the SST anomaly. The fact that the AAM anomaly in 1972/73 is not well correlated with the Niño-3.4 index, then, indicates that SST anomalies outside the conventional El Niño region may be responsible for the low AAM. The uncharacteristically low values of global AAM in 1965/66 and 1972/73 contribute to a low mean for the decade before 1975, which, combined with high AAM in the post-1980 era, produces a significant upward trend in AAM in the second half of the twentieth century. If the weak AAM anomalies during the two pre-1975 El Niños are due to random noise or incidental non-El Niño influences, taking them at face value would result in an overestimate of about 15%–20% in the multidecadal trend of AAM due to boundary forcing alone. Notably, a multidecadal trend in AAM is also simulated in the ensemble mean of the multiple GCM runs, but its magnitude is smaller than the observed counterpart and more consistent with the multidecadal trend of the Niño-3.4 index. The implications of these findings for climate change detection are discussed.

Current affiliation: Lamont-Doherty Earth Observatory, Columbia University, Palisades, New York

Corresponding author address: Dr. Huei-Ping Huang, Lamont-Doherty Earth Observatory, Columbia University, P.O. Box 1000, Palisades, NY 10964-8000. Email: huei@ldeo.columbia.edu

1. Introduction

The global atmospheric angular momentum (AAM) is widely recognized as a useful index for large-scale circulation variability and change related to the Madden–Julian oscillation (Anderson and Rosen 1983; Madden 1987; Weickmann et al. 1997), El Niño (Rosen et al. 1984; Dickey et al. 1994; Mo et al. 1997), quasi-biennial oscillation (QBO; Chao 1989; Abarca del Rio et al. 2000), and global warming (Abarca del Rio 1999; Huang et al. 2001; de Viron et al. 2002). On intraseasonal-to-interannual timescales, the anomaly in AAM is also, to an excellent approximation, proportional to that in the length of day (LOD), due to the near conservation of total angular momentum of the earth–atmosphere system (e.g., Peixoto and Oort 1992; Rosen 1993). This unique property allows one to cross-validate meteorological analyses of AAM with independent geodetic measurements of earth rotation. As is well known and illustrated anew in section 2, variability in AAM on intraseasonal-to-interannual timescales is dominated by that in the relative component of angular momentum, MR, associated with the zonal wind field.

Focusing on El Niño, a casual inspection of the MR time series derived from the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis dataset (Kalnay et al. 1996) shows that AAM typically increases with the Niño-3.4 (5°N–5°S, 170°–120°W) SST index during the warm phase of El Niño, with ΔAAM/ΔSST (Niño-3.4) approximately 1 angular momentum unit (AMU; 1 AMU = 1025 kg m2 s−1) per degree Celsius. As shown in Fig. 1, this ratio holds well for most post-1975 El Niños. (The excessive ΔAAM during 1982/83 can be attributed to the positive phase of the QBO; see section 3.) However, the ratio ΔAAM/ΔSST (Niño-3.4) is much smaller during two early El Niños in 1965/66 and 1972/73, with the global AAM anomalies opposite to or smaller than expected during these events. In contrast, a parallel analysis of a set of climate model simulations made using the NCAR Community Climate Model version 3 (CCM3) with specified SST shows more prominent peaks of positive AAM anomalies during the 1965 and 1972 El Niños (Fig. 2). The AAM anomaly shown in Fig. 2 is the ensemble mean of 12 simulations with identical SST but different initial conditions, while the observed AAM in Fig. 1 is just one single realization. This distinction is not trivial, as will be discussed later.

Motivated by Figs. 1 and 2, we attempt to synthesize available observations, both meteorological and geodetic, and general circulation model (GCM) simulations, to interpret the relatively weak AAM anomalies during the 1965 and 1972 El Niños. The outcome of this effort has implications beyond better understanding these two particular events. First, because the calculation of global AAM requires extensive three-dimensional observations of zonal wind, a concern arises that the apparently weak AAM anomalies in the reanalysis during the 1965 and 1972 El Niños could be due to the relatively sparse nature of early meteorological observations. (The data problem is less likely related to the tropical SST series, which can be faithfully determined with only a few observations in the Tropics.) Some problems in the NCEP–NCAR reanalysis have been reported for the era prior to 1975. These range from a low correlation between reanalysis and in situ observations over the Southern Ocean and Antarctica (Marshall and Harangozo 2000; Hines et al. 2000) to apparently spurious annual oscillations in the sea level pressure in Northern Hemisphere midlatitudes (Yang et al. 2002) and in the global mountain torque (Weickmann et al. 2000; see the pre-1967 portion of their Fig. 1c). On the other hand, Kistler et al. (2001) show that the reanalysis does successfully capture two significant pre-1958 synoptic events. By comparing the reanalysis-based AAM in 1965 and 1972 with independent measurements of LOD, we hope to gain new insights into the usefulness of the early reanalysis data.

Second, the interdecadal trend in the meteorological record for the past 50 years has been intensively analyzed to detect signs of global warming (Houghton et al. 2001). Given that there have been fewer than 10 major El Niños in the last 50 years, each with a substantial contribution to atmospheric circulation anomalies, the atmospheric response to each El Niño noticeably impacts estimates of decadal-to-interdecadal variability and trend. If the AAM responses to the 1965 and 1972 El Niños had been similar to those of the later anomalies in Fig. 1, the decadal mean of AAM from 1965 to 1975 would increase by 0.65 AMU, and the interdecadal trend of the entire record would be reduced (see section 5). Thus, understanding whether, and to what extent, the AAM anomalies during the 1965 and 1972 El Niños are “unusual” is useful for interpreting the observed interdecadal trend of the complete record. In this case, we are concerned with the possibility that the relatively low AAM during the two early El Niños is real, but simply due to noise. Due to the small number of El Niños spanned by the reanalysis, additional information will be drawn from multiple GCM simulations with specified SST to determine the level of noise in the AAM anomalies during these El Niño events.

With this background, we begin with a comparison of AAM and LOD for the early El Niños in section 2, followed by a discussion of the latitudinal structure of the AAM anomalies and the contribution of stratospheric zonal winds in section 3. Section 4 focuses on the GCM simulations of AAM anomalies for the early El Niño events. Discussion and conclusions follow in sections 5 and 6.

2. Cross validation with length-of-day data

To analyze LOD, we use the COMB dataset compiled by Gross (2000) and colleagues, which covers the period from 1962 to 2000. A monthly mean time series was constructed from the (interpolated) daily values in the original dataset. Monthly means of total atmospheric angular momentum, MTOT, were constructed from the 6-hourly NCEP–NCAR reanalysis wind and surface pressure fields, using the spectral coefficients in the full 28 sigma-level data archive. Note that MTOT consists of two components: the relative angular momentum MR, related to the strength and distribution of zonal wind; and the “omega” angular momentum MΩ, related to atmospheric mass distribution (e.g., Peixoto and Oort 1992). The former is the dominant factor that will be extensively discussed.

Assuming conservation of angular momentum in the earth–atmosphere system, an increase in atmospheric angular momentum is proportional to an increase in the length of day, governed by a linear relationship (e.g., Rosen and Salstein 1983; Peixoto and Oort 1992),
MTOT
where the change in angular momentum, ΔMTOT, is in AMU and ΔLOD in milliseconds. In deriving (1), the variability of oceanic angular momentum (due to changes in ocean currents and mass distribution) has been neglected, as it is usually small on the timescale of interest here (Ponte and Stammer 2000). On decadal and longer timescales, changes in LOD are strongly influenced by other factors such as core–mantle coupling (Hide and Dickey 1991; Rosen 1993) or changes in the moment of inertia of the earth (Cox and Chao 2002). Thus, for a useful cross validation based on Eq. (1), low-frequency signals must be removed from the AAM and LOD time series.

Figure 3a shows the unfiltered time series of monthly mean total AAM from 1962 to 2000 and its decadal/interdecadal component. The decadal/interdecadal component is defined as the long-term mean plus the first six harmonics (with periods of 39, 39/2, … , 39/6 yr). Figure 3b shows the mean annual cycle of the MTOT time series, defined as the sum of the annual, semiannual, and third (4-month period) harmonics. After the decadal variability and mean annual cycle are removed from the original time series, we obtain the monthly anomaly shown in Fig. 3c. A similar analysis performed on MR is superimposed in Fig. 3c and confirms that the relative component ΔMR dominates the variability in total AAM, ΔMTOT. In the following sections, we will therefore focus on the MR anomaly and its local structure.

With the removal of decadal variability and trend, the pre-1975 segment of the AAM time series in Fig. 3c is elevated compared to the MR series in Fig. 1, and the AAM anomalies associated with the 1965 and 1972 El Niños emerge as minor local maxima, which, however, remain noticeably smaller than the corresponding peaks in the Niño-3.4 SST index. These two minor positive AAM anomalies in 1965/66 and 1972/73 can be somewhat enhanced by applying a low-pass filter to the monthly mean data (Black et al. 1996; Abarca del Rio et al. 2000). Here, we choose not to filter the data further to allow a more rigorous comparison between the AAM and LOD series. It is worth noting that while the El Niño SST evolves on an interannual timescale, the atmospheric response to the SST anomaly usually exhibits multiple timescales with substantial contributions from synoptic processes (Ponte and Rosen 1999; Barsugli et al. 1999).

Figures 4a–c show the counterparts of Figs. 3a–c for the length of day. The same harmonic analysis was applied to the monthly LOD time series (Fig. 4a) to extract the decadal/interdecadal component and the mean annual cycle (Fig. 4b). The monthly anomaly of LOD, converted to equivalent AAM units using Eq. (1), is shown in Fig. 4c. As expected, the decadal variability of LOD is significantly different from that of AAM. (See Abarca del Rio et al. 2000 for a useful discussion of this difference.) However, once the decadal variability is removed, excellent agreement exists between the LOD and AAM time series for both the mean annual cycle and monthly anomaly. The annual cycle and monthly anomaly of AAM in Figs. 3b,c are superimposed in Figs. 4b,c using Eq. (1). Previous works (e.g., Rosen and Salstein 1985; Abarca del Rio et al. 2000) have also shown the agreement between AAM and LOD anomalies on these timescales.

The close agreement between the AAM and LOD anomalies in the pre-1975 era in Fig. 4c is encouraging. It indicates that, despite their relatively sparse spatial coverage, the meteorological observations made then are reliable enough to construct global angular momentum values. Some discrepancy does exist between the AAM and LOD anomalies in Fig. 4c, which may be due in part to our exclusion of signals in oceanic angular momentum and in part to errors in LOD. [The error in LOD is estimated by Gross (2000) to be around 0.1 ms (0.6 AMU) for the pre-1973 era, improving to less than 0.05 ms by the 1980s.] Overall, however, the cross validation increases our confidence in the reality of the relatively small AAM anomalies during the 1965 and 1972 El Niños.

3. Latitudinal structure of the AAM anomalies

a. Anomaly in zonal-mean zonal wind

To understand why the MR anomalies during the 1965 and 1972 El Niños are not strongly positive, we construct the monthly anomaly of the zonally and vertically averaged angular velocity, u cosθ, where u is zonal wind and θ is latitude, hereafter abbreviated as 〈u cosθ〉 (with the bracket representing both zonal and vertical averages). The monthly anomaly, shown in Fig. 5 as a latitude–time plot for 1962–2000, is constructed in the same manner as the global integrals shown in Figs. 3c and 4c, with the decadal/interdecadal component and mean annual cycle removed from the time series at each latitude. The September–March period of the 1965/66 and 1972/73 El Niños, and the two major events later in 1982/83 and 1997/98, are indicated by bold horizontal bars at the top of the plot. The September–March average of 〈u cosθ〉 for these four events is shown in Figs. 6a,b. As is well known (Pan and Oort 1983; Kang and Lau 1994; Seager et al. 2003, hereafter SHKRM), the most significant feature of the typical atmospheric zonal wind response to El Niño is the acceleration of the subtropical jets in both hemispheres. The angular momentum signals also appear to propagate poleward in both hemispheres (Dickey et al. 1992; Black et al. 1996; Abarca del Rio et al. 2000), although Chang (1998) and SHKRM caution that the “poleward propagation” may result from the superposition of successive standing oscillations with opposite signs. These characteristics are evident in most El Niño events in Fig. 5 but are weak or absent during the 1965/66 El Niño. This can be clearly seen in Fig. 6a, in which there is little enhancement of the subtropical jets during the 1965/66 event. In the 1972/73 event, the atmospheric zonal wind response in the subtropics is quite typical, with enhanced westerly jets in both hemispheres. The relatively low AAM in this case is due to an atypically negative zonal wind response on the equator.

The absence of a canonical El Niño response in AAM during the 1965/66 event could be due to randomly occuring atmospheric noise masking the El Niño signal. Such noise must be sufficiently strong, however, to counter the SST forcing that persisted for several seasons. It is difficult to assess the level of noise by looking at the single realization of the short observational record. We, therefore, turn later to GCM simulations to gain insight into this problem.

Alternatively, the relatively weak AAM anomalies during the 1965 and 1972 El Niños could be due to other organized, recurring, low-frequency phenomena that coincidentally produce a negative AAM response during these El Niño seasons to cancel the canonical El Niño signal. Candidates for these organized low-frequency phenomena include the quasi-biennial oscillation (QBO) and the North Atlantic Oscillation (NAO), both of which have been suggested capable of interfering (though in a relatively minor way) with El Niño in regulating interannual AAM variability [see de Viron et al. (2001) for a discussion about NAO impacts; Chao (1989) and Abarca del Rio et al. (2000) about QBO impacts]. In the 1965/66 winter, a notable negative zonal wind anomaly is present north of 60°N (Fig. 6). This may be related to NAO variability in that season that contributes negatively to the global AAM, but removing this high-latitude signal only slightly increases the global AAM. The most important factor related to the low AAM value in 1965 is clearly the lack of the typical El Niño response in the zonal wind in low latitudes (i.e., the acceleration of the subtropical jets). The possible impact of the stratospheric QBO on the AAM anomalies of 1965 and 1972 deserves a separate analysis, given that QBO has a period close to the El Niño timescale and a maximum amplitude on the equator. We, therefore, next examine whether the small (1965) and strongly negative (1972) zonal wind anomalies on the equator in Fig. 6a can be attributed to the stratospheric QBO.

b. Impact of stratospheric zonal wind anomaly

Figure 7 shows the monthly relative AAM anomaly, constructed in the same way as Figs. 3c and 4c, with and without the contribution from the stratosphere, defined as the uppermost eight sigma levels above σ = 0.10 (approximately 100 mb). To better illustrate the nature of the stratospheric QBO signal in AAM, the monthly anomaly of 〈u cosθ〉 for the stratosphere only is shown in Fig. 8. Negative phases of the QBO signal at the equator peaked just before the buildup of the 1965 and 1972 El Niños. In the early stages of both El Niños, negative AAM anomalies due to the stratospheric QBO remain. Removing the QBO signal would lead to a slight increase of the AAM anomaly in late 1965 and late 1972, as shown in Fig. 7 (indicated by arrows). Quantitatively, however, the contribution from the QBO is minor in both events. Thus, the strongly negative zonal wind anomaly on the equator in 1972/73 is likely of tropospheric origin. Interestingly, the most notable example of the QBO interfering with the El Niño signal in AAM occurs in 1982/83, when the peak of the positive phase of the QBO coincides with the peak of El Niño (Dickey et al. 1994). As a result, the QBO contributes almost 1 AMU to the positive AAM anomaly in 1982/83 (Fig. 7). If this 1 AMU were removed from the AAM anomaly at the peak, the remaining, “true” El Niño signal would agree better with the Niño-3.4 index shown in Fig. 1. Note as well that the strong positive zonal wind anomaly on the equator in 1982/83 (Figs. 5, 6b) is partly due to the positive phase of QBO.

Uncertainties do exist in the stratospheric portion of the reanalysis data. Because of relatively few in situ observations above the tropopause, the quality of the stratospheric analysis depends on the details of the model used to assimilate the data [e.g., the arrangement of the vertical coordinate and the upper-boundary condition; Trenberth and Stepaniak (2002)]. Kistler et al. (2001) show that the largest difference in the zonal-mean zonal wind between the NCEP–NCAR and the European Centre for Medium-Range Weather Forecasts (ECMWF) reanalyses occurs in the tropical stratosphere, with the maximum difference exceeding 7 m s−1 for the 1979–93 climatology. Thus, it will be worthwhile repeating our analysis of stratospheric angular momentum with the so-called ERA40 reanalysis, when this product becomes available from ECMWF. Concerning the dependence of results on the dataset used, it is also worth noting that Black et al. (1996) found relatively stronger peaks of positive AAM anomalies in 1965/66 and 1972/73 by analyzing rawinsonde data. (Their AAM anomalies might have also been enhanced by bandpass filtering, as mentioned earlier, and by omitting data in the stratosphere above 100 mb.) The difference between the rawinsonde and reanalysis data deserves further investigation, but recall that the good agreement between the reanalysis values and the independent LOD series offers confidence in the former.

4. Perspectives from GCM simulations

Two possible explanations remain for the relatively weak AAM anomalies during the 1965 and 1972 El Niños. First, atmospheric noise occurring in a particular El Niño season may occasionally be strong enough to mask the El Niño signal. Second, the SST anomaly outside the conventional eastern Pacific El Niño region (as quantified by the Niño-3.4 index) may sometimes have a significant impact on the tropical/subtropical zonal wind and global AAM anomalies (Hoerling et al. 2001b). On such occasions, the Niño-3.4 index would not agree well with the global AAM anomaly. Here, we explore these possibilities by analyzing a set of GCM simulations for the early El Niños. Note that in multiple atmospheric GCM runs with identical SSTs (but different initial conditions), the standard deviation of the ensemble members may be considered a measure of atmospheric noise. If the standard deviation is small, the simulated ensemble-mean response can be confidently attributed to boundary forcing due to SST anomalies.

We analyzed a set of 12-member runs made using the NCAR CCM3 for 1950–99 with specified (observed) global SST. These simulations produce realistic low-frequency variability and trends in the height and temperature fields (Hoerling et al. 2001a). The ensemble-mean monthly anomaly of the relative AAM of these runs was shown in Fig. 2. Recall that the GCM-simulated AAM anomalies are more positive during 1965 (and, to some extent, 1972) than their observed counterparts. This behavior also exists in the ensemble mean of the six-member Hadley Centre GCM simulations analyzed by Rosen and Salstein (2000, their Fig. 2). The question remains whether a 12-member ensemble is sufficient to reproduce the canonical response to El Niño for individual events, such as 1965/66 and 1972/73. Sardeshmukh et al. (2000, hereafter SCP) show that a 12-member ensemble is needed to detect a signal of 0.8σ with 95% confidence. Here σ is assumed to be the climatological standard deviation of monthly AAM (∼1.0 × 1025 kg m2 s−1). The estimated El Niño AAM signal during northern winter is also ∼0.8σ based on a correlation between AAM and Niño-3.4 of ∼0.7 (see SCP). Thus, a 12-member ensemble is adequate for our purposes, at least on average. However, if the signal is only 0.5σ, a 32-member ensemble would be required to produce the canonical response with high confidence. As will be discussed later, model error also affects the ability to simulate the canonical El Niño response.

Figure 9 shows the 12-member ensemble mean of the NCAR CCM3-simulated monthly mean anomaly of 〈u cosθ〉 defined in the same manner as Fig. 5. The ensemble averaging effectively filters out noise in the individual runs. As a result, the ensemble mean of 〈u cosθ〉 in Fig. 9 looks much smoother than the observations in Fig. 5, which can be thought of as a single realization. The 〈u cosθ〉 plot of an individual ensemble member of the GCM simulations (not shown) is comparable in its structure to the observed field. After filtering out noise, the canonical El Niño response of an acceleration in the subtropical jets of both hemispheres emerges in Fig. 9 for most El Niños, including the 1965/66 event. Notably, a strongly negative zonal wind anomaly on the equator is simulated in the ensemble mean of the 1972/73 El Niño event. Because CCM3 poorly resolves the stratosphere and is incapable of producing a robust QBO, the interannual variability of the model-simulated AAM anomaly does not contain a coherent stratospheric QBO signal. Hence, the presence of the 1972/73 negative equatorial anomaly in Fig. 9 reinforces our assertion that it has a tropospheric origin. (Also, without the QBO signal, the GCM-simulated wind anomaly on the equator in 1982/83 is negative, instead of positive as observed.)

To illustrate the disparity of the zonal wind responses among the individual ensemble members, Fig. 10a shows the September 1965–March 1966 average of 〈u cosθ〉 anomalies for the ensemble mean and 12 individual ensemble members. The ensemble mean shows the canonical El Niño response of subtropical jet accelerations. This behavior is absent or subdued, however, in some of the individual ensemble members. Figure 10b shows two outliers with relatively small AAM anomalies, both with weak zonal wind responses in the subtropics of the Northern Hemisphere. Based on the existence of these two outliers, out of 12 ensemble members, that do not exhibit the canonical El Niño response in zonal wind, we estimate that there is roughly a 2 ÷ 12 probability of occasions when atmospheric noise is strong enough to overcome the canonical signal. This ratio is the same order of magnitude as observed, in that there is one such outlier (the 1965 event) out of about 8 or 9 major El Niños in the plot in Fig. 1. Thus, we interpret the observed low AAM value in 1965/66 as an instance when atmospheric noise masks the canonical El Niño signal.

The noise in the GCM simulations can be measured by the ensemble standard deviation of the individual ensemble members. For the 1965/66 event shown in Fig. 10a, poleward of 30° in both hemispheres, the ensemble-mean response appears insignificant because it is smaller than the ensemble standard deviation. The most robust signal in Fig. 10a is the acceleration of the subtropical jet in the Southern Hemisphere. In the subtropics of the Northern Hemisphere and on the equator, the magnitude of the ensemble-mean response is comparable to one standard deviation of the ensemble members. The situation is quite different for the 1972/73 El Niño, shown in Fig. 11a. In this event, the ensemble-mean responses on the equator and in the subtropics of both hemispheres are highly robust, with a very small ensemble spread compared to the mean response. There are no extreme outliers among the 12 ensemble members for the 1972/73 event. Two moderate outliers, one with the strongest negative anomaly on the equator and the other with the weakest positive response in the subtropics of the Northern Hemisphere, are shown in Fig. 11b. The relatively low level of atmospheric noise in the simulations of this event indicates that the wind anomalies in the Tropics and subtropics in the ensemble mean can be attributed to the SST boundary forcing. In particular, the strong negative wind response on the equator, which also appears in the reanalysis, is likely due to SST forcing and not atmospheric noise. (In the reanalysis, the negative response on the equator is strong enough to compensate for the positive signals in the subtropics, causing a small global AAM anomaly. In the GCM simulations, the positive zonal wind signals in the subtropics are slightly stronger than the negative one on the equator. As a result, the global AAM anomaly is slightly more positive in the GCM simulations.)

If indeed the observed negative response on the equator in 1972/73 is due to SST forcing, then the low correlation between the Niño-3.4 index and the observed AAM anomaly in that event implies that SST anomalies outside the conventional El Niño region are affecting AAM, causing a more negative wind response on the equator. Significant positive SST anomalies do exist over the Indian Ocean and the equatorial and South Atlantic during the 1972/73 El Niño, as shown in Fig. 12a. The atmospheric wind response at 150 mb during this event, shown in Fig. 12b, exhibits unusually strong easterly anomalies over the equatorial western Indian Ocean and the equatorial Atlantic. Together, these unusual features coupled with the typical El Niño signal of an easterly anomaly over the equatorial eastern Pacific contribute to an unusually strong easterly response on the equator, leading to a low value of global AAM despite the canonical El Niño response of subtropical jet acceleration in both hemispheres. The equatorial easterly wind anomaly over the western Indian Ocean (and, to a lesser extent, the equatorial Atlantic) is associated with a pair of strong anticyclones straddling the equator as is clearly shown in the anomalous streamfunction field in the upper troposphere in Fig. 12c. The relationship between these atmospheric responses and the SST anomalies outside the El Niño region deserves further study. Recent work by Hoerling et al. (2001b) attributes the persistent warming of the extratropical atmosphere in 1998–2000 (after the end of 1997/98 El Niño) to the SST anomaly in the warm pool region in the western Pacific and Indian Oceans. Further GCM experiments with SST forcing that is confined to different regions may shed light on the extent to which the low AAM in 1972/73 is due to the influence of SST anomalies in the western Pacific–Indian Ocean region or elsewhere. When interpreting these GCM results, though, one should bear in mind that the SST anomalies in the Indian and Atlantic Oceans are not necessarily independent of the El Niño SST but can be remotely influenced by the latter through an “atmospheric bridge” (Alexander et al. 2002).

5. Implications for trend detection

Although the above analysis has focused on the 1965 and 1972 El Niños, we are also interested in the impact of these two events on analyses of decadal variability and interdecadal trend. Because there are fewer than 10 major El Niños in the second half of the twentieth century (the only period covered by upper-air observations), each one is capable of impacting estimates of decadal/interdecadal trends pertinent to climate change. The AAM series in Fig. 1 is marked by a noticeable transition in the late 1970s, perhaps related to a change then in climate regimes documented by Trenberth (1990) and Trenberth and Hurrell (1994). Thus, the trend of AAM over the second half of the 20th century can be roughly measured by the difference between the means of the post-1980 and pre-1975 periods. The relatively low AAM during the 1965 and 1972 El Niños contributes to a low mean value for the pre-1975 period. As noted earlier, though, if the AAM anomalies in 1965/66 and 1972/73 corresponded to the level of the Niño-3.4 index in these years, the decadal mean of AAM for 1963–73 would increase by 0.65 AMU and the overall trend for 1962–2000 would be reduced by about 15%–20% [this value depends on the choice of the end points for determining the trend; see Rosen and Salstein (2000)].

To illustrate this point, Fig. 13 depicts the 5-yr running mean of the monthly AAM anomaly in Fig. 1, highlighting low frequencies. Also shown in Fig. 13 are the 5-yr running means of the monthly Niño-3.4 SST index and the ensemble mean and standard deviation of the monthly AAM anomaly in the NCAR CCM3 simulations from Fig. 2. In the pre-1975 period, the AAM anomaly from the reanalysis is well below the Niño-3.4 index, partly due to the weak AAM anomalies in 1965/66 and 1972/73. Our analysis of the GCM simulations suggests that the low AAM value in 1965/66 is likely caused by unusually strong atmospheric noise masking the canonical El Niño response. In the post-1980 period, the notable peak in the AAM reanalysis series associated with the 1982/83 El Niño rises significantly above the corresponding Niño-3.4 index series in Fig. 13. Recall from section 3, however, that about 1 AMU of the peak anomaly then is related to the coincidence in phase of the stratospheric QBO and El Niño signals in AAM. Together, the anomalously and randomly low AAM value in 1965/66 and high value in 1982/83 may exaggerate the upward “trend” in AAM in the 1970s. On the other hand, NCAR CCM3 simulations still yield a significant trend in AAM (especially during the late 1970s to early 1980s), albeit not so large as its counterpart from the reanalysis. Notably, the decadal variability and trend in the GCM-simulated ensemble-mean AAM anomaly corresponds well with this variability in the Niño-3.4 index in Fig. 13. The ratio ΔAAM/ΔSST(Niño-3.4) is about 1.2 AMU °C−1, somewhat greater than that obtained from transient global warming simulations made using coupled GCMs with a doubling of the CO2 concentration (Huang et al. 2001; Räisänen 2003).

The debate about determining trends from either a short observational record or a lengthy GCM simulation is ongoing. While we cannot resolve this issue, we argue that one should be cautious in accepting the “observed” trends at face value, given our demonstration that noise during one or two El Niño events can affect the overall trend for the last 50 years. We acknowledge, too, that GCM simulations are also imperfect. Ensemble averaging multiple GCM runs removes some intraensemble noise but not the systematic errors that may be common to all ensemble members. Further complicating the issue is that the systematic error in a GCM is usually determined by validating the GCM run(s) against the single realization of the observation. The “error” obtained in this manner reflects not only the true error in the model physics but also an arbitrary contribution from noise in the short observational record. Further consideration of these points extends beyond the scope of understanding the unusual AAM anomalies in 1965 and 1972 that motivated our work in the first place.

6. Summary and conclusions

Global atmospheric angular momentum typically increases with the tropical eastern Pacific sea surface temperature anomaly during El Niño. Using the NCEP–NCAR reanalysis, the ratio of the monthly AAM anomaly to the Niño-3.4 SST anomaly is approximately 1 AMU °C−1 for most post-1975 El Niños. This ratio is unusually small, however, during the 1965/66 and 1972/73 El Niños, motivating us to investigate these two particular events. We dismiss the likelihood of a data problem in the reanalysis during 1965 and 1972 by cross validating the AAM time series with independent measurements of length of day. We then examine the local structure of the global AAM anomalies in 1965/66 and 1972/73 in both the reanalysis and in a set of general circulation model simulations. Multiple GCM runs with specified SST produce more positive ensemble-mean AAM anomalies in 1965 than observed. The canonical El Niño response of acceleration of the subtropical zonal jets in both hemispheres, absent in the observed 1965/66 event, is produced in the 12-member ensemble mean of the GCM runs. A large spread exists, however, among the individual ensemble members during 1965/66. Although the majority display the canonical El Niño response, 2 of the 12 runs exhibit a very small zonal wind response in the Northern Hemisphere, similar to the observations. Thus, we interpret the AAM anomaly observed during 1965/66 as an outlier, representing an instance when atmospheric noise is strong enough to overcome the canonical El Niño signal.

The low AAM value during the 1972/73 event is related to an unusually strong negative zonal wind response on the equator. This signal is robustly reproduced, although with a slightly smaller amplitude than observed, in the ensemble mean and all individual ensemble members of the GCM simulations. The small ensemble standard deviation and large ensemble-mean response on the equator indicates that this response is due to the lower-boundary SST forcing. The fact that the AAM anomaly in 1972/73 is not well correlated with the Niño-3.4 index, then, suggests that SST anomalies outside the conventional El Niño region may contribute to the low AAM value in 1972/73. Indeed, the twin anticyclones in the upper troposphere of the Atlantic and Indian Oceans during 1972–73 contribute to unusually strong easterly anomalies over the equator and may be a response to local SST anomalies. Finally, the impact of the negative phase of stratospheric QBO on the AAM anomaly is discernable, but otherwise minor, during both 1965 and 1972 El Niños, further pointing to a tropospheric origin for the low AAM value in these events.

The uncharacteristically low values of global AAM in 1965/66 and 1972/73 contribute to a low mean for the decade before 1975, which, combined with a high AAM in the post-1980 era, yields a notable upward trend in AAM in the second half of the twentieth century. If the weak AAM anomalies during the two early El Niños are due to random noise or other incidental influences, accepting them at face value would result in an overestimate of trends in AAM for epochs other than this half-century. It is worth noting that a multidecadal trend in AAM is also simulated in the ensemble mean of the GCM runs, but its magnitude is smaller than observed and more consistent with the trend in the Niño-3.4 index. The implications of these findings for climate change detection deserve further investigation.

Acknowledgments

The work by RDR reported here is supported by the National Science Foundation under Grant ATM-0002688. We thank Gary Bates for providing the GCM output to us. The comments of Steve Feldstein and an anonymous reviewer were also appreciated.

REFERENCES

  • Abarca del Rio, R., 1999: The influence of global warming in Earth rotation speed. Ann. Geophys., 17 , 806811.

  • Abarca del Rio, R., , D. Gambis, , and D. A. Salstein, 2000: Interannual signals in length of day and atmospheric angular momentum. Ann. Geophys., 18 , 347364.

    • Search Google Scholar
    • Export Citation
  • Alexander, M. A., , I. Blade, , M. Newman, , J. R. Lanzante, , N-C. Lau, , and J. D. Scott, 2002: The atmospheric bridge: The influence of ENSO teleconnections on air–sea interaction over the global oceans. J. Climate, 15 , 22052231.

    • Search Google Scholar
    • Export Citation
  • Anderson, J. R., , and R. D. Rosen, 1983: The latitude-height structure of 40–50 day variations in atmospheric angular momentum. J. Atmos. Sci., 40 , 15841591.

    • Search Google Scholar
    • Export Citation
  • Barsugli, J. J., , J. S. Whitaker, , A. F. Loughe, , P. D. Sardeshmukh, , and Z. Toth, 1999: The effect of the 1997/98 El Niño on individual large-scale weather events. Bull. Amer. Meteor. Soc., 80 , 13991411.

    • Search Google Scholar
    • Export Citation
  • Black, R. X., , D. A. Salstein, , and R. D. Rosen, 1996: Interannual modes of variability in atmospheric angular momentum. J. Climate, 9 , 28342849.

    • Search Google Scholar
    • Export Citation
  • Chang, E. K. M., 1998: Poleward-propagating angular momentum perturbations induced by zonally symmetric heat sources in the Tropics. J. Atmos. Sci., 55 , 22292248.

    • Search Google Scholar
    • Export Citation
  • Chao, B. F., 1989: Length-of-day variations caused by El Niño–Southern Oscillation and Quasi-Biennial Oscillation. Science, 243 , 923925.

    • Search Google Scholar
    • Export Citation
  • Cox, C. M., , and B. F. Chao, 2002: Detection of a large-scale mass redistribution in the terrestrial system since 1998. Science, 297 , 831833.

    • Search Google Scholar
    • Export Citation
  • de Viron, O., , S. L. Marcus, , and J. O. Dickey, 2001: Atmospheric torques during the winter of 1989: Impact of ENSO and NAO positive phases. Geophys. Res. Lett., 28 , 19851988.

    • Search Google Scholar
    • Export Citation
  • de Viron, O., , V. Dehant, , H. Goosse, , and M. Crucifix, and Participating CMIP Modeling Groups, 2002: Effect of global warming on the length-of-day. Geophys. Res. Lett., 29 .1146, doi:10.1029/2001GL013672.

    • Search Google Scholar
    • Export Citation
  • Dickey, J. O., , S. L. Marcus, , and R. Hide, 1992: Global propagation of interannual fluctuations in atmospheric angular momentum. Nature, 357 , 484488.

    • Search Google Scholar
    • Export Citation
  • Dickey, J. O., , S. L. Marcus, , R. Hide, , T. M. Eubanks, , and D. H. Boggs, 1994: Angular momentum exchange among the solid Earth, atmosphere, and ocean: A case study of the 1982–1983 El Niño. J. Geophys. Res., 99 , 2392123937.

    • Search Google Scholar
    • Export Citation
  • Gross, R. S., 2000: Combinations of Earth-orientation measurements: SPACE97, COMB97, and POLE97. J. Geod., 73 , 627637.

  • Hide, R., , and J. O. Dickey, 1991: Earth's variable rotation. Science, 253 , 629637.

  • Hines, K. M., , D. H. Bromwich, , and G. J. Marshall, 2000: Artificial surface pressure trend in the NCEP–NCAR reanalysis over the Southern Ocean and Antarctica. J. Climate, 13 , 39403952.

    • Search Google Scholar
    • Export Citation
  • Hoerling, M. P., , J. W. Hurrell, , and T. Xu, 2001a: Tropical origins for recent North Atlantic climate change. Science, 292 , 9092.

  • Hoerling, M. P., , J. S. Whitaker, , A. Kumar, , and W. Wang, 2001b: The midlatitude warming during 1998–2000. Geophys. Res. Lett., 28 , 755758.

    • Search Google Scholar
    • Export Citation
  • Houghton, J. T., , Y. Ding, , D. J. Griggs, , M. Noguer, , P. J. van der Linden, , and D. Xiaosu, Eds.,. 2001: Climate Change 2001: The Scientific Basis. Cambridge University Press, 881 pp.

    • Search Google Scholar
    • Export Citation
  • Huang, H-P., , P. D. Sardeshmukh, , and K. M. Weickmann, 1999: The balance of global angular momentum in a long-term atmospheric data set. J. Geophys. Res., 104 , 20312040.

    • Search Google Scholar
    • Export Citation
  • Huang, H-P., , K. M. Weickmann, , and C. J. Hsu, 2001: Trend in atmospheric angular momentum in a transient climate change simulation with greenhouse gas and aerosol forcing. J. Climate, 14 , 15251534.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and Coauthors. 1996: The NCEP/NCAR 40-year reanalysis project. Bull. Amer. Meteor. Soc., 77 , 437471.

  • Kang, I-S., , and K-M. Lau, 1994: Principal modes of atmospheric circulation anomalies associated with global angular momentum fluctuations. J. Atmos. Sci., 51 , 11941205.

    • Search Google Scholar
    • Export Citation
  • Kistler, R., and Coauthors. 2001: The NCEP–NCAR 50-year reanalysis: Monthly means CD-ROM and documentation. Bull. Amer. Meteor. Soc., 82 , 247267.

    • Search Google Scholar
    • Export Citation
  • Madden, R. A., 1987: Relationship between changes in the length of day and the 40- to 50-day oscillation in the Tropics. J. Geophys. Res., 92 , 83918399.

    • Search Google Scholar
    • Export Citation
  • Marshall, G. J., , and S. A. Harangozo, 2000: An appraisal of NCEP/NCAR reanalysis MSLP data viability for climate studies in the South Pacific. Geophys. Res. Lett., 27 , 30573060.

    • Search Google Scholar
    • Export Citation
  • Mo, K. C., , J. O. Dickey, , and S. L. Marcus, 1997: Interannual fluctuations in atmospheric angular momentum simulated by the National Centers for Environmental Prediction medium range forecast model. J. Geophys. Res., 102 , 67036713.

    • Search Google Scholar
    • Export Citation
  • Pan, Y. H., , and A. H. Oort, 1983: Global climate variations connected with sea surface temperature anomalies in the eastern equatorial Pacific Ocean for the 1958–73 period. Mon. Wea. Rev., 111 , 12441258.

    • Search Google Scholar
    • Export Citation
  • Peixoto, J. P., , and A. H. Oort, 1992: Physics of Climate. American Institute of Physics, 520 pp.

  • Ponte, R. M., , and R. D. Rosen, 1999: Torques responsible for evolution of atmospheric angular momentum during the 1982–83 El Niño. J. Atmos. Sci., 56 , 34573462.

    • Search Google Scholar
    • Export Citation
  • Ponte, R. M., , and D. Stammer, 2000: Global and regional axial ocean angular momentum signals and length-of-day variations (1985–1996). J. Geophys. Res., 105 , 1716117171.

    • Search Google Scholar
    • Export Citation
  • Räisänen, J., 2003: CO2-induced changes in atmospheric angular momentum in CMIP2 experiments. J. Climate, 16 , 132143.

  • Rosen, R. D., 1993: The axial momentum balance of Earth and its fluid envelope. Surv. Geophys., 14 , 129.

  • Rosen, R. D., , and D. A. Salstein, 1983: Variations in atmospheric angular momentum on global and regional scales and the length of day. J. Geophys. Res., 88 , 54515470.

    • Search Google Scholar
    • Export Citation
  • Rosen, R. D., , and D. A. Salstein, 1985: Contribution of stratospheric winds to annual and semiannual fluctuations in atmospheric angular momentum and the length of day. J. Geophys. Res., 90 , 80338041.

    • Search Google Scholar
    • Export Citation
  • Rosen, R. D., , and D. A. Salstein, 2000: Multidecadal signals in the interannual variability of atmospheric angular momentum. Climate Dyn., 16 , 693700.

    • Search Google Scholar
    • Export Citation
  • Rosen, R. D., , D. A. Salstein, , T. M. Eubanks, , J. O. Dickey, , and J. A. Steppe, 1984: An El Nino signal in atmospheric angular momentum and earth rotation. Science, 225 , 411414.

    • Search Google Scholar
    • Export Citation
  • Sardeshmukh, P. D., , G. P. Compo, , and C. Penland, 2000: Changes of probability associated with El Nino. J. Climate, 13 , 42684286.

  • Seager, R., , N. Harnik, , Y. Kushnir, , W. Robinson, , and J. Miller, 2003: Mechanisms of hemispherically symmetric climate variability. J. Climate, in press.

    • Search Google Scholar
    • Export Citation
  • Smith, T. M., , R. W. Reynolds, , R. E. Livezey, , and D. C. Stokes, 1996: Reconstruction of historical sea surface temperatures using empirical orthogonal functions. J. Climate, 9 , 14031420.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., 1990: Recent observed interdecadal climate changes in the Northern Hemisphere. Bull. Amer. Meteor. Soc., 71 , 988993.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., , and J. W. Hurrell, 1994: Decadal atmosphere–ocean variations in the Pacific. Climate Dyn., 9 , 303319.

  • Trenberth, K. E., , and D. P. Stepaniak, 2002: A pathological problem with NCEP reanalysis in the stratosphere. J. Climate, 15 , 690695.

    • Search Google Scholar
    • Export Citation
  • Weickmann, K. M., , G. N. Kiladis, , and P. D. Sardeshmukh, 1997: The dynamics of intraseasonal atmospheric angular momentum oscillations. J. Atmos. Sci., 54 , 14451461.

    • Search Google Scholar
    • Export Citation
  • Weickmann, K. M., , W. A. Robinson, , and C. Penland, 2000: Stochastic and oscillatory forcing of global atmospheric angular momentum. J. Geophys. Res., 105 , 1554315557.

    • Search Google Scholar
    • Export Citation
  • Yang, S., , K-M. Lau, , and K-M. Kim, 2002: Variations of the East Asian jet stream and Asian–Pacific–American winter climate anomalies. J. Climate, 15 , 306325.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

The monthly mean anomalies of global relative angular momentum (black) and Niño-3.4 SST index (red). The anomaly is defined as the departure from the mean seasonal cycle. Both time series are undetrended. Niño-3.4 index is in °C, angular momentum in 1025 kg m2 s−1 (1 AMU). The curves are correlated at 0.52

Citation: Journal of Climate 16, 15; 10.1175/1520-0442(2003)016<2526:UBIAAM>2.0.CO;2

Fig. 2.
Fig. 2.

Same as Fig. 1, but the black curve is the 12-member ensemble average of the global relative AAM anomaly from the NCAR CCM3 simulations. The curves are correlated at 0.61

Citation: Journal of Climate 16, 15; 10.1175/1520-0442(2003)016<2526:UBIAAM>2.0.CO;2

Fig. 3.
Fig. 3.

(a) The monthly mean total atmospheric angular momentum (black) and its decadal/interdecadal component (red) defined as the first six harmonics plus long-term mean. (b) The mean annual cycle of total AAM, defined as the sum of the annual, semiannual, and triannual harmonics. (c) The monthly anomaly of total AAM (black), relative AAM (red), and Niño-3.4 SST index (blue). The AAM curves are defined as the remainder of the monthly AAM time series after the decadal/interdecadal component and mean seasonal cycle are removed. Units are AMU

Citation: Journal of Climate 16, 15; 10.1175/1520-0442(2003)016<2526:UBIAAM>2.0.CO;2

Fig. 4.
Fig. 4.

Same as Fig. 3, but for LOD. The LOD in (c) has been converted to the equivalent AAM using Eq. (1). An arbitrary 5-yr segment of the mean annual cycle of AAM (from Fig. 3b), converted to equivalent LOD, is superimposed in red in (b). The monthly AAM anomaly from Fig. 3c is superimposed in red in (c). Units are ms for (a) and (b), and AMU for (c)

Citation: Journal of Climate 16, 15; 10.1175/1520-0442(2003)016<2526:UBIAAM>2.0.CO;2

Fig. 5.
Fig. 5.

The time–latitude diagram for the monthly anomaly of the vertically and zonally averaged angular velocity u cosθ, based on NCEP reanalysis. The time series of the monthly anomaly at each latitude is constructed in the same way as its globally integrated counterpart in Fig. 3. Regions with the magnitude of the anomaly exceeding 0.5 m s−1 are colored in red (positive) and blue (negative), with shading at an interval of every additional 0.5 m s−1. The four bold horizontal bars indicate the Sep–Mar periods for the 1965/66, 1972/73, 1982/83, and 1997/98 El Niños

Citation: Journal of Climate 16, 15; 10.1175/1520-0442(2003)016<2526:UBIAAM>2.0.CO;2

Fig. 6.
Fig. 6.

The Sep–Mar average of the monthly anomaly of the vertically and zonally averaged angular velocity, u cosθ, for the four El Niño events marked in Fig. 5: (a) 1972/73 (solid) and 1965/66 (dashed), (b) 1997/98 (solid) and 1982/83 (dashed). Units are m s−1

Citation: Journal of Climate 16, 15; 10.1175/1520-0442(2003)016<2526:UBIAAM>2.0.CO;2

Fig. 7.
Fig. 7.

The monthly anomalies of the relative AAM with (black) and without (red) the contribution from the stratosphere, defined as the uppermost eight sigma levels. The anomalies are defined in the same manner as in Figs. 3 and 4. Units are AMU

Citation: Journal of Climate 16, 15; 10.1175/1520-0442(2003)016<2526:UBIAAM>2.0.CO;2

Fig. 8.
Fig. 8.

Same as Fig. 5, but for only the stratospheric portion of the data (the uppermost eight sigma levels) used to construct the vertical/zonal average of u cosθ. Interval for shading is 3 m s−1

Citation: Journal of Climate 16, 15; 10.1175/1520-0442(2003)016<2526:UBIAAM>2.0.CO;2

Fig. 9.
Fig. 9.

Same as Fig. 5, but for the 12-member ensemble average of the u cosθ in the NCAR CCM3 simulations. Shading interval is 0.5 m s−1

Citation: Journal of Climate 16, 15; 10.1175/1520-0442(2003)016<2526:UBIAAM>2.0.CO;2

Fig. 10.
Fig. 10.

The Sep–Mar average of the monthly anomaly of the vertically and zonally averaged angular velocity, u cosθ, for the 1965/66 El Niño, simulated by the NCAR CCM3. (a) The thick black curve shows the 12-member ensemble average; the thin black curves the individual ensemble members. (b) Two outliers of the individual ensemble members with relatively small responses in the Northern Hemisphere are shown. Units are m s−1

Citation: Journal of Climate 16, 15; 10.1175/1520-0442(2003)016<2526:UBIAAM>2.0.CO;2

Fig. 11.
Fig. 11.

Same as Fig. 10, but for the 1972/73 El Niño event. The two selected individual ensemble members shown in (b) are one with a relatively small response in the Northern Hemisphere (solid), and another with the strongest negative response on the equator (dashed)

Citation: Journal of Climate 16, 15; 10.1175/1520-0442(2003)016<2526:UBIAAM>2.0.CO;2

Fig. 12.
Fig. 12.

The Nov 1972–Feb 1973 anomaly of (a) SST based on Reynolds's reconstructed data (Smith et al. 1996); (b) vector wind at 150 hPa (wind direction in arrows, wind speed shaded for anomalies greater than 3, 6, and 9 m s−1); (c) streamfunction at 150 hPa, contour interval 1.0 × 106 m2 s−1. Both (b) and (c) are from reanalysis, and anomalies are based on a 1968–96 climatology

Citation: Journal of Climate 16, 15; 10.1175/1520-0442(2003)016<2526:UBIAAM>2.0.CO;2

Fig. 13.
Fig. 13.

The 5-yr running means of the monthly mean time series in Figs. 1 and 2. The Niño-3.4 SST index is shown in red, global relative AAM of the NCEP reanalysis in blue, and the ensemble mean of the NCAR CCM3-simulated global relative AAM in black. The gray shading indicates one std dev among the 12 ensemble members in the CCM3 simulations. Units are °C for Niño-3.4 index, AMU for angular momentum

Citation: Journal of Climate 16, 15; 10.1175/1520-0442(2003)016<2526:UBIAAM>2.0.CO;2

Save