1. Introduction
Climate variability in the tropical Atlantic Ocean is affected by local factors as well as by the remote influence of the El Niño–Southern Oscillation (ENSO) phenomenon in the tropical Pacific (Saravanan and Chang 2000; Chang et al. 2006). An important manifestation of this remote influence is the strong correlation between ENSO and hurricanes over the tropical Atlantic at interannual time scales. El Niño events tend to suppress hurricane activity over the tropical Atlantic through increased vertical wind shear (Goldenberg and Shapiro 1996; Vitart and Anderson 2001; Shaman et al. 2009) as well as increased stability associated with warmer upper troposphere (Tang and Neelin 2004). In general, increased vertical wind shear is associated with less active Atlantic hurricane seasons (Gray 1984; DeMaria 1996). Hence, one aspect of the ENSO–hurricane relationship can be explained by the remote influence of ENSO on vertical wind shear over the tropical Atlantic, with a warm ENSO event being associated with stronger vertical wind shear and vice versa (Aiyyer and Thorncroft 2006; Shaman et al. 2009).
This ENSO–shear relationship can be quantitatively illustrated by the regression of July–October (JASO) mean vertical wind shear in the tropical Atlantic on the Niño-3 index1 in observations from 1950 to 1999 (Figs. 1a,b). When warming occurs over the Niño-3 region, there is increased vertical wind shear over a region extending from the Caribbean Sea to northwestern Africa. Over the main development region (MDR; 8°–20°N, 65°–20°W), the increased vertical wind shear during El Niño events is evident on the western side, indicating an unfavorable condition for hurricane development. While in the southeastern part of the MDR, there is decreased vertical wind shear during El Niño years (Figs. 1a,b). This implies that during an El Niño, storms that may form will have difficulty intensifying while traveling across the western Atlantic (Kossin et al. 2010). As a result, despite the warmer Atlantic SST anomalies associated with an El Niño (Alexander et al. 2005), fewer hurricanes occur over the tropical Atlantic, especially over the western Atlantic warm pool.
Regression (m s−1) between vertical wind shear and the Niño-3 index during the JASO season for (a) NCEP reanalysis, (b) ERA-40, (c) CCSM3 coupled twentieth-century integration, (d) GFDL CM2.1 coupled twentieth-century integration, (e) CAM3 AMIP integration, and (f) GFDL AM2.1 AMIP integrations. Solid rectangle denotes the MDR region, and shading denotes the 95% significance level.
Citation: Journal of Climate 25, 3; 10.1175/JCLI-D-11-00213.1
By considering the ENSO–shear relationship simulated by a coupled general circulation model (CGCM) and an atmospheric general circulation model (AGCM) and comparing them to observations, we note some interesting differences between the coupled and uncoupled simulations of this relationship. Using observed SST as the boundary condition, the ENSO–shear relationship simulated in some uncoupled AGCMs tends to be opposite of that in coupled model simulations and observations. We hypothesize that these differences arise because of the nonlinear superposition of ENSO-induced anomalous flow upon the mean flow as simulated by uncoupled models. This hypothesis is then validated by constructing a simple regression model for estimating the remote influence of ENSO on the mean zonal winds in the tropical Atlantic.
2. Datasets
For observational data from 1950 to 1999, we use the wind at 200 and 850 hPa from the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis (Kalnay et al. 1996) and the 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-40; Uppala et al. 2005), starting from 1958. We also use the Met Office Hadley Centre Sea Ice and Sea Surface Temperature dataset (HadISST) (Rayner et al. 2003). We analyze simulations from two different CGCMs, the Community Climate System Model, version 3 (CCSM3), and the Geophysical Fluid Dynamics Laboratory Climate Model, version 2.1 (GFDL CM2.1), using 50 yr of data for the period 1950–99, obtained from the Coupled Model Intercomparison Project phase 3 (CMIP3; Meehl et al. 2007), archived in support of the Intergovernmental Panel on Climate Change Fourth Assessment Report (IPCC AR4). In addition to the coupled integrations, we also analyze uncoupled AGCM integrations using observed SST forcing in the Atmospheric Model Intercomparison Project (AMIP). Data for AMIP-style integrations from IPCC AR4 are provided for the Community Atmosphere Model, version 3 (CAM3; the AGCM for CCSM3), from 1978 to 2000 and the GFDL Atmosphere Model, version 2.1 (GFDL AM2.1; the AGCM for GFDL CM2.1), from 1980 to 1999. To increase the sample size for AMIP-style simulations, we conducted an additional 50-yr AMIP integration with the CAM3, which shows similar results as the 23-yr integration for the CAM3 from the IPCC AR4. This implies that the results for 20-yr AMIP-style integrations from IPCC AR4 are robust regardless of the sampling. All data are analyzed from July to October. In this study, we compute vertical wind shear as (|V200 − V850|), where V200 is the horizontal wind vector at 200 hPa and V850 is the horizontal wind vector at 850 hPa (Goldenberg et al. 2001; Vecchi and Soden 2007).
3. Simulated ENSO–shear relationships
The observed regressions between the Niño-3 index and the vertical wind shear in the northern tropical Atlantic are shown for the NCEP and ECMWF reanalyses (Figs. 1a,b). Compared to observations, the coupled models (NCAR CCSM3 and GFDL CM2.1) are fairly successful in capturing the positive regression between ENSO and the vertical wind shear over much of the MDR and the Caribbean (Figs. 1c,d). However, the same ENSO–shear regression for the uncoupled CAM3 integration exhibits a relationship of the opposite sign (Fig. 1e), with warm ENSO events corresponding to decreased wind shear over the entire region between 10° and 20°N. The uncoupled GFDL AM2.1 also predominantly exhibits a weakening of shear over the MDR, although it does not extend into the Caribbean (Fig. 1f).
The vertical wind shear over the tropical Atlantic is connected with the ENSO-related shift of equatorial convection in the Pacific and changes in the Pacific Walker circulation (Pielke and Landsea 1999; Vecchi and Soden 2007). Increased vertical wind shear in the Atlantic is associated with a decrease in the strength of the Walker circulation. The main contribution from the weaker and eastward-shifted Pacific Walker circulation to the increased vertical wind shear is from upper-tropospheric zonal wind anomalies. With stronger convection over the eastern Pacific, there is stronger outflow at upper levels and hence enhanced westerlies over the tropical Atlantic (Klein et al. 1999). Therefore, we carry out regressions of the zonal wind against the Niño-3 index. At the 200-hPa pressure level, we note that warm ENSO events are associated with anomalous westerlies over much of the tropical Atlantic in observations (Fig. 2a). The CAM3 and GFDL AM2.1 regressions also show westerly anomalies at 200 hPa associated with warm ENSO events, even though the amplitudes are weaker (Figs. 2c,e). The regression analysis for the lower-tropospheric 850-hPa zonal wind (not shown) indicates increased easterly trade wind anomalies, consistent with a weakened Pacific Walker circulation.
Regression (m s−1) of zonal wind at 200 hPa against the Niño-3 index during the JASO season for (a) observations (NCEP), (c) CAM3, and (e) GFDL AM2.1. JASO mean zonal wind at 200 hPa for (b) NCEP, (d) CAM3 and (f) GFDL AM2.1. Shading denotes 95% significance level.
Citation: Journal of Climate 25, 3; 10.1175/JCLI-D-11-00213.1
From the above-mentioned analysis, both observations and AMIP-style integrations agree on there being an upper-level westerly anomaly and a weaker lower-level easterly anomaly associated with warm ENSO events, even though they disagree on the change in the total wind shear. Since vertical wind shear is a nonlinear measure, one needs to superpose the anomalous winds over the background winds before computing the shear. To underscore this point, we examine the JASO mean zonal wind at 200 hPa in the observations and the uncoupled model simulations. In observations, the meridional extent of the mean easterly flow at 200 hPa over the tropical Atlantic is about 10° latitude, extending from the equator to 10°N (Fig. 2b). Over the northwestern MDR and the Caribbean, the observed mean flow at 200 hPa becomes westerly. In the CAM3 simulation, the background flow at 200 hPa is strong easterly from the equator to 20°N, with maximum speeds reaching 24 m s−1 (Fig. 2d). Both the meridional extent and strength of the simulated zonal flow at 200 hPa are twice that in the observations. Similar to CAM3, easterlies are also overestimated in the GFDL AM2.1 experiment (Fig. 2f). This difference between the simulated mean flow and the observations indicates that differences in the simulated ENSO–shear relationship may arise from the nonlinear superposition between the ENSO-induced anomalous flow and the mean flow in the tropical Atlantic, especially over the western Atlantic warm pool (Wang and Lee 2007). In the following section, we discuss this mechanism in more detail and validate it using a simple regression model.
4. Nonlinear superposition of mean and anomalous flow
The mean winds over the northern tropical Atlantic are controlled by local parameters, such as the distribution of Atlantic SST and large-scale jet structures, as well as by the subtropical jet and the tropical easterly jet (Aiyyer and Thorncroft 2006). This local flow pattern is modulated by the remote influence of ENSO, which affects the Walker circulation. The superposition of mean flow and anomalous flow is illustrated schematically in Fig. 3. The observed mean upper-tropospheric flow is westerly and the lower-tropospheric flow is easterly. The positive SST anomaly associated with a warm ENSO event induces an anomalous Pacific Walker circulation, which strengthens the upper-level westerlies as well as the lower-level easterlies in the Atlantic, leading to an increase in tropical Atlantic vertical wind shear (Fig. 1a). However, in the CAM3 simulation, the mean upper-level flow is easterly. The upper-level westerly anomalies associated with a warm ENSO event weaken the upper-level easterlies in the CAM3 simulation, whereas the lower-level easterlies increase slightly. Thus, a warm ENSO event is associated with reduced vertical wind shear in CAM3 (Fig. 1e)
Schematic longitude–height section illustrating the nonlinear superposition of the remote influence of ENSO and the mean zonal wind over the western portion of the northern tropical Atlantic region (July–October). Black solid arrows denote the 200- and 850-hPa mean zonal winds in observations. Red dashed arrows denote the zonal winds during a warm ENSO event in the observations. Black dashed–dotted arrows denote the 200-hPa mean zonal winds in the CAM3 simulation. Blue dotted arrows denote the zonal wind during a warm ENSO event, as simulated by CAM3. Note that the warm event leads to increased shear in the observations but decreased shear in the CAM3 simulation, because of the easterly bias in the simulation of the upper-level zonal wind.
Citation: Journal of Climate 25, 3; 10.1175/JCLI-D-11-00213.1
To quantitatively verify the mechanism illustrated in Fig. 3, we construct a regression model to estimate the vertical shear anomalies associated with ENSO. As the zonal wind component makes the largest contribution to vertical wind shear, we compute the vertical wind shear as the difference of zonal wind between 200 and 850 hPa, |u200 − u850|. Next, we decompose the upper-level zonal wind u200 into contributions from the mean flow 
Figure 4a shows the ENSO-induced vertical shear anomaly as computed using the regression model for observations. The regression model compares fairly well with the observed ENSO–vertical shear relationship (Fig. 1a). In particular, the positive correlation in the western portion of the MDR and the Caribbean is captured well. When applied to the CAM3 simulations, the regression results (Fig. 4b) are again consistent with the direct estimates of the ENSO–vertical shear relationship (Fig. 1e). In particular, the regression model demonstrates a weaker vertical wind shear during warm ENSO events over the Caribbean and the MDR. The regression model applied to GFDL AM2.1 (Fig. 4c) also reproduces the negative ENSO–shear relationship over the MDR (Fig. 1f). Furthermore, if we recompute the regression model for CAM3 by replacing the simulated mean flow at 200 hPa with the observed mean flow, we obtain a more realistic ENSO–shear relationship (Fig. 4d). The success of the simple regression model confirms the mechanism for nonlinear superposition between mean flow and anomalous flow schematically illustrated in Fig. 3.
Regression (m s−1) model estimation of ENSO-induced vertical shear anomaly, 
Citation: Journal of Climate 25, 3; 10.1175/JCLI-D-11-00213.1
5. Discussion and conclusions
We have analyzed the ENSO–shear relationship over the northern tropical Atlantic as simulated by coupled and uncoupled models. The uncoupled CAM3 and GFDL AM2.1 atmospheric models forced with observed SST produced an incorrect simulation of the ENSO–shear relationship, with decreased vertical wind shear during an El Niño, as compared to both observations and coupled model simulations. In the observations, stronger vertical wind shear associated with El Niño events occurs from the Caribbean to northwestern Africa, where it is controlled by the upper-level westerlies. In uncoupled AMIP-style experiments, the ENSO-induced flow anomalies are the same as those in the observations, but the simulated upper-level mean zonal wind has an easterly bias over the MDR and the Caribbean. It is in these regions that uncoupled models simulate an incorrect relationship between vertical wind shear and ENSO.
The difference between the uncoupled model simulations and observation indicates that although there is a strong remote influence of ENSO on vertical wind shear, the local background mean flow may alter the sign of the relationship between ENSO and vertical wind shear (Fig. 3). Furthermore, we find that a simple regression model captures important features of the ENSO–shear relationships, both in observations and CAM3 simulations, thus validating the nonlinearity of the superposition of the mean and anomalous flows. To further validate the influence of the mean flow on ENSO–shear correlations, other AMIP-style experiments from IPCC AR4 were also analyzed (Fig. 5). Some models, such as NCAR CAM3 and GFDL AM2.1, exhibit strong easterly biases over the MDR, leading to negative ENSO–shear correlations, while other models, such as L’Institut Pierre-Simon Laplace Coupled Model, version 4 (IPSL CM4), Meteorological Research Institute Coupled General Circulation Model, version 2.3.2a (MRI CGCM2.3.2a), and Hadley Centre Global Environmental Model, version 1 (HadGEM1), tend to have strong westerly biases and exhibit enhanced positive ENSO–shear correlations.
Histogram of biases of mean zonal flow at 200 hPa (m s−1) averaged over the MDR, for nine uncoupled AMIP simulations using IPCC models. Slanted-line shading indicates significant negative ENSO–shear correlation over the MDR, and parallel-line shading indicates significant positive ENSO–shear correlation (no shading indicates weak correlations).
Citation: Journal of Climate 25, 3; 10.1175/JCLI-D-11-00213.1
Interestingly, the analysis of the IPCC coupled model simulations indicates much better agreement between the simulated and observed ENSO–shear relationships (Shaman and Maloney 2012). This is puzzling given that the coupled simulations exhibit equally severe mean flow biases as the uncoupled simulations. Preliminary analyses lead us to hypothesize that the biases in the coupled models’ ENSO simulations may counteract the influence of the mean flow biases on the ENSO–shear relationship. The validity of this hypothesis is currently under investigation and the results will be reported later. Another important issue is the cause of the easterly bias in uncoupled simulations that leads to the wrong ENSO–shear relationship. A likely candidate for this zonal wind bias is model systematic errors in land surface and convection parameterizations, leading to deficient precipitation over equatorial South America and excessive precipitation over equatorial Africa (Richter and Xie 2008; Richter et al. 2012).
The nonlinear superposition of mean flow and anomalous flow has potential implications for climate variability. In addition to ENSO, phenomena local to the Atlantic, such as the Atlantic multidecadal oscillation (AMO), can lead to interdecadal modulations of the mean flow over the northern tropical Atlantic (Knight et al. 2006; Wang et al. 2008). In the context of climate change, there is considerable interest in the trends in vertical shear over the northern tropical Atlantic region (Vecchi and Soden 2007). Our study suggests that biases in the mean states can have a large influence on the simulated shear anomalies. This is true in the MDR and also in the shear enhancement region (13°S–25°N, 40°–90°W), as identified by Vecchi and Soden (2007). Therefore, climate models with significant errors in the mean flow simulations may need to be excluded from any estimate of ensemble-averaged vertical shear trends.
Acknowledgments
This work is supported by NSF Grant ATM-0620624 and NOAA Grant NA09OAR4310135. We thank the anonymous reviewers for their useful suggestions. We thank the various modeling groups for making their data available via the Program for Climate Model Diagnosis and Intercomparison (PCMDI) archive. We thank the Kyoto University Global COE program and Dr. Shigeo Yoden for summer hosting. Ping Chang also acknowledges support from the National Basic Research Program of China (Grant 2007CB816005), the National Natural Science Foundation of China (NSFC) (Grants 40730843 and 41028005), and the Chinese Ministry of Education’s 111 project (Grant B07036).
REFERENCES
Aiyyer, A. R., and C. Thorncroft, 2006: Climatology of vertical wind shear over the tropical Atlantic. J. Climate, 19, 2969–2983.
Alexander, M. A., I. Blade, M. Newman, J. R. Lanzante, N.-C. Lau, and J. D. Scott, 2005: The atmospheric bridge: The influence of ENSO teleconnection on air–sea interaction over the global oceans. J. Climate, 15, 2205–2231.
Chang, P., Y. Fang, R. Saravanan, L. Ji, and H. Seidel, 2006: The cause of the fragile relationship between the Pacific El Niño and the Atlantic Niño. Nature, 443, 324–328.
DeMaria, M., 1996: The effect of vertical shear on tropical cyclone intensity change. J. Atmos. Sci., 53, 2076–2087.
Goldenberg, S. B., and L. J. Shapiro, 1996: Physical mechanisms for the association of El Niño and West African rainfall with Atlantic major hurricane activity. J. Climate, 9, 1169–1187.
Goldenberg, S. B., C. Landsea, A. M. Mestas-Nunez, and W. M. Gray, 2001: The recent increase in Atlantic hurricane activity: Causes and implications. Science, 293, 474–479.
Gray, W. M., 1984: Atlantic seasonal hurricane frequency. Part I: El Niño and 30 mb quasi-biennial oscillation influences. Mon. Wea. Rev., 112, 1649–1668.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437–470.
Klein, S. A., B. J. Soden, and N.-C. Lau, 1999: Remote sea surface temperature variations during ENSO: Evidence for a tropical atmospheric bridge. J. Climate, 12, 917–932.
Knight, J. R., C. K. Folland, and A. A. Scaife, 2006: Climate impacts of the Atlantic multidecadal oscillation. Geophys. Res. Lett., 33, L17706, doi:10.1029/2006GL026242.
Kossin, J. P., S. J. Camargo, and M. Sitkowski, 2010: Climate modulation of North Atlantic hurricane tracks. J. Climate, 23, 3057–3076.
Meehl, G. A., C. Covey, T. Delworth, M. Latif, B. McAvaney, J. F. B. Mitchell, R. J. Stouffer, and K. E. Taylor, 2007: The WCRP CMIP3 multimodel dataset: A new era in climate change research. Bull. Amer. Meteor. Soc., 88, 1383–1394.
Pielke, J. R. A., Jr., and C. N. Landsea, 1999: La Nina, El Niño, and Atlantic hurricane damages in the United States. Bull. Amer. Meteor. Soc., 80, 2027–2033.
Rayner, N. A., D. E. Parker, E. B. Horton, C. K. Folland, L. V. Alexander, D. P. Rowell, E. C. Kent, and A. Kaplan, 2003: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108, 4407, doi:10.1029/2002JD002670.
Richter, I., and S.-P. Xie, 2008: On the origin of equatorial Atlantic biases in coupled general circulation models. Climate Dyn., 31, 587–598.
Richter, I., S.-P. Xie, A. T. Witternberg, and Y. Masumoto, 2012: Tropical Atlantic biases and their relation to surface wind stress and terrestrial precipitation. Climate Dyn., doi:10.1007/s00382-011-1038-9, in press.
Saravanan, R., and P. Chang, 2000: Interaction between tropical Atlantic variability and El Niño–Southern Oscillation. J. Climate, 13, 2177–2194.
Shaman, J., and E. D. Maloney, 2012: Shortcomings in climate model simulations of the ENSO-Atlantic hurricane teleconnection. Climate Dyn., doi:10.1007/s00382-011-1075-4, in press.
Shaman, J., S. K. Esbensen, and E. D. Maloney, 2009: The dynamics of the ENSO–Atlantic hurricane teleconnection: ENSO-related changes to the North African–Asian jet affect Atlantic basin tropical cyclogenesis. J. Climate, 22, 2458–2482.
Tang, B. H., and J. D. Neelin, 2004: ENSO influence on Atlantic hurricanes via tropospheric warming. Geophys. Res. Lett., 31, L24204, doi:10.1029/2004GL021072.
Uppala, S. M., and Coauthors, 2005: The ERA-40 Re-Analysis. Quart. J. Roy. Meteor. Soc., 131, 2961–3012.
Vecchi, G. A., and B. J. Soden, 2007: Increased tropical Atlantic wind shear in model projections of global warming. Geophys. Res. Lett., 34, L08702, doi:10.1029/2006GL028905.
Vitart, F., and J. L. Anderson, 2001: Sensitivity of Atlantic tropical storm frequency to ENSO and interdecadal variability of SSTs in an ensemble of AGCM integrations. J. Climate, 14, 533–545.
Wang, C., and S.-K. Lee, 2007: Atlantic warm pool, Caribbean low-level jet, and their potential impact on Atlantic hurricanes. Geophys. Res. Lett., 34, L02703, doi:10.1029/2006GL028579.
Wang, C., S.-K. Lee, and D. Enfield, 2008: Atlantic warm pool acting as a link between Atlantic multidecadal oscillation and Atlantic tropical cyclone activity. Geochem. Geophys. Geosyst., 9, Q05V03, doi:10.1029/2007GC001.
The Niño-3 index is defined as the regionally averaged SST anomaly over the Niño-3 region (5°S–5°N, 90°–150°W).
