• Blackmon, M. L., 1976: A climatological spectral study of the 500 mb geopotential height of the Northern Hemisphere. J. Atmos. Sci., 33 , 16071623.

    • Search Google Scholar
    • Export Citation
  • Branstator, G., 1992: The maintenance of low-frequency atmospheric anomalies. J. Atmos. Sci., 49 , 19241945.

  • Branstator, G., 1995: Organization of storm track anomalies by recurring low-frequency circulation anomalies. J. Atmos. Sci., 52 , 207226.

    • Search Google Scholar
    • Export Citation
  • Broccoli, A. J., , and S. Manabe, 1992: The effects of orography on midlatitude Northern Hemisphere dry climates. J. Climate, 5 , 11811201.

    • Search Google Scholar
    • Export Citation
  • Cai, M., , and M. Mak, 1990: On the basic dynamics of regional cyclogenesis. J. Atmos. Sci., 47 , 14171442.

  • Chang, E. K. M., 2006: An idealized nonlinear model of the Northern Hemisphere winter storm tracks. J. Atmos. Sci., 63 , 18181839.

  • Chang, E. K. M., , and Y. Guo, 2007: Dynamics of the stationary anomalies associated with the interannual variability of the midwinter Pacific storm track—The roles of tropical heating and remote eddy forcing. J. Atmos. Sci., 64 , 24422461.

    • Search Google Scholar
    • Export Citation
  • Chang, E. K. M., , and P. Zurita-Gotor, 2007: Simulating the seasonal cycle of the Northern Hemisphere storm tracks using idealized nonlinear storm-track models. J. Atmos. Sci., 64 , 23092331.

    • Search Google Scholar
    • Export Citation
  • Chang, E. K. M., , S. Lee, , and K. L. Swanson, 2002: Storm track dynamics. J. Climate, 15 , 21632183.

  • Charney, J. G., , and A. Eliassen, 1949: A numerical method for predicting the perturbations of the middle latitude westerlies. Tellus, 1 , 3854.

    • Search Google Scholar
    • Export Citation
  • Collins, W. D., and Coauthors, 2004: Description of the NCAR Community Atmosphere Model (CAM 3.0). National Center for Atmospheric Research Tech. Rep. NCAR/TN-464+STR, 210 pp.

    • Search Google Scholar
    • Export Citation
  • Collins, W. D., and Coauthors, 2006: The formulation and atmospheric simulation of the Community Atmospheric Model version 3 (CAM3). J. Climate, 19 , 21442161.

    • Search Google Scholar
    • Export Citation
  • Davies, C. A., 1997: The modification of baroclinic waves by the Rocky mountains. J. Atmos. Sci., 54 , 848868.

  • Franzke, C., , K. Fraedrich, , and F. Lunkeit, 2001: Teleconnections and low-frequency variability in idealized experiments with two storm tracks. Quart. J. Roy. Meteor. Soc., 127 , 13211339.

    • Search Google Scholar
    • Export Citation
  • Frierson, D. M. W., , I. M. Held, , and P. Zurita-Gotor, 2006: A gray-radiation aquaplanet GCM. Part I: Static stability and eddy scale. J. Atmos. Sci., 63 , 25482566.

    • Search Google Scholar
    • Export Citation
  • Hayashi, Y., , and D. Golder, 1981: The effects of condensational heating on midlatitude transient waves in their mature stage: Control experiments with a GFDL GCM. J. Atmos. Sci., 38 , 25322539.

    • Search Google Scholar
    • Export Citation
  • Held, I. M., , and M. J. Suarez, 1994: A proposal for the intercomparison of the dynamical cores of atmospheric general circulation models. Bull. Amer. Meteor. Soc., 75 , 18251830.

    • Search Google Scholar
    • Export Citation
  • Held, I. M., , M. Ting, , and H. Wang, 2002: Northern winter stationary waves: Theory and modeling. J. Climate, 15 , 21252144.

  • Hoskins, B. J., , and P. J. Valdes, 1990: On the existence of storm tracks. J. Atmos. Sci., 47 , 18541864.

  • Lee, S., , and H-K. Kim, 2003: The dynamical relationship between subtropical and eddy-driven jets. J. Atmos. Sci., 60 , 14901503.

  • Lee, W-J., , and M. Mak, 1996: The role of orography in the dynamics of storm tracks. J. Atmos. Sci., 53 , 17371750.

  • Mak, M., , and Y. Deng, 2007: Diagnostic and dynamical analyses of two outstanding aspects of storm tracks. Dyn. Atmos. Oceans, 43 , 8099.

    • Search Google Scholar
    • Export Citation
  • Orlanski, I., 1998: Poleward deflection of storm tracks. J. Atmos. Sci., 55 , 25772602.

  • Orlanski, I., , and B. D. Gross, 1994: Orographic modification of cyclone development. J. Atmos. Sci., 51 , 589612.

  • Ringler, T. D., , and K. H. Cook, 1999: Understanding the seasonality of orographically forced stationary waves: Interaction between mechanical and thermal forcing. J. Atmos. Sci., 56 , 11541174.

    • Search Google Scholar
    • Export Citation
  • Seager, R., , D. S. Battisti, , J. Yin, , N. Gordon, , N. Naik, , A. C. Clement, , and M. A. Cane, 2002: Is the Gulf Stream responsible for Europe’s mild winters? Quart. J. Roy. Meteor. Soc., 128 , 25632586.

    • Search Google Scholar
    • Export Citation
  • Simmons, A. J., , and B. J. Hoskins, 1978: The life cycles of some nonlinear baroclinic waves. J. Atmos. Sci., 35 , 14541477.

  • Smagorinsky, J., 1953: The dynamical influence of large-scale heat sources and sinks on the quasi-stationary mean motions of the atmosphere. Quart. J. Roy. Meteor. Soc., 79 , 342366.

    • Search Google Scholar
    • Export Citation
  • Valdes, P. J., , and B. J. Hoskins, 1989: Linear stationary wave simulations of the time-mean climatological flow. J. Atmos. Sci., 46 , 25092527.

    • Search Google Scholar
    • Export Citation
  • von Storch, H., 1995: Spatial patterns: EOFs and CCA. Analysis of Climate Variability, H. von Storch and A. Navarra, Eds., Springer, 227–258.

    • Search Google Scholar
    • Export Citation
  • Wallace, J. M., , G. H. Lim, , and M. L. Blackmon, 1988: Relationship between cyclone tracks, anticyclone tracks, and baroclinic waveguides. J. Atmos. Sci., 45 , 439462.

    • Search Google Scholar
    • Export Citation
  • Wang, H., , and M. Ting, 1999: Seasonal cycle of the climatological stationary waves in the NCEP–NCAR reanalysis. J. Atmos. Sci., 56 , 38923919.

    • Search Google Scholar
    • Export Citation
  • Zurita-Gotor, P., , and E. K. M. Chang, 2005: The impact of zonal propagation and seeding on the eddy-mean flow equilibrium of a zonally varying two-layer model. J. Atmos. Sci., 62 , 22612273.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    (a) Zonal mean distribution for Q for the control experiment, contour interval 0.2 K day−1. (b) Average of Q from 900 to 100 hPa, contour interval 0.5 K day−1. (c) Orography used in control experiment, contour interval 500 m. Shaded regions denote positive values in (a) and (b), and larger than 1000 m in (c).

  • View in gallery

    (a) Zonal asymmetrical part of the time mean 300-hPa streamfunction (contour interval 3 × 106 m2 s−1); (b) standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m) computed from January 1982 to 1994 based on the NCEP–NCAR reanalysis. Shaded regions denote negative values in (a), and larger than 100 m in (b). (c), (d) As in (a), (b) but from 50 months of control model simulation. R, A, and V correspond to pattern correlation between the observed and modeled patterns, normalized amplitude of the modeled pattern, and percentage of observed variance explained by the modeled pattern. See (4) for definitions.

  • View in gallery

    Zonal asymmetrical part of the time mean 300-hPa streamfunction (contour interval 3 × 106 m2 s−1) taken from (a) the NO OROG–FULL HEAT, (b) the NO OROG–TROP, and (c) the NO OROG–NH EXTRA experiment. Difference in the zonal asymmetrical part of the time mean 300-hPa streamfunction (contour interval 3 × 106 m2 s−1) between the control experiment and (d) the FULL OROG–ZONAL HEAT, (e) the FULL OROG–NH EXTRA, and (f) the FULL OROG–TROP experiment. Shaded regions denote negative values; r, a, and v correspond to pattern correlation between the control and sensitivity patterns, normalized amplitude of the sensitivity pattern, and percentage of control variance explained by the sensitivity pattern, respectively. See (5) for definitions.

  • View in gallery

    Zonal asymmetrical part of the time mean 300-hPa streamfunction (contour interval 3 × 106 m2 s−1), taken from (a) the FULL OROG-ZONAL HEAT, (b) the ONLY TIBET–ZONAL HEAT, and (c) the ONLY ROCKIES–ZONAL HEAT experiments. Difference in the zonal asymmetrical part of the time mean 300-hPa streamfunction (contour interval 3 × 106 m2 s−1) between the control experiment and (d) the NO OROG–FULL HEAT, (e) the NO TIBET–FULL HEAT, and (f) the NO ROCKIES–FULL HEAT experiment. Shaded regions denote negative values.

  • View in gallery

    Zonal asymmetrical part of the time mean 300-hPa streamfunction (contour interval 3 × 106 m2 s−1) taken from (a) the stationary wave model experiment forced with orographic forcings only and (b) the experiment forced with eddy vorticity fluxes taken from the FULL OROG–ZONAL HEAT experiment. (c) Difference in the zonal asymmetrical part of the time mean 300-hPa streamfunction (contour interval 3 × 106 m2 s−1) between the control experiment and an experiment similar to the NO ROCKIES–FULL HEAT experiment except that the zonal mean is damped toward the observed zonal mean. Shaded regions denote negative values.

  • View in gallery

    Zonal asymmetrical part of the time mean 700-hPa streamfunction (contour interval 3 × 106 m2 s−1) taken from (a) the NCEP–NCAR reanalysis from January 1982 to 1994: (b) the NO OROG–NH EXTRA, (c) the NO OROG–FULL HEAT, and (d) the control experiment; (e) the stationary wave model experiment forced with orographic and diabatic forcings; (f) similar to (e) but the zonal mean is not damped toward the observed zonal mean. Shaded regions denote negative values.

  • View in gallery

    Standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, values larger than 100 shaded) computed from (a) the control, (b) the NO OROG–FULL HEAT, (c) the NO OROG–TROP, (d) the NO OROG–NH EXTRA, (e) the FULL OROG–ZONAL HEAT, (f) the ONLY TIBET-ZONAL HEAT, and (g) the ONLY ROCKIES–ZONAL HEAT experiment. (h)–(n) Difference in standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, shaded values denote statistically significant at the 95% level based on a standard t test) between each of the rhs panels and the NO OROG–ZONAL HEAT experiment.

  • View in gallery

    Standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, values larger than 100 shaded) computed from (a) the NO OROG–ZONAL HEAT, (b) the FULL OROG–ZONAL HEAT, (c) the FULL OROG–NH EXTRA, (d) the FULL OROG–TROP, (e) the NO OROG–FULL HEAT, (f) the NO TIBET–FULL HEAT, and (g) the NO ROCKIES–FULL HEAT experiment. (h)–(n) Difference in standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, shaded values denote statistically significant at the 95% level based on a standard t test) between the control experiment and each of the rhs panels.

  • View in gallery

    (a) Standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, values larger than 100 shaded) computed from an experiment similar to the NO ROCKIES–FULL HEAT experiment except that the zonal mean is damped toward the observed zonal mean. (b) Difference in standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, shaded values denote statistically significant at the 95% level based on a standard t test) between the control experiment and the panel on the left.

  • View in gallery

    (a) Zonally asymmetrical part of the time mean 300-hPa streamfunction (contour interval 3 × 106 m2 s−1), taken from the ONLY TIBET–NH EXTRA experiment. (b) Standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, values larger than 100 shaded) taken from the ONLY TIBET–NH EXTRA experiment. (c) Difference in standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, shaded values denote statistically significant at the 95% level based on a standard t test) between the ONLY TIBET–NH EXTRA experiment and the NO OROG–ZONAL HEAT experiment. (d) Difference in the zonal asymmetrical part of the time mean 300-hPa streamfunction (contour interval 3 × 106 m2 s−1) between the control experiment and the NO TIBET–TROP experiment. (e) Standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, values larger than 100 shaded), taken from the NO TIBET–TROP experiment. (f) Difference in standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, shaded values denote statistically significant at the 95% level based on a standard t test) between the control experiment and the NO TIBET–TROP experiment.

  • View in gallery

    (a)–(c) As in Figs. 10 a–c but for the FULL OROG–NH EXTRA experiment. As in Figs. 10 d–f but for the NO OROG–TROP experiment.

  • View in gallery

    (a) Zonal asymmetrical part of the time mean 300-hPa streamfunction (contour interval 3 × 106 m2 s−1) taken from the control CAM experiment. (b) Standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, values larger than 100 shaded), taken from the control CAM experiment. (c)–(d) As in (a)–(b) but taken from the CAM experiment without mountains.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 48 48 10
PDF Downloads 47 47 10

Diabatic and Orographic Forcing of Northern Winter Stationary Waves and Storm Tracks

View More View Less
  • 1 School of Marine and Atmospheric Sciences, Stony Brook University, Stony Brook, New York
© Get Permissions
Full access

Abstract

In this study, a dry global circulation model is used to examine the contributions made by orographic and diabatic forcings in shaping the zonal asymmetries in the earth’s Northern Hemisphere (NH) winter climate. By design, the model mean flow is forced to bear a close resemblance to the observed zonal mean and stationary waves. The model also provides a decent simulation of the storm tracks. In particular, the maxima over the Pacific and Atlantic, and minima over Asia and North America, are fairly well simulated. The model also successfully simulates the observation that the Atlantic storm track is stronger than the Pacific storm track, despite stronger baroclinicity over the Pacific. Sensitivity experiments are performed by imposing and removing various parts of the total forcings.

In terms of the NH winter stationary waves in the upper troposphere, results of this study are largely consistent with previous studies. Diabatic forcings explain most of the modeled stationary waves, with orographic forcings playing only a secondary role, and feedbacks due to eddy fluxes probably play only minor roles in most cases. Nevertheless, results of this study suggest that eddy fluxes may be important in modifying the response to orographic forcings in the absence of zonal asymmetries in diabatic heating. On the other hand, unlike the conclusion reached by previous studies, it is argued that the convergence of eddy momentum fluxes is important in forcing the oceanic lows in the lower troposphere, in agreement with one’s synoptic intuition.

Regarding the NH winter storm-track distribution, results of this study suggest that NH extratropical heating is the most important forcing. Zonal asymmetries in NH extratropical heating act to force the Pacific storm track to shift equatorward and the Atlantic storm track to shift poleward, attain a southwest–northeast tilt, and intensify. It appears to be the main forcing responsible for explaining why the Atlantic storm track is stronger than the Pacific storm track. Tibet and the Rockies are also important, mainly in suppressing the storm tracks over the continents, forcing a clearer separation between the two storm tracks. In contrast, asymmetries in tropical heating appear to play only a minor role in forcing the model storm-track distribution.

Corresponding author address: Dr. Edmund K. M. Chang, ITPA/SoMAS, Stony Brook University, Stony Brook, NY 11794-5000. Email: kmchang@notes.cc.sunysb.edu

Abstract

In this study, a dry global circulation model is used to examine the contributions made by orographic and diabatic forcings in shaping the zonal asymmetries in the earth’s Northern Hemisphere (NH) winter climate. By design, the model mean flow is forced to bear a close resemblance to the observed zonal mean and stationary waves. The model also provides a decent simulation of the storm tracks. In particular, the maxima over the Pacific and Atlantic, and minima over Asia and North America, are fairly well simulated. The model also successfully simulates the observation that the Atlantic storm track is stronger than the Pacific storm track, despite stronger baroclinicity over the Pacific. Sensitivity experiments are performed by imposing and removing various parts of the total forcings.

In terms of the NH winter stationary waves in the upper troposphere, results of this study are largely consistent with previous studies. Diabatic forcings explain most of the modeled stationary waves, with orographic forcings playing only a secondary role, and feedbacks due to eddy fluxes probably play only minor roles in most cases. Nevertheless, results of this study suggest that eddy fluxes may be important in modifying the response to orographic forcings in the absence of zonal asymmetries in diabatic heating. On the other hand, unlike the conclusion reached by previous studies, it is argued that the convergence of eddy momentum fluxes is important in forcing the oceanic lows in the lower troposphere, in agreement with one’s synoptic intuition.

Regarding the NH winter storm-track distribution, results of this study suggest that NH extratropical heating is the most important forcing. Zonal asymmetries in NH extratropical heating act to force the Pacific storm track to shift equatorward and the Atlantic storm track to shift poleward, attain a southwest–northeast tilt, and intensify. It appears to be the main forcing responsible for explaining why the Atlantic storm track is stronger than the Pacific storm track. Tibet and the Rockies are also important, mainly in suppressing the storm tracks over the continents, forcing a clearer separation between the two storm tracks. In contrast, asymmetries in tropical heating appear to play only a minor role in forcing the model storm-track distribution.

Corresponding author address: Dr. Edmund K. M. Chang, ITPA/SoMAS, Stony Brook University, Stony Brook, NY 11794-5000. Email: kmchang@notes.cc.sunysb.edu

1. Introduction

It is well known that the climate of the earth is not zonally symmetric. Since the forcing at the top of the atmosphere is zonally symmetric when averaged over a day or longer, the zonal asymmetries in the earth’s climate must be forced by zonal asymmetries in the lower boundary. Zonal asymmetries in the lower boundary include asymmetries in land–ocean distribution, and asymmetries in the location of mountains. Together, these asymmetries give rise to asymmetrical distribution of surface temperature and moisture. Mountains directly impact atmospheric flow by exciting stationary waves, whereas asymmetries in surface temperature (especially SST) and moisture manifest themselves in the asymmetrical distribution of diabatic heating. While the true forcings are due to orography and surface properties, in a modeling context, the problem is often recast as forcings owing to orography and diabatic heating, even though (as will be elaborated below) diabatic heating is dependent on the flow and thus should not be regarded as independent of orographic forcing.

Since the early studies by Charney and Eliassen (1949) and Smagorinsky (1953), numerous studies have been performed to examine the roles of orography and diabatic heating in forcing atmospheric stationary waves, and the readers are referred to the article by Held et al. (2002; hereafter HTW02) for a review of the theory and results of stationary wave modeling. HTW02 suggested that Northern Hemisphere (NH) winter stationary waves are most strongly forced by diabatic heating (with tropical and extratropical heating playing similarly important roles), while orographic forcing plays a somewhat lesser role, with zonal asymmetries in the transient eddy vorticity fluxes playing only a minor role.

While stationary wave modeling has given us a tremendous wealth of insights, there are limitations in the approach. On millennial or shorter time scales, the distribution of orographic forcing can certainly be considered to be an externally given independent forcing for the stationary waves. However, it is well known that transient eddies respond to changes in atmospheric low frequency flow (e.g., Cai and Mak 1990; Branstator 1992, 1995), and thus transient eddy vorticity (as well as sensible heat) fluxes should not really be regarded as an independent forcing. As discussed in HTW02, it is also expected that the distribution of diabatic heating (and even SST) is, to certain extent, dependent on the circulation and, hence, to the distribution of the orography. Thus, diabatic and orographic forcings are not necessarily independent, and diabatic forcing is not independent of the circulation itself. In addition, in most implementations of stationary wave modeling, enhanced damping has to be added to remove near-resonant solutions. Since the final amplitude of the stationary wave solution depends on the exact magnitude of the damping used, there are uncertainties in the quantitative predictions based on stationary wave modeling.

As already mentioned above, the distribution of synoptic-scale transient eddies, which is commonly called the storm tracks (Blackmon 1976), is strongly dependent on the distribution of the stationary waves. While a lot of studies have examined the factors that help to shape the distribution of the storm tracks (for a review, see Chang et al. 2002), a number of fundamental questions still remain unanswered. In particular, during midwinter, the Pacific jet is much stronger than the Atlantic jet, implying much stronger baroclinicity (mainly temperature gradient) over the Pacific than over the Atlantic. Yet the midwinter Atlantic storm track is, on average, significantly stronger than the Pacific storm track.

Several studies have attempted to explain this apparent contradiction. Zurita-Gotor and Chang (2005) and Mak and Deng (2007) both suggested that the midwinter Pacific storm track may be suppressed due to strong upstream damping over Asia, suppressing the upstream seeding of the Pacific storm track. On the other hand, Lee and Kim (2003) suggested that the different characteristics of the two jets, in which the Atlantic jet is an eddy-driven jet while the Pacific jet is largely subtropical in character, may be an underlying reason behind the fact that the Atlantic storm track is stronger than its Pacific counterpart. In Lee and Kim, the subtropical characteristics of the jet are largely determined by the strength of tropical heating, implying that the relative amplitude between the two storm tracks may be mainly determined by tropical heating.

As discussed in Chang et al. (2002), the results of Broccoli and Manabe (1992) showed that, in GCM experiments without mountains, the stationary waves are considerably weaker, and the storm tracks more zonally symmetric, even in the presence of land–ocean contrast. In support of this, Lee and Mak (1996) showed that, in a dry nonlinear model driven by relaxation to the observed winter zonal mean temperature distribution, enhanced baroclinicity over the storm-track entrance region could be maintained just by stationary waves induced by mountains alone, without the need for zonal asymmetries in diabatic heating. These results suggest that orographic forcings could be the most important factor in shaping the storm-track distribution.

The discussions above highlight the fact that we still do not have a quantitative understanding of what forces the observed NH winter storm-track distribution. Hoskins and Valdes (1990) attempted to explain the existence of the Pacific and Atlantic storm tracks by explaining the factors that contribute to the enhanced baroclinicity observed at the storm-track entrance regions. Their results suggested that the enhanced baroclinicity can be largely explained by the distribution of midlatitude diabatic heating. Nevertheless, their results also suggested that midlatitude diabatic heating should tend to force stronger enhanced baroclinicity over the Pacific than over the Atlantic, thus still failing to explain why the Atlantic storm track should be stronger than the Pacific storm track.

To a large extent, this paper can be regarded as an extension of HTW02, Chang et al. (2002), and Chang (2006; hereafter C06). In C06, an intermediate model forced by realistic orography and constant diabatic heating is constructed based on a dry dynamical core (Held and Suarez, 1994). C06 demonstrated that the model can provide realistic simulations of the NH winter stationary waves and storm tracks. In this study, the forcings are modified to assess the degree to which the stationary waves and storm tracks in this model are forced by each piece of the forcings. The model and its climate will be described in section 2. In section 3, orographic and diabatic forcing of the stationary waves will be discussed and results are compared to those of HTW02. In section 4, what forces the distribution of the NH winter storm tracks will be addressed. Further discussions and a summary will be presented in sections 5 and 6.

2. Model formulation and climate

a. Model formulation

The model is based on the dynamical core of the Geophysical Fluid Dynamics Laboratory (GFDL) global spectral model (Held and Suarez 1994). The resolution used in this study is T42 in the horizontal and 20 evenly spaced sigma levels in the vertical. Realistic orography, smoothed to model resolution, is imposed (see Fig. 1c). A land–sea mask is used, with stronger surface friction over land. In this version of the model a linear Rayleigh drag is used, with the drag equivalent to a damping on a time scale of 0.5 days at the surface over land and 2 days over the oceans; the damping decreases linearly to zero at σ = 0.7. Diabatic heating is represented by Newtonian cooling on a damping time scale (τ) of 30 days in the free atmosphere (σ < 0.7), decreasing to 2 days at the surface (σ = 1). The only other damping is a highly scale selective diffusion (∇8), with a damping time scale of 0.1 days on the highest wavenumber.

Apart from the orography, the only other forcing imposed is Newtonian cooling to a radiative equilibrium temperature profile. With this parameterization, the first law of thermodynamics can be written as
i1520-0442-22-3-670-e1
where τ is the radiative time scale; θE can be split into two parts, as follows:
i1520-0442-22-3-670-e2
Here θC can be viewed as the desired model climate and Q to be the (initially unknown) diabatic heating distribution. As described in C06, the heating profile (Q) used to force the model is iteratively determined such that, at the end of the process, the model climate, as given by the time mean three-dimensional temperature distribution, is nearly identical to the desired target temperature distribution (θC). More details concerning the model formulation and the iterative procedure can be found in C06. An earlier version of this model has been used to examine the seasonal cycle of the NH storm tracks (Chang and Zurita-Gotor 2007), as well as to understand factors contributing to the interannual variability of the midwinter Pacific storm track (Chang and Guo 2007).
As discussed in C06, when the model is forced to the observed January climatological temperature distribution, the eddy variances and covariances are found to be much weaker than observed. The study of Hayashi and Golder (1981) showed that condensational heating not only acts as a source of eddy available potential energy (EAPE) but, more importantly, it acts to strongly enhance baroclinic energy conversion. One way of mimicking part of this effect is by reducing the static stability of the model atmosphere. Instead of using the observed temperature profile as the target climate, a profile with reduced static stability is imposed as follows:
i1520-0442-22-3-670-e3
where z(p) is the average geopotential height of the pressure surface. In this study, θobs is taken to be the climatological January potential temperature distribution taken from National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis from the Januarys of 1982 to 1994. C06 found that a value of 1.25 K km−1 for S provides realistic amplitudes of eddy fluxes for January. In the current version of the model, it is found that with higher vertical resolution (20 instead of 10 levels) and a different form of surface friction (linear instead of nonlinear), less reduction in static stability is required to obtain realistic eddy amplitudes, and a value of 0.65 K km−1 has been used for S in all experiments discussed in this paper.1 Note that, as discussed in Chang and Zurita-Gotor (2007), we have found that reduction in the static stability does not seem to significantly distort the dynamics of midlatitude baroclinic waves, consistent with the results of Frierson et al. (2006).

b. Climate of the control experiment

As described in C06, once θC is defined, an iterative procedure can be used to obtain Q, starting with an arbitrary first guess (Q0). The final Q obtained (after about 60 iterations), when Q0 is set to be the diabatic heating distribution obtained from the NCEP–NCAR reanalysis for January 1982–1994, is shown in Figs. 1a,b. This heating distribution is quite similar to the one used by HTW02 (see their Fig. 8). Complete sets of experiments similar to those discussed below have also been conducted by deriving Q from different Q0, including the complete absence of initial heating, heating derived from the 40-yr ECMWF Re-Analysis (ERA-40) and NCEP–NCAR reanalysis based on the thermodynamic equation (similar to the procedure used by HTW02), as well as heating taken from a climate model simulation. As discussed in C06, the final Q obtained starting from different first guesses can be different, especially in the tropics, but the conclusions in this paper (except for the response to the removal of the Rockies, see discussions in section 5c) do not appear to be sensitive to the exact form of Q. Thus, the experiment using the heating shown in Fig. 1 is (somewhat arbitrarily) designated as the control experiment. When forced with the diabatic and orographic forcings shown in Fig. 1, the model climate, in terms of its temperature distribution, differs from Tc [temperature corresponding to θc as defined by Eq. (3)] by less than 0.5°C rms. In this study, the model climate is computed from an 1800-day run, using the average from the final 1500 days (50 months). Several extended runs have also been made to confirm that a stable climate can be obtained based on averages over 1500 days.

The upper-tropospheric stationary waves computed from the control experiment, as indicated by the zonal asymmetrical part of the time mean 300-hPa streamfunction, are shown in Fig. 2c. The corresponding observed streamfunction, computed from NCEP–NCAR reanalysis from the Januarys of 1982 to 1994, is shown in Fig. 2a. In this and subsequent figures, the pattern correlation between the modeled and observed patterns (R), the normalized rms amplitude of the modeled pattern (A), and the percentage of observed variance that can be explained by the modeled pattern (V)2 (see von Storch 1995), are printed on top of the panel. Here
i1520-0442-22-3-670-e4
In (4), O represents an observed quantity (e.g., streamfunction), M is its modeled counterpart, and angle brackets represent area-weighted average over the NH. Alternatively, (1 − V) can be interpreted as the normalized rms difference between the modeled and observed patterns.
Later, results from sensitivity experiments will be compared to those from the control experiments. Under those situations, the pattern correlation between the control and sensitivity experiments, the normalized rms amplitude of the sensitivity experiment, and the percentage of control variance that can be explained by the sensitivity experiment will be denoted by the lowercase letters r, a, and v, respectively. That is,
i1520-0442-22-3-670-e5
Here C represents a quantity from the control experiment, while M is its counterpart from a sensitivity experiment.

Comparing Fig. 2c to 2a, we can see that the model performs quite well in simulating the stationary waves. The main deficiency appears to be over central Asia where the anticyclone to the northwest of Tibet is a bit weak. Overall, the modeled pattern explains 93% of the variance of the observed pattern.

The storm-track distribution is shown by the standard deviation of the 24-h filtered (Wallace et al. 1988) 500-hPa geopotential height. In Fig. 2d, the values of R, A, and V shown are computed from the zonal asymmetrical part of what are shown in Figs. 2b and 2d. The pattern correlation is quite high (0.83) but not as high as that for the stationary waves. The main deficiencies appear to be a slight northward displacement of the modeled storm tracks and a downstream displacement of the peak of the Atlantic storm track. Nevertheless, the model successfully simulates the result that the Atlantic storm track is significantly stronger than the Pacific storm track, not only in terms of rms 500-hPa geopotential height, but also in 24-h filtered 300-hPa meridional velocity variance and lower-tropospheric poleward sensible heat flux (not shown). This is crucial, as one of the questions that we want to address is what forces the Atlantic storm track to be stronger than its Pacific counterpart. C06 noted that this particular feature was not robustly reproduced in experiments using 10 levels in the vertical, and it appears that the increased vertical resolution used in this study is crucial for successfully simulating this feature.

c. Modified forcing experiments

For the sensitivity experiments, the forcings shown in Fig. 1 are modified to test the sensitivity of the stationary wave and storm-track distributions to changes in individual pieces of the forcings. Experiments have been performed using the full orography (FULL OROG), no orography (NO OROG), only Rockies (ONLY ROCKIES), only Tibet (ONLY TIBET), or full orography without either the Rockies (NO ROCKIES) or Tibet (NO TIBET). For the diabatic heating, experiments have been performed with the full heating (FULL HEAT), with only the zonally symmetric part of the heating (ZONAL HEAT), with full heating in the NH extratropics (north of 25°N) and only the zonally symmetric part to the south (NH EXTRA), and with full heating south of 25°N and only the zonally symmetric part to its north (TROP). Note that for experiments that have heating made zonally symmetric, both θc and Q in (2) are symmetrical. Each experiment will be designated by a combination of orographic and diabatic forcings imposed. For example, the experiment FULL OROG–FULL HEAT is the control experiment with all forcings imposed shown in Fig. 2, whereas the experiment NO OROG–ZONAL HEAT (i.e., absence of all orographic forcing, and no zonal asymmetries in the diabatic forcings) can be regarded as the baseline experiment for comparison.

As mentioned above, these experiments can be considered extensions of the stationary wave modeling experiments discussed in HTW02. With this model, when the forcing (orography or heating) is changed, the effects due to adjustment of the eddy fluxes to changes in the circulation are automatically modeled, and, as will be discussed later, the eddy feedback turns out to be significant in some cases. In addition, in these sensitivity experiments, the damping used is the same as that used in the control experiment. However, this model shares the limitation of stationary wave modeling that heating is still regarded as an independent forcing; that is, changes in heating owing to changes in the forcing (e.g., orography) are not modeled. Nevertheless, one can consider this to be an intermediate step between stationary wave modeling and fully coupled GCM simulations.

d. Running the model as a stationary wave model

The model can be run as a nonlinear stationary wave model (i.e., free of transient eddies) by increasing the damping parameters to values similar to those used in HTW02. By enhancing the Rayleigh friction to a damping time scale of 1/3 days at the surface (over both ocean and land), decreasing to 25 days for σ less than 0.8, and changing the diffusion to be less scale selective (∇4), with a time scale of 0.05 days for the highest wavenumber, transient eddy activity becomes practically nonexistent. Note also that as in HTW02, when the model is run as a stationary wave model, the zonal mean flow (zonal mean zonal wind and temperature distribution) is damped toward climatology at a time scale of 3 days.

The forcings used to force the stationary wave model include orography, heating, and the divergence of transient eddy heat and vorticity fluxes, all taken from the control experiment. When all forcings are applied, the stationary wave response (not shown) is quite close to that simulated in the control experiment, with r = 0.90, a = 0.99, and v = 80%, demonstrating that the model dynamics are internally consistent. A full suite of experiments similar to those presented in HTW02 has been conducted, and the results are very similar to those shown there and will not be reproduced here. Results from a few selected experiments will be described later.

3. Stationary wave response

a. Diabatic forcing

Results of the experiment forced with full heating but no orography are shown in Fig. 3a. Consistent with the results of HTW02, the stationary wave pattern forced by heating alone3 is very well correlated with the one forced with all forcings (i.e., diabatic heating plus orography), explaining 74% of the variance of the control experiment. The stationary wave pattern forced by tropical heating alone is shown in Fig. 3b, while that forced by NH extratropical heating alone is shown in Fig. 3c. It is clear that extratropical heating forces a larger portion of the solution than tropical heating does. However, it is also clear that there is significant nonlinearity since the sum of Figs. 3b and 3c does not quite make up the stationary wave forced by the full heating (Fig. 3a).

HTW02 introduced the concept of isolated versus full nonlinear response to a forcing. Let the total forcing be denoted by T, while the forcing being considered is denoted by F. Let the nonlinear solution that is forced by any forcing K be denoted by N(K). The isolated nonlinear response to F is simply N(F)—the response to the forcing F when all other forcings are absent. The full nonlinear response to F is defined to be N(T) − N(T − F), that is, the response due to the addition of F starting from the solution when all other forcings except F are present. Under these definitions, the results shown in Figs. 3a–c represent the isolated nonlinear response to total heating, tropical heating, and NH extratropical heating, respectively.

The full nonlinear response to full heating, tropical heating, and NH extratropical heating are shown in Figs. 3d–f, respectively. All three full nonlinear responses have higher amplitude and are better correlated with the control response than the isolated nonlinear responses, consistent with the results of HTW02. In fact, all the responses shown in Fig. 3 are similar to those shown in HTW02, except that the isolated nonlinear response to tropical heating (Fig. 3b) is a bit weak here. The similarities of the responses shown in Fig. 3 to those simulated by stationary wave modeling (HTW02) suggest that eddy feedback owing to changes in the eddy fluxes is not important for understanding the stationary wave response to heating. Anyway, results shown in Fig. 3 support HTW02’s conclusion that heating is responsible for forcing much of the stationary wave response.

b. Orographic forcing

The results discussed in the preceding section suggest that eddy feedback is not important in shaping the stationary wave patterns forced by diabatic heating. However, it turns out that eddy feedback is important in the response to orographic forcing, giving rise to significant differences between the responses found here and those presented in HTW02.

The isolated nonlinear responses to forcing by the full orography, by Tibet, and by the Rockies alone, are shown in Figs. 4a–c respectively. All three responses are negatively correlated to the control response. Moreover, these responses do not resemble wave trains excited from flow across mountains, as in the responses obtained by stationary wave modeling by HTW02.

To understand why these responses are so different from those shown in HTW02, the model is run as a stationary wave model (see section 2d) forced with orography alone. The results (Fig. 5a) are much more similar to the ones obtained by HTW02. The stationary waves forced by the mountains are now positively correlated with those simulated in the control experiment. One major difference between stationary wave modeling and the approach taken here is that here the eddy feedback is modeled. We investigate whether this may be important by conducting an additional experiment with no stationary wave forcing (i.e., NO OROG–ZONAL HEAT), except that the eddy vorticity fluxes taken from the FULL OROG–ZONAL HEAT experiment is used to force the vorticity equation. The resulting stationary wave is shown in Fig. 5b. Comparing Fig. 5b to Fig. 4a, it is apparent that the eddy fluxes force a substantial portion of the stationary wave shown in Fig. 4a. In particular, the trough–ridge–trough pattern near the date line, and ridge–trough–ridge pattern near 70°W, are both clearly forced by the eddy fluxes instead of being the direct response to the orographic forcing (Fig. 5a). In the next section, we will see that, while the stationary waves forced by orography generally have smaller amplitudes than those forced by heating (cf. Figs. 3 and 4), the storm-track responses to orographic forcing are generally as big as those to heating, and thus the eddy feedback can be more important in shaping the total response in the case of orographic forcing.

The full nonlinear responses to orographic forcings are shown in Figs. 4d–f. The full nonlinear response to Tibet (Fig. 4e) bears significant resemblance to that shown in HTW02 and bears little resemblance to the isolated nonlinear response shown in Fig. 4b, suggesting strong nonlinearities in the model response to orography and strong nonlinear interactions between orographic and diabatic forcings. As discussed in HTW02, currently the origin of this nonlinearity is not yet completely understood. Nevertheless, the full nonlinear response suggests that part of the ridge over northwestern North America may be related to forcing by Tibet. In addition, while the isolated nonlinear responses to orography suggest that orographic forcings do not effectively contribute to the total control streamfunction (v is negative for all three cases), if we consider the full nonlinear response instead, the role of orography becomes much more significant. This is consistent with the results shown in Fig. 3a that the isolated nonlinear response to full heating forces a stationary wave response with amplitude only about 70% of the full stationary wave.

The full nonlinear response to the Rockies (Fig. 4f) appears somewhat anomalous. While the pattern close to North America resembles the full nonlinear pattern obtained by HTW02, the amplitude is much larger, and is larger even than the full nonlinear response to the total orographic forcing (Fig. 4d). In addition, as will be discussed later, the model results also suggest significant changes in the storm tracks. Further examination of the model results reveal that, when the Rockies are removed, the jet in the NO ROCKIES–FULL HEAT experiment experiences a significant narrowing and shifts southward, consistent with the changes in the storm tracks (see discussions in section 4b below). The impact of this jet response can be examined by damping the zonal mean flow of the NO ROCKIES–FULL HEAT experiment toward the observed zonal mean flow with a time scale of 3 days. The stationary wave response for this experiment (Fig. 5c) is now confined to near North America, and is close to the response obtained by HTW02. Some recent idealized modeling studies have also suggested that combinations of orographic and diabatic forcings may sometimes excite highly nonlinear responses to the storm-track distribution (E. Gerber 2007, personal communication). Further studies should be conducted to provide better understanding of the dynamics involved. Nevertheless, care should be taken in interpreting these results, because sensitivity experiments that will be discussed in section 5c below suggest that this anomalously large response may not be robust.

c. Lower-tropospheric streamfunction

Up to now, the focus has been on the upper-tropospheric stationary waves. As the iterative procedure to determine Q for the control experiment guarantees that the mean temperature distribution in the model simulation has to be close to that observed, this immediately implies that the baroclinic part of the modeled stationary wave has to be close to that observed. Since the stationary wave in the upper troposphere is dominated by the baroclinic part, the agreement between the modeled and observed stationary waves in the upper troposphere (Figs. 2a and 2c) is not too surprising. However, the streamfunction in the lower troposphere depends a lot more on the surface pressure (or streamfunction) distribution, and this is shaped by momentum balance in the upper troposphere (nonzero surface wind requires nonzero column mean momentum flux convergence to sustain its existence). Hence examination of the stationary waves in the lower troposphere provides a much more stringent test of the fidelity of the model simulation.

As the parameterization of surface friction used in this study is very simplistic, the model solution within the planetary boundary layer is not expected to be realistic. Hence, the streamfunction at 700 hPa is examined here. The streamfunction taken from the NCEP–NCAR reanalysis is shown in Fig. 6a. The solution is dominated by the Aleutian low over the Pacific and the Icelandic low extending from northeastern North America into the northwestern North Atlantic. The stationary wave from the control experiment, when all forcings are applied, is shown in Fig. 6d. It is clear that the model also does very well in simulating the lower-tropospheric streamfunction, with the model solution explaining 84% of the observed variance. As in HTW02, in Fig. 6, the streamfunction forced by NH extratropical heating alone (Fig. 6b) and that forced by full heating (Fig. 6c), both without orographic forcings, are also shown. Comparing Figs. 6b and 6c, it is clear that, even though the lower-tropospheric streamfunction has very small amplitude in tropical areas, the forcing owing to tropical heating is important. Comparison of Figs. 6c and 6d shows that orographic forcing also plays an important role, consistent with the results of HTW02.

HTW02 mentioned that they were surprised to find that eddy vorticity fluxes are not important in forcing the lower-tropospheric streamfunction since “we tend to intuitively think of the wintertime oceanic lows as being the graveyard of extratropical low pressure systems.” To examine this further, the model is run as a stationary wave model (section 2d), and forced with diabatic and orographic forcings only. This stationary wave solution is shown in Fig. 6e. This solution is clearly well correlated with the observed stationary wave pattern. The oceanic lows are clearly present, even without the application of eddy fluxes as part of the forcing. As in HTW02, inclusion of eddy forcings does not improve the stationary wave simulation in the lower troposphere.

However, part of the stationary wave modeling procedure is forcing the zonal mean flow to that observed. In the atmosphere, the zonal mean flow is forced by a combination of diabatic heating, stationary eddy forcings, and transient eddy forcings. Hence, in the absence of transient eddies, the zonal mean flow will be different. In particular, in the presence of surface friction, surface winds will be much weaker without transient eddy momentum flux convergence within the atmosphere above. Thus, another experiment is performed using the stationary wave model forced with diabatic and orographic forcings, but with the zonal mean flow not damped toward that observed. The stationary wave solution from this experiment is shown in Fig. 6f. Now the oceanic lows are much weaker, suggesting that the zonal mean flow, which transient eddy forcings help maintain, is crucial for the maintenance of these lows. One may argue that during the mature stage of baroclinic wave evolution, eddy momentum flux convergence is strongest (e.g., Simmons and Hoskins 1978; see also Orlanski 1998), and the surface cyclone migrates poleward. Viewed this way, it is this poleward migration of surface cyclones and the associated momentum flux convergence during the mature stage of cyclone evolution that helps to maintain the oceanic lows, consistent with one’s synoptic intuition.

4. Storm-track response

a. Isolated nonlinear response

The storm-track responses to the application of individual or groups of forcings are shown in Fig. 7. The full distributions of the storm tracks under each of the forcing scenarios are shown on the left (Figs. 7a–g), while the differences between the storm track in each experiment from the storm track in the NO OROG–ZONAL HEAT experiment are shown on the rhs panels (Figs. 7h–n). Basically, the rhs panels represent the isolated nonlinear responses of the storm track to each of the imposed forcing scenarios. Note that the storm tracks in the NO OROG–ZONAL HEAT experiment is shown in Fig. 8a. In that experiment, the only zonal asymmetrical forcing is land/ocean difference in surface friction, which leads to weak maxima in the storm-track distribution over the Pacific and eastern Atlantic/western Europe.

The storm track simulated in the control experiment is shown in Fig. 7a. Examining Fig. 7h, one can see that the combined effects of diabatic heating and orography lead to a strong suppression of storm-track activity over Asia, a slight equatorward shift of the storm track over the Pacific, and poleward shift and significant enhancement of the Atlantic storm track.

The storm tracks in the NO OROG experiments are shown in Figs. 7b–d. Figure 7b shows the storm tracks forced with full heating. The isolated nonlinear storm-track response (i.e., response relative to the NO OROG–ZONAL HEAT experiment) to full heating (Fig. 7i) correlates highly with the response to all forcings,4 showing an equatorward shift of the Pacific storm track and poleward shift and significant enhancement of the Atlantic storm track. These features are consistent with the stationary wave forced by heating shown in Fig. 3a: the southward shift of the storm track over the Pacific is consistent with the cyclonic anomaly over midlatitude Pacific, and the poleward shift and southwest–northeast tilt of the Atlantic storm track is consistent with the cyclone–anticyclone pair over northeastern North America and northern North Atlantic. The main missing ingredient is the significant suppression over Asia near Tibet. In addition, in this experiment, the Atlantic storm track is clearly stronger than the Pacific storm track (Fig. 7b).

The storm tracks forced by tropical heating alone (Fig. 7c) show only a weak peak over the Pacific and a weak minimum over Asia. The storm-track response to tropical heating (Fig. 7j) bears little resemblance to the response to all forcings. On the other hand, the storm-track response to NH extratropical heating alone (Fig. 7k) bears significant resemblance to the response to all forcings. Again, these are consistent with the stationary waves forced by these two heating distributions (Figs. 3b,c). Note also that, in the experiment forced with NH extratropical heating alone (Fig. 7d), the Atlantic storm track is slightly stronger than the Pacific storm track.

The storm tracks forced with zonally averaged heating but with orographic forcings present are shown in Figs. 7e–g. The presence of mountains clearly suppresses the storm tracks, both over Asia and North America (Fig. 7l), but it appears that much of the suppression over North America is not locally forced due to the Rockies (Fig. 7n), but is remotely forced by Tibet (Fig. 7m). In all three experiments (Figs. 7e–g), the Pacific storm track is stronger than the Atlantic storm track. Note also that the amplitudes of the storm-track response (a in Figs. 7l–m) to orographic forcings are larger than those to diabatic heating (Figs. 7i–k), even though the amplitudes of the stationary wave response are generally smaller (cf. Figs. 3 and 4).

b. Full nonlinear response

The storm-track responses to the removal of individual or group of forcings from the full forcings are shown in Fig. 8. The panels on the lhs (Figs. 8a–g) show the storm-track distributions after each of these (group of) forcings is removed, and the rhs panels (Figs. 8h–n) show the difference between the storm tracks forced by all forcings (Fig. 7a) and each of the panels on the left. As discussed in section 3a, the rhs panels can be interpreted as the full nonlinear response of the storm tracks to each of these forcings.

The storm-track distribution when all orography and zonal asymmetries in diabatic heating are removed is shown in Fig. 8a; this has been described earlier. By definition, the full nonlinear response to all forcings (Fig. 8h) is equivalent to the isolated nonlinear response to all forcings (Fig. 7h).

The storm-track distribution when full heating is removed is shown in Fig. 8b, which shows that the Atlantic storm track is weaker than the Pacific storm track when zonal asymmetries in diabatic heating are removed. The full nonlinear response to full heating (Fig. 8i) is moderately well correlated with the response to all forcings. When NH extratropical heating is removed, the storm-track distribution (Fig. 8d) is very similar to that shown in Fig. 8b, and the full nonlinear response to NH extratropical heating (Fig. 8k) is also very similar to the full nonlinear response to the full heating (Fig. 8i). On the other hand, the full nonlinear response to tropical heating (Fig. 8j) is much weaker and not as well correlated to the response to all forcings (Fig. 8h), and, in that experiment, the Atlantic storm track is still slightly stronger than the Pacific storm track (Fig. 8c). These results, together with those shown in Fig. 7, suggest that NH extratropical heating is essential in forcing the Atlantic storm track to be stronger than the Pacific storm track.

The full nonlinear response to all orographic forcings (Fig. 8l) shows strong suppression of the storm track near Tibet and some moderate suppression near the Rockies. However, in the presence of zonal asymmetries in diabatic heating, the strong upstream suppression only leads to weak suppression of the Pacific storm track. Even when the strong upstream suppression over Asia is removed (seen by the negative full nonlinear response to orography in Fig. 8e, which indicates weaker storm-track activity when orography is added), the Pacific storm track still comes out to be weaker than the Atlantic storm track. This is also the case when only Tibet is removed (Fig. 8f). Comparing Figs. 8l–m to Figs. 7l–m, it can be seen that the full nonlinear response of the storm tracks to orographic forcing is better correlated with the storm track response to all forcings than the isolated nonlinear response to orography, similar to the results for the stationary waves discussed in section 4.

The full nonlinear response to the Rockies shown in Fig. 8n again appears to be anomalously strong, similar to the full nonlinear response of the stationary waves to the Rockies shown in Fig. 4f. As discussed in section 3b, this highly nonlinear response appears to be related to significant changes in the zonal mean flow when the Rockies are removed. Figure 8n shows that the storm track becomes weaker and shift southward when the Rockies are removed. At the same time, the poleward eddy momentum flux becomes weaker, consistent with the southward shift of the zonal mean jet. The storm-track response in the NO ROCKIES–FULL HEAT experiment in which the zonal mean flow is damped toward the observed zonal mean flow is shown in Fig. 9. This experiment suggests that the Rockies act to damp the storm track over the United States but slightly strengthen the Atlantic storm track (Fig. 9b). Nevertheless, even in this experiment when the Rockies are removed, the Atlantic storm track is still stronger than its Pacific counterpart (Fig. 9a).

5. Discussion

a. Relative strength of Pacific and Atlantic storm tracks

The results presented in Figs. 7 and 8 show that, among the four different forcings considered, the storm-track response to NH extratropical heating (Figs. 7k and 8k) correlates best with the storm-track response to all forcings. The results also show that, when only NH extratropical heating is present, the Atlantic storm track is already slightly stronger than the Pacific storm track (Fig. 7d) while, when only NH extratropical heating is taken away (Fig. 8d), the Atlantic storm track is no longer stronger than the Pacific storm track. Figures 7 and 8 show that NH extratropical heating is the only forcing which does that. These results suggest that, for this model, NH extratropical heating is the forcing that contributes most to forcing the Atlantic storm track to be stronger than its Pacific counterpart.

Among the other three forcings, the one that forces a storm-track response that correlates next best with the response to all forcings is Tibet (Figs. 7m and 8m). The response to the application of Tibet and NH extratropical heating together is shown in Fig. 10. Both the isolated (Fig. 10c) and full nonlinear response (Fig. 10f) to this group of forcings do quite well in simulating the storm-track anomalies. However, it is interesting to note that, in the isolated nonlinear experiment (Fig. 10b), the Atlantic storm track is no longer stronger than the Pacific one.

When NH extratropical heating and full orography are imposed together (i.e., with only tropical heating left out), both the isolated (Fig. 11c) and full (Fig. 11f) nonlinear storm-track responses become very close to the response to all forcings, confirming that tropical heating is not very important in shaping the storm-track distribution. Note that, while the removal of tropical heating has significant impacts on the stationary wave solution, especially in the tropics (Fig. 11a), its impacts on the midlatitude storm-track distribution is relatively minor (Figs. 11b–c).

Taken together, these results do not support the hypotheses that either strong damping upstream of the Pacific storm track (Zurita Gotor and Chang 2005; Mak and Deng 2007) or tropical heating (Lee and Kim 2003) can be considered as the main factors contributing to the relatively weak activity of the Pacific storm track. Rather, they support the conclusion reached by Hoskins and Valdes (1990): that the midlatitude storm tracks are largely maintained by extratropical diabatic heating. Nevertheless, just having two regions of high baroclinicity in the midlatitudes, with one occurring downstream of the other, does not necessarily result in a stronger downstream storm track (e.g., Franzke et al. 2001). Thus, more work needs to be done to investigate which aspects of the extratropical heating distribution are most important in shaping the observed storm-track distribution.

b. Interactions between orography and diabatic heating

As discussed above, in this model, diabatic heating is fixed. In the real atmosphere, the actual heating distribution depends on the circulation. Thus, it is possible that removal of orographic forcing may act to modify the heating distribution, which can then lead to a significant feedback. To assess the full impact of the removal of orography involves simulating how the SST may change. This requires integrating a fully coupled atmosphere–ocean GCM to equilibrium, which is beyond our current capability.

To partially assess the possible impact of the feedback in diabatic heating, as the next step up from the dry storm-track model, several preliminary experiments have been conducted using an atmosphere-only GCM with prescribed climatological SST. In this set of experiments, diabatic heating can respond to changes in the circulation, but the SST cannot change. The model used is the NCAR Community Atmosphere Model version 3 (CAM3) (see Collins et al. 2004, 2006), run at a horizontal resolution of T42, with 26 uneven hybrid sigma-pressure levels in the vertical. The control experiment is conducted with full orographic forcing and perpetual January forcing (insolation and SST), while the NO OROG experiment is conducted with all mountains removed (but with all land use type and subgrid-scale orographic forcing unchanged). Each experiment is run for 30 months, with the mean from the last 24 months used as the model climate.

The stationary wave pattern and storm-track distribution computed from the control experiment are shown in Figs. 12a,b. The model stationary wave is well correlated with that observed (Fig. 2a). The storm-track distribution appears to a bit more zonally symmetric than that observed (Fig. 2b), but, as found in the observations, the Atlantic storm track in this simulation is stronger than the Pacific storm track. In the NO OROG simulation (Figs. 12c,d), even without the existence of mountains, much of the stationary wave remains. The storm-track distribution in this experiment still shows a stronger Atlantic storm track than its Pacific counterpart, in agreement with the results discussed above (Fig. 7b). In fact, comparisons between Figs. 12b and 12d suggest that the main impact of the orographic forcing is to suppress the storm track over Asia and North America, with only very minor impacts over the ocean basins, largely consistent with the results shown above (Fig. 8e). These GCM results confirm that results based on the intermediate model used in this study should provide useful insights regarding what forces the zonal asymmetries in the observed stationary wave pattern and storm-track distribution.

In these CAM experiments, the SST is held fixed. Changes in the orography will lead to changes in the circulation, thus possibly changing the SST. Seager et al. (2002) conducted a set of experiments using CAM3 coupled to a mixed layer model to assess the impact of orography on the zonal asymmetries in the surface temperature distribution. Their results suggest that, if all mountains are removed, surface temperature will be much warmer (by up to 6°C or more) over the eastern parts of the continents. Over the midlatitude oceans and western Europe, the temperature will be significantly cooler (by up to 3°C) owing to a change in surface wind direction from southwest to more westerly. In the CAM experiments discussed above, with the removal of the mountains, the temperature over the eastern parts of the continents also show significant warming, but, with the SST prescribed, no cooling is found over the midlatitude oceans and Europe. Thus, the zonal asymmetries in surface temperature are expected to be further reduced if the SST is allowed to change. However, without conducting a set of coupled experiments, it is difficult to quantify what this may do to the zonal asymmetries in diabatic heating. In addition, changes in precipitation and surface wind stress owing to the removal of the mountains may also lead to changes in ocean circulation and heat transport (and thus SST), so coupling to a full ocean GCM instead of a mixed layer model may be required to assess the full impact of orographic forcing.

Apart from the direct impact of the orography on modifying the diabatic heating distribution, previous studies have also shown that there are strong nonlinear interactions between the responses forced by heating and orography, even if they are treated as independent forcings (HTW02). Diagnostically speaking, such nonlinearities can be regarded as the stationary wave forced by the so-called “stationary nonlinearity” (Wang and Ting 1999; see also Valdes and Hoskins 1989), which includes the effects of heat and momentum fluxes forced by the stationary wave itself, as well as the effects of heating and orography modifying the flow incident on the orography. Ringler and Cook (1999) have found that diabatic heating located locally around Tibet and the Rockies significantly modify the flow incident on the mountains and alter the stationary wave generated by the mountains. Apart from the impacts of local heating, HTW02 pointed out that much of the nonlinearity over the Pacific–North American (PNA) sector appears to be due to nonlinear interactions between Tibet and diabatic heating over the tropical Pacific. The results presented here also show strong nonlinear interactions between the responses forced by the different forcings, as can be seen from the large differences between the isolated and full nonlinear responses to each of the forcings. The similarities between the full nonlinear responses to tropical heating (Fig. 3e) and Tibet (Fig. 4e) over the PNA sector (pattern correlation between Figs. 3e and 4e is 0.84 over the longitude band of 120°E–60°W) is also consistent with HTW02’s suggestion that there are strong nonlinear interactions between the responses to Tibet and Pacific tropical heating. Nevertheless, much work remains before the physical nature of these nonlinear interactions can be fully understood.

c. Full nonlinear response to the Rockies

As discussed in section 3b, in the set of dry model experiments conducted to investigate the impact of orographic forcing, the removal of the Rockies excites a very strong response. This issue is further investigated by running CAM3 with only the Rockies removed. The results (not shown) show much weaker responses than those shown in Figs. 4f and 8n, more in line with the results shown in Figs. 5c and 9. These results are investigated further by conducting a set of experiments using the idealized dry model in which the temperature distribution is iterated toward the climate of the control CAM experiment instead of that observed, starting from the diabatic heating distribution taken from the control CAM experiment. This dry experiment (referred to as the DRYCAM experiment) is again very successful in reproducing the stationary wave pattern taken from the control CAM experiment (Fig. 12a), with a pattern correlation of 0.94, and explains 89% of its variance. The full nonlinear response to the Rockies is then assessed by conducting an experiment with the full heating, but with the Rockies removed. Consistent with the CAM experiments, the results from this experiment show much weaker responses than those shown in Figs. 4f and 8n. These results suggest that the highly nonlinear response to the removal of the Rockies is sensitive to the details of the heating distribution and may not be robust. Further experiments conducted by merging the heating distributions taken from the control dry experiment and the DRYCAM experiment over various regions suggest that the strong nonlinear response is sensitive not only to extratropical heating close to the Rockies, but also to the distribution of tropical heating. More work will need to be conducted to illuminate the nature of this nonlinearity.

In light of the sensitivity discussed in the preceding paragraph, a full set of experiments similar to those discussed in sections 3 and 4 above has been conducted based on the control DRYCAM experiment. The results show that apart from the full nonlinear response to the Rockies discussed in this subsection, all other results are robust.5

6. Summary and conclusions

In this study, a dry global circulation model is used to examine the contributions made by orographic and diabatic forcings in shaping the zonal asymmetries in the earth’s NH winter climate. By design, the model mean flow is forced to bear a close resemblance to the observed zonal mean and stationary waves. The model also provides a decent simulation of the storm tracks (see also C06). In particular, the maxima over the Pacific and Atlantic and minima over Asia and North America are fairly well simulated. The model also successfully simulates the observation that the Atlantic storm track is stronger than the Pacific storm track, despite stronger baroclinicity over the Pacific.

Sensitivity experiments are performed by imposing and removing various parts of the total forcings. The main effects of the following four forcings have been examined: orographic forcings due to Tibet and the Rockies, and diabatic forcings due to zonal asymmetries in the tropical and NH extratropical diabatic heating.

In terms of the NH winter stationary waves, results of this study are largely consistent with those of HTW02. Diabatic forcings explain most of the modeled stationary waves, with orographic forcings playing only a secondary role, and feedbacks due to eddy fluxes probably play only minor roles in most cases. Nevertheless, results of this study suggest that eddy fluxes may be important in modifying the response to orographic forcings in the absence of zonal asymmetries in diabatic heating. However, such a model configuration is rather artificial and may not be important in reality. On the other hand, unlike the conclusion reached by HTW02, it is argued that convergence of eddy momentum fluxes is important in forcing the oceanic lows in the lower troposphere, in agreement with our synoptic intuition.

Regarding the NH winter storm-track distribution, results of this study suggest that NH extratropical heating is the most important forcing, consistent with the suggestion by Hoskins and Valdes (1990). Zonal asymmetries in NH extratropical heating act to force the Pacific storm track to shift equatorward and the Atlantic storm track to shift poleward, attain a southwest–northeast tilt, and intensify. It appears to be the most important forcing responsible for explaining why the Atlantic storm track is stronger than the Pacific storm track. Tibet and the Rockies are also important, mainly in suppressing the storm tracks over the continents, forcing a clearer separation between the Pacific and Atlantic storm tracks. In contrast, asymmetries in tropical heating appear to play only a minor role in forcing the model storm-track distribution.

While this study has provided new insights on what forces the zonal asymmetries in the earth’s climate, many questions still remain unanswered. The NH extratropical heating has been identified to be the main forcing responsible for forcing the Atlantic storm track to be stronger than the Pacific storm track. However, the question of why the Atlantic storm track is stronger despite weaker baroclinicity is still not answered. While orographic forcing is found to suppress the storm-track activity over the continents, the detailed dynamics of how storm tracks interact with orography is still unclear. There are certainly studies that investigate the interactions between individual baroclinic waves with mountains (e.g., Orlanski and Gross 1994; Davies 1997), but it is not entirely clear whether one can extrapolate those results to a statistical equilibrium distribution of baroclinic waves, especially in view of the possibility that eddy feedback owing to these waves may be important in modifying the large-scale flow. In addition, as pointed out by HTW02, there are strong nonlinearities between the atmospheric response to heating and orography, and these nonlinearities are not yet well understood. Moreover, as discussed above, diabatic and orographic forcings are not strictly independent, and to understand the full impact of orography one must also investigate its impacts on the SST and diabatic heating. To answer these and other questions will require further studies ranging from idealized studies using even simpler model configurations, through intermediate model studies like the one conducted here, to studies employing carefully designed full GCM and coupled GCM simulations.

Acknowledgments

The author would like to thank Dr. W. Lin for making the CAM runs, Ms. Y. Guo for helping to set up the stationary wave model experiments, and two anonymous reviewers for suggestions that help clarify the discussions. This study is supported by NOAA Grant NA06OAR4310084.

REFERENCES

  • Blackmon, M. L., 1976: A climatological spectral study of the 500 mb geopotential height of the Northern Hemisphere. J. Atmos. Sci., 33 , 16071623.

    • Search Google Scholar
    • Export Citation
  • Branstator, G., 1992: The maintenance of low-frequency atmospheric anomalies. J. Atmos. Sci., 49 , 19241945.

  • Branstator, G., 1995: Organization of storm track anomalies by recurring low-frequency circulation anomalies. J. Atmos. Sci., 52 , 207226.

    • Search Google Scholar
    • Export Citation
  • Broccoli, A. J., , and S. Manabe, 1992: The effects of orography on midlatitude Northern Hemisphere dry climates. J. Climate, 5 , 11811201.

    • Search Google Scholar
    • Export Citation
  • Cai, M., , and M. Mak, 1990: On the basic dynamics of regional cyclogenesis. J. Atmos. Sci., 47 , 14171442.

  • Chang, E. K. M., 2006: An idealized nonlinear model of the Northern Hemisphere winter storm tracks. J. Atmos. Sci., 63 , 18181839.

  • Chang, E. K. M., , and Y. Guo, 2007: Dynamics of the stationary anomalies associated with the interannual variability of the midwinter Pacific storm track—The roles of tropical heating and remote eddy forcing. J. Atmos. Sci., 64 , 24422461.

    • Search Google Scholar
    • Export Citation
  • Chang, E. K. M., , and P. Zurita-Gotor, 2007: Simulating the seasonal cycle of the Northern Hemisphere storm tracks using idealized nonlinear storm-track models. J. Atmos. Sci., 64 , 23092331.

    • Search Google Scholar
    • Export Citation
  • Chang, E. K. M., , S. Lee, , and K. L. Swanson, 2002: Storm track dynamics. J. Climate, 15 , 21632183.

  • Charney, J. G., , and A. Eliassen, 1949: A numerical method for predicting the perturbations of the middle latitude westerlies. Tellus, 1 , 3854.

    • Search Google Scholar
    • Export Citation
  • Collins, W. D., and Coauthors, 2004: Description of the NCAR Community Atmosphere Model (CAM 3.0). National Center for Atmospheric Research Tech. Rep. NCAR/TN-464+STR, 210 pp.

    • Search Google Scholar
    • Export Citation
  • Collins, W. D., and Coauthors, 2006: The formulation and atmospheric simulation of the Community Atmospheric Model version 3 (CAM3). J. Climate, 19 , 21442161.

    • Search Google Scholar
    • Export Citation
  • Davies, C. A., 1997: The modification of baroclinic waves by the Rocky mountains. J. Atmos. Sci., 54 , 848868.

  • Franzke, C., , K. Fraedrich, , and F. Lunkeit, 2001: Teleconnections and low-frequency variability in idealized experiments with two storm tracks. Quart. J. Roy. Meteor. Soc., 127 , 13211339.

    • Search Google Scholar
    • Export Citation
  • Frierson, D. M. W., , I. M. Held, , and P. Zurita-Gotor, 2006: A gray-radiation aquaplanet GCM. Part I: Static stability and eddy scale. J. Atmos. Sci., 63 , 25482566.

    • Search Google Scholar
    • Export Citation
  • Hayashi, Y., , and D. Golder, 1981: The effects of condensational heating on midlatitude transient waves in their mature stage: Control experiments with a GFDL GCM. J. Atmos. Sci., 38 , 25322539.

    • Search Google Scholar
    • Export Citation
  • Held, I. M., , and M. J. Suarez, 1994: A proposal for the intercomparison of the dynamical cores of atmospheric general circulation models. Bull. Amer. Meteor. Soc., 75 , 18251830.

    • Search Google Scholar
    • Export Citation
  • Held, I. M., , M. Ting, , and H. Wang, 2002: Northern winter stationary waves: Theory and modeling. J. Climate, 15 , 21252144.

  • Hoskins, B. J., , and P. J. Valdes, 1990: On the existence of storm tracks. J. Atmos. Sci., 47 , 18541864.

  • Lee, S., , and H-K. Kim, 2003: The dynamical relationship between subtropical and eddy-driven jets. J. Atmos. Sci., 60 , 14901503.

  • Lee, W-J., , and M. Mak, 1996: The role of orography in the dynamics of storm tracks. J. Atmos. Sci., 53 , 17371750.

  • Mak, M., , and Y. Deng, 2007: Diagnostic and dynamical analyses of two outstanding aspects of storm tracks. Dyn. Atmos. Oceans, 43 , 8099.

    • Search Google Scholar
    • Export Citation
  • Orlanski, I., 1998: Poleward deflection of storm tracks. J. Atmos. Sci., 55 , 25772602.

  • Orlanski, I., , and B. D. Gross, 1994: Orographic modification of cyclone development. J. Atmos. Sci., 51 , 589612.

  • Ringler, T. D., , and K. H. Cook, 1999: Understanding the seasonality of orographically forced stationary waves: Interaction between mechanical and thermal forcing. J. Atmos. Sci., 56 , 11541174.

    • Search Google Scholar
    • Export Citation
  • Seager, R., , D. S. Battisti, , J. Yin, , N. Gordon, , N. Naik, , A. C. Clement, , and M. A. Cane, 2002: Is the Gulf Stream responsible for Europe’s mild winters? Quart. J. Roy. Meteor. Soc., 128 , 25632586.

    • Search Google Scholar
    • Export Citation
  • Simmons, A. J., , and B. J. Hoskins, 1978: The life cycles of some nonlinear baroclinic waves. J. Atmos. Sci., 35 , 14541477.

  • Smagorinsky, J., 1953: The dynamical influence of large-scale heat sources and sinks on the quasi-stationary mean motions of the atmosphere. Quart. J. Roy. Meteor. Soc., 79 , 342366.

    • Search Google Scholar
    • Export Citation
  • Valdes, P. J., , and B. J. Hoskins, 1989: Linear stationary wave simulations of the time-mean climatological flow. J. Atmos. Sci., 46 , 25092527.

    • Search Google Scholar
    • Export Citation
  • von Storch, H., 1995: Spatial patterns: EOFs and CCA. Analysis of Climate Variability, H. von Storch and A. Navarra, Eds., Springer, 227–258.

    • Search Google Scholar
    • Export Citation
  • Wallace, J. M., , G. H. Lim, , and M. L. Blackmon, 1988: Relationship between cyclone tracks, anticyclone tracks, and baroclinic waveguides. J. Atmos. Sci., 45 , 439462.

    • Search Google Scholar
    • Export Citation
  • Wang, H., , and M. Ting, 1999: Seasonal cycle of the climatological stationary waves in the NCEP–NCAR reanalysis. J. Atmos. Sci., 56 , 38923919.

    • Search Google Scholar
    • Export Citation
  • Zurita-Gotor, P., , and E. K. M. Chang, 2005: The impact of zonal propagation and seeding on the eddy-mean flow equilibrium of a zonally varying two-layer model. J. Atmos. Sci., 62 , 22612273.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

(a) Zonal mean distribution for Q for the control experiment, contour interval 0.2 K day−1. (b) Average of Q from 900 to 100 hPa, contour interval 0.5 K day−1. (c) Orography used in control experiment, contour interval 500 m. Shaded regions denote positive values in (a) and (b), and larger than 1000 m in (c).

Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2403.1

Fig. 2.
Fig. 2.

(a) Zonal asymmetrical part of the time mean 300-hPa streamfunction (contour interval 3 × 106 m2 s−1); (b) standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m) computed from January 1982 to 1994 based on the NCEP–NCAR reanalysis. Shaded regions denote negative values in (a), and larger than 100 m in (b). (c), (d) As in (a), (b) but from 50 months of control model simulation. R, A, and V correspond to pattern correlation between the observed and modeled patterns, normalized amplitude of the modeled pattern, and percentage of observed variance explained by the modeled pattern. See (4) for definitions.

Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2403.1

Fig. 3.
Fig. 3.

Zonal asymmetrical part of the time mean 300-hPa streamfunction (contour interval 3 × 106 m2 s−1) taken from (a) the NO OROG–FULL HEAT, (b) the NO OROG–TROP, and (c) the NO OROG–NH EXTRA experiment. Difference in the zonal asymmetrical part of the time mean 300-hPa streamfunction (contour interval 3 × 106 m2 s−1) between the control experiment and (d) the FULL OROG–ZONAL HEAT, (e) the FULL OROG–NH EXTRA, and (f) the FULL OROG–TROP experiment. Shaded regions denote negative values; r, a, and v correspond to pattern correlation between the control and sensitivity patterns, normalized amplitude of the sensitivity pattern, and percentage of control variance explained by the sensitivity pattern, respectively. See (5) for definitions.

Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2403.1

Fig. 4.
Fig. 4.

Zonal asymmetrical part of the time mean 300-hPa streamfunction (contour interval 3 × 106 m2 s−1), taken from (a) the FULL OROG-ZONAL HEAT, (b) the ONLY TIBET–ZONAL HEAT, and (c) the ONLY ROCKIES–ZONAL HEAT experiments. Difference in the zonal asymmetrical part of the time mean 300-hPa streamfunction (contour interval 3 × 106 m2 s−1) between the control experiment and (d) the NO OROG–FULL HEAT, (e) the NO TIBET–FULL HEAT, and (f) the NO ROCKIES–FULL HEAT experiment. Shaded regions denote negative values.

Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2403.1

Fig. 5.
Fig. 5.

Zonal asymmetrical part of the time mean 300-hPa streamfunction (contour interval 3 × 106 m2 s−1) taken from (a) the stationary wave model experiment forced with orographic forcings only and (b) the experiment forced with eddy vorticity fluxes taken from the FULL OROG–ZONAL HEAT experiment. (c) Difference in the zonal asymmetrical part of the time mean 300-hPa streamfunction (contour interval 3 × 106 m2 s−1) between the control experiment and an experiment similar to the NO ROCKIES–FULL HEAT experiment except that the zonal mean is damped toward the observed zonal mean. Shaded regions denote negative values.

Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2403.1

Fig. 6.
Fig. 6.

Zonal asymmetrical part of the time mean 700-hPa streamfunction (contour interval 3 × 106 m2 s−1) taken from (a) the NCEP–NCAR reanalysis from January 1982 to 1994: (b) the NO OROG–NH EXTRA, (c) the NO OROG–FULL HEAT, and (d) the control experiment; (e) the stationary wave model experiment forced with orographic and diabatic forcings; (f) similar to (e) but the zonal mean is not damped toward the observed zonal mean. Shaded regions denote negative values.

Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2403.1

Fig. 7.
Fig. 7.

Standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, values larger than 100 shaded) computed from (a) the control, (b) the NO OROG–FULL HEAT, (c) the NO OROG–TROP, (d) the NO OROG–NH EXTRA, (e) the FULL OROG–ZONAL HEAT, (f) the ONLY TIBET-ZONAL HEAT, and (g) the ONLY ROCKIES–ZONAL HEAT experiment. (h)–(n) Difference in standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, shaded values denote statistically significant at the 95% level based on a standard t test) between each of the rhs panels and the NO OROG–ZONAL HEAT experiment.

Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2403.1

Fig. 8.
Fig. 8.

Standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, values larger than 100 shaded) computed from (a) the NO OROG–ZONAL HEAT, (b) the FULL OROG–ZONAL HEAT, (c) the FULL OROG–NH EXTRA, (d) the FULL OROG–TROP, (e) the NO OROG–FULL HEAT, (f) the NO TIBET–FULL HEAT, and (g) the NO ROCKIES–FULL HEAT experiment. (h)–(n) Difference in standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, shaded values denote statistically significant at the 95% level based on a standard t test) between the control experiment and each of the rhs panels.

Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2403.1

Fig. 9.
Fig. 9.

(a) Standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, values larger than 100 shaded) computed from an experiment similar to the NO ROCKIES–FULL HEAT experiment except that the zonal mean is damped toward the observed zonal mean. (b) Difference in standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, shaded values denote statistically significant at the 95% level based on a standard t test) between the control experiment and the panel on the left.

Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2403.1

Fig. 10.
Fig. 10.

(a) Zonally asymmetrical part of the time mean 300-hPa streamfunction (contour interval 3 × 106 m2 s−1), taken from the ONLY TIBET–NH EXTRA experiment. (b) Standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, values larger than 100 shaded) taken from the ONLY TIBET–NH EXTRA experiment. (c) Difference in standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, shaded values denote statistically significant at the 95% level based on a standard t test) between the ONLY TIBET–NH EXTRA experiment and the NO OROG–ZONAL HEAT experiment. (d) Difference in the zonal asymmetrical part of the time mean 300-hPa streamfunction (contour interval 3 × 106 m2 s−1) between the control experiment and the NO TIBET–TROP experiment. (e) Standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, values larger than 100 shaded), taken from the NO TIBET–TROP experiment. (f) Difference in standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, shaded values denote statistically significant at the 95% level based on a standard t test) between the control experiment and the NO TIBET–TROP experiment.

Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2403.1

Fig. 11.
Fig. 11.

(a)–(c) As in Figs. 10 a–c but for the FULL OROG–NH EXTRA experiment. As in Figs. 10 d–f but for the NO OROG–TROP experiment.

Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2403.1

Fig. 12.
Fig. 12.

(a) Zonal asymmetrical part of the time mean 300-hPa streamfunction (contour interval 3 × 106 m2 s−1) taken from the control CAM experiment. (b) Standard deviation of 24-h filtered 500-hPa geopotential height (contour interval 20 m, values larger than 100 shaded), taken from the control CAM experiment. (c)–(d) As in (a)–(b) but taken from the CAM experiment without mountains.

Citation: Journal of Climate 22, 3; 10.1175/2008JCLI2403.1

1

A full set of experiments has also been conducted using a value of 1.25 K km−1 for S, and the results are qualitatively similar to those presented here.

2

Note that in (4), V takes into account not only the correlation between the two patterns, but also the amplitude difference between the two patterns. Hence, in general, V is not equal to R2.

3

Note that all experiments also include the effects of land–sea contrast in surface friction. However, when forced with this alone (i.e., the NO OROG–ZONAL HEAT experiment), no nonzero contours show up on the stationary wave plot.

4

In the rhs panels of Figs. 7 and 8, the values r, a, and v are computed by comparing each panel to Fig. 7h.

5

As mentioned earlier, results from other sensitivity experiments using Q derived from various first guesses Q0 are also consistent with the results presented in this study.

Save