Influence of Orography on the Extratropical Response to El Niño Events

Sumant Nigam Department of Meteorology, University of Maryland, College Park, Maryland

Search for other papers by Sumant Nigam in
Current site
Google Scholar
PubMed
Close
and
Eric DeWeaver Department of Meteorology, University of Maryland, College Park, Maryland

Search for other papers by Eric DeWeaver in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

The contribution of the interaction between tropically forced circulation anomalies and the extratropicalmountains in the generation of extratropical circulation anomalies during the 1987/88 and 1988/89 winter seasons is diagnosed using a divergent barotropic model that solves for both the zonal-mean and eddy components of the 200-mb rotational anomalies. Barotropic modeling shows that the orographic modulation of the rotational response to the 200-mb tropical divergence anomaly can be substantial over the Pacific–North American region.

  • The modulation consists of a large-scale wave pattern with a ridge in the central subtropical Pacific, a trough over the Gulf of Alaska, and a weak ridge extending across North America from Baja California to Greenland. These features have an amplitude of ∼40 gpm, and the orographic modulation is thus about one-third as strong as the primary wave pattern.

  • The associated 200-mb zonal wind is strongest (∼5 m s−1) in the vicinity of the eastern end of the East Asian jet, thus contributing to the southeastward jet extension during El Niño winters.

  • The Himalayan–Tibetan complex is the major locus of orographic interaction in the model, for it alone accounts for all the features and over two-thirds of the amplitude modulation.

  • The eddy and zonal-mean parts of the tropically forced flow anomalies make comparable contributions to orographic modulation. However, the midlatitude eddy anomalies themselves result, in part, from the interaction of the zonal-mean zonal wind anomaly and the climatological vorticity gradients, that is, from “zonal–eddy”interaction. The strength of this interaction depends on the arbitrarily specified distribution of the compensating zonal-mean subsidence in the model.

These findings indicate the potential importance of secondary orographic interaction in the generation of extratropical circulation anomalies in response to tropical heating anomalies. Experiments with more complete dynamical models that predict both the rotational and divergent components of the flow in response to tropical heating anomalies are clearly warranted.

Corresponding author address: Dr. Sumant Nigam, Department of Meteorology, University of Maryland, College Park, MD 20742-2425.

Email: nigam@atmos.umd.edu

Abstract

The contribution of the interaction between tropically forced circulation anomalies and the extratropicalmountains in the generation of extratropical circulation anomalies during the 1987/88 and 1988/89 winter seasons is diagnosed using a divergent barotropic model that solves for both the zonal-mean and eddy components of the 200-mb rotational anomalies. Barotropic modeling shows that the orographic modulation of the rotational response to the 200-mb tropical divergence anomaly can be substantial over the Pacific–North American region.

  • The modulation consists of a large-scale wave pattern with a ridge in the central subtropical Pacific, a trough over the Gulf of Alaska, and a weak ridge extending across North America from Baja California to Greenland. These features have an amplitude of ∼40 gpm, and the orographic modulation is thus about one-third as strong as the primary wave pattern.

  • The associated 200-mb zonal wind is strongest (∼5 m s−1) in the vicinity of the eastern end of the East Asian jet, thus contributing to the southeastward jet extension during El Niño winters.

  • The Himalayan–Tibetan complex is the major locus of orographic interaction in the model, for it alone accounts for all the features and over two-thirds of the amplitude modulation.

  • The eddy and zonal-mean parts of the tropically forced flow anomalies make comparable contributions to orographic modulation. However, the midlatitude eddy anomalies themselves result, in part, from the interaction of the zonal-mean zonal wind anomaly and the climatological vorticity gradients, that is, from “zonal–eddy”interaction. The strength of this interaction depends on the arbitrarily specified distribution of the compensating zonal-mean subsidence in the model.

These findings indicate the potential importance of secondary orographic interaction in the generation of extratropical circulation anomalies in response to tropical heating anomalies. Experiments with more complete dynamical models that predict both the rotational and divergent components of the flow in response to tropical heating anomalies are clearly warranted.

Corresponding author address: Dr. Sumant Nigam, Department of Meteorology, University of Maryland, College Park, MD 20742-2425.

Email: nigam@atmos.umd.edu

1. Introduction

In observations (Bjerknes 1969; Horel and Wallace 1981; Ropelewski and Halpert 1987) and simulations (e.g., Lau 1985), El Niño-related tropical heating anomalies produce global circulation and precipitation anomalies. However, the stationary waves forced directly by the equatorial Pacific heating anomalies cannot by themselves explain the associated extratropical climate impact (Geisler et al. 1985; Held and Kang 1987): to generate the teleconnected climate anomalies, the directly forced perturbation must undergo secondary interactions with other components of the climate system, such as the East Asian jet, the Pacific stormtracks, the Himalayan–Tibetan and Rocky Mountains, and the underlying extratropical oceans.

The secondary interaction with the Pacific stormtracks and attendant transient vorticity fluxes are increasingly recognized as being important in understanding the generation of seasonal climate anomalies over the Pacific–North American region (Kok and Opsteegh 1985; Held et al. 1989; Lau and Nath 1991; Hoerling and Ting 1994; Branstator 1995), but the cause of Pacific stormtrack displacements, notably during El Niño/La Niña years, remains unclear.

The secondary interactions of the tropically forced perturbations with the underlying extratropical oceans and mountains have, in contrast, received limited attention (Wallace et al. 1990; Alexander 1992; DeWeaver and Nigam 1995; Lau and Nath 1996) as it has been difficult to identify the contributions of these interactions in observations. However, secondary interactions can be significant on the seasonal timescale, particularly if the near-surface part of the tropically forced flow anomalies has significant extratropical amplitude and good correlation with the flow anomalies in the upper troposphere—the level at which tropical–extratropical interaction is best established.

Theoretical analysis (Held 1983) and model calculations (Hoskins and Karoly 1981; Lau 1985) indicate that while the local response of tropical deep heating has baroclinic vertical structure, the far-field response has equivalent barotropic structure (i.e., little, if any, vertical phase variation). Analysis of the vertical structure of the observed El Niño-related northern winter circulation anomalies corroborates these findings (Hsu and Wallace 1985; Dai and Nigam 1998, manuscript submitted to J. Climate). Although the equivalent barotropic structure of the tropically forced extratropical anomalies is not, in itself, a prerequisite, the vertical coherence implied by this structure can lead to significant secondary interactions at the extratropical lower boundary.

In this study, we examine the secondary interactions with extratropical orography. The role of mountains in forcing the prominent zonally varying features of the northern winter tropospheric circulation has been recognized since the pioneering study of Charney and Eliassen (1949). The dynamical influence of orography on the climatological seasonal circulation has been substantiated with a hierarchy of models (e.g., Manabe and Terpstra 1974; Held 1983; Nigam et al. 1988; Chen and Trenberth 1988;1 Valdes and Hoskins 1989), but the role of orography in the generation of seasonal circulation anomalies remains to be assessed.

The objective of this study it to evaluate the contribution of the secondary interaction between the tropically forced perturbations and the underlying extratropical mountains in the generation of the El Niño-related wintertime circulation anomalies in the Pacific–North American region using the divergent barotropic model. Although some of the modeling results have been reported earlier in a letter to Nature (DeWeaver and Nigam 1995), the present study provides a more comprehensive analysis of these results, including several new calculations.

This paper is divided into seven sections. The ECMWF analyses used in generating the zonally varying basic state, and the forcing and target anomalies for diagnostic barotropic modeling are described in section 2a, while the outgoing longwave radiation (OLR) and the optimally interpolated sea surface temperature (OISST) data used in identifying El Niño-related anomalies are briefly described in section 2b. The divergent barotropic anomaly model is described in section 3. Section 4 describes the anomalies between the winters (December–February, or DJF) of 1987–88 and 1988–89, which are dynamically diagnosed in this study—winters that feature opposite phases of El Niño and the zonal index cycle. The barotropic simulation of the 200-mb rotational flow anomaly, along with an off-line assessment of the secondary orographic interaction is also discussed in section 4, whereas a fully consistent model of this interaction is presented in section 5. Experiments conducted to examine the sensitivity of the secondary interaction to the longitude of the equatorial Pacific divergence anomaly are described in section 6. Discussion and concluding remarks follow in section 7.

2. Datasets

a. ECMWF analyses

The twice-daily uninitialized ECMWF analyses archived at the National Center for Atmospheric Research (NCAR) on a 2½° × 2½° latitude–longitude global grid and at 14 pressure levels from January 1985 are used to generate the seasonal rotational (target) and divergent (forcing) circulation anomalies and the basic state about which the barotropic model is linearized. The additional forcing of rotational flow anomalies from the subseasonal transient vorticity flux convergence was also diagnosed from these twice-daily analyses. The orography used in the model calculations was derived from the surface geopotential obtained from the same dataset.

b. OLR and OISST datasets

The 2½° × 2½° OLR and the 1° × 1° OISST monthly datasets were also obtained from the NCAR archives, and these fields were interpolated to a 4° × 5° latitude–longtitude grid for plotting.

3. Barotropic anomaly model

The upper-tropospheric (200 mb) rotational flow anomalies (ψa) during the two winter seasons are diagnosed from that level’s horizontal divergence (χa) and transient vorticity-flux convergence anomalies using a steady-state divergent barotropic anomaly model (e.g., Held and Kang 1987):
i1520-0442-11-4-716-e1
Here, the climatological quantity ( )c is the average of that quantity for seasons 1 and 2, while the anomaly ( )ais the difference. Here ζ is the relative vorticity, χ is the velocity potential, and Vχ and Vψ are the divergent and rotational parts of the wind field. The first two forcing terms on the rhs represent anomalous vorticity generation by vortex-tube stretching and divergent advection of climatological vorticity, while the extreme rhs term represents the forcing by transient vorticity-flux convergence, with the overbar denoting the seasonal average and the primes representing the departure of the quantity from this average. The equation is solved for ψa for specified values of the Rayleigh dissipation and biharmonic diffusion coefficients (r and ν, respectively, in the last term on the lhs).

As the circulation anomalies between contrasting winters are modeled using a zonally varying basic state that is the average of the two periods, the resulting equation is automatically linear in the anomaly field without neglecting any quadratic anomaly terms (Schneider 1988). For example, the anomalous flux of seasonally averaged vorticity is V1ζ1V2ζ2, where ( )1is the average for season 1; this ostensibly nonlinear product can be written as Vcζa + Vaζc, where ( )a = ( )1− ( )2, and ( )c = [( )1 + ( )2]/2, which is linear in the anomaly fields. Thus, although the steady model is posed as a linear matrix equation, the anomaly solution is nonlinearly accurate.

Another unique aspect of the calculations is that boththe zonal mean and the zonally asymmetric (eddy) flow anomalies are determined using the above model. This strategy was employed because of persuasive evidence for the “zonal–eddy” relationships in the upper-tropospheric circulation variability during the northern winter season: Branstator (1984) analyzed the wintertime variability of the monthly 300-mb zonal-mean zonal wind (U300) during 1962–77 and showed that well-defined quasi-stationary circulation anomalies were associated with the leading EOFs of U300. Kang and Lau (1986) presented evidence for similar linkage between the zonal mean and eddy parts of the monthly 300-mb circulation anomalies produced in GCM integrations conducted both with and without the observed SSTs in the tropical Pacific. Nigam and Lindzen (1989) illustrated the existence of zonal–eddy relationships in the context of primitive-equation model dynamics, and recently, Hoerling et al. (1995) and Ting et al. (1996)showed that the interaction of midlatitude U anomalies with the climatological stationary waves account for a significant fraction of the observed interannual variability of the extratropical winter climate.

In view of the observational and modeling evidence for the dynamical relationship between the zonal mean and eddy parts of the extratropical circulation anomalies, it is, perhaps, necessary to model both parts of the anomaly field together. This modeling strategy was, however, adopted in the present study with some reluctance as the diagnosis of U from the zonal-mean zonal momentum equation is a potentially ill-posed problem, particularly when the parameterized zonal-momentum dissipation is very weak or when dissipation is specified as an external forcing (e.g., DeWeaver and Nigam 1997). The zonal-mean zonal momentum equation corresponding to (1) is
i1520-0442-11-4-716-e1a
where Feddy represents the forcing by stationary eddies [included in the operator in (1)] and transients. The diagnosis of Ua is potentially problematic for it is contingent on the parameterization of zonal-momentum dissipation, which is, perhaps, an even more intractable problem. [For further discussion of these issues, see Lorenz (1967).]

The horizontal divergence at 200 mb—an important forcing function in the barotropic anomaly model—was obtained both from the ECMWF analyzed winds and from residual diagnosis using the χ-algorithm for each period (Sardeshmukh 1993); this algorithm uses a variational technique to produce a global divergence field that is consistent with the observed seasonal rotational flow and subseasonal vorticity transients, given the model dissipation parameters. The global mean of the divergence anomaly is ensured to be zero in both anomaly simulation and in the regional forcing experiments. Solutions of the barotropic anomaly model presented here are obtained from a spectral version of the model truncated rhomboidally at wavenumber 15 and with Rayleigh dissipation r = (5 days)−1 and biharmonic diffusion ν = 1016 m4 s−1.

4. Barotropic simulation of the 1987/88 and 1988/89 winter circulation anomalies

The contrasting winter circulations during 1987/88 and 1988/89 were associated with anomalous conditions in the Pacific basin related to the extreme phases of ENSO, as well as with opposite phases of the midlatitude zonal index cycle (see Fig. 4 in Ting et al. 1996); Fig. 1 displays the OLR and OISST anomalies between these winters. The warming of the central–eastern tropical Pacific is accompanied by cooling of the subtropical Pacific basin and an eastward shift and intensification of tropical deep convection at the dateline (180°) in accord with the characteristic structure of El Niño-related SST and OLR anomalies (e.g., Nigam and Shen 1993). The OLR anomalies, displayed between ±30° of the equator, indicate that enhanced deep convection at the dateline is accompanied not only by diminished convection in the western Pacific (a widely recognized feature), but also by reduced convection in the immediate subtropics.2 The attenuation of convection to the north and south, as well as to the west of the equatorial dateline in Fig. 1a, suggests that the canonical depiction of the Pacific basin El Niño heating anomalies by a longitudinal dipole is, perhaps, not quite representative.

The corresponding observed (ECMWF-analyzed) and χ-derived divergence anomalies at 200 mb shown in Figs. 2a and 2b, both of which contain a prominent divergence center in the equatorial dateline region, which is stronger in the observed anomalies. The location and orientation of this divergence feature, including its southeastward extension until 120°W, is rather consistent with the structure of the negative OLR anomaly (Fig. 1a), lending credence to both the observed and the χ-derived tropical divergence anomalies. It is noteworthy that both the observed and, in particular, the χ-derived fields contain notable subtropical convergence anomalies around the divergence center; the convergence over the subtropical central Pacific to the southeast and northwest of the Hawaiian islands is greatly intensified in the χ-derived field.

The structure of the central Pacific divergence and neighboring convergence anomalies in Figs. 2a and 2bsupport our earlier contention regarding the unsuitability of the canonical description of El Niño-related anomalies; the El Niño-related 200-mb divergence anomalies are perhaps better described as a monopole divergence center with subsidence all around it.

The observed and simulated (from χ-divergence and vorticity transients) 200-mb streamfunction anomalies are shown in the lower panels of Fig. 2. A pair of anticyclones straddling the equator with centers located ∼45° eastward of the equatorial divergence center—a tropical feature characteristic of El Niño winters—is present in both observed and simulated anomalies. Extratropical anomalies, including the troughs centered over the Aleutians and the Gulf of Mexico, are also well simulated. Note that the northern subtropical anticyclone and the northern midlatitude response in the Western Hemisphere are marginally weaker in the simulation.

The observed and simulated U anomalies are shown in the right panel of Fig. 2. Despite our earlier expressed concerns about the well posedness of U diagnosis, the chosen values of Rayleigh dissipation (r) and biharmonic diffusion (ν)—the same as in Held and Kang (1987), for example—lead to a remarkable simulation of Ua.3 The U anomalies contain both strengthened subtropical westerlies, which are typically seen during El Niño winters (Nigam 1990; Hoerling et al. 1995), and midlatitude anomalies that increase the subtropical jet but decrease the westerlies poleward of 45°N; these midlatitude U anomalies represent the positive phase of the zonal index [≡U35°NU55°N (e.g., Ting et al. 1996)].

Having simulated the global rotational flow anomaly, we use the diagnostic model to determine how anomalous forcing from a particular geographic region contributes to the generation of global anomalies. The geographic region of greatest interest is, of course, the central–eastern tropical Pacific (the rectangular box in Fig. 2b) where the El Niño-related OLR (heating) anomalies are the strongest. Figures 3a and 3b show the model response when forced exclusively by the divergence in the rectangular box in Fig. 2b. The response shown here is quite similar to the model response forced by the entire tropical strip (not shown), but quite different from the total response shown in Fig. 2d; the striking difference in the Tropics is, of course, the much weaker streamfunction anticyclones, which are now centered on the dateline instead of at ∼140°W. The U forced by the divergence in the tropical box (Fig. 3b)—in excess of 4 m s−1 in the northern subtropics and midlatitudes—on the other hand, resembles the observed U anomalies between the equator and 45°N (right panel of Fig. 2). The strong Ua response in the latitudes of the climatological East Asian and North American jets has important consequences for the coupling of the zonal mean and eddy anomalies, as discussed in section 6.

a. Off-line assessment of secondary orographic interaction

An interesting feature of Figs. 3a and 3b is that both the zonal mean and eddy winds are large in mountainous regions. Furthermore, as the largest mountains are in the extratropics, and thus in the far field of the El Niño-related divergence, the tropically forced flow perturbations are expected to be equivalent barotropic over these mountain ranges based on theoretical and modeling evidence discussed in section 1. In that case, the divergence due to orographic forcing (D) can be estimated in an off-line calculation from the interaction of the scaled-down rotational flow and mountains: D = αVψh/(H0h). Here, H0 (=8 km) is the atmospheric scale height, h is the mountain height, and Vψ is the tropically forced total (zonal mean + eddy) rotational wind anomaly scaled down by a factor α = 0.3 (e.g., Held 1983). To ensure that orographic interaction takes place in the far field of the tropical heating, we set h= 0 for θ < 25°N, although calculations performed using global orography yield similar results. The divergence induced by this extratropical orography is shown in Fig. 3c.

The barotropic model’s response to D (Fig. 3d) contains large-scale features in the Pacific–North American sector: the ridge in the central subtropical Pacific, the low over the Gulf of Alaska, and the ridge extending across North America from Baja California to Greenland. These features, with typical amplitude of ∼40 gpm, can substantially modify the extratropical wave pattern forced directly from the Tropics, as they are about one-third as strong as the features in the latter (Fig. 3a) or in the observed anomalies (Fig. 2c) for that matter. Furthermore, the zonal wind anomaly associated with this streamfunction perturbation, shown later in the context of a fully consistent calculation, is located in the central Pacific overlapping the eastern end of the East Asian jet. Thus, the zonal wind anomaly acts to extend the East Asian jet southeastward, as is typical in El Niño winters.

The role of zonal-mean and eddy wind anomalies in the generation of D is examined by displaying the model’s response to the divergence produced from the interaction of the eddy wind anomalies (Fig. 3a) and orography. This response (Fig. 3e) is about half as strong as the combined response (Fig. 3d), indicating that both the zonal-mean and eddy parts of the tropically forced flow anomalies undergo significant secondary interactions with the extratropical orography.

5. Consistent modeling and sensitivity of secondary orographic modulation

a. Consistent model

The off-line calculation of divergence generated from secondary orographic interaction (D) is a completely literal representation of our physical intuition: tropical heating excites the primary wave pattern, which impinges on the mountains, generating, in a separate calculation, a secondary response that is assumed to have no further interaction with orography. Although well motivated, this model has an ad hoc character that is both undesirable and unnecessary.

A more general model can easily be constructed by including the orographic divergence D in the linear operator rather than treating it as external forcing. The orographic interaction is represented explicitly in the divergent barotropic vorticity equation through the stretching term F = −(∇2ψ + f)αVψ·h/(H0h), which is quadratic in ψ; however, as the anomaly model is formed by taking the difference between the time-mean equations for the two seasons, the anomalous orographic stretching term becomes linear in ψa for reasons stated in section 3:
i1520-0442-11-4-716-e2
With this term in the operator, the model can be written as a single equation that includes both the primary forcing by the El Niño-related divergence anomaly (specified in the deep Tropics) and the secondary effect of orographic modulation. As in the off-line calculation, we set h = 0 for θ < 25°N.

For both the climatological and anomalous flow, (2)assumes that the orographically generated divergence can be separated from the divergence produced by other processes. While such a separation would be difficult in general, the divergence anomaly specified in our experiments is confined to the tropical Pacific, so there is no overlap between the externally imposed and orographically generated divergence anomalies. On the other hand, the climatological divergence is specified everywhere, including the extratropical mountain regions, so the term [αVψch]/(H0h) is omitted in (2), since this divergence must already be included.

The consistent impact of the secondary interaction between the tropically generated wave pattern and the extratropical orography is ascertained by calculating the difference between model solutions obtained with α = 0.3 and α = 0.0. The resulting difference map (Fig. 4a) is remarkably similar to the off-line assessment of this impact (Fig. 3d), suggesting that the wave pattern of Fig. 3d has little further interaction with orography, although the new model does allow for such interactions. The contribution of the extratropical Eastern Hemisphere orography in the generation of orographic modulation is shown in Fig. 4b; the Himalayan–Tibetan complex is evidently the major locus of orographic interaction. This result is somewhat surprising, given that the primary wave pattern of Fig. 3a appears to impinge more directly on the North American land mass. However, equally robust circulation features are generated in the region upstream of the tropical divergence box from zonal/eddy coupling, as discussed in section 6.

The 200-mb zonal wind anomalies in the consistent calculation of secondary orographic interaction are shown in Fig. 5a. Although these anomalies (∼5 m s−1) are about one-third as large as those forced directly from the tropical divergence box (Fig. 5b), they appear to be important in extending the East Asian jet southeastward, as is typical in El Niño winters.

b. Sensitivity of secondary orographic modulation

In this subsection, we examine the sensitivity of secondary orographic modulation to 1) the scaling parameter α, 2) the quality of the 200-mb tropical divergence analysis, 3) the diagnosing model’s horizontal resolution, and 4) the longitudinal position of the tropical Pacific divergence anomaly.

1) Sensitivity to scaling parameter α

The scaling factor between the 200-mb and the mountain-level (e.g., 700 mb) equivalent-barotropic rotational winds (α) usually determines the magnitude of the rhs orographic forcing (F). In that case, the linear solution to orographic forcing is simply proportional to α. However, the orographic stretching term (F) is part of the linear operator on the rhs of (2), and thus it is difficult to anticipate the model’s sensitivity to α. Figure 6ashows the orographic impact obtained when α is tripled (α = 0.9), using a streamfunction contour interval that is also three times larger. A comparison with Fig. 4areveals that increasing α changed the amplitude but not the phase of the solution over most of the domain: outside of the subtropical Pacific and the region surrounding the U.S. Great Lakes, the amplitudes in Fig. 6a are about three times larger—much as would be the case if the α-term were an rhs forcing term.

2) Sensitivity to tropical divergence analysis

The secondary orographic modulation of the primary wave pattern forced from the same tropical Pacific box, but using divergence calculated from the ECMWF analyzed winds rather than the χ-algorithm, is shown in Fig. 6b. The orographic modulation is apparently robust to the methodology-related variations in diagnosis of the 200-mb tropical divergence field, as evident from a comparison of Figs. 4a and 6b.

3) Sensitivity to barotropic model resolution

Next, we evaluate the sensitivity of the orographic modulation by solving the consistent barotropic anomaly model (2) truncated rhomboidally at wavenumber 30 (R30). Figure 6c shows the orographic modulation of the primary wave pattern forced by ECMWF analyzed divergence in the same tropical Pacific box. The orographic modulation at this resolution is quite similar to that displayed in Fig. 6b except for stronger amplitudes in the Pacific–North American sector; for example, the northeast Pacific trough is over 70 m deep in Fig. 6c. The orographic modulation is stronger, partly because the primary wave pattern at R30 resolution (not shown) is itself about 20% stronger over the Tibetan plateau.

4) Sensitivity to the longitude of tropical divergence anomaly

The sensitivity of the orographic effect to the longitude of the El Niño-related tropical heating (200-mb divergence) was assessed by calculating the orographic modulation for an idealized tropical Pacific divergence anomaly centered at 5°S and placed at three different longitudes: 135°E, 180°, and 135°W. The idealized anomaly has zonal and meridional extent of 60° and 20°, respectively; has a maximum amplitude of 6 × 10−6s−1; and is centered on the “+” sign in each panel of Fig. 7. The 200-mb divergence anomaly was idealized as a monopole rather than a dipole, for reasons stated in section 4. In these experiments, the consistent anomaly model (2) was linearized about a 9-yr long (1985/86–1993/94) winter climatology. The results shown in Fig. 7 indicate that even when the divergence anomaly shifts across the Pacific, from 135°E to 135°W, the orographic interaction generates anomalies that are qualitatively very similar.

The insensitivity to tropical divergence location is inherited from the similarity of the three primary wave patterns in the Himalayan–Tibetan region (Figs. 8a–c); these wave patterns are obtained by solving (2) with α = 0.0. A comparison of the primary wave patterns indicates that while the flow anomalies over the Pacific basin are different, circulation features over North America and the Atlantic, and over the Eastern Hemisphere continents are remarkably similar in the three cases. The zonally varying features of the climatological basic-state flow are, of course, crucial for the modeled primary wave structure displayed in Figs. 8a–c. The importance of the zonally varying basic state in the model operator is illustrated by computing the primary wave pattern forced by the Rossby wave source [the first two terms on the rhs of (2)] of Fig. 8b using a model linearized about the corresponding zonally symmetric basic state. This response (Fig. 8d) is quite different from all three of the previous cases. The dynamical basis for the insensitivity of the primary wave patterns (Figs. 8a–c), particularly in the extratropics, to the longitude of the tropical divergence anomaly, is examined in the next section.

6. Cause of insensitivity of the primary wave pattern

The relative insensitivity of the extratropical wave pattern to the longitude of anomalous divergence forcing over the tropical Pacific is an issue meriting investigation by itself. This insensitivity was first noted in the GCM modeling experiments of Geisler et al. (1985), which produced rather similar extratropical winter circulation anomalies for different longitudinal positions of the prescribed tropical Pacific SST anomaly.4 In this study, we discuss aspects of the insensitivity of the extratropical response, and possible reasons for it, in the context of orographic interaction.

a. Rossby wave source

It has been suggested that the insensitivity of extratropical anomalies to the tropical forcing longitude is due to the relative insensitivity of the Rossby wave source (Sardeshmukh and Hoskins 1988). This explanation, however, cannot account for the insensitivity of the extratropical response illustrated in Fig. 8. The Rossby wave source for each of the three positions of the idealized tropical divergence anomaly is shown in Fig. 9, and evidently the Rossby wave source follows the tropical divergence anomaly across the Pacific basin.

The relative invariance of the Rossby wave source can impart some insensitivity to the extratropical response provided the tropical divergence anomaly displacements are confined to the west of the dateline, so that interactions with the climatological vorticity gradients can occur, and provided that these displacements are modest (e.g., ≲30°), as shown by the barotropic calculations of Grimm and Silva Dias (1995). In models that additionally calculate the U anomaly, such as (1)and (2), the extratropical response is evidently insensitive to tropical divergence anomaly displacements extending even eastward of the dateline. As we show in the following sections, this insensitivity results from the interaction of the U anomaly and the climatologicalvorticity gradients.

b. Zonal–eddy interaction

The role of zonal–eddy interactions in generating the circulation anomalies forced by observed divergence in the tropical Pacific box are examined in this section. The eddy wave pattern forced by this divergence anomaly is shown in Fig. 10a; the resemblance with the pattern in Fig. 3a is not surprising given the similarities in divergence in this box in Figs. 2a and 2b.

The wave pattern produced by the zonal-mean Rossby wave source is shown in Fig. 10b. The four-cell pattern over the eastern continents, as well as the three-cell pattern in the western extratropics, are immediately recognizable—the very features of the extratropical anomalies that were relatively invariant in Figs. 8a–c! [Note that the comparison with Fig. 8 is motivated by the structural similarity of the divergence anomaly in the tropical Pacific box (Fig. 2a) and the idealized monopole-type divergence anomaly used in Fig. 8b.] The zonal mean Rossby wave source can be written as
i1520-0442-11-4-716-eq1
where the first term arises from the zonal-mean divergence anomaly, while the second results from the interaction of the eddy components. Here, “*” denotes the eddy part, while the overbar denotes the zonal average. The contribution of the second term was found to be negligible by comparing Fig. 10b to the response obtained with the eddy divergence anomaly set equal to zero (not shown). The substantial eddy amplitudes in the model’s response to essentially the zonal-mean divergence anomaly (Fig. 10b) is indicative of the significance of zonal–eddy interactions in this model.

The wave pattern forced by the eddy component of the Rossby wave source is shown in Fig. 10c, and it is notably different from the pattern depicted in Fig. 10bdue to the absence of wave amplitude in the middle and polar latitudes. The presence of subtropical and midlatitude anomalies in the vicinity of tropical forcing in Fig. 10c, on the other hand, suggests that this part of the solution should be sensitive to the tropical-forcing longitude, and inspection of Figs. 8a–c indicates variability to be indeed confined to the subtropical–midlatitude parts of the Pacific basin.

c. Experiments with an “eddy-only” model

An “eddy-only” model was developed from the matrix elements of the full model (1). This model solves only for the eddies given the zonally asymmetric components of the rhs forcing in (1) (e.g., Held and Kang 1987; Sardeshmukh and Hoskins 1988); in studies of zonal–eddy interaction, the forcing can also include the terms involving the U anomaly (e.g., Branstator 1984). The wave pattern generated by the eddy-only model, when forced only by Ua taken from the full-model solution (for the case shown in Fig. 10a), is depicted in Fig. 10d.

The zonal–eddy coupling can be envisioned to proceed as follows: first, Ua develops in response to υa, with the development governed by (1a); the initial development can occur without any contribution from the meridional stationary eddy vorticity flux —a part of Feddy in (1a). Once Ua attains significant amplitude, its interaction with the climatological vorticity gradients will generate the eddy anomalies, which in turn exercise control over Ua through the meridional eddy vorticity flux.

The above described modeling experiments (Fig. 10) suggest that zonal–eddy interactions in the full anomaly model (1), engendered by the zonal-mean divergence anomalies, impart some insensitivity to the circulation anomalies in the middle and high latitudes with respect to the longitude of tropical forcing. By this reasoning, barotropic anomaly models that assume Ua = 0.0 (e.g., Held and Kang 1987; Sardeshmukh and Hoskins 1988) or else the zonal mean Rossby wave source to be zero (e.g., in the sensitivity experiments of Grimm and Silva Dias 1995) should exhibit greater sensitivity of the extratropical anomalies. Grimm and Silva Dias (1995) do find the extratropical response to be quite sensitive (see their Fig. 9), particularly when the tropical divergence anomaly is displaced eastward of the dateline. Sardeshmukh and Hoskins (1988) display the modeled anomalies only until 40°N (see their Fig. 7), and as such the extent of insensitivity of their middle and high latitude response cannot be gauged.

d. Influence of zonal-mean divergence specification on zonal–eddy interaction

In view of the prominent role of the zonal-mean divergence anomalies in engendering significant zonal–eddy interactions in the full anomaly model (1), we examine the impact of the uncertainties in their specification. The zonal-mean divergence anomalies are uniquely specified in the forcing latitudes (e.g., the tropical Pacific box latitudes), but the compensating subsidence outside these latitudes—necessary for achieving zero global-mean divergence—remains nonunique, since the dynamics/thermodynamics involved in determining where this subsidence occurs is not well understood (e.g., Rasmusson and Mo 1993) and not represented in barotropic models. In barotropic modeling of the regionally forced response, the compensating subsidence is typically spread uniformly over all latitudes (e.g., Sardeshmukh and Hoskins 1988), and the above model calculations use the same strategy.

In this section, we describe a model calculation forced again by the observed divergence in the tropical Pacific box (Fig. 2a), but with the compensating zonal-mean subsidence restricted, perhaps more reasonably, to the 45°S–45°N latitude belt; this calculation is performed at the R30 resolution to facilitate divergence specification.5 Figure 11a compares the zonal-mean divergence anomalies in the latitudinally confined and globally uniform subsidence (Fig. 6c) experiments. A comparison of the tropically forced wave pattern obtained using confined subsidence (Fig. 11b) with Fig. 10a (which is very similar to the “uniform subsidence” solution at the higher R30 resolution) shows considerable diminution of wave amplitudes in the high latitudes, indicating weak zonal–eddy interaction in this calculation. Not surprisingly, Fig. 11b resembles the response to eddy forcing alone (Fig. 10c), more than the response shown in Fig. 10a, which includes the effects of zonal-mean forcing as well. The corresponding U anomalies, both shown at the R30 resolution in Fig. 11d, provide additional evidence for the weakened zonal–eddy interactions; Ua in the confined subsidence case is not even half as strong as the Ua obtained with global subsidence, particularly in the climatological jet latitudes (25°–35°N)—the locus of zonal–eddy interactions.

Figure 11c shows the orographic modulation from the consistent model in the confined subsidence case. Comparison with Fig. 6c shows that the orographic modulation is now about half as strong as that obtained with globally uniform subsidence, consistent with the reduction of primary wave amplitude over the Tibetan Plateau. However, comparison of Figs. 11b and 11c shows that even though the absolute magnitude of orographic modulation is considerably diminished in the confined subsidence case, its strength relative to the primary wave pattern (Fig. 11b) in midlatitudes is not much different from that in the global subsidence case (compare Fig. 6b with Fig. 10a).

The above analysis suggests that when the zonal-mean subsidence that offsets anomalous tropical divergence is more reasonably distributed in latitude, it leads to less vigorous zonal–eddy interaction than it would if the compensating subsidence were spread uniformly over all latitudes (as in extant barotropic modeling of the regionally forced response). The insensitivity of the extratropical response to changes in the tropical forcing longitude, as well as the magnitude of orographic modulation, will be overestimated by the full anomaly model if the zonal–eddy interaction in the model is too strong.

7. Discussion and concluding remarks

This study has attempted to diagnose the contribution of the interaction between tropically forced circulation anomalies and the underlying extratropical mountains in the generation of winter circulation anomalies in the northern extratropics. The diagnosis is undertaken in the context of the contrasting 1987/88 and 1988/89 winter circulations that were associated with anomalous conditions in the tropical Pacific basin as well as with the opposite phases of the midlatitude zonal index cycle.

The diagnosis is performed using a divergent barotropic model. The modeling of orographic interaction with such a model is justified in view of the equivalent barotropic vertical structure of the far-field rotational flow response of tropical heating anomalies in both models and nature. The model is used to diagnose boththe zonal mean and zonally asymmetric flow anomalies because of evidence for zonal–eddy interactions in the upper troposphere during northern winter. The barotropic model is, however, not quite well suited for modeling zonal–eddy interactions that occur in response to tropical heating anomalies, because in this framework, not only must the tropical divergence anomaly be specified, but also the latitudinal distribution of compensating zonal-mean subsidence outside the forcing latitudes—a distribution that is not a priori known and that must therefore be somewhat arbitrarily specified. Additionally, the zonal-mean zonal wind anomaly—central to zonal–eddy interactions—can be sensitively dependent on the dissipation parameterization in the barotropic model.

The search for observational evidence for the diagnosed orographic modulation is complicated by the presence of large zonal index departures in these winters, since substantial eddy streamfunction anomalies are associated with zonal index fluctuations (Ting et al. 1996; DeWeaver and Nigam 1997). However, barotropic modeling indicates this modulation to be substantial over the Pacific–North American region.

  • The orographic modulation of the rotational response to 200-mb divergence (heating) anomaly in the central tropical Pacific consists of a large-scale wave pattern with a ridge in the central subtropical Pacific trough over the Gulf of Alaska and a weak ridge extending across North America from Baja California to Greenland. These features have amplitude of ∼40 gpm, and the orographic modulation is thus about one-third as strong as the primary wave pattern.

  • The 200-mb zonal wind associated with the modeled orographic modulation is strongest (∼5 m s−1) over the central Pacific region overlapping the eastern end of the East Asian jet, thereby contributing to the southeastward extension of this jet, as is typical in El Niño winters.

  • The Himalayan–Tibetan complex is the major locus of orographic interaction in the model, for it alone accounts for all the features and over two-thirds of the amplitude of the resulting modulation pattern.

  • The eddy and zonal-mean parts of the tropically forced flow anomalies make comparable contributions to orographic modulation in the model, but the midlatitude eddy anomalies themselves result, in part, from the interaction of the zonal-mean zonal wind anomaly and the climatological vorticity gradients (associated with the jets), that is, from zonal–eddy interactions.

  • The zonal–eddy interaction in our model (which solves for both the zonal mean and eddy anomalies), induced by the zonal-mean divergence anomalies, imparts considerable insensitivity to the circulation anomalies in the middle and high latitudes (and thus to orographic modulation) with respect to the longitude of tropical divergence anomaly over the Pacific basin.

The importance of zonal–eddy interaction in determining the model’s response to a tropical divergence anomaly depends strongly on the specified distribution of the compensating zonal-mean subsidence. The strategy of spreading this subsidence uniformly over all latitudes in extant barotropic anomaly modeling is not quite satisfactory; for instance, the observed El Niño-related 200-mb divergence anomaly consists of outflow at the equatorial dateline, which is compensated by convergence in the surrounding western Pacific and the central subtropical Pacific regions (e.g., Dai and Nigam 1998, manuscript submitted to J. Climate).

The consistent model of orographic interaction does allow modification of the specified distribution of the compensating zonal-mean subsidence in the extratropics, but the subsidence in nature is likely impacted also by the other secondary interactions of the tropically forced circulation anomalies, such as with the Pacific stormtracks and the midlatitude Pacific SSTs. The thermodynamics of the compensating subsidence regions remains obscure; for example, is the upper-tropospheric diabatic cooling of these regions during ENSO (Dai and Nigam 1998, manuscript submitted to J. Climate) a response or a contributing factor to the generation of subsidence?

Some of these difficulties can be circumvented by modeling the orographic modulation with a global primitive-equation model that solves for both the rotational and divergent components of the flow in response to forcing by the tropical heating anomalies. Orographic modulation experiments with a sigma-coordinate (=p/ps) primitive-equation model are presently being designed.

Acknowledgments

This work was supported by NSF Grant ATM9316278 and NOAA Grant NA46GP0194 to S. Nigam by DOE/OER/CHAMMP Grant DEFG02-95ER 62022 to F. Baer and by the Cooperative Institute for Climate Studies at the University of Maryland, College Park.

The authors would like to thank Ferdinand Baer and Ming Cai for helpful discussions.

REFERENCES

  • Alexander, M. A., 1992: Midlatitude atmosphere–ocean interaction during El Niño. Part I: The North Pacific Ocean. J. Climate,5,944–958.

  • Bjerknes, J., 1969: Atmospheric teleconnections from the equatorial Pacific. Mon. Wea. Rev.,97, 163–172.

  • Branstator, G., 1984: The relationship between zonal mean flow and quasi-stationary waves in the midtroposphere. J. Atmos. Sci.,41,2163–2178.

  • ——, 1995: Organization of storm track anomalies by recurring low-frequency circulation anomalies. J. Atmos. Sci.,52, 207–226.

  • Charney, J. G., and A. Eliassen, 1949: A numerical method for predicting the perturbations of the middle latitude westerlies. Tellus,1, 38–54.

  • Chelliah, M., and P. Arkin, 1992: Large-scale interannual variability of monthly outgoing longwave radiation anomalies over the global tropics. J. Climate,5, 371–389.

  • Chen, S. C., and K. E. Trenberth, 1988: Forced planetary waves in the Northern Hemisphere winter: Wave-coupled orographic and thermal forcings. J. Atmos. Sci.,45, 682–704.

  • DeWeaver, E., and S. Nigam, 1995: Influence of mountain ranges on the mid-latitude atmospheric response to El Niño events. Nature,378, 706–708.

  • ——, and ——, 1997: Dynamics of zonal-mean flow assimilation and implications for winter circulation anomalies. J. Atmos. Sci.,54,1758–1775.

  • Geisler, J. E., M. L. Blackmon, G. T. Bates, and S. J. Muñoz, 1985: Sensitivity of January climate response to the magnitude and position of equatorial Pacific sea-surface temperature anomalies. J. Atmos. Sci.,42, 1037–1049.

  • Grimm, A. M., and P. L. Silva Dias, 1995: Use of barotropic models in the study of the extratropical response to tropical heat sources. J. Meteor. Soc. Japan,73, 765–780.

  • Held, I. M., 1983: Stationary and quasi-stationary eddies in the extratropical troposphere: Theory. Large-Scale Dynamical Processes in the Atmosphere, B. J. Hoskins and R. P. Pearce, Eds., Academic Press, 127–168.

  • ——, and I.-S. Kang, 1987: Barotropic models of the extratropical response to El Niño. J. Atmos. Sci.,44, 3576–3586.

  • ——, S. W. Lyons, and S. Nigam, 1989: Transients and the extratropical response to El Niño. J. Atmos. Sci.,46, 163–174.

  • Hoerling, M. P., and M. Ting, 1994: Organization of extratropical transients during El Niño. J. Climate,7, 745–766.

  • ——, ——, and A. Kumar, 1995: Zonal flow–stationary wave relationship during El Niño: Implications for seasonal forecasting. J. Climate,8, 1838–1852.

  • Horel, J. D., and J. M. Wallace, 1981: Planetary-scale atmospheric phenomena associated with the Southern Oscillation. Mon. Wea. Rev.,109, 813–829.

  • Hoskins, B. J., and D. J. Karoly, 1981: The steady linear response of a spherical atmosphere to thermal and orographic forcing. J. Atmos. Sci.,38, 1179–1196.

  • Hsu, H.-H., and J. M. Wallace, 1985: Vertical structure of wintertime teleconnection patterns. J. Atmos. Sci.,42, 1693–1710.

  • Kang, I. S., and N. C. Lau, 1986: Principal modes of atmospheric variability in model atmospheres with and without anomalous sea surface temperature forcing in the tropical Pacific. J. Atmos. Sci.,43, 2719–2735.

  • Kok, C. J., and J. D. Opsteegh, 1985: On the possible causes of anomalies in seasonal mean circulation pattern during the 1982–83 El Niño event. J. Atmos. Sci.,42, 677–694.

  • Lau, N. C., 1985: Modeling the seasonal dependence of the atmospheric response to observed El Niños in 1962–76. Mon. Wea. Rev.,113, 1970–1996.

  • ——, and M. J. Nath, 1991: Variability of the baroclinic and barotropic transient eddy forcing associated with monthly changes in the midlatitude storm tracks. J. Atmos. Sci.,48, 2589–2613.

  • ——, and ——, 1996: The role of the “atmospheric bridge” in linking tropical Pacific ENSO events to extratropical SST anomalies. J. Climate,9, 2036–2057.

  • Lorenz, E. N., 1967: The Nature and Theory of the General Circulation of the Atmosphere. World Meteorological Organization, 161 pp.

  • Manabe, S., and T. B. Terpstra, 1974: The effects of the mountains on the general circulation of the atmosphere as identified by numerical experiment. J. Atmos. Sci.,31, 3–42.

  • Nigam, S., 1990: On the structure of variability of the observed tropospheric and stratospheric zonal-mean zonal wind. J. Atmos. Sci.,47, 1799–1813.

  • ——, and R. S. Lindzen, 1989: The sensitivity of stationary waves to variations in the basic state zonal flow. J. Atmos. Sci.,46,1746–1768.

  • ——, and H.-S. Shen, 1993: Structure of oceanic and atmospheric low-frequency variability over the tropical Pacific and Indian oceans. Part I: COADS observations. J. Climate,6, 657–676.

  • ——, I. M. Held, and S. W. Lyons, 1988: Linear simulation of the stationary eddies in a general circulation model. Part II: The mountain model. J. Atmos. Sci.,45, 1433–1452.

  • Rasmusson, E. M., and K. Mo, 1993: Linkages between 200-mb tropical and extratropical circulation anomalies during the 1986–1989 ENSO cycle. J. Climate,6, 595–616.

  • Ropelewski, C. F., and M. S. Halpert, 1987: Global and regional-scale precipitation associated with El Niño/Southern Oscillation. Mon. Wea. Rev.,115, 1606–1626.

  • Sardeshmukh, P. D., 1993: The baroclinic χ problem and its application to the diagnosis of atmospheric heating rates. J. Atmos. Sci.,50, 1099–1112.

  • ——, and B. J. Hoskins, 1988: The generation of global rotational flow by steady idealized tropical divergence. J. Atmos. Sci.,45,1228–1251.

  • Schneider, E. K., 1988: A formulation for diagnostic anomaly models. Pure Appl. Geophys.,126, 137–140.

  • Ting, M., M. P. Hoerling, T. Xu, and A. Kumar, 1996: Northern Hemisphere teleconnection patterns during extreme phases of the zonal-mean circulation. J. Climate,9, 2614–2633.

  • Valdes, P. J., and B. J. Hoskins, 1989: Linear stationary wave simulations of the time-mean climatological flow. J. Atmos. Sci.,46, 2509–2527.

  • Wallace, J. M., C. Smith, and Q. Jiang, 1990: Spatial patterns of atmosphere–ocean interaction in the northern winter. J. Climate,3, 990–998.

Fig. 1.
Fig. 1.

OLR and SST anomalies between the winters (DJF) of 1987/88 and 1988/89: (a) OLR anomalies between 30°S and 30°N with a contour interval and shading threshold of 10 W m−2; (b) global SST anomalies are shown with a contour interval and shading threshold of 0.5 K. The zero contour is omitted, and dark (light) shading is for positive (negative) anomalies in both panels.

Citation: Journal of Climate 11, 4; 10.1175/1520-0442(1998)011<0716:IOOOTE>2.0.CO;2

Fig. 2.
Fig. 2.

Observed and diagnosed anomalies in 200-mb streamfunction, divergence, and zonal-mean zonal wind (U) between DJF 1987/88 and DJF 1988/89. (a) ECMWF analyzed divergence anomaly, (b) divergence anomaly diagnosed residually using the χ-algorithm (the χ-problem is solved for each winter using the analyzed divergence as the first guess), (c) eddy streamfunction anomaly calculated from ECMWF analyses, (d) simulated eddy streamfunction anomaly, forced by the χ-divergence and vorticity transient anomalies (not shown). The contour interval and shading threshold for divergence is 10−6 s−1, while for the streamfunction, it is 3 × 106 m2 s−1 and 6 × 106 m2 s−1, respectively. The zero contour is omitted, and dark (light) shading is for positive (negative) anomalies in all four panels. The side panel shows the U200 anomalies corresponding to the eddy streamfunction anomalies displayed in the bottom panels. The open circles depict the ECMWF-analyzed U200 anomaly, while the filled circles show the U200 simulation.

Citation: Journal of Climate 11, 4; 10.1175/1520-0442(1998)011<0716:IOOOTE>2.0.CO;2

Fig. 3.
Fig. 3.

Model response to tropical divergence forcing and subsequent orographic interaction. (a) 200-mb eddy streamfunction, and (b) U200 forced by the χ-divergence anomaly in the tropical Pacific box in Fig. 2b; the contour interval and shading threshold for streamfunction is 2 × 106m2 s−1 and 6 × 106 m2 s−1, respectively. (c) D, the divergence produced from the interaction of the rotational flow depicted in (a) and (b) and the extratropical mountains (θ > 25°N), (d) 200-mb eddy streamfunction forced by D, and (e) 200-mb eddy streamfunction forced by the interaction of the eddy rotational flow shown in (a) and the extratropical mountains; the contour interval and shading threshold for divergence is 2 × 10−7 s−1, while for the streamfunctions in (d) and (e), it is 106 m2 s−1. The zero contour is omitted in all panels. Note that at 45°N, a streamfunction amplitude of 1 × 106 m2 s−1 corresponds to ∼10 m in geopotential height.

Citation: Journal of Climate 11, 4; 10.1175/1520-0442(1998)011<0716:IOOOTE>2.0.CO;2

Fig. 4.
Fig. 4.

Orographic modulation of the 200-mb streamfunction response to tropical divergence, obtained from the consistent barotropic model of section 5: (a) eddy streamfunction modulation forced by the extratropical (θ > 25°N) mountains, and (b) the modulation forced by extratropical mountains of the Eastern Hemisphere. The contour interval and shading threshold for streamfunction is 106 m2 s−1, and the thin solid lines contour orography higher than 500 m with a 500 m interval. The zero contour is omitted in both panels.

Citation: Journal of Climate 11, 4; 10.1175/1520-0442(1998)011<0716:IOOOTE>2.0.CO;2

Fig. 5.
Fig. 5.

The 200-mb zonal wind response to tropical divergence, and subsequent orographic modulation: (a) zonal wind modulation forced by the extratropical mountains, calculated from the eddy streamfunction shown in Fig. 4a and the corresponding zonal-mean component; (b) zonal wind response to tropical divergence (without orographic interaction) calculated from the eddy streamfunction and U200 fields shown in Figs. 3a and 3b. The contour interval and shading threshold is 1 m s−1 in the top panel and 3 m s−1 in the bottom panel, and the zero contour is omitted in both panels.

Citation: Journal of Climate 11, 4; 10.1175/1520-0442(1998)011<0716:IOOOTE>2.0.CO;2

Fig. 6.
Fig. 6.

Orographic modulation of the 200-mb eddy streamfunction: (a) modulation produced when the scaling factor α = 0.9; (b) modulation of the response to the ECMWF analyzed divergence anomaly in the tropical Pacific box in Fig. 2a (with α = 0.3), and (c) as in (b) but with the barotropic model resolution increased to R30. The contour interval and shading threshold is 3 × 106 m2 s−1 in the top panel, that is, three times the interval used in Fig. 4a, and 106 m2s−1 in the other two panels. The thin solid lines contour orography higher than 500 m with a 500-m interval, and the zero contour is omitted in all panels.

Citation: Journal of Climate 11, 4; 10.1175/1520-0442(1998)011<0716:IOOOTE>2.0.CO;2

Fig. 7.
Fig. 7.

Orographic modulation of the 200-mb eddy streamfunction response to an idealized tropical divergence anomaly placed at 5°S and (a) 135°E, (b) 180°, and (c) 135°W. Contouring and shading convention as in Fig. 4a. The divergence anomalies are ellipses centered on the “+”sign in each panel, with zonal and meridional extent of 60° and 20°, respectively, and with maximum amplitude at 6 × 10−6 s−1.

Citation: Journal of Climate 11, 4; 10.1175/1520-0442(1998)011<0716:IOOOTE>2.0.CO;2

Fig. 8.
Fig. 8.

The 200-mb eddy streamfunction response (without orographic interaction, that is, with α = 0) to an idealized tropical divergence anomaly placed at 5°S and (a) 135°E, (b)180°, and (c) 135°W, that is, at the “+” sign in each panel. (d) Response to the Rossby wave source used in (b) in the corresponding zonally symmetric climatological basic state. The idealized divergence anomalies are the same as in Fig. 7. The contour interval and shading threshold for streamfunction is 2 × 106 m2 s−1 and 6 × 106 m2 s−1, respectively, and the zero contour is omitted as before in all panels.

Citation: Journal of Climate 11, 4; 10.1175/1520-0442(1998)011<0716:IOOOTE>2.0.CO;2

Fig. 9.
Fig. 9.

The Rossby wave sources used to produce the model solutions shown in Figs. 8a–c; the source shown in panel (b) is also used to produce the response shown in Fig. 8d. The contour interval and shading threshold is 2 × 10−11 s−2, and the zero contour is omitted in all panels.

Citation: Journal of Climate 11, 4; 10.1175/1520-0442(1998)011<0716:IOOOTE>2.0.CO;2

Fig. 10.
Fig. 10.

Influence of zonal–eddy interaction in determining the model’s 200-mb streamfunction response to tropical divergence in the absence of orographic interaction: (a) response to the ECMWF-analyzed divergence anomaly in the tropical Pacific box in Fig. 2a; (b) response forced by the zonal-mean component of the forcing used in producing (a); (c) response forced by the eddy component of the forcing used in producing (a); (d) response to zonal–eddy interaction, forced by U200 corresponding to panel (a), in a version of the model, which solves only for the eddy components of the streamfunction anomaly. The contour interval and shading threshold is 2 × 106 m2 s−1 and 6 × 106 m2 s−1, respectively, and the zero contour is omitted as before in all panels.

Citation: Journal of Climate 11, 4; 10.1175/1520-0442(1998)011<0716:IOOOTE>2.0.CO;2

Fig. 11.
Fig. 11.

Model response when the compensating convergence that balances the imposed tropical divergence is confined between 45°S and 45°N; the imposed divergence is the ECMWF analyzed anomaly in the tropical Pacific box in Fig. 2a. To better resolve the convergence, the model is solved at R30 resolution. (a) Meridional distribution of zonal-mean divergence when the compensating convergence is either distributed uniformly at all latitudes (open circles), or restricted to |θ| < 45° (filled circles); (b) 200-mb eddy streamfunction response obtained without orographic interaction; (c) orographic modulation of the streamfunction response, with α = 0.3; (d) the zonal-mean zonal wind (U200) response obtained without orographic interaction, when the compensating convergence is distributed uniformly over all latitudes (open circles) or confined to |θ| < 45° (filled circles). The U200 profile shown by the filled circles thus corresponds to the eddy streamfunction response in panel (b). The contour interval and shading threshold is 2 × 106 m2 s−1 and 6 × 106 m2 s−1, respectively, in panel (b), and 106m2 s−1 in panel (c), and the zero contour is omitted in these panels.

Citation: Journal of Climate 11, 4; 10.1175/1520-0442(1998)011<0716:IOOOTE>2.0.CO;2

1

Chen and Trenberth also considered the forcing resulting from the interaction of thermally forced stationary waves with orography. They found, using a primitive equation model, that this additional forcing makes a substantial contribution to the climatological northern winter stationary waves. However, the orographic interaction of the thermally forced stationary waves excited specifically by tropical heating was not discussed in this study.

2

The reduction in subtropical convection is a robust feature that shows up as anomalous 200 mb convergence and as diabatic cooling of the upper (200–500 mb) subtropics in the El Niño related modal structure obtained from a rotated principal component analysis of combined interannual variability (Dai and Nigam 1998, manuscript submitted to J. Climate).

3

The Ua simulation obtained from the ECMWF analyzed divergence (and vorticity transient) anomalies also agrees well with the observed Uain the northern subtropics and midlatitudes, but not in the high latitudes (θ ≳ 60°N); see Fig. 7b in DeWeaver and Nigam (1997).

4

Grimm and Silva Dias (1995) argue that the relative insensitivity of the extratropical response in the Geisler et al. (1985) experiments, at least from the barotropic model’s perspective, results from the occurrence of suppressed rainfall (or upper-level convergence) in the subtropical central Pacific in all their GCM integrations, irrespective of the longitudinal position of the specified SST anomaly (in the central–eastern tropical Pacific basin). Thus, from the barotropic model’s view point, the issue is not one of insensitivity to the divergence anomaly position.

5

Compensating convergence is confined to |θ| < 45° using an iterative method in which convergence is set to zero for |θ| > 45°. Removal of convergence at higher latitudes introduces a positive global-mean divergence, which is subtracted equally from each latitude, including the higher ones, by transforming to the spectral domain and removing the global coefficient. After each iteration, subsidence poleward of 45° is reduced by a factor of 3.

Save
  • Alexander, M. A., 1992: Midlatitude atmosphere–ocean interaction during El Niño. Part I: The North Pacific Ocean. J. Climate,5,944–958.

  • Bjerknes, J., 1969: Atmospheric teleconnections from the equatorial Pacific. Mon. Wea. Rev.,97, 163–172.

  • Branstator, G., 1984: The relationship between zonal mean flow and quasi-stationary waves in the midtroposphere. J. Atmos. Sci.,41,2163–2178.

  • ——, 1995: Organization of storm track anomalies by recurring low-frequency circulation anomalies. J. Atmos. Sci.,52, 207–226.

  • Charney, J. G., and A. Eliassen, 1949: A numerical method for predicting the perturbations of the middle latitude westerlies. Tellus,1, 38–54.

  • Chelliah, M., and P. Arkin, 1992: Large-scale interannual variability of monthly outgoing longwave radiation anomalies over the global tropics. J. Climate,5, 371–389.

  • Chen, S. C., and K. E. Trenberth, 1988: Forced planetary waves in the Northern Hemisphere winter: Wave-coupled orographic and thermal forcings. J. Atmos. Sci.,45, 682–704.

  • DeWeaver, E., and S. Nigam, 1995: Influence of mountain ranges on the mid-latitude atmospheric response to El Niño events. Nature,378, 706–708.

  • ——, and ——, 1997: Dynamics of zonal-mean flow assimilation and implications for winter circulation anomalies. J. Atmos. Sci.,54,1758–1775.

  • Geisler, J. E., M. L. Blackmon, G. T. Bates, and S. J. Muñoz, 1985: Sensitivity of January climate response to the magnitude and position of equatorial Pacific sea-surface temperature anomalies. J. Atmos. Sci.,42, 1037–1049.

  • Grimm, A. M., and P. L. Silva Dias, 1995: Use of barotropic models in the study of the extratropical response to tropical heat sources. J. Meteor. Soc. Japan,73, 765–780.

  • Held, I. M., 1983: Stationary and quasi-stationary eddies in the extratropical troposphere: Theory. Large-Scale Dynamical Processes in the Atmosphere, B. J. Hoskins and R. P. Pearce, Eds., Academic Press, 127–168.

  • ——, and I.-S. Kang, 1987: Barotropic models of the extratropical response to El Niño. J. Atmos. Sci.,44, 3576–3586.

  • ——, S. W. Lyons, and S. Nigam, 1989: Transients and the extratropical response to El Niño. J. Atmos. Sci.,46, 163–174.

  • Hoerling, M. P., and M. Ting, 1994: Organization of extratropical transients during El Niño. J. Climate,7, 745–766.

  • ——, ——, and A. Kumar, 1995: Zonal flow–stationary wave relationship during El Niño: Implications for seasonal forecasting. J. Climate,8, 1838–1852.

  • Horel, J. D., and J. M. Wallace, 1981: Planetary-scale atmospheric phenomena associated with the Southern Oscillation. Mon. Wea. Rev.,109, 813–829.

  • Hoskins, B. J., and D. J. Karoly, 1981: The steady linear response of a spherical atmosphere to thermal and orographic forcing. J. Atmos. Sci.,38, 1179–1196.

  • Hsu, H.-H., and J. M. Wallace, 1985: Vertical structure of wintertime teleconnection patterns. J. Atmos. Sci.,42, 1693–1710.

  • Kang, I. S., and N. C. Lau, 1986: Principal modes of atmospheric variability in model atmospheres with and without anomalous sea surface temperature forcing in the tropical Pacific. J. Atmos. Sci.,43, 2719–2735.

  • Kok, C. J., and J. D. Opsteegh, 1985: On the possible causes of anomalies in seasonal mean circulation pattern during the 1982–83 El Niño event. J. Atmos. Sci.,42, 677–694.

  • Lau, N. C., 1985: Modeling the seasonal dependence of the atmospheric response to observed El Niños in 1962–76. Mon. Wea. Rev.,113, 1970–1996.

  • ——, and M. J. Nath, 1991: Variability of the baroclinic and barotropic transient eddy forcing associated with monthly changes in the midlatitude storm tracks. J. Atmos. Sci.,48, 2589–2613.

  • ——, and ——, 1996: The role of the “atmospheric bridge” in linking tropical Pacific ENSO events to extratropical SST anomalies. J. Climate,9, 2036–2057.

  • Lorenz, E. N., 1967: The Nature and Theory of the General Circulation of the Atmosphere. World Meteorological Organization, 161 pp.

  • Manabe, S., and T. B. Terpstra, 1974: The effects of the mountains on the general circulation of the atmosphere as identified by numerical experiment. J. Atmos. Sci.,31, 3–42.

  • Nigam, S., 1990: On the structure of variability of the observed tropospheric and stratospheric zonal-mean zonal wind. J. Atmos. Sci.,47, 1799–1813.

  • ——, and R. S. Lindzen, 1989: The sensitivity of stationary waves to variations in the basic state zonal flow. J. Atmos. Sci.,46,1746–1768.

  • ——, and H.-S. Shen, 1993: Structure of oceanic and atmospheric low-frequency variability over the tropical Pacific and Indian oceans. Part I: COADS observations. J. Climate,6, 657–676.

  • ——, I. M. Held, and S. W. Lyons, 1988: Linear simulation of the stationary eddies in a general circulation model. Part II: The mountain model. J. Atmos. Sci.,45, 1433–1452.

  • Rasmusson, E. M., and K. Mo, 1993: Linkages between 200-mb tropical and extratropical circulation anomalies during the 1986–1989 ENSO cycle. J. Climate,6, 595–616.

  • Ropelewski, C. F., and M. S. Halpert, 1987: Global and regional-scale precipitation associated with El Niño/Southern Oscillation. Mon. Wea. Rev.,115, 1606–1626.

  • Sardeshmukh, P. D., 1993: The baroclinic χ problem and its application to the diagnosis of atmospheric heating rates. J. Atmos. Sci.,50, 1099–1112.

  • ——, and B. J. Hoskins, 1988: The generation of global rotational flow by steady idealized tropical divergence. J. Atmos. Sci.,45,1228–1251.

  • Schneider, E. K., 1988: A formulation for diagnostic anomaly models. Pure Appl. Geophys.,126, 137–140.

  • Ting, M., M. P. Hoerling, T. Xu, and A. Kumar, 1996: Northern Hemisphere teleconnection patterns during extreme phases of the zonal-mean circulation. J. Climate,9, 2614–2633.

  • Valdes, P. J., and B. J. Hoskins, 1989: Linear stationary wave simulations of the time-mean climatological flow. J. Atmos. Sci.,46, 2509–2527.

  • Wallace, J. M., C. Smith, and Q. Jiang, 1990: Spatial patterns of atmosphere–ocean interaction in the northern winter. J. Climate,3, 990–998.

  • Fig. 1.

    OLR and SST anomalies between the winters (DJF) of 1987/88 and 1988/89: (a) OLR anomalies between 30°S and 30°N with a contour interval and shading threshold of 10 W m−2; (b) global SST anomalies are shown with a contour interval and shading threshold of 0.5 K. The zero contour is omitted, and dark (light) shading is for positive (negative) anomalies in both panels.

  • Fig. 2.

    Observed and diagnosed anomalies in 200-mb streamfunction, divergence, and zonal-mean zonal wind (U) between DJF 1987/88 and DJF 1988/89. (a) ECMWF analyzed divergence anomaly, (b) divergence anomaly diagnosed residually using the χ-algorithm (the χ-problem is solved for each winter using the analyzed divergence as the first guess), (c) eddy streamfunction anomaly calculated from ECMWF analyses, (d) simulated eddy streamfunction anomaly, forced by the χ-divergence and vorticity transient anomalies (not shown). The contour interval and shading threshold for divergence is 10−6 s−1, while for the streamfunction, it is 3 × 106 m2 s−1 and 6 × 106 m2 s−1, respectively. The zero contour is omitted, and dark (light) shading is for positive (negative) anomalies in all four panels. The side panel shows the U200 anomalies corresponding to the eddy streamfunction anomalies displayed in the bottom panels. The open circles depict the ECMWF-analyzed U200 anomaly, while the filled circles show the U200 simulation.

  • Fig. 3.

    Model response to tropical divergence forcing and subsequent orographic interaction. (a) 200-mb eddy streamfunction, and (b) U200 forced by the χ-divergence anomaly in the tropical Pacific box in Fig. 2b; the contour interval and shading threshold for streamfunction is 2 × 106m2 s−1 and 6 × 106 m2 s−1, respectively. (c) D, the divergence produced from the interaction of the rotational flow depicted in (a) and (b) and the extratropical mountains (θ > 25°N), (d) 200-mb eddy streamfunction forced by D, and (e) 200-mb eddy streamfunction forced by the interaction of the eddy rotational flow shown in (a) and the extratropical mountains; the contour interval and shading threshold for divergence is 2 × 10−7 s−1, while for the streamfunctions in (d) and (e), it is 106 m2 s−1. The zero contour is omitted in all panels. Note that at 45°N, a streamfunction amplitude of 1 × 106 m2 s−1 corresponds to ∼10 m in geopotential height.

  • Fig. 4.

    Orographic modulation of the 200-mb streamfunction response to tropical divergence, obtained from the consistent barotropic model of section 5: (a) eddy streamfunction modulation forced by the extratropical (θ > 25°N) mountains, and (b) the modulation forced by extratropical mountains of the Eastern Hemisphere. The contour interval and shading threshold for streamfunction is 106 m2 s−1, and the thin solid lines contour orography higher than 500 m with a 500 m interval. The zero contour is omitted in both panels.

  • Fig. 5.

    The 200-mb zonal wind response to tropical divergence, and subsequent orographic modulation: (a) zonal wind modulation forced by the extratropical mountains, calculated from the eddy streamfunction shown in Fig. 4a and the corresponding zonal-mean component; (b) zonal wind response to tropical divergence (without orographic interaction) calculated from the eddy streamfunction and U200 fields shown in Figs. 3a and 3b. The contour interval and shading threshold is 1 m s−1 in the top panel and 3 m s−1 in the bottom panel, and the zero contour is omitted in both panels.

  • Fig. 6.

    Orographic modulation of the 200-mb eddy streamfunction: (a) modulation produced when the scaling factor α = 0.9; (b) modulation of the response to the ECMWF analyzed divergence anomaly in the tropical Pacific box in Fig. 2a (with α = 0.3), and (c) as in (b) but with the barotropic model resolution increased to R30. The contour interval and shading threshold is 3 × 106 m2 s−1 in the top panel, that is, three times the interval used in Fig. 4a, and 106 m2s−1 in the other two panels. The thin solid lines contour orography higher than 500 m with a 500-m interval, and the zero contour is omitted in all panels.

  • Fig. 7.

    Orographic modulation of the 200-mb eddy streamfunction response to an idealized tropical divergence anomaly placed at 5°S and (a) 135°E, (b) 180°, and (c) 135°W. Contouring and shading convention as in Fig. 4a. The divergence anomalies are ellipses centered on the “+”sign in each panel, with zonal and meridional extent of 60° and 20°, respectively, and with maximum amplitude at 6 × 10−6 s−1.

  • Fig. 8.

    The 200-mb eddy streamfunction response (without orographic interaction, that is, with α = 0) to an idealized tropical divergence anomaly placed at 5°S and (a) 135°E, (b)180°, and (c) 135°W, that is, at the “+” sign in each panel. (d) Response to the Rossby wave source used in (b) in the corresponding zonally symmetric climatological basic state. The idealized divergence anomalies are the same as in Fig. 7. The contour interval and shading threshold for streamfunction is 2 × 106 m2 s−1 and 6 × 106 m2 s−1, respectively, and the zero contour is omitted as before in all panels.

  • Fig. 9.

    The Rossby wave sources used to produce the model solutions shown in Figs. 8a–c; the source shown in panel (b) is also used to produce the response shown in Fig. 8d. The contour interval and shading threshold is 2 × 10−11 s−2, and the zero contour is omitted in all panels.

  • Fig. 10.

    Influence of zonal–eddy interaction in determining the model’s 200-mb streamfunction response to tropical divergence in the absence of orographic interaction: (a) response to the ECMWF-analyzed divergence anomaly in the tropical Pacific box in Fig. 2a; (b) response forced by the zonal-mean component of the forcing used in producing (a); (c) response forced by the eddy component of the forcing used in producing (a); (d) response to zonal–eddy interaction, forced by U200 corresponding to panel (a), in a version of the model, which solves only for the eddy components of the streamfunction anomaly. The contour interval and shading threshold is 2 × 106 m2 s−1 and 6 × 106 m2 s−1, respectively, and the zero contour is omitted as before in all panels.

  • Fig. 11.

    Model response when the compensating convergence that balances the imposed tropical divergence is confined between 45°S and 45°N; the imposed divergence is the ECMWF analyzed anomaly in the tropical Pacific box in Fig. 2a. To better resolve the convergence, the model is solved at R30 resolution. (a) Meridional distribution of zonal-mean divergence when the compensating convergence is either distributed uniformly at all latitudes (open circles), or restricted to |θ| < 45° (filled circles); (b) 200-mb eddy streamfunction response obtained without orographic interaction; (c) orographic modulation of the streamfunction response, with α = 0.3; (d) the zonal-mean zonal wind (U200) response obtained without orographic interaction, when the compensating convergence is distributed uniformly over all latitudes (open circles) or confined to |θ| < 45° (filled circles). The U200 profile shown by the filled circles thus corresponds to the eddy streamfunction response in panel (b). The contour interval and shading threshold is 2 × 106 m2 s−1 and 6 × 106 m2 s−1, respectively, in panel (b), and 106m2 s−1 in panel (c), and the zero contour is omitted in these panels.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 156 21 3
PDF Downloads 36 14 0