Tropical–Extratropical Interactions Associated with an Atlantic Tropical Plume and Subtropical Jet Streak

Peter Knippertz Department of Atmospheric and Oceanic Sciences, University of Wisconsin—Madison, Madison, Wisconsin

Search for other papers by Peter Knippertz in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

Tropical plumes (TPs) are elongated bands of upper- and midlevel clouds stretching from the Tropics poleward and eastward into the subtropics, typically accompanied by a subtropical jet (STJ) streak and a trough on their poleward side. This study uses ECMWF analyses and high-resolution University of Wisconsin–Nonhydrostatic Modeling System trajectories to analyze the multiscale complex tropical–extratropical interactions involved in the genesis of a pronounced TP and STJ over the NH Atlantic Ocean in late March 2002 that was associated with extreme precipitation in arid northwest Africa. Previous concepts for TP genesis from the literature are discussed in the light of this case study.

Analysis of the upper-level flow prior to the TP formation shows a northeastward propagation and a continuous acceleration of the STJ over the Atlantic Ocean equatorward of a positively tilted upper-level trough to the west of northwest Africa. Both dynamic and advective processes contribute to the generation of the accompanying cloud band. The northern portion of the TP consists of parcels that exit a strong STJ streak over North America, enter the deep Tropics over South America, and then accelerate into the Atlantic STJ, accompanied by strong cross-jet ageostrophic motions, rising, and cloud formation. The southern portion is formed by parcels originating in the divergent outflow from strong near-equatorial convection accompanying the TP genesis. A local increase in the Hadley overturning is found over the tropical Atlantic and east Pacific/South America and appears to be related to low inertial stability at the outflow level and to low-level trade surges associated with the cold advection, sinking, and lower-level divergence underneath two strong upper-level convergence centers in the eastern portions of both a subtropical ridge over North America and an extratropical ridge over the North Atlantic Ocean. Evidence is presented that the convective response lags the trade surge by several days.

Corresponding author address: Peter Knippertz, Institute for Atmospheric Physics, University of Mainz, Becherweg 21, 55099 Mainz, Germany. Email: Knippertz@uni-mainz.de

Abstract

Tropical plumes (TPs) are elongated bands of upper- and midlevel clouds stretching from the Tropics poleward and eastward into the subtropics, typically accompanied by a subtropical jet (STJ) streak and a trough on their poleward side. This study uses ECMWF analyses and high-resolution University of Wisconsin–Nonhydrostatic Modeling System trajectories to analyze the multiscale complex tropical–extratropical interactions involved in the genesis of a pronounced TP and STJ over the NH Atlantic Ocean in late March 2002 that was associated with extreme precipitation in arid northwest Africa. Previous concepts for TP genesis from the literature are discussed in the light of this case study.

Analysis of the upper-level flow prior to the TP formation shows a northeastward propagation and a continuous acceleration of the STJ over the Atlantic Ocean equatorward of a positively tilted upper-level trough to the west of northwest Africa. Both dynamic and advective processes contribute to the generation of the accompanying cloud band. The northern portion of the TP consists of parcels that exit a strong STJ streak over North America, enter the deep Tropics over South America, and then accelerate into the Atlantic STJ, accompanied by strong cross-jet ageostrophic motions, rising, and cloud formation. The southern portion is formed by parcels originating in the divergent outflow from strong near-equatorial convection accompanying the TP genesis. A local increase in the Hadley overturning is found over the tropical Atlantic and east Pacific/South America and appears to be related to low inertial stability at the outflow level and to low-level trade surges associated with the cold advection, sinking, and lower-level divergence underneath two strong upper-level convergence centers in the eastern portions of both a subtropical ridge over North America and an extratropical ridge over the North Atlantic Ocean. Evidence is presented that the convective response lags the trade surge by several days.

Corresponding author address: Peter Knippertz, Institute for Atmospheric Physics, University of Mainz, Becherweg 21, 55099 Mainz, Germany. Email: Knippertz@uni-mainz.de

1. Introduction

Continuous bands of upper- and midlevel clouds, stretching for more than 2000 km from the Tropics poleward and eastward into the subtropics, are a frequently observed feature in IR imagery at all seasons and have been defined as tropical plumes (TPs) by McGuirk et al. (1987, 1988). TPs typically form on the downwind side of a positively tilted (southwest–northeast-oriented) upper-level subtropical trough at low latitudes and are usually accompanied by a subtropical jet (STJ) streak (e.g., McGuirk et al. 1988; Kuhnel 1990; Ziv 2001). Climatological studies for the NH have shown occurrence maxima in the central and eastern Pacific and (to a lesser degree) the Atlantic Ocean during the transition seasons (McGuirk et al. 1987; Kuhnel 1989; Iskenderian 1995). These regions have been termed “westerly ducts” because the refractive properties of the basic flow allow a penetration of Rossby waves from the extratropics into the Tropics and even to the other hemisphere (Webster and Holton 1982; Matthews and Kiladis 1999). Most studies on TPs (or TP-like events) and their dynamics concentrate on the NH Pacific and the cool season (e.g., McGuirk et al. 1987, 1988; Kiladis and Weickmann 1992b, hereafter KW1992b; McGuirk 1993; Mecikalski and Tripoli 1998, hereafter MT98; Blackwell 2000), while other regions and seasons are covered less (e.g., Kuhnel 1990; Geb 2000; Ziv 2001; Knippertz et al. 2003).

Commonly the poleward advection of convectively generated cirrus from the ITCZ by the strong winds at the outflow level is regarded as the dominant factor for the cloud band genesis. A progression of the clouds across 15°N is defined as the “beginning” of a TP (McGuirk et al. 1987). Tropical plume progression rates, however, do not always match the wind speed at the outflow level, and TP cloud bands have been observed to result from consolidation due to both development and advection (McGuirk et al. 1987, 1988; MT98). The TP definitions given by McGuirk et al. (1987) and Iskenderian (1995) explicitly include cloud bands of extratropical origin that penetrate equatorward of 15°–20°N. So far a clear distinction between the different factors involved in TP cloud generation has not been made.

With respect to TP dynamics, several papers (e.g., McGuirk et al. 1988) postulate that quasigeostrophic (QG) forcing associated with large-scale Rossby waves triggers the eruption of the cloud plumes. This view is consistent with the work of Liebmann and Hartmann (1984), Kiladis and Weickmann (1992a, hereafter KW1992a), KW1992b, Iskenderian (1995), and Kiladis (1998), who use averaged, filtered, or composited data of the circulation at upper and midlevels, and outgoing longwave radiation (OLR), to demonstrate a statistical relationship between enhanced tropical convection/TP formation and wave trains from the extratropics. They consistently find strong signals in the westerly duct regions. For the central tropical Pacific, Kiladis (1998) finds characteristics of baroclinic wave developments in the extratropics, but more complicated structures at low latitudes. Prior to the cloud band events he observes a rapid amplification and southeastward propagation of a 350-K potential vorticity (PV) trough downstream of a pronounced ridge and a strong Asian jet. Kiladis (1998) claims that the convection is triggered through vorticity advection and the decrease of static stability ahead of the positive PV anomaly. The resulting convective updrafts drive an anomalous local Hadley circulation, and the upper-level horizontal momentum fluxes downstream of the positively tilted trough accelerate the STJ in the region of the cloud band and the downstream ridge (see also McGuirk et al. 1987). After TP initiation, the anomaly eventually weakens quickly as a consequence of PV destruction caused by convective latent heating.

Being predominantly an upper- and midlevel phenomenon, relatively few studies have considered the low-level circulation associated with TPs. The studies of Kiladis and Weickmann (1997, hereafter KW1997), KW1992b, and Kiladis (1998) reveal anticyclones to the north or northwest of the TP that are associated with trade surges on their eastern sides, leading to near-surface convergence in the TP genesis region. Various studies observe mid- and low-level cold advection/anomalies and a low-level frontal band to the north of the tropical convection and along the northwestern edge of the TP (McGuirk et al. 1988; KW1992b; Kiladis 1998), where an interaction with extratropical synoptic systems is often observed (Thepenier and Cruette 1981).

Another characteristic feature of TPs is a very dry (dark) area in water vapor (WV) imagery to the northwest (McGuirk and Ulsh 1990; Blackwell and McGuirk 1996), where a sharply defined line of upper-level convergence forms at the junction between the divergent outflow from tropical convection and from the storm-track cyclone farther to the northwest (Kiladis 1998). Attendant with this drying, KW1992b, Kiladis (1998), and Blackwell (2000) find moderate to strong subsidence at midlevels. Motivated by these observations Blackwell (2000) used a dry barotropic shallow-water model to demonstrate that TP-like circulations can be initiated downstream of a prescribed convergent forcing region placed in the eastern portion of the wintertime climatological upper ridge over the central Pacific Ocean. In his numerical experiments the advection of vorticity through the convergent winds leads to an amplification and zonal contraction of the trough downstream of the ridge, associated with a steering of extratropical flow toward the deep Tropics. This flow streams up the gradient of geopotential, decelerates, then sharply turns cyclonically and accelerates back into the extratropics, forming an anticyclonically curved downstream STJ streak. Experiments with additional divergent forcing farther upstream, mimicking a Walker-type zonal circulation, generally produce stronger TP events. The effect of Rossby wave excitation by advection of vorticity through the divergent wind has been termed a Rossby wave source (Sardeshmukh and Hoskins 1988). Blackwell (2000) does not explicitly address the cause for the convergent forcing in his study.

MT98 proposed a dynamical concept with a more active role of convective-scale processes. They observe that at least part of the convection in the TP genesis region is relatively remote from the upper-level high-PV anomaly associated with the TP-inducing trough and argue that the associated QG forcing does not penetrate deep enough into the troposphere due to the small Coriolis parameter at low latitudes (chapter 3 in Hoskins et al. 1985; see also section 4a in Kiladis 1998). MT98 suggest that this convection responds to the changes in inertial stability caused by large-scale northeastward advection of low-PV air to the east of the upper PV anomaly. Introducing a new diagnostic parameter termed “inertial available kinetic energy” (IAKE), they quantify the kinetic energy lost or gained when convective outflow, driven by the energy gained in the updraft, expands quasi-horizontally into the surrounding upper-tropospheric environment. IAKE is highly sensitive to convective-scale processes like the difference in vertical stability between the region above the convective heating and the surrounding environment, and the convective transport of horizontal momentum. Since the effect of low IAKE can only come into play when convection is present, MT98 consider a combined measure of low IAKE, high surface equivalent-potential temperatures Θe, and near-surface convergence as an indicator for TP genesis.

The present paper investigates the evolution of a pronounced TP/STJ over the NH Atlantic Ocean in late March 2002. As reported by Fink and Knippertz (2003, hereafter FK), extraordinary rainfalls were observed underneath the extratropical end of this particular TP over arid northwestern Africa. The investigations are based on European Centre for Medium-Range Weather Forecasts (ECMWF) analyses and on data from a simulation with the University of Wisconsin–Nonhydrostatic Modeling System (UW-NMS), which allows for the calculation of high-resolution trajectories to trace back the origin and characteristics of TP parcels, to distinguish between dynamical and advective cloud generation, and to analyze the acceleration of the STJ streak accompanying the TP. To the best of the author’s knowledge this technique was not yet used for the analysis of TPs. Subsequent to this introduction section 2 provides basic information on the datasets, computational methods, and the UW-NMS model used in the analysis. The investigation part of the paper consists of a description of the TP clouds and the upper-level jet in both observations and model data (section 3), a detailed analysis of upper-level trajectories (section 4), an examination of the circulation at lower levels on the basis of isentropic flow charts and trajectories (section 5), and an analysis of the large-scale tropical–extratropical interactions based on isentropic PV charts (section 6). The results will be summarized in section 7 and further discussed in the concluding section 8.

2. Data

For the representation of all large-scale upper-air atmospheric variables twice daily (0000 and 1200 UTC) ECMWF Tropical Ocean and Global Atmosphere (TOGA) operational analyses are used in 2.5° × 2.5° horizontal resolution on standard pressure levels (Trenberth 1992). The General Meteorological Package (GEMPAK) has been employed for the interpolation to isentropic levels and the computation and display of streamlines and various other meteorological parameters. Divergent wind vectors were calculated from the gradient of the velocity potential χ (with maxima in χ corresponding to convergence). Infrared and WV images from the Geostationary Operational Environmental Satellite (GOES) and the Meteorological Satellite (Meteosat) have been obtained through the Space Science and Engineering Center of the University of Wisconsin—Madison.

To obtain high-resolution data for the trajectory analysis a simulation with the UW-NMS (Tripoli 1992) was carried out. The model operates on a single 160 × 97 points grid with 75-km grid spacing and 40 vertical levels up to 22.6 km. The model domain spans 15°S–50°N, 84°W–24°E. The simulation was initialized with the ECMWF analysis of 0000 UTC 26 March 2002 and was then run for 8.5 days. Updating the model with ECMWF data every 12 h at the boundary, a physically consistent “interpolation” of the analyses was produced. The simulated precipitation is gratifyingly similar to station observations (Knippertz and Martin 2005, hereafter KM). More details on the model configuration can be found in Mecikalski and Tripoli (2003). The visualization software VIS5D (Hibbard et al. 1996) was used for trajectory computation from the hourly model output and for a three-dimensional data display.

3. TP and STP streak evolution

a. Observations

For the representation of the STJ in Figs. 1 and 2 the 345-K isentropic surface is used, where highest wind speeds are analyzed throughout most of the TP development phase. At 0000 UTC 26 March, 5 days before the TP reaches its mature stage, high upper-level wind speeds are found at the crests of the ridges over North America and North Africa, while over the subtropical Atlantic only a weak local maximum of 34 m s−1 is observed to the northeast of the Lesser Antilles (Fig. 1a). From 26 to 30 March the Atlantic jet maximum propagates northeastward and intensifies by about 10 m s−1 per day (Fig. 1). At 0000 UTC 28 March the jet core (of now 52 m s−1) is located to the northwest of the Cape Verde Islands (Fig. 1c). At this time, no prominent cloud features were observed in IR imagery in the entire region between the Lesser Antilles/northern South America and West Africa (not shown). Between 0000 UTC 28 and 1200 UTC 29 March the jet maximum further increases to 66 m s−1 and slowly propagates eastward; at this time first indications of cloud formation along the anticyclonic shear side of the jet appear in IR imagery (Fig. 2a). Full wind streamlines at 345 K show that the jet maximum is located in the southwesterly flow between a strong zonally elongated anticyclone centered over southwestern West Africa and a trough with an axis reaching from the west of Ireland southwestward into the subtropical Atlantic Ocean to about 25°N (Fig. 2a). The jet entrance region is located close to the Caribbean coast of Venezuela and Colombia (∼13°N), where strong confluence occurs between a broad trough close to the Lesser Antilles and the southerly (partly cross-equatorial) flow from the convectively active region over eastern tropical South America. More convection is found in the region of southeasterly flow over the tropical Atlantic Ocean between the equator and 5°N.

Between 1200 UTC 29 and 0000 UTC 30 March the large-scale flow pattern changes very little, but the jet maximum increases by 8 m s−1 to 74 m s−1 and approaches the African continent over the west Sahara (Figs. 1e and 2b). The clouds just to the south of the axis thicken, in particular in the right entrance region and to the south of the wind speed maximum (Fig. 2b). During the 12 h between 1200 UTC 30 March (Fig. 2c) and 0000 UTC 31 March (Figs. 1f and 2d) the trough over the Atlantic Ocean and the anticyclone over West Africa move slowly eastward. At the end of this period the STJ maximum reaches its highest value of almost 80 m s−1 at the crest of the ridge over the High Atlas Mountains in Morocco. By that time the TP has further thickened and now forms an almost continuous, ∼6000 km long band of clouds from near South America to North Africa north of 30°N where some indications of the TP-typical anticyclonic curvature are evident (Fig. 2d). Strong convection is observed under the southerly or southeasterly upper-level winds over eastern South America and the tropical Atlantic between the equator and 5°N during the end of the acceleration phase (Figs. 2c and 2d).

By 1200 UTC 31 March, the jet maximum has moved to just downstream of the ridge crest over Tunisia–Algeria and weakened slightly (Fig. 2e). It is not until this time that the TP under investigation clearly fulfills the definitions given by McGuirk et al. (1987) and Iskenderian (1995). The sharp poleward edge of the TP marks almost exactly the axis of the jet all the way from its entrance near South America to the crest over North Africa as observed in other TP cases (Thepenier and Cruette 1981; Iskenderian 1995; KM). From 1200 UTC 31 March to 0000 UTC 1 April the trough over the Atlantic Ocean weakens and the TP starts to move away from the Tropics (Figs. 2e and 2f). Meanwhile the anticyclone over the tropical Atlantic/Africa splits into two separate centers, potentially as a reaction to the strong outflow from convection over the western tropical Atlantic over the previous days. In the course of 1 April the TP weakened substantially (not shown).

According to FK the TP-related rainfalls began on 30 March over northern Mauritania and western Algeria and spread into southern Morocco, reaching peak intensities under the thick clouds over northwestern Africa at 0000 UTC 1 April (Fig. 2f). The patchy wind maxima in Figs. 2e and 2f might be the result of outflow from the strong convection while the rapid decay of the upper trough is likely related in some way to diabatic PV destruction through latent heat release (cf. Kiladis 1998). Under the axis of the trough localized convection is evident over the Atlantic Ocean and the Canary Islands (Fig. 2e).

b. UW-NMS simulation

In the following the satellite observations shown in Fig. 2 will be compared to UW-NMS output. “Clouds” in the model are defined as regions where the logarithm of total water condensate mixing ratio (graupel, pristine crystals, rain, etc. in grams per kilogram, LCMR hereafter) is greater than zero. With this definition the precipitation zones underneath the cloud base are included in the clouds. Using VIS5D-generated three-dimensional isosurfaces of LCMR = 0.1, the horizontal extensions of the TP clouds were determined and are displayed in Fig. 3 with the same map projection used in Fig. 2 (for examples of LCMR = 0.1 isosurfaces, see Figs. 4a and 5a). Usually the TP had an easy to identify boundary, but the tropical end was at times not very clear-cut.

In accordance with Fig. 2a first indications of TP clouds are simulated for 1200 UTC 29 March (solid line in Fig. 3a). By 0000 UTC 30 March the model TP has thinned and stretched out into North Africa. At this stage the average propagation speed of the “leading edge” of nearly 60 m s−1 (or more than 2500 km in 12 h) is near the magnitude of the jet core (Figs. 2a and 2b). Between 0000 UTC 30 and 0000 UTC 31 March the TP thickens slightly, and the subtropical end starts to move poleward and show an anticyclonic curvature (Fig. 3). By 0000 UTC 1 April the TP has finally moved away from the genesis region in the Tropics, increased the cyclonic curvature in its central part, and widened substantially over the precipitation region in northwestern Africa (short dashed line in Fig. 3b). Overall reasonable agreement of the model cloud evolution with the satellite observations is found. The same is true for the model STJ that reached 78 m s–1 at 1400 UTC 30 March in agreement with ECMWF analyses (Fig. 2c).

The three-dimensional display of the model clouds also allows an estimate of the vertical extension of the TP, which is impossible to derive from satellite images in regions of extensive cirrus. Figure 4 shows an example for 1200 UTC 30 March. While the model clouds almost span the depths of the troposphere in the convectively active “TP genesis region” over South America, the entire long axis of the TP from ∼8°N northeastward to the coast of Africa merely consists of high clouds (7–14 km; Fig. 4b) in agreement with the absence of a discernable surface front (see below) and precipitation (KM). On 30 March deep clouds develop over northwestern Africa under the high clouds of the TP (Fig. 4b) in connection with moisture transports from tropical Africa (FK; KM). These results demonstrate that, even though TPs usually appear as continuous cloud bands in satellite IR imagery, vertical cloud structure and precipitation might vary substantially along the TP.

4. Trajectory analysis

To identify the origin of air parcels involved in the TP formation, a large ensemble of backward trajectories was started from various locations within the upper-level clouds (9–12 km) over Africa and the Atlantic Ocean at different times during the TP development phase (29–31 March). Three-dimensional winds from the model simulation back to 0000 UTC 26 March were used for the trajectory computation. The trajectory ensemble could be divided into two clearly distinguishable clusters (not shown). One exemplary trajectory from each cluster is shown in Fig. 5a, together with the model clouds at 0000 UTC 31 March, and total and ageostrophic (only north of 5°N) winds. Figures 5b–i depict time series of various meteorological parameters along the trajectories. The gray shading indicates the passage of the respective trajectory through the TP cloud band (based on LCMR values greater than zero; Figs. 5b and 5f).

a. Trajectory 1

Trajectory 1 enters the model domain at 234 hPa (11.4 km) on the 345-K isentropic level at 1200 UTC 26 March in a strong (39 m s−1) northwesterly flow over the cloud-free eastern side of a pronounced ridge over southern North America. The small angle between the relatively weak ageostrophic and the total wind suggests a flow in near-gradient wind balance. The parcel’s low PV points to an origin in the tropical troposphere. By 0000 UTC 27 March the parcel has entered the deep Tropics (∼12°N) and has decelerated to 26 m s−1. An almost equal ageostrophic component of 24 m s−1 nearly perpendicular to the flow indicates a loss of rotational balance. After another 24 h (0000 UTC 28 March) the parcel reaches its southernmost point over eastern Colombia (just north of 5°N) and its minimum total wind speed of only 7 m s−1. An ageostrophic component of more than double the total wind speed characterizes the now highly buoyancy-driven flow. By 0000 UTC 29 March the parcel has moved northeastward and strongly accelerated to 22.5 m s−1, accompanied by ageostrophic motions of up to 30 m s−1 between 1200 UTC 28 March and 0000 UTC 29 March. The strong acceleration of the flow in this region is reflected by downgradient crossing of contours of Montgomery potential (see Figs. 8c and 9c). At 1800 UTC 28 March the parcel reaches its lowest point at 279 hPa (10.3 km). Since the parcel approximately conserves its Θ, the 345-K isentropic surface must slope downward from the subtropics toward the Tropics in this region (see also Fig. 8b). Figure 5b reveals that the parcel’s passage over South America is accompanied by a substantial increase in mixing ratio (MR) up to nearly 0.3 g kg−1, accompanied by a strong increase in RH to over 50%. The former is presumably caused by convective mixing.

Between 0000 and 1200 UTC 29 March the ageostrophic component of the wind more than halves, indicating the beginning of a recovery of geostrophic balance. Before this process can be completed, however, the parcel reaches the right entrance region of the STJ over the Atlantic Ocean (Fig. 2a) and ageostrophic winds reintensify to over 20 m s−1 on 30 March, further accelerating the parcel to a maximum wind speed of 65 m s−1 just to the west of West Africa on 0600 UTC 30 March. The slightly negative PV values the parcel obtains on 29 March as far away from the equator as 15°N indicates inertial instability. This is supported by the observation of strong upper-level divergence along the jet axis (Fig. 4c in FK). Along this portion of the trajectory the parcel approximately maintains its Θ, but rises (positive w) until it reaches its minimum pressure of 203 hPa (maximum height of 12.1 km) at the crest of the ridge over Africa at 1200 UTC 30 March (filled circle in Fig. 5a). At this point the flow appears to have regained gradient wind balance as indicated by the almost parallel ageostrophic and full winds. Despite a slight decrease in MR the adiabatic cooling associated with the flow on upward-sloping isentropes (see also Fig. 9b) is sufficient to increase RH to over 70% and generate a TP cloud, as indicated by the positive values of LCMR between 1200 UTC 29 March and 1200 UTC 30 March. Because of the very low MR at upper-tropospheric levels, the condensational heating within the TP hardly alters the parcel’s Θ. Shortly before 1200 UTC 30 March the parcel passes the region of precipitation over the Sahara. After the ridge crest trajectory 1 decelerates, sinks, and the cloud dissolves, giving rise to the characteristic anticyclonic curvature of the TP in the subtropics.

b. Trajectory 2

Trajectory 2 starts at midlevels (∼467 hPa, 6.4 km) over the convectively active region in eastern Brazil (10°S, 47°W; Fig. 5). With positive w between 0000 UTC 26 and 0000 UTC 28 March the parcel rises to 245 hPa (11.2 km) while slowly approaching the equator (horizontal wind speeds are less than 10 m s−1 throughout this period). Relative humidity of more than 75%, LCMR values greater than zero, and a strong decrease in MR between 0000 UTC 26 and 0600 UTC 27 March indicate that the parcel passes through a precipitating cloud. The strong increase in Θ (15 K during the first 36 h), combined with only a small increase in Θe, points to a basically moist-adiabatic process. Slightly negative PV (–0.1 to –0.2 PVU; 1 PVU = 10−6 m2 s–1 K kg−1) is consistent with the parcel’s SH origin. Although single air parcels realistically pass much faster through convective updrafts, the author believes that the model’s vertical motion field used for the trajectory computation yields useful information, since it reflects the larger-than-convective-scale (horizontally averaged) impact of the convective updrafts. Until 1200 UTC 29 March the parcel continues to slowly move northward to 6°N (horizontal wind speeds merely exceed 10 m s−1).

Between 1200 UTC 29 and 1200 UTC 31 March the parcel steadily accelerates northeastward until it reaches a maximum of 47 m s−1 at the crest of the ridge over Africa (Figs. 5a and 5i). During the entire acceleration phase strong ageostrophic winds perpendicular to the flow are observed until gradient wind balance is finally achieved around 1200 UTC 31 March. While approaching the right entrance region of the jet, the parcel’s PV changes sign, probably due to mixing with the nearby NH air (note the converging trajectories in Fig. 5a). Satellite images reveal that this mixing takes place in a region where numerous “ripples” in the cloud structure are indicative of strong gravity wave activity (Figs. 2c and 2d).

Variations in the parcel’s Θ, Θe, RH, MR, LCMR, pressure, and w between 0000 UTC 28 March and 0600 UTC 29 March are likely to result from convective processes alone (up- and downdrafts, mixing). During the 30 h following 0600 UTC 29 March, however, the evolution is dominated by the strong horizontal acceleration into the jet over the Atlantic Ocean and shows similarities to trajectory 1. During this time the parcel maintains its Θ, Θe, and MR, but rises slightly to 221 hPa (11.8 km). Though small in magnitude the associated cooling and RH increase initiate cloud formation on the southwestern flank of the TP. This cloud is advected quasi-horizontally northeastward until the strong subsidence to the east of the ridge crest over Africa and the associated decrease in RH leads to evaporation of the condensate. In accordance with the changes in PV along this portion of the trajectory, the decrease in both Θ and Θe suggests a mixing with cooler and drier air, possibly supported by radiative cooling. The smaller slope of the 345-K isentropic surface farther equatorward (and the associated smaller lifting; cf. Figs. 5d and 5h) explains the commonly observed thinner clouds along the southern flank of the TP.

5. Evolution at lower levels

In this section the low-level circulation accompanying the TP evolution is analyzed with the help of isentropic charts and trajectories. The focus is on the subtropical anticyclones and trade surges as described by Kiladis (1998). In the absence of latent heat release and on a time scale of days (which is too short for radiative cooling to exert a substantial influence) the low-level flow from the subtropics into the Tropics basically follows the downward sloping isentropes. Only in the immediate vicinity of the surface, sensible heat fluxes and friction cause considerable deviations from adiabatic movement. The 300–K isentropic surface slopes downward from about 500 hPa in the midlatitudes to below 900 hPa in near-equatorial regions (or even below ground over the hot tropical and subtropical continents), and therefore seems most adequate for this analysis (Fig. 6). Given the decreasing validity of the geostrophic approximation in the Tropics, a full wind streamline representation is preferred to Montgomery potential.

At 1200 UTC 26 March, 3 days before the first indications of TP clouds (Fig. 2a), 300-K streamlines show a weak ridge near northeastern North America (Fig. 6a), associated with a small surface anticyclone as shown by UW-NMS generated sea level pressure (SLP) distributions (Fig. 7a). A large SW–NE tilted anticyclone and a small downstream cyclone are found to the west of Africa (Fig. 6a). The region in between the two features is characterized by northerly flow, divergence (shading), subsidence, and cold advection (crossing of isobars by the streamlines). By 0000 UTC 29 March the western anticyclone has moved to the central North Atlantic and reaches its largest north–south extension. Again a small cyclone is located on its eastern flank, and low-level divergence and cold advection are observed in the northerly flow between the two systems (Figs. 6b and 7b). The clouds close to the tip of the upper-level trough in Fig. 2a appear to be associated with this low. There is at best a marginal westward tilt between the low-level features and the upper-level PV ridge and trough (Fig. 1d). At 0000 UTC 31 March the low-level anticyclone develops a SW–NE tilt and thereby cuts off the formerly strong cold advection to the west of the weakening low (Fig. 6c). By that time the cloud band associated with this system has dissipated over the Atlantic (Fig. 2d). The divergent region between the cyclone and anticyclone in Fig. 6c has spread out and moved southward. Low-level baroclinicity (as measured by the slope of the 300-K surface) is only modest along the Atlantic portion of the TP (Fig. 6c). On 1 April the surface low reintensifies and moves toward the Mediterranean (see Fig. 3b in FK).

An ensemble of UW-NMS trajectories, started near 22.8°N, 24.1°W (see the black bars in Figs. 6a and 7a), illustrates the trade surge associated with the strong low-level divergence at 1200 UTC 26 March. Flowing southward on the downward sloping 305-K isentropic surface the trajectories start to diverge at around 20°N and circa 750 hPa (2.5 km; Fig. 7a). While the eastern branch turns cyclonically around the trough, the other trajectories turn anticyclonically or follow the northeasterly trade winds over the Atlantic Ocean. The latter is in agreement with the production of negative relative vorticity in a diverging shrinking atmospheric column (assuming PV conservation). After 0000 UTC 29 March the trade wind trajectories subside to below ∼950 hPa over the tropical Atlantic (5°–9°N). A decrease in Θ to around 300 K and a strong increase of Θe are indications for cooling and moistening of the adiabatically warmed and dried parcels over the ocean surface. The fact that the parcels finally approach the region of enhanced near-surface convergence shown in Figs. 6b and 6c suggests a contribution of this trade surge to the formation of the convection whose outflow feeds into the TP (Figs. 1d–f and 2a–d; see section 6).

An ensemble of trajectories started from the divergent region around 27°N, 36°W (see Fig. 6b) at 0000 UTC 29 March shows the same bifurcation into a cyclonic and a trade wind branch at about 20°N. Entering the model domain from the north with Θ values around 305 K, the trajectories subside from around 390 hPa (∼7.5 km) to 640 hPa between 0000 UTC 28 and 0000 UTC 29 March (Fig. 7b) in accordance with midtropospheric descending motions of up to 0.1 m s−1. On 30 March the trade trajectories penetrate into the lowest kilometer of the atmosphere, again accompanied by surface energy fluxes from the ocean. It is not before the end of 31 March that these trade parcels enter the deep Tropics and contribute to an enhanced near-surface convergence in the ITCZ. The resulting increase of convection and upper-tropospheric outflow occurs too late for an impact on the TP evolution.

6. Large-scale tropical–extratropical interactions

Hoskins et al. (1985, p. 936) promoted the use of maps of PV on isentropic surfaces for the analysis of tropical–extratropical interactions. Horizontal gradients of this quantity allow inferring the inertial stability of the flow and its influence on the evolution at upper levels. Divergent winds have commonly been employed to determine the influence of outflow from tropical convection on the large-scale flow (e.g., Sardeshmukh and Hoskins 1988). These two parameters will be used here to address more far-reaching questions related to the previous analysis.

a. Convergent forcing over the Caribbean Sea

In section 4a it was demonstrated that parcels from the anticyclonic shear side of the STJ over North America are forced to enter the deep Tropics over South America, before they finally accelerate into the STJ over the Atlantic Ocean. This subsection is meant to further examine the nature of this forcing. At 0000 UTC 26 March, 12 h before trajectory 1 in Fig. 5a enters the model domain from the west, a strong anticyclonically curved STJ streak is observed at the crest of a large PV ridge across southern North America (Fig. 1a). As suggested by satellite images, this jet is related to a subtropical cloud band over the eastern Pacific that formed during the previous days (not shown). Two effects contribute to the pronounced convergence centered over the Caribbean Sea: 1) Convection over equatorial western South America and the adjacent Pacific Ocean directs its divergent outflow into the region of low inertial stability, indicated by small positive values of PV. 2) The deceleration and the strong curvature in the right exit region of the jet (Fig. 1a) induce convergent ageostrophic motions associated with moderate cold advection at upper levels (300–150 hPa) over the Caribbean Sea and around Florida (not shown). The lack of cross-contour flow at 300 K (Fig. 6a) suggests that the cold advection is restricted to the upper troposphere. A moderate lowering of geopotential heights to the north of Venezuela was observed during the following days, which increased the gradient toward the convectively generated high over South America and thereby helped to accelerate tropical outflow into the TP genesis region. It is exactly during the passage around the northeastern side of this convergence center on 27 March where trajectory 1 in Fig. 5a begins to become highly ageostrophic and to decelerate into the deep Tropics.

Between 26 and 28 March a cutoff positive PV feature off the California coast induces a subtropical cloud band (see the low brightness temperatures in Fig. 8a), strong divergent winds, a downstream STJ streak at the crest of the ridge over the United States (the acceleration to the south of the PV trough and the deceleration near the jet exit are clearly marked by 345-K flow crossing the contours of Montgomery potential in Fig. 8c), and finally a trough to the east of the West Indies (Figs. 1a–c). These features are very similar to the ones connected to the previous cloud band over the eastern Pacific described in the preceding paragraph. Again strong convergence is found in the jet exit region over the Caribbean Sea on 28 and 29 March (Fig. 1c and 1d). Cold advection was moderate at upper levels over the Caribbean Sea, the Gulf of Mexico, and Florida (not shown; in addition the cross-contour flow in Fig. 6b indicates weak lower-level cold advection).

The WV image for 0000 UTC 28 March shows extremely high brightness temperatures near the maximum velocity potential χ associated with the convergence (Fig. 8a), indicating subsidence and drying. A SW–NE vertical cross section through the χ maximum depicts the locally enhanced Hadley overturning in this region (Fig. 8b). Lower-level convergence, rising, and strong upper-level divergence is found near the equator in correspondence with the widespread divergent outflow over the equatorial eastern Pacific and South America during 28–30 March (Figs. 1c– e; see also Fig. 8a). It is possible that this is in part a lagged response to the enhancement in near-surface convergence caused by the weak trade surge (see Fig. 6a) associated with the convergence maximum on 26 March (Fig. 1a). The gray shading in Fig. 8c shows first evidence of a beginning acceleration of the STJ over the Atlantic through flow crossing contours of Montgomery potential.

b. Development over the Atlantic Ocean

In section 5 a possible relation between surges in the trade winds over the Atlantic Ocean and enhanced tropical convection was discussed. Here the focus of the analysis is on the upper-level forcing for and the response to these trade surges. The two together form an important part of the tropical–extratropical interactions involved in the TP genesis. At 0000 UTC 26 March, 12 h before the beginning of the first trade surge (Fig. 7a), the PV distribution at 345 K over the Atlantic Ocean shows remnants of a weak subtropical ridge–trough pattern to the west of Africa. Moderate upper-level convergence is found to the west of the trough (Fig. 1a), above the region of strongly descending trajectories (Fig. 7a). Contributions from tropical convection to the convergence are weak, since the large PV values to the southwest of the trough inhibit further northward penetration of divergent outflow. Instead, strong cold advection is found at mid- and upper levels to the west of the trough axis (not shown). Somewhat weaker low-level cold advection is found around the southern edge of the trough (see the cross-contour flow in Fig. 6a).

The resulting enhanced trade winds over the tropical Atlantic Ocean during the following days (Fig. 6b) are likely to increase the convective transport of easterly momentum, which, according to the IAKE concept of Mecikalski and Tripoli (1998, 2003), supports a poleward ventilation of convective outflow. In addition to that, the equatorward amplification of the trough to the north of Venezuela discussed above (Figs. 1a and 1b) initiates a northeastward advection of weakly positive (or even negative) PV air on its eastern side. The resulting PV ridge over the Atlantic Ocean provides a region of low inertial stability, additionally favoring the ventilation of convective outflow from South America and the adjacent Atlantic in this direction (Figs. 1b– e). This is most distinct in the regions of negative PV in the NH over the Atlantic Ocean.

At 0000 UTC 29 March, when the second trade surge was initiated, pronounced upper-level convergence is found between the PV ridge and the downstream trough over the Atlantic Ocean (Fig. 1d), again right above the region of quickly descending trajectories (Fig. 7b). Figure 9a shows that the associated 330-K χ maximum at 1200 UTC 29 March is just upstream of the region of the highest WV brightness temperatures, indicative of midlevel dryness (note that a different grayscale was used than in Fig. 8a). A dryline along the western side of the PV trough indicates a stratospheric intrusion. The two dry features are also evident in the IR image for the same time (Fig. 2a). Figure 9b displays a vertical cross section through the upper-level convergence maximum at 0000 UTC 29 March. The strong flow down the sloping isentropes around 30°N reveals troposphere-wide cold advection (see also Fig. 6b). Underneath the upper-level convergence strong midlevel subsidence, lower-level divergence, and enhanced trade winds are found. The active tropical convection (most likely connected to the prior trade surge; see section 5) is reflected by strong low-level convergence, rising, and upper-level divergence. The subtropical branch of the enhanced Hadley cell is marked by weak upper-level convergence, sinking, and low-level divergence (see also Fig. 6b), further supporting the strong trade flow over the tropical Atlantic. Confluence is observed between the tropical outflow and the extratropical flow near the maximum in wind speed normal to the cross section at about 20°N and 200 hPa (∼345 K), which corresponds to the jet in Fig. 1d. Figure 9c shows that it is within this region where the kinetic energy for the acceleration of the STJ is produced through flow crossing the contours of Montgomery potential at 345 K.

7. Summary

The previous analysis of the dynamics involved in the formation of a vast NH Atlantic TP in March 2002 demonstrated that complicated two-way interactions between tropical convection and the extratropical circulation are responsible for the formation of the TP clouds and the acceleration of the accompanying strong STJ streak. The investigations were based on satellite imagery, ECMWF analyses, and high-resolution trajectories computed from output of a numerical simulation performed using the UW-NMS. A comparison of model-generated clouds and winds with observations showed satisfying agreement. Figure 10 shows a schematic depiction of the most important features involved in the TP evolution. Two major airstreams can be identified that eventually form the TP cloud band (thick gray arrows in Fig. 10a):

  • (i) The poleward portion of the TP originates from the anticyclonic shear side of the STJ surrounding the crest of the upstream PV ridge over North America (Fig. 10a). Trajectories passing to the northeast of the strong convergence in the eastern portion of the ridge are forced to enter the deep Tropics, lose their rotational balance, strongly decelerate, and finally accelerate down the gradient of geopotential back toward the subtropics in association with strong ageostrophic motions. Two factors are important for generating the convergent forcing: 1) Convection within a subtropical cloud band over the Pacific Ocean accelerates the downstream STJ streak (Fig. 10a) and creates a strongly curved jet exit region. 2) The low inertial stability in the ridge (hatching in Fig. 10a) favors a north- or northeastward ventilation of tropical convective outflow. Since the PV ridge is quasi-stationary, the tropical convection is given enough time to establish a locally enhanced Hadley overturning and to maintain the convergent forcing for days (see the weak trade surge in Fig. 10b).

  • (ii) The equatorward portion of the TP originates from outflow of near-equatorial convection over South America and the adjacent Atlantic Ocean (Fig. 10a). Parcels within this airstream rise in convective updrafts and flow slowly northward into the entrance region of the emerging STJ over the Atlantic Ocean. The northeastward acceleration is promoted by enhanced gradients in geopotential toward the trough downstream of the ridge over North America and by low inertial stability caused by large-scale advection of low-PV air to the east of this trough (Fig. 10a). Repeated strong baroclinic developments in the extratropics produce high-amplitude waves over the North Atlantic (see KM), associated with strong cold advection, upper-level convergence, subsidence, and trade surges (Figs. 10a and 10b). The resulting enhanced near-equatorial easterlies are likely to cause increased convective transports of easterly momentum, which would further promote a northward ventilation of convective outflow (MT98). Again the relatively stationary flow at low latitudes and the repeated forcing from the extratropics allow a response of the local Hadley overturning (Fig. 10a).

Figure 10c depicts the mature phase of the TP evolution. North of about 10°N parcels from both contributing airstreams show a similar behavior. The parcels accelerate away from the Tropics and enter the baroclinic zone on the southeastern flank of a pronounced positively tilted subtropical trough to the west of West Africa resulting from a Rossby wave breaking in the extratropics. The strong pressure gradient in this region leads to even stronger acceleration and eventual inertial instability. Gradient wind balance is only (re-) gained close to the jet maximum around the ridge crest over North Africa (Fig. 10c). The strong, highly ageostrophic, prolonged northeastward acceleration is associated with an upslope flow on isentropic surfaces, cooling, and the formation of cirrus clouds along the tropospheric side of the jet axis at upper levels (7–14 km). Simple advection of cirrus from tropical convection appears to be of minor importance for this particular TP. Over Africa heavy precipitation is generated underneath the TP (Fig. 10c; see KM).

8. Discussion and conclusions

The presented characteristics of TP genesis reveal basic agreement with a number of previous studies. The behavior of the subtropical airflow contributing to the TP gives observational support to the results of Blackwell (2000) who was able to stimulate TP-like circulations through prescribed convergent forcing in a dry barotropic model. Similarly to the present case, he observed a penetration of extratropical flow to low latitudes and an equatorward amplification of the trough downstream of the forcing. Blackwell (2000) assumed that large-scale cold advection from the extratropics causes the convergent forcing, but the present case suggests an additional contribution from the subtropical branch of the locally enhanced Hadley overturning. As previously observed, this study also finds the convergent forcing region characterized by a conspicuous dark (i.e., dry) area in WV imagery (McGuirk and Ulsh 1990; Blackwell and McGuirk 1996). The behavior of the tropical airflow contributing to the TP shows agreement with the IAKE concept proposed by MT98 (see section 1). Advection of low PV by the large-scale wave in the subtropics and enhanced convective transports of easterly momentum favor a northward ventilation of convective outflow. It should, however, be pointed out that the representation of convective-scale processes in the model does not allow a detailed investigation of this effect as in Mecikalski and Tripoli (2003). Nevertheless the presented study demonstrates that the “dry dynamics” results of Blackwell (2000) and the convection-based concept of MT98 (see introduction) portray two different views on the same phenomenon that can be integrated into one more complete picture.

One important result of this study is the predominantly dynamical cloud generation (as opposed to advection) that explains the occasionally observed discrepancy between TP extension speed and wind velocity (McGuirk et al. 1987, 1988; MT98). Dynamical reasons are also responsible for the subsidence and the associated disintegration of the TP clouds to the east of the wind speed maximum giving rise to the typical anticyclonic curvature of the TP in the subtropics noticed by several studies. Quasigeostrophic forcing, which has been postulated to be responsible for TP genesis in various studies (e.g., McGuirk et al. 1988), does not seem to be of major importance along the tropical end of the TP in agreement with the argument of diminishing penetration depth of the influence of an upper-level PV anomaly in the Tropics (chapter 3 in Hoskins et al. 1985; Kiladis 1998; MT98). Over northwest Africa, however, QG forcing is one contributing factor to the generation of heavy precipitation under the northernmost portion of the TP, as shown by KM (see also FK). The strong cold advection, subsidence, upper-level convergence, and trade surges documented in this paper show agreement with the work on Pacific TPs by, for example, McGuirk et al. (1988), KW1992b, and Kiladis (1998).

Several previous studies demonstrated a statistical relationship between the penetration of an extratropical Rossby wave train to low latitudes and tropical cloudiness (Liebmann and Hartmann 1984; KW1992a; KW1997; Kiladis 1998). Knippertz and Martin (2005) showed that the subtropical trough to the west of the mature TP examined here is in fact related to an upstream wave. They demonstrate that the diabatic destruction of upper-level PV associated with prolonged frontogenesis over the eastern United States and convection ahead of the PV trough close to the West Indies (see Fig. 10a) is crucial for the amplification and the breaking of the PV wave over the Atlantic Ocean, and therefore for the subsequent formation of the sharp positively tilted downstream trough in the subtropics (Fig. 10c; cf. LC1 life cycle in Thorncroft et al. 1993). However, the idea of a linear response of tropical cloudiness to nondivergent barotropic Rossby wave forcing, although a statistically significant result in filtered or averaged data (KW1992a), does not account for all the interactions between the Tropics and the extratropics responsible for TP generation identified in the present study. Important feedbacks such as the lagged response to trade surges and modifications of upper-level outflow conditions are not considered. In particular, the different response times and directions in the Tropics and extratropics as well as at upper and at lower levels require a more complex analysis of the various interactions. This is further supported by the argument that Rossby waves only propagate in westerlies and therefore do not penetrate to lower levels in the Tropics (Tomas and Webster 1994). Moreover the high wave amplitudes, and the observed strong zonal and meridional gradients render a view in terms of basic flow and superimposed wavelike disturbances difficult. Another critical point is the common use of OLR, which does not allow a distinction between convection and cirrus, a point addressed here by considering the depth of model-generated clouds (Fig. 4).

The present study unveils possible factors why the east and central Pacific, and the Atlantic Ocean, are the major TP regions in the NH (Iskenderian 1995): First, the cool-season quasi-permanent baroclinic zones along the continents to the northwest generate wave disturbances that eventually penetrate into the Tropics at the eastern ends of the basins. Second, the maxima of tropical convection over the western Pacific and the Amazon basin, respectively, force upper-level ridges that occasionally produce situations of convergent forcing. An important difference between the two basins with respect to precipitation under the extratropical end of TPs might be the deeper layer of tropical moisture over continental Africa during the transition seasons (see also Knippertz et al. 2003). Finally it should be pointed out that not all TPs show the strong east–west orientation, slow evolution, predominantly dynamic cloud generation, and clear distinction between an upper-level subtropical ridge–trough pattern (here over Central America) and its vertically deeper extratropical pendant (here over the Atlantic Ocean) found for the present TP (e.g., KM). More cases should therefore be examined to assess the relative importance of single influence factors and to better understand seasonal variations of TP occurrence (see the final discussion in KM). Given that ENSO considerably modifies the westerly duct and TP activity over the eastern Pacific (McGuirk et al. 1987; McGuirk and Ulsh 1990; Iskenderian 1995; Matthews and Kiladis 1999), it would be interesting to study how ENSO-related circulation changes affect the downstream Atlantic TPs.

Acknowledgments

The scientific results presented in this paper were accomplished during a two-year postdoctoral research scholarship from the German Science Foundation (DFG; Grant KN 581/1–1) at the Department of Atmospheric and Oceanic Sciences of the University of Wisconsin—Madison under the supervision of Prof. Jonathan E. Martin and Prof. emer. Stefan L. Hastenrath. I wish to thank Dr. John R. Mecikalski for his help with running the UW-NMS model and for numerous interesting discussions on TP formation, and Pete Pokrandt for his assistance in data and software issues. I would also like to acknowledge fruitful discussions with Prof. emer. Francis P. Bretherton and Dr. Andreas H. Fink, as well as the helpful comments by Dr. George Kiladis and an anonymous reviewer.

REFERENCES

  • Blackwell, K G., 2000: Tropical plumes in a barotropic model: A product of Rossby wave generation in the tropical upper troposphere. Mon. Wea. Rev., 128 , 22882302.

    • Search Google Scholar
    • Export Citation
  • Blackwell, K G., and J P. McGuirk, 1996: Tropical upper-tropospheric dry regions from TOVS and rawinsondes. J. Appl. Meteor., 35 , 464481.

    • Search Google Scholar
    • Export Citation
  • Fink, A H., and P. Knippertz, 2003: An extreme precipitation event in southern Morocco in spring 2002 and some hydrological implications. Weather, 58 , 377387.

    • Search Google Scholar
    • Export Citation
  • Geb, M., 2000: Factors favouring precipitation in North Africa: Seen from the viewpoint of present-day climatology. Global Planet. Change, 26 , 8596.

    • Search Google Scholar
    • Export Citation
  • Hibbard, W., J. Anderson, I. Foster, B. Paul, R. Jacob, C. Schafer, and M. Tyree, 1996: Exploring coupled atmosphere–ocean models using Vis5D. Int. J. Supercomput. Appl., 10 , 211222.

    • Search Google Scholar
    • Export Citation
  • Hoskins, B J., M E. McIntyre, and A W. Robertson, 1985: On the use and significance of isentropic potential vorticity maps. Quart. J. Roy. Meteor. Soc., 111 , 877946.

    • Search Google Scholar
    • Export Citation
  • Iskenderian, H., 1995: A 10-year climatology of Northern Hemisphere tropical cloud plumes and their composite flow patterns. J. Climate, 8 , 16301637.

    • Search Google Scholar
    • Export Citation
  • Kiladis, G N., 1998: Observations of Rossby waves linked to convection over the eastern tropical Pacific. J. Atmos. Sci., 55 , 321339.

    • Search Google Scholar
    • Export Citation
  • Kiladis, G N., and K M. Weickmann, 1992a: Circulation anomalies associated with tropical convection during northern winter. Mon. Wea. Rev., 120 , 19001923.

    • Search Google Scholar
    • Export Citation
  • Kiladis, G N., and K M. Weickmann, 1992b: Extratropical forcing of tropical Pacific convection during northern winter. Mon. Wea. Rev., 120 , 19241938.

    • Search Google Scholar
    • Export Citation
  • Kiladis, G N., and K M. Weickmann, 1997: Horizontal structure and seasonality of large-scale circulations associated with submonthly tropical convection. Mon. Wea. Rev., 125 , 19972013.

    • Search Google Scholar
    • Export Citation
  • Knippertz, P., and J E. Martin, 2005: Tropical plumes and extreme precipitation in subtropical and tropical West Africa. Quart. J. Roy. Meteor. Soc., in press.

    • Search Google Scholar
    • Export Citation
  • Knippertz, P., A H. Fink, A. Reiner, and P. Speth, 2003: Three late summer/early autumn cases of tropical–extratropical interactions causing precipitation in northwest Africa. Mon. Wea. Rev., 131 , 116135.

    • Search Google Scholar
    • Export Citation
  • Kuhnel, I., 1989: Tropical–extratropical cloudband climatology based on satellite data. Int. J. Climatol., 9 , 441463.

  • Kuhnel, I., 1990: Tropical–extratropical cloudbands in the Australian region. Int. J. Climatol., 10 , 341364.

  • Liebmann, B., and D L. Hartmann, 1984: An observational study of tropical–midlatitude interaction on intraseasonal time scales during winter. J. Atmos. Sci., 41 , 33333350.

    • Search Google Scholar
    • Export Citation
  • Matthews, A J., and G N. Kiladis, 1999: Interactions between ENSO, transient circulation, and tropical convection over the Pacific. J. Climate, 12 , 30623086.

    • Search Google Scholar
    • Export Citation
  • McGuirk, J P., 1993: Impact of increased TOVS signal on the NMC global spectral model: A tropical-plume case study. Mon. Wea. Rev., 121 , 695712.

    • Search Google Scholar
    • Export Citation
  • McGuirk, J P., and D J. Ulsh, 1990: Evolution of tropical plumes in VAS water vapor imagery. Mon. Wea. Rev., 118 , 17581766.

  • McGuirk, J P., A H. Thompson, and N R. Smith, 1987: Moisture bursts over the tropical Pacific Ocean. Mon. Wea. Rev., 115 , 787798.

    • Search Google Scholar
    • Export Citation
  • McGuirk, J P., A H. Thompson, and J R. Schaefer, 1988: An eastern Pacific tropical plume. Mon. Wea. Rev., 116 , 25052521.

  • Mecikalski, J R., and G J. Tripoli, 1998: Inertial available kinetic energy and the dynamics of tropical plume formation. Mon. Wea. Rev., 126 , 22002216.

    • Search Google Scholar
    • Export Citation
  • Mecikalski, J R., and G J. Tripoli, 2003: Influence of upper-troposheric inertial stability on the convective transport of momentum. Quart. J. Roy. Meteor. Soc., 129 , 15371563.

    • Search Google Scholar
    • Export Citation
  • Sardeshmukh, P D., and B J. Hoskins, 1988: The generation of global rotational flow by steady idealized tropical divergence. J. Atmos. Sci., 45 , 12281251.

    • Search Google Scholar
    • Export Citation
  • Thepenier, R-M., and D. Cruette, 1981: Formation of cloud bands associated with the American subtropical jet stream and their interaction with midlatitude synoptic disturbances reaching Europe. Mon. Wea. Rev., 109 , 22092220.

    • Search Google Scholar
    • Export Citation
  • Thorncroft, C D., B J. Hoskins, and M E. McIntyre, 1993: Two paradigms of baroclinic-wave life-cycle behaviour. Quart. J. Roy. Meteor. Soc., 119 , 1755.

    • Search Google Scholar
    • Export Citation
  • Tomas, R A., and P J. Webster, 1994: Horizontal and vertical structure of cross-equatorial wave propagation. J. Atmos. Sci., 51 , 14171430.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K E., 1992: Global analyses from ECMWF and atlas of 1000 to 10 mb circulation statistics. NCAR Tech. Note NCAR/TN-373+STR, National Center for Atmospheric Research, Boulder, CO, 191 pp.

  • Tripoli, G J., 1992: A nonhydrostatic numerical model designed to simulate scale interaction. Mon. Wea. Rev., 120 , 13421359.

  • Webster, P J., and J R. Holton, 1982: Cross-equatorial response to middle-latitude forcing in a zonally varying basic state. J. Atmos. Sci., 39 , 722733.

    • Search Google Scholar
    • Export Citation
  • Ziv, B., 2001: A subtropical rainstorm associated with a tropical plume over Africa and the Middle-East. Theor. Appl. Climatol., 69 , 91102.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

PV (hatching/shading according to legend; units are in PVU), isotachs (contours are 30, 50, 60, and 70 m s−1), and divergent wind vectors greater than 2 m s−1 (scale in the bottom-left corner) at 345 K for 0000 UTC on (a) 26, (b) 27, (c) 28, (d) 29, (e) 30, and (f) 31 Mar 2002.

Citation: Monthly Weather Review 133, 9; 10.1175/MWR2999.1

Fig. 1.
Fig. 1.

(Continued)

Citation: Monthly Weather Review 133, 9; 10.1175/MWR2999.1

Fig. 2.
Fig. 2.

Meteosat IR satellite images with superimposed 345-K isotachs in white (contours are 30, 50, 60, and 70 m s−1) and full wind streamlines in black for (a) 1200 UTC 29 Mar, (b) 0000 UTC 30 Mar, (c) 1200 UTC 30 Mar, (d) 0000 UTC 31 Mar, (e) 1200 UTC 31 Mar, and (f) 0000 UTC 1 Apr 2002.

Citation: Monthly Weather Review 133, 9; 10.1175/MWR2999.1

Fig. 3.
Fig. 3.

Horizontal extension of the TP clouds in the UW-NMS model based on isosurfaces of LCMR = 0.1 (see text) for the analysis times used in Fig. 2: (a) 1200 UTC 29 Mar–1200 UTC 30 Mar and (b) 0000 UTC 31 Mar–0000 UTC 1 Apr 2002.

Citation: Monthly Weather Review 133, 9; 10.1175/MWR2999.1

Fig. 4.
Fig. 4.

(a) Isosurfaces of LCMR = 0.1 calculated from UW-NMS model output for 1200 UTC 30 Mar 2002. (b) West–east vertical cross section along the thick black line in (a) showing isopleths of LCMR in 0.5 increments starting with 0.1. Dashed horizontal lines depict height above sea level in km.

Citation: Monthly Weather Review 133, 9; 10.1175/MWR2999.1

Fig. 5.
Fig. 5.

(a) Tracks of two exemplary UW-NMS trajectories are shown as solid (dashed) gray lines before (after) 0000 UTC 31 Mar; 0000 and 1200 UTC trajectory positions are marked by gray filled circles; additional 0600 and 1800 UTC positions are marked by open circles; black numbers indicate the calendar day (Mar–Apr 2002) for the 0000 UTC positions; shading represents isosurfaces of LCMR = 0.1; total and ageostrophic winds are indicated by black and gray vectors, respectively, according to the scale in the lower-right corner. (b)–(i) Meteorological quantities along trajectories (left) 1 and (right) 2 from 0000 UTC 26 Mar–0000 UTC 1 Apr 2002. (b), (f) LCMR (solid), RH (dashed), and MR (dotted); (c), (g) vertical velocity (denoted w; solid) and PV (dashed); (d), (h) Θ (solid), Θe (dashed), and pressure (p; dotted); (e), (i) total (V; solid), zonal (u; dashed), and meridional (v; dotted) wind. Gray shaded areas indicate the passages through the TP clouds [based on LCMR > 0, see (b) and (g)].

Citation: Monthly Weather Review 133, 9; 10.1175/MWR2999.1

Fig. 6.
Fig. 6.

Full wind streamlines (thin black lines), divergence (shading/hatching according to legend; units are in 10−6 s–1), and isobars (every 100 hPa) at 300 K for (a) 1200 UTC 26 Mar, (b) 0000 UTC 29 Mar, and (c) 0000 UTC 31 Mar 2002. No values are plotted where the 300-K surface is underground. Black bars indicate the starting regions for the trajectories in Fig. 7.

Citation: Monthly Weather Review 133, 9; 10.1175/MWR2999.1

Fig. 7.
Fig. 7.

(a) SLP for 1200 UTC 26 Mar 2002 (thin dotted lines; contour interval is 5 hPa; highs and lows are indicated) and UW-NMS model trajectories (thick solid lines). The trajectories were started at 3-km height, 22.8°N, and between 21.1° and 27.1°W (black bars) at 1200 UTC 26 Mar 2002; trajectory paths are shown for the period from 0000 UTC 26 to 1200 UTC 31 Mar. (b) Same as (a), but for 0000 UTC 29 Mar, trajectories started at 3.5 km, 27°N, and between 33.8° and 38.8°W, and the time period from 0000 UTC 28 Mar to 0000 UTC 1 Apr. The height of the trajectories (in hPa) is indicated by gray shading.

Citation: Monthly Weather Review 133, 9; 10.1175/MWR2999.1

Fig. 8.
Fig. 8.

(a) GOES WV satellite image for 0000 UTC 28 Mar 2002 with superimposed 345-K full wind streamlines in black and velocity potential in white (contour interval is 2 × 106 m2 s−1). (b) Vertical cross section along the line BB–AA shown in (c) for the same date. Displayed quantities are Θ (solid lines in 10-K increments), wind speed normal to cross section (dashed lines in 10 m s−1 increments), divergence (hatching/shading according to legend, units are 10−6 s−1), and tangential winds/vertical velocity omega as vectors. The scale for the horizontal wind is in the upper-left corner; maximum ascent is 0.34 Pa s−1. (c) The 345-K Montgomery potential M (black lines every 8 m2 s−2) and generation of kinetic energy through “cross-contour flow” −υ · ∇M (hatching/shading according to legend, units are 10−2 W kg−1) for the same date.

Citation: Monthly Weather Review 133, 9; 10.1175/MWR2999.1

Fig. 9.
Fig. 9.

(a) Meteosat WV satellite image for 1200 UTC 29 Mar 2002 with superimposed 330-K full wind streamlines in black and velocity potential in white (contour interval is 2 × 106 m2 s−1). (b) Vertical cross section along the line DD–CC shown in (c) for 0000 UTC on the same day. Displayed quantities are Θ (solid lines in 10-K increments), wind speed normal to cross section (dashed lines in 10 m s−1 increments), divergence (hatching/shading according to legend, units are 10−6 s−1), and tangential winds/vertical velocity omega as vectors. The scale for the horizontal wind is in the upper-left corner; maximum ascent is 0.54 Pa s−1. (c) The 345-K Montgomery potential M (black lines every 8 m2 s−2) and generation of kinetic energy through “cross-contour flow” −υ · ∇M (hatching/shading according to legend, units are 10−2 W kg−1) for the same date as (b).

Citation: Monthly Weather Review 133, 9; 10.1175/MWR2999.1

Fig. 10.
Fig. 10.

Schematic depiction of TP formation: (a) Upper-level (∼345 K) and (b) lower-level (∼300 K) flow prior to TP genesis; (c) as in (a) but for the mature TP (about 2.5 days later). Thin black arrows show selected streamlines of the divergent (full) wind at upper (lower) levels; short black arrows depict upper-level jet maxima; long, thick black arrows in (c) depict the trajectories from Fig. 5a; gray shaded areas indicate the major regions of high clouds as inferred from IR satellite images; C and D mark centers of con-/divergence, H and L mark SLP extrema; hatching indicates regions of low inertial stability (small PV values away from the equator) in (a) and cold advection in (b); the 1- and 4-PVU contours at 345 K and the height of the 300-K isentropic surface (in 100-hPa increments) are displayed as thick solid and dashed lines, respectively. See text for further explanations.

Citation: Monthly Weather Review 133, 9; 10.1175/MWR2999.1

Save
  • Blackwell, K G., 2000: Tropical plumes in a barotropic model: A product of Rossby wave generation in the tropical upper troposphere. Mon. Wea. Rev., 128 , 22882302.

    • Search Google Scholar
    • Export Citation
  • Blackwell, K G., and J P. McGuirk, 1996: Tropical upper-tropospheric dry regions from TOVS and rawinsondes. J. Appl. Meteor., 35 , 464481.

    • Search Google Scholar
    • Export Citation
  • Fink, A H., and P. Knippertz, 2003: An extreme precipitation event in southern Morocco in spring 2002 and some hydrological implications. Weather, 58 , 377387.

    • Search Google Scholar
    • Export Citation
  • Geb, M., 2000: Factors favouring precipitation in North Africa: Seen from the viewpoint of present-day climatology. Global Planet. Change, 26 , 8596.

    • Search Google Scholar
    • Export Citation
  • Hibbard, W., J. Anderson, I. Foster, B. Paul, R. Jacob, C. Schafer, and M. Tyree, 1996: Exploring coupled atmosphere–ocean models using Vis5D. Int. J. Supercomput. Appl., 10 , 211222.

    • Search Google Scholar
    • Export Citation
  • Hoskins, B J., M E. McIntyre, and A W. Robertson, 1985: On the use and significance of isentropic potential vorticity maps. Quart. J. Roy. Meteor. Soc., 111 , 877946.

    • Search Google Scholar
    • Export Citation
  • Iskenderian, H., 1995: A 10-year climatology of Northern Hemisphere tropical cloud plumes and their composite flow patterns. J. Climate, 8 , 16301637.

    • Search Google Scholar
    • Export Citation
  • Kiladis, G N., 1998: Observations of Rossby waves linked to convection over the eastern tropical Pacific. J. Atmos. Sci., 55 , 321339.

    • Search Google Scholar
    • Export Citation
  • Kiladis, G N., and K M. Weickmann, 1992a: Circulation anomalies associated with tropical convection during northern winter. Mon. Wea. Rev., 120 , 19001923.

    • Search Google Scholar
    • Export Citation
  • Kiladis, G N., and K M. Weickmann, 1992b: Extratropical forcing of tropical Pacific convection during northern winter. Mon. Wea. Rev., 120 , 19241938.

    • Search Google Scholar
    • Export Citation
  • Kiladis, G N., and K M. Weickmann, 1997: Horizontal structure and seasonality of large-scale circulations associated with submonthly tropical convection. Mon. Wea. Rev., 125 , 19972013.

    • Search Google Scholar
    • Export Citation
  • Knippertz, P., and J E. Martin, 2005: Tropical plumes and extreme precipitation in subtropical and tropical West Africa. Quart. J. Roy. Meteor. Soc., in press.

    • Search Google Scholar
    • Export Citation
  • Knippertz, P., A H. Fink, A. Reiner, and P. Speth, 2003: Three late summer/early autumn cases of tropical–extratropical interactions causing precipitation in northwest Africa. Mon. Wea. Rev., 131 , 116135.

    • Search Google Scholar
    • Export Citation
  • Kuhnel, I., 1989: Tropical–extratropical cloudband climatology based on satellite data. Int. J. Climatol., 9 , 441463.

  • Kuhnel, I., 1990: Tropical–extratropical cloudbands in the Australian region. Int. J. Climatol., 10 , 341364.

  • Liebmann, B., and D L. Hartmann, 1984: An observational study of tropical–midlatitude interaction on intraseasonal time scales during winter. J. Atmos. Sci., 41 , 33333350.

    • Search Google Scholar
    • Export Citation
  • Matthews, A J., and G N. Kiladis, 1999: Interactions between ENSO, transient circulation, and tropical convection over the Pacific. J. Climate, 12 , 30623086.

    • Search Google Scholar
    • Export Citation
  • McGuirk, J P., 1993: Impact of increased TOVS signal on the NMC global spectral model: A tropical-plume case study. Mon. Wea. Rev., 121 , 695712.

    • Search Google Scholar
    • Export Citation
  • McGuirk, J P., and D J. Ulsh, 1990: Evolution of tropical plumes in VAS water vapor imagery. Mon. Wea. Rev., 118 , 17581766.

  • McGuirk, J P., A H. Thompson, and N R. Smith, 1987: Moisture bursts over the tropical Pacific Ocean. Mon. Wea. Rev., 115 , 787798.

    • Search Google Scholar
    • Export Citation
  • McGuirk, J P., A H. Thompson, and J R. Schaefer, 1988: An eastern Pacific tropical plume. Mon. Wea. Rev., 116 , 25052521.

  • Mecikalski, J R., and G J. Tripoli, 1998: Inertial available kinetic energy and the dynamics of tropical plume formation. Mon. Wea. Rev., 126 , 22002216.

    • Search Google Scholar
    • Export Citation
  • Mecikalski, J R., and G J. Tripoli, 2003: Influence of upper-troposheric inertial stability on the convective transport of momentum. Quart. J. Roy. Meteor. Soc., 129 , 15371563.

    • Search Google Scholar
    • Export Citation
  • Sardeshmukh, P D., and B J. Hoskins, 1988: The generation of global rotational flow by steady idealized tropical divergence. J. Atmos. Sci., 45 , 12281251.

    • Search Google Scholar
    • Export Citation
  • Thepenier, R-M., and D. Cruette, 1981: Formation of cloud bands associated with the American subtropical jet stream and their interaction with midlatitude synoptic disturbances reaching Europe. Mon. Wea. Rev., 109 , 22092220.

    • Search Google Scholar
    • Export Citation
  • Thorncroft, C D., B J. Hoskins, and M E. McIntyre, 1993: Two paradigms of baroclinic-wave life-cycle behaviour. Quart. J. Roy. Meteor. Soc., 119 , 1755.

    • Search Google Scholar
    • Export Citation
  • Tomas, R A., and P J. Webster, 1994: Horizontal and vertical structure of cross-equatorial wave propagation. J. Atmos. Sci., 51 , 14171430.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K E., 1992: Global analyses from ECMWF and atlas of 1000 to 10 mb circulation statistics. NCAR Tech. Note NCAR/TN-373+STR, National Center for Atmospheric Research, Boulder, CO, 191 pp.

  • Tripoli, G J., 1992: A nonhydrostatic numerical model designed to simulate scale interaction. Mon. Wea. Rev., 120 , 13421359.

  • Webster, P J., and J R. Holton, 1982: Cross-equatorial response to middle-latitude forcing in a zonally varying basic state. J. Atmos. Sci., 39 , 722733.

    • Search Google Scholar
    • Export Citation
  • Ziv, B., 2001: A subtropical rainstorm associated with a tropical plume over Africa and the Middle-East. Theor. Appl. Climatol., 69 , 91102.

    • Search Google Scholar
    • Export Citation
  • Fig. 1.

    PV (hatching/shading according to legend; units are in PVU), isotachs (contours are 30, 50, 60, and 70 m s−1), and divergent wind vectors greater than 2 m s−1 (scale in the bottom-left corner) at 345 K for 0000 UTC on (a) 26, (b) 27, (c) 28, (d) 29, (e) 30, and (f) 31 Mar 2002.

  • Fig. 1.

    (Continued)

  • Fig. 2.

    Meteosat IR satellite images with superimposed 345-K isotachs in white (contours are 30, 50, 60, and 70 m s−1) and full wind streamlines in black for (a) 1200 UTC 29 Mar, (b) 0000 UTC 30 Mar, (c) 1200 UTC 30 Mar, (d) 0000 UTC 31 Mar, (e) 1200 UTC 31 Mar, and (f) 0000 UTC 1 Apr 2002.

  • Fig. 3.

    Horizontal extension of the TP clouds in the UW-NMS model based on isosurfaces of LCMR = 0.1 (see text) for the analysis times used in Fig. 2: (a) 1200 UTC 29 Mar–1200 UTC 30 Mar and (b) 0000 UTC 31 Mar–0000 UTC 1 Apr 2002.

  • Fig. 4.

    (a) Isosurfaces of LCMR = 0.1 calculated from UW-NMS model output for 1200 UTC 30 Mar 2002. (b) West–east vertical cross section along the thick black line in (a) showing isopleths of LCMR in 0.5 increments starting with 0.1. Dashed horizontal lines depict height above sea level in km.

  • Fig. 5.

    (a) Tracks of two exemplary UW-NMS trajectories are shown as solid (dashed) gray lines before (after) 0000 UTC 31 Mar; 0000 and 1200 UTC trajectory positions are marked by gray filled circles; additional 0600 and 1800 UTC positions are marked by open circles; black numbers indicate the calendar day (Mar–Apr 2002) for the 0000 UTC positions; shading represents isosurfaces of LCMR = 0.1; total and ageostrophic winds are indicated by black and gray vectors, respectively, according to the scale in the lower-right corner. (b)–(i) Meteorological quantities along trajectories (left) 1 and (right) 2 from 0000 UTC 26 Mar–0000 UTC 1 Apr 2002. (b), (f) LCMR (solid), RH (dashed), and MR (dotted); (c), (g) vertical velocity (denoted w; solid) and PV (dashed); (d), (h) Θ (solid), Θe (dashed), and pressure (p; dotted); (e), (i) total (V; solid), zonal (u; dashed), and meridional (v; dotted) wind. Gray shaded areas indicate the passages through the TP clouds [based on LCMR > 0, see (b) and (g)].

  • Fig. 6.

    Full wind streamlines (thin black lines), divergence (shading/hatching according to legend; units are in 10−6 s–1), and isobars (every 100 hPa) at 300 K for (a) 1200 UTC 26 Mar, (b) 0000 UTC 29 Mar, and (c) 0000 UTC 31 Mar 2002. No values are plotted where the 300-K surface is underground. Black bars indicate the starting regions for the trajectories in Fig. 7.

  • Fig. 7.

    (a) SLP for 1200 UTC 26 Mar 2002 (thin dotted lines; contour interval is 5 hPa; highs and lows are indicated) and UW-NMS model trajectories (thick solid lines). The trajectories were started at 3-km height, 22.8°N, and between 21.1° and 27.1°W (black bars) at 1200 UTC 26 Mar 2002; trajectory paths are shown for the period from 0000 UTC 26 to 1200 UTC 31 Mar. (b) Same as (a), but for 0000 UTC 29 Mar, trajectories started at 3.5 km, 27°N, and between 33.8° and 38.8°W, and the time period from 0000 UTC 28 Mar to 0000 UTC 1 Apr. The height of the trajectories (in hPa) is indicated by gray shading.

  • Fig. 8.

    (a) GOES WV satellite image for 0000 UTC 28 Mar 2002 with superimposed 345-K full wind streamlines in black and velocity potential in white (contour interval is 2 × 106 m2 s−1). (b) Vertical cross section along the line BB–AA shown in (c) for the same date. Displayed quantities are Θ (solid lines in 10-K increments), wind speed normal to cross section (dashed lines in 10 m s−1 increments), divergence (hatching/shading according to legend, units are 10−6 s−1), and tangential winds/vertical velocity omega as vectors. The scale for the horizontal wind is in the upper-left corner; maximum ascent is 0.34 Pa s−1. (c) The 345-K Montgomery potential M (black lines every 8 m2 s−2) and generation of kinetic energy through “cross-contour flow” −υ · ∇M (hatching/shading according to legend, units are 10−2 W kg−1) for the same date.

  • Fig. 9.

    (a) Meteosat WV satellite image for 1200 UTC 29 Mar 2002 with superimposed 330-K full wind streamlines in black and velocity potential in white (contour interval is 2 × 106 m2 s−1). (b) Vertical cross section along the line DD–CC shown in (c) for 0000 UTC on the same day. Displayed quantities are Θ (solid lines in 10-K increments), wind speed normal to cross section (dashed lines in 10 m s−1 increments), divergence (hatching/shading according to legend, units are 10−6 s−1), and tangential winds/vertical velocity omega as vectors. The scale for the horizontal wind is in the upper-left corner; maximum ascent is 0.54 Pa s−1. (c) The 345-K Montgomery potential M (black lines every 8 m2 s−2) and generation of kinetic energy through “cross-contour flow” −υ · ∇M (hatching/shading according to legend, units are 10−2 W kg−1) for the same date as (b).

  • Fig. 10.

    Schematic depiction of TP formation: (a) Upper-level (∼345 K) and (b) lower-level (∼300 K) flow prior to TP genesis; (c) as in (a) but for the mature TP (about 2.5 days later). Thin black arrows show selected streamlines of the divergent (full) wind at upper (lower) levels; short black arrows depict upper-level jet maxima; long, thick black arrows in (c) depict the trajectories from Fig. 5a; gray shaded areas indicate the major regions of high clouds as inferred from IR satellite images; C and D mark centers of con-/divergence, H and L mark SLP extrema; hatching indicates regions of low inertial stability (small PV values away from the equator) in (a) and cold advection in (b); the 1- and 4-PVU contours at 345 K and the height of the 300-K isentropic surface (in 100-hPa increments) are displayed as thick solid and dashed lines, respectively. See text for further explanations.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 1358 858 309
PDF Downloads 274 44 4