The Threefold Structure of the Atmosphere and the Characteristics of the Tropopause

By FR. DEFANT and H. TABA, International Meteorological Institute in Stockholm

(Manuscript received March zs, 1957)

Abstract

All available soundings for January I, 1956 have been studied with respect to their tropo- spheric and stratospheric temperature structure and with emphasis on the height and form of the tropopause. It was found that the latitudinal influence on the total vertical structure of the soundings was not dominating, but rather the location of soundings with respect to the two main westwind maxima (polarfront jet and subtropical jet) of the westerlies seemed to be important. A selection of soundings according to this principle (north of polarfront jet, between the two jets, south of subtropical jet) made it possible to distinguish between different types of soundings. This fundamental difference is discussed and will form the basis for future studies of world wide circulation changes. Furthermore the distribution of tropopause height and temper- ature has been mapped and studied on a hemispheric scale. Each of the above mentioned types possesses a typical vertical temperature structure and a characteristic tropopause heigth. Two rapid changes in tropopause height (tropopause breaklines) connected with the main westwind belts separate areas possessing an almost uniform tropopause level. Inside each of these areas a remarkable quasi-barotropic state of the upper tropospheric layers is shown. Similar discussion has been devoted to the nature of soundings inside or close to the jets or the breaks in tropopause height. Here the soundings have a more complicated vertical structure and their interpretation becomes difficult.

Introduction There are certain important features in the up er atmosphere (upper part of the Until the present time synoptic and theore- troposp ere and the stratosphere), which tical work in meteorology has been more or should Kbe closely studied. We find at these less restricted to the study of atmospheric levels the polarfront jet connected with the conditions up to the 300 mb surface (about polarfront and the subtropical jet located 9 km). This is especially true for the daily mainly at about 30° lat. and at a height of routine synoptic work. Although numerous about 200 mb. These two jets are the main studies have included the stratospheric condi- windmaxima of the broad westerlies. While tions, these studies are mainly restricted to the polarfront jet has been studied rather certain areas. Worldwide hemis heric in- thoroughly and its hemispheric existence has vestigations for stratospheric levef s are ex- been shown, the subtropical jet has not been tremely rare in meteorolo y. This is partly due investigated enough. It is not shown et, to the fact that colpplete f emispheric data for that the subtropical jet is a continuous Belt levels above 300 mb are not available, and around the hemisphere, however, climatolo- also because the contribution of stratospheric gical studies and certain vertical cross-sections develo ments to that at lower levels has not in different meridional planes suggest its more generaP ly been considered very important. or less continuous existence around the earth. Tellus IX (1957). 3 260 FR. DEFANT AND H. TABA

Fig. I. 200 mb contour map of January I, 1956 (0300 2) (Thick contours for every IOO dyn. m., thin contours for every 20 dyn. m.; Temperatures in C" and minus sign omitted. Windspeed in kts.)

Another very important feature of the higher plotted and studied thoroughly. This period atmosphere is the variation of the tropopause was selected as hemispheric maps from the height. The te erature field in the strato- surface up to the 300 mb level showed certain sphere has not eger been studied enough. periodic and pronounced changes of the world- The pury of this paper is to investigate wide patterns. This paper deals only with the behawour of the atmosphere from the January I, 1956. During the yparacon of tro opaw bvel upward. Special emphasis the material, it was found that t e Russian and wdbe to the variation of the tropo- Chinese radiosondes did not extend above 300 pause wtin vertical and in latituhal mb, so that no, or only little information was directiov. This is being done for a period available in these areas. The analysis covers beg- I January 1956. All available sound- therefore the Pacific Ocean, North America, ings the northern hemisphere have been the Atlantic Ocean and Europe. Tellur IX (1957). 3 THE THREEFOLD STRUCTURE OF THE ATMOSPHERE 261

I. General synoptic situation on January I, 1956 indicating that the tropopause intersects the 200 mb level and slopes upward to higher It is clear that the 200 mb level is most altitudes. Still further south the temperatures suitable for identifymg the subtropical jet, rise slightly and refer now to the higher as the hi hest wind s tro osphere. This can best be seen over the occurs ck ose to this At Pantic area where the subtropical jet runs can also be noticed exceptionally far north. In this area temperatures velocity remains still strong and the gradient lower than - 60" C are observed, indicating of the contour lines shows a well marked that most of the Atlantic is covered by a concentration at the 200 mb level (upper rather cold air of subtropical origin with the olarfront). A further advantage is the well tropopause above 200 mb. Here the tempe- Enown reverse of the meridional temperature ratures belong to the higher part of the gradient, so that the temperature field is troposphere. also a guide in distinguishing the two separate This discussion indicates a threefold structure wind belts. This will be shown in detail later. of the tropopause height which will be As an exampel Fig. I gives the situation on discussed in the following. January I, 1956 (0300 2). This analysis clearly shows a strong meandering circulation pattern. Particularly strong troughs can be noticed 2. The tropopause map of January I, 1956 just off the American East Coast, over Europe To study this threefold structure in tropo- and over the Japanese Islands. A weaker pause height, it was decided to prepare trough appears south of Alaska. The wind separate hemispheric tropopause maps. For values over the North American Continent this purpose about zoo soundings were together with the gradient of the contour plotted and checked for every day and for a lines show the upper polarfront jet running period of seven days. The height of the across Northern Canada and oing around tropo ause was computed for each individual the trough at the East Coast oB America. At sound? ing. For most of the soundings it was the same time, the high wind velocities over easy to select the tropopause level by applying the Southern United States and the Caribbean the standard method. (Either a noticable Islands show the subtropical jet quite well. change of tropospheric lapse rate to an iso- Over the Atlantic the polarfront jet seems to thermal layer or to an increase of temperature run from New Foundland to the southern part with height.) Sometimes it was difficult, of Greenland, to the north of Iceland and into however, to find a well defined tropopause. the middle of Europe. This was almost always true for soundings, The subtro ical jet in this area continues which met either the polarfront jet or the from the Carig bean Islands towards the north- subtropical jet close to their wind maxima. east, crosses the northern part of the Atlantic In case of soundings running through the olar- Ocean and enters England. Over France the front jet no distinct tropopause cou Pd be two jets nearly join, running southwest to found. In the case of soundings going through the African West Coast and show up again the subtropical jet, two different breaks in over North Africa. lapse rate appeared, suggesting a double Fig. I also shows the reversed temperature structure of the tropo ause. Both of these gradient at 200 mb. North of the upper cases will be discussed Pater. polarfront relatively high temperatures appear Fig. z shows the tropopause map of January on the chart (sometimes higher than - so" C). I, 1956 (0300 2). In this map the two separate This indicates that in the area surrounded by west wind belts stand out clearly in the the upper polarfront the tropopause is below radent of the contour lines of the tropo ause the 200 mb level and the temperatures belong Lvel. Two different bands picturing the 1reak to the polar stratosphere. South of the upper in the tropopause hei ht connected with the polarfront temperatures between - so and existence of the polar sront-and the subtrop - 60" C appear, where the tropopause is still. ical wind maxima can be seen all around below the 200 mb surface. Going further the hemisphere. This double break in tro o- south the temperatures become relatively low, pause separates, in a clear manner, three dii$ er- Tellus IX (1957), 3 262 FR. DEFANT AND H. TABA

Fig. 2. Hemispheric map of tropopause height (in mb) of January I, 1956 (0300 Z). Breaklines of tropopause height are shown by concentrations of isolines. Dashed lines over Asia give approximate continuation of the breaklines of tropopause height. ent areas with different characteristic tropo- In the area between the two breaklines pause heights. the tropopause height stays around zso mb. South of the subtropical break line the It is rather astonishing that the height of the height of the tropo ause is mostly around tropopause does not vary by more than so TOO mb or even higR er. Only in those parts mb inspite of the fact that the stations are where this southern breakline goes far towards located in widely differing latitudes. This can the north (in this ma over the Atlantic Ocean) best be seen over the Northern American Con- does the height of t Re tropopause drop down tinent where the two break lines are far apart. to about 17s mb, which can still be considered In the area north of the polar break line, an exceptionally high tropopause at these the tropopause heights range between 300 latitudes according to synoptic experience. and 400 mb and in certain locations even as Tellur IX (1957), 3 THE THREEFOLD STRUCTURE OF THE ATMOSPHERE 263

Fig. 3. Hemispheric map of tropopause temperature of January I, 1956 (0300 Z). (For each station tropopause height in mb (three figures) and tropopause temperatures in Co (two figures) are plotted. Minus sign for temperatures omitted.) Note the striking similarity with Fig. 2. low as 500 mb. This low tropopause hei ht gether with the location of both jets. One not only shows above cold olar vortices, fut that the tropopause map gives an can also be found above coP d polar outbreaks canimme 2 'ate three-dimensional picture of the in relatively low latitudes; see e.g. the troughs state of the atmosphere. In this connection one off the East Coast of America, to the south should pay attention to a remarkable feature of Alaska, over Northern Japan or even the of this map, namely the location of the sub- weak trough east of the Rocky Mountains. tropical breakline over the Atlantic Ocean. The tropo ause map has the advantage to Here the subtropical jet has been displaced far show in a cearP manner the distribution of towards the north. The discussion of this cold and warm tropospheric air masses, to- feature will be left for a separate paper. Tellus IX (1957). 3 264 FR. DEFANT AND H. TABA

Fig. 4. The distribution of stations selected with respect to the tropopause breaklines. (The triangles (a)= sta- tions north of the polar breakline, squares (0)= stations between the two breaklines, full circles (0) = sta- tions south of the subtropical breakline. Circles with a cross (0)= stations inside polar breaklines, circles (0) = stations inside the subtropical breakline.)

3. The tropopause temperature map of January temperatures south of the southern breakline 1, 1956 remain almost constant between -67 and - 74' C. The same is true for the middle area Fig. 3 shows a similar analysis of the tropo- between the two break lines where the tropo- pause temperature. In comparing the pattern pause temperature ranges between - 56 and of this temperature field with the pattern of - 60" C. North of the norrhern breakline the the tropopause height Presented in Fig. 2, tropopause temperatures are mostly higher we notice a marked simi arity. The breaklines than - 5o'C and in dee depressions of in tropopause height and strong concentrations tropopause height (e.g. at t K e American East in the distribution of the tropopause tempera- Coast) may be as high as - 37' C. This kind of ture agree closly in location. The tropopause temperature analysis offers a rather clear Tellus IX (19J7). 3 THE THREEFOLD STRUCTURE OF THE ATMOSPHERE 265 distinction between three very different parts d) Stations close to or in the polar break- of the northern hemisphere which are separ- line (soundings through polarfront jet) ated by the main wind belts or, what is the e) Stations close to or in the subtropical same, by the breaklines in tropopause height breakline (soundings through subtropical jet) or by strong bands with rapid change in the tropopause temperature. With the exception a) Polar group of soundings of the very northern latitudes (75 to 90" N) the close relation between tropopause height The polar soundings and their mean are and tropopause temperature seems to be shown in Fig. 5. In these soundings the tropo- almost perfectly fulfilled in the sense that pause usually is formed below the 300 mb level. an increase of height goes parallel with a The tropopause height is more changeable decrease in temperature or vice versa (com- than for the two types b and c discussed later. pensation principle). In some cases (near the center of polar vortices and inside deep troughs) the tropopause is lower than 400 mb, sometimes even below 4. The presentation of different groups of 500 mb (see e.g. No. 3, IS and 16, Mould soundings Bay, Portland and Nantucket). In the lowest The threefold structure in the tropopause layers of the troposphere a nearly isothermal heigth mentioned above will now be shown structure or strong surface inversion is found. by presenting different series of soundings The higher tropospheric layers possess a selected according to their location with nearly uniform lapse rate and in the strato- respect to both breaklines. Fig. 4 shows sphere an almost constant temperature with the distribution of the stations selected. The height or even a slight decrease (above 200 mb) stations are divided into five groups: is observed. The total structure (tropos here a) Stations north of the polar breakline plus stratosphere) is similar in almost ah the (polar group of soundings) soundings. This typical form of the polar b) Stations between the two breaklines soundings is best seen in the mean sounding (middle group of soundings) on the right side of Fig. 5. Table I presents c) Stations south of the subtropical breakline vertical lapse rates of the mean polar sound- (subtropical group of soundings) ing.

-..- Fig. 5. The vertical temperature structure of soundings north of the polar break- line (polar group). Note the location of the tropopause below 300 mb. The - IOO, -zoo, - 30° C a.s.0. are indicated by I. 2, 3 resp. Mean temperature distribution of these soundings is given on the right side. Tellus IX (1957), 3 266 FR. DEFANT AND H. TABA

Table I (hundredth of "C/roo m) layer (mb) from 940 920 900 880 860 840 820 800 780 760 740 720 940 to 920 goo 880 860 840 820 800 780 760 740 720 700

22 26 14 22 18 18 24 22 24 35 38 39 layer (mb) from 700 680 660 640 620 600 580 560 540 520 500 480 to 680 660 640 620 600 580 560 540 520 500 480 460 41 61 65 55 67 59 81 71 78 70 58 58 layer (mb) from 460 440 420 400 380 700 to 440 420 400 380 360 360 40 55 54 35 39 I Mean 61 layer (mb) from 360 340 320 300 280 260 240 220 zoo 180 160 140 120 360 to 340 320 300 280 260 240 220 zoo 180 160 140 120 IOO 1 I00 - 00 05 10 05 08 04 08 02 or 08 10 ox 00 Mean 05 From Table I one may see that the lower to 850 mb, vary considerably from one polar tropos here up to 700 mb is rather sounding to another. Often strong surface stable (mean ?apse rate - 0.25" C/IOOm) while inversions are observed which partly are low the upper rroposphere tends towards a more level polarfront intersections. The higher trop unstable stratification (mean lapse rate - 0.6" ospheric layers up to the tropopause show C/IOO m, sometimes -0.7 to -0.8). a surprisingly uniform vertical temperature The total mean tropospheric lapse rate is distribution. The tropopause is found at - 0.46"/100 m. In the polar stratosphere about 240 mb and in no case seems to be almost isothermal conditions are present (mean above zoo or below 300 mb. At the tropopause lapse rate -0.046"/100 m). The mean tropo- the nearly constant vertical lapse rate of the pause height is 370 mb (approx. 8 km) and higher troposphere changes to a more or the tropopause temperature is - 49" C. less constant temperature with height in the stratosphere. However in the higher strato- b) Middle group of soundings spheric layers (from 150 mb upwards) the Fig. 6 a, b shows all the soundings from sta- tem eratures are mostly - 60" C and lower. tions situated between the two breaklines togeth- In tR is group of soundmgs only those some- er with their mean. These soundings in general what closer to the subtropical breakline show look rather different in their vertical structure a small inversion just above the main tro o- from the ones mentioned above. The lowest pause. Here also slight indications of an aB di- parts of these soundings, approximately up tional rather high tropopause above 100 mb

Table I1 (hundredth of 'C/IOO m) layer(mb)from 980 960 940 920 goo 880 860 840 840 980 to 960 940 920 900 880 860 840 820 800 1 860 00 00 00 00 00 00 00 12 16 Mean 14 layer (mb) from 800 780 760 740 720 700 680 660 640 620 600 to 780 760 740 720 700 680 660 640 620 600 580

layer (mb) from 580 560 540 520 500 480 460 440 420 400 380 360 340 320 300 280 260 to 560 540 520 500 480 460 440 420 400 380 360 340 320 300 280 260 240 240 72 78 70 73 69 63 73 80 86 83 73 85 75 76 61 62 401 Mean 72 layer (mb) from 240 220 zoo 180 160 140 120 IOO 240 to 220 200 180 160 140 120 IOO 80 608o 60150 50 +18+11 00 15 24 19 16 01+06+031 Mean 12 Tellur IX (1957), 3 mi

1 m Y 268 . FR. DEFANT AND H. TABA can be noticed (e.g. No. 7, 9, 10, 11, 12, 16; around620mb (approx. 4.0 km). The tropopause Bismark, North Platte, Omaha, Columbia, appears in all soundings at rather high altitudes Dodge City, Albuquerque). (near the IOO mb level). The vertical lapse Table I1 presents vertical lapse rates for rate seems to be quite uniform in the whole different height intervals of the mean sounding troposphere and at the tropopause level tem- of this group. peratures of - 70' C or lower are observed. The lowest layers show rather variable A sharp break in lapse rate appears at the conditions and the mean lapse rate of - 0.14"C/ tropopause and from there on the temperature IOO m is only a result of the averaging process. increases with height up to 50 mb. At $0 mb Above 800 mb the lapse rate of the mean the observed temperatures are mostly around middle soundmg increases with height up to - 60" C. There seems to be no indication of a about 340 mb and decreases slightly until constant temperature with height in the the tropopause is reached (240 mb). A total stratosphere, at least not in the layer between mean lapse rate of the troposphere is - 0.64"/ IOO and 50 mb. Above $0 mb the observational IOO m. The tropopause temperature is - 56°C. evidence is rather poor, therefore nothing Even in the mean sounding a slight inversion conclusive can be said about it. A study of appears above the tropopause and from 170 the mean sounding shows the above described mb upwards the temperature decreases only structure in a clear manner. Table 111 presents very little with height ( - 0.12' C/IOO m) vertical lapse rates for different layers of up to 90 mb, where a secondary tropopause this mean sounding. level can be found. From there on the tem- In the lower troposphere (1,000-600 mb) perature remains nearly constant. where the temperature varies more from sounding to sounding, the mean vertical c) Subtropical group of soundings lapse rate is - 0.46" C/IOOm and from there Fig. 7 shows all the soundings from stations on a rather constant lapse rate of - 0.73' C/ south of the subtropical jet, to ether with IOO m extends upwards to 250 mb. At this their mean sounding. These soun%i ngs possess level the lapse rate changes a little and the mean again a rather similar vertical temperature struc- value from 250 mb to the tropopause is about ture. Even in the lower layers the differences in the same as in the lowest layer ( - 0.41" C/IOO temperature are rather small. The 0" C level is m). In the subtropical stratosphere the mean

Fig. 7. The vertical temperature structure of soundings south of the subtropical breakline (subtropical group). Note the high level tropopause (above 150 mb). Location of - 10'. -zoo, - 30' C a.s.0. are indicated by I, 2, 3 resp. Mean tempe- rature distribution of these soundings is given on the right side. Tellur IX (1957). 3 THE THREEFOLD STRUCTURE OF THE ATMOSPHERE 269 Table III (hundreth of 'C/IOO m)

layer (mb) from 240 220 zoo 180 160 140 120 240 to 220 200 180 160 140 120 IOO I00 47 39 41 60 46 44 10 Mean 41

P.IW J.I er-, Fig. 8. The vertical temperature structure of soundings inside the polar breakline. The tropopause is not well defined and the temperature continues to decrease upward. The mean of these soundings is presented on the right side of the figure. The - 1o0, -zoo, - 30' C a.t.0. are indicated by I, 2. 3. Tellur IX (1957), 3 270 FR. DEFANT AND H. TABA vertical increase of temperature with height indications may be seen. The form of these was computed to be +0.27' C/IOOm. The soundings and the breaks in lapse rat, depend mean tropopause is a little above 100 mb very much on the exact location of the surface (exactly at go mb) and the tropopause sounding with respect to the polar jet and on temperature is - 72' C. the steepness of the upper polarfront. One In the area south of the subtropical breakline has to be rather careful not to mix the tropo- a typical sounding shows therefore a more or pause with breaks in lapse rate appearing less uniform temperature decrease in the whole when the sounding passes the polarfront layer. troposphere, which extends upwards to higher The mean sounding also shows the continuous altitudes (near to IOO mb). From there on a temperature decrease with height. A broad noticeable increase of temperature with height layer with a slight change in lapse rate sepa- is observed. The tropopause level is in most rates the troposphere from the stratosphere. cases well defined. It must be clearly under- stood that soundings farther south would show e) Subtropical jet group of soundings the tradewind inversion. In addition the In Fig. 9 soundings through the subtropical tropo ause would show up in still higher jet are presented. In these soundings it is also altitu I! es (between 90 and 60 mb) and the difficult to select a single tropopause. But tropopause temperature would still decrease (in contrary to the case of polar jet soundings, the equatorial region even as low as - 80" C, two different tropopause levels are generally see e.g. No. 7, Khartoum). present. From these, the lower one is formed around 200 mb level or a little above, while d) Polar jet group of soundings the higher one shows up at about IOO mb. In soundings which intersect the wind This twofold tropopause structure can clearly maximum connected with the polarfront be seen in the mean sounding. Immediately (Fig. 8) no clear tropopause is observed. The after passing the lower tropopause, a frontlike temperature decreases mostly with height both temperature inversion can be noticed in most in tropo- as well as in stratosphere. This can cases. This lower tro opause seems to be be seen in the soundings of Berlin, Munich, closely related to the Peve1 of highest wind Dresden and Albany (No. 6, 8, 7 and 10). speed of the subtropical jet. Above this inver- In some of these soundings slight tropopause sion the temperature falls off again sharply

Fig. 9. The vertical temperature structure of soundings inside the subtropical breakline. Note the double tropopause level. The - 10,- 20, - 30' C a.s.0. are indicated by I, 2, 3. The mean of these soundings is given on the right side of the figure. Tellur 1X (19>7), 3 THE THREEFOLD STRUCTURE OF THE ATMOSPHERE 271 and at the higher tropopause this decrease changes abruptly to an increase with height. Here also it depends critically on the exact location of the sounding with respect to the subtropical jet maximum whether the lower or the higher tropopause is dominating.

5. Comparison of the mean soundings Fig. 10presents the five mean soundin s side by side. From this picture it is rather of vious that besides the strong baroclinici troposphere between polar and midd7 e sound-in the ings concentrated mainly in the polarfront zone an equally important baroclinicity is pres- ent in the stratosphere between middle and subtropical group, except that the horizontal

,.."A". ,.."A". I I... (01001, ..I., I..,.11,.....1 .*. 1.5 .I,,., temperature gradient is reversed. However, ~"....~YI."""I"...... *.. ..". ..". C... -11.1.1. 0. .C... l the stratospheric differences in temperature between polar and middle grou are rather Fig. 10. Mean vertical temperature distribution fo the SUC five different groups. The - 10, -20, -30' C a.s.0. are small. For a circulation ty e 1 as that of indicated by I, 2, 3. January I, 1956, where tR e subtropical jet

er UI 07 m IIW ..mw .in. -mi. .L* .I,,, *I. V."" tY....i,*,

Fig. 11. Vertical wind profiles of January I, 1956, from stations close to the subtropical tropo- pause breakline. The mean vertical wind profile is presented on the right side of the lowest row. Each small interval on the abscissa indicates 10kts and the main vertical lines intervals of so kts. Tellus IX (1957), 3 272 FR. DEFANT AND H. TABA remains mainly near latitude circle 30' N and wind maximum can be noticed ( acksonville, the polar jet is mostly far apart .the two Ft. Worth and Lake Charles), th at is when baroclinic zones arc also well separated. the sounding did not hit the jet center. There When the subtropical system moves north- are also some examples of extremely sharp ward, as e.g. over the Atlantic Ocean, then wind maxima in the vertical, indicating that the tropospheric and the stratos heric barocli- the sounding met the center (e.g. Crawley, nicity are nearly superimposecf In this case Liverpool (Fazakerley), Hemsby, Shiono- the middle part disappears and the whole misaki a.0.). The vertical wind profiles of baroclinicity exists between subtro ical and Liverpool and Crawley are rather beautiful polar group in tropo- and stratosp ere, but examples of a well defined jet restricted to in the stratosphere in the reversed sense.K a layer of a small vertical extent of about 50 mb (1.5 km). There exists a strong vertical 6. Vertical wind profiles of subtropical jet wind shear above and below the jet center. region By making a mean profile from these 20 In the foregoing discussion it has been stations, the phenomenon of the subtropical mentioned that the obvious breaks in tropo- jet in its hemis heric extent is well shown. pause height are closely related to the main That this jet p Kenomenon is not connected wind belts. To show this more clearly the with the polarfront can be seen from the vertical wind profiles of stations close to, temperature distributions presented in Fig. 9. or in, the southern breakline of the tropopause It was not possible to present a similar wind (group e) are presented in Fig. 11. From these profile for the polar breakline of tropopause vertical profiles it is rather evident that the height, as stations over Canada did not have subtropical breakline in tropopause height is good enough wind information and over close1 related to the subtropical jet, which in the Atlantic the two jet belts were so close all orthese profiles appears around the zoo together that both phenomena showed up in mb level. In some of these examples a broad vertical profiles.

Fig. I?,. Mean meridional cross-section of actual temperature for January I, 1956. Note the threefold structure of the tropopause. 1, = location of polarfront jet, Is = location of subtropical jet. Tellus IX (1957). 3 THE THREEFOLD STRUCTURE OF THE ATMOSPHERE 273 7. Summary of the foregoing discussion and has an average height of about 90 mb and presentation of a mean meridional cross- seems to rise towards the equator. A close section study of the middle soundings indicates that the subtropical tropopause continues also Fig. 12 a mean meridional cross-section towards north, increases in height and grad- based on the mean values of temperature ually disappears. obtained above is presented. The purpose of So far we have not mentioned anything this figure is to give the reader a general about the temperature field in the stratosphere. picture of the important features of the A study of the Fig. 12 shows that the strato- temperature field of the atmosphere in meri- spheric temperature difference between the dional direction. The threefold structure of polar and the middle part is not very pro- the tropopause can easily be seen. nounced. However, this difference becomes The foregoing discussion summarized in quite noticeable between the stratos here of Fig. 12 divides the atmosphere from the polar the middle part and the upper suE tropical to the tropical region into three widely differ- troposphere. In ordinary cases when the two ing parts represented by their mean soundings. main westwind maxima are separated by the Each part has its own typical vertical temper- middle part the stratospheric meridional ature structure in troposphere and stratosphere temperature gradient is not very pronounced. together with a characteristic tropopause height. But whenever the subtropical system is The polar and the middle part are separated displaced northward and the middle part in the troposphere by the polarfront, which disappears the total meridional temperature extends upwards to about 300 mb. There difference of the stratosphere concentrates in exist a definite break in the tropopause height a frontlike zone above the polarfront but with related to the polarfront jet (polar breakline an opposite inclination. in tropopause height; mostly between 250 From this discussion we may now easily and 350 mb). As mentioned before the understand why a hemispheric presentation of tropopause height as well as tropopause temperature makes it possible to visualise the characteristic features of the three dimen- sional structure of the atmosphere. The study sidered almost of the atmospheric field of temperature by cially for the upper troposphere. means of certain isobaric levels can lead to The middle and the subtropical, tropical part serious misinterpretations. It is also rather resp., are not separated by any frontal surface in difficult to draw any conclusion in this way. the troposphere. However, there exists a small In this discussion we have so far excluded baroclinicity between the almost quasibaro- the area from about 80 to 90' N (inner tropic stratification of the middle and sub- cap). This region is affected by d:!: tropical troposphere. This is indicated by factors such as advection in different layers the inclination of the isotherms in Fig. 12 and from different directions or radiational around lat. 25-30"N. On top of this weak influences. Here no representative mean state baroclinic zone another break in the tropo- can be found. Ths does not mean that we pause is present (subtropical breakline in tropo- underestimate the importance of this part of pause height; mostly between 240 and 100 the atmosphere. mb). The vertical extent of this breakline is Fig. 13 presents the meridional distribution much larger than that of the polar breakline. of potential tem erature of the section shown The middle tropopause seems to continue in Fig. 12. In tK e troposphere north of the southward and increases its height to about polarfront the isentropic surfaces rise towards 180 mb. The subtro ical jet is located in the the pole with an almost constant inclination. mean just beneath t E's continued part of the The concentration of these surfaces starting middle tro opause. This is why the soundings at the surface around lat. 6s" and continuing through tK e subtropical jet indicated in all towards the pole up to a height of about 650 cases a lower tropopause as well as the higher mb is an indication of the so called arctic front. subtropical one. The subtropical tropopause The strongest concentration can be seen inside Tellus IX (1957), 3 2-7 0 3 3 6 5 ..- the boundaries of the polarfront. The sharp slope of the isentropes immediately south of the olar jet (about lat. 45O) seems to be closeP y related to the existence of the jet. Between 45 and 3ooN the isentropic levels are almost horizontal. This is an indication of the quasibarotropic state of the middle part of the atmosphere described before. From - _. 30’ southward to about 2ooN a further ..I..... -I- downward slope of the isentropes indicates Fig. 14. Examples of complicated double jet structure the difference between middle and subtropical in the vertical wind profiles of Keflavik (Iceland) and air. Further south the atmosphere is again Narsarssuak (Greenland). jusi barotropic. Here the concentration of this stratospheric frontal slope becomes more t e isentropes especially in the upper tropo- pronounced and may be interpreted as a sphere has a close relation to the existence of stratospheric continuation of the polarfront. subtropical jet near 200 mb. In such cases the two jet maxima approach In the upper troposphere of the tropical and nearly superimpose, resulting in a double region between the equator and 20°N the wind maximum in the vertical wind fields. stratification is again barotropic. Above the Fig. 14 shows an example of such a double tropical tropopause (tropical stratosphere) and jet over southern Greenland and Iceland, in the middle stratosphere (from about 20’ N) where the two jets are close to each other (see the isentropes slope downwards until about also Figs. 2 and 3). 65O N. From there on towards the pole they incline upwards. In the space where the middle Acknowledgements tropopause rises and gradually disappears (northern side of subtropical jet) a frontlike The authors wish to express their thanks to concentration of the isentropes can be seen. Mrs. Birgitta Lorensson and Mrs. Birgit Dahl- Whenever the subtropical jet moves far borg for helping with the preparation of the northward and comes close to the polar jet, material and the figures. Tellur IX (1957), 3