OCTOBER-DECEMBER 1963 MONTHLY WEATHER REVIEW 605
SUMMARY OF HIGH LEVEL TURBULENCE OVER UNITED STATES
DeVER COLSON
US. Weather Bureau, Washington, D.C.
[Manuscript received May 6, 1963: revised July 3, 19631
ABSTRACT
A comprehensive study was ninclc of pilot-reported occurrences of high level turbulence (more coiiiriionly referred
to as CAT) over the United States from September 1960 through August 1962. The data consist of 12,126 reports
of moderate or severe turbulence above 15,000 f t . The tabulations presented include the distribution of turbulence
by months, seasons, altitude, and geographical location, and its association with such mcteorological items as: the
position with respect to the jet stream, curvature of the jet stream, strength of the isotach center, streamline curvature,
troughs, closed or cut-off LOWS, and mountain waves. Some additional data are included for levels as high as 70,000 ft.
1. INTRODUCTION
The U.S. Weather Bureau has undertaken a compre-
hensive study of high le\rel turbulence using a iiiuch larger
aiid far more complete set of pilot reports than has been
andable for previous studies. The actual tabulation and
analyses have been limited to reports of moderate or
greater turbulence above 15,000 ft. m.s.1. Pilot reports of
turbulence in or near towering cumulus clouds or thunder-
storms have been eliminated from the tabulations aid
analyses.
It is fully recognized t8hat this analysis is limited to the
areas of reported turbulence. I t is not known whether
there was no turbulence in the other areas or simply that
there were no planes flying over these areas at the time or
altitude that turbulence occurred. The pilot reports are
highly subjective and qualitative, as the intensity of the
turbulence felt is dependent among other things on the
type and weight loading of the plane and the air speed
and angle of attack. Even so, these reports constitute a
valuable set of data and every attempt should be made to
extract as many useful conclusions from them as possible.
Pilot reports were collected from weekly summaries
tabulated by all Weather Bureau Airway Forecast Centers.
from teletypewriter data on Air Force circuits, and from
additional data on a monthly basis from the National
Weather Records Center at Asheville. Daily tabulations
were made of all reports of moderate or greater turbulence
occurrences above 15,000 ft., exclusive of those in or near
towering cumulus clouds or thunderstornis. Since Oc-
tober 1, 1961, data pertaining to the occurrences have been
placed on punch cards and standard print-outs are avail-
able t o any interested research or operational group.
2. RESULTS
GENERAL DISTRIBUTIONS
Over the 24-month period from September 1960 through
August 1962, there were 12,126 reports of moderate or
greater turbulence occurrences available for the study.
The mean monthly distribution (fig. 1) shows a ninximurn
in February and a miniinum in Siigust. It is interesting
to note that there were 5,623 reports in the last 12 months
as compared to 6,503 in the first 12 luonths, in spite of the
fact' that the number of high lerel flights and the effi-
ciency of the collection of the pilot reports is likely t o have
increased gradually throughout the entire period. The
decrease could have been caused by fewer or smaller tur-
bulence areas in the atmosphere, but it could haw resulted
from improved forecast'iiig so that the pilots could better
aroid these high threat areas or could fly through these
areas at reduced flying speeds which would reduce the
frequency of reports of the more serious turbulence.
The ox-erall distribution of the occurrence of moderate
or greater turbulence with altitude is shown in figure 2.
The frequency of turbulence reports by altitude is highly
dependent on the frequency of flights by altitude. Un-
fortunately, accurate information on the latter is difficult
to obtain. However, discussions with aircraft operatioid
groups suggest a inaxiinunl of flight frequency around
30,000 to 31,000 ft. and then a rapid decrease. Thus, i t
appears that the peak at 35,000 to 36,000 ft. is not due to
any bias in actual flight frequencies but is associated with
the average height of the maximum wind and the tropo-
pause.
The aver age monthly height of the reported turbulence
occurrences showed a regular -cariation throughout the
year ranging from a minimum of 30,200 ft. in March to a
maximum of 34,000 f t . in July. The monthly median
height ranged from a minimum of 30,900 ft. in March
to a maximuni of 34,300 ft. in July. It is difficult to
attribute the above monthly trend to any bias in the flight
frequencies and i t appears more likely to be associated with
changes in meteorological conditiolls, such as a higher
tropopause and level of maximuru winds, during the
summer months.
606 MONTHLY WEATHER REVIEW OCTOBEB-DECEMBER 1963
FIGURE 1.-Mean monthly distribution of moderate or severe
turbulence Septcmber 1, 1960, through August 31, 1962.
The distribution o€ over 12,000 reported occurrences of
moderate or greater turbulence during the %year period
is shown in figure 3 using 5O latitude-longitude “squares.”
Obviously this distribution is greatly biased by the flight
routes. However, it appears that the concentration
along and to the lee of the Rocky Mountains is due
more likely to the influence of the mountains in producing
mountain waves and their associated turbulence than to
any flight bias in this region. The peak over Kentucky,
West Virginia, and western Virginia may also have some
validity since it is slightly south of the more frequent
flight routes into New York.
The average number of reported occurrences of moder-
ate or greater turbulence for each month and season is
shown in figure 4 for seven arbitrarily chosen regions:
Far West, Rocky Mountains, Southwest, Midwest,
South-Central, Northeast, and Southeast. There was a
winter maximum in each region except in the Northeast
and Southeast where the maxima are in the spring.
There was a smnmer rnininiuin in each region except in
the Northeast where the minimum was in the fall. The
results in general confirin and greatly extend some of the
earlier findings by Clem [l].
These variations may be associated with actual wind
patterns rather than with some flight bias. In all regions
except the Rocky Mountains and the Far West the num-
ber of occurrences in the fall was considerably less than
half the number in the winter months. In the Rockies
and the Far West, the fall months appear to be relatively
inore important than in the other regions which may be
due to the increase in the mountain wave activity over
the Rocky Mountains in the fall. As mentioned before,
spring appears to be relatively more important in the
Northeast and Southeast than in the other regions.
It is evident from all of the above distributions of
43-44 :.::
)
PERCENTAGE
FIGURE 2.-Distribution of moderate or severe turbulence by
altitude.
turbulence occurrences that there can be some moderate
or severe turbulence expected along all flight routes, at
all altitudes, and a t any time of the year.
SPECIFIC METEOROLOGICAL PATTERNS
This section of the paper deals with the association of
the reported occurrences with various meteorological
pat’terns from September 1, 1961, through August 31,
1962. The reports of moderate or greater turbulence
were tabulated as to whether they occurred (I) within 150
mi. to the left of the jet, (2 ) within 150 mi, to the right of
the jet, or (3) more than 150 mi. €rom any jet inwhich case
they were listed as non-jet cases. The jet positions were
those analyzed by the National Meteorological Center
(NMC) and transmitted on the Maximum Wind Analysis
facsimile charts. Left and right sides were defined looking
downwind along the jet. For the occurrences associated
with the jet, tabulations were made with respect to whether
the curvature of the associated segment of the jet was
cyclonic, anticyclonic, or straight. Occurrences between
two jets (with the jets not over 300 mi. apart) were tabu-
lated and further subdivided as to whether the jets were
parallel, converging, or diverging.
The occurrences of moderate or greater turbulence were
tabulated with respect to whether the curvature of as-
sociated height contour lines (300-mb. facsimile chartj
was cyclonic, anticyclonic, or straight. Occurrences
which were associated with strong wind flow normal to the
Rocky or Appalachian Mountains were tabulated as
mountain-wave cases. Occurrences within 150 mi. of a
trough line (as shown on the NMC 300-mb. facsimile
OCTOBER-DECEMBER 1963
Season
MONTHLY WEATHER REVIEW 607
Within Within 150 mi. 150 mi. Non-jct to left to right
FIGURE 3.-Distribution of moderate or severe turbulence ovcr thc
United States.
chart) were listed with a further breakdown as to whether
the occurrences were ahead of or behind the trough line.
Tabulations were niade of the occurrences associated with
closed or cut-off Lows.
Jet Stream. Association.-Of the 5,623 reports of nioder-
Fall (Sept., Oct., Nov.) ................................. 34% !I% 35% Winter (Dee., Jan., Feb.) ............................... 32
Summer (June, July, Aug.) ............................. 29 41 Spring (Mar., Apr., May) ............................
i i: I 4: I
38
ate or greater turbulence during the year, 36 percent were
within 150 mi. to the left or the cold side of the jet, 28
percent were within 150 mi. to the right of the jet, and 36
percent were more than 150 mi. from any jet (non-jet
cases). In any study or forecast of turbulence limited to
150 mi. to the left or cold side of the jet, 64 percent of all
the occurrences would have been neglected.
The seasonal distribution, as shown in table 1, indicates
that the association between the reported occurrences and
the left or cold side of the jet was most important in the
winter and spring when the jet streanis were stronger and
better organized, and less effective in the summer. During
t,he summer months, 71 percent of all the reported occur-
rences were either on the riglit side or more than 150 mi.
I~IGURE 4.-Distribution of moderate or severe turbuleiice by moiiths and seasons.
608 MONTHLY WEATHER REVIEW Oc-I OSER-DECEJI I ~E B 1063
TABLE 2.--Xeasonal distribution of curvature of the jet associaled with
the reported turbulence occtirrences -
I I I
Bcason
Of the occurrences reported within 150 mi. on the left
side of the jet, the curvature of the associated segment of
the jet stream was cyclonic 29 percent, anticyclonic 30
percent, and straight 41 percent of the time. Of the
occurrences reported within 150 mi. to the right side of the
jet, the curvature of the associated segment of the jet
stream was cyclonic 37 percent, anticyclonic 26 percent,
and straight 37 percent of the time. This tabulation
indicates that when turbulence was reported, the segment
of the associated jet stream was cyclonically curved only
slightly more often than it WVRS curved anticyclonically.
Certainly, in a large majority of the cases, the segment of
the jet associated with the reported turbulence was either
anticyclonically curved or straight and these situations
cannot be neglected in t~ study or forecast of high level
turbulence.
The distribution of the currature of the jet stream asso-
ciated with all the reported occurrences within 150 mi. on
either the left or right side of the jet, as shown by table 2 ,
indicates that cyclonically curved jets were more fre-
quently associated with the occurrences in the suninier
and fall months, anticyclonically curved jets in the spring,
and long straight jets in the winter months.
Of the total number of reported occurrences of moderate
or greater turbulence, 43 percent were located within the
100-kt. isotach and 8 percent were within the 150-kt.
isotach. A seasonal breakdown with respect to the occur-
rences within the 100-kt. isotach indicated that 56 percent
of all the occurrence reports in the winter months were
within the 100-kt. isotach and only 12 percent of all the
suiiinier reports were within the 100-kt. isotach center.
Of all the moderate or greater turbulence reports, 11
percent were between two jets (jets 300 mi. or less apart).
A further breakdown indicated .5 percent occurred between
parallel, 3 percent between converging, and 3 percent
between diverging jets.
Circulation Patterns.-Of all the reported nioderat e or
greater turbulence occurrences during the pear from
September 1, 1961, through August 31, 1962, 35 percent
T.QBLE 3.-Seasonal distribution of contour curvature associated with
the reported turbulence occurrences
, , I
1 Cyclonic!
1
Snti- 1 Straight
SCaSOll (percen t) csdonic! (percent) (percent)
Fttll.. . -... . . .... ... . .....__. . ........ . ....... . .~.
Winter -.......
~
....-......-..
~
....
~~
.... ~~...~...
Spring. ....
~ ~
......
~
-......- ~.. ..~~ .....
~
..... .- .-
Summer ___....... . .-. ..... - .-~.. ... . ....... .-.-..
33 47
34 24
~-
eo 45 33 e9
were associated with cyclonically curved, 29 percent with
anticyclonically curved, and 36 percent with straight
height contours. Again, there was only a slightly greater
frequency of reported occurreiices with cyclonically rather
than anticyclonically curved height contours. If only the
turbulence occurrences associated with cyclonically curved
contours were considered, 65 percent of the turbulence
sit'uations would have been neglected.
The seasonal distribution of the curvature of the height
contours associated with the reported turbulence occur-
rences (table 3) shows the cyclonic patterns are relatively
much more frequent in the fall and suninier months and
the long straight contours are relatively more frequently
associated with the reported turbulence during the winter
months. In the spring months, the reported occurrences
seem equallj- likely to be associated with cyclonic, nnti-
cyclonic, or straight contours.
I n this study, it was found that 21 percent of all the
reported occurrences were tabulated as possible moun-
tain-wave type. Of all the reported occurrences, 16 per-
cent were found to be located within 150 mi. ahead of or
to the rear of a trough line (as indicated on the NMC
300-nib. facsimile chart). I n the trough study, there were
as many occurrences ahead of as there were to the rear of
the trough line. This study indicated that only 3 percent
of all the reported occurrences were associated with :L
closed or cut-off Low.
The results in table 4 indicate that the associatioii of
reported turbulence with troughs is greater in the fall than
in the winter and spring when the association with the left
or cold side of the jet is more pronounced. The percent of
occurrences associated with closed Lows is quite low
throughout the year but is relatively greater in the fall and
spring. The table shows considerable mountain-wave
type occurrences all during the year but that thej- were
relatively more important in the winter than in the other
seasons.
HIGHER LEVEL TURBULENCE
At the present time, there is an increasing interest in
the problem of turbulence a t levels well above 40,000 ft.,
especially in connection with the proposed supersonic jet
transport operations. Some turbulence data above 40,000
ft. from a high-altitude sampling program (March 1958
through August 1959) were available at the time of this
study.
It is interesting that the percent of flights with no tur-
bulence was considerably higher in the 60,000-70,000-ft.
layer than in either of the other layers. I n the highest
These results are shown in table 5.
TABLE $.-Seasonal distribution of the association of tvrbulence occurrences with troughs, closed Lows, and naountain waves
Srason (perccnt)
OCTOBER-DECEMBER 1963 MONTHLY WEATHER REVIEW 609
TABLE 5.-Higher level turbulence data
I I
Flights witli-
Altitude (thsd. It.)
I Nizk:o‘ 1
No
I
Light or
I
Noderate turbulence greater or greeter (percent) (percent) (percent)
Distribution with respect to iriteiisity
Altitude (tlisd. Et.) IOecurrenccrl Light
1
Moderate
1
Severe (percent) (percent) (percent)
,-.---
40-5OL. ~. .. . . ~~... . -. . . . . .. . . ~.. . . . . .
50-60L. .. ~. . . ......
~
...... . .. ~-.. . . ..
60-70 ...
~
~. .-.
~
. . . . ~-. . . . . . . . ~. ...- ~.
44 48 8
layer, 45 percent of all the reported occurrences were
moderate or greater as conipared to 59 and 56 percent i n
the other two layers. However. 25 percent of all the
fliqhts in t,he 60,000-70,000-ft. layer encountered moderate
or greater turbulence. This indicates that turbulence
may be 01 considerable importance in the stratospheric
levels.
Most of the reported turbulence occurrences were above
the tropopause and well into the stratosphere. The mech-
anism for the production of turbulence in the stratosphere
may be considerably different from that involved in the
troposphere. The actual wind speeds and wind shears in
the lower stratospheric levels are not nearly as strong as
in the troposphere near the jet stream levels. However,
there are temperature and density discontinuities about,
which wave motions and turbulence may be generated in
ways similar to those suggested by Reiter [9]. Many
more turbulence occurrence data will be needed before
this stratospheric turbulence can be understood.
3. SUMMARY AND CONCLUSION
This study provides confirmation for ideas which have
been presented by other workers in this field. The data
on the association of turbulence with the jet stream and
its associated vertical and horizontal wind shear support
earlier work by Harrison [8] and George [6]. The data
on the association of turbulence with air flow over the
niountains support earlier works by Clodman et al. [a ].
Harrison [S], and Colson [3]. The data on the association
of turbulence with troughs confirm earlier findings by
Colson and Rustenbeck [4] and Colson [5 ]. This work
shows some relation between turbulence and upper-level
closed Lows which has been mentioned by Harrison 171,
Colson and Rustenbeck [4]; however, the percent of such
occurrences was found to be quite low in this study.
While this study confirms the above-mentioned associa-
tions, it is apparent from this and earlier studies, Colson
131, 151, that no one meteorological pattern or model will
be sufficient for the understanding and forecasting of
high level turbulence over the entire country and through-
out the year. There appears to be more than one mecha-
nisni effective in the production of smdl-scale eddies
which affect aircraft in flight. These include (1) strong
jet stream configurations with their associated horizontal
and vertical wind shears, (2 ) sharp inajor and even minor
troughs moving across the country with their associated
directional shears, and 13) airflow over mountain barriers
with t8he proper wind shear and st,ability, Other niecha-
nisms may be involved in the turbulence occurrences
associated with cirrus cloud lormations and with high
pressure or anticyclonic circulation pat terns.
In general, it will be necessary to make a complete and
detailed analysis of the atmospheric conditions and to
forecast the changes in these conditions in order to make
an accurate forecast of high level turbulence on ti local
scale which will be useful for aviation operations. High
level turbulence is not a simple phenomenon and it will
not be possible to arrive at any simple criterion or inapic
number which mill solve the problem of forecasting high
level turbulence accurately. ‘.
ACKNOWLEDGMENT
The author wishes to express his thanks to the late
Dr. Harry Wexler who suggested that this research pro-
gram be undertaken and who furnished very helpful
suggestions and encouragement during the study.
REFERENCES
1. L. H. Clem, “Clear Air Turbulence from 25,000 to 45,000 Feet
Over the United States,” Sir Wcathcr Service, Technical
Report 105-47, July 1957.
2. J. Clodman, G. M. Morgan, Jr., and J. T. Ball, “High Level
Turbulence,” Research Division, College of Engineering,
Xcn? York University, report 011 contract AF19(604)-5,208,
Sept. 1960.
3. D. Colson, “Ana!ysis of Clear Air Turbulence During April
1960,’’ Monthly Il’ealhm Review, vol. SS, No. 3, Mar. 1961,
pp. 94-98.
4. D. Colson aiid J. D. Rustenbeck, “Clear Air Turbulence Over
Western United States October 8-12, 1960,” U.S. Weather
Bureau manuscript, June 1961.
5. D. Colson, “Analysis of Clear Air Turbulence Data for March
1962,” ,fi407tth1y Weather Review, vol. 91, No. 2, Feb. 1963,
pp. 73-82.
6. J. J. George, A Method f o i the Prediction of Clear Air Turbulence,
Eastern Air Lines, Inc., Atlanta, Aug. 1960.
7. H. T . Harrison, “Cold Lows and Clear Air Turbuleiice,” United
Air Lines, Inc., Information Bulletin, Denvcr, Jan. 19, 1949.
8. H. T. Harrison, “Progress in Forecasting High Level, Clear Air
Turbulence,” United Air Lines, Inc., Meteorology Circu.Zar
No. 52, Denver, Mar. 1, 1961.
9. E. R. Reiter, “Nature aiid Observation of High Level Turbulence
Especially in Clear Air,” Technical Paper NO. 41, Colorado
Statme University, Fort Collins, Dec. 1962.