satellite views of hurricane camille

8/7/2019 Satellite Views of Hurricane Camille

http://slidepdf.com/reader/full/satellite-views-of-hurricane-camille 1/82

NASA TECHNICAL NOTE NASA TN 0-7547

LMJ

(NASA-TN-D-7 5 4 7) SATELLITE VIEWS OF

HURRICANE CAMILL- S- ELLITE VIE SC (NASA) P HC $4.00 N74-20194

CSCL 04B

-1/20 34577

SATELLITE VIEWS OF HURRICANE CAMILLE

by William E. henk and Edward B. RodgersGoddard Space Flight Center

Greenbelt, Md. 20771

NATIONAL AERONAUTICS AN D SPACE ADMINISTRATION * WASHINGTON, D. C. * APRIL 1974



1. Report No. 2. Government Accession No. 3. Recipient's Catalog No.

4. Title and7utitle 5. Report Date

Satellite Views of Hurricane Camille APRIL 19746. Performing Organization Code

650

7. Author(s) 8. Performing Organization Report No.William E. Shenk an d Edward B. Rodgers G-7405

9. Performing Organization Name an d Address 10 . Work Unit No.

039-23-01-01Goddard Space Flight Center 11. Contract or Grant No.

Greenbelt, Maryland 20771

13. Type of Report an d Period Covered

12. Sponsoring Agency Name and Address

Technical NoteNational Aeronautics and Space Administration

Washington, D. C. 20546 14. Sponsoring Agency Code

15. Supplementary Notes

16. Abstract

Three periods within the life cycle of Hurncane Camille (1969) are studied with radio-

metric and camera measurements from Nimbus-3 and camera information from ATS-3

in conjunction with conventional information. These periods are the deepening phase,the interaction of Camille with midlatitude westerlies, and the excessive rain producing

period when the cyclone was over the central Appalachian Mountains. Just prior to

significant deepening, the Nimbus-3 Medium Resolution Infrared Radiometer (MRIR)

showed that a pronounced feeder band had formed southeast of the center which was

associated with the rapid transport of moisture into the storm circulation. During the

rapid deepening phase the MRIR measurements indicated the development of large-

scale subsidence throughout the troposphere northwest of the center. When Camille

was over the lower Mississippi Valley it acted as an obstruction to the environmentalwind. A region of widespread subsidence was created west and north of the cyclone

center. Increased cloud-top elevations, back to the levels reached when Camille wasan intense cyclone over the Gulf of Mexico, were estimated from the Nimbus-3 High

Resolution Infrared Radiometer (HRIR) measurements on August 20, 1969, whenCamille produced rains or major tiood proportions near the east slopes of the

Appalachians in central Virginia.

17. Key Words (Selected by Author(s)) 18. Distribution Statement

Hurricane analysis; Radiation

measurements; Geophysics;

MeteorologyCat. 20

19. Security Classif. (of this report) 120. Security Classif. (of this page) 121. No. of Pages 22. Price*

Unclassified Unclassified I 1 $4.00oo

*For sale by the National Technical Information Service, Springfield, Virginia 22151.

j



CONTENTS

Page

ABSTRACT . . . ........... ............... i

INTRODUCTION . . . .. . . . . . . . . . . . . .. . . .. . 1

AVAILABLE DATA AND MAPPING PROCEDURES ... ...... . 2

RESULTS . . . . . . . . . . . . . . . . . . . . . . . . .. . 4

CONCLUSIONS ............ ............. 73

ACKNOWLEDGMENTS ...... ................ 76

REFERENCES ... ................ . ..... . 77

APPENDIX-RELATIONSHIP BETWEEN BRIGHTNESS AND

CLOUD-TOP TEMPERATURE . ......... . . . . 79

1G1PAG

iii



SATELLITE VIEWS OF HURRICANE CAMILLE

William E. ShenkGoddardSpace FlightCenter

an d

Edward B. Rodgers

EnvironmentalResearch and Technology, Inc.

INTRODUCTION

From a meteorological point of view, Hurricane Camille was the most important Atlantic

hurricane of record (Reference 1). As it crossed the Gulf Coast the highest storm tides

on record were produced (7.5 m (24.6 ft)). No recording anemometer equipment was able

to survive in the area of maximum winds as the cyclone moved ashore. Velocities approach-ing 175 kn were estimated from an appraisal of the damage within a few hundred yards of

the coast. Th e minimum central pressure of 905 mbar was second only to the Florida Keys

storm of 1935 (892 mbar) for Atlantic hurricanes. Whereas the Florida Keys storm proba-

bly ha d higher maximum winds than Camille, the diameter of hurricane winds was less than

100 km (62 miles) as compared to 160 to 240 km (100 to 150 miles) for Camille.

Camille began to develop near the island of Grand Cayman from a tropical disturbance

which moved westward across the Atlantic from the African Coast (Figure 1). The cyclone

deepened rapidly as it passed over the western tip of Cuba into the Gulf, and it reached

maximum intensity by 0000 GM T on August 17, 1969. This intensity was maintained

with little apparent fluctuation as observed by reconnaissance aircraft until the center

reached the coast at 0430 GM T on August 18. Within 12 hours the winds had decreased

to less than hurricane strength as the cyclone moved northward through Mississippi. For

the next 36 hours the remains moved north, then east, with the winds and rains diminishing

in the typical history of a tropical cyclone moving away from its moisture source. Suddenly,

excessive rainfall began over the central Appalachians in response to an injection of extreme-

ly moist ai r at low levels, an unstable lapse rate, and topography (Reference 2). This

portion of Camille's history is similar to that of Hurricane Diane in 1955 that produced

massive flooding in New England after Diane had weakened considerably when it moved

across the Carolinas and Virginia. Camille briefly regained tropical storm strength after

moving eastward off the Atlantic Coast near Norfolk, but interaction with a cold front

caused the cyclone to become extratropical.

Throughout the important phases of Camille's life cycle the storm was observed by

meteorological satellites such as Nimbus-3, ATS-3, and operational satellites. This out-

standing coverage, along with conventional measurements, permitted the examination in

some detail of three important events in the history of Camille which were: (a) the rapid

deepening phase; (b ) the interaction with the midlatitude westerlies as the cyclone moved

inland; and (c) the development of the heavy rains over the central Appalachians.

1



105 8 100 95, 90* 85' 80* 75* 70' 65 60 550' 45W 40* 35* 3W 25*

TROPICAL DEPRESSION AAMP PO

2TROPICAL STORM N

.35*N

intesoio Ie

3010N

80W 75 70 65 60 55 0 45W

Figure 1. Track and development stages of Hurricane Camille from August 9 to 22, 1969. (Extracted from

Monthly Weather Review, Vol. 98, No. 4, April 1970, p.294.)

AVAILABLE DATA AND MAPPING PROCEDURESAAAAA

Radiemetric measurements with a 55-km subsatellite-track spatial resolution were taken at

shows the layers of a mean tropical cloudless atmosphere that contribute to the sensed

radiance in each of the three infrared spectral intervals. The emission in the 6.5- to 7.0-pm

region primarily comes from the upper troposphere, whereas the 20 - to 23-pm emissionemanates mostly from the middle troposphere, and the 10- to 11-pm radiance nearly all

comes from the surface and lower troposphere.

Nimbus-3 High Resolution Infrared Radiometer (HRIR) nighttime emission measurements

were made at 3.5 to 4.1 pm with a 9-km subsatellite-track spatial resolution. Further

2



40-

---- 6.5-7.Oi0m

......... IO - II m35-- 20 -23 Lm

30-

25 -

o 20-

15 P----o--

-- *....'"

o"""*, I I

.01 .10 .20

Figure 2. Weighting function (I) curves (p units are w/cm2 ster cm

-1

km normalized to unity) determined from radiative transfer theory

based on a mean tropical atmosphere for three Nimbus MR IR channels

(6.5 to 7.0 pm, 10 to 11 pm, and 20 to 23 pm).

details of the Nimbus-3 satellite and the instrumentation is provided in the Nimbus-3

Users' Guide (Reference 3).

Visible measurements were obtained at 11-min intervals from the Multicolor Spin Scan

Cloud Camera (MSSCC) on ATS-3. Due to the frequent observations, it was possible

to track the motions of individual cloud elements to determine the wind. The clouds

were classified as high or low, depending on their appearance an d motion; the high clouds

most likely indicating the air motion near 200 mbar and the low clouds near 850 mbar.

Some middle clouds may have inadvertently been included, especially in the high-cloud

sample. Hubert and Whitney (Reference 4) have discussed cloud motion-wind accuracies

from ATS data. More information on the ATS-3 satellite and the MSSCC can be found in

the ATS Users' Guide and catalog publication series.

3



Conventional radiosonde and surface observations were obtained every 12 hours from

0000 GM T on August 14 to 1200 GM T on August 20, 1969. The radiosonde information

was analyzed over the Gulf of Mexico, the Caribbean, an d the United States at the stand-

ard pressure levels (850, 700, 500, 300, and 200 mbar), except for the observations of1200 GM T on August 18 an d 0000 GMT on August 19. At these times, charts were con-

structed over the United States every 50 mbar from the surface to 100 mbar and at

the 70-, 50-, 30-, and 20-mbar levels. This permitted a more detailed examination of

Camille and its interaction with the westerlies after the cyclone had moved inland.

The Nimbus radiation measirements were mapped in the stereographic horizon map

projection (Reference 5) where the center of Camille was always placed at the center of

the map. Thus, Camille was mapped (but not viewed) in the same perspective throughout

the period of interest. A correction was applied to the measurements from the water vapor

channels (6.5 to 7.0 pm an d 20 to 23 pm) to compensate for limb-darkening effects.* Th e

correction varied from zero to several degrees Kelvin for equivalent blackbody temperatures

(TBB 's) measured at a nadir angle of 500 (the maximum allowed).

RESULTS

Rapid Deepening Phase

The first significant event in the life of Camille was the rapid deepening that occurred

before and after the cyclone center passed over the western tip of Cuba. An interpretation

of the time history of the minimum central pressure curve is shown in Figure 3. There is a

suggestion of two periods of deepening. The first primarily occurred between 1200 GM T

on August 14 an d 15. Four closely spaced reconnaissance aircraft reports centered at about

1200 GMT on August 15 allreported a minimum pressure of about 965 mbar. Shortly there-

after the center crossed the southern coast of Cuba near the western tip. It took approxi-

mately 3 hours for the eye to move to the northern coast. During this time the shape of

the minimum pressure curve is more uncertain, but since most of the cyclone's circulation

renained over the water it is likely that the rise, if any, in the central pressure was small.

There were no reconnaissance penetrations of the eye for 18 hours after Camille left the

Cuban Coast. At 0500 GM T on August 16 Nimbus-3 passed overhead with an almost

vertical view of the storm.

Figure 4 depicts the cloud shield of Camille with HRIR measurements (1:2,000,000 map

scale) 4 hours after the center moved off the Cuban Coast. Two features suggest that

further deepening ha d begun. First is the appearance of a well-defined eye as indicated by

TBB's as high as 287 K (the insert in Figure 4 depicts a small area near the center mapped

at 1:500,000). This measurement could have been caused by either near transparent cirrus

*V . V. Salomonson, personal communication, 1972.

4



(997)1639-

1642 (985)

1647

1001649

00(965) (951) (911) (905) (905) (920)1674- 1560- 1667-1653 1660 1667- 16764- 1680-2 1697-

1655 1662 1669

980

n 960 )- RECONNAISSANCEAIRCRAFT

Z OVER CUBA 0- LAND STATION

j 940

S920 -

CENTER CROSSES

U.S. COAST LINE

14 15 15 16 16 17 17 18 18

12GMT 00GMT 12GMT 00GMT 12GMT OOGMT 12GMT 60GMT 12GMT

AUGUST. 1969

Figure 3. A trace of the minimum central pressure for Hurricane Camille from

1200 GM T on August 14 to 1800 GM T on August 18, 1969. The estimated centralpressures and the Nimbus-3 orbit numbers are shown above each of the arrows thatindicates the time of a Nimbus-3 overpass.

and/or low clouds, or a partially filled radiometer field of view with any of the above cloud

combinations. Second, a ring of very low T, 's (< 210 K), indicating the opaque stormcloudiness, surrounds the eye except to the west. This ring represents the highest clouds

associated with the wall cloud where the lowest (and probably the weakest) portion is on

the west side.

With the satellite evidence of a nearly clear eye (in the radiometric sense) an d a strong wall

cloud that almost encircled the eye, the rapid deepening was shown in Figure 3 to com-mence immediately after the storm moved off the Cuban Coast. No reconnaissance reports

were available until 1800 GMT on August 16 when Camille had reached severe status, an

intensity that the aircraft central pressure measurements would indicate was maintained

until the cyclone crossed the U. S. coastline. Thereafter, the central pressure rose rapidly;

the Jackson, Mississippi, minimum pressure was 978 mbar as the center passed about

35 km east of the station at 1200 GMT on August 18.

Perhaps the best day and night satellite observations of Camille during the rapid deepening

phase were made by the Nimbus-3 MRIR. By examining the information from the three

infrared channels at 12-hr intervals and the visible channel data every 24 hours, it was

anticipated that the changes in the structure of Camille as it became a severe cyclone could

5



86.0 W 85.0 W 84.00 WI I290 260

23.9 2- 28 2 2 20

2.923.9 N

210 22 2.9 N

21 0 10 21 0I 022.9 N -22§ n 229 N

S205 220

210

21.9 N 21.9

1 0 20 20

230

210 50 2 200

270 /

195

22.90 N -200 200 -22.90

N

21019

210

85.30

W 85.00W

Figure 4. Two views of the interior portion of the cloud struc-

ture of Hurricane Camille at 0500 GMT on August 16, 1969,

as depicted by the 3.5- to 4.1-,m equivalent blackbody temp-

eratures (TBB's) from the Nimbus-3 HRIR. The top picture

was mapped at a 1:2 million scale while the insert over the eye

seen in the bottom picture (the area within the dashed box inthe top picture) was mapped at a 1:500,000 scale.

be determined. The daytime observations were combined with the ATS-3 cloud motions

to obtain a more complete picture of the circulation.

6



August 14, 1969

On August 14, Camille advanced from a tropical disturbance to a tropical storm. Th e first

MRIR observations were taken between 0300 an d 0700 GMT on August 14 (a composite

of three orbits) and Figure 5 shows the 10- to 11 -pm TaB's over a large region surrounding

the center at 180 N, 800 W. Already the cyclone cloud pattern exhibits a definite spiral

configuration. However, there appear to be only two small regions to the east of the

center where deep convection is occurring. The deep convection associated with the

intertropical convergence zone (ITCZ) is seen to extend from the ITCZ toward Camille.

Another band of intense cloudiness about 1000 km north of Camille is connected with a

strong easterly wave approaching Florida. A feeder band with clouds reaching the middle

and upper troposphere is located to the southeast an d the south of the cyclone center.

In the semicircle to the west of the center there is a large region where 10- to 11 -pm

TB's are > 290 K with some as high as 295 K, which, when corrected for the effects of

a tropical atmosphere by adding 6 to 8 K, are very close to the 302- to 304-K sea surface

temperatures reported by ships. Other large areas of > 290-K TBB's are found northeast

and northwest of Camille well away from the cyclone circulation. These high TBB's indi-

cate that the amount and/or the vertical development of low level cumulus was being

suppressed, probably due to subsidence in the lower troposphere.

This method of inferring subsidence has the limitation that once the clouds have been

substantially suppressed or eliminated there is no further information on the strength of

the subsidence. Once the clouds are gone it is not possible to use cloud tracking from

geosynchronous satellite measurements to implement kinematic techniques of calculating

the vertical motion. However, the information from the water vapor channels offers apossible method for inferring subsidence in clear air in the middle and upper troposphere.

This is possible because the emission will come from lower in the atmosphere as the temp-

erature and dewpoint spread increases, and therefore the TBB's will also increase. Theseregions of dry air, characterized by high 6.5- to 7.0-pm TB B 's, are associated either with

areas of subsidence or horizontal advection (Reference 6). Subsidence is indicated when

a local change in clear air TBB to a higher measurement has occurred where the possibility

of the advection of the dry air was small. Even in the advection case, the vertical motion

is almost certain to be small or downward since upward motion would produce clouds or

a more moist atmosphere.

Figures 6 and 7, respectively, depict the 6.5- to 7.0-pm and 20- to 23-pm measurements

for the same time as Figure 5. Th e relatively high 6.5- to 7.0-pim TB, areas generally

coincide with the areas of high TBB's in the 10- to 11-pm channel. However, there are

some regions where this does not occur. Whereas the high 10- to 11-pm TBB's are about

the same for the regions where at least low level subsidence is inferred, the 6.5-to 7.0-Am

T,,'s are not nearly as high west of Camille as to the northeast and northwest of the

cyclone. The strongest upper tropospheric subsidence is inferred northwest of Camille

where there is evidence of 200-mbar horizontal speed convergence (Figure 8) over the

Southern Plains and New Mexico, and an area of relatively high temperatures at 300 mbar

7



00 2 05 2 100 2" 95' 210 2 as ' 0 752 70- 2 65' 60* 55

210 22 210

218

256

2 90 20 2 4 0 2 2O2

-023 2 - 4 22 26

G4 27g

M2050 3M 5

or ite

27127 27 n

_ -210

sesr h stem r ie nKadteccln etri niae ytecrldcos



S95- 3,05* 100,9* 230 Bul, (I n 75- 70' 65o 60o C 5

a235 CT 235 235 235 40

235 235-

225 230 0 35 240

a3 23 5 3

2 245 20240 235 240 235 30'

245 24 230

2455k4 3 2 5/3

24 20 2464 3 20e1 2 , 3 22530 202

245 3 23 2524023 25 5 2 3 245 23 2035*2

235 225 235

3 230 210 24033

'S 230 235 235 -2822 23 210 ) 23 2 5 235 5023

23 302 23g5 220 23

31 35 225 225 22*325 .

235 20 24 238

220 5 230 223 225 -285 21 2205 CT _22 23. 232 5 5 C 2 CM 03

21 23353 228 220 238 0 " 15 2 215 0 13

238 2 5 23, 2 22120

(2335 20 220

~ ~0 2V2-2515 -a 232535 235- 3

C23 22 7 2355

22 2 2 2 2 230 23

25 21 35 22 3 25 3 5 5 _4 1 240 24

218_ 230-e

2423 248 25 235 235 03 0,3 21,24

2402

M 5 235

22rGb

230 Th 20cr

~o 235 2353

6/(951230 20 -24

C24035 235 24.)?~7

k-J 2/4.:R

240ot C' ,

Fiue6 uriae a ilea 00 t 70 M o ugs 4,169 s eite y .- to 0-mTB' ro heM Isesr h sthrsaegvn nKadteccln etri idctdb hecrldcos



255 85- 25 70- 6 C 2"IL1~i~ ~ 00 60':M 55- 2 1 *\ , OO 95* 9 2 2 5 2 2 2 5* 25 6 6

245~~5 ( 2510 255025

255216 20250 2 5 245 24 2 2M

25 5 55 253 26 22 25 5

2003 23 24 245M403025M

5245

255 C250 2% 2 25*2

24so 250 24025

20 is24 0

D0 20 2 245-.02

2524C54 M 0 8

2 52502 5 265

255 55205 245 240 2605

245 2230 0 210 255 23MO3025 0

25 20 220 15 2s0 250

2255

235 2 5 0 15 1242 2 25 2825

24 248b 2 5 2454 25 245255 2 6V

2401

() 25 250-4 2

to - 155 47

Figure. 2452rian Ca il at5 030 to 500Go ut1,16,a eitdby2-t 3p B' rmteM I esr

2 eed

2401

2 250 250 255 21

211 45 kQ _255I S245 250 250

2 1125

-- 2115

Fiue7 uriaeCail a 30 o070G T nAgut1,16, sdpitdby2-to2-umTB'sfomteMRIRsesr

Th sthrsae ie n n heccoe etr sidcae yth icedcos



110- 105. 100- 95- 90 6 85' 80* 75- 70' 65- 60- 55- 40 50-

2 3 35-

-3.9 -54.0, 5 1-3

4f2 -54.4 4

39 53. -57 -53. -51.40 53 -5 7 30-

-52. 3 54. -5z.1

-53.0

5 4 -825

-52.35 1

-53. 0- 9

-52.9

20-

6 -30

-52.0-53.3 I

-4A 53.6-54.6 15 Is

- 5 7 5& 010

3

5 .0

Figure 8. 200-mbar radiosonde data and streamline analysis for 1200 GM T on August 14, 1969.



below the higher level horizontal convergence. Northeast of Camille the inferred subsidence

was most likely produced by a ridge line between the strong southerly flow over Florida and

the northerly flow at Bermuda. Whereas the 10- to 11-pm measurements indicated a continu-

ous band of high TBB measurements at the northern end of this area, the high 6.5-

to 7.0-pAm TBB's shows that the driest air was mostly west and east of the cloud circulation.

The relatively high 6.5- to 7.0-pm TBB's west of the main cloud-shield associated with

Camille were not even high enough to be reasonably certain that cirrus clouds were not pres-

ent. A comparison of high-altitude aircraft cloud photography and Nimbus-3, 6.5- to

7.0-pim TBB's over the tropics has indicated that when the TBB range was 235 to 237 K the

cirrus probability was 40 percent (Reference 7). Thus, the subsidence, if any, west of

Camille was probably in response to the low-level inflow and rising motion within the inner

portion of the circulation.

From the analysis of the aircraft photography, it was found that 6.5- to 7.0-pm TBB's of

< 230 K in the tropics were associated with cirrus clouds. Thus the area of < 230-K Ta B

measurements outlines at least the minimum area of the cirrus shield which covered a much

larger region than the areas of deep convection that were noted in Figure 5. The 200-mbar

chart for 1200 GMT on August 14 (Figure 8) shows the southerly flow from Jamaica

northward to the Carolinas with a break in the data continuity over Cuba. It is important

to know if this southerly flow is continuous and if the upper tropospheric air ejected from

Camille had begun to move northward in response to what appears to be a favorable pattern

for the ventilation of the upper tropospheric air away from the storm. The radiation data

cannot determine the continuity of the wind field but the cirrus shield pattern shows a

minor break between the storm shield and the cirrus associated with the easterly wave

further north. This suggests that a substantial outflow had not begun in that direction.

Essentially the same large-scale features are seen in the 20- to 23-ym measurements

(Figure7)

asin the 6.5- to 7.0-pm data.

Twelve hours later (1500 to 1900 GMT on August 14) Camille had almost reached tropical

storm status. The 10- to 11-pm TBB map (Figure 9) indicates that the cyclone was more

organized with an almost circular cloud shield where the deepest convection was in a

curved line north and west of the storm center. The area of < 234 K TB 's extended over

a much larger region than before. However, despite the areal increase in cloud development,

the highest cloud tops (lowest TBB 's) reached to about the same levels. Equivalent black-

body temperatures of > 290 K had persisted on the western semicircle, reflecting the low

cloudiness suppression noted earlier (the > 300-K measurements northwest of Camille

are caused by strong heating of the land during the daytime).

The 6.5- to 7.0-pm (Figure 10) and 20- to 23-/pm channels (not shown) indicate that the

12-hr change in the inferred weak subsidence region west of Camille's center had expanded

and strengthened slightly. Measurements of 238 to 240 K appeared in contrast to the

235- to 237-K range 12 hours earlier. Based on the aircraft flights, the cirrus proba-

bility had dropped to about 15 percent. Figure 10 also shows that the cirrus north

of the center was continuous and a little thicker (and/or higher) than before, which sug-

gests that the upper tropospheric ventilation had become more effective. Also, the northern

12



10 2 05- 260 100- 20 25, 2 2 6 90o M a 7(r7 65 / 60. 551---

loosoo 2a 2**300:--00 -2 -2020 20 2 0 22510..202 020 2770 2 N

B20 2

-3DOO

31270 -2 50 2M062075

400 2000

26g~ 270 ?9" ... 210n "O - 2

__30 22702

200 C6 I I

370 30 Soo 2702 2g 240 0

277

290 26 M225

20240 290 2e 2070

290 5

2 01

260 -7 M 290 2402

:lo2d 27 29D 2T,,n (

290 260 40

qS290 N 2N21

r

210

210 260 27

290I

2504 25 20 0 So 2 2u 1

2es 2heds

lr0 70 50

290,

290 -- 275

250

Figure 9. Hurricane Camille at 1500 to 1900 GM T on August 14, 1969, as depicted by 10- to 11-Arn TBB's from the MRIRw; sensor. The isotherms are given in K and the cyclone center is indicated by the circled cross.



]1()- 24 0 2oo 2 230 225 2 90. 235 85 1i7570- 65 i

340240

-a a 2

C 22 C240 240 (2..

30-

24 0 5 2 0 20 3 4 3 2 405

244 0 248 2A0 30 Z430 D 230 230

245 C52 2S 242

245 4 45 28 20 240 2023 2 2

2n 4523 225 5M

235 (:145 2323 2 40 ~ 2 5 5 ( 241 225 20 2352 25 2 2

-23 I 2 2 30 2

2 23 2 48 25

U2 22 2 4 5 21 2223 25 C 235 3 2 425 240 25 24.15

.23 255 2M20

2425 - 230325 3 2O5252X2 25 25 0 202

235 022 23 25 222..52

19 5230 2M 235D520

2220235

2 5 - 2 30 23521 21 2 5 3 2 52 0 4

2n 0 230 C 240

23 23 235 238 225 235 23 j P 235 2 40 15

2 22 5 232Z 2x0 2400

VO 22 25--235 (::: 35 25 3 2 0 24024

F 0 r C a7T's5 25 Z2 2

'23.nhe im 3i0 2n 230C2

C2 235 23 3522

C 23 I Q 20 30 30 23235 25 233240D

3523 05 10C -31C25 0 0 40 5

238 22 LCZ 20 20 2

C24 20 7 46 240CC23 20 24

235 225 1 240') o

Figure 10. Hurricane Camille at 1500 to 1900 GM T on August 14, 1969, as depicted by 6.5- to 7.0-um TBB's from the MR I

sensor. The isotherms are given in K and the cyclone center is indicated by the circled cross.



100- 95' 20*5- NOM f 75 - 2-70- 65 00i an 2 3

27 252-1

052* 27 M@ M 2

20) N 270 290

2@ 2@0

25 M C) 2W 40 7

nom no U

2N 27 7 2 5 2 28

2~ 2 M

230 2W 2M 24 5 2022--

soo se sot a2,

2M 2 2N

2202425

2 7 9 2 22 2 2N 2

2gg27 25080 2

M~ 230

M4 10*

2N M 25020 so 270 2750

272 200

M5 270 2 288

2 g2 20230 270

1,217 231 22 20 230

D 248 210 M M

C M 0 70260

220 M5 702

2182" Y ~ i~ ~"

Figure~~~~~ 240 248ian 2aml0t 0400 25080 0MnAgs258160a-eice 1-to1-mTBs rmteM

s0 20 20 2dero s

280 2 250 0 23270 10 40 - OD

r* 240

"270 70 20 21N-28 M s

~~250

no 2w 0

2 ZU 290

Figue 1. Huricne amile040 to080 GMTon ugut 15 199, s deictd by10-to 1 -A TB' sfromtheMRIsesr.Te sthrsgv n nK n teccln cneri idcae b hecrce co s



110o 105* 100* 95* 90* 85o 80* 75* 70* 65* 60* 55*

35* ::, no 2 . 35PN

otl~ ,e * *

30o 30*

25* "o 2a " 25

.c- -I Ob

ifoasC

2.

20 * 2 2 20 *

6I a

2i .3 o~ss as

t'g'n

15* 24 15

IO* --_ =* ' , so E- IO*~50

IO5W 100* 95* 90* 85* 800 75* 70* 65* 1O*W

Figure 12. Surface data and streamline analysis fo r 1200 GMT on August 15. 1969.



110 21250 105' 2100 2 95* 2 5/90 5. 80. 30 - 75. 255 1 55- 35*

235235

5230

245 24

240 245 240 U 24030

40 22355245

22

C 245 22050 235 O 2s 2 352s a

2253 23552 230 r5

20240 00 225 235

D 2 40 U 3 235

2403 24 25 230 235 35

30 235 230 225 22352220 2 5O

20 230 232 2152 5 25 5 3 25 22

22 20a3 240 03

23023 2152

(25C2202 25 21 225

242 25320

0~4 220 232? 5 23330c

235 0 5O2

23 20120 0230 230T5

230

2 2 0 230 235 2502 1

3 235 23525 225 235

25 22 25 5 2 230 202520(3

U35 235 r35- 23022 2102120 235 C230

235 D523 230 220 225 238

235 235 235 223522 8 2352 25 25 220 2

230 23Z - 231 2 5 5 240

2235

4,

-A~

Figure 13 . Hurricane Camille at 0400 to 0800 GM T on August 15 , 1969, as depicted by 6.5- to 7.0-pm TBB' s from the MRIR




110, 5 0 7' 2b

no j7 I~ M V2

7 z

Mni i, M Q Mr-

AID 31~

%19 (-21.

11eu, t2

t. , I

2V "0 2M,0

2 . 2 0 2

270 C:7 II 27

Ma

211

M 2 1 U196

Figure 14 . Hurricane Camille at '1400 to 1800 GM T on August '15, '1969, as depicted by 10 - toll-Jim TBB' s from the MRIR

sensor. The isotherms are given in K and the cyclone center is indicated by the circled cross. Superimposed on the analysis

of the radiation data is the surface analysis and cloud motion vectors derived from low clouds tracked from the ATS-3 satellite

\e images. The cloud motion directions are indicated by the arrows and the speeds are in knots.



indicative of the continued cloud development in conjunction with the low-level confluence

that was shown by the low cloud motions. The band of cloudiness was continuous from

the ITCZ to Camille. The low cloud motions that were located between the two principal

developing cloud bands were generally a few knots higher than to the northeast of the

bands, even though the pressure gradient from conventional analysis was about the same

inside and outside the bands.

The presence of a surface depression along the leading edge of a band would increase the

pressure gradient between the band and the environment and thus increase the winds just

ahead of the leading edge. Tatehira (Reference 9) has indicated that a depression of about

1 mbar occurred with the passage of the leading edge of an outer spiral band of a typhoon.

Th e band that is south-southeast of the center of Camille appears to be the most active ofthe two bands. If this is true, the greatest pressure gradient would be between an d parallelto the bands (along the east side of the strongest band). That is where the area of consis-

tently high velocities (26 to 29 kn) are located. The band formation will also increase the

low-level convergence and therefore maintain the vigorous cloud growth within the bands.

There is a possibility that the surface pressure is being lowered along the band by tropo-

spheric warming caused by latent heat release. Gentry (Reference 10) has found that the

temperatures are lower in hurricane bands at a considerable distance from the eye wall than

the immediate surrounding environment du e to entrainment. The average width of the

precipitating bands in his investigation was around 20 to 40 km . It is not possible to ac-

curately estimate the precipitating width of the strongest band south-southeast of the centerof Camille, but the width of extensive cloudiness averages more than 350 km. This suggests

that the precipitating width is considerably greater than 20 to 40 km. Under these con-ditions entrainment would probably have a small effect. Whether the effect is sufficient

to produce a warm core band cannot be proven with these data. Simultaneous aircraftflights making temperature measurements would be necessary.

Outflow from the strongest cluster (as measured by the lowest TBB's along the most active

band centered at 130 N, 780 W) was probably augmenting the velocity of two clouds

(28 to 29 kn) just to the north of the cluster by about 10 kn. A cloud motion further

to the north (19 kn) was most likely more representative of ambient conditions in the

band.

The development of the clouds in the confluence zone could be an important factor for amore rapid transport of moist air into the inner circulation of Camille. Conventional data

indicate that the winds between the surface and 850 mbar had about the same radial com-

ponent toward the stormbetween 0000 GM T on August 15 and 0000 GM T on August 16.The low-level winds were as high as 35 kn transporting moist air at 850 mbar with a large

radial component 350 km southeast of the center at 0000 GMT on August 16, preceded by

30-kn and 35-kn winds at similar locations 12 and 24 hours earlier. Thus, with a similarenvironment for 24 hours the cloud development was able to continue.

20



A small area of TBB's > 290 K was located west of the storm, which indicated some suppres-

sion of low-level cloud development. To the east was a larger area that was most likely

associated with upper horizontal convergence produced by the outflow, where the strength

and position changed little on the 200 mbar analysis between 1200 GM T on August 15 and

0000 GM T on August 16.

North of the cyclone center, the cirrus shield persisted as seen in the 6.5- to 7.0-pm

measurements (Figure 15), indicating that ventilation was still present. The ventilation isclearly shown by the middle and high cloud motions from ATS-3 where a strong southerly

current exists at the cloud levels for at least 1100 km. Also, the cloud motions indicate

local outflow from one of the clusters in the band southeast of Camille that was assumed

to be the most active from the analysis of the 10- to 11-pm measurements. The local out-

flow that is typical of a vigorous cluster gives further support to the earlier assumption

regarding the relative strength of the two bands.

The 200-mbar chart for 0000 GM T on August 16 (Figure 16) indicates that the flow oversouthern Florida ha d changed from southerly to westerly. This change was most likely

an expansion of the outflow area of Camille. East of the cirrus there is a well-defined

ridge line demarcated by 6.5- to 7.0-/pm T B's as high as 244 K and the cloud motions.

The 200-mbar analysis cannot confirm the presence of the line since the high TBB's were

east of Florida. It is likely that some of the cirrus ejected by the outflow dissipated in the

dry air associated with the ridge line. West of Camille the relatively high TB area ha dweakened an d moved little with respect to the center, and the upper horizontal conver-gence as shown by the cloud motions is coincident with the highest TB 's.

A chart was prepared combining the measurements from the two water-vapor channels and

the reflectance channel to more clearly delineate the regions of most intense cloudiness and

the areas where the strongest subsidence could be inferred. Th e mapped values were gen-erated from the expression T, + T4 + (100 - R), where Ti and T4 are the channel 1 an d 4

T ,'s respectively, and R is the normalized reflectance. Thus, a dry atmosphere over the

ocean would be characterized by relatively high T, 's an d a low reflectance. For example,if T, = 24 5 K, T4 = 265 K, and R = 10, then the mapped value would be 600. At the other

extreme, with deep convection, the same set of measurements could be T, = 210 K,T4 = 210 K, an d R = 70 , and the mapped value would be 450.

Figure 17 depicts the combined values for the daylight satellite observations on August 15.

There are two regions where the values are > 580 and these are about where the two areas

of maximum TaB's were seen in Figure 15. Values of > 600 occur only over a very limitedarea in the western region and in the center of the ridge line in the eastern region. Thischart indicates that the air was probably drier throughout the tropospheric column in the

eastern region. This same result could have been deduced from the examination of the data

from each of the three channels, but the new chart presents the information in a more

convenient form. Near the storm center values of < 440 indicate the two portions within

the cloud shield where the most intense cloudiness was located. These values are compar-

able to those observed in the most intense portion of the ITCZ on the same day.

21



110- 105- 109- 5- 10. 85- _/ so, 7 To- 55 *6~

t2.

,DD

45 r2 M CT M Q 2A 2B 23

211

20 a 3

2 20

3.

2 0 2 8 30 . 302 3 23 2

Fiue1.Hu 0 0 .1 n0 o10 MTo uut15 99 s eitdby65 o70-mTBSfrmteMI

sesr.Tei 25 2 35 -40 14 24cylnecneridctd yth icldcos.Speipoe n h nayi

ofte aiain aacod oio etosdeiedfomhghad idl lod rakd rm T-3s3Meiagsan h

stremlie aalyissedon he loudmotonsand200-barrawns.The lou moiondiretios ad seedsareindcatd

30 C90 23 3 2 .2. 240re4



115- 110. 105- 100. 95* 90* 8 s*Bo* 75- 70- 65* 60. 55'

U

40-

51.9 9 '4,47.3-545/ -55, O

-52.5-52.8 04 54.0 4

1.8 -56.0 -54.3U

5 4

51.8 6_5-51 - 3.

-53.9

4 425*

-54.2 -524 -53. .4

3

-55.3

-5120*

1 - .4

-527 -54.0 -551 -54.9

-54.0 - 5 0j 184~ 0 -55.0

.15

,/ } 55.9

-53.9 -54.9

9

Figure 16 . 200-mbar radiosonde data and streamline analyses for 0000 GM T on August 16, 1969.



110 6 105 100 95* 90* 0 85 7 5 7 0 60* 5 2Goo- 550 5

50Q5 0 50 525 5 565 5560 to

540 500

560

Goo

582 50 a0

58 540ase e

80 8 680 5 4 02

544 50

55,

Figure 17. Hurricane Camille at 1400 to 1800 GMT on August 15, 1969, as depicted by the combination of the B

measurements (in K) from the two water vapor channels and the normalized reflectance measurements from therefectance

channel (spectral albedo in pet-cent) of the MRIR sensor. The cyclone center is snown by the circled cross.

55 Soor i

46A0 5208 $4 5 So54

so 556 Sool

580104

5 So) 5 2 gg 5401

so 541 15,;

52 5854 g 004

520 o.

41 0 0.2020 og 2

F i u r 1 70 4 0ian 4 0l 41 4 0 ts80 G T oouu t 5@9 9 s d p c e y h o b n t o f t e T

measuement 0i 5)fow ae ao hnesadtenraie elcac esrmnsfo h elcac

4h a nn (m-ta0iei

inn m n0 h R R s n o .T ec con e tri n w yt ecrld cos



August 16, 1969

Twelve hours later, Camille had probably begun the most dynamic deepening phase of its

life cycle. The changes in the structure of Camille as seen in satellite observations also

reach a climax. The 10- to 11-/pm TBB's (Figure 18) show thatthe

sizeof the

active cloudi-ness, as defined by the 235-K isotherm, had been reduced by about a factor of three from

the last satellite observation. Therefore, during the early portion of rapid deepening, the

cloud canopy contracted sharply. The cloud bands in the converging moist air southeast

of the center were still strong and extended 1100 km from the center. The outer limit

of the cloud bands was almost exactly the distance from Camille's center where the sur-

face streamlines began to converge at 1200 GMT on August 16 (Figure 19). There were no

rawinsonde measurements close to the center but the surface reports show the continued

inflow where the wind speeds were 20 to 30 kn. West of the center, the area covered by

T BB > 290 K had expanded considerably in 12 hours. This large expansion represents

a substantial increase in the region where the cloudiness was being suppressed. This would

indicate that the subsidence surrounding Camille over a large area was beginning to affect

the cloudiness in response to the increased vigor of the inner circulation.

The 6.5- to 7.0-pm and 20- to 23-Mm measurements (Figures 20 and 21) also indicate the

subsidence increase west an d northwest of the center at higher levels. Twelve hours earlier

there had been a weakening band of high TBB's oriented northwest-southeast that was most

likely advected from north of the storm. The high TBB's at this time encircle the center in

the western semicircle, which suggests that they were produced by the hurricane. They

cover approximately the same area as that covered by the > 290-K isotherm in Figure 18.

It is interesting that the time lag between the increased vertical motion in the inner portion

of Camille and the subsidence appeared to be small. Th e cyclone center had moved off the

northern Cuban coast for only about 4 hours. A more exact assessment of the lag would

require geosynchronous satellite measurements.

By 1500 to 1900 GM T on August 16, Camille had become a very severe hurricane. At

1800 GM T a U.S.. Air Force reconnaissance aircraft measured a central pressure of

908 mbar. Therefore, Camille deepened about 60 mbar in 24 hours. The 10- to 11-Mm

measurements (Figure 22 ) show that the intense portion of the Camille cloud canopy

(TBB < 235 K) had expanded by about a factor of four in 12 hours. There were two areas

of great activity, on e near the center with TBB's between 200 an d 204 K (the lowest record-

ed during the entire life cycle) and another of 205 to 20 9 K imbedded in the broad band

southeast of the center. The low TBB measurements near the eye were almost perfectly

centered, thus suggesting an intense circular wall cloud. This is verified by the concurrent

high resolution (3 km) image dissector camera system (IDCS) image (Figure 23) in

which the eye is clearly seen and the wall cloud is like an inner ring surrounded by the

rest of the cloud system that is separated by a ring that ha d a lower brightness than the

clouds forming the rest of the canopy or the wall cloud. These features suggest that the

cloud tops within the wall cloud ring had reached a higher level than the surrounding

clouds an d that the air was descending immediately once it moved outside of the wall

25



115 110- 105- 100- 95 20590 20 20 2 0 5 75* 7'221 5 60250 250 M/ 2 00 220 210 20200

220M 210

20 2 290

so 290 270 230*

210

248 217

255 200~221g 270

S270' 20-

2201 2 M "

Figure 18i Cmi a 0 to 0A 1 9 a 0 t by 10

senso.

2 5 2 0 2 218 2 40 25029

2502M 290

2M -M2 22 27 20 5*

2280: 2 27M 29 5

270 p 240 4

no 210 270 280 2"

25 0 M27 20M 0

I . /- 48 2o '27

IU- 271 M 2D 220 2 na 702 18 H e C e 20t,

c0

sensor. The4 ic th

-20 71-210 22

20 250 C2,i

250 M 2 212. 260L

ZW 28 2 20' C 0

Figure 18. Hurricane Camille at 0300 to 0700 GMT on August 16, 1969, as depicted by 10- to 11 -,m T,,IS from the MRIRsensor. The isotherms are given in K and the cyclone center is indicated by the circled cross.



35 1100 1050 100' 95* 90* 850 800 75* 700 65" 600 55*W

a '35*

11 23 2: 35* N

i9 24

to t

30* 2. 2 2

3o*2. 2 2

31 3/~~ 1 0

225

6)6

20**

2. 2I. 23

2102

15* 2; 6, 24 2as s 2.0'.&150

22 212 a50 ~ ~~~ zel ; o3

50* Nt5.-N

105* 100 950 9Q0 850 800 750 70. 650 600

W

Fgr 1

F g r 19.i n. A

20'21

2 2.

10*

50 YNo

105*00 950 90 85 0 800750 70065* 60*

Fiue1.Srfc aaad temieanlssfr120GTo ug s 6 99



115- 110* - 105* 100- 95* 90-238 8 230 23080 75- 230 70- 20V 25 60- 55*

23D S 230

035 235 23035 3 3 23235

25 V 2 1 28 2 53

23 8 3 8 28202353530

2 8 255 230 2323 3 5 235 2 5

2 82 20 235\) \\ 220 235

2458 U5 230202M 30230 235 235 235

n as as 221ss 2no 2ae S 23 235 23 A3 2 235 235

22' 225 23 :< /24. 13 230 3 Q!

305 402 5 2 2523

nol 25 2n 231

anne,, 525 235"o

23 23 13 225230 3U 35

23 230 22 s

21 230 215 Cus23

30g 225 235 2525

Fiue20 uriaeCail 21 0230 Is 0700 GMT 23 25us 2306adeitdb6-o70-mT sfrmheMis-

nded231 23 225 2.. , 220 20 15230 C20 G23

225- 31 21 230 230 25 0 M~,\/I !

22 , 25 22 2300 210 21 2 22V23 1 A2

220 22 21.1230

230_15 220 5

23822J 20 2 20 2

225 5 20Il

Q 2 322523 23 2

Figure 20. Hurricane Camille at 0300 to 0700 GM T on August 16, 1969, as depicted by 6.5- to 7.0-prn T, , s fror the M R I




115W 110W 105°W 100°W 95°W 90°W 85°W 80°W 75°W 70°W 65W 60"W 55°W

-3 / 35ON

5255N

20go

cz 3 Q. Q.~ 0~

ca cd- a. 10"N iOON~B

Figure 21. Hurricane Camille at 0300 to 0700 GMT on August 16,1969, s depicted by 20- o 23-#m TBB's from th e M RIR sensor.

The isotherms are given in and th e cyclone center is ndicated by th e circled cross.

26o Q

230 255"

Po 2 0"

24o w 2a

~a, m 15oN

2151-, 235 0 Gh m i,

Figure 2l. Hurricane Camille at 0300 to 0700 GMT on August '16, 1969, as depicted by 20- to,23-~m T,, s from the MRIR sensor.5: The isotherms are given in Kand the cyclone center is indicated by the circled cross.



s a 110' 05 = MO3 10 95-() M 9 2M0 28 1m 70 - 65' so-

255*

70 7O

WGH35

31

Q Goo 30WV

0 30nd

s r s ei cros S

1131

1861717n

of the radiation data is the surface analysis and cloud motion vectors derived from low clouds tracked from the ATS-3 satellite

imanes. The cloud motion directions are indicated by the arrows and the soeeds are in knots.



> 290 K occurred over the Gulf of Mexico. Thus, the suppression of cloudiness (probably

low-level) had become more widespread and had intensified in some of the areas to the

point where all cloudiness had probably disappeared. After correction for the atmosphere,a 295 K TBB was within 2 K of the sea surface temperature. The cloud motions showed

evidence of low-level divergence over portions of the area where TB B > 290 K. Also, the0.2- to 4.0-Mm reflectance (Figure 24) of < 10 percent covered nearly the entire Gulf west

of the Yucatan Peninsula.

The water vapor channels exhibited some remarkable changes (Figures 25 and 26). In the

entire western semicircle over a radial width of about 50, higher TBB 's were found with

the maximum measurements of 245 K in the 6.5- to 7.0-gm channel northwest of the

center, and 268 K in the 20- to 23-Am channel further away and west of the eye. Massive

subsidence can be inferred as the increased circulation within Camille became reflected

in the large ejection of air to the periphery. This is particularly impressive when the

evidence from the conventional data indicates that the 200-mbar flow east of the storm

also increased between 0000 GMT and 1200 GM T on August 16 and did not diminish in

the following 12 hours. The region of maximum upper convergence appears to be alonga line 50 northwest of the center where the 6.5- to 7.0-gim TBB's were 245 K. Air wasapparently flowing northwestward to converge with the eastward moving air around the

base of the trough over the central United States. A combination of the 200-mbar rawin-sonde reports for 1200 GM T and the high an d middle cloud motions from AT S supports

that interpretation of the radiation map (Figure 25). The analyzed line of upper horizontal

convergence was within 100 km of the area of 245 K TBB's northwest of the cyclone center.

Because most of the emission in the 6.5- to 7.0-gm channel came from the 300- to

500-mbar layer, which was just below the region of probable maximum horizontal conver-gence at about 200 mbar, the subsiding air in this layer was most likely not fa r from the

position where the air initially began its descent. The location and shape of the maximum

20- to 23-gAm TBB's suggests that the air did not descend vertically. The maximum descentin the 400- to 700-mbar layer is inferred to be further west an d over a larger region. With

the lack of conventional measurements it was not possible to correlate the 12- or 24-hr,6.5- to 7.0-pm an d 20- to 23-gm TBB changes with computed vertical motions. However,these TBB changes should be related to the relative difference between the vertical motion

in the two layers. Over the area 6 to 80 W of the center, where clouds were probably

absent for the 24-hr interval, the A TBB's in the 6.5- to 7.0-gm and 20- to 23-gim channels

were about 60 an d 100, respectively. Thus, an estimate of the ratio of the subsidencestrength would be 10/6, with the greatest sinking in the 400- to 700-mbar layer. Since the

maximum vertical motion usually occurs at the level of nondivergence and this level is

about 600 mbar, the result was not surprising.

The three-channel combination chart is shown in Figure 27. West of Camille the large-

scale subsidence had produced widespread values that were at or slightly below 600,

in contrast to a small area 24 hours earlier. A value of 418 near the center was the mini-mum for Camille during its life cycle and was close to the minimum observed for any cloud

mass anywhere on the charts during the Camille period.

32



230

q

2o5*

10 N

1 20'~~2-~~\) \\~~

Figure 24. Hurricane Camille at 1500 to 1900 GM T on August 16, 1969, as depicted by 0.2- to 4.0-pem normalized

reflectance measurements from the MRIR sensor. Th e isolines are spectral albedo in percent and the cyclone center

w is indicated by the circled cross.



5 105 l00 95 25 no 90* As . 75 70- 65" 60

2 ; 30

c~ 2M

e\Ci~ jcB3 \o25*

AUG 16/ 2969 040

theI sam mn as in Figure2./

mMM2

2350 20

C34~ 0- 2 40/

AUG~ 126

e3 I LL--~JRno 5 L 4 35i~~Bd

2 2 15 ) 341-

OT A 1816'),"

Figure 25 . Hurricane Camille at 1500 to 1900 GM T on August 16, 1969, as depicted by 6.5- to 7.0-m TBB's from the MRI Rsensor. The isotherms are given in K and the cyclone center is indicated by the circled cross. Superimposed on the analysis

of the radiation data are cloud motion vectors derived from high and middle clouds tracked from ATS-3 satellite images an d the

streamline analysis based on the cloud motions and 200-mbar rawins. Th e cloud motion directions and speeds are indicated in

the same manner as in Fiaure 14.



110- 105- 100- 95- no 90- 85.2. , 80- 75- 70- 65- 60-

_260 C 26a 2

Cf~C

D2. 26 26 "o30

20 650i~J V

'i[o

2.'-'( no 1-2 "

24oi

- 21 21 2. S

2. 2

255(

(21) 0 36I C,

-2 255CI GJC301~ ,-2.

w Fiure26.Hurican Caill at1500to 900GMTon Agus 16 199. a deictd b 20-to 3-Em TBIS romth2.0I

vl snsor Th isoherm ar givn i K ad th cylonecentr i indcatd bytheo -240ss



" 10 r 100. 95 o / sS 75' "5 So

I m U " 2. .

on I HUm

albedo in percent) of the MRI R sensor.



August 17, 1969

Thirteen hours later (0400 to 0800 GMT) Camille gave the appearance in the satellite

data of a storm that had weakened. There were no reconnaissance ey e penetrations near

the satellite observation time, but the aircraft monitoring the storm on radar and making

wind estimates outside the most active region reported that the cyclone appeared to main-

tain its peak strength. The low 10- to 11-Mm TBB's (Figure 28) near the center covered asharply reduced area from 12 hours before. Southeast of the center there was still evidence

of a low-level moisture inflow where the TBB's associated with the broad cloud bands were

generally not as low as 12 hours earlier.

The 6.5- to 7.0-/am measurements (Figure 29) show that a new surge of cirrus had appar-

ently moved northeastward in response to southwesterly flow over western Florida, Ala-

bama, an d Mississippi. The area of dry air west of the center had become smaller and the

TBB's were not as high as 12 hours before in both the 6.5- to 7.0-ym and 20- to 23-1m

measurements (Figure 30), which indicated that there had been similar reduction in the size

of the relatively clear area. Thus, the subsidence had most likely weakened.

It is unfortunate that there were no reconnaissance aircraft measurements of central pres-

sure taken at the time of the satellite observations early on August 17. The strong surge in

the conversion from potential to kinetic energy that accompanied the rapid deepening

could have surpassed an equilibrium state during the da y on August 16. Then, 12 hours

later in response to the equilibrium overshoot, the reverse was observed where the circula-

tion was below the equilibrium state. The details of any possible oscillations in energy level

could be specified more definitively with frequent measurements in the same spectral

intervals from a geosynchronous satellite.

In the last Nimbus satellite view of Camille before it crossed the U.S. coastline (1500 to

1800 GM T on August 17), the 10- to 11-pm TBB's (Figure 31) indicated that the cyclone

cloud shield had again expanded both horizontally and vertically, but not to the extent

that it had near midday on August 16. Perhaps equilibrium had now been reached. Also,

the cloud bands southeast of the center had become narrower than before. This suggests

that the rapid inflow of high moisture content air from that quadrant was diminishing.

Camille had now increased its forward movement to 11 to 12 kn in the same direction

as the flow in the southeast quadrant. Thus, even with the same windspeeds, the net in-

flow would be smaller. There appeared to be a cloud band almost due south of the center

that was separated from the other cloudiness. Streamline analyses show that the maximum

surface an d 850-mbar convergence was the greatest in this region. Since the distance

between the ITCZ and Camille was increasing, the entrainment of drier air was becoming

a greater possibility, further reducing the chances that the influx of moisture from that

source could be continued at the previous rate.

Despite these circumstances, Camille was still a severe hurricane (minimum central pressure

of 905 mbar). A probable major contributing factor was that the sea surface temperatures

surrounding the storm were 2 to 3 standard deviations above normal (304 to 305 K).

37



I15 0O00' BS* 801 09 75* . 0.- I 70' 65-

a260"" -2 288 2 2100 621 270

227

28 M 80 --- 2.5202

2M20 'm002

210 M2

22M0

214 258290 200

maaNO0 0 2 25

22 250

:5 2 50

so 2 0

200 270 2M MM20507

2502405 2 0 27 29 21

2;2D 2302809 270,

25 s 2 5S A ---260 2214

24 1 21 1 23 2 6%23 2 2. 250- 250 28,2

2 4g 210 240 2 70

24

240 2 0 - 2 2 0 5 0 4

Figur 2840rcaeCml00 2o 080 G M onAs 1 7 199 27e.c.e b21-to1-mT sfo/teM

( nan 240 DOr 2.

2 01-

240 700 - -230 80 20 25 25O7 26 27

258 26010 U so Us' - 02. 260 29

23026 90

2 MM 0026 210 ?o0

27 IO i" ,7:1C?,

Fiue2.HriaeCmil t00 o00 M o uut1,16,a dpce y1-t 1-UmTB'sfo h Rsesr h sthrsaegvn nKadteccln etri idctdb hecrldcos



11 110' 2 -105 -100 3-'95-3 0 U35o *s*

S50 235 0 40

255 2 Q50 235 230

U 235

240 224 2 2 4035 235 023 2 2

2352524J 235 252Q

224 ) 2. .23240 240 40230 240 230:

235 240 24 24 - 25 2 '240 2.s* 2

2450 225 24 2440 2

2330

245 220

24 25 0 24 230 C)2025 2 245 20akO

245 245

2 55 352 22350 22240 2452 0 223. 240

2411 240: 225 22 2M 4

n3 220 235n 2

Fiur0 29. 2Hricn 2e 2 G2amle 245 2 f t 0sesr The i 2i3 2

242 5 245 225225 23

5~~i2330302

2022225 2 550025

235 22530 0 220 5 2 40 225 2 2 20 23 5 2302

202 2250 230 225240 225 C I3 230 225 0223 2 15 2 35

25 245 C2m 225 2 z 2 35 1

- 230 23 225 '20

2252524 231 22 45 2213 0 3 230 25 g 21 3

28225 220 Q 2 2 2 r) 202 2 5 21 2510

22522 230 2022 24O 125"2 C 235 22 25'

22230 2t--'

220 5 22 215 25 2 511

23Q 20 2 20 330 2302300 2 120 230 20 02230 225 22 25 22303 252

22 2 )e

C235 235D 225 C230 32 r 235

220 220 230 25 2153 13N

235 2 20 U 22

251

235D 2511

Figure 29 . Hurricane Camille at 0400 to 0800 GMT on August 17, 1969, as depicted by 6.5- to 7.0-jum T, 's from the MRIR




2" 25 M 1245 40-

N,5 o

as~l On

T a0 255

24 245 25

45 " 225M 24 i2

5M 2 02 M 24 2525 2

Ctr'.

2iur 3 Hurcn i01t

2 enor T 23m =vKc I

Fiue3. urcn ZU 27 2M 2o0GT nAgs 1.16, sdpctdb 0 t 3~ Ta"fomteMI

senor.Theisohems re ive i K nd he yconecener s ndiate byth cicle45s



C2W

noG

310 310 no -o~ ?OCf>_C1 C

00

00 0.

AUG 1, 19690

Figure 31 . Hurricane Camille at 1500 to 1800 GM T on August 17, 1969, as depicted by 10- to 1-m T 's from the MRIR

enar. Th e imotherrne ame niun in Kand cyvone center is indicated by the circled cross.



The 6.5- to 7.0-pm measurements (Figure 32) indicate that dry ai r produced by the trough

to the north of Camille had moved southward to the northwestern Gulf of Mexico, South-

ern Texas, and Northeastern Mexico. Since the advected dry air was close to Camille, it

was not possible to assess the contribution of any possible subsidence caused by Camille

to explain the processes that caused the dry atmosphere. However, some subsidence stillshould have been present because the line of upper horizontal convergence had persistedwest of the center as determined from the 200-mbar rawinsondes and the high and middle

cloud motions.

August 18, 1969

At about 0430 GMT, Camille crossed the Mississippi coast with maximum wind speeds

estimated from damage surveys at 175 kn, an d was observed by Nimbus-3 1 hour later.

There ha d been a sizeable 12-hr decrease in the intense portion of the cloud area as

seen in the 10- to 11-pm measurements (Figure 33), although the maximum cloud-top

heights reached are about the same level near the eye. The separation distance between the

cyclone an d the ITCZ had continued to increase, and the cloud patterns suggest that the

main movement of the moist ai r tongue was passing to the east of Camille with a cloudband south of the center that was weaker than 12 hours before. The surface air southeast

of the center was still flowing almost directly towards the center, but at 850 mbar the

flow ha d a smaller component to the west and this air was being carried east of the main

portion of Camille. The principal change in the 6.5- to 7.0-pm measurements (Figure 34)

had been that the T ,'s were lower in the dry air advected southwest of the center. At

the same time, the 6.5- to 7.0-pm TBB's associated with the dry air northwest of Camille

had remained about the same. Th e middle and high cloud motions on August 17 and the

0000 GM T 200-mbar chart indicated that upper tropospheric horizontal convergenceappeared to be increasing in this area as the southwest flow along the outer boundary of

the outflow moved northward into an area of northwest flow.

Over the United States

Interactionwith the Westerlies

Camille rapidly weakened as it moved inland. By 1200 GM T the storm had reached Jack-

son, Mississippi, with maximum wind gusts of about minimum hurricane force. Nimbus-3observed the cyclone near midday when Camille was at tropical storm strength. Figures

35 through 38 show the 10- to 11-pm, 0.2- to 4.0-pm, 6.5- to 7.0-pm, and 20- to 23-pm

channel measurements, respectively. There had been a sharp decrease in the maximum

cloud-top levels (Figure 35). The minimum 10- to 11-im TBB is 225 K, and assuming that

the clouds were opaque and filled the field of view of the radiometer, this temperature is

located at 200 mbar on the 1200 GMT, Jackson, Mississippi, sounding. Twelve hours ear-

lier the minimum cloud top TB,'s were 205 to 209 K, which is near 150 mbar on both the

0000 GMT, August 18 soundings at Jackson an d Shreveport, Louisiana. The 2-km drop

42



1AU0. 1. 10517,969 952

350

32 2ae 4

sensor. The it re v iKa te c n c t i i d h c i

22 2

40 2 25

t m20

301

7 630 Ga 30

45 12 30 00 24

Fiue3.Hrian2 3 2t100t0GTo Ags 7 99 a eitdb .5 o70~mTBI rmteMR

sesreiohem regvninKadth yloe25P8 2ctdbytecrce ros uermoedo h aayi

ofteraito dt r 20 50 M 3 Z4ivdfomhg ndmdl cod takdfrmA S stlit mgs n h45 + 26 24 H 3 0 Z4 2odmin ad20-br ain.Te ludm tondreton n sedsaeiniatdi

thesa0 3 3 122 20ur4



120- 115- YIOy 105- ' -1270J 95 260. O'd2m aV*0 To70 20 \-720' 2027Z10 7 0 0 Z27

72500 2 0 270 2 2 7

40~ 23 26 '70 C

2102

-270 T 8 250 260

2020 . 2 2 027 2 0

2 28o 2 0 0 2 60

260 2.

270 26070

210 24 20 0

2,, 20 2 70

22 2 40

2603

2 0 2 27 8 2 26

7* 270 25 20

2M 280 22. 270

:D0 240 23 60

2200 21 270 25

2262

\ 270 29 n

2365 2901022g 3 G 8

?280

2 00 224 0

290 24 220

Figure 33. Hurricane Camille at 0400 to 0600 GM T on August 18, 1969, as depicted by 10- to 11-m Tu'Is from the MRIRsensor. The isotherms are given in K and the cyclone center is indicated by the circled cross.



120' 115- 110- IO5 o00 95. 2403 8 5 *0o* 2.

K5 23 1

148 2o 5 23 2 o

235230/

225 30

2. ]1

235 23 2a2

2 0 .23

240230 23

2 D -5- ' N 3 1 3

2x~~~ j 23 20451 20

245

2 so 1za 244

20 02 s

24

23 2 5 25 2 2254

30

2235

2 0 2 0 4014

I\ U 220 245 V

245 2 230 25

P Figure 34. Hurricane Camille at 0400 to 0600 GM T on August '18, 1969, as depicted by 6.5-to 7.0-GemT 'Is rom the MRIRvl sensor. Th e isotherms are aiven in K and cyclone center is indicated by the circled cross.



125" 120- 115" 110' 105' 100* 95* 90* 85 o* 75' 70" 65' 60 55

so28 280 255

S2 0 ' 2 2 5w 280

300 3M 2

208 2"0 2B200 240

280 200 2

so 0(20 6 5 2 0 "2 025

3 1. 0210

0

700 201

30020 2o20020

2

270 270S

b21 so*

0a

2 0

so'

GO 21025*

5'290

22400 2 0 290 29

2.0 1

20

2n2

M3

2 0 224

so300 2 530 40 40 22 20 15o

4 260

Figure 35. Tropical storm Camille at 1600 to 1900 GM T on August 18 , 1969, as depicted by 10- to 11-ym TBB'S from the MRIR




125. 120. 115. 110. 105* 100. 95* 90, 85. 80, 75' 70' 6. 60, 55

45'

1o 25

Q.~ID,-

30C. 40 0 00 0 1i40

00 30 a3 15 W 0

300 400 . to 21303

4005 0e 30/

u 38oreoin

20r

20~ ~c,

2

0.

(10, 20

1 20 1. 1

0 15c

24

1 C,

13 10 3 20.oss



125' 120 115" 110. 105' 100. 95 90 85e

0 750 70 651 60 55*

2Bo0

2n

236 3

30as30

30 0

0

305

Z2-v,l 3.3/6,6 2-- 2;

',13 4 3

Figure 37. Tropical storm Camille at 1600Oto 1900 GMT on August 18, 1969, s depicted by 6.5-to 7.0-pm Ta's from the MRIR

sensor. The isotherms are given in and th e cyclone center is ndicated by the circled cross. Superimposed on the analysis of the

radiation data are cloud motion vectors derived from high and middle clouds tracked from ATS-3 satellite images and the stream-

line analysis based on th e cloud motions and 200-mber rawins. The cloud motion directions and speeds are indicated in hesame manner as in igure 14.

/ 30,3

A 3. 4

,2A,

20 36 2

01~

0 36l

12

6

4 %

'1101. 120 yt

Figue 3. Topialor Caill at160to 100 MT n Agus 18 199. s dpiced b 6.- t 7.-~t Te'" romt21I

sesr heiohrs ie i n heccoecetr 2 20.becrld rs.Speipsdonteaaysso h

Fiur3 . roialstr Cmll a 60 t 90 MTonAgut 8199,a dpctd y6.-t .01m ,,sfrm h MI



N 125* 120* 115, 110* 05 100* 95, 90* IS 85- s* T5 70' 65 so, 55*

2 2 4 250 24045

25D

M MRs 5524a

cm~

o"o

2455 235

M5 265

2425 22

245 255-2Po45*

265

2453

255 240

26 455 25 -5 222

2 25 225 250

2 40

5 4 255 4 4 5 23 520152525 233 260.22 .

4 264 35 22

M 56

Figure 38. Tropical torm Camille at 1600 to 1900 GM T on August 18, 1969, as depicted by 20- to 23-prm TBB's from the

MRIR sensor. Th e isotherms are aiven in K and the cyclone center is indicated by the circled cross.



in maximum cloud-top height was most likely due to the diminished vigor of the upward

vertical motion in the most intense activity near the center.

Normalized reflectances as high as 70 percent were found near the center (Figure 36), wherethe highest reflectances were associated with low 10- to 11-MAm TBB's. The reflectances

that were near 60 percent were accompanied by a wide range of TBB's, some as high as260 K. This suggests that there may be a relationship between reflectance an d the rangeof concurrent 10- to 11-Mm TBB's when the reflectance is high. Th e results of a smallstudy investigating this possible relationship are shown in the Appendix.

Perhaps the most interesting satellite observations made of Camille during this phase of its

lifecycle were taken by the two water vapor channels (Figures 37 and 38). Surrounding

the center to the north and west was a band of high TBB's which were higher than 12 hours

before. This was surprising since the cyclone had weakened so much, and therefore, it

should follow that any storm-produced subsidence would diminish. Upper horizontal

convergence between the outflow and the westerlies over the lower Mississippi Valley that

was first noted from the cloud motions and 200-mbar measurements on August 17 (Figure37) had become more extensive and intense. Th e cloud motions on August 18 best illus-trated the horizontal convergence zone, with the clouds within the outflow moving from

the south and the clouds moving from the northwest in toward Camille in the area north-

west of the outflow.

To obtain a better understanding of this area of high TBB's, both the 1200 GMT, August 18and 0000 GMT, August 19 constant pressure charts were analyzed every 50 mbar from thesurface to 100 mbar and at the 70-, 50-, 30-, an d 20-mbar levels to examine the synoptic

features. The pressure charts were examined individually and by making a time-lapse moviefor each time, vertical continuity could be observed. Selected constant-pressure charts

will be shown that highlight the results of viewing the movies.

A quasi-geostrophic adiabatic I0-level diagnostic model was employed for the same period

to examine the middle- an d upper-tropospheric dynamics and the water vapor budget. Themodel computes vertical velocity from the quasi-geostrophic adiabatic omega equation fora midlatitude synoptic scale atmosphere at ten levels (1000 mbar to 100 mbar at 100-mbarincrements), with a lateral grid spacing of 169 km (Reference 11). The model was later

modified to compute the local change of water vapor in an adiabatic atmosphere (Reference12). Since the model will only handle adiabatic processes, results from the model are most

realistic outside the tropical cyclone.

Figure 39 depicts the ground-observed present weather and cloud cover an d types at 1800GMT (less than 1 hour after the satellite overflight), superimposed on the 6.5- to 7.0-gm

TBB's. Th e cloud information assists in the interpretation of the 6.5- to 7.0-gm measure-ments where the regions of clouds can be identified that could contribute along with the

water vapor to the sensed radiance. Clouds beneath the cirrus level had a negligible effect.

Th e areas covered at least in part by cirrus clouds are shown within the dashed lines in

Figure 39. In general, cirrus clouds were not present when TBB > 240 K. Where the

50



125* Io 115 110" 105" 100* 93* 90" 85" 80 75* 0" W e0O, SS5

o35

DC

.g Om

. .. .~~~~C- 0jy:: -ON\

35*

MO0

2o *

Figure 39. Tropical storm Camille at 1600 to 1900 GM T on August 18 , 1969, as depicted by 6.5- to 7.0-pm T B's from theMRIR sensor superimposed upon the ground-observed cloud cover an d type an d present weather. Th e isotherms are given in K.Areas covered by some cirrus cloud cover are shown within the thin dashed line.



TBB's of > 240 K and cirrus coincide, the occurreince could be explained by satellite

gridding errors, areas of cirrus much smaller than the radiometer field of view, or cirrus

near the local horizon of the observer. Only low and middle clouds occurred in the areaof maximum TBB's northwest of the center. Cirrus were almost always present when the

TBB's were < 235 K. Thus, the only TB B range where considerable interpretation ambiguity

existed was 235 to 239 K.

Figure 40 is the superposition of the August 18, 1200 GMT, 300-mbar chart on the 6.5- to

7.0-pm measurements that were taken 5 hours later. Immediately to the northwest of the

114.0 108.0 102.0 96.0 90D 84.0 78.0 72.0 66.0

43.0 N- b-

S255 -40.0 N

37.0- 0C240

6Jm (2 o- 34.0

31.0- 1

02 0

19.0 NC

105.0W 99.0 9.0 87D 81.0 75.0W

Figure 40. Superposition of the 300-mbar geopotential heights, dewpoint depressions, and winds at 1200

GM T on August 18 , 1969, upon the 6.5- to 7.0-/um TB'S from the MRIR sensor at 1600 to 1900 GMT.

The 6.5-to 7.0-/m (thin solid lines) and dewpoint depression (dashed lines) isotherms are given in Kand the

geopotential heights are in meters. The areas where the dewpoint depressions are <5 or >15 are accentuatedwith a stippled border. Additional stippling depicts the regions where the 6.5- to 7.0-pm TBB's are > 245 K.

cyclone is the southwest flow that is part of the outflow. In the dry ai r further to the

northwest the environmental current is from the northwest. The northwest flow has a

vertical extent from 450 mbar to 150 mbar an d the maximum angle with the outflow

52



was at 300 mbar. When the northwest flow reaches the edge of the outflow the air

must go over or around the storm. Since the outflow extends to, or near, the tropopause,

the majority of the air must have flowed around the cyclone. Therefore, the outflow

from Camille was acting as a barrier to the ambient winds.

The barrier effect was at its maximum north-northwest of the center where the outflow

an d the ambient flow were almost perpendicular. Since the 6.5- to 7.0-pm TBB's were

high and had risen north and west of the center in the past 12 hours, the interaction

was most likely producing subsidence. Some of the dry air was carried east-northeast-

ward from the point of maximum upper horizontal convergence, but the majority of the

air was transported toward the southwest. The dewpoint depression (T - Td) from radi-

osonde measurements was > 15 K over a broad region that included the high TBB's

north and west of the center. Cirrus overriding the dry air, inaccuracies in the radio-

sonde humidity element at cold temperatures, and the 5-hr time difference between

the conventional and satellite observations are probably the main reasons why there were

some TB B 's of < 230 K within the area where T - Td > 15 K. Where the upper hori-zontal convergence occurred between the outflow and the environment, the time-lapse

analysis of the moisture and wind field indicated that the dry air was penetrating beneath

the outflow layer in the middle an d upper troposphere.

To obtain further insight into the causes producing the increasingly warm equivalent TBB's

in the 6.5- to 7.0-pm MRIR water vapor channel west an d northwest of Camille, two

procedures were followed. The first was to examine the upper tropospheric dynamics,

an d the second was to determine how these dynamics affected the upper tropospheric

water vapor.

Figure 41 depicts the 350-mbar vertical motion patterns superimposed upon the 300-mbar

geopotential height and radiosonde wind observations. There are two areas of interest. Thefirst is located behind the weakening trough north of Camille. Subsidence of > 1 Pbar/s

had been induced mainly by differential advection of negative vorticity. Cross section A

in Figure 42 , which is located west-east vertically through the base of the trough (desig-

nated by line A in Figure 41), delineates the vertical extent of this subsidence between

longitudes 850 to 950 W. It can be seen that the greatest subsidence was in the upper

troposphere. The second area of interest is the obstruction effect previously mentioned in

the synoptic discussion. Here subsidence of > 0.5 pbar/s is observed an d is attributed to

both the Laplacian of horizontal cold ai r advection an d to differential advection of neg-

ative vorticity. Since the boundary between the outflow and the ambient flow was much

smaller than the spatial resolution of the grid mesh in the model, it is possible that

higher vertical velocities could be concentrated along the boundary than were calculatedby the model. In quantitatively investigating the horizontal convergence within this area,

it was found from vertical mass influx computations at all 10 levels for 1200 GMT,

August 18, 1969, that the level of maximum horizontal convergence was above the

350-mbar level with horizontal divergence below. Thus, as was suggested in the synoptic

discussion, Camille's upper tropospheric outflow was converging with the westerlies with

53



12 GMT 18 Aug '69 300 ml

934 92

94

300xjOgpm

w3 50 x 10- 4

mb/sec

Figure 41. 300-mbar geopotential heights and winds and 350-mbar verticalmotions for 1200 GMT on August 18, 1969. Lines A and B denote crosssections A and B which are shown in later diagrams.

subsidence below. Figure 43 depicts the 300-mbar horizontal divergence field superim-posed upon the 300-mbar geopotential height and radiosonde wind observations. An areaof horizontal convergence of < -5 X 10-6 s- 1 is seen northwest of Camille. As seen incross section B in Figure 44 which is located west-east vertically through the obstruction

effect area (designated by line B in Figure 41), the subsidence again extends throughout

the troposphere between longitudes 880 to 980 W, but with weaker magnitudes in the

upper troposphere as compared with the upper troposphere subsidence behind the trough

as seen in cross section A.

To best illustrate how the two areas of dynamical interest support the tongue of warm

TBB's measured northwest of Camille at 1800 GMT on August 18, three-dimensional tra-

jectories were computed at 3-hr intervals between 1200 GM T on August 18 to

0000 GM T on August 19 at the 350-mbar level. Figure 45 depicts four such trajectories

superimposed upon the field of 6.5- to 7.0-pm TBB's for 1800 GM T on August 18. Tra-jectories 1 and 2 best outline the upper tropospheric flow from the base of the trough

southwestward into the obstruction effect area. It is seen that the parcels were first

under the influence of subsidence behind the trough, but as they moved out of thetrough and into the area of the obstruction effect, the parcels were forced to continue

to subside. Further downstream, southwest of this obstruction effect as outlined by tra-

jectory 4, the air parcel continued to subside during the first 6 hours as it moved south-

westward, but at a much slower rate. This subsidence was completely induced by the

54



I2GMT 18 Aug '69 Cross Section A

VERTICAL MOTION =10- 4

mb/SEC

SPECIFIC HUMIDITY q= g/Kg

200 5

30 0

5001.0

700 - ,/6800 -

900 10900 - so -

I I I I I IIIOW O100W 90 0 W 80 0 W

HURRICANE CAMILLE 32.50N 900°W

Figure 42. Cross section A, depicting vertical motions (solid lines) and spe-

cific humidity (dashed lines) for 1200 GMT, August 18, 1969.

Laplacian of the horizontal advection of colder air from the northeast. After 6 hours the

parcel moved into an area of slight ascending motion. Trajectory 3 illustrates an air parcel

moving through the base of the trough. During the first 9 hours the parcel experienced

subsidence induced mainly by differential negative vorticity advection from behind the

trough. However, once the parcel moved out in front of the trough it moved into an area

of ascending motion. Thus the trajectories suggest that there were two tongues of subsid-

ence that were associated with the tongue of warm TBB's: one moving cyclonically north-

eastward through the base of the trough and the other moving anticyclonically southwest-

ward out of the trough and into the area of enhanced subsidence below the obstruction

effect area.

To obtain a better comprehension of what the 6.5- to 7.0-m water vapor channel measure-

ments indicated, an examination of the effects of the upper troposphere dynamics upon

water vapor amounts at this same level was miade. In determining the water vapor budget

the local change of specific humidity for an adiabatic atmosphere was calculated. Assuming

55



12 GMT 18 Aug '69 300 mb

/ 0 H 940-5

946

-I--55

VH x9520 sec- t 5

958

0 2

5q-

(qw)967 970

or

964

atnerines)or 1200 GMT on August 18,196(2)9.

5656



where

aq/at = local change of specific humidity

-V - Vq = horizontal advection of specific humidity

aw

S = vertical advection of specific humidity

a(qw)

ap

P = precipitation

E = evaporation

All three terms on the right side of Equation 2 were computed from data from 1200 GMT

on August 18, 1969, for the interior grid points from 800 mbar to 300 mbar and were thenalgebraically added to obtain the local change of specific humidity. Figures 46 and 47

12 GMT 18 Aug. '69 Cross Section B

VERTICAL MOTION W=104 mb/SEC

SPECIFIC HUMIDITY q=g/Kg

P

200 -

00 o

300- A

,-5II

oo ;-

600- 5 0 60

-40700 I 5

-40iII Ie

1100

W 1000

IW 900

W 80nW

HURRICANE CAMILLE 32.50 N 90.00 W

Figure 44. Cross section B, depicting vertical

motion (solid lines) and specific humidity (dashed

lines) for 1200 GMT on August 18, 1969.

57



18 GMT 18 August 69 NIMBUS 3 MRIR 6.5-7.0 pm TBB

230 23

240 240 23

240

20

20

42230

A B4 _' 66 240 A

20 4240

2 0 24 240

Figure 45. Three-dimensional trajectories between 1200 GM T on August 18, 1969, and 0000 GM T on

August 19, 1969, initiated at the 350-mbar level, superimposed upon the 6.5- to 7.0-pm TBB's at 1600 to

1900 GMT, August 18, 1969, from the MRIR sensor. The crosses represent the position of the air parcels

at 3-hr intervals and the number to the side represents the pressure height.

represent these water i apor budget computations for a small segment of cross sections A and

B, respectively, at 1200 GMT of August 18. Figure 46 depicts the vertical intersection

through the area of warm TBB's near the trough north of Camille, and Figure 47 depicts the

vertical intersection through the area of warm TBB's in the obstruction effect area north-

west of Camille. Th e graph located above the local change of specific humidity calculation

represents the 6.5- to 7.0-pm and 20- to 23-pm TBB's along the cross section as observed by

the MRIR sensor approximately 6 hours later. It can be seen from the local change of

specific humidity computations in Figure 46 that the upper troposphere became drier with

a maximum decrease of specific humidity of about 5 X 10s g/kg s- .This drying trend is

attributed mainly to subsidence and to a lesser extent to the horizontal advection of drier

air from behind the trough. This is what is inferred by the 6.5- to 7.0-pm TB B trace along

the same segment of cross section A 6 hours later. The location of the maximum TBBcorresponds closely to the location of the greatest upper tropospheric drying.

58



Cross Section A Cross Section A

12GMT 18 Aug. '69 Nimbus 3 MRIR for 18 GMT

18 August 1969Local change of specific humidity assuming 20-23m8ugust 9no precipitation or evaporation I K 6.5-7pm --- K

E 270 - 250Values given in10-5 g/Kg sec

- 1E 260 2500

2600 -0- -2400

00 2500 2300

aw -aqw aqq ap - Vvq ap at

P P

300 - -00

Slii l oo

400 -. .400/

/0 0 2 500

600- 2 600

700 --. 700

95°W 90°W 85°W 950 W 90 0 W 850 W 95OW 90°W 85OW 950 W 90°W 85°W

Figure 46. Segment of cross section A depicting the calculations of the local change of specific humidity

and its partitioned terms from 1200 GMT, August 18, 1969, data and the 6.5- to 7.0-gm and 20- to 23-gm

TBB traces for the same segment of cross section A at 1800 GMT on August 18, 1969.

The same results were found in Figure 47. The upper troposphere also became drier, as seen

in the local change of specific humidity calculations, but at a somewhat faster rate as com-pared to the calculations represented in Figure 46 for the same level. Again this is attribut-

ed mainly to subsidence, but as before, horizontal advection of drier air from the northeast,

particularly in the middle troposphere, cannot be ignored. The 6.5- to 7.0-pAm TB B at

1800 GMT again shows that the location of the maximum TBB corresponded favorably to

the position of the upper tropospheric segment that had maximum drying.

From the examination of the upper tropospheric dynamics and water vapor budget, it can

be concluded that the dry air that had descended to the west of the trough had moved both

northeastward and southwestward. The southwestward moving branch continued to de -

scend under the influence of enhanced subsidence induced by the upper tropospheric con-

vergence between Camille's outflow and the westerlies.

East of Camille there was another area where T - Td > 15 K in the middle troposphere. This

appears to have been produced by a shear line between two anticyclones east of the storm.

The time-lapse movie that showed the shear line sloped from east to west with decreasing

altitude as the western anticyclone, probably created by the outflow, weakened and disap-

peared below the outflow layer. Cirrus overriding the dry ai r prevented the occurrence

59



Cross Section B Cross Section B

12GMT 18 Aug. '69 Nimbus 3MRIR for 18GMT

Local change of specific humidity assuming 18August 1969

no precipitation or evaporation K K '

Values given in I0- 5

g/Kg sec-

, 2700 250" ~E 2600 2400 3C 250' 20-23 m - e 230

°0

aw -aqw eq

P ap -VVq ap at P

300 - \- -- 30 0

400- - 400

500 - C- - 500

600- -600

700- 700

800 - 800Camille at 32.5*N 90.0*W

I 1I 1 il

IOOOW 900

W IOOOW 90°W IO00W 900

W IOOOW 90°W

Figure 47 . Segment of cross section B depicting the calculations of the local change of specific humidity

and its partitioned terms from 1200 GMT, August 18, 1969, data and the 6.5- to 7.0-pm and 20- to 23-Gm

TBB traces for the same segment of cross section A at 1800 GM T on August 18, 1969.

of high TBB's except for the southern portion where no cirrus was indicated by a few sta-

tions over Florida where TBB's as high as 240 K were observed. Higher TB B 'S were not

measured because the T - Td was not consistently large throughout the 300- to 500-mbar

layer. Below 300 mbar, T - Td generally varied between 10 an d 15 and the 235 to 2401measurements compare favorably to the 236 K TBB that would be expected from a clear

atmosphere using radiative transfer theory and the 1200 GMT temperature and moisture

profile from Tampa.

The superposition of the 500-mbar analysis for August 18 at 1200 GMT on the 20- to

23-pm TBB field is shown in Figure 48 . North and west of Camille is the broad dry zone

that was observed in the 6.5- to 7.0-pm measurements. Most of the TBB 'Sof > 265 are

within the T - Td > 20 isoline. A small 500-mbar anticyclone was located almost under-

neath the area of maximum horizontal convergence at higher levels. Most of the dry ai r in

the middle troposphere created by the subsidence was transported southwestward by the

broad current of northeast winds which was similar to observations in the upper troposphere.

The T- Td > 20 measurements east of the center, produced by the shear line, were under-

neath the cirrus moving eastward from Camille. Therefore, this dry area was not detectable

with the 20- to 23-/pm radiances.

In examining the TgB measurements in the 20- to 23-pim water vapor channel, the same

procedures were followed for the examination of the dynamics and the water vapor budget

60



108.0 102.0 96.0 90.0 84.0 78.0 72.0

0255

.* 15 240

-,5 24 40 -34.0

31.0- 0 225

5 .. ..260 2 0

- 25 555 2635 250

50 ai eh , 1io a 24920

25.0- 20

_0260 222

25 245

19.0

1050 W 99.0 930 870 810 75.0

1200 GMT on August 18, 1969, uoon the 20- o 23-Ezm TBB'S from th e MRIR sensor at 1600 to 1900 GMT.The 20- to 23-pm (thin solid lines) and dewpoint depression (dashed lines) isotherms are given in K and the

geopotential heights are in meters. The areas where the dewpoint depressions are < 5 or > 20 are accentuated

with a stippled border. Additional stippling depicts the regions where the 20- to 23-gm TBB'S are > 265 K.

as were used for the measurements in the 6.5- to 7.0-jm channel. However, since the emis-

sion measured in the 20- to 23-pm channel emanates mostly from the middle troposphere

for clear-sky conditions, the dynamics an d water vapor budget must be examined in both

the upper and middle troposphere.

The two areas of greatest dynamical interest are the area behind the trough and the area of

the obstruction effect. Figure 49 shows the 550-mbar vertical motion field superimposed

upon the 500-mbar geopotential heights, radiosonde wind observations, and integrated pre-cipitable water amounts above 800 mbar.

The subsidence associated with the trough in the upper troposphere had created an area of

horizontal divergence that can be seen from the 500-mbar wind observation north of

Camille. This is the small middle tropospheric anticyclone mentioned in the synoptic dis-

cussion. Emanating from this area was a southwest-northeast orientated tongue of

61



12 GMT 18 Aug'69 500 mb

15

50ooX 10Opm E> 0mm precipitable water above 800mb

S55 0x IO - 4 mb/sec E3< 10mm precipitable water above 800mb

Figure 49. 500-mbar geopotential heights and wind velocities, 550-mbar

vertical motions, and integrated precipitable water above 800 mbar for

1200 GM T on August 18, 1969. Lines A and B denote cross sections A and B

which are shown in later diagrams.

subsidence of the magnitude of >0.5 pbar/s. As seen in cross section A (Figure 42), subsid-

ence extended throughout the middle troposphere and was induced by both the differen-

tial advection of negative vorticity and the Laplacian of horizontal cold air advection.

Further southwestward, under the area of the upper tropospheric obstruction effect, subsid-

ence still existed, but the upper tropospheric convergence was not the dominant reason

for its existence. The tongue of maximum subsidence was further west, away from the ob-

struction effect, and was induced mainly by the Laplacian of horizontal cold ai r advection

from the northeast. Cross section B between longitudes 90' to 1000 W in Figure 44 best

illustrates the westward displacement of the middle tropospheric subsidence from that in

the upper troposphere.

Figure 50 shows four middle tropospheric trajectories superimposed upon the MRIR 20- to

23-gm TBB field for 1800 GMT on August 18. All of the trajectories originated at the

550-mbar level at the same points as those at the upper tropospherelevel. A descent of near-

ly 12 mbar in the first 6 hours for trajectory 1 shows that the air that moved out of the

trough continued to subside in the middle troposphere. Trajectories 2 and 3 illustrate themiddle tropospheric divergence north of Camille. Air moving out of the area of divergence

subsided during the first 3 hours, but as seen in trajectory 3, ai r parcels moving out ahead of

the trough moved into an area of ascending motion. Trajectory 4 depicts the subsidence

further downstream west of Camille. Here, under the influence of the Laplacian of horizon-

tal cold air advection, the air continued to subside throughout the period. Thus, in the

62



18 GMT 18 August 69 NIMBUS 3 MRIR 20-23 pm TBB

26055

A .. A

2 2.5B B

240 2504

Figure 50. Three-dimensional trajectories between 1200 GMT on August 18, 1969, and 0000 GMT on

August 19, 1969, initiated at the 550-mbar level superimposed upon the 20- to 23-ym TBB's at 1600 to1900 GMT on August 18, 1969, from the MRIR sensor. The crosses represent the position of the air

parcels at 3-hr intervals and the number to the side represents the pressure height.

middle troposphere, the trajectories suggest that there were again two tongues of subsidence

that coincided with the warm TBB's,one moving northeast of the divergent area ahead of the

trough and the other moving southwestward out of the trough.

Examining the local change of specific humidity for the same segment of cross sections A

and B (Figures 46 and 47) that was examined for the upper troposphere, it is shown in cross

section A that dry air extended downward to 600 mbar, where there was dry air resulting

from subsidence. However, the horizontal advection of moist air from the northwest below

600 mbar essentially compensated for the drying by subsidence. Figure47 shows that the

loss of moisture in the middle troposphere was contained by both horizontal and vertical

advection. As was found in the 6.5- to 7.0-pm TBB's, the warm 20- to 23-pm TBB 's

correspond favorably with the position of maximum middle level subsidence.

There are two regions where T - Td < 5 K. One is associated with Camille and the other is

over a portion of the northeast United States. In the western third of the latter area the

low TBB 's show the high clouds associated with thunderstorms over Pennsylvania. In view

63



of the westerly current, the small T - Td values east of the thunderstorms were probably the

result of these storms and the westerly current that advected the moisture downstream.

Figure 51 indicates the amount of compressional heating caused by the subsidence at the

600-mbar level. In the dry ai r the temperatures were 277 to 278 K as compared with 274 K

0.0 43.3 - .0

0.2 o 0.0

2. 0.2 0.8 0. 0

3.0 4 1.0

H

L 4240S2.0

600 MB

Figure 51. 600-mbar geopotential heights (in meters), dewpoint depression (in K) analyses, and radiosonde

stations showing 600-mbar temperature and winds for 1200 GMT on August 18, 1969.

around the periphery of Camille. The dry ai r extended down to the 850- to 900-mbar layer

underneath the region of maximum upper tropospheric horizontal convergence. Camille

still appeared to be a warm core system in the middle troposphere, because the 277.5-K

temperature measurement at Jackson, Mississippi, was 2 to 4 K warmer than the temper-atures near the edge of the circulation. The time-lapse movie showed that the area of

T - Td < 5 K increased from low to middle levels with the most noticeable effect be-

tween 700 and 600 mbar. This could have been caused by the outward transport

of moisture in response to high momentum air at low levels moving upward to a region

of lower momentum near the center of a warm vortex.

64



In Figure 51 the 600-mbar wind at Little Rock had a pronounced radial component away

from the storm center.

The time-lapse movie of the radiosonde measurements taken at 0000 GM T on August 19

showed that Camille had weakened further by the time the cyclone had reached ex-

treme northern Mississippi. By then Camille had been downgraded to a tropical depression.

The strong obstruction effect had almost disappeared, with the ambient upper tropospheric

ai r moving eastward north of the cyclone and an anticyclone west of the center building

eastward with a small ridge north of the storm center. These features are seen in the

300-mbar chart for 0000 GM T on August 19 (Figure 52). Despite the substantial reduction

of the obstruction effect, the dry ai r created earlier was still present. With the change to a

more zonal circulation north of Camille more of the dry air (T - Td > 15) advected

300 MB Ar

Figure 52. 300-mbar geopotential heights (in meters), dewpoint depression (in K) analyses, and radiosonde

stations showing 300-mbar winds for 0000 GMT on August 19, 1969.

65



eastward over Indiana, Ohio, and some of the Middle Atlantic States. West and southwest

of the center there was a large region of dry ai r which was most pronounced in the middle

troposphere. This would indicate that no large-scale upward motion had begun. The

500-mbar chart for August 19 at 0000 GM T (Figure 53) illustrates the extent of this dry

air and also that Camille acquired more of the characteristics of a cloud field seen with

extratropical cyclones with the cloud mass principally northeast of the center.

50 0 MB

Figure 53. 500-mbar geopotential heights (in meters), dewpoint depression (in K) analyses, and radiosondestations showing 500-mbar winds for 0000 GMT on August 19, 1969.

At 0000 GMT on August 19, 1969, the dynamics of the upper and middle troposphere had

changed. With Camille weakening further and becoming more embedded into a more zonal

westerly flow, the interaction between Camille and the westerlies had become weaker.

Horizontal divergence calculations at all 10 levels revealed that there was an area of hori-

zontal convergence at the 300-mbar level north of Camille; however, it is not part of the

obstruction effect that was seen 12 hours earlier. The vertical motion patterns at 350 mbar

as seen in Figure 54 indicate that the subsidence below the area of this maximum upper

66



00 GM T 19 Aug.'69 30 0 mb

o0 4

964/ 0p -10

3 5o x10-4 b/sec

sections A and B.

-567

0

A300x Ogpm

'350 xIO-4 mb/sec

Figure 54. 300-mbar geopotential heights, winds, and 350-mbar vertical

motions for 0000 GMT on August 19. Lines A and B denote cross

sections A and B.

tropospheric horizontal convergence wa s quite weak. This subsidence can be attributed

mainly to th e differential advection of negative vorticity from around the anticyclone to

the west. Cross section A (Figure 55)* indicates that th e subsidence wa s considerably weak-

er and had shifted further west. The vertical extent, however, still extended through th e

troposphere and as depicted by the isolines of specific humidity, the column was still dry.Futher south, to the west an d northwest of Camille (the area of the obstruction effect 12

hours before), subsidence was still present, but was mainly induced by the Laplacian of

horizontal cold air advection. This area of subsidence does not extend as fa r southwest-

ward as it did 12 hours earlier. Cross section B (Figure 56 ) reveals that th e upper tropos-

pheric subsidence had also weakened considerably, which indicates that the obstruction

effect has decreased. However, the subsidence extended throughout the troposphere which

continued to keep th e column dry as seen in th e isolines of specific humidity.

In the middle troposphere, vertical motion patterns associated with the interaction of

Camille with the westerlies and with the trough north of Camille had also changed. Figure

57 indicates that subsidence at the 550-mbar level associated with the dissipating trough

northwest of Camille wa s still present but ha d weakened. This subsidence was induced by

both the Laplacian of horizontal cold air advection and th e differential negative vorticity

*Cross sections A an d B in Figures 55 an d 56 respectively, intersect the same areas as in Figures 42 and 44 .

67



00 GMT 19 Aug '69 Cross Section A

VERTICAL MOTION w = 104

mb/SEC

SPECIFIC HUMIDITY q= g/Kg

0

00

-5

5S

lOW IOO*W 90°W 80*W

HURRICANE CAMILLE 35.0*N 90.0*W

Figure 55. Cross section A, depicting vertical motions (solid lines)

and specific humidities (dashed lines) for 0000 GMT, August 19, 1969.

advection. Cross section A (Figure 55) depicts this middle tropospheric subsidence. Further

southwestward under the area of the obstruction effect that was present 12 hours earlier,

both the vertical motion patterns on Figure 57 and cross section B in Figure 56 reveal that

the subsidence had also decreased. The strong subsidence further west that is best seen in

cross section B (Figure 56) was not associated with Camille or its interaction with thewesterlies but was induced by the Laplacian of horizontal cold ai r advection from around

the ridge west of Camille.

Excessive RainsOver the CentralAppalachians

As Camille moved northeast an d then east across western Tennessee and Kentucky it was

a tropical depression producing moderate to locally heavy rainfall at stations close to the

track of the center of 5 to 10 cm. However, as the storm center reached West Virginia,

heavy rainfall began in west central Virginia around 0000 GM T on August 20 and

continued for approximately 12 hours. Schwarz (Reference 2) has reported that the

extreme rainfalls were associated with near record low-levelmoisture that was not subject

to depletion by upwind mountain ridges. The Appalachians contributed to the large local

amounts (as high as 69 cm ) because the low-level flow was from the southeast, perpen-

dicular to the mountain range. Upper tropospheric horizontal divergence is another

important condition that would favor regeneration of the cloud mass and lead to the pro-

duction of heavier precipitation. A strong field of horizontal divergence is present just

68



00 GM T 19 Aug.'69 Cross Section B

VERTICAL MOTION w=I-4

mb/SEC

SPECIFIC HUMIDITY q=g/Kg

/

/

.00 20 0

15 0I

I OW O100W 900

W 800

W

HURRICANE CAMILLE 35.O0N 90.O°W

Figure 56. Cross section B, depicting vertical

motion (solid lines) and specific humidity

for 0000 GM T on August 19, 1969.

northeast of the cyclone center as shown by the 200-mbar chart in Figure 58 for

1200 GMT on August 19 (about 12 hours before the high rainfall rates began). Both

speed and direction divergence were present.

The regeneration of the cloud mass was best seen by the Nimbus-3 HRIR which ob-

served Camille's cloud canopy at 24 hour intervals for 2 days before and during the

excessive rain period.

Figure 59 provides a standard of reference for the two later HRIR views of Camille

as the cyclone was moving ashore on the Mississippi Coast. A large spiral band is seen

south of the storm center and the lowest 3.5- to 4.1-jpm TBB's are < 210 K over a

small area near the center. A maximum cloud-top height of 125 mbar (15.4 km) anbe inferred from the minimum TBB and the 0000 GMT, August 18, Lake Charles

temperature profile.

Twenty-four hours later the greatly weakened cyclone was over western Tennessee. The

HRIR view (Figure 60) indicates that the cloud pattern was more disorganized with the

69



00 GM T 19 Aug.'69 500 mb

#500 xlOgpm E >20mmprecipitable water above 800 mb

o550 xliO- 4 mb/sec [< IOmm precipitable water above 800 mb

Figure 57. 500-mbar geopotential height, wind, 550-mbar vertical

motion, and integrated precipital water above 800 mbar for 0000 GMT

on August 19. Lines A and B denote cross sections A and B.

highest clouds (lowest TBB's) about 300 km northeast of the center. There has been an

increase in the minimum TBB to about 215 K. Using the 0000 GMT, August 19 sounding

at Nashville, this TBB measurement would mean the maximum cloud top was at 180 mbar

(13.2 km). Thus, there was an apparent reduction in the maximum cloud top height of

2.2 km in 24 hours.

The third and final HRIR view of the series (Figure 61) is at the time of the excessive rain-

fall. Most of the intense cloudiness (as outlined by the area where the TBB,'s were < 230 K)

was north and east of the center similar to the position of the cloud shield for an extra-

tropical cyclone. There were several areas where the TBB values were < 210 K with a min-

imum of 20 3 K. Therefore, the maximum cloud top height had returned to the 125-mbar

(15.4-km) level which was the maximum cloud top level when Camille was still a severe

hurricane, shortly after crossing the Mississippi coast.

Extratropical secondary storm development has occurred when there was a significant ver-

tical cloud growth within a large existing cloud mass; this growth can precede the detection

of the new cyclone over the ocean with conventional data (Reference 13). In the case of

Camille, the vertical cloud growth to high levels was noticeable when the heavy rains werein progress. Since the HRIR observation interval is 24 hours it is not possible to determine

whether or not the clouds reached the observed 0600 GMT, August 20 levels before or at

the onset of the period of maximum rainfall rate. If infrared measurements would have

been available on a nearly continuous basis from a geosynchronous satellite, it is possible

70



1100 1000 90 * so * 70' 60

LOW4

420

30 30 4

3242

2v 20 5 50 41~ 42 3

16 36 426-42 4 6 22 3

20

22 8 3 220 0 83 2

12 2

Is 26 1 20 W\ Lo

\ h/ HIGH 32I

is 30

20

Figure 58 . 200-mbar streamlines as derived from radiosonde wind observation and ATS middle and upper cloud motion fo r 1200 GMT on August

~ 19, 1969. The cloud motion directions are indicated by the arrows and the speeds are given in knots.



95.00W 92.0*W ss.0W 85 5W 83.0*W

33.00N 33w0_

92.0W 89.0oW 85.5oW 8300

" 21 2-

72

J5.°.0

tion interval.

72



97.O0W 94.O0W 91.0oW 88.0oW 85.0W 82.0W

26 0 2 76 7P 22 76 2 20 6

Qoo

260 C?2 2 0 2$0

2 2722 2,0 8 O

27 6 82 : 2 1 (2290S

C" 2W0 T 26008.

0: ' i o c io2

6 39 29 0 Cia93t [C.

i u 270 2o 700-26,4D1-#m

25 0--z _29

Prioroterpddeeigpaeo ai ther was th evlp e o f no-ee

35.o~n JiJ s'

280 ~ go 2970 (~ 276 2 j

32.0N -2- [5 290 C776, - 32.o5

2 7 0 2 9 0C0 12 0 " 8 n

s 0

K0 0 0

w 0t0 t260 29o 27d l op 9 N

2 20 2 6 27,

(3 n ~ ~ 6640I 2 296 7 C6

2967 26o9 }6 2o-,6

91.OW 94.O'W 91.0W 88.OW 85.OW 82.OW

Figure 60. Tropical depression Camille at 0450 GMT on August 19. 1969. as depicted by 3.5- to 4.-m

oa from the HRIR. The isotherms are given in K.

CONCLUSIONS

Prior to the rapid deepening phase of Camille there was the development of strong low-level

inflow of high moisture content air from the region near the ITCZ moving over a sea surface

with temperatures 1 to 3 standard deviations above normal. The concurrent developmentof large-scale cloud bands southeast of the center extending to the cirrus level was the satel-

lite evidence of the inflow. Just before the onset of the rapid deepening, the cloud bands

showed a broad connection to the cloudiness within the ITCZ, and the 6.5- to 7 .0 -am water

vapor channel indicated the outflow of cirrus to the north of the center which was associat-

ed with the necessary ventilation of air away from the cyclone. It was possible to ascertain

73



87.0W 84.0 W 81.0W I78.0OW 75.0W 72.00

W

2g 6 280 1 n9 Q M t280o4.! m

22 280 (280828

280 r9

43.0N (5:8 2 43.0'N

Ts H so r a i

25 240

-- 2211.

7.0°0°80 7° 3.0-N

Figure 61. Tropical depression Camille at 0550 GM T on August 20, 1969, as depicted by 3.5- toe4.1-/im2'srom the HRIR. Th e isotherms are give in K.

the relative strength and position of the important features such as upper tropospheric shear

least the low cloud level as inferred from the measurements made by four of the MRIR

channels. Neither the expanded cloud shield nor the large subsidence region seemed to

74

Figue 6.opial eprssin Cmile a 050 GT o Auust20,196, a deictd b 3.- t)4.Sot2 25 2eHIRiohrm regvn nK

thltvesregh n o27o29' 29ntfatrs uh supe roophrc ha

liesad atel ?0 200 260 CtinNibsi RI hanesan oniury

Duin herai depnngphs te ludbnd suh2Bto te2 260inedt

pr

sist At he bginingth raid~depenngthe ize f th inenseclou r (!90 C300te



persist for more than 24 hours, leading to the hypothesis that that the rapid deepening was

associated with a sudden surge that extended past an energy level that was sustainable.

There was a 2- to 3-km reduction in maximum cloud top level about 12 hours after Camille

moved inland in response to the probable weakening of the vertical circulation within the

inner portion of the cyclone. The outflow acted as an obstruction to the environmentalflow in the upper troposphere northwest of the center of Camille. This interaction produced

horizontal convergence and subsidence. The effects of the subsidence could be seen in

the two water vapor channels where the subsidence areas were demarcated by TB B 's that

were initially relatively high and were increasing with time. A smaller region of dry air

produced by a shear line east of Camille was not seen as clearly in the satellite measure-

ments as the area northwest of the center du e to the cirrus clouds in the outflow region

overriding the dry air.

As Camille turned eastward across the lower Ohio valley, the cloud pattern looked like that

of a small vigorous extratropical storm with maximum cloud tops at 13 to 14 km , and the

interaction between the outflow and the environmental flow had diminished considerably.

Speed an d direction divergence in the upper tropospheric wind field determined from con-

ventional and cloud motion measurements was an indication that conditions were favor-

able for renewed vertical cloud development as early as 1200 GMT on August 19. When

the excessive rainfall was in progress over Virginia, the cloud tops had risen to 15 to 16 km,

which was the level reached when the cyclone was still a severe hurricane just after

crossing the Mississippi coast.

This investigation of Camille has attempted to indicate the knowledge that can be gained

with (a) multispectral analysis of registered satellite measurements of reflected and emitted

infrared radiation; (b) the combination of polar orbiting and geosynchronous satellite and

conventional information; and (c) the use of the time domain.

Some of the hypotheses that emanated from the interpretation of the satellite data could

not be confirmed due to the lack of supporting conventional information. With the flight

of multispectral sensors on future geosynchronous satellites, the increased temporal reso-

lution will undoubtedly improve the understanding of tropical cyclone circulation features

beyond what was possible with a 12-hour observation interval. Th e observations with the

spacecraft instrument should be combined with the concurrent collection of in situ, remote

earth-based or aircraft measurements of sufficient quality to more effectively interpret the

satellite information.

75



ACKNOWLEDGMENTS

The authors wish to thank Professor Tetsuya T. Fujita, University of Chicago, who com-

puted the cloud motions, performed streamline analyses from the results, and madenumerous helpful suggestions. The analyses of the radiation and conventional charts by

David Simmes, Allied Research Associates, and Mark Smith, GSFC, are also appreciated.

Goddard Space Flight Center

National Aeronautics and Space Administration

Greenbelt, Maryland, August 10, 1973

039-23-01-01-51

76



REFERENCES

1. Simpson, R. H., A. L. Sugg, and Staff, "Atlantic Hurricane Season of 1969," Mo.

WeatherReview, 98, 1970, pp. 307-314.

2. Schwarz, F. K., "The Unprecedented Rains in Virginia Associated with the

Remnants of Hurricane Camille," Mo. WeatherReview, 98, 1971. pp. 851-859.

3. Nimbus Project, Nimbus-3 Users Guide, National Space Science Data Center, Goddard

Space Flight Center, Greenbelt, Md., 1969, p. 237.

4. Hubert, L. F., and L. F. Whitney, "Wind Estimation from Geostationary-Satellite

Pictures," Mo. WeatherReview, 99, 1971, p. 665.

5. Shenk, W. E., H. Powell, V. V. Salomonson, and W. R. Bandeen, "MeteorologicalUses of the Stereographic Horizon Map Projection," J. Appl. Meteor., 10, 1971,

pp. 582-589.

6. Rodgers, E. B., and V. V. Salomonson, "Upper Tropospheric Dynamics as Reflected

in Nimbus-4 THIR 6 .7-pm Data," NASA Technical Note D-7493, 1973, p. 27 .

7. Shenk, W. E., and R. J. Holub, "A Multispectral Cloud Type Identification Method

Using Nimbus-3 MRIR Measurements," PreprintVolume of the Conference on

AtmosphericRadiation, 1972.

8. U. S. Naval Oceanographic Office, OceanographicAtlas of the NorthAtlantic Ocean,

Pub. No. 700, Washington, D. C., 1967.

9. Tatehira, Ryozo, "A Study of Rainband," JapanMeteorologicalAgency, 34, 1967,pp. 115-137.

10. Gentry, Cecil, "A Study of Hurricane Rainbands," National Hurricane Research Project

No. 69 , 1964, p. 85.

11 . Barr, S., P. E. Long, and I. A. Miller, "Atmospheric Sensing and Prediction Projection

Project," Scientific Report No. 2, Air Force Cambridge Research Laboratories, Air

Force Systems Command, U. S. A. F. , Bedford, Mass., 1970.

12. Palmen, E., and C. W. Newton, Atmospheric CirculationSystems, Academic Press,

New York and London, 1969, pp. 379-386.

13. Shenk, W. E., "Meteorological Satellite Infrared Views of Cloud Growth Associated

with the Development of Secondary Cyclones," Mo. Weather Review, 98, 1971,

pp. 861-868.

77



APPENDIX

RELATIONSHIP BETWEEN BRIGHTNESS AND CLOUD-TOP TEMPERATURE

From the comparison of the 0.2- to 4.0-pm normalized reflectances and the 10- to 11-pm

TBB 's on August 18 over the Camille cloud canopy, it was observed that when the reflec-

tances were near 70 the TBB's were around 230 K, whereas when the reflectances were

near 60 they were concomitant with TBB's from 225 to 260 K. From these observations

it was hypothesized that there might be a relationship between reflectance an d the range

of 10- to 11-pm TBB 's, where, at extremely high reflectances, this range might be small.

When reflected energy channel. spatial resolution is much higher than that of the infrared

window channel, a relationship could be useful in isolating the regions where the highest

clouds occur and for the interpretation of a registered infrared measurement. A specific

example of a large resolution difference between these two channels will be on the Syn-

chronous Meteorological Satellite (SMS) where the visible and infrared channel spatial res-

olutions will be 0.9 and 9.0 km, espectively.

The MRIR reflectance window channel relationship shown in Figure A-i was established

through the use of daytime data for August 18 that was associated with the ITCZ south

of Camille. This region was chosen because a relatively large percentage of the data sam-

ples had normalized reflectances of > 60 . Figure A-1 indicates that the spread of the

10- to 11-pm measurements decreases as reflectance increases. The range of TBB's when

R > 73 is only 15 K and the maximum TBB was 212 K. An upper boundary could be

drawn in Figure A-i that would connect the highest TBB 's for a given reflectance. Thus,these data suggest that extremely high reflectances can provide some estimate of the

window TBB's in regions of deep convection.

PRECEDING PAGCE BLANI NOT FI1 -ED

79



90

~80

- *

7 - * A MS,

N 60 - • *

0z 50

250 240 230 220 210 200 190

10-11/.m EQUIVALENT BLACKBODY TEMPERATURE (oK)

Figure A-1. Plot of concurrent normalized reflectance measurements and 10- to 11-pm TBB's in the

Intertropical Convergence Zone south of Hurricane Camille during the daytime on August 18, 1969.

satellite views of hurricane camille

Documents