sonde experiments for comparative measurements of optical turbulence · sonde experiments for...
TRANSCRIPT
SAFGL-TR-LR?.0079
INtSTRUM9NTATION PAPERX. NO. 310
Sonde Experiments for Comparative Measurementsof Optical Turbulence
J.H. BROWNR.E. GOODP.M. BENCHG. FAUCHER
24 FEBRUARY 1982
Approved for public release; distributlor "uIlmitod.
C)
AERONOMY DIVISION PROJECT 6687
AIR FORCE GEOPHYSICS LABORATORYHANSCOM APB, MASSACHUSETTS 01731
AIR FORCE SYSTEMS COMMAND, USAF
144
REPORT DOCUNENTATIO?4 PAGE j FFRE (t0MLNSTICG FOR
",~iclentiui, . 11 1 U'lis).A v A I: I-X~ II (NJ* )L ( 0A)MPAI 1 LX iVMI.ASUIII;MYNTS 0V 1"I'M AI 'L TUIMlULVYfi1% Tn "'Y..
I H-' No. 31'4~~ CONTR., ,ACT OR CRANr nUM3FR
.1. HI. Brown G,. 1'aucilurT1%. Goodl
P. M. 1pc
Air Force Geophysics Laboratory (1,1W) I .. *R NIOP
Hjanscom A Vi) (;2 11) F'M a s.a chusetts 0 j,:31 fifr 8- 0T8
Aitr Yarce Geophyipics I-aboratory (LKI)) 2 4 February CBHanscom AFB F ;MR E 0,Massachusetts 01731 ________46 ____
M7*, , '4 .' A.' ¶ O'i S 'fI4**4 7, M~n.,I, . "-f ,- SECURIT4, CL4ASS. 'ýI 11
.1W 10 - N' IE
I'lefr'active Htrur~ure cous.tatit Scintilloinnftrilarlay, turbulerice RadiosondeIIerl-rIIsrl)dJIre T'urbulenice
t . ~AV(IU thesiistd art() aizx-raft tr ox-u -thermnal rnfaSUrerncnt5. DescribedI:le11i thle radiousrinde and thurtios'jnduý etectrunieni arid their calibrations.
Thie data i-'v'Ju,:tiun kS (Intr k f descriloed and a b~rif-f program desr:ription isgiveti. Altisungh the differersnt lnsti'urijunt.j gaive mecasu-remeýnts somewhat
DD 1473 Unclassified
UnclassifiedSECURITY CLASSIFICATION OF THIS PAOE(WIhe DlI*e EntIr'ed)
20. Abstract (Continued)
separated in space and time, first results show good agreement between theradar, scintl.ometer, and thermosonde measurements.\ý
UnclassifiedSECUAITY CLASSIFICATION OF THIS PACOE(Whs., Date Entered)
-A
.1
i-'-mu w - , ' . - - .... .
i .. ei"n".: , ...
IT
\.)
~i~ivt
Preface
This program owes its success to the dedication of numerous individuals.
These include:
Thermosonde construction and Mr. Hans Laping, LCCRadiosonde modification Mrs. Eovelyn Hnyni-, L].(
MSGT N. K. Newijian, LCCMr. G. Murphy, Tri-Con Assoc., Inc.
Radiosonde tuning and tracking Mr. J. Griffin, LCRMr. K. Walker, LCR
Radar observations Dr. Brenton Watkins, Univ. of AlaskaMr. G. Loriot, NEROCDr. Prabhat Rostogi, NEROC
Scintillometer observations Dr. A. Quesada, LED
Radiosonde balloon launch Mr. G. McPhetres, LCCMSGT J. Gordon, 1,CCC
Tracking data reduction Mr. Brian Sullivan, Boston College
Theoretical discussions Dr. Edmond Dewan, LKD
We sincerely thank the many named and unnamed individuals who in large or
small part contributed of themselves so willingly.
3
Contents
i. INTRODUCTION 7
2. RAWINSONDE 8
2. 1 hawinsond( Moditications 92. 1. 1 Met Circuit Card 92. 1. 2 Transponder Receiver 92.1.3 Transmitter 10
2.2 Resistance Calculation 102.2. 1 Temperature Calibration 112.2.2 Humidity Calibration 112. 2. 3 Pressure Calibration 12
3. THERMOSONDE 14
3. 1 Thermosonde Description 143.2 Sensor Heating Effects 173.3 Sensor Frequency Response 203.4 Therrnosonde Calibration 213.5 Thermosonde Environment 21
4. DATA REDUCTION 23
4. 1 Time Period Records 244.2 Reference Frequency ID 244. 3 Segment Window ID 244.4 Tracker Data 25
5. PROGRAM DESCRIPTION 26
6. IIESULTS 31
REFERENCES 45
knzczg~iMG PAGE RANK-NOT FlJ
S.. . ... .... .. , **,., T Ir-Illustrations
1. Sample Output of Reduced Radioaonde/Thermosonde Data iS2. Photograph of Thermosonde and Micro-thermal Probe 16
3. Schematic of Thermosonde Sensor Convective Heat TransferTest Circuit 18
4. Heat Transfer Coefficient Curve of Thermosonde Sensor as aFunction of Altitude 19
5. Calibration Curve of Thermosonde Sensor AR as a Function
of Radiosonde Frequency 226. Calibration Curve of Thermosonde (AT)RMS as a Function of
Radiosonde Frequency 22
7. Sample Output of Reduced Tracker Data 278. Typical Temperature Profile Plot 319. Typical Relative Humidity Profile Plot 32
10. Typical Balloon Velocity Analysis - East Profile Plot 3311. Typical Balloon Velocity Analysis - North Profile Plot 33
12. Typical Horizontal Vector Shear Analysis Profile Plot 3413. Expanded View of Horizontal Vector Shear 34
14. Power Spectral Density Analysis of Horizontal Vector Shear 35
15. Raw ThermosoNde C2 Analysis -- Profile Plot 37
16. Expanded Vie- of Raw Thermosonde C• Profile Plot 38
17. Smoothed Thermosonde Cý Profile Plot 39
18. Cumulative ProbabUity Distribution Curve of Log C2 4019. Comparison of Radar C2 and Scintillometer C2 41
N N Profiles 4
20. Scintillometer Weighting Functions 42
21. Comparison of Thermosonde C2 and Scintillometer C2
Profiles N 4322. Comparison of Thermosonde C2 and Radar C2 Profiles 44
24
Tables
1. Optical CR (Turbulence) Measurement Program 28
61
I
Sonde Experiments for Comparative
Measurements of Optical Turbulance
I. INTRODUCTION
Atmospheric properties ot temperature, pressurc, humidity, and wind speed,
as measured by meteorological radiosondes, offer a means of estimating the fluc-
tuations in the atmospheric optical index of refraction. Fluctuations of the index
of refraction, characterized as C are important to numerous DOD optical sys-
tems being both planned and developed. Direct measurements of C are possible
with optical systems such as those used to evaluate sites for astronomical tele-
scopes. These measurements have been limited to a few locations, and this tech-
nique is not suitable for assembling a world wide data base of C2 . Given that CN
determined from meteorological data can be an accurate measure of optical turbu-
lence, the best promise for economically assembling a large data base is through
the use of the archived meteorological sounding data obtained from sondes.
The present rawinsonde measurements are usually achieved with altitude
resolution of 300 m or larger. This scale is generally accepted as being larger
than the vertical scales of turbulence that produce the C2 . Thus, in order to de-
velop an analytical model predicting CN from the standard rawinsonde, it is nec-
essary to have a statistical base relating the actual turbulence scales to the rawin-sonde observed scales. New sondes (artsondes) have been developed by VIZ
Corporation for the National Weather Service to measure the meteorological
(Received for publication 19 February 1982)
7
S-t.AV''W----X
variables at smaller altitude resolution. These sondes were modified at AFGL to
provide data for a CN intercomparison test. The program objective is to compare
C as determined from standard rawinsondes, the AFOL modified artsondes,
AFGL built thermosondes. stellar scintillometers0 aircraft-mounted commercial
thermosondes, and UHF radar.
The dttailed description of the AFGL artsonde and thermosonde is presented in
this report. Also presented is the calibration procedure and the computer analysis
used to derive the meteorological values from the recorded telemetry signal. The
sonde tracking and ranging system is described along with the data processing pro-
cedures used to smooth the data and determine the winds as a function of altitude.
A brief description of experimental results is shown as comparisons of C 2 meas-AN
urements obtained with radar, optical scintillometer, and thermosonde.
2. RAWINSONDE
The rawinsonde used in these experiments consists of meteorological measuring
elements coupled to a radio transmitter and assembled into a small lightweight
plastic case. The device is carried aloft by a 1200-gm balloon filled with heliumgas. Included in the train is a small parachute to slow the descent of the device
after the balloon bursts. An automatic launching reel is attached to the train, which
lowers the radiosonde until the train is fully extended to about 20 m. As the balloon
ascends at about 5 m/see, measurements of pressure, temperature, and relative
humidity are transmitted to a ground station where the data are recorded. As the
balloon rises, it is followed electronically by radio direction-finding equipment that
tracks the transmitted signal. Measurements of the balloon position relative to the
ground station are recorded in terms of azimuth, elevation, and slant range. These
measurements ax'e converted to X, Y, and Z distances. The data are filtered as
described in Section 4. 4 to remove extraneous signals and then time differenced
to obtain wind velocity profiles.
Fifty (50) state-of-the-art radiosondes were purchased from VIZ Corporation,
Philadelphia, Pa. These! are called by them the "Artsonde-Phase 11, " and they
carry VIZ part number, 1499-311. This is a 1680-MHz amplitude modulated sonde
with a 403-MHz receiver used for tra.nsponder ranging. The transmission ofmeteorological (MET) data serially switches by solid-state circuitry from pressure
to temperature, to relative humidity, to reference, and back to pressure with a
segment dwell time of 250 msec. Premium sensor elements are used for the
aneroid baroswitch, thermistor, and hygristor. Depending upon the sensor inputresistance, a frequency of 30 to 1000 Hz is produced that modulates the 1680 MHztransmitter. The reference frequency is set close to 1000 Hz. It provides the
A i4 h
• . • '1• " , , ...... , •i: " .... : ..... " .. . •.. . .| " • " " .. ... ...i . ....... •.......... 'i .. ... ... ',
identification of segment position, and it allows for data correction of tenipcraturu
induced frequency drifts.
2.1 Ralinsornde NMoifications
After delivery of the radiosondes, sv,'.,. in-house modifications were re-
quired. Because greater temporal and spatial data resolution were desired than
were available from a standard radiosonde and its associated GMD tr-acker, special
payload and ground equipment changes were implemented. Additionally, a sepa-
rate instrument to mneasure small temperature fluctuations (thermosonde) was
attached physically and electrically to the radiosonde. The modification included
changes in the radiosonde transmitter, receiver, solid-state switching, and MET
oscillator circuit boards.
.MET CIRCUIT CARD
In order to transmit the additional thermosonde data, the number of commuta-
tion data segments was increased from the standard four to eight. To accomplish
this, a second quad-bilateral switch (p/n 4060) was added to the MET circuit iboard. This was facilitated uy the existing 4017 IC clock, which allows serial
switching of up to ten segments. Two of the additional segments were used for
thermosonde information and two were used as auxilliary segments. The auxillary
segmcnts allowcd supcr-commutation, uf tim MET dta ou r the reporting of electron-
ics temperature.
To accomplish calibration and ground-base checks the eight lock-up and one
common test points were brought to an eight-position rotary switch. During cali-
bration the normally free-running segment switch was stopped on a segment cor-
responding to the position of the switch. The frequency of each segment was
examined. in turn, by rotating the switch from one position to the next. The
measured frequency was then recorded along with the measured temperature,humidity, pressure commutator position, and thermosonde test value. The baro-switch that forms one part of the MET board transmits one of four unique fre-
quencies depending upon the position of the baroswitch armature. The original
separation of frequencies was only 10 or 20 Hz. In order to provide a separation
of 100 Hz between each frequency, three new resistors replaced existing ones with
appropriate values. 'This reduced confusion over contact identification.
2, 1.2 TIRANSPONDER RECEIVER
The balloon borne 403-MHz transponder receiver is used to derive the slant
range in the following manner: Ground-base equipment generates a PCM code that
is synchronized with the start pulse of a time interval counter. This code modulates
the ,k03-MHz uplink transmitter. The receiver transponder uses this code to
9 1
So.. .... ..... ..., ..... . ....:''- 1 . ... " . ... .... -'" ...
modulate a 70-kHz subcarrier oscillator. This subcarrier is mixed with the MET
data and used to modulate the balloon-borne transmitter. Ground-base equipment
retrieves and decodes the PCM code, which then @ends a stop pulse to the inter-
val counter. Since slant range is calibrated vs interval, a direct digital coded
range output it produced.
2.1.3 TRANSMITTER
Normally the transmitter is 100 percent amplitude modulated b" .Ae MET data
pulses. When the MET data pulse is high, the transmitter is -hut down. This
results in an AM pulse repetition frequency that is measured by ground receiving
equipment to determine the sensor input resistance.
The on-board 403-MHz receiver frequency modulates the transmitter. Since
this information is lost when the MET signal is high (because the transmitter is
shut down), difficulty In holding synchronization to the ranging signal develops.
Consequently the evaluation of slant range data contains associated errors. To
eliminate this problem, the ac-coupling AM input to the transmitter was removed.
Instead, the MET data was ac-coupled to the transmitter in the same manner and
to the same junction point as the ranging signal. This results in MET data fre-
quency modulation of 30 to 1000 Hz and ranging modulation of 70 ldiz.
2.2 Resktance Calculation
This section presents the MET data transfer equations as given by VIZ
Corporation (document nos. VIZ 800721 and 761101B). The form of these ex-
pressions as well as the constants depend on the sensor type and manufactu.ring
lot number. The input resistance to the radiosonde corresponding to the measured
MET frequency is:
Rinput e R it (refifmeas- D
where fref is the average measured reference frequency and I raas is the average
measured segment frequency when the baroswitch in on a space (that is, between
two discrete contact points).
The thermistor or pressure resistance is the input resistance when on a tern-
perature or pressure segment.
The input resistance when on a humidity segment is the equivalent parallel
resistance of a I-M resistor and the hygristor resistance; thus,
Rinput X 10 6
1RH - 6_- X 10 -input
10
S. .. ..... - ... i ..... i l' ........i .. .i ...... ...... i.... i" " i 'i= T " - ' ' ....
2.2.1 TEMPERATURE CALIBRATION
Temperature is calculated from L,.t. thermistor resistance hiT as follows:
Ii -273.15t 3
ý A k Ii, M T/R30)
ko
where R30 is the acculok lock-in resistance at 300 C. The coefficients for pre1Cit..i,
temperature element lot P6,7, as provided by VLZ are:
A 3.2987 E-03
A,= 4. 7610 E-04
A2 = 2.8417 E-06
A 3 = 1.5691E-06
2.2.2 HUMIDITY CALIBRATION
Relative humnidity is calculated from the hygristor resisiance R as follows:
}'•. k [ fIRH)H69.RH 11 A-
k0 RH
where
f(R RH)= f(t) in (1RH/R 33)
R is the acculok lock-in resistance at 33 percent 1L1 and 25
and
3
f(t) = W C. (t)k (t = °Celsius)
k=o
The coefficients fo-" premium humidity element lot 2265 as provided by VIZ are:
for RHRH/R 3 3 _: 1: for ,RH/R33 '1 1:
C = 7.885 E-01 C = 9.243 E-01o 0
C 9.286 E-03 C1 = 3. 059 E-03
C2 = -2.462 E-05 C2 = -1. 188 E-06
C3 = -3. 368 E-07. C3 = 0.
3I
it i I ..I I 1... ..i ...i i. . '
r
for f(R]RH) 2t -0.2: for MR.H) < -0.2:
P i. a -2. 0909 E00
D! * 7. 1454 E-01 D, a 9.2406E00
D 2 & -7. 9163 E-02 D2 = 2.1693 E01
D13 a 3.2555 E-02 D3 = 2.2214 E01
D4 a 3. 9255 E-03 D4 - -6. 1926 BOO
A * 102. A - 0.
2.2.3 PRESSURE CALIBRATION
Each radiosonde is provided with pressure calibration values for the leading
edge of 160 contact positions of the baroswitch by VIZ Corporation. Four discrete
frequencies can be transmitted by the baroswitch, depending upon at which of the four
classifications of contacts the baroswitch is located. The four frequencies were
adjusted to provide approximately 500-, 600-, 700-, and 800-Hz signals. Approxi-
mately 500 Hz is transmitted when the baroswitch is located between any two dis-crete contact points. Approximately 600 Hz is transmitted for the following contact
numbers: 30, 45, 60, 75, 00, 105, 120, 135, 140, 145, 150, 155, and 160. These
start at contact number 30 and increase by 15 for contact numbers less than 135 arid
increase by five fui contact numbers grester than 135. Approximately 700) Hz is
transmitted for the following contact numbers: 5, 10, 15, 20, 25, 35, 40, 50, 55,65, 70. 80, 85, 95, 100, 110, 115, 125, 1301, 136, 137, 138, 139, 141, 142, 143,
144, 146, 147, 148, 149, 151, 152, 153, 154, 156, 157, 158, and 150. These start
at contact number 5 and increase by five for contact numbers less than 135 and
increase by one for contact numbers gieater than 135, Approximately 800 lfz is
transmitted for all other contacts. Before each launch the baroswitch is manuallyset to the contact position corresponding to the ambient station pressure. During
analysis the computer identifies groups and subgroups of contacts and defines aparticular contact within a larger group. Thus, if transmission is lost for a short
time, the computer redefines the location of the baroswitch and estimates the pres-
ent contact number. Pressure is defined at the leading edge of each contact and is
logarithmnically interpolated between contacts based on altitude.Sample results of the pressure, temperature, humidity, and CN calculations
are presented in Figure 1. The first column is the record number. "Timne" isdefined as seconds after launch. "CTEMP" is the corrected temperature based un
inferred time lags. No correction is made to humidity below -400C. Pressure
contacts are split into classes 0 to 3. Column 11 shows that a contact number is
defined at the first encounter of a new contact. Columns 14 and 16 show that
12iA
• *..... - .i.iT T~•= -'•~ " i I •Ii I ii " !'4
-~ ~ ~~~ . .JW J W M 4 b f .1 .114 ?U c14-F.ClM0~l'@~lt ~ V4C, -1 V *O 0- 010 t0410Oi4C 0 0 Q
-00-01 000 0- 0 0 -c ý0 M -fl c 0 ~ ~ .0 m ..... . I00 n000In 000-00 I
- 0 0 0 - 0 .-"Ta 0 00 0 O *- 0 O 0 T02
o S. n - - -- - - 0 0 "
Illý II! I!11 I!I !InI
a wow*EI.4-~-U il.0lM 1 *l2t.a*11g~e~ *S,10C~i* *ro-. ono~q..c~q oc Orlq * 0C~nO . 7-04*00*l M o-o oooV cn -- 'M O T c.-oo M .0 0 . .pwc~
* . .4 . .* . . a ... . (.Ad
I a.2))0lW ~ W 0 .4a41. a7.iC'041 -- a%.C74 0 q0a0-a,4 . V
ooooooooooo -oo~ oooooo C(a qqqn~~~oo~o~~r~r~r..-----t I00 0 7 1 70 6 4 4 0 W Nrl--N41O 4 W
wz7- m74 mmm 212ýl0 4 mw
Me- 000P)12 00 0 912C7.0470.0' q141.41C'70-C,)fl21217444 0 n0) 44 0
15 w 0 . 0 7 7 4 . 1 7 ? S 2 c 1 J 0 v r o z ~ - b t t 2 0 c ' 1 . ~ ~ ~ -7* oo vv voo
0 o~ 1241414o41A .7m " " m
o4) 1 1 o 0 b 2 4 0 ) t 0 t 4 0 0 t ) 4 1 m0l
Cl I ~ ~ ~ ~ N .... I2 00 0 0 0 0 0 0 00 0 0 0 0 0 0 0 000 0 0 0 0 0 0
7 . -13
( rT)mras is undefined for values less than 0. 0020. The "* ' in the last column
flag values of (AT)rms low gain that are greater than 0. 035°. The corresponding
high gain value in not valid for this condition.
3. THERMOSONDE
The thermosonde is a small, lightweight, temperature-sensing electronic
unit that attaches to the radiosonde. It measures the difference In temperature of
the ambient atmosphere between two probes mounted horizontally I m apart. As
the balloon ascends or descends, the temperature differences are processed and
routed to the radiosonde MET segments. The received data is used to calculatethe atmospheric temperature structure parameter C2 . These data, together with
I Tpressure and temperature, are used to calculate the refractive index structure
parameter CN. Fundamentally, C2 is defined as: 1
C x [n() - 2)]/r
for n(r.) the index of refraction at point r1 and r = ri--i I. This parameter in thit
key unknown in estimating the performance cf laser systems through the clear.2atmosphere. Among those -xplicit or indirect dcpendencie-s on the Cpr'ile ar.. ..
log-amplitude of light fluctuations, wave structure function, receiver diaieter for
heterodyne detection, and modulation transfer function.
In flight, the thermosonde passes its temperature difference measurcre,:,t (if
T 2 - T 1 through a root mean square (rms) module. Ground-based procensi, g con-
verts this to the mean square temperature fluctuation (T, - T1)2, which over oric2 1
meter separation equals Cj. This is converted to CjN through the equation:
2 = 79.9 P(h)X X 10 2CN(h) CT Q()
where P(h) is pressures in mbar and h is altitude. 1
The original lightweight thermosonde system was developed by GTE' Sylvania
Inc. , Electro-optics Organization for the Goddard Space Flight Center in 117 1.2 . 3
1. Tatarski, V.I. (1961) Wave Propagation In a Turbulent Medium, McGraw-11111Co., New York.
2. 1litterton, P. J. , Mallery, L. E. , and Arken, T. A. (1.971) IhtwegyhtTh'ermosonde System., GTE Sylvania, Inc., l"inal l ~eport-NA•-114 93.
3. GTE Sylvanda, Inc. (1972) Lighhtieight Therrnosondc System Model 140(Instruction Manual), Mountain View, Caornia, fl4i-0
14
S. . . ... . . . .. .
. , . . .. :• . . .. .... , ,..
Thisi jaxotttyj'. rsynti''nt Witt)flow IJWIy till Witilomi w0iof. till Al,(It 1i,of
1 U N'uvcrnl ru 11171 I Lt U.bl.', 11it. hod li iio ith nx 'twult~. wniklyZit it fy .11,'•k I .. IHilie'l
of the Goddard Space Flight Center, NASA. 4
Mudlbatlonu 1',m fi Wilvi -)utit( ojprittlon ojf thiin hypitiri, v, ert railfti In If7'?" by tlt-i
j~Alit Fuitt cWcapo:'i Luji'vz miiy. liizti I mr Avii, V.. M tx. %&iUk cnntintanrie orji
ft orrlnul rl-illu JDyliaaitt lonl Cl~ '~wallot'. A imw L-1 Ovml 't6~ A oard I t yout afil cutin-
juiuit.Iirt 1t:julk- i'u t til 4. In wt '1 1''* ii.t. di Ilt A I(;I In l Io fii,- £ t Itttrx% 1"1 1
id I 'liltuI 1:t vwith it !P1 tll fI111,imfl v tf't t
t001 W ~lt (-it qut 4 -di ' It W.14n 11 f it' tnfty Wt tuft1
ritl f .L Urej'I;1A;:Ill V.; It' Ioill u r Lnjif ;yipta 0 (Hymir: 2) vib 10, ; gfu i nign fi,;'ift'ai -
tit Ituti 11th Y'C2( tmhfitI.Il.Wd waut tuatjtltyti lt- tirall1rtzIrg, n'rv1'vntu ;it lm hu f
tilt- A)-(.c;sIU liuiivi brartn' 1.
:i. I1 In ,frmintifusei Ir 1-11 uijt'llu
The theirmomonde monsists of lt.u fise-wirs twigastan probes vxpoflrd tu thit
ambient atmospherv and ueparated by I m, Each probe ham a reaiubanoo definedlit ulp I Ille [IIi I i' a ~ ttt? diffi ci' ti I.* IV 1-t . till `LAu I~ 6eTh,A'
In (it 't-i ,iiio (I fj n; , if w,(,u l o I'i vi-d duff': i, * ;i It. vltn utme'
0.1'li ki I . litU Ahtb t. . I 'I hliti n -t I 11jut li Lit Li'~ ' t -tI . 1 9.' iHiIf aw1ý 0 it it'
nunt Iu Iii tin 's Ar a ili.I blttru I flu"' t'atal''t hi .i II. lit hI'i l1j ie''J'lit V, titI' l
jtari tVId It oJuw ti; 'J:u;,''lt ';til ,1- t1 iIttiia 1, .'li', 1.1: i'ft, I t it't li-uj:.hii, itt P Iiol, it
;w m uit te 14cl't i' t It' Ti ft ' t ' ,V II I'll tii','ha l t- 1 h'Ill. t l 111' 1 (t Utlj T ll to tt I 1' IN'
42 Hu'i' i~ttz vI Y U01 tai l '311 v1.Y -)kfhl filtt I li't 'fI Y. I'N'.t: -3 01J1 1 utolii'L tfi'I itta MIl 11
4 li fto, .1. 1 , ('17:5) 1 tiit I itie Jiridf Alv'tt I ii Wto bidtltt;'t !Iata v.l it
til--i ut jowt Ill njtli' il' -t Ill~I tilt~ t AS ol ' Nli 711.7
7. Mu: jihy, P;. (;~i V, 'I 'It Itleirti- Mill IStV ' th b I fillQ~j bt/u~~~~
AIt1ý1 ~io~, 5
mOp,
1F:I'ure 2. Phot',graph of "T'hermoFbonde and Micro .Thermal Probe, Tungsten wire(Iatrnitr!r 3.45 min. Effective length is 0. 45 cm. Nominal ice point resistance is27 uht,, awl itst t e.i.pcrature co efficient of resistance is 0. 0u375OC-1. Sensorwireretou:hed for Iettor photographing
IwoO 1iz and 0.2 liz, respectively. A 1-sec rma average is performed on the
filtered signal.
In tile cas e of a squar'e wave Fignal, the rms resistance is one-hall the peak-
to-peak rensitance and Eq. (2) becomes
rms )(/(2crl })
T'ihe fijri 99. f8 zi , rCeit palre tungtiten wire is m anufactured by the: Sigm ond
Colc, wCorp,. Mt. Vernon, N.Y. , and the wire is attached to a probe assembly
by Datumetricus, Inc. The normil at wirv diaine'ier Is 3.45 pn, A 0.45-cm length
of the wire is mounted hetwccrn two 0. 5-cm long wire supports extending out from
the probe holder (Figure 2). Its nominal ice point resistance, R10, is 27 ohms at
0()C. "J'hin bets thie elctrical resistivity at 5. 9 uQ2 - em. The cemperature coef-
fitient of 2,11 tarlcC was imiCEIsured at 0. 00375)C-1. Note that the electrical
If-;
.1.
r%.y.
resistivity and temperature coefficient differ from tabulated values. This is pro-
bably due to a low annealing temperature of the drawn wire. 8
Opthnum performance of the thermosonde can be achieved by passing a current
through the wire that is large enough to enhance the temperature response, but
small enough to avoid significant velocity response. Since a cool ambient fluidflowing over a warmer body will lower the temperature of the body toward the
fluid temperature, it is desirable to keep the current-Induced temperature dif-ference below the minimum detectable value of 0. 0020 C rms. The maximum cur-
rent allowed for not exceeding this temperature difference can be estimated from
a heat transfer analysis. The current generated temperature difference, tTw.can be determined as
•P.
ATwYKJi * (4)
where1 w P is the electrical power input between two conditions
J is the mechanical equivalent of heat 4. 18 (Joule/Cal)
As is the surface area of the wire = 5 X t0-4 (cmr2 )
h is the heat transfer coefficient (Cal - es . - LI).
The power input is P = 12(RA+-P,) where RA is the sensor resistance in the ab-
sence of an applied current and AR is the change in resistance due to applied
current.The heat transfer coefficient of the actual thermosonde probe was determined
in the laboratory by the temperature difference between two power levels. A spe-
cial bridge test circuit (Figure 3) was built to provide a constant current across
the bridge. This current alternatively can be stepped high and low, depending upon
the input voltage E. The bridge output, E0 , was amplified by a factor of 83. 33.The additional wire resistance, R, caused by the current heating was determined
from the observed voltages and current as
t1 - E /83.331i (5)
The current converter input voltages were stepped from 0. 5 V to 5 V, which pro-
duced pjibe currents, Ii. of 0. 5 mA and 5 m.A, respectively. The observed
8. Sm itnc2)ls, C.J. (1952) Tungsten, A Treatise on Its Metallurgy, PropertiesAnd Appblcat ions, Chapman & Hall, London, 3rd ed.
I'r
17J
I
°106k 160K-
lOOK 1 2KK,.-K A
50RIDGE " W500 OHM
-
Figure 3. Schematic of Thermosonde Sensor Convective Heat Transfer TestCircuit. For a given input voltage, Ein, the LH004I circuit produces a con-stant probe current of (Ein/(2 X500)) amps. The 725 amplifier provides an out-put, E - 83 33 X Bridge potential difference. Thus the differential resistanceof the gridge due to probe 12 R heating is: R = Eo/83. 33 Ii
outputs were 0. 01 V and 0. 32 V. The temperature difference produced ini the
wire, ATW, is
ATw = (R5.0 - 0. 5 )/ aR , (6)
where in this case R =26.28 0, AR 5.0 -R0.5 H 0.528 0, and ATW = 5.360C.
The difference in power between the two test conditions was determined as
AP z P5.0 - PO. 5 = (24.75 RA + 60 E 5. 0 - 6E 0 . 5 )puW with a measured RA = 28.25 Q,
AP = 718 pW. From Eq. (4), the experimentally determined heat transfer coef-
ficient is found, h = 0.064 cal , s"1 . cm- 2 . oK-1.
In flight, using a probe power of 1/3 MW rms produces a self-heating temper-
ature rise [using Eq. (4)] of
AT = 0.00250C (rms)
Consequently, self-heating effects set the thermosonde noise level at 0. 002 C. The
actual flight condition heat transfer coefficient depends upon air velocity and den-
sity. A study was performed to determine the convective heat transfer coefficient
18
• .a. -s
as a function Reynolds number. The heat transfer coefficient h is described in
terms of the Nusselt number Nud as h a Nudk/d, where k is the thermal conductivity
of the air and d is the characteristic length. The Nusselt number can be calculated
from the Reynolds number Red over limited regions.9 Calibrations of h vs pressure
and wind velocity were performed in a bell jar/blower vacuum system at conditions
corresponding to the surface level, that is, p = 760 mmHg, T = 24°C,
velocity= 5m - s- (corresponding to the vertical balloon ascent rate) Re 1 1. 2;
h is evaluated to be 0. 12 cal - s-I • cm"2 . o°KI or double the "no-wind" condition.
At lower pressures, the bell jar density is correlated with the standard atmosphere
to obtain altitude. An altitude profile of h is constructed from these calibrations
(Figure 4). Above 29-km altitude molecular flow occurs. Here, the free molecular
heat conductivity given by Dushman is used.' 0 It is clear that an increase in velo-
city (Reynolds number) at standard conditions enhances the value of h and lowers
self-heating even below the specified 0. 0020 C. At the maximum expected altitude
of about 30 kIn, h 0.031 cal. s- cmI 2' °.K', and self-heating can be ex-
pected to approach a noise value of 0. 005 0 C.
4 0 \%
4f4
0.031
30 :I
t320-
10 REYNOLDS<EFFECTLOWER LIMIT FOR----
VISCOSITY FLOW
0-3 10-2 0.064 10-I
HEAT TRANSFER COEFFICIENT, (vol. s-, cn" 2- kOli)
Figure 4. Heat Transfer Coefficient Curve of Thermo-sonde Sensor as a Function of Altitude. Bell jar cali-bration. At STyf, h is _measured at0.064, cal ' - cm-2 °K-I. With air §peed5 m "-, h- 0.12 oa1l cm- - °K-` at 30 km,h .0.31 cal. s- •cm .K . Above 30kin,free molecular conductivity applies
9. Maron, L. (1981) Fluid Dynamics, Academic Press, New York, Vol. 18,Part A, p. 27 1.
10. Dualhman. S. (1962) Scientific Foundations of Vacuum Techniques, John Wiley,New York, p. 47, mair
19
I3.3 Seno Frequency Respons
If the temperature of the environment T. suddenly changes, one wishes to know
how fast the sensor will follow the environmental changes. Assume that the wire
initially has a temperature Ti and is to be brought to an environment temperatureTe, The time for the wire temperature to recover a given fraction of the environ-
mental temperature is given as:I1
0 -2abt (7)r
where.
00 Tmax - T e
0i Ti - Te
r wire radius (cm)
a thermal diffusivity of the wire (cm 2 /see)
b h/k
h - convective heat transfer coefficient (cal s- 1 . cm 2 °K'l)
- the maximum temperature in the wire after time, t
k = thermal conaurtivityof tungsten (cal -s- I cm- 1 o K-l)
x 0.426 cal- s- cm-1- OK-1
a --
Pw tungsten density - 19.3 g cm 3
a heat capacity of tungsten = 0. 032 cal• g 1 OK-1
2.-a 0. 69 cm s
b 0. 16 cm 1
t - time (sec)Thus, the response time required to recover 0./0i 50 percent and 0 0 /9 = 90 per-
cent is 0.4 msec and 0. 72 msec, respectively.
This was verified experimentally. The sensor was placed in , bell jar vacuumchamber. A current pulse train of 0. 5- to 5- mA amplitudes was applied to the
sensor and the ac bridge signal response was displayed on an oscilloscope. It wasobserved that the bridge output approached 1/2 of its full-scale change in 0. 3 msec.
This gives a frequency response fl/ 2 - 3570 Hz. It is clear that the sensor has an
adequate temperature frequency response for the purpose of turbulence
meaaurements.
11. Jakob, M. (1949) Heat TransfeL, John Wiley, New York. Vol. 1, p. 291.
20
IA
3.4 Thermosonde Calibration
The thermosonde measurement is transmitted as a frequency value, nominally
ranging between 30 and 700 Hz. Each frequency is related to a resistance which
is defined by the observed temperature difference using Eq. (3). The relationship
between frequency and resistance is determined by actual observations. A thermo-
sonde test box was fabricated that alternately switches a test resistor in and out
across one of the pair of matched 27 Ii resistors to simulate the probes. Any one of
12 precision resistors ranging from 1. 4 M 0 to 20 k Q can be inserted for the test
resistor. These parallel resistors establish a known equivalent resistance. The
differential resistance (AR) across the bridge is then,
,AR R -Rps eq
where R is the simulated probe resistor 27 Q, andps
R XRR eq RPs +a
ps s
where Rs is the inserted test resistor. Thus,
2
R = R
ps s
For each thermosonde a table is constructed for AR vs radiosonde frequency.
6Trms is calculated from Eq. (3) and a polynomial fit is performed on this data
in order to provide the transfer function. Typical calibrations are shown in Fig-
ures 5 and 6.
3.5 Thermoeonde Environment
The thermosonde was suspended 18 m below the sonde balloon. The balloon
diameter expands from approximately 2 m at launch to about 10 m at 30 krn.
A working rule is to locate experiments at least five diameters below the bal-
loon to avoid wake effects. The 18-m suspension length is a compromise. A
longer length can enhance a pendulum motion that would interfere with measure-
inents. Thus, it is expected that the therroosonde will be located inside the balloon
wake except during instances of large shears. However, even with the thermosonde
inside the balloon wake, it is not clear what type of thermosonde error, if any, the
wake would produce.
21
'I
R91
.008 R
"I RG.007
' R7 HIGH GAIN
S-R6 ,1RIP.-z__ . 5 RII1- -.05
-.04R5 oR0.04
R5-
.001 HIG GAIN4 .01
.0 R6 RI3
I00 200 300 400 500 600 700FREQUENCY (Hz)
Figure 5. Calibration Curve of Thermosonde Sensor6F% as a Function of Radiosonde Frequency. Testresistors are switched quickly across simulatedprobes. This prnvide.s a calibration frequency foreach lmown differential resistance
.05 - .25h e HIGH GAIN CALIBRATION
.4 LOW GAIN CALIBRATION.20
LOW GAIN IS-.03 15
-�3 nHIGH GAIN 10
.01 _ .05
/]Hr I 1I, I I J
100 200 300 400 500 600 700FREQUENCY (Hz)
Figure 6. Calibration Curve of Thermosonde(AT)rms as a Function of Radiosonde Frequency. Fora probe of given resistance and temperature coeffi-cient, the resistance vs frequency calibration curveis converted to (OT)rms vs frequency
22
L I --
The existence of R wake behind a slow moving balloon (Re - 105) could gener-
ate a temperature fluctuation due tc two mechanisms. The first is due to turbulent
mixing of the atmospheric ternperFture structure over some vertical scale length.
The second could arise as a companion to compressible flow fluctuations of the
velocity in the wake.
In the troposphere, the sonde is usually in supercritical flow, and the sepa-
rated flow produces organized vortices. At about ll-kms. the flow becomes sub-
critical and the flow around the balloon turns into a randomized separation and a
uniform turbulent wake. Analysis of the thermosonde data reveals no significant
differences between tropospheric and stratospheric data. The thermosonde data
indicates the probability of C2is log normally distributed. Power spectral analy-
sis of CT shows that no frequency spikes or features exist over the measurement
range of 2- to 200-sec periods. This implies that any wake effects arising from
atmospheric structure would have to arise from scales smaller than 10 m. The2thermosonde observes layers of high CT over scales larger than 20 m as well as
undetectable C T (S/N < 1) for scales larger than 10 m. This strongly suggests
that the existence of a wake does not cause false signals.
The explanation of why there is no wake effect could be due to the rapid dis-
sipation of wake gradients. The small scale of the wake (length less than 10 m)
allows rapid mechanical mixing to a still smaller scale followed by rapid thermal
dissipation. We have made measurements of the rate of energy dissipation, F, in
the wake of high altitude aircraft. Film density analysib of aircraft wakes marked
by creating a smoke trail of titanium oxide hydrates shows the energy decay rate-4
to follow a t dependence. Thus, for a vertical rise rate of 5 m/see, the balloon
wake energy should be expected to dissipate to 1/256 of its original value by the
time it reaches the thermosonde.
Confidence in the absence of wake effect will be developed as more of the data
is analyzed and C2 comparisons are made with the scintillometer and radar.
4. DATA REDICCION
The "raw" transmitted MET signal is bandpass filtered at 400 Hz and 5000 Hz,
after which it triggers a "one shot" comparator. This results in a "clean" train
of constant-amplitude/constant-width pulses equivalent to the MET frequencies.
A Hewlett-Packard model 5326b timer/counter measures the time period of every
other frequency period and transfers these times in a binary coded decimal format
to a Honeywell computer, model316. The computer packs a magnetic tape eight
bits per word in a pre-determined record size and format. These tapes are then
unpacked and converted to CDC 6600 computer compatible 9 track tapes.
23
.1
4.1 Time Period Records
N Running time is determined by setting the algorithm time e Header time +
2 E T where header time is the actual Zulu time at the start of each compu-im true'
ter record. The factor 2 appears because the counter measures every other peri-
od. The true time period is:
T tape speed input to countertrue =tape" apeed of original recording X measured counter time
For this series, the tape speed counter input was 1/2 the original tape speed. If
any T true is less than 0. 91 msec, it is considered a data glitch and dropped from
the summation and analysis.
4.2 Reference Frequency ID
The unpack routine is initialized to the measured reference frequency at launch.
A forward test searches for three successive data periods with a frequency of the
initialized value : 5 percent. When this criterion is achieved, a backwards test
searches in reverse from the first of the three found frequencies to the first en-
counter of the initialized reference frequency ±50 Hz. Once aloft, the reference
frequency is allowed to drift slow"' 'om its initialized value.
4.3 Segment Windnw ID
After locating the last reference period of the previous reference segment, the
time interval of the entire eight segment commutation frame is measured by sum-
ming each period and multiplying the result by two as described before. This de-
fines the frame time. Since there are eight segments, the frame time is divided by
eight to identify each segment interval. Since the frame rate is set at approximate-
ly 2.0 sec, the segment dwell time is approximately 250 msec per segment. Only
the center 60 percent of each time segment is used in the frequency determination
for that segment. The first and last 20 percent of the segment time is dropped and
the remaining periods are averaged. Inverting the average period provides the
average frequency for the segment. The standard deviation of each segment inter-val is calculated, and if aref exceeds t 10 Hz, or if a temp a.., or apress exceeds
± 5 Hz, the two smallest time periods are dropped. In most cases this removes
spurious pulses. A new average and a are redetermined, and if aref exceeds
+ 10 Hz, the entire commutation frame is dropped from analysis. If utemp, 0 RH,
or Upress exceeds ± 5 Hz, the calculated segment frequency is considered erron-
eous, and individual bad segments are dropped from analysis. In addition, if at
least two time periods are not found in a particular segment, that segment is
dropped from analysis. The segment sequence per frame is, in most cases:
24
• =• . ... . I .. . -
reference, thermosonde high gain, temperature, relative humidity, thermosonde
low gain, pressure, temperature, and relative humidity. When the thermosonde is
not flown, those segments usually contain second samples of pressure and relative
humidity.
4.4 Tracker Data
Due to the intrinsically low signal sensitivity of the on-board receiver, and
to tracker hunting errors, large fluctuations occurred in the instantaneous position
measurements of azimuth, elevation, and slant range. These data were refined
after each flight on a CDC 6600 computer. The data. which were recorded every
0. 1 see, was edited by passing through a first difference software routine. Dif-
ferences between adjacent points greater than two standard deviations from the
average deviation were flagged. The average first difference is defined by:
n (X -xi..1
7- X n- IP2
Where n is the number of points and X. is the value of the ith point. The standard
deviation is:
-n ]
a = 1 - 1 -2 ý Xi xi I(i i- il) i=2
i2: -L
To smooth through the flagged data, a search was made to find four statistically
acceptable points on either side of the flagged values. Once found, least squares
polynomials were generated through this eight point array. Root-mean-square and
Durbin-Watson criteria were used to automatically select the optimal fit, up to
fifth degree. Polynomial interpolation then replaced the flagged data,
Smoothing of data within ± 2 a was performed by replacing centered data with
a 3 X 3 moving average of neighboring points. The average is defined by:
X -/9Xi- 2 + 2/9xi. 1 + 1/3Xi + 2/9Xi+1 + 1/9X11+2
Simultaneously, a second arithmetic running average was performed on the
resulting data to create 10-sec averages spaced at 2-see intervals. Furthermore,
the data went through a low-pass digital filter having a -3 dB band edge at 0. 01 Hz
that drops to -60 dB at 0.02 Hz.
25
The smoothed azimuth, elevation, and slant range were then transposed to X,
Y, and Z Cartesian and geodetic coordinates. Using data time spacings of 30 sec,
north, east, vertical, horizontal balloon speeds, and horizontal wind shears were
calculated.
Sample results of the balloon position, velocity, and horizontal wind shear
calculations are presented in Figure 7. The data are interpolated for even spac-
ings of time every 2. 0 second.
5. PROGRAM DESCRIPTION
The general researci, objective of this program is to compare the measure-
ments of atmospheric turbu!ence of several different experimental designs. The
primary instruments consisted of the MIT Lincoln Laboratory 440 MHz 12 turbu-
lence scatter radar and an Air Force Weapons Laboratory Stellar Scintillometer.1 3
In addition to the primary instrumentation, radiosonde, thermosonde, and air-
craft instrumentation were used to provide a broader comparison base. AFGL
operated the sondes and scintillometer, while personnel from the Northeast Radio
Observatory Corporation (NEROC) operated the raar. Aircraft measurements
were provided by Airborne Research Associates, Bedford, Mass., under contract
to AFGL.
From 10 Feb 1081 until 8 Aug 1981, over 30 radiosondes were frown. Twenty-
nine of these are listed in Table 1. The preliminary qualification flights, up until
21 May 1981, were launched from AFGL. The remainder of the experimnents (22)
were launched near the NEROC Haystack Observatory at Millstone Hill, Westford,
Mass., during the month of July.
Scheduling conflicts, time of day, and equipment availability dictated, to some
extent, the particular experiment on a given day. The scintillometer could be
operated only on cloudless nights. The aircraft required co-ordination with the
FAA, and it was further constrained by bad weather and other scheduling difficulties.
The radar was shared by other groups at Millstone, and it was available for these
measurements only on those dates pre-arranged weeks in advance. The radiosonde
*As of this printing, a detailed description of the positional data was beingprepared for publication as an AFGL technical memorandum. The title is:Analysis of Balloon Trajectory for the Calculation of Wind Velocities and Shears,by Sullivan, B., et. al.
12. Watkins, B.J. (1!81) Millstone Hill Doppler Radar Observations of Winds andTurbulence, Middle- trpheric Program, Handbook for M•P, Extendedabstracts from Int. Sym. on Middle Atm. Dyn. 2:30-36.
13. Ochs, G.R., Lawrence, 1.S.,. and Wang, T. (1976) Stellar scintillationmeasurement of the vertical profile of refractive--index turbulence in theatmosphere, Proc. Soc. of Photo-optical Inst. Eng. 75:48-54.
26
,:$
aa
I A ... . . . . . .. 0.. . . .. . . ... .... . . . . .. U. . . . . . , . .. . . . . . . .
uim. . m... .. e. . o.. .. u... ... . . ... o... .. . ..... .. .... .. . . "...
r -rErmmm mmpm ewam oiasmmwem~ eaaaaamam aaaam ammmmaaa
ASP
Li . .... ... ... .. . . ...oe m m -rc ..... ... ........ .............. ~nflV
o mm
5 .p. . . . . . .V W IS.. . . . . . ... .lC.. . .. . .. . . ... .. l + . . . . . 0~... -f, l aS •pmr m nnN nnmn. 44 ummwtnwwmpmr-r-me-,---e--w-mm~m~
S * AVt st Mosns 1 11 1 mill mas Nss git. si owl1a111111na
ftJ r4I 40M1, n
S -• Ct
N- UNM N -n N vnMan tA V I n r-9 T In0WW .I r -t-fl l sts - t-a awct nn#-tns- Ma
a-- - T- -- c -- - - - r- .- -t~9~ene -tOf? -' O-VU-w
.413 *
In~~~~O t a0mw i~ its a m o ~ o O O~ 1-
*-!
23 . .W. .wwwo 1,waaCC weVa ...... i-N• tz 5g 0 5 55 5 5 till ii Is il 56 itl l ,5 issis
w Am
flN5.~C .-oo O . o .ooooo.oo.o a C
ho fI--N - -- - - -- - - - --- --- NI-- NI - -N I--N--
W~~~ -e-I-- N~-~N NNNC-N.Nr, N-.-l?~t-0cN~~p.NlI-np~I-lncNCInt~-.enC j
o..... gnink ein tm wwm mw wi &nata nc, uw)mkinin nL nL n inmW wi6nOviwi mw ommmin W WVn U an ino ni
40toac! . . .. .. . C! e~ce~c!.'ý .'ýlcý' R
A' n nWAn nWVVVVAU-WVQVAWUVWvWv~vWWVWWWVV vvmWUW tvVaQYaWaaW
Wa -oQ c ~!1 ~!c ~ C!it nfn tnN -q o.N - tinN.- n0In0I04-"Int at mk-Ain , r. "nt--W~U 0 C Am
oft.m. mmm* mg m maowmaoaoaommaamQsnmgo om m o o c mmom*nanonccio o b
-9 - - -- c" q," a " e""I""c
NW' N . N f9-. tw dieNwaiN qc.-drq n .w.c% NvNfl n Nmn nt ml c e-NN-4sflflt ft
%j~ ~w aoc 27
-C- mamagmwwwvwW a n in c W W C -- tNNtt
20 Au Iwo 1 VI u I al * i 2*
|fl u 0- .- ... 5 . i-N C - .
1-0 on a X. 10.. t o U 0 N , , 0 . . . ,vW.0 W.t• -- ' Is4' :• ,
-,-- ac- -- )-N,-- --,- -I -- - I- -C -I --
b3•1l• - -oN Io . O O
*I• , ! , : I, . 1 ,* . 9- '
K I-O • . .
•I -, o l • ' • • I • -"" U' : ' •
H. . . ..O----I-.... r--- -k- --1K- -K j jv-r;E !~ d d d , ' !d•
- I
v ' I a '"I-4-------I - , -,-.----- H n - H- .. H. H--. -r- H- • -
' I ' I I II0 X E'
LLI
r; t= Wi , o go a) I :•
u w D ' I
S* I - II ~ ul 0. 0 I-Ii
C, t w 1 c j -
' ZI I•P > Q.
• L-- -•-L --- •--L --- H--- - •.. -- •--H---.. .. •: ' • • ,
C- -. '
' 7i :" • I I • ;• ' • " • , '
I ' ' ~ ~ I' o ..Q, oI: cr I . 01 iy C
28
S< =l •I • ' ' ' ' ' , ; ?
--- - -, - - --- -, - ,--- -- I- ... ------ t ... iI ;.a. ; -.aMI •, , ' •, 's. -I-
"I.:; ; Ap l ii Aa il
I - I . 04.. , I
9I , . I ,p ,. I • . I .. . , .
IKol i II •., -,k-I v i I ib .. , I' f) 1,
jI * I , , ,
I T - - I... l.. . *--I - - - -- 4- - .. .9-9- - ,- -I .. L ' I :L : .,L.. N 4
, , ,F MLII , V V L Li I .I , , : , , , I , I
r- .. .• - Is . .I. . . . i ..... 4 4 I - ! . . -IF - . - -I
'. 4 '4 I i ' 4 '4 i'" ' i
. II I II ',* I .- • •, •
.. . ... I .. . . . I C.. .. , .
.-. ,. - , , . I .. .. . . . -
I- I-. . . .. - - -. . . • . . . . F.. . • . . . .. . • - '.. . .. 4 --,. . .
• ,,s : ' i .. : : o , .- ; A
! , . , ., • • ., • . . , ., I , • l2-,
1I
i PI .. n '.. "
i2*'o, Q Q- lt I ,.,<*• o5I '<I I I +
I• , - " > I . : , I ' ,I
I i
1,1,
8 ---- 1 -- --; ----- - -- H-- -- ;.-H--H -- •-I-- --k--I
Q , ' : :. :. . . . .-.
•. -; .. ,
"I-_, ;- • [ ... . . . _ • _ • • + ,C, ' ' I
'- -------H--- • -------- ---- ----- I- -- L- F- --- -- I
ۥ- " .4<1
I'i • -, r-" -, • I F-. " -~*-• - - - - - • - - - - - - H r- - - -H----
* .t-•H m H- . ,
L' • I ' I-d- .. . .. T.. - - - • - ' . . .'I-
- " Ii ... ,+ I I , I
C , I ,- ' , C , ,
I ,,
...I I . . .II I . .. . .. .. .II. . .. ..I
and thermosonde had no constraints other than availability of tracking personnel.
As it turned out, the scintIllometer and radar operated together only on one even-
ing. 31 July 1981. Good quality experiments using three techniques, radar,
thermosonde, and aircraft were accomplished twice. Good quality radar and radio-
sonde measurements were completed on fifteen occasions. Good quality radiosonde
and thermosonde experiments were operated on eighteen occasions. Radar, radio-
sonde, and thermosonde operated successfully on eleven occasions.
6. RESULTS
Results of typical experimental runs are presented in the following plots.
Figures 8 and 9 show temperature and relative humidity profiles obtained 31 July.
A small temperature inversion is found at 3 km that roughly corresponds with a
high negative gradient in relative humidity. A more pronounced temperature
inversion is found at 12 km, defining the tropopause.
QO2O MT DTA
24.0-
21.0-
18.0-
I S. a- ,
S12.0"
9.0"
6.0-
3.0
-si -4.0 -31.S -A-00.5 1.o0 1.5 2s .0TEMP (DEG C)
Figure 8. Typical Temperature Profile Plot. Processed result
31
• (
¢QM~ MET CPTYA1.0. T
7.0
6.0
7.0
6.0
S.0
S4.D
.- , 3.O
2.0
10 20 30 ZO so 60 70 so
RELATIVE HUMIDITY (PERCENT)
Figure 9. Typical Relative Humidity Profile Plot. Processedresult
Figures 10 and 11 show the east and north balloon velocity component profiles.
The maximum horizontal wind occurs at 10.7 km at 35 m/sec. The detailed wind
structure is clearly evident.
Figure 12 shows the horizontal vector wind shear, SHR profile to 30 km. SH
is defined as the resultant of the east and north horizontal component shears, or
SH2 =N + S2. Figure 13 is an expanded view of Figure 12 in the altitude range-1
from 20 krn to 25 lun and shows the mean shear at about 0.02 sec . This plot
shows structuring of the wind shear down to the spatial filter wavelength of about
120 m. The power spectral density of the winds are plotted in Figure 14. The
slope is about -2. 5, which compares well with other measurements. 14
14. Endlich, R. M., and Singleton, H. C. (1969) Spectral analysis of detailedvertical wind speeds profiles, j. Atmos. Sci. 26:1030-1041.
32
QIX002 FINIA TRIO4rT 4NOLVSIS
6.7O
S11.z
3.7,
S-is -o -is A -'s w a mVELOCITY - EAST (M/SEC)
Figure 10. Typical Balloon Velocity Analysis -East Profile Plot. Edited and filtered
QXOOM FINAL Tq••T ANALYSIS
•I•.O
3 .S!1.2
-AS A-0 -is tj its 1' 4VELOCITY - N0RTH (M/IS:ll
Figure 11. Typical Balloon Velocity Analysis -North Profile Plot. Edited and filtered
33 i
XOMOO FINAL TRlMT ANAYSIS
3.7
-. i 6)1 .&si 5.6-m 6.8-w 0.WS 5.1% . o.8- co
514 ,VERTICAL 1141HAR OF I4OIZONTAL WIM
Figure 12. Typical Horizontal Vector Shear Analysis.Profile plot. Shear is calculated over a 30-sec dataspacing
OX0M0 FINAL TRAOT ANALYSIS
M HVETIA OHEA OFM ZNTLWN
Figur 13. xaddVewo7oiotl etr er
Th sher aliuersltoni bu 2
Z3.IZ34
ZZ.~ so,
02(20 MGO 126 POINT FFT. *IHING S'TOOIEO.
to
S101
(-I
-. . . .. It,-
WAVE NO. (CY/M1)
Figure 14. Power Spectral Density Analysis of HorizontalVector Shear. The slope iss about -2. 5
Figure, 15 shows "RAW" thermosonde C2 values as a function of altitude. The
2 N
mean C• tends to decrease logarithmically with altitude. Figure 16 is an expanded
N
view of Figure 15 in the altitude range from 10 km to 15 kin. A large-scale max-
mum shows up at 11. 5 km while the fluctuations in C- Npert edet h ml2
sampling interval. C N smoothed to 120 m is plotted in Figure 17 for comparison.Figure 18 plots the cumulative frequency distribution of C 2. Th~e data follows a
UT
straight line, except at the endpoints, which confirms the log normal distributionof C2N"
IL The purpose of the optical turbulence program was to compare various tech-
niques of observing C N. Sonde measurements were important in the low altitude
comparison between radar, scintillometer, and thermosonde. The radar-received2
power, scattered by refractive index fluctuations, can be related to a CN. This
C• is caused by turbulent fluctuations in both temperature and water vapor concen-N 2
trations. As the comparisons are based on optical wavelength CN the contribution
of water vapor must be removed from the radar C N.
35
lam
WAV .O IC/N
Vetr ha. h soel;ao t -2.5 1Figure.14. Poe SpcrlDniyAayi s fHrzna ,-
The fluctuations of the refractive index across a layer is given 1' 15 as:
M x -77. 6 X10' 6 1+ 15,50i4 - 15, 5004j 8
where
j
P x pressure (mb)
T g temperature (Kelvin)g , gravity (m •s2
q s specific humidity (gm/gm)2 = squared Brunt-Vaisala frequency (s-2) r (g/6) 0'
qI = vertical gradient of specific humnidity (m" ')0' = vertical gradient of potential temperature (Kelvin/m)
The radar CN can be corrected to dry conditions, CN (dry) by evaluating thebracketed terms in Eq. (8) on the basis of the sonde measurements of P, T, and q.
2 2 1 _ _rC'N(dry) - CN(rmdar) ft1,00 -15, 500g ý j2
2
The comparison between the CN (dry) derived from radar measurements and the
optical scintillometer is shown in Figure 19. The optical scintillometer 1 3 obtains1 ata over very large weighting iunctions. The scintillometer i/e half-width varies
from 1.4 km at an altitude of 2.2 to 4. 9 km at an altitude of 14 km. The weighting
functions are shown in Figure 20. The comparison between radar and scintiliom-
eter shown in Figure 19 is obtained by passing the C2 (dry) through the six weight-
ing functions.
A comparison of thermnosonde data and the scintillometer is shown in Fig-
ure 2 1. Here the thermosonde data shown in Figure 15 is passed through thescintillometer weighting functions. This agreement is good. The large smoothing
of the weighting functions hides whatever variations may exist due to the space andtime differences in the respective measurements. The scintillometer viewed the
sky vertically while the thermosonde followed a slant path.Thermosonde and radar comparison can be made over very nearly the same
volume and space by the fact that the radar can be pointed in the direction of the
sonde. The radar C 2 (dry) is obtained over one-minute periods and averaged over
15. VanZandt, T. E., Gage, KS.., and Wjirnock, J.M. (1981) An Improved Modelfor the Calculation of Profiles of CN and e in the Free Atmosphere fromBackground Profiles of Ternerature, and Humidity, 20th conf. onRadar Meteorology, (3" Nov--3 Dee181
36
4j
L -,o -,, .
five minutes. A composite of the radar results is made to match the sonde altitude
and times. The comparison of radar and thermosonde CN values is presented in
Figure 22. The thermosonde data has been averaged over 0. 50-krn intervals.
The comparisons shown indicate the thermosonde radar and scintilloneter
are in quite good agreement. The data shown represents the results from just
one day's measurements. Detailed comparisons involving all the data sets will
be analyzed and presented in separate scientific reports.
26.2-
22.&
15.t
11.2
•- 7. 5-,I-- 7*
3.7-
c? (RAW)2
Figure 15. ]Raw Thermosonde C Analysis - Profile Plot. Low andhigh gain measurements have bein combj~ned into a single best meas-urement. The thermosode measures Cfq from ground level to 30 krn.C is calculated from CN and the radiosonde measurements of pres-sure and temperature
37
"l aI" " " . -
F ----.--.--. O--
01(20
14 ,
12.
r't. i i [•
k--
10.
1071
c (RAW)N2
Figure 16. Expanded View of Raw Thertnosoride CN Profile plot. TheI altituide r-esolution of (~is about 10 m. Bach data point is, a result ofthe rms value of CT taken over 1 sec
38
11.*1-j 9.%
C],Cps
12.7.
• 11.4-
S10.2
I,--
-J 6.
7.6
CN2 (SMOOTHED)
Figure 17. Smoothed Thermosonde CN Profile Plot.CK is smoothed over 120 rn
39
I4<NSTRUMENTNOISE LEVEL
-44
-3-
cli
X
0
0.00-00-0-2 .5 2 1 203040 0 0 7 8 9095 98 9 9ý899
CUMULTIVEPROBAIL-T
8a C uaiePc'c'Jlt itiuinC.-eo ,G TenuTc
l~itur obbi .t Cu2 ltv Prbblt Ditiuio uw x O i.. Tero
probbiliy cT is 4. 5 X( 1Uo-0U
40)
15.0
5 MIN AVERAGE0205-0210
12.5
.- '--- .
€.,. * ! .c4 (DRY)
RADAR WITHPATH AVERAGING
5,0 .. "', .
STELLAR -'•42.5 02-08 ",
0020
-o_-0 -19.0 -io o170 16.0 -15.0 -14.0LOG C11
2 (rr2/ 3 )
Figure 19. Comparison of lBadar C and Seintillometer C2Profiles. The path averaged radar and scintillometer pro-fiJes show good agreement. T'he points depict radar(Cn measurements. The solid line through the radar pointsrerqwsent 5 -min radar averages. The seven scintillometerineasurements are shown as • with straight lines drawnb'tween thejn. 'Fhe dashed curve represents the radaraveragId C, corrected for moisture and passed through thescintiliometir weighting functions of Figure 19
41
FILTER PEAK ALTITUDE2 2 3.4 525 73 94 41•
1.0.
4x
.2;
0 2 4 6 to 12 14 1 8 2
ALTITUDE {km)
Figure 20. Scintillometer Weighting Functions. The optical scintillom
eter obtains data over these large weighting functions
42
20.00
17.50
OX20
15.00 I
12.50 - h THERMOSONDE
10.00W
;d 7.5f STELLAR
5.00
2.50
0.00 . • -... .- -
-20.00 -19.00 -18.00 -17.00 -16.00 -15.00
LOG C2 (m-2/3
Figure 2 1. Corn arison of Thermosond C2 andScintillometer C1, and ScirtillometernCa Profiles.The Thermosonde "raw" C.1 data was passedthrough the ocintillometer weighting functions ofFigurC 1n. Although the scint!1orneter vie-wed thesky vertically while the thermosonde followed aslant path, the agreement is good
"4I3
0238
/ • _ ozz620 0234
0230
/ •.-" MIN 0223
TH[ERiOSOWDE R - RADAR AVtRAGE 0
50015 COMPOSITE OF 029,
LOG AVERAGE r 5 PROFILES/ _0215
0210
0206
bE
0204
LAUNCH: OsbO r
- o 0020
II
31 JUL 01Th e•0o d -0 1e50 1
0-GZ RADAR
:* (DRY) 01,67
-15-20
LOG CN ( 2~
C2 2
Figure 22. Comparison of Tberrnosorlde 1 1, and Radar CN Profiles.
The herosode data is averaged over 0. 50 Ian. The R adar data is
D, time and space composite which attempts to sYnchronize the ie
space coordinates of the Radar and Thermosonde.
44
I
References
1. Tatarski, V.1. (1961) Wave Propagation in a Turbulent Medium, McGraw-HillCo., New York.
2. Titterton, P..T. , Mallery, L.E., and Arken, T.A. (1971) LightweightThermosonde .System, GTE Sylvania, Inc., Final Report, NAS5-11493.
3. GTE Sylvania, Inc. (1972) Lightweight Thermosonde System Model 140(Instruction Manual), Mountain View, California, B440.
4. Bufton, J. L. (1973) Correlation of microthermal turbulence data wi'th
meteorological soundings in the troposphere, J. Atmos. Sci. 30:83-87.
5. Bufton, J. L. (1975) A Radiosonde Thermal Sensor Technique For Measure-ment of Atmospheric Turbulence, NASA TN D-7867.
6. Curatola, C. (1975) Modifications to GTE Sylvania Thermesonde For ImprovedOperation, Final Report, AFWL-TR-75-180, ADA 018227.
7. Murphy, G. P. (1981) New Techniques and Devices For Measuring StratosphericTurbulence, AFGL-T---81-0128, ADA 102680.
8. Smithells, C.J. (1952) Tungsten, A Treatise on Its Metallurgy, PropertiesAnd Applicationc, Chapman & Hall, LozAdun, 3rd ed.
9. Marton, L. (1981) Fluid Dynamics, Academic Press, New York, Vol. 18,Part A, p. 271.
10. Dushman, S. (1962) Scientific Foundations of Vacuum Techniques, John Wiley,New York, p. 47, 2nd ed.
11. Jakob, M. (1949) Eeat Transfer, John Wiley, New York, Vol. 1, p. 291.
12. Watkins, B. J. (1981) Millstone Hill Doppler Radar Observations of Winds andTurbulence. Middle ,tmospheric Program, HandbooFfor MAP, Exte-Te--deabstracts from Int. Sym. on Middle Atm. Dyn. 2:30-36.
13. Ochs, G. R. , Lawrence, B.S., and Wang, T. (1976) Stellar scintillationmeasurement of the vertical profile of refractive - index turbulence in theatmosphere, Proc. Soc. of Photo-optical Inst. Eng. 75:48-54.
45
14. Endlich, R. M. , and Singleton, R. C. (1969) Spectral analysis of detailedvertical wind speeds profiles, J. Atmos. Sci. 26:1030-1041.
15. VanZandt, T.E., Gage, K.S., and Warnock, J.M. (1981) An Improved Modelfor the Calculation of Profiles of C2 and i in the Free Atmosphere from IBackground Profiles of Wind, Temperature, and Humidity, 20th Conf. onRadar Meteorology, (30 Nov. - 3 Dec. 1981).
416
III
if