AEROCHEM TP-343 NASA CR-145079

ENVIRONMENTAL EFFECTS
OF SPACE SHUTTLE SOLID ROCKET
MOTOR EXHAUST PLUMES
By

BAOCHUAN HWANG AND HAROLD S. PERGAMENT

> :* ••

i I

| 'R» t
Prepared under Contract No. NAS1-14271
By

Research Laboratories, Inc.

Princeton, New Jersey

for

f\JASA
National Aeronautics and
Space Administration

>, ->
•/.95^y
 BIBLIOGRAPHIC DATA 1. Report No. S.N^ecipient's Accession No.
SHEET
4. Title and Subtitle '5. Report Date
July 1976
ENVIRONMENTAL EFFECTS OF SPACE SHUTTLE SOLID ROCKET 6.
MOTOR EXHAUST PLUMES
7. Author(s) 8- Performing Organization Rept.
No
BaoChuan Hwang and Harold S. Pergament - AeroChem TP-343
9. Performing Organization Name and Address 10. Ptoject/Task/Work Unit No.
AeroChem Research Laboratories, Inc.
P.O. Box 12 11. Contract/Grant No.
Princeton, New Jersey 08540 NAS1-14271

12. Sponsoring Organization Name and Address 13. Type of Report & Period
Covered
National Aeronautics andJSpace Administration Final Report
T.T.-,„I-,•
Washington, „„<-,,,, r>'r>
DC Or>£/.£ '<"' . }', i "X •' i" •'•• s"*-- ,-•. •
20546 14.

15. Supplementary Notes

16. Abstractsj-he deposition of NOX and HC1 in the stratosphere from the Space Shuttle solid
rocket motors (SRM) and exhaust plume is discussed. A detailed comparison between stratc
spheric deposition rates using the baseline SRM propellant and an 'alternate' propellant
which replaces ammonium perchlorate by ammonium nitrate, shows the total NOX deposition
rate to be approximately the same for each propellant, ranging from ^0.018 kg/m at 15 .
km to fy 0.0036 kg/m at 30 km. For both propellants the ratio of the deposition rates of
NOX to total chlorine-containing species is negligibly small.
Rocket exhaust ground cloud transport processes in the troposphere are also ex-
amined. A brief critique of the NASA/MSFC Multilayer Diffusion Models (presently used
for predicting pollutant deposition in the troposphere) is presented, and some detailed
cloud rise calculations are compared with data for Titan III C launches. The results
show that, when launch time meteorological data are used as input, the model can reason-
ably predict measured cloud stabilization heights.
17. Key Words and Document Analysis. 17a. Descriptors

Rocket exhaust plumes

Rocket exhaust ground clouds
Solid propellants
Computer codes
Stratosphere
Troposphere

17b. Identifiers/Open-Ended Terms

Space Shuttle
Atmospheric Interaction Plume Program
NASA/MSFC Multilayer Diffusion Model

17c. COSATI Field/Group

18. Availability Statement 19..Security Class (This 21. No. of Pages

Report)
UNCLASSIFIED 51
20. Security Class (This 22. Price
Page
UNCLASSIFIED
FORM NTIS-35 ( R E V . 3-72) USCOMM-DC 149S2-P72
THIS FORM MAY BE REPRODUCED
itentionally
Left Blank
 TABLE OF CONTENTS

Page

1. SUMMARY 1

II . INTRODUCTION 2

III. NOX AND HC1 DEPOSITION IN THE STRATOSPHERE - BASELINE

AND ALTERNATE PROPELLANTS . V 3
. . -f-
A. Combustion Chamber Analysis of NOX Concentrations 3
B. Influence of Altitude on NOX Deposition Rates 4
C. Comparison between NOx and HC1. Deposition Rates 6

IV. COMPUTATION OF EXHAUST. PLUME PROPERTIES 7

A. Delivery and Use of the AIPP Code 7

B. Calculations Using the AIPP Code 7
C. Heterogeneous Chemistry, Aerosol Formation,
Particle Growth and Coagulation " 8
1. Heterogeneous Chemistry 8
2. Aerosol Formation, Particle Growth and Coagulation 9
a. Particle Growth 9
b. Particle Coagulation 11

V. SRM GROUND CLOUD RISE AND ATMOSPHERIC DIFFUSION MODELS 12

A. NASA/MSFC Multilayer Diffusion Model (MDM) 13

1. Cloud Rise 13
2. Atmospheric Diffusion Models 15
3. Program Output 15
a. Concentration and Dosage 16
b. Instantaneous Concentration Distribution 16
c. Computational Grid -- 16
4. Diffusion Coefficients 16
5. Code Operation 17

B. Calculations and Comparisons with Data 17

1. Meteorological Conditions 17
2. Stabilization Height, Cloud Path and Comparisons
with Data 18
3. Cloud Width and Cloud Shape Calculations and
Comparison with Data 18

VI. CONCLUDING REMARKS " 20

VII. REFERENCES 21

111
 LIST OF TABLES

Table Page

I PROPELLANT COMPOSITION 24

II EQUILIBRIUM CHAMBER PROPERTIES 24

III INITIAL CONDITIONS FOR COMBUSTION CHAMBER NONEQUILIBRIUM

CALCULATIONS 25

IV REDUCED NOX REACTION SET 25

V INITIAL CONDITIONS FOR AFTERBURNING CALCULATIONS 26

VI INFLUENCE OF SHOCKS AND AFTERBURNING ON PLUME ENHANCEMENT

OF NOX 26

VII NOX DEPOSITION IN THE STRATOSPHERE 27

VIII COMPARISON BETWEEN NOX AND HC1 DEPOSITION RATES 27

IX REACTION MECHANISM FOR AIPP CALCULATIONS 28

X. HETEROGENEOUS REACTION MECHANISM 29

LIST OF FIGURES

Figure

1 Approach to equilibrium of NOX in SRM combustion chamber 30

2 Influence of afterburning on plume temperatures and NOX

enhancement 31

3 SRM nozzle exit plane pressure and gas temperature

distributions 32

4 SRM nozzle exit plane gas velocity and flow angle distributions 33

5 SRM nozzle exit plane particle temperature and velocity

distributions • 34

6 " SRM exhaust plume structure 35

7 SRM exhaust plume centerline temperature and Mach number

distributions 36

8 SRM exhaust plume centerline pressure and NOX mole fraction

distributions 37

iv
Figure Page

9 Radial NOX profiles in SRM plume at 30 km 38

10 Configuration of stabilized cloud of exhaust products used

fof model 3 calculation (for a simulated normal launch of
Space Shuttle vehicle on 21 October 1972). 39

lla Schematic formation of plume from superposition of individual

instantaneous concentrations at subsequent times 40

lib Schematic spreading-disc plume model neglecting X-diffusion 40

12 Schematic formation of time integrated concentration 41

13 Vertical profiles of temperature, wind speed, wind direction

and relative humidity at Kennedy Space Center from rawinsonde
sounding measurement for 20 May 1975, 10:05 EDT 42

14 Comparison of the path of rising cloud obtained from optical

measurements and the NASA/MSFC cloud rise model for a
Titan III C vehicle on 20 May 1975 43

15 A comparison between calculated and measured crosswind cloud

widths 44

16 Stabilized cloud configuration obtained from NASA/MSFC cloud

rise models 3 and 4 45

17 Cloud configuration 17 min. after Titan III C launch,

20 May 1975 46

18 Effective cloud configurations calculated from NASA/MSFC

cloud rise model 3 using different inversion heights 47
 ENVIRONMENTAL EFFECTS OF SPACE SHUTTLE SOLID ROCKET MOTOR

EXHAUST PLUMES

BaoChuan Hwang and Harold S. Pergament

AeroChem Research Laboratories, Inc.

I. SUMMARY

This report discusses the effects of the Space Shuttle solid rocket
motor exhaust plumes on stratospheric NOX and HC1 deposition rates. The
stratospheric exhaust plume predictions were made using both the AeroChem
Atmospheric Interaction Plume Program (AIPP) and Low Altitude Plume Program
(LAPP) . The former code treats the coupled effects of nonequilibrium chem-
istry, gas/particle nonequilibrium and plume shocks (including the Mach
disc) ; the latter code specifically treats the influence of nonequilibrium
chemistry in the mixing/afterburning regions and is used to determine after-
burning effects at the lower stratospheric altitudes. Detailed comparisons
are made between stratospheric deposition rates using the baseline SRM pro-
pellant and an 'alternate' propellant which replaces ammonium perchlorate
by ammonium nitrate in order to reduce the total amount of HC1 in the ex-
haust. It was found that the total NOX deposition rate is approximately the
same for each propellant, ranging from = 0.018 kg/m at 15 km to =» 0.0036
kg/m at 30 km. For both propellants the ratio of NOX to total chlorine-
containing species deposition rates is negligibly small, on the order of
10~2 for the baseline propellant and ranging from 10"1 to 10~2 for the
alternate propellant. It is concluded that the alternate propellant, al-
though it contains a large amount of ammonium nitrate, will not increase
stratospheric NC^ deposition rates

The results of detailed calculations, using the AIPP code, are given
for the SRM plume at 30 km altitude and demonstrate the unique capabilities
of this new code.

Rocket exhaust ground cloud transport processes in the troposphere are

also examined. A brief critique of the NASA/MSFC Multilayer Diffusion Models
(presently used for predicting pollutant deposition in the troposphere) is
presented, and some detailed cloud rise calculations are compared with data
for Titan III C launches. The results show that, when launch time meteoro-
logical data are used as input , the model can reasonably predict measured
cloud stabilization heights.
 II. INTRODUCTION

The work reported herein is, in part, a continuation of our previous

studies1'2'3 on determining the total amount of NOX deposited in the strato-
sphere by the Space Shuttle solid rocket motors (SRM). Reference 1 reported
initial NOX deposition calculations, in which the effects of (1) gas/particle
nonequilibrium and nonequilibrium chemistry in the nozzle, (ii) plume shocks
(including the Mach disc) and (iii) nonequilibrium chemistry in the mixing/
afterburning regions and downstream of the Mach disc were taken into account.
Reference 2 extended this work to study the influence (on NOX deposition
rates) of: (1) uncertainties in A1203 particle sizes, size distribution and
particle/gas drag and heat transfer coefficients, (ii) intersection of the
two SRM plumes and (iii) mixing (and potential combustion) of rocket exhaust
products and ambient air in the shuttle base recirculation region. As a
result of thse studies the total mass flow of NOX 'leaving' the plume at
30 km altitude was determined to be 4 kg/s with a possible error factor of
x* 3. For a vehicle velocity of 1140 m/s this yields a nominal NO deposi-
X
tion rate of about 3.5 x 10~3 kg/m.

In this report we continue the above studies to determine the influence

of altitude on NOx and HC1 deposition rates for both the 'baseline' and so-
called 'alternate' propellants, the latter using mostly ammonium nitrate as
oxidizer in place of ammonium perchlorate in order to reduce the total amount
of HC1 in the exhaust. Next we discuss the use of the AeroChem Atmospheric
Interaction Plume Program (the AIPP code, which has been delivered to NASA/
Langley) to make Space Shuttle SRM plume calculations, present detailed
results for the 30 km case and show how to incorporate heterogeneous chemis-
try, aerosol formation and particle growth into the AIPP code. Finally, we
turn our attention to an important problem involving the environmental effects
of the SRM exhaust plume in the planetary boundary layer, i.e. the ground
cloud generated by the SRM launch, which may result in considerable ground
level pollutant deposition. A detailed examination is made of the model presently
used by NASA to determine the path of the cloud and ground level pollutant
(e.g., HCl) deposition rates.

The authors would like to acknowledge the assistance of Scott Wagner,

Rich Gomberg, Dick Bendura, and Gerry Gregory of the NASA/Langley Atmospheric
Environments Branch in explaining to us the intricacies of the Titan III
ground cloud data which they obtained. In addition, Briscoe Stephens of
NASA/MSFC and Arnold Gbldford of Science Applications, Inc., were helpful in
sending us the Uhivac version of the NASA/MSFC Multilayer Diffusion Model
(MDM Code). Diana Buckeye of Battelle and J.R. Bjorklund of H.E. Cramer Co.,
explained some of the details of the code that we needed in order to convert
it from Univac to operation on a CDC machine.
 III. NOX AND HC1 DEPOSITION IN THE STRATOSPHERE - BASELINE
AND ALTERNATE PROPELLANTS

This section discusses the influence of altitude on NOX and HC1* deposition
rates in the stratosphere for both the baseline and alternate A propellants. ** A
comparison between the propellant formulations (Table I) shows that, in the alter-
nate propellant, ammonium nitrate has replaced most of the ammonium perchlorate
which should result in substantially decreased HC1 concentrations at the nozzle
exit plane. Interestingly enough, although we might expect that the large amount
of ammonium nitrate might result in significantly larger amounts of NOX produced
in the chamber, equilibrium calculations (Table II) show that this is not the
case. In fact, the NO mole fraction is observed to be about a factor of 50 less
than that for the baseline propellant, primarily because of the substantially
decreased chamber temperature. However, the question arises of whether, because
of the ammonium nitrate in the propellant,''' equilibrium values of NOX will in
fact be achieved in the chamber. (It is well known5'6 that in laboratory flames
and furnaces, above-equilibrium levels of NOX are reached.) This question is
addressed below.

A. Combustion Chamber Analysis of NOX Concentrations

The model used for these calculations assumed that all exhaust products,
except the nitrogen-containing species, reached their equilibrium levels at the
computed adiabatic flame temperature. For conservatism, we assumed that 25% of
the N in the propellant was initially converted to NOf (note that at equilibrium
only 0.1% of the N in the propellant is converted to NO), the remaining N going
to N2. A constant pressure, constant temperature, nonequilibrium chemistry
calculation was then made in which the nitrogen-containing species concentrations
were determined by solving the species conservation equations using the initial
conditions given in Table III and the reaction mechanism in Table IV. Figure 1
shows the computed NO mole fraction (XJJQ) as a function of time (labeled FIRST
ASSUMPTION). It is seen that XNO approaches its equilibrium value in about
5 ms . In order to determine whether equilibrium will be achieved in the cham-
ber the above reaction time must be compared with a characteristic flow time.
For an average gas velocity in the chamber of 91.5 m/s§ the time to traverse

* Almost all the chlorine in the exhaust plume is tied up as HC1, although a
small percentage appears as Cl and C12. When the term HC1 is used in the text
it refers to the total chlorine-containing species.

t Of equal importance is that the atomic mass fraction of N in the alternate

propellant is about a factor of 2 larger than the corresponding value for the
standard propellant.

t Because of this assumption some 0 had to be removed from the remaining 0-

containing species. This was accomplished by arbitrarily reducing the mole
fraction of H20 by an appropriate amount.

§ This was computed from the SRM mass flow, the density of the combustion pro-
ducts and the approximate cross sectional area of the chamber.
the approximately 61 m length of the chamber is about 700 ms. Therefore we
expect that NOx equilibrium will be achieved under these conditions. In order
to get an approximate upper limit to X^Q in the chamber the rate coefficients
were arbitrarily decreased to their estimated lower bounds (as noted in Table IV),
These results, labeled in figure 1 as, MOST CONSERVATIVE, show that within about
20 ms (1.8 m) equilibrium is reached in the chamber.

In order to determine the reasonableness of the assumptions used in this

model, two additional calculations were made in which the temperature was
arbitrarily reduced to 2000 K and 1500 K. These results, also shown in figure
1, indicate that at temperatures typical of laboratory flames and furnaces, NO
is computed to freeze at above equilibrium values, in accord with what is known
about these lower temperature systems.s»6 These results demonstrate that NO
reaches equilibrium primarily because of the high rocket chamber temperature,
and we conclude that the computed equilibrium chamber value of XJJQ for the alter-
nate propellant is a reasonable starting point for the subsequent nozzle and ex-
haust plume calculations.

B. Influence of Altitude on NOY Deposition Rates

A calculation procedure similar to that described in reference 1 to deter-

mine the NOX deposition rate at 30 km was used to determine the corresponding
value at 15 km. Qualitatively, the principal differences in the overall plume
structure at these two altitudes are (i) at the lower altitude the effects of
afterburning are more severe and (ii) the plume is less underexpanded at 15 km,
changing the shock locations, the size of the Mach disc and therefore the mass
flow of exhaust gases passing through the Mach disc.
As discussed in reference 1 the amount of NOX 'leaving* the plume can be
greater than that leaving the nozzle (the so-called plume enhancement) due to
both afterburning and shocks. The total NOX deposition rate can be related to
the mass flow of NOX leaving the nozzle by defining an overall plume enhance-
ment factor (OPEF),

(m
NO }T
OPEF = (1)
(m }
NO ex
x

where m,Trt is the mass flow of NOx, subscript T signifies the total amount leav-
NOx
ing the plume and subscript ex signifies the amount leaving the nozzle. For the
purpose of determining the influence of altitude on NOX deposition it is reason-
able to assume that the NOX produced by shocks and afterburning are independent
quantities and can be separately calculated. Then the predicted NOX deposition
rate obtained by adding the NOx produced by shocks and afterburning individually
will give a conservative upper bound to the real NOx deposition. Defining total
mass flow of NOx separately for shocks and afterburning gives,

(2a)

(2b)
Then, defining individual OPEF's for afterburning and shocks gives

(OPEF) = .X . (3)
SH (m_ )ex
x

I.AFT
(OPEF)
AFT ' -tf±l <4)
j ex
x

From the above definitions it is evident that,

OPEF ^ (OPEF) + (OPEF) +1 (5)

oil AJ! L

Equation (3) can be rewritten as

(Y
NO
(OPEF)gH = 1 + (Y } I g^ I (6)
N0x ex \ ex

where (YJJQ ) and (YJJQ ) are the mass fractions of NOX downstream of the
Mach disc and at the exit plane respectively and m^ and mex are the mass
flows passing through the Mach disc and leaving the nozzle, respectively. The
use of equation 6 alone to establish the enhancement of NOX due to shocks im-
plies that the major increase is due to the first Mach disc. The other shocks
(exhaust gas and air shocks and subsequent Mach discs) are assumed to contri-
bute a negligible amount to the total plume NOX enhancement. This assumption
is quite adequate for the plumes of interest (see, e.g. ref. 2).

An examination of equation 6 clearly shows that (OPEF)~u depends on two

factors; (1) the increase in NOX mass fraction due to the high post-shock
temperature (approximately equal to the combustion chamber temperature) and
(2) the amount of propellant exhaust product mass flow passing through the Mach
disc. The first factor is primarily controlled by thermodynamics, since we
have already.shown1 that the NOX mass fraction rapidly achieves a state of near
local thermochemical equilibrium at the pressure and temperature downstream of
the Mach disc. The second factor is controlled by the gas dynamic plume struc-
ture, and is a function primarily of the nozzle exit to ambient pressure ratio
and the nozzle exit Mach number. To get a feeling for numbers, if there was
an order of magnitude increase in YNQ through the Mach disc and 10% of the
X
total mass flow passed through the Mach disc, (OPEF)gjj = 2.

The OPEF for afterburning can be written as,

}
AFT
(7)

5
 4
(OPEF)AFI - 1 +"£•
pu Y -
rdr
T (8)
NO ex ex
x

where p and u are the gas density and velocity respectively and r is the radial
distance from the axis. From equation 8 it is observed that the influence of
afterburning on plume NOx enhancement roust be determined via a radial integra-
tion of NOX profiles through the plume.

Initial conditions for afterburning calculations at 15 km made using the

AeroChem LAPP code7 are given in Table V. The radius of 8.23 m is an 'expanded'
radius for an equivalent single nozzle plume—expanded from the actual exit
condition, p =105 Pa, R =1.83 m to the ambient pressure of 1.21 x 10* Pa. . Figure 2
shows the large temperature increase due to afterburning at this altitude for
both the baseline and alternate propellants (compare these results with figure
11 of reference 1 which shows no temperature increase at 30 km for the baseline
propellant). The local plume enhancement factor due to afterburning (which is
simply equation 8 evaluated as a function of downstream distance) is shown to
be considerably larger for the alternate propellant than for the baseline pro-
pellant. This is primarily because (YJJQ )ex ^^ ^s about a factor of 50 less
x '
than (YNQ ) BASE* ^e actual NOX mass fractions produced in the afterburning
X
plumes of both propellants are about the same.

Table VI gives the results of the OPEF calculations showing the relative
effects of shocks and afterburning for both the baseline and alternate propel-
lants. As expected, because of increased reaction rates resulting in greater
heat release in the plume, the effects of afterburning are greater at 15 km
than at 30 km. The enhancement for the alternate propellant is greater than
that for the baseline primarily because the mass flow of NOX leaving the nozzle
is very much less than that of the baseline propellant (see exit plane composi-
tion in Table V).

The absolute amounts of NOX deposited in the stratosphere are shown in

Table VII. It is observed that these are about the same for each propellant,
the values being about a factor of from 3 to 5 higher at 15 km than at 30 km.

C. Comparison between N0y and HC1 Deposition Rates

Since both NOX and HC1 participate in the ozone destruction cycle (NOX
less efficiently than HC1), of primary interest is the ratio of NOX to HC1
deposition rates in the stratosphere. Since there are no major sources or
sinks for the total amount of chlorine contained in the exhaust products, the
mass flow of HC1 leaving the plume is approximately equal to that leaving the
nozzle.. Table VIII gives both the total NOX deposition rate (in kg/m for use
in statospheric diffusion models which use pollutant deposition as a source
term) and the ratio of NOX to HC1 deposition. These results show that the
amount of NOX deposited in the stratosphere is negligibly small compared to
the HC1 deposited,* the NOX/HC1 ratio ranging from values on the order of
10"1 to the order of 10~3.

* Note that this is the case even using the upper bound calculation described
above.
 IV.. COMPUTATION OF EXHAUST PLUME PROPERTIES

This section discusses the use of and results from a new computational
tool developed by AeroChem to compute rocket exhaust plume properties, taking
into account the coupled effects of nonequilibrium chemistry, turbulent mixing,
gas/particle dynamic and thermal nonequilibrium and plume shocks. This code,
the Atmospheric Interaction Plume Program (AIPP)8»9 was written with support
from the Air Force Rocket Propulsion Laboratory and has been applied here to
the determination of stratospheric SRM plume properties.

A. Delivery and Use of the AIPP Code

The AIPP code was delivered to the NASA/Langley Research Center and
demonstrated to operate on the CDC 6600 computer facility. Complete documenta-
tion of the code (both user oriented and program oriented) is given in AeroChem
TP-302a and TP-328, which have been updated and submitted as the Computer Pro-
gramming Documentation, together with a source language deck of the code and
test case input and output.

Use of the code requires that detailed nozzle exit plane gas and particle
properties be specified as initial conditions together with the vehicle flight
velocity and altitude, a suitable chemical reaction mechanism and rate coeffi-
cients. The proper nozzle exit plane conditions can be obtained via the
AeroChem fully-coupled nozzle code (FULLNOZ)10 which outputs exit plane gas and
particle properties in a form directly usable as input to the AIPP code.

B. Calculations Using the AIPP Code

The AIPP code has been used to calculate complete SRM exhaust plume
properties at 30 km for a flight Mach number of 3.8. Calculations were made
for an equivalent single nozzle (with a mass flow equal to the total mass flow
of the two SRM's). Distributions of gas and particle properties at the nozzle
exit plane (computed via FULLNOZ) are given in figures 3 to 5. The rather
large decrease in gas temperature toward the nozzle lip occurs at the limiting
particle streamline (i.e. there are no particles at radii greater than the
radius at which the limiting particle streamline occurs). Table IX gives the
chemical reaction mechanism and rate coefficients used for the plume calculations.
The shock structure in the plume is given in figure 7, together with the computed
mixing/afterburning region and the dividing streamline (DSL) in the subsonic/
supersonic (SS) region which separates the exhaust gas passing through the Mach
disc from the exhaust gas which has passed through both the exhaust gas shock and
reflected shock. The DSL in the SS region is seen to resemble a convergent
nozzle until the point where the initially subsonic flow turns supersonic. Fig-
ures 7 and 8 give the centerline distributions of temperature, Mach number, pres-
sure and NOX mole fraction. As expected, the pressure drops well below the
ambient pressure just upstream of the Mach disc and then increases to about a
factor of 2 greater than ambient after passing through the normal shock. The
subsequent pressure distribution is observed to rapidly approach the ambient
level.*

* In reference 9 it is pointed out that the AIPP code does not handle multiple
normal shocks downstream of the Mach disc.
 The temperature immediately downstream of the Mach disc increases to
approximately the combustion chamber value and then, because the flow is
far out of nonequilibrium, rapidly dissociates with a large temperature
decrease. The centerline mole fraction of NOX (fig. 8) increases to a peak
value that, because of the rapid dissociation, is less than the equilibrium
value at the post-shock pressure and temperature.

The radial plots of X^oX (fig- 9) are indicative of the total NO,, mass
^oo .
flow, i.e., mwn
wwx « jo
\ pu XNn
wux rdr, at any axial location. This mass flow is
then divided by the exit plane NOX mass flow to get the Local Plume Enhancement
Factor (LPEE). Finally, the asymptotic value of the LPEF is the OPEF, which
gives the total NOX deposition rate. A comparison between the OPEF computed
via the AIPP code and that computed from the 'equivalent AIPP' calculations of
reference 1 shows that the results are similar.

C. Heterogeneous Chemistry» Aerosol Formation, Particle Growth and Coagulation

1. Heterogeneous Chemistry

In this work we have concentrated on developing a formal technique

for incorporating heterogeneous chemistry within the AIPP code, particularly
radical recombination reactions taking place on the surface of A1203 particles.
These reactions, by increasing the rate of recombination of e.g., H, OH, Cl,
could reduce these species concentrations to a level low enough to 'suppress'
afterburning.

Our heterogeneous chemistry model is a phenomenological one which

relates the recombination rate at the particle surface to the arrival rate of
molecules and the surface sticking probability. Thus we define a 'heteroge-
neous rate coefficient' for the following general type of one-way reaction

RI +• R2 + S -»- P + S

where RI and R2 are radicals, P is the recombination product and S represents

the particle surface. The rate equation then becomes/

d[R2]
dt riET

where [Ri] and [R2] are the number densities of RI and R2 respectively and [S]
is the particle number density. If we assume that the particle diameters are
small compared to the gas mean free path, then the arrival rate of molecules
at the surface is simply the random flux of molecules across a surface, deter-
mined from kinetic theory. Using this assumption, the rate equation becomes,

where rp. is the radius of the ith particle group, mR is the mass of Rx, k
is the Boltzmann constant and T is gas temperature.
Utilizing equations 9 and 10 and defining a sticking probability, ns» to
account for the fact that not all molecules will stick to the particle surface
gives the following expression for

'a /_8kT_\ 1/2/_8kT_V/2

ns
^IET Pi \ k TT mRi ) \ TT mRz y (11)

In a practical sense, all radicals in the flow can recombine at particle sur-
faces and a heterogeneous reaction mechanism (Table X) must be added to the gas
phase reactions given in Table IX.
The formal method of incorporating a heterogeneous reaction scheme
into AIPP is given below:

1. A subroutine must be added to the code9(to be called from CHEM) to

evaluate the heterogeneous rate coefficient, kjjEx, for each reaction
and each particle group. This routine must take input data on rig f°r
each reaction, as well as particle number densities and species concen-
trations (which are available in COMMON).
2. The reactions are input as one-way, two-body reactions of type 9 (see
ref. 9) since [S] and kjjE^ will be lumped together as one term which
will multiply the particle concentrations.

A similar procedure has been used by AeroChem1x to analyze ion/elec-

tron recombination on liquid droplets with application to the injection of water
and electrophilic liquids into re-entry plasma sheaths to alleviate communica-
tions blackout.

2. Aerosol Formation, Particle Growth and Coagulation

In addition to the influence of solid particles in enhancing radical

recombination, they can also act as nucleation sites for the condensable species
in the exhausts. Condensation can also, of course, be initiated via homogeneous
nucleation in the plume, i.e., the kinetic formation of 'critical size' nuclei.
However, considering the large particle number densities and ion densities
(which can also induce nucleation), it is likely that heterogeneous nucleation
will prevail. Heterogeneous nucleation rates can be calculated in a manner
similar to that described in Section IV.C.I above, assuming that each collision
at a particle surface results in the condensible species being assimilated into
the particle, i.e., the sticking probability is unity. The rate of disappearance
of the gas phase can then be written as,
1/2

dt P\ m L
where mL is the mass of the condensable species.
a. Particle Growth - When the condensate particles are small, e.g.,
formed by heterogeneous nucleation via ions, the condensation rate and therefore
the rate of particle growth is controlled via chemical kinetic processes. When
the particles are sufficiently large that the Knudson number (ratio of mean free
path to particle diameter) in the flow falls below unity, the particle growth
(condensation) process is controlled by diffusion of the condensing species
through the boundary layer surrounding the droplet. Basically, the problem
reduces to one of determining the rate of heat transfer from and mass transfer
to a spherical droplet within a surrounding gas stream — exactly the opposite of
the 'evaporating* droplet problem.

A chemical kinetic approach to the growth phenomena was suggested by

Jensen12 who studied the growth of carbon particles, formed by homogeneous
nucleation, in a liquid rocket motor containing the monopropellant , isopropryl
nitrate. Jensen suggested several possible 'growth1 species, defined as "mole-
cules which are assimilated into a particle, physically or chemically, upon
collision of the molecule with the particle surface." In formulating a 'rate
coefficient' to calculate the total rate of depletion of the growth species it
was assumed that each collision of a growth species at a particle surface re-
sults in assimilation of the growth species into the particle (similar to the
assumption leading to equation 12) .
The problem of calculating particle growth when the particles are
large enough to have a significant boundary layer and their growth is controlled
by diffusion of the condensing gas species is one of extreme mathematical diffi-
culty which has not been solved 'exactly' except for either very small or large
Reynolds numbers. There have been, however, some successful approximations
(see, e.g. ref. 13). The, more or less, classical approach is to relate a
set of dimensionless parameters containing the heat and mass transfer rates to
the flow properties. These parameters are Reynolds number (Re), Nusselt number
(Nu) , Sherwood number (Sh) , Prandtl number (Pr) , and Schmidt number (Sc) , de-
fined as:
pud
Re = -*f- (13)

' TE ' T

yc
Pr = - (16)

Sc
'

The following nomenclature is used in the above equations:

d is particle diameter, u is gas viscosity, qo is heat transfer to the

particle surface, k is gas thermal conductivity, T is gas temperature, h is the
standard heat transfer coefficient, m is the mass transfer rate, D is the diffu-
sion coefficient, C is species concentration, Cp is specific heat and subscripts
'e' and 'o' refer to the edge of the boundary layer and the particle surface,
respectively.

10
 From similarity it can be shown that,

Nu = fiCRe, Pr) (18)

Sh = f2(Re, Sc) (19)

Employing the above formulation reduces the basic problem of determining the depo-
sition of a condensing vapor on a particle to that of determining the functions
fi and f2. Considerable theoretical and experimental work has been done on this
fundamental problem covering a wide range of Reynolds numbers (see, e.g. ref. 14).
For 102 < Re < 10s, a laminar boundary probably exists over the front face of a
spherical particle; the flow behind the particle is the 'usual' type of wake flow.
At higher Re, transition to turbulent flow will occur (the precise value depend-
ing in part on the 'freestream' turbulence level). A reasonable semi-empirical
expression for the condensation rate can be obtained from13

Sh = 2(1 + B Re1/2 Sc1/3) (20)

which is analogous to the familiar heat transfer expression

Nu = 2(1 + B Re1/2 Pr1/3) (21)

where B has been empirically determined to be about 0.3.

The implementation of expressions for the Sherwood number (similar to

equation 20) to calculate the actual growth of droplets within a solid propellant
exhaust plume is relatively straightforward within the framework of a computer
model using a finite difference formulation of the governing equations (e.g. , the
AeroChem AIPP code9). Combining equations 15 and 17 gives, for a spherical
particle,

m = 27rd(Ce-C0)D[l + B Re1/2 Sc1/3] (22)

Although the particle diameter is increasing with time the quasi-steady approxi-
mation, which assumes the time-dependent rate of change of surface area, 8A/3t,
is much smaller than that computed from the condensation rate m, may safely be
employed. Using the local flow conditions and initial droplet diameter, d^, a
new droplet diameter df at the end of a given integration step, At, can be
computed from

| tdf3 -<*i3] Pp = mAt (23)

•where p is the particle density. In this step by step manner the increase of
droplet diameter can be determined throughout the plume. In the AIPP code9 this
analysis should be added to subroutine PART.

b. Particle Coagulation - It is expected that condensate particles

will coalesce or stick together if they come in contact. This coagulation pro-
cess is well known and it can be shown15 that the rate at which particles dis-
appear by coagulation can be expressed by,

11
 --»>•
3
««>
where n is the number of particles per cm . The classic approach to determining
the kinetics of coagulation is the theory of Smoluchowski16 who proposed that
particles collide due to Brownian motion and that the particles stick to each ~
other in a certain fraction of those collisions. For all particles of the same
size the Smoluchowski expression reduces to

(25)

where R is the universal gas constant, N is Avogadro's number and s is the ratio
of the sphere of influence to the radius of the particle (which, if greater than
unity, assumes that particles will coagulate even before touching). A routine to
determine the coagulation of particles can readily be added to subroutine PART
in the AIPP code.9

V. SRM GROUND CLOUD RISE AND ATMOSPHERIC DIFFUSION MODELS

This section reports on the initial results of a long term study to deter-
mine the accuracy of existing models to predict both the path of the SRM ground
cloud and the deposition of pollutants along the path. The study is aimed at
assessing the environmental impact of Space Shuttle operations with respect to
the influence on the troposphere and on ground level pollutant concentrations
(particularly HC1 and particulates) in the vicinity of the Kennedy Space Center.
The need to understand and predict these environmental effects led to the devel-
opment of the NASA/MSFC Multilayer Diffusion Model (MDM). This model is de-
scribed in great detail in reference 17 and some calculated results of ground
cloud rise, cloud path and HC1 downwind deposition for some Titan III launches have
been reported in e.g. reference 18. However, the assumptions incorporated into
the MDM (reviewed in Section V.A below), led to concern that there might be
wide error bands on the predicted results. Since it is anticipated that the
MDM is to be used during the Shuttle launch countdown in order to determine
whether a potential launch hazard exists, it is necessary to minimize these
error bands. Thus, our work is directed toward the determination of (1) those
meteorological conditions for which the MDM is likely to give results that
either underpredict the actual hazard or are too conservative and (2) a method
by which the present model can be modified to give more realistic predictions
of the ground cloud path and pollutant deposition rates.

Our overall approach toward accomplishing the above goals is to first,

make a detailed, critical analysis of the MDM code in order to understand its
strengths and weaknesses and its explicit and implicit assumptions; second,
to compare calculated results from the MDM code with more exact 'numerical'
models (e.g. the Lawrence Livermore ADPIC code19) and the available data (e.g.
ref . 20) on Titan III launches* under a wide variety of meteorological con-
ditions and third, as a result of the above steps determine the magnitude of

* Titan III is used for the environmental effects study because it uses the same
propellant as the SRM and is about 1/2 the thrust. The interpretation of these
data is therefore very significant with regard to the SRM ground cloud study.

12
the errors in predicting the concentration and dosage fields (for different
weather conditions) resulting from transport of the Shuttle ground cloud ex-
haust products as they move downwind in the earth's surface mixing layer. In
this section it is assumed that the reader has a general familiarity with both
rocket exhaust ground cloud dispersion problems and with the MDM code. Addi-
tional information on these topics can be found in references 21 and 22.

The work reported below is the first part of the overall program to validate
the MDM code in the manner described above; it will be continued over the next
year under Contract NAS1-14504.

A. NASA/MSFC Multilayer Diffusion Model (MDM)

The basic code includes two programs which describe two different physical
processes; the first describes the buoyant force dominated cloud rise while the
second describes the diffusion processes controlled by atmospheric turbulence.

1. Cloud Rise

The hot exhaust products formed during rocket launches rise due to
buoyancy. The exhaust cloud of pollutants mixes with ambient air because of
the velocity shear between the cloud gases and the air. This entrainment of
cool air has a critical effect on cloud rise since it reduces the buoyant force.
In stable and neutral atmospheres, the cloud will rise to a stabilization height
where it is in thermal equilibrium with the ambient environment, following which
atmospheric turbulence will dominate the subsequent transport process, i.e., fur-
ther motions of the cloud will be controlled by both the wind velocity and turbu-
lent diffusion.

The NASA/MSFC models for predicting cloud rise are based on the work
of Briggs23'2* who developed formulae for stack plumes. The NASA/MSFC MDM in-
corporates two types of cloud rise models: (1) instantaneous cloud rise models
which assume spherical entrainment (designed to apply to normal launches of
solid-fueled rocket vehicles, e.g., Space Shuttle, Titan III, Minutemen II, etc.)
and (2) continuous cloud rise models which assume cylindrical entrainment (used
for. the cloud rise from vehicles with liquid-fueled first stages). For a vehicle
using both liquid and solid fuel engines, e.g., the Delta-Thor, the authors of
the code17 recommend an average of the rise predicted using the instantaneous
and continuous cloud rise models.

Since the net effect on buoyant cloud rise of atmospheric entrainment

due to turbulence is not well known and the large velocity fluctuations due to
convective eddies in unstable air cause severe measurement problems, no adequate
formula for predicting cloud rise in unstable conditions presently exists. Thus,
the cloud rise program utilizes the same formula for both unstable and neutral
atmospheres.
Derivations of all models in the NASA/MSFC cloud rise program can be
found in the report by Dumbauld et al.22 The original Briggs model is based
on (1) the conservation equations for buoyancy and momentum, (2) the assumption

13
 that the plume grows linearly* (the entrainment assumption), (3) the equivalency
between the mean horizontal speed of the plume in the bent over stage and the
ambient wind speed, and (4) a constant stability parameter S. The entrainment
constant for the linear growth of the radius of the rising cloud is still "a
controversial number.25 The value of 0.64 used in the NASA/MSFC models is based
on a best fit for a bulk of empirical measurements on stack plumes. Certainly,
a check of this value against the rocket exhaust cloud is needed since the
entrainment constant may not be the same for rising plumes which have been gen-
erated by different means.26

In addition to incorporating the Briggs model for the path of the mass
center of the rising cloud, the NASA/MSFC cloud rise models also provide calcula-
tions of the dimensions and spatial positions of the ground cloud at stabilization
and the distribution of exhaust products which are required input in the NASA/MSFC
diffusion models. At present only the input information from MDM models 3 and 4
(see Section V.A.2 below) are given in the cloud rise models. It is assumed that
the pollutant concentration distribution of the stabilized cloud is Gaussian. The
center of the Gaussian distribution is at the cloud stabilization height, calculated
using Briggs' formula. Assuming that the pollutant concentration at a distance of
one radius, r, from the cloud stabilization height, 2^, is 10% of the concentration
at Zjjj, the standard deviation of the Gaussian distribution can be given as r/2.15.
From the entrainment assumption in Briggs1 formula r = Y-Z = 0.64Z, where y is the
entrainment constant defined above. Since the portion of the cloud above the
mixing layer height (also referred to as the inversion height) is dominated by a
convective mechanism, only the portion of the cloud beneath the inversion height
is considered in the NASA models, i.e., the portion of the Gaussian distributed
cloud below the inversion height is used for the diffusion calculation. For
model 3, it is assumed that the portion of the cloud below the inversion height
is completely contained within an elliptical volume, the center of which is lo-
cated at a so-called effective height. Determination of this effective height is
illustrated in figure 10, in which H is the effective height, the shaded area is
the effective ground cloud geometry for model 3 and 2^ is the cloud stabilization
height. In model 4, the stabilized ground cloud in the mixing layer is assumed to
be contained in a number of cylindrical volume sources extending from the ground
to the top -of the inversion height. The distribution of exhaust product concen-
trations in each sublayer for model 4 is assumed to be Gaussian in the alongwind
and lateral directions. The total amount of product in each sublayer is given
as the integration of the corresponding portion of the vertically ascending
distribution cloud. Thus the predicted downwind concentrations will be very
strongly dependent upon the value chosen for the inversion height.

Finally, it should be pointed out that the entrainment constant, y» may

strongly influence the predicted downwind concentrations (in addition to the
selected value of the inversion height) because it influences the cloud radius
which in turn affects the initial pollutant concentration distribution. Therefore
the sensitivity of predicted ground level pollutant deposition to variations in
entrainment constant must be studied, in addition to analyzing the available
cloud rise optical data.

Measurements of the mean velocity distribution in a cross section of a rising

cloud from stacks show the cloud to be a double vortex. The greatest vertical
velocity and buoyancy occur near the center of the cloud. As the gases encoun-
ter ambient air above the cloud, vigorous mixing occurs across the entire top
of the cloud and causes the cloud diameter to grow approximately linearly with
height as it rises.

14
 2. Atmospheric Diffusion Models

The MDM is derived basically from Gaussian plume theory coupled with
the Cramer diffusion coefficients. The program was originally developed for
hazard estimates for the U.S. Army27 and then adopted by Dumbauld et al22 for
NASA operation. Fundamentally, there are four models included in the program:
the cylindrical distribution model 1, the static-plume model 2, the ellipsoidal
distribution model 3 and the multi-layered distribution model 4. In addition,
three options can be utilized within each model: precipitation scavenging, gra-
vitational settling and surface absorption.

Briefly, model 1 assumes that the vertical distribution of pollutants

is uniform and that the distributions in the alongwind and crosswind directions
are Gaussian. Model 2 utilizes the same expressions as model 1 except that the
dispersion process is static, i.e., the diffusion coefficients in the Gaussian
distribution form are assumed constant. Model 3 is the typical three-dimensional
ellipsoidal Gaussian distribution with mirror reflections on both the top and
bottom boundaries. In model 4 the concentration distribution for a layer, L,
is expressed as an integration of the Gaussian distributions of sublayers in-
cluded in L. A damping factor is coupled with the reflection terms in models 3
and 4 to account for boundary absorption. The pollutant concentrations in
different independent atmospheric physical layers can be calculated simultaneous-
ly by choosing either models 1, 2, 3, or 4 for each layer.

All models in the MDM code assume no flux of material across the
physical layer boundaries. Any flux of material (which would occur if the damp-
ing factor for boundary absorption is not set equal to 1) is assumed to be
instantaneously depleted. Thus the material across a layer boundary will not
be accounted for in the diffusion process in the next layer when multilayer
distribution is calculated simultaneously.

The Gaussian distribution assumption is only adequate for modeling

the diffusion process in a quasi-homogeneous and quasi-stationary flow field
with a uniform mean wind profile.28'29 Unfortunately, the real flow field in
the atmospheric boundary layer does not usually satisfy this condition. There-
fore it is natural to ask whether layering the non-homogeneous or shear layer
into numbers of sublayers which are nearly homogeneous and with uniform mean
wind will give more accurate predictions. The answer is probably no. Although
model 4 does account for the mixing of material from sublayers, its formulation
is such that the material in a sublayer will spread by following its original
dispersion rate in that sublayer. Thus the mixing model is not in accord with
the actual physics and using model 4 with a layered atmospheric mixing layer in
shear or non-homogeneous conditions will probably not give more accurate results.

3. Program Output

It should be pointed out that reference 17 gives little practical

guidance on how to properly use the models for a real problem. All decisions
on the operation of the code depend upon the user's familiarity with it and his
knowledge of atmospheric science relating to the problem under consideration.

15
 a. Concentration and Dosage - The total time integrated concentration
at a receptor position (defined as dosage) is not directly integrated from the
instantaneous concentration distribution. This is probably because it is not
analytically integrable nor is it easily integrated by numerical techniques.
Therefore it is important to closely examine how the dosage calculation is for-
mulated within the code in order to clearly identify the implicit assumption
associated with its use. In fact, the formulation can be obtained by either
of two possible explanations. The first utilizes the classical assumption that
diffusion in the alongwind direction is negligible in comparison to the gross
transport by the mean wind.30 In general, however, this assumption is not true
for modeling the diffusion of self-contained rocket clouds.31 Figure 11 demon-
strates the differences in the dosage formulation between using and not using
the above assumption. The lower cigar-like figure (lib) (in three dimensions
it looks like a pancake) shows the model used in the MDM for the dosage cal-
culations, which are obtained by neglecting diffusion in the X direction. The
second method of interpretation which will give the identical dosage formulation
is to assume that the effective instantaneous cloud during the time the cloud
passes over an observer has a constant distribution which is identical to the
distribution in the cloud at TQ. TQ is the time at which the cloud centroid
is at the same downwind position, x, as the receptor. The latter method is
also used for the formulations in MDM of the time mean alongwind concentration
which is the average partial dosage from time To-TA/2 to time TQ + TA/2 over
the average time TA. Figure 12 illustrates the second assumption used for the
dosage formulation. This assumption, however, is not correct because the cloud
will diffuse and cannot remain uniform in turbulent flow fields. The assumption
will result in a high predicted value in the near field of the launch pad. It
is apparent that the time mean alongwind concentration multiplied by the time
T. is equivalent to the dosage when T is larger than the cloud passage time.
A A

b. Instantaneous Concentration Distribution - The program does not

produce as output an instantaneous cloud concentration distribution. Therefore
any direct comparison between results calculated using the model and measured
data must be made very carefully. For example, the HC1 measurements for a
Titan III C launch on May 20, 1975 show that concentration data at ground level
were recorded at receptor locations P-9 and P-4,20 well before the cloud centroid
actually reached these locations. Therefore it is impossible to make a direct
comparison between these data and the analytical results.

c. Computational Grid - The spatial resolution for calculations with

the MDM code is designed for a maximum of 27,000 grid points on cylindrical co-
ordinates, which allows the grid size to be about AX % 1000 m, Ay £ 1500 m, and
AZ£ 2.5 m.

4. Diffusion Coefficients

The MDM code is based in large part on the Cramer diffusion coeffi-
cients,32-3A which are used to model the physics of atmospheric turbulent trans-
port. Cramer assumed a power law for the diffusion coefficients which, mathemat-
ically, are the standard deviation of the pollutant particle dispersion
displacement,

16
 oy = aQ x P (26)

tfz = 0-4, x 1 (27)

where ay and az are diffusion coefficients in the crosswind and vertical direc-
tions, respectively and OQ and a^ are the standard deviations of the azimuth
and elevated angles, respectively.

Two sets of experimental data, from the Prairie Grass35 and Round
36
Hill diffusion studies, were used to support Cramer's diffusion coefficients.
However it should be mentioned that both the Prairie Grass and Round Hill diffu-
sion experiments were carried out in planetary surface layers, which are usually
on the order of only tens of meters. Thus, it is questionable whether the Cramer
coefficients are suitable for investigating rocket ground clouds in the atmo-
spheric mixing layer with heights on the order of hundreds to a thousand meters.
At present the standard deviations ae and a^ of azimuth angle.and elevated angle
used in the MDM are calculated from an ad hoc formula22 based on the measured
a. only at a reference height Z near the ground. These values need to be
verified both theoretically and experimentally.
5. Code Operation

The total program contains about 6000 Fortran statements and 44 sub-
routines, most of which are for output purposes. Our experience has shown that
modifications to the code will be quite difficult and could cause operational
problems.

Both the cloud rise (preprocessor) program and the main MDM program
were written in Fortran IV, originally for a Univac computer. Since the pres-
ent operating computer system at NASA/Langley is a CDC 6600 and since there are
some differences in Fortran IV software among the various computer systems, a
conversion of the code to the CDC machine from the Univac system was needed.
This has been accomplished on version 5 of the code and several test cases have
been run successfuly. using different models. The CDC version of the NASA/MSFC
MDM program can be obtained from either NASA/Langley or AeroChem upon request.

B. Calculations and Comparisons with Data

A set of calculations for the cloud stabilization height, crosswind cloud

width, cloud shape at stabilization and its path during cloud rise for a
Titan III C launch on 20 May 1975 was made using the NASA/MSFC cloud rise models
and the MDM preprocessor computer code. The results of these calculations have
been compared with the available measurements carried out by the NASA/Langley
Atmospheric Environments Branch.

1. Meteorological Conditions

The meteorological parametric data used in the cloud rise calculations

were obtained from a rawinsonde sounding released from the Cape Canaveral weather
station at launch time, 10:05 EOT. The weather was fair on the day of
the launch with high pressure. The atmospheric vertical profiles of wind

17
velocity, temperature, wind direction and relative humidity are shown in fig-
ure 13. From these measured meteorological data collected near the launch
complex at the time of launch, the inversion height appears to be at about
2203 m from the ground. In fact, the consistent sharp changes of humidity,
temperature and wind velocity near 2203 m demonstrate the existence of the inver-
sion layer. It is evident that at launch time the atmosphere was in a slightly
stably stratified condition. Using Tabata's formula,37 in which the effect of
relative humidity is considered empirically, the potential temperature gradient
turns out to be about 0.011 degrees K per meter.

2. Stabilization Height, Cloud Path and Comparisons with Data

Using the meteorological data mentioned above and the NASA/MSFC cloud
rise models, the cloud was calculated to be stabilized at a height of 1527 m
and 1664 m away from the launch pad. The total time to reach the stabilization
height was about 8 min. 45 sec. The temporal history is estimated from the cloud
rise formula for stable conditions in the MSFC instantaneous cloud rise models.
For a Titan III C normal launch vehicle, the expression is,

arc cos (1 - 9.94883 x 10~12-Az<(.-ps«Z3'316) > (28)

where t(sec) is the time for the exhaust cloud to rise to an altitude Z (meter).
Pg (g/m3) is the air density at the surface and Az<|> is the potential temperature
gradient.
The stability parameter S is,

• S = •*- AZ<J) (29)

where g is the gravitational constant and Tg is the surface ambient temperature.

The temporal history of the ascent of the cloud mass center and the stabiliza-
tion height, as well as the measured data by NASA/Langley using the Askania
tracking camera, are shown in figure 14. The stabilization height and the path
obtained from the NASA/Langley measurements are based on the geometric center
of the cloud; the MDM code predicts the motion of the center of mass. If we
assume however that the center of mass is located near the center of geometry
(this occurs when the real cloud geometry is close to symmetric) direct comparisons
between the calculations and data can be made because the concentration distribution
in the modeled cloud is assumed to be Gaussian.

3. Cloud Width and Cloud Shape Calculations and Comparison with Data

The growing history of the cloud crosswind width during cloud rise is
calculated by using the following formula, given by the NASA/MSFC cloud rise
models,

r = yZ for Z < Zmm (30)

where r is the cloud radius at any height Z less than the cloud stabilization
height Zm and y is the entrainment constant (0.64).

18
 The cloud crosswind width history is shown in figure 15 in which the
shaded area is the envelope of cloud widths from the NASA/Langley measurements
for four launches.* The predicted cloud crosswind width'in the initial stage
apparently grows much faster than the average measured values but goes into the
envelope near the cloud stabilization height.

Figure 16 shows the cloud shapes at the stabilization height. The

stair-like irregular shaded area is the exhaust cloud calculated by using model
4 in the MDM preprocessor and the ellipsoidal area is from model 3. In both
calculations the inversion height is set at 2203 m.

Figure 17 is scaled from real time IR measurements in a very rough

manner and is given here to give a feeling for the real cloud shape. A compar-
ison with figure 16 shows that the cloud shape calculated with model 4 gives a
reasonable approximation to the measured cloud shape.

Finally, because the inversion layer cannot be exactly determined, it

is important to understand the influence of using different selected values for
the inversion height on the stabilized cloud shapes. When using model 4, the
only change in the shape is in neglecting that portion of the cloud above the
selected inversion height.. When using model 3 however, the cloud shape itself
will vary. As shown in figure 18 for an inversion layer at 1829 m, the
cloud is compressed into a much narrower ellipse than for an inversion layer
at 2203 m.

The four launches are February 11, 1974, May 30, 1974, December 10, 1974
and March 14, 1976.

19
 VI. CONCLUDING REMARKS

Based on our study of the deposition of NO and chlorine^containing

species in the stratosphere it is concluded that:

1. Using an 'alternate' propellant in the SRM, which replaces

ammonium perchlorate by ammonium nitrate in order to reduce HC1
in the exhaust products, will yield approximately the same NO
deposition rate as the baseline propellant.

2. For both baseline and 'alternate' propellants, the ratio of

N0x" to chlorine-containing
NO chlorine-contain species deposited at 15 km and 30
km is negligibly small.

3. The Atmospheric Interaction Plume Program (AIPP) is a useful

tool for computing detailed SRM plume properties.

In our study of the NASA/MSFC Multilayer Diffusion Model (MDM) for

predicting SRM rocket ground cloud rise and downwind pollutant deposition it has
been found that:

1. Stabilization heights calculated using the Briggs formula within

the MDM code are in reasonable agreement with cloud measurements.

2. The value selected for the inversion height and the entrainment
constant y may strongly influence the downwind deposition prediction.
A detailed investigation of these two parameters is needed.

3. The Cramer diffusion coefficients need to be validated for the

diffusion study of the rocket generated exhaust cloud in the
mixing layer.

4. The dosage and concentrations are inaccurately formulated.

5. The code requires a more definitive guideline on how to set up

input data for the cloud dispersion studies.

20
 VII. . REFERENCES

1. Pergament, H.S. and Thorpe,R.D., "NQ Deposited in the Stratosphere.by the

Space Shuttle," Final Summary.Report* Phase I, AeroChem TN-161, NASA
CR-132715, July 1975.

2. Pergament, H.S., Thorpe, R.D. and Hwang, B., "NO Deposited in the Stratosphere
by the Space Shuttle Solid Rocket Motors," Final Summary Report, Phase II,
AeroChem TP-333, NASA CR-144928, December 1975.

3. Thorpe, R.D., Pergament, H.S. and Hwang, B., "NOX Deposition in the Stratosphere
by the Space Shuttle Solid Rocket Motors," JANNAF 9th Plume Technology Meeting,
C.P1A Publ. 277 (Applied Physics Lab., Johns Hopkins Univ., Silver Spring, April
1976) pp. 317-352.
4. Space Shuttle Environmental Workshop on Stratospheric Effects, JSC-11137,
24-25 March 1976

5. Tokagi, T., Ogasawara, M., Fujii, K., Dazo, M., "A Study on Nitric Oxide
Formation in Turbulent Flames," Fifteenth Symposium (International) on Combustion
(The Combustion Institute, Pittsburgh, 1975) pp. 1051-1059.

6. Altenkirch, R.A. and Mellor, A.M., "Emission and Performance of Continuous

Flow Combustors," Fifteenth Symposium (International) on Combustion (The
Combustion Institute, Pittsburgh, 1975) pp. 1181-1189.

7. Mikatarian, R.R., Kau, C.J. and Pergament, H.S., "A Fast Computer Program for
Nonequilibrium Rocket Plume Predictions," Final Report, AeroChem TP-282, AFRPL
TR-72-94, NTIS AD 751 984, August 1972.

8. Pergament, H.S. and Kelly, J.T., "A Fully-Coupled Underexpanded Rocket Plume
Program (The AIPP Code). Part I. Analytical and Numerical Techniques," Final
Report, AeroChem TP-302a, AFRPL-TR-74-59, NTIS AD A006 235, November 1974.

9. Pergament, H.S. and Kelly, J.T., "A Fully-Coupled Underexpanded Afterburning

Plume Program (The AIPP Code). Part II. Program User's Manual," Final Report,
AeroChem TP-328, AFRPL-TR-75-52, NTIS AS A019 411, October 1975.

10. Pergament, H.S. and Thorpe, R.D.., "A Computer Code for Fully-Coupled Rocket
Nozzle Flows (FULLNOZ)," AeroChem TP-322, AFOSR-TR-75-1563, NTIS AD A019 538,
April 1975.

11. Pergament, H.S. and' Kau, C.J., "A Computer Code to Predict the Effects of
Electrophilic Liquid Injection on Re-Entry Plasma Sheath Properties," Final
Report, AeroChem TP-308,. AFCRL-TR-74-0074, NTIS AD 782 023, January 1974.

12. Jensen, D.E., "Chamber Non-Equilibrium and Smoke Predictions for a Liquid
Monopropellant Rocket Motor," JANNAF 7th Plume Technology Meeting, CPIA
Publ. No. 234, June 1973.

13. Fuchs, N.A., Evaporation and Droplet Growth in Gaseous Media (Pergamon Press,
London, 1959).

21
 14. Johnstone, H.F. and Eades, D.K., Ind. Eng, Chem. 42_,. 2293 (1950),

15. Zebel, G., ''Coagulation of Aerosols," Aerosol Science, C.N. Davies, Ed.
(Academic Press, London, 1966) Ch. II.

16. Smoluchowski, M. von, Phys. Z. _17_, 557, 585 (1916).

17. Dumbauld, R.K. and Bjorklund, J.R., "NASA/MSFC Multilayer Diffusion Models
and Computer Programs - Version 5," NASA CR-2631, 1975.

18. Stephens, J.B., "Atmospheric Diffusion Predictions for the Exhaust Effluents
from the Launch of a Titan 3C, December 13, 1973," NASA TM X-64925, 27
September 1974.

19. Lange, R., "ADPIC, A Three Dimensional Computer Code for the Study of Pollutant
Dispersal and Deposition under Complex Conditions," Lawrence Livermore Lab.,
UCRL-5KL62, 1973.

20. Gregory, G.L., Wornom, D.C., Bendura, R.J. and Wagner, H.S., "Hydrogen Chloride
Measurements from Titan III Launches at the Air Force Eastern Test Range, FL,
1973 thru 1975," NASA TM X-72832, 1976.

21. Slade, D.H., Ed., "Meteorology and Atomic Energy," USAEC-TID-24190, 1968.

22. Dumbauld, R.K. et al, ""NASA/MSFC Multilayer Diffusion Models and Computer
Program for Operational Prediction of Toxic Fuel Hazards," NASA CR-129006,
1973.

23. Briggs, G.A., "Plume Rise: A Critical Survey," Air Resources Atmospheric
Turbulence and Diffusion Lab., TID-25075, 1969.

24. Briggs, G.A., "Some Recent Analyses of Plume Rise Observations," presented
at the Second International Clean Air Congress, Washington, DC, December
6-11, 1970, Paper ME-8E.

25. Briggs, G.A., "Discussion; Chimney Plumes in Neutral and Stable Surroundings,"
Atmos. Env. £, 511-512 (1972).

26. Baker, P.J. and Jacobs, B.E.A., "Pulsed. Emission Chimney," Civil Engineering
Publics Work Rev., 199-200 (1971).

27. Cramer, H.E. and Dumbauld, R.K., "Experimental Designs for Dosage Prediction
in CB Field Tests," Geophys. Corp. of America, GCA 68-17-G, 1968.

28. Corrsin, S., "Limitations of Gradient Transport Models in Random Walks and in
Turbulence," Advances in Geophysics, Vol. 18A (Academic Press, New York, 1974)
pp. 25-60.

29. Hwang, B., "Concentration Distributions in Turbulent Shear Flow," presented at

1975 Amer. Phys. Soc. Fluid Dynamic^ Division Annual Meeting, Univ. of Maryland,
College Park, MD, 1975. (Paper submitted to J. Fluid Mech.)

22
30. Frenkiel, F.N., "Turbulent Diffusion: Mean Concentration Distribution in a
Flow Field of Homogeneous Turbulence," Adv. Appl Mech, _3, 61-107 (1953).

31. Stephens, J.B. and Stewart, R.B., ''"Rocket Exhaust Effluent Modeling for
Tropospheric Air Quality and Environmental Assessment," Preliminary Report,
NASA, 1975.

32. Cramer, H.E., "A Practical Method for Estimating the Dispersal of Contaminants,"
Proceedings of the First National Conference on Applied Meteorology (Amer.
Meteor. Soc., Hartford, 1957) pp. C-33 to C-35.

33. Cramer, H.E., "Engineering Estimates of Atmospheric-Dispersion Capacity,"

presented at Annual Meeting of the American Industrial Hygiene Association,
Chicago, IL, April 1959.

34. Cramer, H.E. et.al, "Meteorological Prediction Techniques and Data System,"
Geophysics Corp. of America, GCA-64-3-G, 1964.

35. Barad, M.L., "Project Prairie Grass. A Field Program in Diffusion," Geo-
physics Research Papers No. 59, Vols. I and II, AFCRC-TR-58-235, July 1958.

36. Cramer, H.E., Record, F.A. and Vaughan, H.C., Final Report, AFCRC-TR-58-
239, 1958, p. 70.

37. Tabata, S., "A Simple but Accurate Formula for the Saturation Vapor Pressure
Over Liquid Water," J. Appl. Meteor. 12, 1410 (1973).

23
 . TABLE I

PROPELLANT COMPOSITION

Wt% AH°(kcal/mole)
—r
Standard Alternate

NH<.C10A 69.6 10.0 -70.69

NH3HN03 44.0 -87.27
Al 16.0 15.0 0.0
C H ON 12 0 -12.0
TJ rt \T O f\
C 6«13*l6«97vi%i7lNo,O3 £• U -28.3
Fe203 0.4 -197.3
C<,H8N808 17.0 17.9
Ci 7 . 6H2s. eOo ..2 12.0 13.55
CaaH«04. 2.0 -311.6

TABLE II

EQUILIBRIUM CHAMBER PROPERTIES

Standard Alternate
Pressure, Pa (N/m2) 4.14 x 106 4 .14 x 106
Temperature , K 3400 2690
Mass Fraction A1203 Particles 0.30 0.283

Gas Composition, Mole Fraction

CO 2.49(-l) 2.7K-1)
C02 1.74 (-2) 8.63(-3)
HC1 1.43(-1) 1.92 (-2)
H 3.80(-2) 5, 47 (-3)
H2 2.77(-l) 4.19C-1)
OH 9.1K-3) 2. 84 (-4)
H20 1.52(-1) 8.80(-2)
N2 9. 92 (-2) 1.87C-1)
N 6. 19 (-6) < 1.0 (-9)
NO 7. 05 (-4) 1.60C-5)
0 7.45(-4) 2.45(-6)
02 1.62(-4) < l.OC-9)
Cl 1.30(-2) 1.73C-4)

24
 TABLE III

INITIAL CONDITIONS FOR COMBUSTION CHAMBER

NONEQUILIBRIUM CALCULATIONS

Pressure 4.14 x 10s Pa (N/m2)

Temperature 2690 K
Gas Composition, M.ole Fraction
NO 9.1(-2)
N2 1.3C-1)
N 1.0(-10)
CO 2.6(-l)
C02 1.7(-4)
HC1 1.8(-2)
OH 2.7(-4)
0 2.3(-6)
02 2.2(-11)
Cl 1.6(-4)
H 2.2(-2)
H2 4.7(-l)
H20 1.4(-4)

TABLE IV

REDUCED NO,,
A.
REACTION SET

Kf = AT~N exp(B/RT)

B Upper, Lower
N (cal/mole) Bound Bound
N + N +M = N2 + M 5.5 (-30) 1.0 0.0 10 10
0 + N2 = NO + N 1.3 (-10) 0.0 -76000 3 3
N -h 02 ^ NO + 0 2.2 (-11) 0.0 6250 3 3
NO + H i: N + OH 1.5 (-10) 0.0 -47000 5 5

cm-molecule-sec units
multiply Kf by upper bound to get upper limit of rate coefficient
divide K,. by lower bound to get lower limit of rate coefficient

25
 TABLE V

INITIAL CONDITIONS FOR AFTERBURNING CALCULATIONS

15 km, Mo, = 1.9

Standard Alternate
Pressure, Pa (N/m ) 2
12.21 x 103 1.22 x 10"
Temperature, K 1350 1070
Velocity, m/s 2710 2410
Radius, m 8.23 8.23
Composition, Mole Fraction

CO 2.28C-1) 2.39(-l)
CO2 2.47(-2) 1.18 (-2)
Cl 4. 08 (-3) 5. 25 (-5)
HC1 1.44(-1) 1.85(-2)
H 8. 75 (-3) 1.2K-3)
OH 7. 21 (-5) 2.20(-6)
H2 2.74C-1) 4.03(-1)
H20 1.46C-1) 8. 18 (-2)
N2 9.45(-2) 1.72(-1)
0 2. 09 (-6) 6. 63 (-9)
02 7.76(-7) 9. 62 (-10)
Al203(s) 7. 46 (-2) 7. 21 (-2)
N 9. 95 (-10) 9. 62 (-IP)
NO 1.99 (-4) 4. 36 (-6)

TABLE VI

INFLUENCE OF SHOCKS AND AFTERBURNING ON PLUME ENHANCEMENT OF N0?

Overall Plume Enhancement Factors (OPEF)

15 km 30 km
Baseline Alternate Baseline Alternate
Shocks 3.4 216 1.5 46

Afterburning 2.5 80 1.2 10

Total 4.9 295 1.7 55

26
 TABLE VII

NOX DEPOSITION IN THE STRATOSPHERE (kg/s)

Nozzle 15 km 30 km

Baseline 2.5 (300 ppm) 2.5 (300 ppm)

Alternate 0.055 0.055

Plume Enhancement

Baseline 9.5 1.5

Alternate 16.0 3.0

TOTAL

Baseline 12.0 4.0

Alternate 16.055 3.055

TABLE VIII

COMPARISON BETWEEN NOX AND HC1 DEPOSITION RATES

15 km; Velocity = 561 m/s

30 km; Velocity = 1140 m/s

NO Deposition Rate
X ...

(kg/m)
15 km 30km
Baseline 0.0207 0.0036
Alternate 0.0279 0.0027

Ratio of NOj£ /HC1 Deposition Rates

Baseline 6; 009 0.0045

Alternate 0.099 0.0243

27
 TABLE IX

REACTION MECHANISM FOR AIPP CALCULATIONS

K f =AT~" exp(B/RT)
B
A N (cal/mole)
0 + 0 + M -> : 02 + M 1.0 (-29) 1.0 0.0
0 + H + M ->•OH + M l.OC-29) 1.0 0.0
H + H + M' -»• H2 + M 5.0(-29) 1.0 0.0
H + OH + M -»• H20 + M 2.0(-28) 1.0 0.0
CO + 0 + M ->•C02 + M 1.0(-29) 1.0. -2500
OH + OH ->• H20 + 0 l.O(-ll) 0.0 -780
OH + H2 •* H20 + H 3. 6 (-11) 0.0 -5200
0 + H2 ->• OH + H 2.9(-ll) 0.0 -9460
H + 02 -»• OH + 0 3. 7 (-10) 0.0 -16800
CO + OH -»• C02 + H 9.0(-13) 0.0 -1080
H + Cl + M -»• HC1 + M 5.5(-31) 1.0 0.0
HC1 + OH ->• H20 + Cl 7.2(-12) 0.0 -3250
H + HC1 -»• Cl + H2 8. 8 (-11) 0.0 -4620
OH + Cl •*• HC1 + 0 3.0(-11) 0.0 -5000
0 + N2 -»• NO + N 1.3(-10) 0.0 -76000
N + 02 •*• 0 + NO 2.2(-ll) 0.0 -6250
NO + H -»• N + OH 1.5(-10) 0.0 -47000
N + N + M -»• N2 + M 5.5(-30) 1.0 0.0

28
 TABLE X

HETEROGENEOUS REACTION MECHANISM

(one way reactions only)

H + OH + S ->• H20 + S
0 + H+S -»• OH + S
0 + 0 + S -»• 02 + S
H + H + S -»• H2 + S
CO 4- 0 + S -»• C02 + S
H + Cl + S -»• HC1 +. S
Cl + Cl + S -»• C12 + S
N + N + S -J- N2 + S
N + O + S -»• NO + S

29
 76-102

TEMPERATURE t K
1500
2000

10-2

z
o
h-
o
a:
u_
LU I0~3
_i
o
FIRST ASSUMPTION

MOST CONSERVATIVE

io-4

EQUILIBRIUM

0 15 20 25 30 35
TIME, milliseconds

Figure 1. Approach to equilibrium of NOX in SRM combustion chamber.

30
 76-101

I n r
3000 120
TEMPERATURE
LPEF
U.
UJ
Q_
2600 100
ALTERNATE
o
h-
CJ
2200 80

Lul UJ
CK 2
UJ
o
< 1800 60
LU
Q_
X
2 UJ
UJ
h-
UJ
1400 40 ^
ZD
BASELINE _l
Q.

o
1000 20 o

600
20 40 60 80

DISTANCE FROM NOZZLE EXIT, x/R

Figure 2. Influence of afterburning on plume temperatures and NOX

enhancement.
Baseline and Alternate propellants, 15 km

31
 76-107

2200

TEMPERATURE
.8XI05
~\J > ^ * "'''
^~ 2100

I.6XI0 5
2000

CVJ

1900

WI.2XI05
PRESSURE LU
CO 1800
CO
Ld LU
o:
Q.
I0

1700

8XI0 4

1600

1500
0 0.5 1.0 1.5 2.0 2.5
RADIAL DISTANCE FROM AXIS,m

Figure 3. SRM nozzle exit plane pressure and gas temperature

distributions.

32
 76-109

0.25 -VELOCITY

0.20
o
o>

0.15 IO
O

o FLOW ANGLE

I o.io O
O
_l
LiJ

0.05

RADIAL DISTANCE FROM AXIS, m

U)
U) Figure 4. SRM nozzle exit plane gas velocity and flow angle
distributions.
 76-106

2.5 2500

2.4 PARTICLE RADIUS, p. 2400

u
CO
« 2.3 2300
UJ
10
tr
O
2.2 2200 eg
Q.

O UJ
O
_l
LU
2.1 2100
u
UJ I-
O cc
a: 2,0 2000
£
QL

1,9 PARTICLE VELOCITY 1900

PARTICLE TEMPERATURE

1.8 1 1800
0.5 1.0 1.5 2.0 2.5

RADIAL DISTANCE FROM A X I S . m

figure 5. SRM nozzle exit plane particle temperature and velocity

distributions.

34
 76-110

AIR SHOCK
120
MIXING/AFTERBURNING REGION
X
<
REFLECTED SHOCK
o SUBSONIC/SUPERSONIC
(T
UL DIVIDING STREAMLINE
80
UJ
O EXHAUST GAS SHOCK
h-
Q
40
O
cr

0
0 40 80 120 160 200
DISTANCE FROM NOZZLE EXIT PLANE, m

Figure 6. SRM exhaust plume structure. 30 km, vehicle Mach

number = 3.8. Mixing region outer boundary defined
as ambient N2 less 5% of the difference between ambient
U) and exhaust product N2; inner boundary is exhaust
Ul product N2 plus 5% of the difference.
 76-111

MACH NUMBER

(T
UJ UJ
o: TEMPERATURE CD
ID 4000

IT
UJ
Q_ 3 5
5
UJ

2000
MACH NUMBER

0
40 80 120 160 200
DISTANCE FROM NOZZLE EXIT PLANE,m

Figure 7. SRM exhaust plume centerline temperature and Mach

number distributions.
 76-108

1 I I I

PRESSURE
10 -2 2
O
I-
o
Q_
AMBIENT PRESSURE (30km)
UJ
UJ
CO
_l
C/5
o
UJ
a: X
Q_
O

10 10'3

•NOX MOLE FRACTION

I I I I 10-4
80 160 240
DISTANCE FROM NOZZLE EXIT PLANE,m

Figure 8 SRM exhaust plume centerline pressure and NOX mole fraction
distributions.

37
 76-114

10 -3

t
I

O 10-4
J-
o

u_
UJ
_l DISTANCE FROM
o NOZZLE EXIT, m
— — — — 140
x
O
40

10-6
10 20 30
RADIAL DISTANCE FROM AXIS, m

Figure 9. Radial NO profiles in SKM plume at 30 km.

X

38
 76-105

3200

2800 -

2400-

~2000 -

1600-

LU
X 1200-

800-

400-

1200 800 400 400 800 1200

WIDTH, m

Figure 10. Configuration of stabilized cloud of exhaust products used

for model 3 calculation (for a simulated normal launch of
Space Shuttle vehicle on 21 October 1972). Shaded area
represents effective cloud dimensions, Hm is inversion
height.

39
 76-94

INSTANTANEOUS
CONCENTRATION
ISOPLETH

Figure lla. Schematic formation of plume from superposition of individual

instantaneous concentrations at subsequent times.

dxnrdt

Figure lib. Schematic spreading-disc plume model neglecting X-diffusion.

40
 76-100

RECEPTOR

Uk

Figure 12. Schematic formation of time integrated concentration. Constant

concentration distributions assumed during the time cloud passes
the receptor.
N>

76-92

•V-i
2400

2000

1600

1200
UJ

800

400
-9.8°C/km

0
10 15 20 25 3.5 7.0 10.5 2 4 6 8 40 60 80 999
1
' TEMPERATURE, °C WIND DIRECTION, WIND SPEED, m sec' RELATIVE HUMIDITY, percent
rod.

Figure 13. Vertical profiles of temperature, wind speed, wind

direction and relative humidity at Kennedy Space
Center from rawinsonde sounding measurement for
20 May 1975, 10:05 EDT.
 76-97

i i i i i 1 I I I

CACULATED
LU STABILIZATION HEIGHT
Q

tj<
Q
O
_J —— — NASA/LANGLEY MEASUREMENTS
O
(ASKANIA TRACKING UNIT)

CALCULATED FROM NASA/MSFC

CLOUD RISE MODEL
I I I I i i L
0 8 12 16 20 24
TIME AFTER LAUNCH, minutes

Figure 14. Comparison of the path of rising cloud obtained

from optical measurements and the NASA/MSFC
cloud rise model for a Titan III C vehicle
on 20 May 1975.
OJ
 76-95

ENVELOPE
NASA/LANGLEY DATA

I
h-
O

Q
D
o
_J
o
Q

CO
CO
o
<r
o

FROM NASA/MSFC
CLOUD RISE MODEL

2 4 6 8 10
TIME AFTER LAUNCH, minutes

Figure 15. A comparison between calculated and measured crosswind

cloud widths.

44
 76-99

3500 l
INVERSION LAYER AT 2203 m

3000 NASA/MSFC CLOUD RISE MODEL 4

NASA/MSFC CLOUD RISE MODEL 3

2500

2000
;

1500

LJ
X
1000

500

0
-1000 0 1000 2000 3OOO 4000
DISTANCE FROM LAUNCHING PAD, m

Figure 16. Stabilized cloud configuration obtained from NASA/MSFC

cloud rise models 3 and 4.

Ol
O\ 76-96

3000

0 1000 2000 4000

DOWNWIND DISTANCE.m

Figure 17. Cloud configuration 17 rain, after Titan III C launch,

20 May 1975. (from NASA/Langley IR measurements)
 76-98

2500
INVERSION LAYER ASSUMED AT'
2203 m

2000

1500

1000
LJ
X

500

0 I I I I I
0 1000 2000 3000
DISTANCE FROM LAUNCHING PAD, m

Figure 18. Effective cloud configurations calculated from

NASA/MSFC cloud rise model 3 using different
inversion heights.