AICE-AIR-72-13
AICE# SURVEY OF USSR AIR POLLUTION LITERATURE
Volume -Xlii
TECHNICAL PAPERS FROM THE LENINGRAD INTERNATIONAL SYMPOSIUM
ON THE
METEOROLOGICAL ASPECTS 0T ATMOSPHERIC POLLUTION
PART II
Edited By
M. Y. Nuttonson
The material presented here is part of a survey of
USSR literature on air pollution
conducted by ths'Air Pollution Section
AMERICAN INSTITUTE OF CROP ECOLOGY
This survey is being conducted under GRANT 1 ROl AP00786 — APC
OFFICE OF AIR PROGRAMS
of the
U.S. ENVIRONMENTAL PROTECTION AGENCY
•AMERICAN INSTITUTE OF CROP ECOLOGY
809 DALE DRIVE
SILVER SPRING, MARYLAND 20910
1972
-------
TABLE OF CONTENTS
Page
PREFACE vii
AUTOMATION OF INFORMATION PROCESSING INVOLVED IN EXPERIMENTAL
STUDIES OF ATMOSPHERIC DIFFUSION
A. S. Zaytsev 1
MICROMETEOROLOGICAL CHARACTERISTICS OF ATMOSPHERIC POLLUTION
CONDITIONS
T. A. Ogneva 8
STUDY OF THE INFLUENCE OF IRREGULARITIES OF THE EARTH'S SURFACE
ON THE AIR FLOW CHARACTERISTICS IN A WIND TUNNEL
S. M. Gorlin, I. M. Zrazhevskiy, and S. P. Ziborova 18
USE OF PARAMETERS OF EULERIAN TURBULENCE FOR ESTIMATES OF
LAGRANGIAN CHARACTERISTICS
V. I. Ivanov 33
METHOD OF EVALUATING ATMOSPHERIC DIFFUSION FROM
TURBULENT CHARACTERISTICS
N. L. Byzova 49
SCATTERING OF SMOKE FROM A HIGH-LEVEL POINT SOURCE
Ye. K. Garger 64
DIFFUSION FROM A POINT SOURCE OF FINITE TIME OF ACTION
Yu. S. Osipov 77
USE OF SURFACE OBSERVATIONS FOR CHARACTERIZING THE STATE OF
THE SURFACE ATMOSPHERIC LAYER
G. B. Mashkova 85
SULFUR DIOXIDE AND DUST MEASUREMENTS IN MEASURING NETWORKS
OF THE HYDROMETEOROLOGICAL INSTITUTE
0. Muller 91
EXPERIMENTAL STUDIES OF ATMOSPHERIC POLLUTION
IN INDUSTRIAL AREAS
B. B. Goroshko and E. N. Zasukhin 98
FIELD STUDIES OF AIR POLLUTION IN THE AREA OF
THE SKAWINA ELECTRIC POWER PLANT
W. Parczewski 109
EFFECT OF METEOROLOGICAL CONDITIONS ON AIR POLLUTION
IN CITIES OF THE SOVIET UNION
E. Yu. Bezuglaya and L. R. Son'kin 112
lii
-------
CONTENTS OF PARTS I AND III
of the AICE Translations of the
TECHNICAL PAPERS FROM THE LENINGRAD INTERNATIONAL SYMPOSIUM
ON THE
METEOROLOGICAL ASPECTS OF ATMOSPHERIC POLLUTION
Part I
ABSTRACT
M. Ye. Berlyand
INTRODUCTION
M. Ye. Berlyand
STATUS AND PROSPECTIVE DEVELOPMENT OF METEOROLOGICAL STUDIES
OF ATMOSPHERIC POLLUTION
M. Ye. Berlyand
EFFECT OF THE STABILITY OF THE ATMOSPHERE ON THE DISSEMINATION
OF GASEOUS POLLUTANTS
V. Parchevsky
METHOD OF DETERMINATION OF AVERAGE IMPURITY CONCENTRATION NEAR
AN ELECTRIC POWER PLANT BY MEANS OF AN ELECTRONIC COMPUTER
D. Sepeshi
METHODS OF CALCULATION OF THE SURFACE CONCENTRATION OF A
GASEOUS IMPURITY DISCHARGED FROM AN ELEVATED SOURCE
L. Nemets
RESULTS OF EXPERIMENTAL STUDY OF SMOKE PLUMES FROM THERMAL
POWER PLANTS
B. Bern
ATMOSPHERIC DIFFUSION AND STRUCTURE OF THE AIR FLOW ABOVE A
NONUNIFORM UNDERLYING SURFACE
M. Ye. Berlyand and Ye. L. Genikhovich
PROCEDURE FOR CALCULATING THE POLLUTION OF THE ATMOSPHERE WITH
DISCHARGES OF INDUSTRIAL PLANTS AND THERMAL POWER PLANTS
R. I. Onikul
STATISTICAL FORECASTING AVERAGE ATMOSPHERIC POLLUTION
A. Kaspshitzky
iv
-------
METHOD OF CALCULATING THE DEGREE OF ATMOSPHERIC POLLUTION
K. Budzinsky
ON THE DETERMINATION OF DIFFUSION PARAMETERS FOR
ACTUAL LOCATIONS
A. Lehmann
TURBULENCE IN THE LOWER 500 M LAYER AND DIFFUSION OF IMPURITIES
I. V. Vasil'chenko and P. A. Vorontsov
ATMOSPHERIC TURBULENCE AT SMALL HEIGHTS
N. Z. Pinus
Part III
ON THE REMOVAL OF IMPURITIES FROM THE ATMOSPHERE BY CLOUDS
AND PRECIPITATION
Ye. S. Selezneva and 0. P. Petrenchuk
SOME ASPECTS OF THE ADOPTION OF AUTOMATIC METHODS OF DETERMINING
ATMOSPHERIC POLLUTANTS
N. Sh. Vol'berg, G. V. Gal'dinov, and V. Z. Al'perin
RECORDING OF SULFUR DIOXIDE CONTENT AT THE OUTSKIRTS OF A CITY.
COMPARISON OF MEASUREMENT RESULTS FOR A VALLEY AND AN ELEVATION
H. Mrose and W. Warmbt
THEORETICAL AND EXPERIMENTAL STUDY OF THE ASPIRATION COEFFICIENT
OF AEROSOLS
S. P. Belyayev, V. M. Voloshchuk, and L. M. Levin
MECHANISM OF PHOTOCHEMICAL POLLUTION OF THE URBAN ATMOSPHERE
M. T. Dmitriyev, N. A. Kitrosskiy, and V. A. Popov
PROCEDURE FOR DETERMINING THE CONTENT OF TRACE ELEMENTS IN
PRECIPITATED WATER
T. N. Zhigalovskaya, R. I. Pervunina, V. V. Yegorov,
E. P. Makhon'ko, and A. I. Shilina
CONTENT OF HEAVY METALS IN THE AIR OF CERTAIN REGIONS OF THE USSR
T. N. Zhigalovskaya, V. V. Yegorov, S. G. Malakhov,
A. I. Shilina, and Yu. P. Krasnopevtsev
ON THE DESIGN OF A MEASURING NETWORK FOR AIR POLLUTION IN THE
GERMAN DEMOCRATIC REPUBLIC
W. Warmbt
v
-------
CONTENT OF PHOTOOXIDANTS IN URBAN AIR
Yu. G. Fel' dmari
STUDY OF AIR POLLUTION AND ATMOSPHERIC PRECIPITATION RESULTING
FROM ARTIFICIAL MODIFICATION OF CLOUDS
Sh. G. Gavasheli
MICROCLIMATIC CHARACTERISTICS AND HYGIENIC EVALUATION OF THE
RELATIVE EMPLACEMENT OF INDUSTRIAL AND RESIDENTIAL COMPLEXES
IN THE REGIONS OF SIBERIA
L. I. Koldomasov and M. T. Zenin
NUMERICAL CHARACTERISTICS OF METEOROLOGICAL CONDITIONS ASSOCIATED
WITH PERIODS OF HEAVY ATMOSPHERIC POLLUTION IN WESTERN SIBERIA
I. A. Shevchuk and L. I. Vvedenskaya
EXPERIENCE IN SIMULATING THE PROPAGATION OF NOXIOUS SUBSTANCES IN
THE SURFACE ATMOSPHERIC LAYER OVER PLANT SITES AND SURROUNDING
GROUNDS
V. M. El'terman
SPECIAL CASES OF VERTICAL CURRENTS
I. G. Diaconescu, M. Frimescu, I. Moroianu, and A. Moroianu
SYNOPTIC CONDITIONS OF FORMATION OF A VERY STABLE ATMOSPHERIC
BOUNDARY LAYER
F. Rein
vi
-------
PREFACE
The present volume constitutes Part II* of translations of papers
presented in Leningrad during July 1968 at the International Symposium on
Meteorological Aspects of Air Pollution. The original papers delivered at
the Symposium were edited by Prof. M. E. Berlyand and published as a report
in 19 71 by the Hydrometeorological Publishing House in Leningrad.
The report contains a total of 40 papers, 37 of which are in Russian
and three, in German, together with their accompanying abstracts. In this
volume we present a collection of translations of a considerable number of
papers from the original report. It is planned to publish translations of
the remaining papers of this report in a subsequent volume.
A review article by T. A. Ogneva, "International Symposium on Meteoro-
logical Aspects of Atmospheric Pollution" (Izv. VGO**, Vol. 101, No. 4, 1969,
pp. 395-396) provides interesting background material regarding the Symposium.
For this reason we present our translation of Ogneva's paper in the following
paragraphs.
"At the present time, considerable attention is being given to the study
of atmospheric pollution in order to improve the sanitary status of water and
air reservoirs in urban areas and workers' settlements and to intensify
measures aimed at the preservation of nature. In working out the measures
designed to decrease the pollution of air by noxious impurities, a special
role is played by the consideration of meteorological factors, that substan-
tially determine the behavior (dispersal) of impurities in the atmosphere.
This problem was the subject of an International Symposium on Meteorological
Aspects of Atmospheric Pollution, held in Leningrad at the A. I. Voeykov
Main Geophysical Observatory on 22-31 July 1968. The organizers of the Sym-
posium were the State Committee of the USSR Council of Ministers on Science
and Technology (Scientific Council on the Problem 'Protection of the Air
Reservoir from Pollution by Noxious Substances') and the Main Administration
of the Hydrometeorological Service, Council of Ministers of the USSR.
"The Symposium was attended by scientists and specialists (meteorologists,
hygienists, chemists, power engineers, metallurgists, etc.) from Bulgaria,
Hungary, Vietnam, the German Democratic Republic, Poland, Rumania, Czechos-
lovakia, and the Soviet Union. Forty-five reports were delivered, and there
was an animated discussion on the subjects of physical principles and methods
of calculation of dispersal of industrial emissions in the atmosphere; meteoro-
logical parameters determining atmospheric diffusion and intensity of atmos-
pheric pollution; procedures and equipment for observing atmospheric pollution,
and experimental studies of pollution of the atmosphere by noxious impurities.
* Part I can be found in Volume XII of the AICE Survey of USSR Air Pollution Literature.
** Izvestiya Vsesoyuznogo Geograficheskogo Obschestva, Tom 101, vip. A, Iyul' - Avgust 19&9«
vii
-------
"On the first day, M. Ye. Berlyand (Main Geophysical Observatory)
reported on the status and prospects of development of meteorological studies
of atmospheric pollution. He noted the following basic problems in the devel-
opment of further research: study of meteorological conditions in the presence
of heavy air pollution, including the mechanisms of turbulent exchange in the
presence of inversions, a complex topography, and urban buildings; development
of automatic and recording apparatus; study of the vertical distribution of
noxious impurities; accumulation of observational data and development of
methods of forecasting pollution, etc.
"In the report of I. L. Varshavskiy (State Committee of the USSR Council
of Ministers on Science and Technology), 'Basic Trends of Scientific Research
on Protection of the Air Reservoir from Pollution', the problems mentioned
as the major ones involved the study of the biological action and hygienic
importance of atmospheric pollutants, the influence of meteorological condi-
tions on the distribution of noxious impurities in the atmosphere, the removal
of noxious impurities from waste gases of industrial enterprises, and the
removal of toxic components from the exhaust gases of internal combustion
engines.
"Considerable interest was elicited by the report of M. S. Gol'dberg
(A. I. Sysin Institute of General and Communal Hygiene) on the 'Hygienic
Standards for the Maximum Permissible Content of Noxious Substances in Atmos-
pheric Air.' He noted that hygienic standards worked out on the basis of
experiments and actual laboratory and clinical studies constitute the scien-
tific basis of environmental improvement measures in the struggle with
atmospheric pollution, and make it possible to evaluate the results from the
standpoint of providing the optimum living conditions for the population.
"Papers by a number of members of the Main Geophysical Observatory pre-
sented studies dealing with the physical validation of the procedure for
calculating the pollution of the atmosphere with emissions of industrial
enterprises and steam power plants, an investigation of turbulence in the
lower 500-meter layer, an experimental study of pollution in industrial areas
on the territory of the Soviet Union, the establishment of the relationship
between meteorological conditions and air pollution in urban areas, the devel-
opment of automatic methods of determination of "atmospheric pollutants, etc.
There were also reports on studies conducted in collaboration with Moscow
University (S. M. Gorlin) in which the influence of irregularities of the
earth's surface on the characteristics of the air flow was investigated in
wind tunnels in connection with the problem of impurity diffusion.
"Members of the Institute of Experimental Meteorology (USSR) presented a
series of reports on the study of diffusion of impurities and characteristics
of turbulence from a 300-meter meteorological mast in Obninsk. Individual
reports were also delivered by members of the Institute of Applied Geophysics
(Ye. N. Teverovskiy), the Central Aerological Observatory (N. Z. Pinus), the
viii
-------
F. F. Erisman Scientific Research Institute of Hygiene (R. S. Gil'denskiol'd
and B. V. Rikhter), and others.
"Specialists from Czechoslovakia (A. I. Vesecky, head of delegation),
the German Democratic Republic (W. Warmbt, head of delegation), Poland
(W. R. Parczewski, head of delegation) introduced the participants of the
Symposium to the broad range of experimental research on the problem of
atmospheric pollution. D. Szepesi (Hungary) reported on studies of meteoro-
logical conditions of turbulent diffusion, and G. I. Diaconescu (Rumania), on
studies of thermal stratification of the atmosphere in connection with the
pollution problem.
"All the participants of the Symposium came to the conclusion that at
the present stage of the struggle with atmospheric pollution, the methods of
purification and the construction of high stacks should be combined for the
purpose of effectively utilizing the dispersing capacity of the atmosphere.
"The need for an extensive introduction of meteorological investigation
of industrial projects and for the forecasting of air pollution was noted.
One of the primary problems noted was that of validating and standardizing
the methods of calculation of atmospheric pollution, the meteorological and
hygienic procedures in studying the chemical composition of the atmosphere,
and the meteorological parameters determining this composition."
It is hoped that the papers presented in this volume will be conducive
to a better appreciation of the meteorological-air pollution investigations
conducted in the USSR and in a number of the Soviet-block countries. As the
editor of this volume I wish to thank my co-workers in the Air Pollution
Section of the Institute for their valuable assistance.
M. Y. Nuttonson
April 1972
-------
AUTOMATION OF INFORMATION PROCESSING INVOLVED IN
EXPERIMENTAL STUDIES OF ATMOSPHERIC DIFFUSION
A. S. Zaytsev (USSR)
From Glavnoe Upravlenie Gidrometeorologicheskoy Sluzhby Pri Sovete Ministrov SSSR. (Chief Administration of
the Hydrometeorological Service Under the Council of Ministers of the USSR.) "Meteorologisheskie Aspekty
Zagryazneniya Atraosfery". (Meteorological Aspects of Air Pollution.) Sbornik dokladov na mezhdunarodnon
simpoziume v Leningrade - Iyul' 1968 g. (Reports delivered at the International Symposium in Leningrad -
July 1966.) Pod redaktsiey d-ra fiz.-mat. naufc U. E. Berlyanda. . (Edited by Prof. M. E. Berlyand.)
Gidrometeorolgicheskoe izdatel'stvo, Leningrad, p. 130-136 (l97l). (Hydrometeorological Publishing House,
Leningrad, (1971).)
Experimental studies of atmospheric diffusion involve the necessity of
recording the change of meteorological parameters and impurity concentrations
in time and space. They include measurements of temperature, wind, turbulence
characteristics, and concentration of a large number of noxious impurities.
The data obtained require various types of treatment, which consumes consid-
erable time when manual methods are employed. Modern experimental studies
have become such that a relatively short observation time (of the order of
minutes and hours) requires processing of the data obtained that occupies a
considerably longer period of time (of the order of days and months).
In the organization of effective control of atmospheric pollution, par-
ticular importance is assumed by the efficiency of the treatment of the in-
formation obtained. The problem of automating the measurements of noxious
impurity concentrations also includes the creation of automatic gas analyzer
instruments, but this paper will discuss only the automation aspects of the
processing. In view of the fact that in studies of atmospheric diffusion
the primary information obtained is in different forms (electric signal,
weighed sample of dust on filter, stereopairs of smoke plume photographs,
etc.), the solution of the problem of automating the processing of such in-
formation, in contrast to, for example, aircraft meteorological information
[1], must be carried out by using different methods.
There are two approaches to solving the problem of automation of the pro-
cessing. One involves primary processing directly under field conditions and
the presentation of information in a form that is suitable for further machine
processing. The second involves only the recording of measurable quantities
under field conditions followed by processing under stationary conditions.
Data on the time and space variability of atmospheric parameters make it
possible to evaluate the detail with which their measurements should be
carried out (for example, in airplane studies, to obtain details of the distri-
bution of meteorological elements, it is necessary to have information every
- 1 -
-------
20-30 m along the vertical and every 1-2 km along the horizontal [l]). For
impurity concentration fields under urban conditions, the frequency of dis-
crete measurements is difficult to establish at the present time, so that it
is necessary to strive for a continuous recording.
In view of the above, it appears expedient to use the second approach
to solving the problem of automating the processing of data on atmospheric
pollution and some other meteorological parameters.
Since the chief forms of information carriers are graphical tracings on
photographic paper or a strip chart, one cannot help reading the ordinates
with a certain discreteness. Industrial devices designed for this purpose
use optical reading methods employing electron tubes or photodiode matrices
and are designed for automatic reading without the participation of an oper-
ator. Also known are analog computers that solve individual problems with
the aid of recordings on tape. A considerable limitation is imposed on the
use of automatic devices by requirements regarding the quality and quantity
of the tracings on strip charts. The number of tracings is usually limited
to one or two, and they should not intersect. The line thickness is strict-
ly standardized, there must be no grid of any kind on the chart, the record-
ings must not be discontinuous, should be sufficiently smooth, etc. It
follows from the above that the experimental data obtained under expedition-
ary conditions, which are frequently quite complex, can satisfy the enumer-
ated requirements in only rare exceptions. It thus becomes necessary to
develop methods of semiautomatic reading of graphical information, which in-
volve the participation of an operator who follows the recording on the chart.
Known techniques of this kind are very limited [2, 3, 9] and do not solve
the problem of automation of the processing as a whole. Considering the possi-
bility of direct input into a digital electronic computer and the relative
simplicity of the development and execution of the scheme, a method of en-
coding in digital code on telegraph tape was chosen [4] involving the use of
industrial schemes and units, which made the system more versatile. In line
with technical requirements, the measurement accuracy is ±0.57» of the measur-
ed value, the measurement frequency, 4-10 Hz, the tape speed, 1-5 m/hr (the
speed being limited chiefly by capabilities of the operator), and the tape
width, up to 300 mm.
The use of the method of semiautomatic digitizing made it possible to
remove a number of substantial limitations on the recording. It is possible
to process the tapes in the presence of a large number of intersecting trac-
ings, which can even be discontinuous, and there are no limitations on the
thickness, steepness and color of the tracings. The system consists of a
series of units: 1) conversion of the ordinate to an electrical analog (or-
dinate-analog unit), 2) unit of measurement of the electrical analog with
representation of the data in digital code, 3) conversion of digital codes
and synchronization, and 4) recording unit. The recording unit employed was
to tape puncher. In view of the fact that in many problems it is sufficient
- 2 -
-------
to obtain the ordinates in the form of digits, a variant of the recording
scheme was developed involving digital printing on paper tape in the decimal
system with a frequency of 2 ord/sec. The use of this device substantially
accelerates the processing. Thus, whereas several hours is needed for manual
processing of an oscillogram of overloads for a single airplane flight, only
half an hour is necessary when the semiautomatic device is used. The evalua-
tions performed showed that on the average, the variant with digital printing
accelerates the processing by a factor of 10, and with perforation, by a fac-
tor of 40.
As was stated above, experimental studies of atmospheric diffusion yield
different types of information, including those requiring preliminary process-
ing before their processing can be automated. Thus, the preliminary process-
ing of data of the stereogrammetric photography of smoke plumes is carried
out in a photographic laboratory, then the coordinates of the smoke plume
boundaries are digitized on stereocomparators, the information being present-
ed in a form suitable for input into digital computers (on punched cards or
punched tape) for subsequent calculations. The developed methods of auto-
matic recording of dust concentration [5] also assume a preliminary photo-
metering of tape filters under laboratory conditions, following which the in-
formation obtained can be automatically processed by using the methods describ-
ed above.
The second approach to automation - initial treatment of information under
field conditions with output on punched tape - can be effectively used for
recording the wind velocity profile in the surface layer (for example, calcu-
lation of contacts, use of moving averages for a given time segment, encoding
and output of mean velocities on punched tape), and also for automating pilot
balloon (particularly reference observations [10]).
Switching immediately to the results of processing and calculations by
means of the above-described methods, we should note that the recordings of
airplane overloads have been the ones processed most effectively at the pre-
sent time. Measurements of overloads of the airplane's center of gravity
are carried out in expeditionary studies of atmospheric pollution. An analy-
sis of data from observations in the area of a number of large-capacity emis-
sion sources is given in adequate detail by P. A. Vorontsov [6]. For this
reason, only some aspects of the semiautomatic processing of the data will
be described below. The procedure used in the flights usually includes sound-
ing of the atmospheric boundary layer with ascent and descent in plateaus,
at each of which the airplane overloads are recorded for 3 minutes. The over-
load gauge used is an MP-66 instrument, with loop oscillographs being used as
recorders. The observational material in the area of the Shchekino State Re-
gional Electric Power Plant (SREPP) (1963) and the Moldavian SREPP (1965)
were digitized semiautomatically, and computations of the turbulence charac-
teristics (using the Lyapin-Dubov method) were carried out on the "Ural-A"
digital computer:
- 3 -
-------
k = g , L = 2xV, (1)
where k is the turbulence coefficient, andl^'ljis the mean value of fluctu-
ations of the vertical wind velocity component:
13)' —
1
1
PF [H, o>) N
N
V
\h,-h |
(2)
where hj[ is the running ordinate of the overloads, R is the mean ordinate,
F(H,
-------
Fig. 1, Vertical profiles of Fig. 2. Me^n profiles of the
fluctuations of the vertical size ol turbulent disturbaoc-
irind velocity component for es for.areas of the Moldavian
tew hrct5 arc Shr-hekiro
'Fig, 3. Frequency of sizes of turbulent disturbances in tfce area of
Sfichekino EfttPP during tlie day CU and during interasiiate hours
- 5 -
-------
Analysis of data on the turbulence coefficient showed that in the course
of the expeditions, the characteristic values k«25-30 m^/sec. Maximum values
were observed at a height of approximately 500 m.
Semiautomatic methods were used to process other measurements as well.
Experimental studies of atmospheric diffusion involve measurements of oscil-
lations of the wind direction. The gauge used is an M-45 wind recorder.
Wind direction recordings obtained in the course of a study of atmospheric
pollution near a synthetic fiber plant were processed by using a variant of
the above-discribed unit with output of the data on a digital printer. In
addition, calculations of the variance of the wind velocity were carried out
for different averaging intervals, and the values obtained were compared with
the stability parameters of the atmospheric surface layer. Results of the
analysis are given in [8],
In working out the procedure for determining hydrogen sulfide and carbon
disulfide [7], a dynamic sulfur dioxide concentration dispenser and a GKP-1
automatic gas analyzer were used. Digitizing of the recording of concentra-
tions before and after the absorber was carried out semiautomatically and
demonstrated the high effectiveness of the use of such processing methods in
studies involving gas analyses.
In conclusion, it should be noted that work on the automation of measure-
ments and processing of observational materials on atmospheric pollution and
necessary meteorological parameters has only begun and will be continued as
part of a program of development of methods and instrumentation aimed at or-
ganizing an effective control of atmospheric pollution and meteorological
support of such work.
LITERATURE CITED
1. B a bh n ob B H, MaueiOB P. A. I"Io,nyaBTOMaTimecKoe KOAiipoBaime rpa-
((miecKoii iiM^opMauiui Tpy^bi fTO, Bbin 234, 1968
2 BopoiiuoB n A TypGy/ieiiTiiocTb n Bep-rHKa.ibiibie tokii b norpaimwiiOM c.ioe
aTMOcifepbi riupoM€Tcoii3flaT, J1, 1966
3 Apeiiep A A Ilo.iyaDTOMaTiriccKoc cmiTbioaniie rpaimecKoii imtfiopMamiii.
MeTeopojiorim ii rii/ipo.ioriiH, jYs 5, 1968
4. KoHbKOB C. A O periiCTpaumi xoimeiiTpamiii iih^h b aTM0C(J)epe Tpyabi ITO,
Bun 234, 1968
5. 0 r ii c b a T A MiiKpoMeTeopo.ionisccKiic xapaKTepiicTiiKii yc.iooiii'i 3arpn3iic-
IMB aTMOC(|>epbl (cm llaCT cCopilllk)
6. IlaBJiciiKo A. A , K y 3 b m ii ii a T A K MCTO/iaiu onpeac.iciinn ccpoBOAopoAa
ii cepoyr.nepo.na Tpyaw rTO, bum 234, 1968
7. nana B T.TaiapeiiKOBE B CTpyKTypiibiii ana.ni3aTop Tpyaw HOAH,
t. XXV, 1964
8. riiiiiyc H 3. ii ap OciioBHbie npimminbi aBTOMani3amiii o6pa6oTK» caMO.icT-
iioft MeTeopo.ionmecKoii iiH())opMamiH MeTeopo^onin ii Tiiiipo^oriifl, Afs 9,
1964
9Bo^b6eprH III, T a a b a h h o b I* B., A.ibnepiinB 3. Bonpocw biic-
ApetuiH aoTOMdTimecKiix mctoaod onpeAe/iemia a r.Moc^epi'bix 3arpH3Mennii
(cm. tiacT cCopmiK).
- 6 -
-------
10 Bullet T and Dupriez I L A Semi-Automatic Render for the Reading
of Comcntional Climatological Diagram Symposium on the Data Processing
for Clitnatologic.il Purposes. Acheville, 1968 "
II. Rachclc H and Duncan L D Desirability of using a fast sampling rate
for compul;ng wind wclocity from pilot-bailoon data. Monthly Wcalher Rev.
I vol 95, No 4, April 19G7
(
A. S ZA1TSEV
THE AUTOMATION OF INFORMATION TREATAAENT IN CONNECTION WITH
EXPERIA\ENTAL INVESTIGATIONS OF ATMOSPHERIC DIFFUSION*
Experimental investigations of atmospheric diffusion include the
measurements of meteorological and turbulent air current characte-
ristics and also pollutant concentrations by quick-response sensors
The automation of data treatments is connected with primary treat-
ment, information memorization, and subsequent treatment on elec-
tronic computers
The employment of helicopters and cars supplied by automatic
instruments for air pollution investigation leads to the necessity of
prompt processing of information in stationary conditions. For this
purpose, the method of automatic and semiautomatic treatment of
information read out from oscillographic tapes is used
For some problems, memorization of coded measured quantities
is possible directly in field conditions. In connection with this, the
methods of automatic reading of sensors, information memorization,
calculation of averages for a definite time period and data coding on
punched tape are used
With this, the final data treatment is carried out on electronic
computer according to beforehand compiled programs {calculations
of turbulence coefficient, wind characteristics, etc.). The calculations
and analysis of turbulent coefficient and of the parameter which cha-
racterizes the wind direction variability are carried out on obtained
measurement results, and records of automatic gas analyzers are
treated, too.
* 'Editor's notes The abstract is presented as given in English with the original Russian
-------
MICROMETEOROLOGICAL CHARACTERISTICS OF ATMOSPHERIC POLLUTION CONDITIONS
T. A. Ogneva (USSR)
From Glavnoe Upravlenie Gidrometeorologicheskoy Sluzhby Pri Sovete Ministrov SSSR. (Chief Administration of
the Hydrometeorological Service Under the Council of Ministers of the USSR.) "Meteorologisheskie Aspekty
Zagryazneniya Atraosfery". (Meteorological Aspects of.Air Pollution.) Sbornik dokladov na raezhdunarodnom
simpoziume v Leningrade - Iyul1 1968 g. (Reports delivered at the Irfcernational Symposium in Leningrad -
July 1968.) Pod redaktsiey d-ra fiz.-mat. nauk U. E. Berlyanda. (Edited by Prof. M. E. Berlyand.)
Gidrometeorolgicheskoe izdatel'stvo, Leningrad, p. 157-146, (1971). (Hydrometeorological Publishing House,
Leningrad, (l97l).)
The dispersal of impurities in the atmosphere depends considerably on
the meteorological conditions in its lowest, surface layer. This paper
discusses some results of studies made at the A. I. Voyeykov Main Geophy-
sical Observatory on the meteorological characteristics of the surface
layer, sometimes called the micrometeorological characteristics [8].
I. In the Soviet Union, systematic micrometeorological observations
are being conducted at so-called heat balance stations. The set of obser-
vations includes measurements in the 0.5-2 m layer of differences in air
temperature, water vapor pressure, wind velocity, and also measurements of
the temperature and humidity of the upper 20-centimeter layer of the soil.
These observations make it possible to supplement the existing experimental
data (obtained chiefly during short-term expeditionary observations) on the
time and space relationships (within the boundaries of the Soviet Union)
of the chief characteristics of turbulent mixing - the thermal stratifica-
tion of the surface layer and the turbulence coefficient at a height of 1 m.
The distribution of stations whose observational material was used in
this paper is shown in Fig. 1.
It should be noted that these observations are conducted on platforms
of weather stations located mainly in open, comparatively flat areas on
surfaces with grassland vegetation characteristic of this particular land-
scape zone. All the data considered here pertain to such surfaces. Table 1
gives characteristic values of temperature differences at the level of
0.5-2 m (At) in the warm season for daytime (1 P.M.) and nighttime (1 A.M.)
in the main landscape zones. These values may be considered characteristic,
since they were obtained on the basis of daily observational data averaged
over several years (from 4 to 10) and taken at individual stations. The
following regularities of variation of characteristic values of At in time
and space may be noted.
In the daytime, the sign of At is positive during all the months of
the warm season, and this promotes turbulent mixing; on the contrary, at
night, an inversion is observed that decreases the turbulence intensity.
- 8 -
-------
In the daytime, the lowest values of At, which do not exceed 0.5°C.,
are characteristic of the forest zone; to the north of the latter (in the
tundra) and to the south (in steppes, deserts), the values of At increase,
and in the deserts exceed 1°C.; at night, the zonal distribution is less
distinct, and in the forest, steppe and desert zones, the inversions do
not exceed 0.5°C.; they reach 1°C. only in the zone of piedmont steppes and
in floodplains.
In comparison with the vertical temperature gradients of the free atmos-
phere, the characteristic gradients of the surface layer in the warm half of
the year are two orders of magnitude (50-100 times) greater, and are pro-
nounced superadiabatic gradients in the daytime and inverted gradients at
night.
60 70 80 90 tOO 110 120 130
Fig. 1. Map of the stations.
The sufficiency of characteristic At values can be seen from Fig. 2.
As an example, the latter shows diagrams of the seasonal variation of
frequencies of different At values in the daytime and at night for weather
typical of a given zone. They pertain to stations located in the desert,
a typical steppe, and in zones of coniferous and mixed forests. From the
diagram it is evident that at night, independently of the landscape zone,
in 60-70% of cases there is a predominance of At's which in absolute value
do not exceed -0.5°C., and that their values amounting to less than -1°C.
occur in only 10-15% of the cases; this proportion is also characteristic
of the seasonal variation. In the daytime, the frequency of the different
values shows a more pronounced variation in seasonal course and depends on
the landscape zone. In the desert, At>l°C. is observed in 60-70% of the
cases from May to October. In the steppe zone, such At values are observed
in 60% of the cases in August only, and in all of the remaining months,
- 9 -
-------
Table 1
Characteristic Values of At in Different Landscape Zones.
Zone
V
VI
VII
VI11
IX
X
V
VI
VII
VIII
IX
X
1 A.
u.
1 P.!
a.
Pretundra thin forest
-0.6
-0,8
-0,3
0.9
0,7
0.4
Coniferous forest
-0.3
-0,5
-0,5
—0.4
-0,3
-0,2
0,5
0,5
0,5
0,4
0.3
o.l
kLixeti. forest
-0,5
-0,6
-0.4
-0.4
-0,4
—0,2
0.4
0,4
0.4
0.3
0,2
o.l
Deciduous forest
-0,4
-0,4
-0,4
-0.4
-0,4
-0,4
0,4
0,5
0,5
0.4
0.4
0.3
Floodplains
-0,5
-0,6
-0.7
-0.7
—0.5
—0,4
0,4
0,6
0.6
0,5
0,4
0,3
Steppelike grassland
-0,6
-0,6
-0.6
-0,6
-0,5
-0,4
0,6
0,6
0,8
0.8
0.7
0,5
Piedmont steppes
-0,9
-1.1
-1,1
-1.0
-1,0
-0,8
0,6
0,5
0,7
0.9
1.0
0.8
Grassland steppes
-0.4
-0.5
-0.5
-0.4
-0,4
-0,3
0,6
0,7
0,6
0.5
0,5
0.4
Typical steppes
-0,4
-0,4
-0,4
-0,4
-0,5
-0,3
0,7
0,9
0,9
0,8
0,7
0.4
Deserts
-0,3
-0.4
-0,5
-0,5
-0,5
-0,4
1.1
1.2
1.2
1.1
1.0
0.7
-------
100
(a)
, v Variable
vc; cloudiness
1 P.M.
100
V VI VII VIII IX X V VI VII VIII IX X
(b)
1 A.M. Clear 1 P.M.
V VI VII VIII /XX V V! VII VW IX X
, v Variable
Clear cloudiness
1 A.M.
1 P.M.
50-
1 A.M.
m
<¦-1
-0,5,-1
0,-0,S
>0
1 P.M.
>1
0,5-1
0-0,5
<0
V VI VII I'llI IX X V V! VII VIII IX X VI VII VIII IX VI VII VW IX
Fig. 2. Frequency of different values ofAt.
a - steppe, b - desert, c - zone of mixed forests, d - zone of coniferous forests
-------
their frequency is much lower. In the forest zone, the seasonal course of
the frequency of the values At>l°C. is also pronounced, but its greatest
values do not exceed 50%. The frequency of the values At< 0.5°C. also
changes with the zone, increasing from the desert to the forest zone. As is
evident from the diagram, positive At values at night and negative ones
during the day are observed very seldom.
Extreme values of At selected from observational data for 3-5 years
(Table 2) show that Atm do not exceed 5°C. and on the whole, depend little
on the landscape zone. Only a tendency toward an increase of Atmax in the
zones of insufficient humidification (steppes, deserts) can be observed.
The minimum At's are greater than the maximum ones in absolute value. As a
rule, extreme At values are observed during the summer months.
Table 2
Extreme Values of At
Zone
At,
max
Month
Month
Pretundra thin
forest 2,5 VII -3,7 VIII
Coniferous forest 3,4 VIII —4,3 VI
Mixed forest 3,3 Vll —3,7 VII
Deciduous forest 2,4 VIII —3,6 IX
Floodplains 3,2 VI —3,7 VI
Steppelike grasslai ds 3,3 VI —3,6 IX
Piedmont steppes 3,7 V 4,6 V
Grassland steppes 3,7 VI —4,7 VI
Typical steppes 3,8 VI —4,7 V
Desert 4,5 VIII -4.3 I
As we know, the formation of any given values of temperature (lapse
rates) is determined by the proportion of the components of the heat balance
and its main component, the radiation balance. The proportion of the com-
ponents of the heat balance may be characterized in a simplified manner by
the radiation balance, wind speed and state of humidification of the surface.
Analysis of the relationship between At and the radiation balance [6]
shows that the temperature differences increase with increasing radiation
balance and decrease with increasing wind speed in the zone of forests and
steppes for both positive and negative values of the radiation balance.
In the desert zone, the largest positive differences do not arise at weak
wind speeds, but at a speed of 2-4 m/sec. A change of sign of the difference
in the forest zone takes place at higher values of the radiation balance than
in the zone of steppes and deserts. As the wind speed increases, the change
of sign is also observed at a higher value of the radiation balance as com-
pared with weak speeds.
- 12 -
-------
II. When characterizing the turbulent exchange, it is necessary to
consider not only the values of the differences in the temperature of
the surface air layer, but also the wind speed. A complex characteristic
of the thermal stratification of the atmosphere is the Ri number or its
analog, the parameter At/u^ which is widely used in the USSR [7]. Its
characteristic values for the main landscape zones during the day and at
night in the warmest month (July) are shown in Table 3. On the whole, the
parameter At increases from the forest zone to the desert zone from 0.10 to
U2
1
0.25 in the daytime and from -0.09 to -0.17 at night (its values were calcu-
lated for cases of more or less developed turbulence, when u-^>l m/sec). This
means that as a result of the influence of thermal stratification, the char-
acteristic values of the turbulence coefficient at a height of 1 m during
the day increase 1.5-2-fold in the forest zone, 2-2.5-fold in the steppe
zone, and 2.5-3-fold in the desert zone. At night, the effect is approxi-
mately the same.
III. The observational materials of the heat balance stations make it
possible to obtain the distribution of the turbulence coefficient k^ in
space and time at a height of 1 m. V. P. Gracheva [4] plotted a map of
characteristic values of k-, over the territory of the Soviet Union for July
during hours of strongest heating of the surface (1 P.M.). The k^ values
were calculated by the heat balance method for a period of 5-10 years.
The turbulence intensity is distributed by zones, increasing from north
to south. The highest values of k^, up to 0.25 m^/sec, are characteristic
of the areas of steppes and deserts; in the
forest zone, the turbulence coefficient de-
creases to 0.15-0.10 m^/sec. The mean monthly
k, values are fairly stable; they change only
slightly from year to year, in the range of
10-15% (the standard deviations amount to
0.02-0.03). July is the month of the strongest
turbulence, as is evident from Fig. 3 [3],
which shows the annual course of k^ for the
steppe and forest zones in the daytime and at
night. No marked differences in the annual
course of daytime k-^ values are observed in
the different zones; there are differences
only in absolute values and in amplitude, which in the forest zone is greater
than in the steppe zone. The nighttime values of k-^ have an annual course
opposite to that of the daytime ones, i.e., in summer a minimum is observed,
and in winter, a maximum. These relationships are determined by both the
influence of thermal stratification and the character of the wind speed.
The method widely used in the Soviet Union, of determining k^ on the
basis of measured lapse rate data (using the results of the semiempirical
Table 5
Limits of A t/u| Values in
Different Landscape Zones in July.
Zone
1 A.M.
1 P.M.
Forest
Steppe
Desert
-0,03-0.16
-0,13-0,16
-0,15-0,17
0,10-0,15
0,17-0,19
0,21-0,25
- 13 -
-------
theory of turbulence or the heat balance equation) makes it possible to
obtain the turbulence intensity for specific periods of time, in particu-
lar, when conducting surveys of pollution in cities or at individual
industrial facilities. Figure 4 shows a graph of the daily course of k^
obtained by G. P. Rastorguyeva from data of observations in the forest
and forest-steppe zones in summer and winter. In summer, the daily ampli-
tudes of k^ are twice as high or more than those in winter. The largest
values of ky occur in the midday hours.
IV. Wind is one of the chief meteorological parameters determining
the diffusion of impurities. This parameter (as compared with other
meteorological elements) is subject to the strongest variations under the
influence of different kinds of obstacles such as topography of the terrain,
building system, and vegetation. Wind variations caused by the topography
are studied experimentally and are characterized by certain general regu-
larities. A number of observational results are reported in the monograph
[5]. The latter gives a table showing that the greatest speed variations
K,
0,2
II HI IV V VI VII VIIIIX X XI XII
Fig. 5.
Annual course of in the
steppe zone during the day (l), at
night (2) and in the forest zone dui
ing the day (ft), and at night (3).
K, ra^/seo
0.2Sr
0J0
0,15
0J0
nos
r > I I
I 3 5 7 9 II 13 15 17 13 2123 1
Time
Fig. b. Daily course of in
summer (l) and winter (2).
are observed at the tops and in the upper portions of windward slopes,
where the speed increases by 30-70%, and in valleys and lower parts of
slopes through which the wind does not blow, where it decreases by 20-30%.
At night, the speed variations are greater than during the day. As the wind
speed increases, its differences in different topographical forms decrease.
The wind direction also' undergoes substantial changes in a dissected area,
manifested in particular forms of the wind rose.
In surveys of pollution of specific areas, the program of work includes
special observations of the wind variation. I. I. Solomatina conducted such
observations in the area of the Shchekino SREPP during the winter and summer
- U -
-------
periods and obtained microclimatic corrections for this area. As illus-
trated in Table 4, they amounted to 30-50% in the concave topographic
forms (points 4-6) and lower parts of the slopes. The speed differences
decrease with height over flat and dissected terrain.
Table 4
Ratio of Wind Speed at Observation Point to Speed
over Open Flat Terrain.
Wind
Direction
leight,
m
Points
1
2
3
4
5
6
NE and NNE
1
0,8
1.1
0,9
0.9
0,5
0,7
4
0.9
1,2
1.0
1.0
0.6
0,9
SW and SSW
1
0,8
0,9
0.8
0.8
0.7
1.0
4
0,9
1.0
0,9
0,9
0,7
1.0
V. In analyzing and calculating the diffusion of impurities, it is
essential to have information on the character of the stability of the
wind direction as an indicator of turbulence. Experimental studies of
this parameter in expeditions investigating atmospheric pollution and the
data correlation carried out by V. P. Gracheva [2] made it possible to
draw the following conclusions:
1) the standard deviation of the wind direction a changes over a
range of 20° in the course of 20 minutes;
2) C decreases with increasing wind speed;
3) the thermal stratification has a sub-
stantial influence on CT; as the instability
¦ increases, the wind gustiness increases
sharply, and as the inversion intensifies,
it decreases;
4) as the averaging period increases,
¦ 0 grows, this growth being most pronounced
for short periods (for example, as the period
changes from 20 to 40 minutes, O increases by
25%, and from 40 to 60 minutes, by only 10%;
5) correlation of data for different
areas of the Soviet Union shows that these
relationships are generally universal and fit
into the dependences shown in Fig. 5 (the
figure illustrates the relationship of a to
u^ for different values of At.
15
1 2 3 4 5 6 7 u,
Fig. 5. Relationship of to for
different t's.
- 15 -
-------
VI. The development of conditions of impurity dispersal in cities
Ls greatly influenced by the meteorological regime of the city. There
are known results of studies of changes in thermal regime which are due
to the influence of the urban building system and of changes in transpar-
ency conditions which are due to industrial emissions, as a result of
which a heat island forms above the city. Such studies are now also being
conducted during surveys of atmospheric pollution in specific cities.
The data of Table 5 were obtained by G. P. Rastorguyeva in an industrial
city with a modern building system and a topography with some points exceed-
ing others by 100 m, and with isolated depressions containing water reser-
voirs. The table lists values of air temperature differences at a height of
2 m measured at a weather station (located on the outskirts of the city)
and in different parts of the city at night. These data confirm the pres-
ence of warmer air in the city as compared'with the outskirts. It is
essential to note that a marked lowering of the air temperature was ob-
served in lower topographic forms with a natural surface.
Table 5
Values of Air Temperature Difference at a Height of 2 m at Niglit
Based on Data of Observations from a Weather Station and at
Various Points of a City
Place of
Observation
tt
Place of
Observation
iIf
Suburburban
Areas
City streets
City squares
1 1
J—©
to ¦
.
Low relief
(soil)
Low relief
("asphalt)
-3,7
-2,1
The distribution of the air temperature at night in the city surveyed
shows that its highest values coincide with the most built-up parts of the
city and with the highest places, and the lowest values, with low places.
The temperature contrast within the confines of a city with linear dimen-
sions of 20 x 20 km was 7°C. when its values in depressions were taken into
account, and 3°C. when they were neglected.
- 16 -
-------
LITERATURE CITED
]. ByAUKo M H Hcnapewie b ccicciBeiriibix yc.iODiisix riwpoMeieoii3flaT, JT.
1948
2. Fpaieoa 3 [1 06 ycTOiViiinocTii iinnpanjicmifl BeTpa e npiueMHOM cnoc arao-
ctjjcpbi Tpyiibi rrO, own 158, 1964
3. r paioo a B. n. Hcc/icAooamic TypOy.iciiTiioro pokiima npnacMiioro c.ion doj-
ayxa b paa^itmibix rcorpa^imccKiix pafiona\ Tpyau ITO, Bbin 185, 1966
4. TpancsaB fl TcorpatymeeKoc pacnpcae;ieimc Hos^^nmiCiiTa Typ6yjieimiocT>i
b npiiscMiiOM c.aoc aTMOcifcpbi .hctom b fliicoiioe spend. Tpyflu ITO, dwi 207,
1968
5. MiiKpoKTinMnT CCCP. rioA pea H A TojibuGcpr niflpoMeicoii3aaT, J], 1967.
6. K a c b a ii H A.OriieBaT A, TeptxoBa K M 3aKoiiOMCpiiocrii ii3Mcne-
iiim TeMncpaTypiioro rpaaiicirra b npiiacMiioji c.ioe eosjyxa iia TeppiiTopint
CCCP. Tpyabi rrO, bnii. 234, 1968
7. Onicna T. A HeKotopuc oco6ciuioctii tcmoboto Ca.iauca acnfc^uioi'i tioiscpx-
hocth. rjiapoMCTcom^aT, .1, 1955
8. Ccttoii O T. MuKpOMCieopo.ionin rnflpOMCTCOH3AaT, Jl., 1958
T A. OGNEVA
MlCROiNETEORO LOGICAL CHARACTERISTICS OF AIR POLLUTION
CONDITIONS*
Some investigation results of micrometeorological characteristics
which influence the pollutant dispersion in the atmosphere are con-
sidered in this report. Information on air temperature gradient in the
lower layer of two meters, received on systematic measurement datar
is cited. Its variation in annual and daily course in different geogra-
phical zones on the territory of the Soviet Union and frequencies
of different values of extreme temperature gradient are shown Close
connection of gradients with radiation balance is revealed.
The regularities of turbulence coefficient variation on the height
of lm in time and space are considered, a map of its values at
1:00 p. m. in July is given, from which latitudinal variations are
seen, turbulence coefficient in southern regions of the country being
twice as large as in nothern regions.
A correlation table is given which characterises wind velocity
changes in different relief forms. Wind direction stability as an index
of turbulent mixing intensity is discussed. A diagram built on expe-
rimental data received in different regions is cited. It is shown that
wind direction variation increases with instability growth and de-
creases with wind velocity increasing
Some data on variation of air temperature regime under the effect
of urban regions are reported. The largest differences in comparison
with rural areas are observed in building quarters and also on con-
cave relief forms.
~ Editor's note: The abstract is presented as given in English with the original Russian article.
- 17 -
-------
STUDY OF THE INFLUENCE OF IRREGULARITIES OF THE EARTH'S SURFACE
ON THE AIR FLOW CHARACTERISTICS IN A WIND TUNNEL
S. M. Gorlin, I. M. Zrazhevskiy, and S. P. Ziborova (USSR)
From Glavnoe Upravlenie Gidrometeorologicheskoy Sluzhby Pri Sovete Ministrov SSSR. (Chief Administration of
the Hydrometsorological Service Under the Council of Ministers of the USSR.)_ "Meteorologisheskie Aspekty
Zagryazneniya Atraosfery". (Meteorological Aspects of Air Pollution.) Sbornik dokladovna mezhdunarodnom
simpoziume v Leningrade - Iyul' 1968 g. (Reports delivered at the International Symposium in Leningrad -
July 1966.) Pod redaktsiey d-ra fiz.-met. nault U. E. Berlyanda. (Edited by Prof. M.E. Berlyand.)
Cidrometeorolgicheskoe izdatel'stvo, Leningrad, p. l4?-l6l, (1971). (Hydrometeorological Publishing Houset
Leningrad. (l97l).)
Introduction
In studies of turbulent diffusion it is usually assumed that the under-
lying surface is comparatively flat and the air flow parameters, which influ-
ence the diffusion of an impurity, depend only on the height. Such a distri-
bution of the parameters is appreciably altered by the presence of various
types of obstacles on the earth's surface and especially by a relief. The
collection under natural conditions of data that permit one to take Into con-
sideration the influence of the topography on the impurity diffusion processes
involves major difficulties. For this reason, studies are now being conducted
on the application of numerical methods and electronic computers to the study
of the influence of obstacles on the characteristics of air flow [1].
Significant possibilities in the solution of this problem are due to the
use of wind tunnel simulation, which, as demonstrated by many years' exper-
ience, has proved itself in applications to problems of aerohydromechanics.
However, this path was found to have its own fundamental and technical obstacles.
In the first place, they pertain to the finding of similarity criteria neces-
sary and sufficient for the transition from tests on models in a wind tunnel
to natural conditions. In the second place, it was found very difficult to
produce in a wind tunnel a flow with a variable intensity of a turbulence simi-
lar in structure to atmospheric turbulence. This applies in particular to the
reconstruction of high turbulence intensities or of a gradient turbulent flow
in a considerable portion of the test section of the tunnel while retaining
a spectral composition similar to an atmospheric turbulent flow.
This paper presents results pertaining to the creation in wind tunnels
of a variable turbulence of the forward flow to the influence of turbulence
on the aerodynamic characteristics of various bodies [2], and also to studies
of disturbances of the velocity field and turbulence during flow past different
kinds of obstacles. The experiments were conducted by the authors* in the large
A-6 wind tunnel of the Scientific Research Institute of Mechanics, Moscow State
University, and published in part in [3].
•Measurements of turbulence intensity were carried out under the direction of L. T. Arafailova.
- 18 -
-------
Influence of Initial Turbulence on the
Aerodynamic Characteristics of Various Bodies
Y~un
Problems of the influence of turbulence intensity z——-— ( of the
forward flow on the aerodynamic characteristics of various bodies have long
attracted the attention of researchers. One problem of particular interest
was that of the influence of turbulence intensity and Reynolds Re number
on the characteristics of the lift of an airplane wing. In 1930-1940, these
questions were widely studied in this country and abroad. However, for a
number of reasons, including a small range of variation of Re numbers and £,
a generally inadequate level of aerodynamic measurements, etc., it was im-
possible to find any general regularities in the influence of these para-
meters. However, in recent years, this agreement began to be observed in
the non-coincidence of results of experiments on physically identical models
in tunnels, even on models identical in aerodynamic contour. A systematic
application of principles of the similarity theory to the Reynolds equation
indicates the existence of a dependence of the stream pattern on the degree
of turbulence of external flow. These facts made it necessary to return to
a detailed and thorough study of the problems of influence of turbulence
intensity and Re number.
First, the methods of determination of the degree of turbulence and,
in particular, of its longitudinal component were analyzed. The analysis
showed that in the case of measurement of the intensity with a hot-wire
anemometer, the frequency characteristic of the latter should have a very
wide range, from 0 to tens of kilohertz. When a hot-wire anemometer is not
used to measure low frequencies (up to 200-300 Hz), the measured E may be
smaller than the true value, by a factor of more than 2. When the turbu-
lence intensity is determined by an indirect method from results of sphere
tests for drag or from the difference in pressures in front of this sphere
and behind it, the sphere diameter Dg (for small values of e) should be no
less than 0.1-0.12 of the diameter of the test section of the tunnel, Dts«
The dependence of the turbulence intensity £ in percent of the criti-
cal Reynolds number Recr found experimentally for a sphere has the form
A refinement of the method of determining the turbulence intensity
made it possible to account for the difference in characteristics obtained
in different tunnels for the same model. The change in the turbulence of
the forward flow was found to have a substantial effect on the stream pattern
around the body. For example, as £o changes by 0.2% (from 0.1 to 0.3), the
'' ' "" ' " " m" """ ' " * """ ' the maxi-
- 19 -
-------
In the past, a series of experiments involving measurement of the
critical Reynolds number Recr for a sphere in the atmosphere, in particu-
lar, the known studies of Millikan and Klein [4] showed the measured Recr
to be independent of the weather conditions. The value of Recr obtained
from measurement data is greater than all the measured values of this
number in wind tunnels and comes quite close to the number given by measure-
ments in a wind tunnel with a very small degree of turbulence. For this
reason, wind tunnels are designed to produce the smallest possible degree
of turbulence of the outer flow. The methods of reducing the initial flow
turbulence essentially amounted (for a closed wind tunnel) to decreasing
the turbulence at the approaches to the honeycomb and nozzle by installing
nets in front of the honeycomb. This made it possible to reduce the ex-
ternal degree of flow turbulence to a minimum.
However, the prewar studies of Taylor [5] and other authors revealed
that the drag of bodies in air depends not only on the magnitude of the
fluctuation velocities but also on the turbulence structure. J. I. Taylor
proposed that the ratio of the sphere diameter to the characteristic length
of turbulence be taken as the characteristic of the turbulence structure
for flow around a sphere. From these points of view, one can account for
the mentioned experiments of Millikan and Klein by the very minute values
of this ratio in atmospheric experiments.
Our objective includes the simulation of flow around obstacles commen-
surate in size with the mean characteristic scale of turbulence in the atmos-
phere. Therefore, the influence of the degree of outer flow turbulence will
be entirely different. Experimental data on the nature of the influence of
the degree of turbulence on the stream pattern in the case of comparable
size of obstacles and size of a typical eddy will be given below. In our
studies, a wide variation of turbulence intensity in the test section was
achieved by installing nets of different drags at the exit from the wind
tunnel nozzle. In addition, it was possible to obtain a uniform velocity
field over the entire test section and a very small gradient of turbulence
change along the length of the test section. This fact was particularly
important and set our experiments apart from work done in the past.
The relationship between the intensity of the longitudinal turbulence
component e (in percent) to the drag coefficient of the nets mounted on the
edge of the nozzle may be expressed by
e = 2,5552._
where £ is the drag coefficient.
To determine the turbulence scale and the spectral energy distribution,
measurements were made with a DYSA hot-wire anemometer with frequency cutoff.
The turbulence scale L_ was calculated from the formula , Ho where uo
t L T = _; >
/ max
- 20 -
-------
is the velocity in m/sec, and f is the frequency at which the energy
TTlclX
maximum is located. As is evident from Fig. 1, obtained from experiments
in an A-6 wind tunnel, the turbulence scale at small values of C has a size
of the order of the tunnel diameter. As the degree of turbulence increases,
it decreases sharply, then remains practically constant. This is due to
the mounting of nets at the entrance, since in this case the turbulence
scale will be determined not only by the tunnel diameter, but also by the
mesh size of the nets. Knowledge of the dependences similar to the one
shown in Fig. 1 makes it possible to estimate the ratio of the size of the
obstacle studied to the turbulence scale, and to compare it with the ratio
prevailing in natural conditions.
L
4
3
2
1
0
Fig. 1. Dependence of turbulence scale on turbulence intensity.
t
1
\
N
5 10 " e %
When it is necessary to produce a vertical gradient of the velocity
field and turbulence, one can use the same nets and low drags in the form
of corners located at irregular intervals along the vertical. The vertical
gradient character necessary for experiments with land obstacles was found
on the basis of experiments with a screen of variable surface roughness at
different Reynolds numbers (up to Re = 5 x 10&) and different turbulence
intensities of the forward flow.
The most complete evaluation of the structure of flows may be made by
comparing their power spectra. Treatment of the results of such measure-
ments showed a weak dependence of a considerable portion of the spectrum
(for different values of the wave number k) on the type of measurement,
perhaps with the exception of the case of small degree of turbulence at
the entrance. For comparison, Fig. 2 shows data (f(k) being the spectral
function and V the viscosity) obtained in a meteorological wind tunnel, a
tidal basin, and in the atmosphere above a wavy sea surface [6] and also the
results of our measurements in the A-6 tunnel for the following cases:
(a) hollow tunnel, £0 = 0.2%, u = 20-30 m/sec; (b) Eo = 5%, u = 10-20 m/sec;
(c) £q = 5%, screen with abrasive x = 1480 mm, uav = 10 m/sec; (d) E„ =
5%, smooth screen, x = 3500 mm, = 10 m/sec; (e) £0 = 5%, screen with
- 21 -
-------
abrasive, x = 3500 mm, u = 10 m/sec.
av
.log
F(k) u'
(e vs)I
Fig. 2. Dependence of spectral function
on wave number.
1 - wind tunnel, 2 - tidal basin,
3 - atmosphere above wavy sea surface,
it - A-6 wind tunnel.
-2
X
N..
V*
<
»S5bt,
•
<
• n
it *>
t
1
I v
t.
~ <
t
N°'«
A,
S*
s.
*1
• 2
o 3
°2o
$
9
-t, -3 .2 -1 0
log
{e \'3H
O 12 3*56 7 8 9 10 t%
Fig. 3. Dependence of relative resistance of various bodies on the turbu-
lence intensity of forward flow.
1 - sphere on screen (Re = 't x 10^), 2 - flat plate, 3 - cylinder above
stall (Re = 4 x 10^), k - sphere above stall (Re = r x 105). 5 - cube ,
(Re = 4 x 105), 6 - beam (Re = 3 x 105), 7 - car on screen (Re = k x 10°),
B - sphere on screen, 9 - cylinder below stall (Re = 2 i 10?), 10 - sphere
below stall (Re = 2 x 105); circles denote flat plate (old experiments).
- 22 -
-------
As is evident from the figure, the results obtained are in mutual
agreement, but our measurements give an appreciable scatter of points
due to large values of the bands in the band-pass filter employed.
In order to study the influence of the degree of turbulence of the
forward flow in the presence of comparatively small changes in the ratio
of the obstacle size to the characteristic turbulence scale, we studied
a broad category of bodies including streamlined (wings, airplanes) and
bluff ones (sphere, cylinder, plate, cube, etc.)* These experiments
(Fig. 3) showed the presence of a stall of aerodynamic characteristics with
respect to the £ number, somewhat analogous to the stall with respect to
the Re number. They also showed that for a number of bodies, simulation
with Re number may be replaced by simulation with the e number. For example,
at low Re numbers and high c values one can obtain drag values for a sphere,
cylinder, etc. corresponding to large (above-stall) Re numbers and small
e values. This fact substantially extends the possibilities of study in
tunnels where it is difficult to achieve substantial Re numbers.
The influence of turbulence on aerodynamic characteristics depends on
the shape of the body and the Re number. However, the general tendency
lies in the fact that the drag of bodies in a flowing medium with sharp
stalling (for example, a flat plate placed perpendicular to the flow or a
model of a house with a pointed roof) is not sensitive to the initial flow
turbulence.
Study of Flow Around Land Obstacles - Results of Simulation
The A-6 MGU wind tunnel has an open test section of elliptical cross
section measuring 4.0 x 2.4 m with a length of 4.0 m. The parameters of
undistorted forward flow in the test section of the tunnel are characterized
by the following quantities: the nonuniformity of the velocity does not
exceed -0.5%, the sidewashes in the horizontal and vertical planes are equal
to -0.3, and the critical Re number for the sphere Dc = 150 mm. Re„,_ = 3.7 x
r b C1
10 (from hot-wire anemometer measurements, z = 0.1-0.15%). Models of ground
installations were mounted on a flat screen simulating the underlying surface.
Different turbulence and velocity conditions in the flow incident on the
model were produced by the indicated method using nets mounted on the edge
of the nozzle. Measurements of the velocities and sidewashes in the vicinity
of the models were made with a six-wave TsAGI probe, and those of the turbu-
lence intensity, with a hot-wire anemometer.
Model of an isolated (two-dimensional) hill. The model studied had a cross
section in the shape of an isosceles trapezoid with the following dimensions:
length of base, 1500 mm; height, 100 mm; length of upper side, 500 mm; span,
2000 mm. The model was made of wood and polished, which corresponds to a
height of roughness prominences of approximately 10-15 pm. A diagram of the
- 23 -
-------
placement of the model in the test section is given in Fig. 4.
When a weakly turbulent uniform flow (e0 = 0.3%) flows around the model,
the velocity change in the vicinity of the model follows the configuration
of the hill. On the forward slope, an acceleration of the flow reaching up
to 10% of the velocity of the forward flow on the upper plane is observed.
In the rear part of the hill, because of the diffusor effect, a deceleration
is observed reaching' 20-30% near the "ground". The zone of decelerated flow
extends over approximately one "height" of the hill. A change in the range
of 1 x 106-4.1 x 10^ in Re numbers calculated from the length of the base of
the hill had practically no effect on the nature of the velocity distribution.
Studies of the velocity distribution in the immediate proximity of the
hill surface were made by means of a special rack of 30 Pitot tubes which
made it possible to obtain the velocity profile starting at 8 mm from the
surface. Results of experiments at Re numbers of 2.1 x 10^ and 3.15 x 10*>
were in mutual agreement and showed that, as was to be expected, near the
surface the deceleration of the velocity intensifies, reaching 15-25% on
the forward slope, up to 30% on the upper surface, and up to 50% of the
velocity of the forward flow in the afterbody. The regions of strong decel-
eration extend to a height equal to 0.5 of the height of the model (in the
afterbody) and up to 0.2 on the forward slope of the hill.
The vertical mean velocity profiles near the surface (Fig. 5) differ
substantially in the zones of low deceleration and acceleration from the
velocity profile for the case of a turbulent boundary layer (according to
the "1/7" law), shown in this figure (z being the height and 6 the height
of the boundary layer). The deviations are particularly appreciable in the
Fig. Diagram of arrangement of coiners, nets and model of isolated hill
in A-6 wind tunnel, and dimensions of cross sections in which the flow
parameters were measured.
- 24 -
-------
lower half of the boundary layer; in the upper part, this difference does
not exceed 10% of the velocity of the forward flow.
The direction of the flow as a whole follows the configuration of the
hill. In the forward (confusor part), the sidewash is positive (up to
SjjlO0); in the diffusor part, negative (up to 8°), and equalized very
rapidly.
The change of the initial turbulence is monotonic in character and
barely noticeable. In the region of acceleration, a certain decrease of E
is observed, and in the diffusor region, there is an increase. Here, as in
the case of velocity distribution, no dependence on the Reynolds number is
observed. Results of experiments using the same hill model but with an
appreciable turbulence of the forward flow (up to e«8-9%) and different Re
numbers (up to Re = 2.1 x 10®) are shown in Fig. 6.
Fig. 5. Distribution of wind
velocities in different cross
sections above the hill.
1 - norm, 2 - behind the hill,
3 - forward slope, 4 - upper sur-
face , 5 --in J"ront of thejnodal,
As is evident from the figures, in the case
of large initial turbulence, the disturbances
introduced by the model extend over a smaller
region than for small Eo values. The influence
of Re number is also very slight in this case.
This may be explained chiefly by the fact that
the dimensions of the model in the direction
normal to the flow are small compared with its
longitudinal dimensions. An analogous picture
obtains in the case of flow around thin wings
under small angles of attack. For example, as
Re numbers change from 1.8 x 10^ to 1.65 x 10^
and e simultaneously changes from 0.3 to 3%, the
value of the lift coefficient Cx of such wings
remains unchanged, attesting to a virtual con-
stancy of the distribution of circulation, and,
in particular, of the velocity above the wing.
In our case, local velocity changes different
from zero are "reinforced" by the screen and
manifest themselves perceptibly only in regions
along the normal that are bounded by the "height"
of the model. Let us note in conclusion to this
part of the experiments that studies under natural
conditions also indicate the smallness of the dis-
turbances introduced by gentle hills into the air
flow above them.
- 25 -
-------
-------
0 5 101%
T 1 1
tunnel axis SOO
e<,
600
to
-j
eC'8,3%
WO
300
200
150
100
50
3C
WO
SCO
1200
1600
2000
2400 X mm
Fig. 6B. Distribution of longitudinal component of turbulence above the hill model at different velocities
of forward Ilo».
-------
The velocity field of the longitudinal component of turbulence was
also studied above a three-dimensional hill. The model consisted of a
truncated elliptical cone and had as its base an ellipse with axes of 1340
and 1080 mm, and in the upper part, an ellipse with axes of 880 and 660 mm;
the height of the hill model was 60 mm. The model was made of wood and
polished.
Fig. 7 shows the distribution of the degree of turbulence along the
middle of a three dimensional hill for Go values of 0.2 and 4%. As is
evident from the figure, at a height of 0.3H, the influence of the initial
turbulence in front of the obstacle is clearly visible. It is interesting
to note that behind the hill, the maximum degree of turbulence is the same
0,3 H
*00 800
1S00 x mm
Fig. 7. Distribution of the degree of turbulence at heights
of 0.3 H and 0.8 H for Eo= fi* l) and E0 = °«2# 2) and of
the normal surface pressure for E0 = 3) and £. = 0.2
4) along the middle of a three dimensional hill.
in both cases. Comparison of the distribution of the velocity and degree of
turbulence shows that the velocity minima correspond to a maximum of the
degree of turbulence, and the velocity maxima, to a minimum of the degree of
turbulence. Comparison of the zones of velocity field and degree of turbu-
lence above the three-dimensional and the two-dimensional hill and also above
a wedge-shaped protrusion shows that, as expected, the disturbances introduced
by the three-dimensional hill are more monotonic in character, and the distor-
tion zone is somewhat smaller.
Models of urban building systems. The study was made on models of two
types of urban building systems: "flat" houses of a few stories (in the old
blocks of Leningrad) and modern buildings (multistoried, with large gaps
between them) (Fig. 8). The experiments involved the study of the influence
of the models on the character of the flow above them as a function of the
turbulence of the forward flow; the velocities within city blocks with modern
building systems were found as a function of the wind direction and turbulence
of the forward flow.
- 28 -
-------
In the flow around the model of the system of "flat" buildings of a
few stories, the deceleration of velocity and change of the turbulence
intensity were found to be insignificant. The disturbance includes a
relatively small region above the model, below a straight line with an
inclination of 4°, with the origin of coordinates at a distance of 0.25 of
the length of the model from its forward part. In the flow around the
model of the modern building system, the deceleration of the velocity is
more substantial, and the region of increased turbulence is located above
the entire model.
Fig. 8. Model of urban building system (second type).
At an increased turbulence of the forward flow (e = 10%) , the turbu-
lence intensity above the second model increases sharply, reaching 80-100%,
and the approximately rectilinear boundary separating the disturbed from
the undisturbed region has a large inclination (up to 6-7°). As was shown
by the analysis, the angle of inclination of the boundary depends on the
relationship between the mean height of the irregularities of the earth's
surface and the turbulence scale L , which is related to the turbulence
£ 7
intensity of the incoming flow.
Fig, 9« Dependence of angle of inclin-
ation 9f boundary of turbulent disturbance
on ratio of surface roughness to turbulence
scale.
- 29 -
-------
As the turbulence intensity of the forward flow or more exactly of
the ratio hQ/Lt increases, the angle of inclination of the boundary has a
tendency to stabilize (Fig. 9). This fact may turn out to be very useful
in the solution of certain problems related to the diffusion and determina-
tion of regions of disturbances above different ground obstacles.
Fag. 10. Distribution of relative velocities in flow around trapezoidal depression.
Studies of the velocity distribution inside blocks with modern buildings
showed the presence of zones with an increased velocity (with "straight-through
blowing") and also stagnant zones and zones with a velocity directed against
the wind direction. However, this information on the "nonaerodynamicity" of
the urban building system studied is not the most important part. More impor-
tant is the fact that the turbulence intensity of the forward flow (wind)
substantially affects the velocities within the blocks. Special experiments
showed that in this sense, for a part of the block, an increase of turbulence
(up to £q = 5%) is equivalent in its effect to the erection in front of the
"city" of a continuous wall (barrier) of a height equal to the mean double
height of the houses in the block. These facts make it necessary to use par-
ticular caution in setting up and conducting experiments in wind tunnels with
models and obstacles of complex shape and also in evaluating the results of
known experiments. Particular attention should be given to the measurement of
the initial turbulence of the incoming flow, to which the experimental results
are referred.
Results of the study of a current above a depression are shown in Figs.
10 and 11, on condition that in the incoming flow e0 =0.2. As is evident
from an analysis of these figures, velocity and particularly turbulence dis-
turbances are particularly sharp in character, owing to turbulent flows within
the depression itself. However, at a distance (along the vertical) of only
about 3-4 "depths" of the depression, the flow becomes practically normalized.
On the whole, wind tunnel studies of flow around models of a hill, urban
building systems and other ground obstacles yielded data on the disturbance
- 30 -
-------
3001-
200
m
cut
epnaa « cipyKiypa
B03,ayiuiioro noTOKa nafl. neoflHopofliioii rcoflCTHJiaiomeii noBcpxHOCTbio. (Cm
HacT cSopniiK )
2. Top/iKEi C. M B/iKHinrR Ka^a/ibiioii TypGy/icHTHocTJi na oGTCKaiuie pa3.ni[Hiibi<<
Te/i TpeTjiii BcccoiODiitiii cbesji no rccpcTimccKoii u npnK.na.anon jicxaiuiKc
M, 1968
3 rop^HH C M , 3 p a » e b c k ii il H, M Hay^efiiie o6renaHHH MOiie;icfi pcflt>c(]ia
H ropojicKoj'i 33CTpoiiKu a aspottKHasuiHecKoit tpySe, Tpyiu rrO, awn, 234,
1968.
4. Millikan C B, Klein A. L. The effect turbulence Aircraft Eng. 169
(August 1933)
5. 7 a y I o r G t Statistical theory of turbulence Part 5 Proc Roy. Soc. Lon-
don A 151. 307-317, 1936
6. Sandborn V. A, Marshall R. D Local isotropy in wind tunnel turbu-
lence. Research Memorandum No 1. Fluid Dynamics and Diffusion Labora-
tory, Colorado State University, 1965.
- 31 -
-------
S. M. CORLIN, i. M ZRA/EVSKY, S. P. Z/BOROVA
THE INVESTIGATION OF INFLUENCE Or GROUND SURFACE ROUGHNESS
ON AIR FLOW CHARACTERISTICS IN WIND TUNNEL*
The investigation of concentration field and air currents in the
complex relief conditions can be fulfilled with the help of laboratory
modelling and also on the basis of laboratory experiments and nume-
rical investigations.
The Institute of Mechanics of Moscow State University and the
Main Geophysical Observatory carry out the works on simulation of
flow pattern over topographical features of the earth in wind tunnel.
Substantial attention is paid to the questions connected with the ef-
fect of Reynolds number and air flow turbulence intensity on current
pattern. The experiments are carried out in some wind tunnels under
different upstream conditions (velocity and turbulization). The com-
parison with data obtained in natural conditions is carried out
In the case, when characteristic turbulence scale is of order of
the size of studied obstacle, the effect of upstream turbulence
intensity is substantial. For all this, the character of the effect de-
pends on the form of the body and considerably influences current
pattern over complex obstacles. As a rule, the increase of intensity
of turbulence leads to more monotonous velocity distribution and to
the increase of turbulized zone.
The change of Reynolds number does not influence practically
speaking on the whole current structure in rather large range, except
the thin layer directly adjacent ot obstacle.
As a result of these investigations, velocity field and intensity
of longitudinal turbulence component over flat hills with different
slope and shape, hollows, three-dimensional hill, two types of resi-
dential buildings and some other obstacles were studied The change
of turbulence intensity in the region of hill model with a slope up
to 16° is not very large.
A decrease of turbulence intensity is observed in the regions of
flow convergence, i. e. in the front part of the hill, and more substan-
tial increase is found in its back part. Velocity braking behind the hill
with a slope of 12° reaches 50% from upstream velocity and includes
the region equal to half the height of the model.
Disturbances brought into the flow by the hollow have especially
sharp character, but the current is practically normalized already at
the height equal to 3 - 4 hollow "depths". Turbulence intensity over
residential building model increases sharply. Eddy zone has rectili-
near boundary with large slope up to 6—7 . The angle of slope of this
boundary depends on ratio of mean building height and characteris-
tic turbulence scale.
The obtained results and their comparison with observational data
in natural conditions show that they are in good agreement.
* Editor's notes Ihe abstract is presented as given in English with the original Russian article.
- 32 -
-------
USE OF PARAMETERS OF EULERIAN TURBULENCE
FOR ESTIMATES OF LAGRANGIAN CHARACTERISTICS
V. I. Ivanov (USSR)
From Clavnoe Upravlenie GidrometeoroLosicheskoy Sluihby Pri Sovete Ministrov SSSR. (Chief Administration of
the Hydrometeorological Service Under the Council of Ministers of the USSR.)^ "Meteorologisheskie Aspekty
Zagryazneniya Atroosfery". (Meteorological Aspects of Air Pollution.) Sbornik dokladov na roezhdunarodnom
sinpoziume v Leningrade - Iyul' 1968 g. (Reports delivered at the International Symposium in Leningrad -
July 1968.) Pod redaktsiey d-ra fiz.-mat. naift K. E. Berlyanda. (Edited by Prof. M. E. Berlyand.)
Gidrometeorplgicheskoe izdatel'stvo, Leningrad, p.162-173, (l97l). (Hydrometeorological Publishing House.
Leningrad, (l97l).)
It is well known that the process of turbulent diffusion is determined
by the turbulent characteristics of the velocity field. We will discuss
the principal methods of determination of diffusion parameters in the lower
atmospheric layer from the standpoint of the use of turbulent characteristics
in such calculations. Major attention will be given to the possibility of
using for this purpose turbulent characteristics that at the present time
can be determined relatively simply by experimental means.
Two different approaches to the solution of turbulent diffusion problems
may be indicated at the present time.
1. Use of a semiempirical equation. To calculate the concentrations
of a passive impurity it is necessary to know in this case such parameters
of the lower atmospheric layer as the coefficients of turbulent diffusion
k , k and k , and also the wind velocity in the layer under consideration,
x y z
Of the above-enumerated coefficients of turbulent diffusion, the most acces-
sible (from the standpoint of the possibilities of its practical determina-
tion) is k . For a weightless passive impurity, a measure of k may be the
z z
turbulent viscosity coefficient kz, defined by the known ratio
(1)
~sr
sffi
where T is the tangential frictional stress and J__ is the mean wind
C &i
velocity gradient.
I flu
While the determination of is comparatively simple in the surface
_dz
layer, its evaluation at heights greater than 50-100 m involves a large error.
This is explained first of all by the difficulty of measuring small values of' &u'
' <32
- 33 -
-------
As an illustration, we present data on the magnitude of mean wind velocity
gradients obtained from results of two-year observations on the Obninsk
meteorological mast of the Institute of Experimental Meteorology [11]
(Table 1).
Table 1
Mean Vertical Wind Gradients According to Periods and Seasons
( m/sec per 100 m)
dz
Layer, m '
Time, hours
I
7
13
19
I
7
13
19
Winter
Spring
0-100
6,2
6,0
5,4
5.9
6,4
5,2
5,6
6,0
100-200
1.0
1,7
1.3
1.1
1.5
1.2
0,5
1.3
200-300
0.0
0,2
0,4
0,6
-0,1
0,0
0,5
0.5
Fall
Summer
0-100
6,3
5,5
5.5
5.9
5,5
4,0
5,2
5.1
100-200
1,5
1,9
0,9
1.8
1.5
1,4
0,6
1.0
200-300
0,6
1,2
0,5
0.6
0.8
0,5
0,5
0,5
It follows from these data that the errors in the determination of
dz
and hence, k for heights greater than 100 m may reach several hundred per-
cent with the existing accuracy of the measurements of •
i dz
It should be noted also that the determination of the tangential fric-
tion t makes it necessary to use measuring devices that permit simultaneous
measurement of two components, the longitudinal and the vertical components
of the wind velocity.
The turbulent diffusion coefficients k and k for a boundary layer can-
not be determined from a ratio analogous to (1). ^The procedure for determin-
ing these coefficients from the meteorological parameters of the lower atmos-
pheric layer has not yet been described in the literature. As a rule, they
are evaluated only from results of measurement of the turbulent diffusion
itself. In theoretical calculations, various relationships that in many cases
have no adequate physical justification are used in calculations of these
quantities. Thus, an urgent problem is to develop a method of their determ-
ination on the basis of characteristics of the lower atmospheric layer.
2. Use of statistical theory of turbulent diffusion. To evaluate the
distribution of an impurity concentration in the simplest case of a uniform
- 34 -
-------
isotropic turbulent velocity field, it is necessary to know the structural
function of the coordinates 02 as a function of time t and other parameters.
This structural function is expressed via the velocity correlation function
R^(t) in Lagrangian variables:
i
c2(0 = 2j(/-.)/?t(t)rfx. (2)
o
Assuming that the distribution of the impurity particles obeys the normal
law, by using a^ one can carry out the necessary calculations. Relation
(2) was derived for homogeneous and isotropic turbulence. However, approxi-
mate methods of calculation may be proposed that may be used for an inhomog-
eneous and nonisotropic velocity field [2].
Thus, the use of statistical theory presupposes the knowledge of the
velocity correlation function ^("O and its dependence on stratification and
other meteorological parameters. However, for the lower atmospheric layer,
the character of the correlation function has been practically unstudied
(with the exception of the relationships in the inertial interval). One can
indicate an approximate procedure used by the authors of [5], based on the
hypothesis of similarity of the Eulerian RE(x)( and Lagrangianj/?L(t) ivelocity
correlation functions:
^£(t) = ^(P")- (3)
Although the dependence of the functions /?£(t) and on the time
shift T is generally different, this assumption makes it possible to carry
out approximate calculations of a^ along different directions if the numerical
value of the parameter 8 is known. However, at the present time, the value
of 8 itself is known only with a rough approximation, which makes it impossible
to obtain reliable data on Rl(t). The difficulties of this approach are com-
plicated even further by the fact that this procedure is not based on any
physical hypotheses that would facilitate the finding of the dependence of B
on meteorological and turbulence parameters.
In addition to the methods enumerated above, mixed ones are also used,
in which statistical theory and semiempirical equations are used simultane-
ously [9]. Thus, to describe diffusion in the vertical direction, a semi-
empirical diffusion equation is used, and in other directions the distribution
is given by using Lagrangian representations. From the point of view under
consideration, this method contains no new turbulence parameters other than
those discussed above.
In this connection, theoretical and experimental studies were carried out
at the Institute of Experimental Meteorology for the purpose of working out a
method of computation [1, 2]. The latter is based on the use of similarity
theory, the principles of which are given in the works of Batchelor [14] and
- 35 -
-------
Obukhov and Mania [11].
We will begin the discussion with estimates of the turbulent diffusion
coefficients. As we know, the_ turbulent diffusion coefficient k may be de-
fined by the expression . 1 daz
2 dt '
Introducing the integral time scale of the Lagrangian velocity cor-
relation function; RL(x),, and examining this quantity for each of the velocity
components {u,} = {u, v~w), ,we obtain an estimate of the turbulent diffusion
coefficients along""different directions. For a sufficiently long time T,
{^i} — {&xt kkz} may be represented in the form
where ^2" is the turbulent energy of the i-th velocity component.
„ t \
It should be noted that expression (4) does not require the knowledge
of wind velocity gradients or other mean values. However, expression (4)
contains the time scale ~ , which in contrast to ^ requires a determination
from the standpoint of the possibilities of its measurement in the experiment.
To evaluate this quantity, one can make use of dimensional theory considera-
tions. From the standpoint of its physical meaning,Represents the character-
istic lifetime of energy-carrying turbulent inhomcgenei'ties. If we know the
total turbulent energy of a velocity component.and also that part of it
which dissipates per unit time, e (i.e., the dissipation of turbulent energy),
the characteristic lifetime^] should be given by the relation
— \ v2.
\.=4~f- «>
where a is some numerical constant.
Like vz, the dissipation of turbulent energy £ is a quantity accessible
to measurements, and therefore relation (5) may be used to evaluate.rLt , from
experimental data of Eulerian turbulence. Thus, by taking (5) into account,
the turbulent diffusion coefficient k^ may be evaluated by means of the
relation
* /t »
(6)
In order to be able to use formula (6) for practical estimates of k^, it
is necessary to refine the value of the numerical constant a. To this end,
we will compare the value of the turbulent viscosity coefficient ^Skp given
by formula (1), with k^» given by formula (6). This may be done on the basis
of data on the best-studied part of the boundary layer - the atmospheric sur-
face layer.
- 36 -
-------
As we know, in a surface layer with indifferent stratification kz=w+z,
where is the dynamic velocity, andjv. | is von Karman's constant. Sub-
stituting into formula (6) the expression for the dissipation of turbulent
energy in the surface layer v* > and also expressing the turbulent
fc = , .
energy of the vertical component in terms of the dynamic velocity itus= v\ , [8],
we obtain
which is in complete agreement with the expression given above for kz in the
surface layer for the proportionality coefficient a = 1.
Because of a lack of experimental data for the atmospheric boundary layer,
we are unable to draw a similar comparison for k^ and k^, which would enable
z us to estimate the numerical value of a. However,
this may be done for the boundary layer on a flat
plate, a layer that is similar in physical proper-
ties and that has been studied in a wind tunnel
[10]. Figure 1 shows the ratio (for cases of
- k' '
0,<
0,2
—O i>
o
o
c
° c
<
o
2¥,
Fig. 1. Ratio.of turbulent dif-
fusion coefficient kz to turbu-
lent viscosity coefficient k^
for boundary layer on flat plate.
1) — =0,045, 2) -£- = 0,04.
v, »•
equal to 0.04 and 0.045, where u is the veloc-
'
ity of undisturbed flow).
As follows from the above estimates, the
equality of k^ and k^ makes it necessary that
the constant entering into formula (6) be approxi-
mately equal to 0.5-1.0 when averaged out over the
layer, this being in agreement with estimates of
this constant in the surface layer. As shown by
calculations of a based on other considerations,
this quantity is also close to unity [2, 6, 7].
Thus, for the vertical component, the numerical
coefficient a entering into expression (6) is
close to unity. It may be expected that the
numerical coefficients entering into the expres-
sion for the diffusion coefficients kx and k^ are
also close to unity.* With this assumption, ex-
pression (6) may be used for estimating the tur-
bulent mixing coefficients kx, ky and kz>
* Generally speaking, it is possible that this constant may be slightly different for the coefficients
k along different directions. However, this possibility will not be considered at this stage.
- 37 -
-------
The representation of v\ and £ as functions of stratification makes it
possible in turn to find the"dependence of kx> k^ and kz on the stratification.
Using relation (6), one can estimate the degree of anisotropy of turbu-
lent mixing along different directions in the lower atmosphere. We will use
for this purpose data on the mean turbulent energy whose values for the
lower layer of the atmosphere are known (in particular, for the surface layer).
The magnitude of dissipation of turbulent energy entering into (6) is
the same for the coefficients along all the directions. As follows from the
data given in [8], the ratio . j/^. j/"^ is approximately proportional
to 2.4:2:1. It follows that the turbulent diffusion coefficients of (6) have
to stand approximately in the ratio 30:15rl. This indicates a significant
anisotropy of mixing in the atmospheric boundary layer along different direc-
tions. An important fact is that kx and ky surpass kz by more than one order
of magnitude. This fact is in accord with other estimates.
We will now consider the application of similarity theory to the evalua-
tion of the structural function of the coordinates <7^. in Batchelor's papers,
such estimates are made for the inertial range for the relative diffusion of
a pair of particles [3]. If this theory is applied to a Lagrangian acceler-
ation field j (t), the acceleration spectrum S,(f) is independent of frequency
in this case: and the correlation function for the inertial range
will have the form [4]
(7)
v2
where t = , and c. is a numerical constant close to 0.7*.
C |B 1
n
Using (7) and (2), one can estimate the variance of the coordinates cr
also by using the parameters of Eulerian turbulence ^2 and E.
Let us consider certain experimental data on direct measurements of the
Legrangian velocity correlation function in the lower atmosphere. These
results were obtained in [15], where constant-level balloons (whose trajec-
tories were recorded by means of radar) were used to evaluate, The re-
sults obtained show that in the lower atmosphere at heights of 100-300 m,
the correlation function RL(x) is close to linear for time scales of 35 minutes
(Fig. 2).
However, from a practical standpoint it is more convenient to check,
not the expression itself for the velocity correlation function (7), but its
consequences.
* Instead of relation (?), an extrapolation formula for nay be proposed in the form of an exponent
[2 ^ Expression (7), ihich applies in the.inertial,range, will in this case be made up of the first terms of
the expansion of the exponential correlation function.
- 38 -
-------
Thus, by substituting expression (7) into (2), we find that the
structural function of the coordinates for one particle has the following
form:
2 / \ ~2 2 «i _ 3 /o\
oi(T)=Kt g- st to;
If the plume from a point source is treated as complex motion consist-
ing of the motion of its instantaneous center of gravity and the motion
relative to this center, then the structural function describing the change
of the instantaneous center of gravity will be
(9)
Since expressions (8) and (9) are a.consequence of the fact that the corre-
lation function RL(x)\thas the form (7) and is determined by the parameters
|^2, and £, their agreement with experimental results should indicate that
the' form of the correlation function and of the parameters determining it is
correct.
A check of relations (8) and (9) was made at the Institute of Experi-
mental Meteorology. The results obtained are given in [4]. Transverse dif-
fusion from a point source located at a h_eight of 100 m was studied. Simul-
taneously, the turbulent characteristics and e were measured with the aid
of the measuring system of a high-level meteorological mast. Comparison of
(j2 values calculated by using the parameters of the turbulent velocity field
^I'and e and also measured directly in observations of smoke plumes indicates
- 39 -
-------
satisfactory agreement to within a time tl of the order of 500-800 sec
(spatial scale of 1500-2000 m).
A check of relation (9), made in similar fashion, showed that there is
also satisfactory agreement between the theoretical estimates and the experi-
mental data. It should be noted that because of the difficulty of selecting
stationary conditions in conducting the experiment, a somewhat greater scatter
is observed in this case.
To illustrate the degree of agreement between the two methods of esti-
mating 0^, we reproduce the data on and £ given in [4], obtained from
direct measurements and by estimating-these parameters from a smoke plume in
the inertial interval (Table 2).
Table 2
Values of and ' Obtained from Direct Measurements
with £ and l/^j2 Obtained by Evaluation from a Smoke Plume.
6 g,
Experi-
ment
ecra^/sec'
om^/seo'
V^2 m/seo
~Vv'g m/sec
1
0,65
1,0
0,18
0,22
2
13,0
20,0
0,66
0,64
3
2.2
4,0
0,21
0,26
4
0,4
0,5
0,23
0,25
5
4,8
5,8
0,39
0,31
Thus the proposed method may be considered to give a satisfactory
agreement with results of experimental measurements and shows that it may be
used as the basis for a method of calculating the parameters of turbulent dif-
fusion. Its application to the case of inhomogeneous velocity fields naturally
contains certain errors, but, as shown by the data of Table 1, they are insig-
nificant for a height of 100 m in relation to transverse diffusion.
Let us find the relationship between the method of calculation discussed
in [5] in the approximation of similarity of the Lagrangian and Eulerian cor-
relation functions (cf. formula (3)) and the above-discussed results of simi-
larity theory. Such a relationship may be established fairly simply for the
range of turbulent inhomogeneities best studied to date, i.e., the inertial
range. Let us consider for this range the relationship between the Lagrangian
and Eulerian time scales.
As we know, the Eulerian velocity correlation function in the inertial
range has the form
= do)
where u is the mean velocity and c is a numerical constant (practically equal
- 40 -
-------
to unity for the longitudinal correlation function). We take the quantity
from (5) introduced earlier and use it to express /?£(t)^ from (10):
= (ID
where
r=r \ —
'-+m
and c^ is a numerical constant entering into expression (7).
The expression for the coefficient A entering into (11) shows that
A 1 (since1 Vl? ) • This indicates a slower decrease of the Lagrangian
u 1
correlation function to zero as compared with the Eulerian function.
We find the Eulerian 1 to£ ^ and Lagrangian1^ turbulence scales by defining
them as the time interval corresponding to a twofold decrease of the normal-
ized correlation function:
Ago^o.S; M;°f) =0,5. (12)
It is assumed that in such a decrease of the correlation function, we do
not overstep the boundaries of the inertial range in either case.
Use of conditions (12) together with expressions (11) and (7) (using
numerical values of the constants c^ = 0.7 and c = 1) leads to the following
time scales:
xog 0.25^ * - ; (I3)
II
*0L— 0,5-L-
Thus, B, which represents the ratio of the time scales, will be equal to
Expression (15) indicates a possible dependence of 6 on the stratifica-
tion: in the presence of unstable stratification, it should be less than in
the presence of indifferent stratification or inversion.
Let us consider some operational aspects connected with the experimental
determination of the turbulent energy and its dissipation e in the lower
atmosphere. The procedure for determining the turbulent energy of the wind
velocity components is known. In the surface layer, its evaluation may be
- 41 -
-------
made by using the similarity theory [13], for which it suffices to know
the mean profiles of wind velocity and temperature. The characteristic
feature of its measurement at heights of 50-100 m lies in the fact that it
may be measured with satisfactory precision by means of a fairly coarse
gauge, for example, one having a time constant of 2-5 sec (excluding stable
stratification). The Obninsk, meteorological mast uses propeller gauges
(1-sec lag) and also bivanes (1-4-sec lag) for this purpose. Estimates shew
that the error in the determination of turbulent energy due to limitation of
the transmission band by the lag of the gauges does not exceed 5%.
Let us now consider certain questions connected with the determination
of the dissipation of turbulent energy E. Most widely used for the determ-
ination of £ at the present time are the regularities of the inertial range,
where the structural function of the velocity field D(t) has the form
D (t) = [u (t) ~ a (t + -)]2 = 2cz'> (16)
Here 2c is a numerical constant equal to 1.95 for the longitudinal component.
The spectral density is described by the expression
S(w)= (17)
where crs0,5c.
Formulas (16) and (17) are equivalent from the standpoint of the possi-
bilities of determination of £. However, the determination of £ from structural
function (16) has certain advantages. First, this is due to a smaller quanti-
ty of arithmetic operations, since the determination of S(w) makes it necessary
also to carry out a transformation of D(t) or to filter u(t), which is a very
laborious operation. Secondly, the following fact should be pointed out. As a
rule, in measurements of the fluctuation components of the wind velocity, it
is necessary to use gauges having a certain lag. If the impulsive admittance
function h(t) of the measuring system is in the exponential form —
h(t) = e\ rac t0
where T0 is the time constant, then instead of expressions (16) and (17) for
u(t) measured with the aid of these gauges, other relations will hold:
(a) for the structural function
D (^) =: 2czh rt7' ("Vl -- *o' ¦ 0,9); (16a)
(b) for the spectral density
S(«0 = C¦ (17a)
The expression for the structural function is approximate and applies
starting with time T > 3T0.
- 42 -
-------
Let us note that if in the determination of spectral density (17a) the
distortions due to lag of the gauge are different at different frequencies,
then for the structural function the influence of the lag amounts only to
the appearance of the constant additional ternij ~0,9t^, jwhich is independent
of the shift T. It is easy to show that the dissipation of turbulent energy e
calculated from expression (16a) will be determined by a relation that is in-
dependent of the time constant x0:
where AD(t) is the difference in values of the structural function at times
and , lying in the inertial range, and satisfying the condition r2, ti^3to
When the time constant of the gauge is known only approximately, and
also when the time constant itself is a function of the wind velocity (for
example, for propeller and cup gauge), and therefore the introduction of a
correction into the expression for S(w) is considerably complicated, the most
convenient method of determining e is to evaluate it from the structural
function (18). The method given above is now being used on the Obninsk high-
level meteorological mast. It makes it possible to evaluate c, not with
special quick-response gauges (hot-wire anemometer, acoustic anemometer), but
with commercial M-49 propeller gauges used by meteorological networks.
Let us examine some experimental data on turbulent characteristics,
obtained in the lower 300-meter atmospheric layer with the aid of the measur-
ing system of the high-level meteorological mast of the Institute of Experi-
mental Meteorology (Obninsk). The initial turbulent characteristic used for
estimates of diffusion parameters were values of the turbulent energy and
its dissipation e. We will give the experimental data on profiles of these
characteristics in the lower 300-meter layer (the measured values of u? per-
tain to the longitudinal component of the wind velocity u^). Each of the
profiles of T? and e shown in Fig. 3 was obtained by averaging over 5-7 sep-
arate series of observations, each lasting 30 minutes, pertaining to lower-
layer conditions similar in stratification. The Monin-Obukhov parameter was
used to estimate the stratification of the atmosphere [13]:
where g is the acceleration due to gravity, T is the mean temperature, c is
the heat capacity at constant pressure, p is the air density, Q is the tBrbu-
lent heat flow, and K is von Karman's constant.
The values given in Fig. 3 were determined as the mean value L for the
given group of profiles. As shown by the results obtained, a monotonic
(18)
- A3 -
-------
2 m
0 10 20 for L-"35
Fig. 3. Profiles of turbulent energy (a) and its
dissipation E (b) for different stratifications,
l) h = -155 m, 2) h = -24 ra, 5) L = + 35 m, 4) L = -6m.
dependence i? and £ on the stratification is observed: for a stable strat-
ification (L = +35), the dissipation and the turbulent energy are substantially
less than for an indifferent and unstable stratification. Attention is drawn
to the fact that the dimensional values of u^_and £ in the presence of a marked
instability (L = -6) and slight instability (L = -100) are close to each other.
This is because in the_presence of a strong instability, the mean wind velocity
is low, and the group L = -100 corresponds to a high wind velocity. Although
the causes of generation which make the chief contribution to the turbulent
energy and dissipation are different in these states, their absolute values
are close to each other. In dimensionless coordinates they should differ from
each other.
The values of u^ and £ in the presence of an instability occupying an
intermediate position (L = -24), turn out to be maximum in absolute value.
- 44 -
-------
This is due to the fact that the series studied are characterized by a
higher wind velocity (dynamic turbulence), and also an appreciable inflow
of heat.
zm
Fig. 4. Estimates of lagrangian time scale "—*"or different
stratifications. For notation see Fig.
The indicated experimental data on values of the longitudinal component
of turbulent energy u^ and £ were used to calculate the parameters of turbu-
lent diffusion in accordance with the relations given above. Figure 4 shows
data on an estimate of the Lagrangian characteristic time scale given by
formula (5) with constant c = 0.7. They show that the values of are similar
for different groups pertaining to a weak and strong instability. For heights
of 100-300 m, 5 min, which is in good agreement with results of direct
measurements shown in Fig. 2, and also in [4], In the presence of an inversion,
is slightly larger, which also agrees with the experimental estimates given
in the above-mentioned paper.
Let us consider one more quantity, which is determined by the Lagrangian
time and constitutes an estimate of the distance along the flow over which
the principal process of decay of energy-carrying eddies takes place. This
parameter is given by the relation }. = uxL. contrast to X^ depends
substantially on the stratification (Fig. 5). The estimates obtained show
that this distance changes considerably with a change in the degree of insta-
bility of the lower atmosphere: X = 1.5 km for L = -6 and X = 6-8 km for
L = -100.
We will estimate another important characteristic in calculations of turbu-
lent diffusion parameters, i.e., the coefficient of turbulent mixing, by using
formula (6) and a = 0.7. Results of calculations of m^/sec are given
in Fig. 6.
- 45 -
-------
Experimental data on kx values are not given in the literature, and
for this reason the numerical kx values obtained cannot be compared with
results of estimates by other methods. Use may be made only of the fact
that k : kz = 30 : 1 for an indifferent stratification. For kz we than
obtain Kzsil05 cm^/sec, which in order of magnitude agrees well with esti-
mates obtained by other methods for the lower atmosphere [14].
In conclusion, we will cite the estimate of still another parameter, 8,
calculated in accordance with formula (15) (Fig. 7). From the results cited
above it follows that $ should depend substantially on the stratification:
8^5 for an unstable stratification and 8?z2Q for a slight (close to indif-
ferent) instability. As was shown by an estimate of the numerical constants,
the values cited not only give a close enough description of the range of
scales which enters into the inertial range, but can also be used to carry
out the appropriate methodical treatment for estimates in the entire spectral
range.
The proposed method of approaching the evaluation of turbulent diffu-
sion parameters in the lower layer of the atmosphere is very promising.
It contains no additional constants other than those used in the theory of
turbulence and, in contrast to some other methods, permits a sufficiently
detailed consideration of the dependence of the diffusion parameter on
stratification. Although the method is approximate in principle, nevertheless
in view of the fact that the accuracy required for solving practical problems
is also relatively low, it is fully applicable to specific calculations.
Its development was started relatively recently and it is important to bring
it to the attention of researchers.
200-
6000 A. m'
Fig. 5. Estimate of Lagrangian space scale
for different stratifications.
For notation see Fig. 5.
0 200
Fig. 6. Estimate of turbulent diffusion coef-
ficient kx for different stratifications.
For notation see Fig. 3.
- 46 -
-------
Further development of this method should include an experimental
determination of the correlation functions in Lagrangian variables for
large shifts, and also a study of their dependence on stratification.
Another important objective is to refine the constants entering into the
above relations, and to allow for the influence of anisotropy and inhomo-
geneity of the lower atmosphere on the statistical characteristics. The
development of this method will undoubtedly result in an accumulation of
experimental data and will promote the creation of a theory of inhomogeneous
and nonisotropic turbulence in Lagrange variables.
Fig. 7. Estimate of the parameter ; 3— '0L
zOE
For notation, see Fig. 3.
The creation of such a theory will promote the solution of a large number of
problems currently facing researchers in turbulent diffusion.
LITERATURE CITED
1. EusoDa H JI. MeTOA oueiiMi aiMoc^cpiioi'i antjx^ysiiii no T)p6y/ienTHbi,\t xa-
paKTCpiiCTiiKaM Cm. tiacTOsumn't c6cpnnK
2 E bi 3 o b a I I Jl , H d a 11 0 d B H OuciiKa napamcTpoD noncpeniiofi Aiic{>(t>y3iiii
b HiDKiiCM cjioc aTMoccpu no TypGy/iciiTHWM xapaKTcpucTiiKaM Msb AH
CCCP, KCBbi\ xapan-
TCpiiCTMKax TypfiyjieimiocTii Msb AH CCCP, ccp. reoit3 , N° 10, 1963
5. H d a ii o b B. H Oueiina xapaKTcpiicTiiK Typ6yvieiiTiioro nepcMeuJiiBai'iiH b hiihv-
iicm c.ioe aTMOctJicpu H3D AH CCCP, ccp reo(j>n3, Ars 12, 1964
6. JI a m 11 n a 11 0 b c k h ft F. A. ATMOc^epiian Typ<5\.ienTiiocTb H/1.A1.,
1966
- 47 -
-------
7. /] a ii x tm nil JX /I u3iiKa norpnimnioro c-non atMOc^cpw riapoMeni3aaT.
Jl, 1961.
8. Ma t bcc a /I T. Ocitonu oGiucsi Meieopo.ionni rjinpoMeieoiiajiT, Jl, 19&5
9 MauiKODa T B XapaKTcpticTiikii MCTCopo.nonmocKOro po/KHMj un>tuieri>
300-wcTpoQoro c.icfi no jBj'x.iermiM iia6.naaeiiiinn ra Bbicornoii .uaiTepHaH Aii({)(J)y3iiri
h 3arpH3iiciine B03ay
-------
METHOD OF EVALUATING ATMOSPHERIC DIFFUSION
FROM TURBULENT CHARACTERISTICS
N. L. Byzova (USSR)
From Glavnoe Upravlenie Gidroraeteorologicheskoy Sluzhby Pri Sovete Ministrov SSSR. (Chief Administration of
the Hydrometeorological Service Under the Council of Ministers of the USSR.) "Meteorologisheskie Aspekty
Zagryazneniya Atmosfery". (Meteorological Aspects of Air Pollution.) Sbornik dokladov na raezhdunarodnoo
simpoziume v Lemngrede - Iyul' 1968 g. (Reports delivered at the International Symposium in Leningrad -
July 196s.) Pod redaktsiey d-ra fiz.-mat. nault 11. E. Berlyanda. (Edited by Prof. M. E. Berlyand.)
Gidrometeorolgichesltoe izdatel'stvo, Leningrad, p. 170.101 (1971). (Hydrometeorological Publishing House,
Leningrad, U971J.) ¦ '
1. General Principles
The diffusion of a passive impurity in the atmosphere is determined
by the mixing properties of the turbulent atmosphere and hence may be cal-
culated from atmospheric characteristics of turbulence. The statistical
method makes it possible to calculate the diffusion characteristics directly
from the turbulent characteristics, without using the diffusion equation.
At the same time, it makes it possible to introduce clarity into the problem
of magnitude of coefficients of turbulent diffusion. In particular, it
enables one to determine under what conditions they may be considered constant
and in what cases they change with the duration of diffusion, and, in the
presence of wind, with the distance from the source (for example, in the known
schemes of Sutton, Bosanquet and Pierson [17]). A few examples of such treat-
ment of certain problems are analyzed below.
Let us recall the basic initial assumptions of the method. We consider
a smoke plume from the stack of an industrial enterprise (in the case of
absence of thermal ascent) in the way we observe it (Fig. 1). As a rule,
this plume expands with the distance from the stack, and its axis continually
changes its position in space. An instantaneous plume may be treated as a
collection of an infinite number of individual puffs successively emitted by
a stationary source and carried by the mean wind, and a statistical model of
a stationary plume is obtained by superimposing an infinite number of such
instantaneous plumes (a similar scheme was first discussed by Gifford [4] and
later by Ivanov and Stratonovich [7]). The motion of individual puffs con-
sists of the dispersion of the puff relative to its instantaneous center (o,)
and random wandering of the centers (c^) , causing the sinuosity of the visible
plume. This pattern is observed along the vertical direction as well as the
horizontal direction perpendicular to the wind.
Let the dimensions of an instantaneous puff be 6, and the dimension of
the stationary puff (plume) formed by the superposition of instantaneous ones,
2a^. Both the puff and the plume consist of an infinitely large number of
particles whose coordinates will be denoted by yi(x), where T is the time and
- 49 -
-------
-£ the particle number. The coordinate of their center of gravity at every
instant of time (average coordinate with respect to the number of particles)
will be denoted by y0(T), and the coordinate of the initial center of gravity,
by y0 = y0 (0). In this case, the variance of the coordinates of the particles
relative to the initial center of gravity
W*=<[y,M-y0]3>
represents the square of the half-width of the stationary plume; the relative
dispersion of two particles
W=<[yi(0-y/ (2)
We also determine the dispersion of the coordinates of the particles
relative to their instantaneous center of gravity
°^)---<[y,(-) -y0]a> (3)
and the dispersion of the center of gTavity coordinates
a5(")=([y0(")-yo]2)- (4)
The angle brackets signify averaging over the statistical ensemble of
the realization of the plume.
It is known [A, 7, 17) that the following relationship holds:
°i (x) — Go(") = 32 (") = . (5)
- 50 -
-------
which signifies that the dispersion of a puff relative to its instantaneous
center and the wandering of the centers are independent, and in sum, their
dispersions amount to the square of the half-width of the stationary plume.
The quantities entering into relation (5) are expressed via Lagrangian
velocity correlation functions of one particle and a pair of particles.
For one particle
RtV. (6)
the dispersion of one particle is determined from Taylor's formula
t ¦:
a*(7)
0 0
The velocity correlation function of a pair of particles amounts to
W- *"• ro) = • (8)
Here ik(t') is the Lagrangian velocity of the i-th particle at time t", and
r0 is the initial distance between the i-th and j-th particles.
The dispersion of an instantaneous puff is determined by the combination
of these two correlation functions
IT t 1
S3(-)=2f J]"rG)d-:d-". (9)
06 0 0
The concrete form of the Lagrangian velocity correlation function of one
particle is well-known in the inertial range, where this function depends
linearly on the correlation interval l/x~«xl=1 and is determined by two
quantities, the root-mean-square velocity fluctuation {y2)fand the Lagrangian
scale T . This function is frequently approximated by means of an exponential
function [15], which upon a series expansion for small r's gives a quali-
tatively correct dependence in the inertial region
(O ^ (O) = u •
(T=|t"—t'1). This leads to the expression
= + lj.
(10)
(11)
- 51 -
-------
In [10], using a generalization of the law of damping of equilibrium
fluctuations and the method of random forces, both formula (10) and an
expression for the correlation function of relative velocities of a pair
of particles were obtained; for t">t'
< K (*') - Vj (-/)! (-/'} - Vj (t")| > =
4 (v-) e H sh —J- D(r0)e *L . (12)
Here D(rQ) is the Eulerian spatial structural function of velocities,
where rQ is the initial distance between the particles. Hence, using the
relation
<[WiCO-MO] =
= 2 <«,(*') vt (-.")> - 2 (yl (x')ii;(-")>¦ (13)
one can readily obtain the correlation function of two particles. For the
inertial range, the structural function of two particles ([y, (t') was
obtained in [15] in implicit form. The paper [10] also gives an expression
for the correlation function of the coordinates of a pair of particles in
the case where this function no longer depends on the initial distance between
the particles:
2* «
T=2<®»>^
(u)
whence correspondingly
o?(,)=,£]-
-------
According to [2, 8, 10]
_ Cf2)
'£ — '
where £ is the dissipation of energy and c^ is a dimensionless constant.
Thus, having the measurement data for {v2) and £, one can calculate both
k and T^, and also, if the mean drift velocity (wind velocity) is known,
the scale along the wind direction 'a2}[/ The formulas cited make it
xL = .
Ci e
possible to evaluate this range of distances and the time during which the k
values should depend on the time (or distances) and where they may be assumed
constant:
for | or ,x - = -jf,
fori t oo
Fig. 2. Isolines k^ (z, x) (a) and kj (z< x) (b) and typical
profiles of and '*2 at distances of 0.2 ltin (l) and It km 12).
Considering the wind velocity profile in the surface layer to be loga-
rithmic (indifferent stratification), one can calculate the dependence of kQ
and Tl on height [8]. Figure 2 shows isolines of k for vertical diffusion,
calculated for the atmospheric surface layer to a height of 100 m. Above that,
- 53 -
-------
the layer is considered homogeneous. Solid lines indicate k^ (for a sta-
tionary plume), and dashed lines, - for an instantaneous puff. In the
right corner are shown the profiles of k^ and k^ at distances of 0.2 km (1)
and 4 km (2) from the source. At the 4 km distance, the usual profile of k
(with a discontinuity) is observed, but at a close distance it differs
markedly from the limiting one. Saturation takes place faster near the earth
than in the upper layers, since x increases with height.
The magnitude of xT in the layer above 100 m in indifferent stratifica-
tion amounts to 500-1000 m for vertical and 2000-5000 m for transverse dif-
fusion. That is why the change of the vertical diffusion coefficient with
distance is usually neglected, something that cannot be done with the hori-
zontal diffusion coefficient.
A similar method of calculating the dispersion of particles by turbulent
fluctuations [method of Khey (Hay) and Pasquill (14, 17)] is widely employed
in American and British studies. They use the spectral approach, which is
based on some assumptions concerning the relationships between the Eulerian
and Lagrangian spectra of wind velocity fluctuations. Although these assump-
tions are not rigorous, they yield satisfactory results in practical calcula-
tions. The key quantity in this approach is the ratio of the Lagrangian to
the Eulerian time scale
t,U
P = ^£ = -^—. (17)
E
which depends on the character of turbulence, i.e., on stratification and
other conditions, and requires further experimental studies.
Let us now consider certain special problems that are solved by applying
the above method. Let us note at once that these solutions are approximate
mainly because of an inadequate knowledge of the properties of the velocity
correlation functions. In this paper, we present solutions that were obtained
on the basis of reduced expressions for these functions, and therefore good
quantitative agreement between the theoretical calculations and the experi-
mental data cannot be expected in all cases. However, it will be possible to
refine these solutions after more complete information is obtained on Lagrangian
velocity correlation functions under conditions of real atmospheric turbulence,
and such experimental studies are being extensively conducted at the present
time.
2. Diffusion from a Point Source of Finite Time of Action
As our first problem we will consider the calculation of the dispersion
of impurity particles from a source with a finite time of action, or, what
amounts to the same thing, the dispersion of impurity particles from a station-
ary plume, but with a finite' measurement interval. This problem is frequently
- 54 -
-------
of practical importance, since for a normal particle distribution in the
plume, other things being equal, the axial, i.e., maximum possible concen-
trations are inversely proportional to the horizontal and vertical dimensions
of the plume. Obviously, if the diffusion along the wind direction is neglec-
ted, the total dispersion of the "shortened" plume (i.e., formed during a
finite time) is determined by the superposition of instantaneous plumes during
this time interval regardless of whether this interval is the time of action
of the source or the time of observation.
The scheme of the calculation*, based on the above considerations, is
shown in Fig. 3. In the latter, y (t, t) is the coordinate of a particle
that has left the source at time t and wandered during a time T. Having used
the expression for the dispersion of a random quantity calculated from a
Fig. 3. Diagram for the calculation
of the dispersion of particles from a
source of finite time of action.
statistical ensemble of shortened realizations of duration Tj for a fixed T,
we obtain
0
T
=ci (x) - JL {cr - ?)'/?, (€. *) (18)
0
where 0^(t) is determined from (7) and (11),
T
¦=)= W. *)>. (19)
The coordinate y (t,x) may be obtained by integrating the particle
velocity from the time of its escape from point (xq, y ) to time T determining
its distance to the source. Having denoted the velocity of such a particle by
U (xQ/t, s), where t signifies that at point xq the particle was located at
* Coincides with the scheme used in J iy
- 55 -
-------
time T, we have
whence
y({- x) = j u (*<)/'• s)ds-
0
x t
Ry (5. t) = J { dS' CIS"
(20)
0 0
The function under the integral sign is the velocity correlation function
of two particles, and at the instant the first of these passed through point
xq, the second was located at a distance rQ from it; under the assumption of
frozen turbulence, it may be considered that ro = t/£. , Having used expressions
(12) and (13) to find it, and carried out the integration, we obtain
(21)
where ^(f) is determined from (15). Substituting (21) into (18), we have
<4 (5) = (0 - °o (0 9 ft) = °2 ft) + °o ft) [1 — ? fa)l
(22)
D(r0)
2<"2)
>R{r9) ^being the normalized Eulerian spatial correlation function);
* ft) =-3-J ft-«)*('£. 5)«,
' o
(23)
r being the Eulerian spatial scale.
Ci
Fig. 4. The function
-------
Thus, according to (22), the shortened dispersion is equal to the dif-
ference between the dispersion of the stationary plume and the dispersion
of the centers of gravity, multiplied by a correction factor dependent on
T. By giving a different form to the function D (rQ), or, what amounts to
the same thing, R (rQ), one can obtain the concrete form of this factor.
In particular, for D (r) = C (era)'1* (inertial interval), fl(r) = e-r'r£1 and
R (f) ==jexp(—rlrE)tu 1 we obtain
, . r (cUT)'*>
= C «,*) ' (24)
?(v) = -jp-(i + e~*~ ])-
f (i) = -^r [(1 + V') e~T''' + ® ~ 1 '
where
a
Fig. 5. versus £ for different values
°f Cn).
The functions (ri) ¦+ 1, and long ones 4>(ri) -*¦ 0, we have
"P" T'—oo,
0r CO — °2 (x) "P" T 0-
- 57 -
-------
Let us note that when E, 1 (i.e., for short diffusion times near the
source), expression (22) takes the simple form
(25)
where
T
(v\) = (v2)~wUr-VK®d*
0
is the dispersion of the velocity fluctuations, calculated from the shortened
Eulerian realization of length T.
When 1 and 1, i.e., for small T and T, we have
°r ®= 2 3? ~'3 + 0,45W '2- (26>
If the second term in expression (26) predominates, the source may be con-
sidered quasi-permanent; in the opposite case, when o 675 f—-Cl "*"**
\ tc / T L 1
quasi-instantaneous. Figure 5 gives the dependence of on E, for different
values of (n).
Results of experiments conducted at the Institute of Experimental Meteor-
ology for different times of action of the source [11] show a fairly good
qualitative agreement with the results of calculations, but no quantitative
agreement was obtained. This appears to be due to the influence of large-
scale eddies of quasi-periodic structure which, according to several studies
[9], are observed in the atmosphere mainly with an unstable stratification.
The correlation functions used in the calculation do not take this into
account.
3. Dispersion of Heavy Particles in a Turbulent Atmosphere
In his study [16], M. I. Yudin formulated three mechanisms of the influ-
ence of gravity settling of particles on diffusion in a turbulent medium.
The first mechanism - displacement of the center of gravity of the particle
cloud - is considered by a suitable term in the semiempirical equation of turbu-
lent diffusion. The second mechanism involves the fact that because of their
size and inertia, the particles do not fully acquire the high frequency fluctu-
ations of the medium (this mechanism was examined in several studies [17, 12];
even in the diffusion of fairly heavy, but fine particles in a turbulent atmos-
phere, the effect of this mechanism may be neglected. The third mechamism is
associated with the fact that during its fall, a heavy particle intersects the
trajectories of the surrounding air particles, falling successively into the
- 58 -
-------
range of influence of different eddies. A calculation of the influence of
this effect was made for a continually acting source by Yudin [16] and for
a cloud of individual particles by Smith [13, 17]. The present study gives
a certain generalization of the results of [16].
In conformity with Yudin's ideas, the calculation will be made by
using expression (7), where the velocity correlation function
Rl (~)= (-^o> t 4" x)y (27)
should be taken along the mean trajectory of the motion of the scattering
particles. (Here v(xQ, t) is the fluctuation velocity of the particle at
time t, particle that at t = 0 was located at a point with coordinate xQ).
Let us consider two limiting cases. If the gravity settling rate w of
the particles is small compared with the fluctuation velocity in the direction
of diffusion v, so that during the correlation time the particle is unable to
excape to any appreciable extent from the effective range of these eddies
(which were acting on it at the instant of escape from the origin of coordin-
ates), the particle may be considered practically weightless, and the Lagrangian
time correlation function Rl (t) may be used in expression (27).
In the other limiting case, when w^>v , the particles escape along the
direction of action of the gravity force so rapidly that the loss of the
Lagrangian correlation will not be able to manifest itself. The reason for
loss of the correlation in this case will be the particle displacement in space,
and in expression (27) it is necessary to take the Eulerian space correlation
function
^E(r)~, (28)
where u (x, z, t) is the Eulerian particle velocity.
In the intermediate case, the velocity correlation function must be taken
with a certain approximation along the mean path of the incident particles:
*) = <•"(¦*<»¦ z> 2 + + (29)
where r = wt. Expression (29) constitutes a mixed space-time velocity correla-
tion function of a fluid particle. Its dependence on the time and coordinates
is known only in the inertial range along the coordinate axes:
R„(r, 0) = /?£(r) = [l - (-7j)V>]. (30)
*.<0, = (31)
where r^ and are the spatial Eulerian and temporal Lagrangian turbulence
scales. It is assumed that (u2) = (u2)~
- 59 -
-------
Taking Rw(r> 0) = 0 for r>rE> and 1^(0, t) = 0 for r>rL, Yudin con-
structed the majorant and minorant of the function I\j(r, x), which made it
possible to calculate the dependence of j ^az
kw = ~2' If
from w in two variants,
J-+00
with the true values of 1^ located between these variants. The relatively
inconvenient form of correlation functions (30) and (31) made it impossible
to obtain this dependence in the analytical form for any w; in the limit,
^ f°r w 00 (A is some constant determinable from the relationship of
W w
the Eulerian and Lagrangian turbulence scales).
In order to obtain more obvious results, it is expedient to set approxi-
mately
RE(r) = (v2)e~'lr'£.
(32)
(33)
Expression (33) is obviously a poor approximation of the "two-thirds law"
for small r's, but in view of the fact that it enters into the final result
under the integral sign, it is not very important. (To preserve the scale
relationship, it is necessary to set l,25r^)
It may be approximately assumed that
$u(r. *)=<
(34)
As can be readily ascertained, the expression
_ _L_ r
Qw(r. TL~~71
(35)
corresponds to the interpolation over the minorant used in [16].
Setting r = wt and integrating expression (7) with Rt values from (34)
or (35), we obtain
w
2 "I
where
?2 (s*)
x?=xi/?(Sv); v=
_L._1+„-*/
ZL V
- 60 -
-------
For variant (34), cp(sv) = "/l +sV.
For variant (35), } =J 4-s v-
Comparing (36) with (11), we see that the expression for differs from
w
the analogous expression in the case of w = 0 only in the fact that the
scale is replaced by
For the turbulent diffusion coefficient of heavy particles k^, we have
in this case for V -*¦ «:
4
__ 1 _J rE (17\
A0 T (-sv) st wiL '
which agrees with the limiting dependence, in conformity with (16) and (13).
Assuming, according to the frozenness hypothesis, that Fe^Uie, i where U
is the mean wind velocity, we have
1,25U *
Figure 6 gives the dependence of k /k on the parameter V. At the same
It? ^
time, it gives the scale v = - with the provision that the value of s is de-
termined only by the inertial interval (in this case sasl.8). The figure also
gives the curves taken from [16], obtained from the majorant and minorant;
the latter dependence nearly coincides with the dependence (37), with condi-
tion (35). Figure 6 shows that the influence of W on K begins to be manifested
in the range of 10% for sv = 0.3, and above 50% for V = 2.
K
0 0,5 ftO w in/sec
Fig. 6. Dependence of on U and sU .
- 61 -
-------
To check the relations obtained, it is necessary to carry out experi-
ments on the spreading of particles at different gravity settling rates from
a point source under the same meteorological conditions. In order to detect
the effect described, the precision of the experimental set-up should be
sufficiently high. There are almost no experimental results available that
satisfy these conditions and that could be compared with calculated data,
so that at the present time the comparison can only be qualitative. Figure 6
also shows results of calculations of the vertical turbulent diffusion coef-
ficient as a function of w, obtained in [6] on the basis of experiments
described in [5]. The scale of w was chosen so that the experimental points
fell as nearly as possible on the calculated curve. Knowing the mean wind
velocity, one can determine from the scale relationship the approximate
value of ft = —, which in this case is found to be approximately equal to 10.
LITERATURE CITED
1. 5bt30Ba II Si O Dw6ope KO3(j)(J)imiieHT0B AiHjx^yaiiit npii pctucmm no.iysvnn-
piiiccKoro )paBnei:iia a/ih Toienioro iicTOimiita Tpyflbi HFIT, gun -1, 1967.
2. Bbt30Ba H JI , HsaHOD B H OueiiKa napaMeTpoB nonepc>uiofi
b hidkhcm c.ioe aTMoc(j)epbi no Typ6y^enTJibiM xapaKTepHCTimaM. H3B AH
CCCP, ccp aTjioc(Ji ii OKeana, t III, Ms 5, 1967.
3 BwTiicp 3 K 3aBiici[MOCTb bc-ihiiiiiij fliicncpciiii MacTiiu, ncnyckac\ibix hg-
npcpuDHbi.M hctoiiikkom, ot KiiTcpDiiJia BpcMciui HaSjiiOiiciiHH. Tpyjiu rro.
Bbin 150, 1964
4. T h (J) (}) o p a . CTaTHCTiitecKaH Moae/ib flbiMOBoii cipyn C6 «ATMOccpnasi
A»(J)(J)y3im ii aarpH3HeHnc B03flyxa» HJ1, M, 1962
5. AyucKaAB h jp Oceaanne rpy6oancncpcnoro asp030.ast H3 npioeMiioro
cjioh aTMOct|)cpbi na noACTH^aiouiyio noBepxnocTb seiMH Tpyflbi ITO, Bun 185,
1966
6. JlyHCMiflB . He3ji»poBa H C. O h h k y ji P H 0 pacieTe paccen-
Bamin oceaaiomeii npiiMecii or .ntnciinoro HCTOmiiiKa b norpaminioM c.ioe
aTMoc({)cpbi Tpyau rro, Bwn 207, 1968.
7. MsaiiOB B H.CTpaTOiioBimP J1 K aonpocy o jiarpaHweBhix xapaKTe-
piicniKax TjpSy.icEiTiiocTH H3B AH CCCP, cep., reoii3, 10, 1963
8. H b a h o b B H OuenKa xapahTepiicTiiK Typ6y;ienTHoro nepeMemiiBaiiHH b hiijk-
Hesi c.ioe aTMoat>epbi M3B AH CCCP, cep reocjws , As 12, 1964
9 dBaHDB B H.UpflaHOBimA E CneKTpbi ckopocth Berpa npn HeycTofi-
hhboh cTpaTinjuiKamtit b HH3K043CT0TH0M wiana30ne H3B AH CCCP, cep
H3 aTHoccJ) ii OKeana, t III, Afs 8, 19G7
10. Hobiikob E A Mctoa c^ynafliiux ch/i b Teopim TypSy.ieiiTHOCTU }K3T,
44, 6, 1963
11. OciinoB [O. C flii(j)t{)y3iifi ot Tonemioro iiCTOUiiiKa Konemioro Bpe.\tenn fleii-
CTBHfl Cm HaCTOHLUIlfl cCOpHHK
12. C it h e n b m » k o b B. C O K034>c{)imi!ChTe Typ6y;ieHTKCifi micf>4)y3!:n lacnn
B3Bec». JKypiia.i npHK.ia.aiiofi Mc\amiKii n TexiiimecKofi 4>ii3iikii, j\? 6, 1967.
13. Cmiit . B TypOy/ieimioc pacceiiBauue o6;iaKa T«>he.nbrx lacum C6 «Atmo-
c<()epiiaH ,ni(})4)y3ii!i ii 3arpn3iiciinc Bosajxa* HJI, M, 1962
14. X e ii /l>h C, n a c k b ii ji XIii(J)(j)y3iifT ot nenpcpbiBiioro hctoihiiio b 3aBii-
chmoctm ot cneKTpa h MacuiTa6a TypSyjieimiocTii C6 *ATMOC$ep»aH
3iin ii 3arps!3iie!iiie B03nyxa» HJI, M, 1962
15. X ii h u e P >K TypC>7ienTiiocTb imiaTrii3, M, 1963
16. KDahh M H epiian flH(()<})y3iifr ii 3arpn3iiDiuie B03flyxa» MJ1, At , I962
17. Pasquill F. Atmospheric Diffusion London, 1962
- 62 -
-------
N. L BYZOVA
THE METHOD OF ATMOSPHERIC DIFFUSION EVALUATION
ON TURBULENT CHARACTERISTICS*
The pollutant distribution in turbulent atmosphere is discussed
with the help of Lagrangian method Approximate connection between
Lagrangian and Eulcrian turbulent characteristics is used for eva-
luation of turbulent diffusion parameters The diffusion parameter
estimations for boundary air layer are given. The comparison with
semi-empirical methods of pollutant distribution calculation is done.
This method allows to consider some specific problems, in particular,
to evaluate the influence of gravitational sedimentation on diffusion
parameters taking into account correlation loss because of particle
displacement in the process of vertical motion, to obtain formulas
for dispersion calculation of diffusible pollutant from a point source
of finite operation time (this problem is equivalent to the problem of
finite time of diffusion measurement from the source of infinite ope-
ration time).
* Editor's note: The abstract is presented as given in English with the original Russian article.
- 63 -
-------
SCATTERING OF SMOKE FROM A HIGH-LEVEL POINT SOURCE
Ye. K. Garger (USSR)
From Glavnoe Upravlenie Gidrometeorologicheskoy Sluzhby Pri Sovete Ministrov SSSR. (Chief Administration of
the Hydrometeorological Service Under the Council of Ministers of the USSR.) "Heteorologisheskie Aspekty
Zagryazneniya Atraosfery". (Meteorological Aspects of Air Pollution.) Sbornik dokladov na mezhdunarodnom
simpoziume v Leningrade - Iyul' 1968 g. (Reports delivered at the International Symposium in Leningrad -
July 1966.) Pod redaktsiey d-ra fiz.-mat. naufc M. E. Berlvanda. (Edited by Prof. M. E. Berlyand.)
Gidrometeorolgicheskoe izdatel'stvo, Leningrad, p. 194-206 (l97l). (Hydrometeorological Publishing House,
Leningrad, (197l).)
In calculations of the scattering of an impurity from a point source,
the greatest difficulty lies in calculating the scattering in the cross-wind
direction. In this connection, both theoretical and experimental studies of
this phenomenon are being conducted at the Institute of Experimental Meteor-
ology. Ivanov [5] and Byzova [2] have presented an outline of results of
theoretical studies which made it possible to develop a procedure for calcu-
lating the scattering. In the present report, an evaluation of the proce-
dure and its scope of applicability is made on the basis of experimental data.
The concepts and arbitrary notation fully correspond to the studies [5, 2],
Experimental Procedure
In the present study, the investigation of the scattering of a light im-
purity in the horizontal plane from a high-level point source involved the use
of oblique photography of smoke plumes, made from a 300-meter meteorological
mast.
To obtain the scattering parameters, it is necessary to use some model
of a smoke plume. We used as the basis Gifford's statistical model [4], which
corresponds most closely to actual plumes. The statistical model consists of
a set of discs with a Gaussian distribution of matter; the centers of gravity
of the discs are subjected to random oscillations relative to the wind direc-
tion. Scattering along the x axis is neglected. It follows from this model
that the dispersion of the impurity distribution for an averaged plume is
equal to the sum of the dispersion of the impurity distribution relative to
the instantaneous plume axis and dispersion of the coordinates of the in-
stantaneous axis relative to the averaged plume axis. The x axis coincides
with the average wind direction, the y axis is perpendicular to it, and the
z axis is directed vertically upward.
For dispersion along the y axis, we have
0? = 02 + °o- (1)
- 64 -
-------
Photography of the smoke plumes gives direct information on the geometric
shape of the plume, which can be converted to dispersion values. Thus, using
the expression for the instantaneous distribution of the smoke concentration
' = Q [2-u (olyat)2
1 "l-i -
(y~°y)2 (*-0z)2
2°;
2y
2o*f
(2)
where Q is the output of the source, u is the average wind velocity at the
source level, and Dy and Dz are the coordinates of the axis of the instan-
taneous plume; one can obtain a^y by integrating with respect to 2 and
assuming that the concentration is constant everywhere over the contour of
the plume, in accordance with Robert's "transparancy theory" [12], In this
case, we obtain
?o = Q[2-32] 2 u x e 2"2y = const.
(3)
Taking logs of (3) and differentiating, we get the expression
1 <*<=2
4 «
(y — Dy)2 d-\ 2 (y — Dy) a
dt
-w &-*>,)•
(4)
from which it follows that at time t, when y - Dy is maximum,
(y — — °2'
(5)
Using the values of q0 at the end of the plume, and also relations (3) and
(5), we finally obtain
lib
(y-Z)y)2
(6)
Hence, one can study as a function of t without specifying the form of the
2 2
function
-------
of the terrain over which it extends [1], Fig. 1 shows how the perspective
representation of a smoke plume is made, and also the basic elements of the
central projection, where P is the object plane, or plane of propagation of
the smoke plume; K is the plane of the photograph; S is the projection center
0 is the principal point of the picture; SO is a perpendicular to the plane
of the picture, equal to the focal distance of the camera; SN is the distance
from the projection center to the object plane, called the terrain clearance
and equal to the height of the camera above the smoke source (SN = H); n is
the nadir point, i. e., the trace of the ray SN on the plane of the picture;
a is the angle formed by the principal direction SO with the plumb line and
equal to the angle between the picture plane and the object plane.
By passing the vertical plane through rays SO and SN at right angles to
planes K and P, its intersection with the picture plane gives the prime ver-
tical, and the intersection with the object plane gives the photographic line
of direction NO. In Fig. 1, the x axis of the picture coincides with the
direction of the prime vertical, the y axis is parallel to the line of visibl
horizon hh, and the X axis coincides with the photographic line of direction.
When the origin of coordinates of the picture is chosen at the nadir point,
the coordinates of the points of the smoke plume are related to the points
corresponding to them on the picture by the following simple formulas (4):
Fig. 1. Oblique representation of smoke plume
cosset
C0S<1 Sill a
/
cos a sin a
COS a
/
(7)
- 66 -
-------
which describe the variation of the image scale with the angle of inclination
of the camera.
To estimate the relative values of X and Y after taking logs and differ-
entiating, we have
~~ = + 2tan'a da + cos«) .
" x f — x Sin a cos a '
-y- = -^--H»ngrfa+-^+ tf(-rsln.cosq)
' H y f~*r sin o cos a '
It follows from (8) that the errors in the difference of height H and in
the value of (f-x sin Ci cos Ol) cause a regular distortion of coordinates X and Y;
errors in the measurement of x and y cause proportional errors in the measure-
ment of X and Y; the error due to the angle of inclination is proportional not
only to the precision of measurement of the angle, but also to the tangent of
the angle of inclination, which increases rapidly with increasing C(. The
most convenient angle of inclination of the camera at H = 200-300 m turns out
to be 45°, since in this case the field of the picture is used to a maximum
extent, and the line of the visible horizon is recorded; this line makes it
possible to refine the angle of inclination of the camera to one minute,
based on the magnitude of segment 01 in Fig. 1, with the aid of the formula
01 = f/tan a,
the error due to a being determined only by the magnitude of da.
Formulas (8) do not consider all the errors inherent in the method. For
example, errors in the determination of X and Y may be due to an inadequate
flattening of the aerial film and deformation after the developing and fixing;
in particular, there may be errors in the determination of the plume edge,
owing to a lack of sharp image contrast. Substantial errors in the deter-
mination of the plume width and axis coordinates are introduced by oscillations
of the plume in the vertical plane (Fig. 2).
The relationship between the height of the
plume above the horizontal plane h, which
we took as the object plane, and the dis-
placement of the image 6rh on the picture
is given by the formula
*,h= -77-(l - -^-costpsin2aJ, (9)
where r is the radius vector drawn from
the nadir point; <(> is the angle made by
the radius vector with prime vertical.
Fig. 2. Displacement of plume image due
to oscillation in vertical plane.
- 67 -
-------
In processing the pictures, one takes the cross sections of the smoke
plumes along the y axis at predetermined distances x on the picture calcu-
lated from the first of formulas (7) and corresponding to fixed X values.
Hence, vertical oscillation of the plume introduces errors only into the
widths of the plume and into the axis coordinates; as the plume rises above
the horizontal plane, the image of the plume (Fig. 2) moves away from the
nadir point, and closer as the plume descends. In a stable atmospheric
stratification, this error will be minimal; thus, for h<10 m, Ot = 45°,
H = 200 m, f = 70 mm, and for maximum values = 20° , r = 130 mm, 6 ^ is equal
to about 0.27 mm, or approximately 1%.
The relative error in the determination of the smoke plume edge reaches
its largest value at the end of the plume, where the latter has least con-
trast and where for this reason a contour is outlined which introduces a
certain arbitrariness and smoothing of the plume edge. In some cases, the
error caused in this manner may reach 8-107.. The total maximum relative
error in the determination of the plume width and ordinates of the axis de-
pends on the portion of the plume and ranges from 3.5 to 11.57.. In an un-
stable atmospheric stratification, when the plume oscillations in the ver-
tical plane are large, the accuracy of the method decreases considerably,
and it is preferable to use stereophotogrammetry, which gives all three
coordinates of the object studied, but is more complex and time-consuming.
The random error in the determination of 02 and oo» in contrast to the
systematic error, whose sources and evaluation are indicated above, is much
larger and also increases with the distance from the impurity source because
the number of recorded plumes of the same length decreases in the course of
the survey. The figures shown in this study give the total mean square errors
of the measurements.
The experiments involved the use of an AFA-37 topographic aerial camera
with f = 70 mm, a diagonal angular field of 122 , and a frame size of 18 by
18 cm. The AFA-37 takes pictures automatically at predetermined time inter-
vals of 2 to 90 sec, and is equipped with a set of color and neutral filters
and a vacuum pump to flatten the aerial film when the photographs are taken.
To enable it to rotate around the vertical and horizontal axes, the camera is
attached to a special tripod mounted at the 300-meter level of a high-level
mast. The source of smoke (smoke charges) was placed at one of the levels
(75, 100, 125 m) and was set out on a mast ledge perpendicular to the flow.
Experimental Results
Table 1 lists data on the conditions of diffusion of the experiments, all
of which were carried out in the evening or morning hours during a stable strat-
ification of the atmosphere and at weak or moderate wind speeds in an hg layer
of about 2.0 m. The magnitudes of the Monin-Obukhov scale L indicate that in
- 68 -
-------
all the experiments at source height hs, the flow may be considered quasi-
uniform. The initial size of the plume yQ was equal to an average of 0.4 m.
Table 1
Experiment
number
Date
Time
Hr rain
T, min.
At,
sec.
Height,
h , a
act
0
d>
O
C
1
L, m
B
1
20/VII 1967
18 30
16,5
30
121,6
0,8
3,0
2,3
0,03S2
2
14/IX
18 00
15,0
30
75,0
0,0
1,7
0,28
0,6440
3
26/[[]
11 00
7,0
15
97,0
2,7
4,6
7,5
0,1134
4
1S/IV 19GS
19 00
9,0
15
97,0
1,4
4,0
40
0,0065
5
5 ;v
19 21
16,0
15
97,0
1,0
2,6
7,1
0,0378
6
0/V
05 13
11,25
15
97,0
1,4
7,0
20
0,0133
An important parameter characterizing the relative scattering of par-
ticles in the plume is the relative dispersion of the Impurity distribution,
which we determine directly from experimental data by using formula (6). Ex-
pressions for the relative dispersion are known in partial cases of short
and long time of particle diffusion (t - t0). Reynold's number Re is large,_
and the investigated scale of the phenomenon satisfies the conditions ,
where lofV'hg-'/t is the internal scale of turbulence (v is the viscosity of
air and £ the dissipation of energy), and L is the external scale. The latter
regime corresponds to the inertial range of the turbulence spectrum and in
this case, on the basis of similarity theory, Batchelor [10] found that
is determined by the rate of dissipation of turbulent energy E, by the time
t, and by the initial plume size yQ. If the initial plume size is small and
the condition 'O2—t/a^>y0, is fulfilled, after a certain time interval the
particles "forget" the initial distance, and O2 will then depend on E and
t in such a way that
2 ,1
3i —- est,
(10)
where c is a constant close to 0.4, in accordance with [3]
Experiments on relative diffusion carri-ed out by different authors and
compiled by Gifford [11] confirm the presence of this regime. The latter is
also confirmed by our data, but, in contrast to the results of [11], they
made it possible to follow the regime up to a period one order of
magnitude greater, and to establish its limits. In Fig. 3, where the values
of 02 and x are plotted on a log-log scale; the scale of this period for a
stable stratification of the atmosphere at a height of about 100 m is 500-
800 sec, and the distances are 2000 m. In some experiments, it was possible
to follow the variation of up to the limiting regime
particles move independently, and
2
ct2~/ , when the
5 * *
where 0^ is the structural function of a single particle.
- 69 -
-------
Pig. 3. ff2 vs. X.
1-6 — numbers of experiments*
- 70 "
-------
In [3], a type of correlation function of the distance between two
"liquid particles" discussed in [9] was employed, and on the basis of simi-
larity theory, an expression was obtained for the relative dispersion of the
impurity distribution. This dispersion is determined by the two main para-
meters of turbulent flow, the turbulent energy v2 and its dissipation rate £ :
C2=2l>M
*1"l —5 \r e~2'/v + 2e~ ]. (11)
Here tl = is the Lagrangian time scale.
cte
<$i /VTrk
Fig. I*, cr,/ V~2 rL vs,//u.
- 71 -
-------
The quantity v? was determined from measurements of 0-^ by using Taylor's
limiting formula
3| = V t
2
for small t, and £ was determined by interpolating measurements of 0^ with
formula (10). In Fig. 4, the experimental data are plotted on a log-log
scale. The solid line shows function (11), and the dashed line, the asymp-
1 /—
tote 02= (ce/3)'^ . The magnitude of the diffusion scale is equal to )/ vix .
Comparison of the theoretical curves with the experiment shows their sat-
isfactory agreement up to */*£ = 1,0, with the exception of the fifth experiment,
in which deviation from the asymptote 3/2 occurred sooner, when t/tl = 0,35. A
certain discrepancy at short diffusion times can be explained by the presence
of a considerable transition period from thec22~-/2to the regime,
recorded by measurements, and also explained by the influence of the aerody-
namic shadow of the high-level mast.
- 72 -
-------
In contrast to ^2# the dispersion of the ordinates of the instantaneous
axis ao is measured directly from data of the photographs without any addi-
tional hypotheses. This is a distinct advantage of the proposed procedure and
makes it possible to study the oscillations of the plume axis in pure form.
Whereas relative diffusion makes it possible to study the influence of
small-scale eddies comparable to the plume width on impurity scattering, os-
cillations of the instantaneous plume axis make it possible to evaluate the
influence of eddies of much larger size.
2
In [6], an expression was obtained for Oq based on the similarity theory
for the inertial range, the extrapolation form of which is
4 = 2?4(i- _ + i- e-'Vi). (12)
From (12) it follows that at low t's, the growth of is determined by
2
v2, and afterward it slows down and tends to r in the limit. This practi-
Lj
2
cally means that when/2>Ti,, the contribution of is small, and the split-
ting of the scattering processes in two is no longer valid. This may serve
as an independent estimate of tt.
Lj
In Fig. 5, the experimental data are given on a log-log scale for re-
lative coordinates, and the solid curve gives function (12). We should point
out the good agreement between the calculated and experimental data for small
values of i/Tf For large '//Tt , despite a certain scatter, all the experi-
mental points fall slightly below the calculated curve, which may be accounted
for by an insufficiently long period of observation at the corresponding dis-
tances.
Dispersion of the Impurity Distribution of an
Averaged Smoke Plume
2 2
Using measured data on and G^ and formula (1), one can obtain the quan-
2
tity which characterizes the scattering of an impurity in an averaged
smoke plume. The following expression is derived in [3, 6, 7]:
=27*! (*/*,+*-'"'-I). (13)
Figure 6 represents plots of relative values of Cf^ obtained experimen-
tally, and theoretical curves corresponding to a-, and a?*
/ J- £
- 73 -
-------
From a comparison of the experimental and theoretical data it follows
that the experimental points fall satisfactorily on the curve of for
small < tj%L\ i since the observation periods for these distances are suffi-
cient to give reliable estimates of At distance approaching xfX[-j= ,
the experimental data fall between the two limiting curves of and Oj
corresponding to dispersions from a continuous and an instantaneous source
of impurity; this is due to the effect of the limited duration of observa-
tion periods of
Table 2 shows values of £ and obtained from direct measurements
of the longitudinal wind velocity component (a) from the high-level mast and
from diffusion experiments (b). The anisotropy coefficient was taken as
0.75 [3]; the procedure for calculating z is described in [8], We can point
- 74 -
-------
out the good agreement of ' values and a somewhat poorer agreement of
values. The latter, obtained from diffusion data, are an average of 1.4 times
smaller than c values established from structural measurements on the mast.
Table 2 also gives values of and k^, which can be used to estimate
these quantities in the boundary layer of the atmosphere for a stable strati-
fication.
Table 2
+5
c
61 OO*
2.
/sec
m/sec
sec
, n
kp u
^/sec.
H
U
8. .
X o
a
b
a
b
a
b
a
b
a
b
CJ ==
1
0,65
1,0
0,18
0,22
850
736
2490
220S
27,0
35,0
2
0,21
—
0,11
—
941
—
1500
—
11,0
.
3
13,0
20,0
0,66
0,64
568
348
2610
1600
250,0
140,0
4
2,2
4,0
0,24
0,26
436
288
1750
1110
25,0
20,0
5
0,4
0,5
0,23
0,25
2200
20S0
5720
5420
120,0
130,0
6
4,8
5,8
0,36
0,31
450
370
3100
2600
58,0
35,0
LITERATURE CITED
1. 5 o 6 it p H 51 OoiorpaMMCTpHH reoaeaioaaT, M, 1956
2. Bbi30Ba H J1 MeToa oueiiMt aT.\ioc<|>epHori no Typ6y.nciiTHbiM
xapaKTepHCTiiKasi (Cm HacT c6)
3. Bbi30BaH Hsaiios B H OneiiKa napanteTpoB noncpeinioii A>i^i(J)y3iiti
b iihjkhcm cjioc aT,\ioc(()cpu no TypSy.iciiTHLiM xapanTepixTHKaM H3B AH
CCCP, ccp (J)H3 aTMOctfi h 0Keana, t 3, N° 5, 1967
4 rH({)<|)opa <> CTaTiiCTHnecKaa Moae/ib AbiMonoii CTpyit C6 «ATMOC(jJcpna>i
iiiik-
hcm c.noc aTMOCihcpLi H3B AH CCCP, cep rcocJ>i!3 , Ars 12, 1961
8 H b a ii o b B H fincciinaiuifl Typ6yjieim[ofi 3iicprini b aiMOC(J)cpe H3B AH
CCCP, 9, 1962
9 Hobiikob E A AIctoa cjivmafmux cu.n s Tcopim Typoy.ieimiocTii ^Kypna.a
3KcnepiiMeiiTa.nbiiofi ii Tec pern iccKOfi <}w3hkii, ]\.r° 5, 1963
10 Batchelor G K Diffusion in a field of homogeneous turbulence II The
relative motion of particles. Proc of the Cambridge Phil Soc , v 48, p. 2,
April 1952
11 Gilford Tr. Relative atmospheric diffusion of smoke puffs J Met, 14, 1957
12 Robertso F. T The theoretical scattering of smol
-------
E K. GARGER
ON SA\OKE DISPERSION FROM AN ELEVATED POINT SOURCE*
Measurements of smoke plume dispersion in horizontal plane were
carried out from elevated point source situated at a height of 300m
Photogrammetric method of smoke plume fixing was used in this
case
The values of pollutant dispersion, for instantaneous and averaged
smoke plume and of dispersion of axis ordmates for instantaneous
plumes were received. On the diffusive experimental data it was car
ried out an estimation of velocity of turbulent energy dissipation,
Lagrangian time scale, and coefficients of turbulent mixing in cross-
wind direction. The results were compared with analogous parameters
received from structural measurements fulfilled at the same time with
measurements of smoke plume dispersion.
"Editor's note: The abstract is presented as given in English with the original Russian article.
- 76 -
-------
DIFFUSION FROM A POINT SOURCE OF FINITE TIME OF ACTION
Yu. S. Osipov (USSR)
From Glnvnoe Upravleme Gidrometeorologicheskoy Sluzhby Pri Sovete Nlinistrov SSSR. (Chief Administration of
the-Hydrometeorological Service Under the Council of Ministers of the USSR.)_ "Meteorologisheskie Aspekty
Zncryoisnoniya Atmosfory". (Motooroloeiool Aopooto of Air Pollution.) SborniK doklodov no niozhdunorodnom
simpoziume v Leningrade - lyul' 1968 g. (Reports delivered at the International Symposium in Leningrad -
July 1966.) Pod redaktsiey d-ra fiz.-mat. nauk U. E. Berlyanda. (Edited by Prof. M.E. Berlyand.)
Gidrometeorolgicheskoe izdatel'stvo, Leningrad, p. 207-214, (1971). (Hydrometeorological Publishing House,
Leningrad, (197l).)
For a whole series of practical problems it is important to know the
ways in which the average concentration of an impurity will vary with chang-
ing length of the observation interval. It is well known, for example, that
even brief but substantial excess amounts over the maximum permissible con-
centration of an impurity in the atmosphere may cause enormous damage to
human health and have a harmful effect on vegetation and crops. As was shown
by experimental data, variations in impurity concentration with time are
fairly substantial, and at certain moderate distances from the source, the
maximum concentration may surpass the average value calculated from a 20-minute
observation interval by one order of magnitude or more.
As an example, Fig. 1 shows the variation with time of the density of
a deposit of aerosol particles 45-65 ym in size on the earth's surface at a
distance of 300 m from a point source 75 m high. Sharp increases in density
of the deposit with time are due to the movement of the impurity plume axis
across the observation point. At the same time, one can see how much the
20-minute average, drawn on the graph as a dashed line, may be below the
maximum values.
Within the framework of the statistical theory of atmospheric diffusion,
the concentrations of a diffusing substance are determined via the dispersions
of its distribution. To find the relationships between the maximum and aver-
age concentrations from a continuously acting source for different time inter-
vals, it is necessary to know the regularities of the variation of this dis-
tribution as a function of the length of the observation interval.
This problem was discussed theoretically by I. Ogura [10, 11], E. K. Byutner
[6], and N. L. Byzova [3], but the relationships they found are applicable only
to rough estimates. For this reason, it is of interest to study the empirical
characteristics of the dispersion of particles emitted by a point source of
finite time of action. Such characteristics were found to be obtainable from
the results of model experiments with fluorescent small particles 45-100 ym in
size, conducted in 1966-67 from the 300-meter meteorological mast of the Insti-
tute of Experimental Meteorology. The procedure of such experiments has been
described in detail in [1, 2].
- 77 -
-------
A distinctive characteristic of the experiments of these two series
was the fact that at the same level of the mast, two sources of an aerosol
of different colors began to operate simultaneously, spraying the same
amount of impurity over markedly different time intervals. The samples
were collected on sticky boards previously emplaced in the zone of the pre-
sumed deposit. The dispersion and and the axial densities of the
b 1
deposit and P^ (subscripts b and 1 refer to the brief and lasting sources)
were determined from the density of the impurity deposited on the b oards,
with the assumption that a normal distribution of its concentration takes
place in the direction across the mean wind.
If the diffusion in the direction of the mean wind is neglected, and
the adhesion of the impurity to the earth's surface is assumed to be complete,
then in the board method of sampling, the time of action of the source may
be considered equal to the time of deposition of the particles on the board,
this being analogous to the time of averaging of the sample during its col-
lection. Indeed, the time when we observe any particle at a certain distance
from the source differs from the time of its escape from the source by the
flight time t, and therefore the time interval during which we observe the
particle from the first to the last at this distance is equal to the time
interval during which the particles leave the source, with a shift by an
amount t. This "makes it possible to treat the time of action of the source
as an analog of the time of observation of diffusion from a continuously acting
source.
Fig. 1. Variation with time in density of deposit of
impurity aerosol particles on the earth's surface at a
distance of 300 m from a point source 73 m high.
The experimental results showed that as the duration of action of the
source increases, so does the width of the pollution zone, and the axial
densities of the deposit decrease, which is natural, since the total amount of
impurity was the same in the brief and long emissions.
If the reduction in the axial density of the deposit were completely
- 78 -
-------
determined by the increase in transverse scattering of the impurity, we
would have [5]
"k
However, as the duration of action of the source increases, the verti-
cal scattering of the impurity also increases. For the isotropic case,
~ A
(2)
If we consider that the contribution of vertical diffusion to the re-
duction of the axial density of the deposit is less than the contribution
made by the increase of scattering in the cross-wind direction, since for
the atmospheric boundary layer the size of turbulent eddies in the vertical
direction is restricted by the earth's surface, then
Pl C„
ir< — <-r- (3)
fl; r k ca
A
Table 1 gives average values of the ratios entering into (3), obtained
from our experimental data. It is evident from the table that (3) is
adequately confirmed.
Table 1
Series
Period of
Experiment
Number
of
Paired
Experi-
ments
A
°b
S1
1
2
3
4
VII l-X 1963 r.
VI—VIII 1966 r.
11-111 1967 r.
VI—VII
2
10
6
4
0,42
0,32
0,50
0,70
0,56
0,54
0,56
0,64
0,64
0,57
0,71
0,84
Average total
0,48
0,58
0,69
Simultaneously with the model experiment, measurements of the meteoro-
logical characteristics were made on the high-level meteorological mast. In
addition to the mean velocity profiles (averaged over 10-minute observation
intervals), wind direction, and temperature, measurements were also made on
the kinetic energy of the longitudinal component u^ of the wind velocity
fluctuations and on the rate of its dissipation £. The procedure and appara-
tus for measuring the turbulence characteristics were worked out at the
Institute of Experimental Meteorology and reported in several papers [7, 8].
Since we were interested in diffusion in the direction across the mean wind,
the analysis necessitated the use of the kinetic energy of the transverse
component v2 of wind velocity fluctuations, which was not measured directly.
- 79 -
-------
It is well known that in the surface layer, and are linearly
related to the square of the dynamic friction velocity v2. The numerical
coefficients of this relation were obtained in a whole series of experimental
studies and correlated in [9]. According to [9], it may be considered that
on the average, the transverse component amounts to 3/4 of the longitudinal
component. This made it possible to estimate v^ from the available u^ values.
The turbulent characteristics of v^ £ were used in an analysis of
experimental results to carry out the normalization, which made it possible
to reduce the results of different experiments to a comparable form. The
normalization factor for the time of action of the source T and time of
flight of the impurity t was the Lagrangian time scale of turbulence in
the direction across the mean wind, a scale defined in terms of the available
turbulence characteristics by the formula
C,E
(4)
0 2
where c, is a constant taken as 0.6; the dispersions and a were normal-
? ^—o o
ized to the square of the linear Lagrangian scale r~ = v^t*.
L L
The turbulence characteristics of the atmosphere were measured on a
series of levels of the high-level meteorological mast, ancl the normalization
factors and r^ were calculated by using the quantities v^ and £ (average
in the layer of propagation), the numerical values of which were obtained as
weighted averages over vertical profiles v^(z) and e(z).
Some basic characteristics of the series of experiments are shown in
Table 2.
Table 2
Series
Number
of
Paired
Experi-
ments
Height of
Source,
Time of
Action of
Source,
7", min
Limits of
Distances (m)
Up to Y/hich
2 1
o and o
were Measured
4
12
6
4
25-50
25-50
2
73-100
1,0-30,0
1,0—15,0
0,5-10,0
2,0-50,0
50-250
20-200
100-450
1000-1300
1000-2000
100-200
1000-2500
A substantial drawback of experiments of series 1 (1963) is the lack of
structural measurements of wind velocity fluctuations, which made it impos-
sible to carry out the necessary normalization of experimental data.
- 80 -
-------
1
Fig. 2. Experimentally obtained variation of standard devia-
tion O /r^ «ith diffusion time c,.
1) T) > 10,0, 7) 11-3,5, J) >1-1,5. 0 >)-0.75
Pig. 2 shows results of 30 experiments conducted in the summer of 1966
and 1967. The state of atmospheric stability was close to indifferent or
unstable, and the height of the impurity source in the experiments ranged
from 25 to 100 m. Values of are laid off along the ordinate axis, and
the normalized diffusion time £=t/xalong the abscissa axis; the curves
were drawn through the experimental points for different times of action of
the source rpT/x^. Most justified statistically is the upper curve 1, drawn
by averaging the results of 15 experiments. A considerable scatter of the
experimental points is observed in this case, but it is within the limits of
combined accuracy of determination of the dispersions (the accuracy of deter-
mination of O in the experiments was 10-15% [2]) and normalization factors.
In this group of experiments, ri varied from 11 to 40, and an attempt to sep-
arate them in more detail according to n failed to give any satisfactory
results. Curves 2 and 3 were drawn by averaging three experiments. The
scatter of the experimental points did not exceed *25% in either case.
Curve A was obtained by averaging the results of nine experiments, and the
scatter of the points did not exceed *15%.
Fig. 3 correlates the experimental results of 12 experiments conducted
in the winter of 196 7. A distinguishing characteristic of this series of
experiments was the low height of the source (March 2, above a snow cover),
and therefore the distances at which the dispersion of the transverse distri-
bution of the density of the impurity deposit was measured were smaller.
The same quantities as in Fig. 2 were laid off along the coordinate axes.
According to the time of action of the source, the experiments were
divided into three groups; curve a was drawn by averaging six experiments,
and curves b and c connect three experiments each. For comparison, Fig. 3
also represents a plot of the initial portion of the curves from Fig. 2.
- 81 -
-------
Considering the accuracy of the experimental data, the agreement may be
regarded as satisfactory. However, it should be noted that in the winter
series of experiments, the normalized values of the dispersions are sys-
tematically slightly higher than in the preceding series. This may be due
to certain operational differences in the experiments, inhomogeneity of
the surface atmospheric layer, or characteristics of measurement of the
normalization factors.
o_2
n
Fig. 3. Experimentally obtained dependence of standard
deviation o/r^ on diffusion time t,.
a) ti. b) n-2.5. c) 11-8,0, I. 2, 3, 4 — cm Fig. 2.
The experimental curves shown in Figs. 2 and 3 make it possible to
evaluate the width of the pollution zone (2a) as a function of a high-level
point source of impurity, taking the duration of its action into account.
As already stated, this problem is analogous to the problem of determining
the width of a pollution zone from a continuously acting source, allowing
for the finiteness of the observation interval. The accuracy of such an
estimate amounts to an average of *25-30%. Considering the insufficient
theoretical treatment of the problem of influence of averaging time on
turbulent diffusion, a problem that is of decisive importance in evaluating
the relationship between the average concentration during a given time inter-
val and the maximum possible concentration of a pollutant in the atmosphere,
the empirical relationships obtained are of definite practical interest.
Fig. A shows on a log-log scale the dependence of the ratio 0^/0^ on
the ratio of the times of action of a long and a brief source T^/T^, obtained
from our experimental data. It follows from the graph that
-------
On the basis of (5) and also by considering (1), (2) and (3), we can
see that the decrease in the maximum density of the deposit with increasing
duration of action of the source may be different depending on how active a
part is played by vertical diffusion in this process. If its contribution
is small and relation (1) applies, then
— I II I £
1 2 3 5 7 10 20 30^
Fig. 4. Increase in width of pollution zone with in-
creasing time of action of high-level point source.
In the other limiting case, when relation (2) may be considered valid,
P T-'/«
max
Three basic conclusions may be drawn from the above study:
1. The expansion of the pollution zone observed during increasing dur-
ation of action of a high-level point source and the resulting decrease in
the axial densities of the deposit were quantitatively estimated. On the
average, for the case of isotropic turbulent diffusion, the following relation
is confirmed:
•!l
-------
LITERATURE CITED
1. A^CKtaHApoBa A K McTOAiiKa iiccjiejoBamifl pacnpocTpaiiemin iickjcct-
Demioro aapo30.iH b npnaeMUOM c.noc C<3 «H3yiCHite norpamiinoro c.tosi st-
MOC(J>epu c 300 >ieTpoBOfi MeTeopo.ionmecKoii 6aiumi» H3A AH CCCP, M,
1963
2 A/icKcaiiapooa A K, Eusosa H Jl, M a m k o d a T B Onuibi no
pacnpocTpaucmiio ocawAaiomeftCH npiiMecii ot TOMe>inoro itcTomimca d hhw-
hcm cnoe aTMoc(])epbi C6. cHcc-neaoBamie HHHtHero 300-MeTpoDoro c/ioh aTMO-
cepiiOH flmfujijaiin no Typ6y.nenTiibiM xa-
paKTepiiCTiiKasi Cm Hacronuiiu'i cfiopmiK
4. EuaoBaH /I , HnaiiOB B H OueiiKa napaMeTpoB nonepcmioii jnr(fnJ>y3;in
b imwHCM cjioc aTMOC^epw no Tvp6y/ienTHbiM xapaKTepiicfiinaM H3B. AH
CCCP !!3 aiMoct}) h oneaiia t III, 5, 1967
5 E bi a o b a H /1, 0 c it n o b 10 C 0 paccefliimi npiiMecii ot Toicmioro hctom-
HiiKa b Hanpan-ieiiHH, nonepe'iiiOM BCTpy Tpyaw WIT, Bbin 4, 1967
6. EioTnep 3 K 3aBHCHM0CTb oeiiMMniibi AHcnepcmi nacTim, ncnycKaeMwi nenpe-
pblBllblM HCTOHIIEIKOM, OT HHT€pB3J13 BpCMCllll OnblTa TpyAbI ITO, Bbin 150,
1964
7. H b a ii o b B H Typ5y.ieHTnan SHepriia 11 ee jnccimamtn b hidkhcm c.noe aTMO-
cepu H3B AH CCCP, cep reoij>n3, ;\s 9, 1964
8. H b a ii o b B H HcnojibaoBaiiiie napameTpoa siijiepOBOii Typ6y.ieHTHOCTn xuih
oueHOK ^arpanjheobix xapaKTepHCTiiK (Cm HacT. cSopmiK)
9. /I a m .n h JXx , n a ii o b c k i: ii T CTpyKTypa aT.\toc<{)epHoii Typ6y^eHTiiocTH
«Mnp», M, 1966
10 Orypa H Omtcamic ziH
-------
USE OF SURFACE OBSERVATIONS FOR CHARACTERIZING
THE STATE OF THE SURFACE ATMOSPHERIC LAYER
G. B. Mashkova (USSR)
From Clavnoe Upravlenie Gidrometeorologicheskoy Sluzhby Pri Sovete Ministrov SSSR. (Chief Administration of
the Hydrometeorological Service Under the Council of Ministers of the USSR.) "Meteorologisheskie Aspekty
Zagryazneniya Atraosfery". (Meteorological Aspects of Air Pollution.) Sbornik dokladov na raezhdunarodnom
sirapoziume v Leningrade - Iyul' 1968 g. (Reports delivered at the International Symposium in Leningrad -
July 1966.) Pod redaktsiey d-ra fiz.-mat.' naufc U. E. Berlyanda. (Edited by Prof. M. E. Berlyarid.)
0ldrometeorolgicheskoe lzdatel'stvo, Leningrad, p. 215-220, (l97l). (Hydrometeorological Publishing House,
Leningrad, (1971).)
The stratification of the surface air layer plays an essential part
in the determination of the capacity of the atmosphere to scatter noxious
industrial emissions [1].
In the absence of an elevated inversion, lapse rate observations to
a height of 10 m provide reasonably good information on the stratification
of the entire surface layer. However, there are few points of lapse rate
observations. For this reason, surface meteorological measurements assume
a considerable importance. If the accuracy requirements for the determina-
tion of stratification are not very rigorous, Pasquill's classification can
be used for conversion to characteristics of the state of stratification.
However, this classification requires comparison with lapse rate observations.
The object of the present study was to draw such a comparison with
materials of observations from the high-level mast of the Institute of Experi-
mental Meteorology (lapse rate observations up to 8 m, weather station data,
profiles in the 300 meter layer). Pasquill's classification based on the use
of weather station data was compared with a classification based on the sta-
bility criterion B^ At in the surface layer, where At is the temperature dif-
v^
ference at heights of 4 and 1 m and v is the wind velocity at a height of 2 m.
Based on 1967 data, the frequency of different types of stability was
calculated after Pasquill on the basis of fixed times for each individual
month. Examples of the daily variation of different types of stability for
January, April and July are given in Fig. 1. The times of sunrise and sunset
are indicated by arrows.
Fig. 1 shows that in winter, an indifferent state prevails in the course
of 24 hours, being twice as frequent during the day as at night. At night,
a slight and moderate stability is observed rather frequently, but a marked
stability is seldom observed. By spring, the number of 24-hour periods with
a marked stability observed at night increases substantially, and an indiffer-
ent state is most characteristic of hours close to sunrise and sunset. In the
- 85 -
-------
daytime, the instability increases: the frequency of days with a slight
instability increases markedly; days appear with a moderate instability,
and near midday, with a marked instability. In summer, an indifferent
state is observed mainly near sunrise and sunset, and somewhat less often
near midday (the latter situation is apparently due to an increase in the
quantity of cumulus clouds during these hours). The frequency of days with
Nunber 'of cases
1 - very unstable state, 2 - unstable, 5 - slightly unstable,
4 - neutral (indifferent), 5 - slightly stable, 6 - stable,
7 - very stable.
- 86 -
-------
a slight and moderate instability is almost as high as the number of days
with an indifferent state. On some days, a marked instability is observed,
most frequently during hours close to midday. At night, a marked or moder-
ate stability prevails. In autumn, the picture gradually changes and comes
to resemble that of spring (October is an almost complete analog of April).
All the observations for which the value of criterion B could be cal-
culated were arbitrarily divided into groups according to the magnitude and
sign of B (Tables 1 and 2). As one of the characteristics of these groups
one can use the character of the temperature profile in the lower 300-meter
layer as a function of the sign of the lapse rate in the 0.5-10.0 m layer
and above [2].
Table 1
Frequency {#) of Temperature Profiles of Different Types According to Groups
of Parameter B
4)
¦t
+3
2
a.r-f
B
u
o
e e.
O 3 0)
< -0,01
-0.009,
-0,002
-0,001,
+0,001
0,002-
0,009
0,010-
0,040
>0,050
E-h
0) <*-i
Oh a.
o
0)
a
>»
e-»
jJ i—1
ca h
o u
•HO) O
.p a. Ui
« 5 a.
B 0)
0JH
o
1
Marked
and Very
Marked
Instabil-
ity
Slight
and
Moderate
Instabil-
ity
Indiffer-
ent State
and One
Close
To It v
Slight
Stability
Moderate
Stability
Harked
Stability
1
\
100
88
63
—
—
—
2
—
12
4
—
—
—
3
>
—
—
29
70
36
—
4
~
—
—
4
30
64
100
Type 1 assumes the presence of a vertical temperature lapse rate y in
the entire 300 meter layer (Y>0); type 2, a positive lapse rate in the lower
surface layer (y>0) and a negative one above it (Y<0) ; type 3, a negative
lapse rate in the lower surface layer (Y^) and a positive one (y>0) above
this layer; type 4, stability (Y<0) in the whole (or nearly whole) 300-meter
layer. The frequency of the different types of temperature profiles for dif-
ferent range groups of B are given in Table 1, compiled from observational
material for several years in the warm half of the year. The table shows
that in 33% of the cases, an indifferent state in the surface layer corres-
ponds to temperature profiles with a changing sign of the lapse rate in the
300-meter layer (types 2, 3), i.e., corresponds to certain transition states
in the layer considered> mainly in the morning and evening hours. A substan-
tial portion of such profiles (70 and 36% respectively) are observed during
- 87 -
-------
a slight and moderate stability, and a much smaller portion (12%), during a
slight instability. Examples of the temperature and wind velocity profiles
for different values of B are given in Fig. 2.
ruble 2
Frequency ($) of Different Types of Instability
According to Groups of Parameter B
B
Types of
Instability
Hissed on
<-0,04
-0,030,
-0,010
-0,009,
-0,002
-0,001,
+0,001
0,002-
0,009
0,01 —
0,040
5:0,050
Surface,
Data
Very
Marked
1 Insta-
bility
' (1)
Marked
Insta-
bility
(2).
Slight,
and
Moderate
, Insta-
bility
(3)
Close to
Indiffer-
ent
State
.(O
Slight
Stability
,(5) ¦
Moderate
Stabil-
d)y
Harked
Stability
(7)
1
2
3
4
5
6
100
48
20
32
5
24
70
3
6
B1
9
1
10
51
n
12
12
24
11
53
15
20
10
15
40
20
/
Number of
cases
11
59
54
105
49
34
Table 2 gives the frequency of different types of stability after
Pasquill (surface observations) according to groups of parameter B (lapse
rate observations). As is evident from the table, the Pasquill types are
in most cases similar to the types based on criterion B. The maximum fre-
quency in five cases falls on the same type; in two other cases, on neighbor-
ing types. The latter situation is due to transition moments in the morning
and evening periods, when a rearrangement of the stratification of the entire
lower portion of the boundary layer takes place. For a very strong instabil-
ity corresponding to the first type according to surface observations, we
were unable to obtain a single value of B, since this type is characterized
by a very low wind velocity at the wind vane, corresponding to a calm at a
height of 2 m, so that B cannot be calculated. However, it follows from the
same consideration that B under these conditions is much larger than 0.03 in
absolute value, and hence, the first instability group according to B corres-
ponds to the first instability group according to surface data.
Table 2 summarizes all the data obtained for the spring of 1967. Analysis
of these data with separation into observations in the warm season (208 cases)
and cold season (124 cases) gave a picture that was no different from the dis-
tribution of frequency according to types for the year. The only character-
istic feature of the distribution was the fact that the number of cases of
marked instability and marked stability in the cold half of the year was much
lower than in the warm half, and the number of cases with slight and moderate
stability was greater (as can be seen from Fig. 1).
- 88 -
-------
300
200
B"0,008
t' v'
Fig. 2. Examples of profiles of temperature t and wind
velocity v for different values of parameter B.
Thus, experience in comparing the two classifications, based on obser-
vations of an individual year, shows that the classification based on standard
surface meteorological observations is apparently reduced to the classifica-
tion based on lapse rate observations. Consequently, the characteristics
obtainable with the aid of lapse rate observations may be justifiably extended
to much larger areas when network meteorological observations are taken into
consideration.
- 89 -
-------
LITERATURE CITED
1. Bonpocw aiMoc^epjioij Am])i)>y3ii)r ji 3arpH3ncui!ji so3jyxa. Iloji pen. M E. Bep-
.1 hh A a Tpyflbi rrO, awn. 207 I9G8
2. H a in k o d a T B. 0 npo<|)ii;iHx TeMnepaTypbi B03Ayxa h BCTpa b iinmnei'i Hacni
norpantiMHoro won aTMoc
-------
SULFUR DIOXIDE AND DUST MEASUREMENTS IN MEASURING
NETWORKS OF THE HYDROMETEOROLOGICAL INSTITUTE*
0. Muller CSSR (Czechoslovak Socialist Republic)
From Clavnoe Upravlenie Gidrometeorologicheskoy Sluzhby Pri Sovete Ministrov SSSR. (Chief Administration of
the Hydrometeorological Service Under the Council of Ministers of the USSR.) "Meteorologisheskie Aspekty
Zagryazneniya Atmosfery". (Meteorological Aspects of Air Pollution.) Sbornik dokladov na mezhdunarodnom
simpoziuro v Leningrade - Iyul' 1968 g. (Reports delivered at the International Symposium in Leningrad -
July 1966.) Pod redaktsiey d-ra fiz.-mat. nautc M. E. Berlyanda. (Edited by Prof. M. E. Berlvand.)
Gidrometeorolgicheskoe izdatel'stvo, Leningrad, p. 221-226, (1971). (Hydroneteorological Publishing House,
Leningrad, (l97l)«)
In most industrial countries, a significant increase
in air pollution has been recorded in recent years.
Czechoslovakia, too, has not been spared from the negative
results of industrialization. In particular, the bad effects
of air pollution have today become an economic and sociological
problem of the first rank. Some areas are even today limited
in their further industrial development by the air pollution
factor. At present, unfavorable air purity conditions
prevail in the North Bohemian coal basin, one of the greatest
regions of industrial concentration of our country. It is
maintained that this area is one of the most polluted regions
of the world. The origin of the air pollution is for the
most part due to the burning of low-quality lignite
produced in the North Bohemian open pit mines. It serves as
the energy source for all the industry in the region. This
lignite contains from 1.5 to 6% sulfur in its dry material.
In combustion processes, this sulfur is mostly exhausted to
the atmosphere as sulfur dioxide. The greatest SO2 emitters
in the region are large power plants, which are assigned the
greatest proportion of the air pollution.
Also of decisive importance in the pollution of the air
layers near the ground are industry, house fires, steam-powered
transporation, and burning coal seams.
The danger of very high air pollutant concentrations is
due to the unfavorable meteorological situation of the North
2
Bohemian basin. This region, covering an area of about 3,500 km ,
is bounded on" the northwest by the Erzgebirg chain and on the
* Editor's note: This report was originally presented in German. Its English translation has been supplied
by APTIC.
- 91 -
-------
southeast by the offshoots of the Central Bohemian Range. It
is a poorly ventilated area, in which weather conditions with
poor exchange frequently occur, leading to accumulation of
injurious materials.
Results of air pollution include, among others, extended
forest damages, decreased yields, in agricultural production,
and increased corrosion damage. But it is mostly the population
which must suffer from the air pollution. Aside from stress
due to dust and smoke, suspicious differences in the state of
health of population groups from industrial cities and from
regions without industry have been related to air pollution.
The extensive reports available on the bad effects of
polluted air on humans, animals, plants, and economic goods
are so far contrasted to only occasional exact data on the
concentrations of harmful materials in the air. But the
measurements undertaken have shown that the maximum immission
concentration values for SO2 are far exceeded in many places.
In the last few years, the hygienic service has performed
long-term and large-scale measurements of sulfur dioxide and
dust in the North Bohemian basin by means of semiquantitative
methods. Thus SO2 has for several years been determined by
a summation method based on spontaneous absorption of SO2 in a
sodium hydroxide solution. To be sure, the absorption rate
is strongly dependent on the wind velocity and the temperature.
The method provides relative values which are quite comparable
with each other, but which do not allow direct conversion into
SO2 concentrations. At present, this measuring network includes
some 300 test points.
At the same test sites, the dust deposition is determined
by a sedimentation method. Amounts of dust deposited in collection
vessels are determined gravimetrically every month. This method
also provides only approximate values, as it is well known that
the larger particles of airborne dust predominate by far and so
- 92 -
-------
are determined preferentially.
In 1967 the Hydrometeorological Institute of the CSSR was
given the mission of establishing a measuring network for
measurement of sulfur dioxide and dust concentration in the
North Bohemian coal basin. The goal of this research
can be summarized briefly in the following points:
1. The experimental values are to provide information
on the present state and further development of air
pollution.
2. The SO2 and dust concentration values determined, along
with meteorologic measurements, are to serve as bases
for study of the relations between air pollution,
weather, and climate.
The practical application of the measurements lies primarily
in the following areas:
1. The values determined for SO2 and dust concentration
serve as bases for remedial measures.
2. The effectiveness of the measures carried out for decreasing
emissions will be evaluated by means of the measure-,
ments.
3. The concentration values will be bases for the evaluation
of sites for possible construction of more large, emitters,
especially power plants.
4. The measurements will be made available to all institutions
directly interested in air pollution.
Determination of the relation between air pollution, weather
and climate should contribute to knowledge about the dissemination of
harmful material in the boundary layer, serve to make diffusion
calculations more precise, and finally, contribute to the
possibility of air pollution prognosis, which would be in the
form of a warning service for the controlled operation of large
emitters of the greatest importance.
- 93 -
-------
About 85 fixed long-term measuring sites are planned in
2
the whole region covering some 3,500 dm . The measuring points
do not form a geometric network. Rather, they are selected
so that the basin> proper as well as the slopes and ridges of
the Erzgebirge range are rather evenly covered by measuring
sites. To exclude local effects on the measurement as much
as possible, and to obtain a larger range of representative
values, we attempted to set up measuring sites away from
the larger settlements and industrial plants. In following
the level of harmful material in a large area measurement
area, we cannot go into the often more complex air pollution
conditions in settlements or on factory land. Thus only
some control measuring points are planned in cities.
As already mentioned, the measuring program proper is
limited to measurement of concentrations of SO2 and dust.
At all the measuring sites, the sampling period for SO2
and dust is 24 hours. For about 10% of the measuring sites,
a special program is planned, by which the time variability of
the dust and SO2 concentration will be followed with continuously
recording analyzers.
The basic method for sulfur dioxide is the generally well
known colorimetric determination according to West and Gaeke.
An absorption solution is aerated for 24 hours in an impinger
washing bottle with a measured amount of outside air. The
sulfur dioxide content of the air sample is then determined
in the laboratory with a color reaction. The color resulting
from the reaction is measured with a photometer. The intensity
of the color is proportional to the SO2 content of the sample.
To determine the dust concentration, a standardized membrane
filter method recommended by the hygienic service is used. The
amount of suspended dust caught on the filter is determined by
weighing.
- 94 -
-------
The automatic analyzers for sulfur dioxide and dust planned
for the special measuring sites are both devices developed in
the CSSR.
The sulfur dioxide trace analyzer developed by Novak works
on the polarographic-coulometric principle. Reaction of the SO2
with an absorption solution quantitatively produces a reaction
product which is electrolytically decomposed in the device.
The flow thus produced at an electrode pair is a measure of
the SO2 concentration in the air being studied. It is recorded
continuously by a recorder. The recorder charts are evaluated
over 30-minute intervals.
The Aerosol Concentrometer of Polydorova is used for
continuous recording of dust concentration. The air under
study is sucked through a continuously transported membrane
filter tape. Following that, the decrease in the intensity
of a light beam is measured. This is proportional to the
blackening of the filter and thus to the dust concentration.
Thirty-minute averages are taken from the continuous recording.
At the time, SC^ experimental measurements are being done
to a small extent. Routine measurements in the entire net
will be started at the beginning of 1969. It should be mentioned
that the SO2 and dust orientation measurements of the hygienic
service mentioned above are coordinated with the measuring pro-
gram of the Hydrometcorological Institute. It is planned to
perform all measurements at the same points.
After the description of the network for air pollution
measurements, we shall briefly report on the meteorological
companion measurements.
In this region, there are already available about 35
meteorological stations and two meteorological masts 80 m high.
The program of the meteorological stations contains measurements
- 95 -
-------
of air temperature, relative humidity, and wind direction and
velocity. The meteorologic masts have an extensive special
program directed toward obtaining measuring values for a meso-
climatic evaluation of the limiting layer with respect to
diffusion of harmful materials. Measurements and observations
are carried out uninterruptedly on the masts, and evaluated at
hourly intervals. At the same time, the masts are special
measuring sites for measurement of harmful materials.
Less extensive measuring networks are planned for other
industrial regions in the near future. Part of them are already
under construction. These are the measuring networks in the
districts of Frydlant (North Bohemia, with 12 measuring sites),
Ostrava (North Moravia, about 25 measuring sites), Bratislava
and Kosice (Slovakia, with a total of about 20 measuring sites).
For the next stage, we plan measurements of other harmful
materials. We will be particularly concerned with the reaction
products from SO2. Sulfur dioxide is generally assigned the role
of an indicator for air pollution by the waste gases from heating
plants, but SO2 is unstable. During its transport through the
atmosphere it undergoes extensive physical and chemical reactions.
It is obvious that the effect of weather conditions on oxidation
should be followed, and the reaction products, such as total sulfur
or sulfate, should be included in the measuring program.
DETERMINATION OF CONCENTRATIONS OF SULFUR DIOXIDE AND DUST
BY A MEASURING NETWORK OF THE HYDROMETEOROLOGICAL INSTITUTE* (ABSTRACT)
0. Muller (Czechoslovak Socialist Republic)
In recent years, the level of atmospheric pollution in Czechoslovakia has risen
considerably. Particularly unfavorable conditions arose in one of the major indus-
trial regions of the country, the North Bohemian coal basin, whose open pits are
used to mine low-grade brown coal, which is the chief source of power for the indus-
try in that region. The high sulfur content of brown coal (from 1.5 to 6%) causes
considerable amounts of sulfur dioxide to be discharged into the atmosphere during
the combustion of coal. .Large sulfur dioxide concentrations in air persist for a
long time as a result of the geographical characteristics of the location of the
* This is a translation of the original Russian abstract accompanying this German paper.
- 96 -
-------
coal basin, which occupies a vast territory bordering on mountains where
stagnant situations frequently arise.
In the last few years, Czechoslovaks's sanitary service conducted
studies of air pollution over the territory of the coal basin, where a
network of stations consisting of 300 points was created. This network per-
formed measurements of mean daily sulfur dioxide concentrations and of dust
deposits by using semiquantitative methods. The sulfur dioxide concentra-
tion was determined by the addition method, based on spontaneous adsorption
of the gas in a solution of sodium hydroxide. Since the magnitude of the
adsorption substantially depends on the wind speed and temperature, it is
difficult to convert the relative values obtained to actual sulfur dioxide
concentrations. The magnitude of the dust deposit was determined by a
gravimetric method.
In 196 7, a study of air pollution in the region of the North Bohemian
coal basin was initiated by the Hydrometeorological Institute of CSSR,
which organized a measuring network consisting of 85 stationary points
located at a certain distance from major population centers and industrial
plants. The program of studies involves measurement of mean daily concentra-
tion of sulfur dioxide and dust in air.
The West-Gaeke method is used to determine the sulfur dioxide concen-
tration. The dust content is determined by means of membrane filters.
Continuous monitoring of the sulfur dioxide and dust concentrations is carried
out at 10% of the measuring points by means of automatic analyzers. A Novak
gas analyzer based on the polarographic-coulometric principle is used to
record sulfur dioxide. The dust concentration is recorded with a Polidorov
aerosol concentrometer based on measurement of the attenuation of the intensity
of a light beam passing through the moving belt of a membrane filter through
which the air studied is drawn. The tracings obtained on the belts of the
two analyzers are averaged over a 30-minute interval.
In the coal basin region, in addition to measurements of air pollution
at 35 meteorological stations and two 80-meter meteorological towers, meteoro-
logical observations are carried out that include measurements of temperature
and relative humidity of air and wind velocity and direction.
An expansion of the studies of atmospheric pollution is being contemplated
in the immediate future. The creation of a network of measuring points in
other industrial areas and an increase in the number of ingredients measured
have been proposed.
- 97 -
-------
EXPERIMENTAL STUDIES OF ATMOSPHERIC POLLUTION
IN INDUSTRIAL AREAS
B. B. Goroshko and E. N. Zasukhin (USSR)
From Glavnoe Upravlenie Gidrometeorologicheskoy Sluzhby Pri Sovete Ministrov SSSR. (Chief Administration of
the Hydrometeorological Service Under the Council of Ministers of the USSR.) "Meteorologisheskie Aspekty
Zagryazneniya Atmosfery". (Meteorological Aspects of Air Pollution.) Sbornik dokladovna mezhdunarodnom
simpoziurae v Leningrade - Iyul1 1968 g. (Reports delivered at the International Symposium in Leningrad -
July 1966.) Pod redaktsiey d-ra fiz.-mat. naulc U. E. Berlyanda. (Edited by Prof. M. E. Berlyand.)
GidrometeorolgichesKoe izdatel'stvo, Leningrad, p. 227-237, (l97l). (Hydrometeorological Publishing House.
Leningrad, (1971)*)
An experimental study of the spreading of noxious substances from
individual high sources and of the influence of meteorological conditions
on the concentration field were begun by the A. I. Voyeykov Main Geophysical
Observatory in collaboration with the Moscow Institute of Hygiene im.
F. F. Erisman and other agencies in 1961. Since that time, extensive expe-
ditions have been carried out every year in the areas of individual sources
with different emission characteristics, in different seasons and in differ-
ent climatic zones. The results obtained make it possible to draw a compar-
ison with data of methods for calculating the scattering of impurities in
the atmosphere and to establish the mechanism of propagation as a function
of the meteorological conditions and emission parameters. They are also of
great importance in the development of methods studying the pollution of a
city's air reservoir as a whole.
In the study of the concentration field in the region of individual
sources, an extensive series of measurements were formulated that include
meteorological observations, the determination of the concentration field,
and the measurement of emission parameters [3, 4,7]. A distinctive feature
of these studies is the fact that they involve the measurement of not only
the concentrations of pollutants, but also of all the main elements determ-
ining the spreading of impurities under natural conditions.
A significant influence on the scattering of an impurity in the atmos-
phere is exerted by the distribution of the wind velocity, temperature,
wind direction and other meteorological elements, the greatest variation of
which in both height and time is observed in the surface layer. For this
reason, the most detailed measurements were carried out in this layer (when
the indicated expeditionary work was organized). The wind speed was measured
at eight levels of a telescopic mast up to a height of 17 m by means of con-
tact anemometers with recording on electromagnetic counters. This method
made it possible to perform an automatic averaging of the wind speed over the
necessary time and to carry out remote measurements.
- 98 -
-------
In view of the fact that the duration of the sampling was 20 minutes,
the wind speed was also averaged over 20 minutes, with subsequent averag-
ing over 1 hour. The temperature and humidity of air were measured at six
levels from 0.2 to 17 m with resistance thermometers, and measurements of
the solar radiation were made with a balansometer, actinometer, and
albedometer. The soil evaporation and temperature were measured at five
levels to a depth of 20 cm. All the meteorological elements were recorded
round the clock, every hour during the day, every 3 hours at night, and
every hour during the sampling.
in
Fig. 1. Telescopic masts with installed wind velocity and
temperature gauges.
Fig. 1 shows 17-meter telescopic masts with wind velocity and temperature
gauges.
The data obtained permit the calculation of the turbulence coefficient
and the elimination of the influence of individual meteorological elements on
the degree of atmospheric pollution. Continuous recording of the variation of
- 99 -
-------
wind direction by means of an anemorhumbograph permits an evaluation of the
degree of dissolution of the plume in the horizontal plane.
The presence of a high stack through which the noxious pollutants are
discharged into the atmosphere, a considerable exit velocity of gases from
the stacks, and for hot sources, overheating of the gases relative to the
ambient air leads to an effective ascent of the plume. Thus, for a high-
capacity state regional electric power plant (SREPP) with a stack height
of 120 m, the effective ascent is 30 to 200 m, and under certain meteoro-
logical conditions, as high as 1000 m. For this reason, it was important
to determine the size of the plume in width at various distances from the
source, the range of propagation of its visible portion, and the height of
ascent. The measurements were made with the aid of an airplane or helicopter
and by photographing in the plane perpendicular to the plume.
To measure the meteorological elements at heights up to 500 m, use was
made of the balloon sounding method, and up to 1000-1500 m, of a specially
equipped airplane or helicopter.
One of the main areas in the overall program of study consisted in
determining the surface concentrations of a number of major pollutants.
The most commonly employed chemical methods of analysis were used for this
purpose with only slight modifications. The accuracy of most of these methods
was 20%. Basically, the sampling for all gaseous ingredients was made by
the suction method by drawing a given volume of air through an absorbing instru-
ment.
The dust was collected by using filters of FPP-15 cloth, and the concen-
tration was determined gravimetrically. The sensitivity of the determination
of dust was 0.1 mg/m^. The air, which was analyzed for the soot content, was
drawn through paper filters. The soot concentration was determined from the
degree of blackening of the filter, and the sensitivity of the determination
was 0.1 mg/m3.
Gases and substances in vapor form were trapped in liquid absorbing
media. Absorbing instruments with a porous glass plate were used for this
purpose. The sulfur dioxide concentration was determined by absorbing the
gas in a potassium chlorate solution, then determining the sulfate ion formed
by turbidimetry. The sensitivity of the method was 0.12 mg/m^. Nitrogen
dioxide was determined with Griess reagent after absorption in potassium
iodide solution. The sensitivity of the determination was 0.02 Carbon
monoxide was determined by using a TG-5 gas analyzer in which the oxide was
burned to the dioxide, then determined titrimetrically. The sensitivity of
the determination was 1.5 mg/m^. Carbon disulfide was determined from the
reaction of formation of a colored compound, copper diethyldithiocarbamate.
To eliminate the influence of hydrogen sulfide and sulfur dioxide, tubes
filled with dry adsorbents were used. The sensitivity of the determination
- 100 -
-------
was 0.02 g/m^. .Phenol was determined from the reaction with diazotized
paranitroaniline. The sensitivity was 0.01 mg/m3. Hydrogen sulfide was
trapped with a suspension of cadmium hydroxide and determined from the
reaction of formation of methylene blue. The sensitivity of the method
was 0.006 mg/rn^.
In the survey of individual high-output sources, the samples were col-
lected under the plume at about 20 points simultaneously at 3-5 distances
and sometimes more. At each distance, from 3 to 5 sampling points were
established perpendicular to the plume at a distance of 50-400 m from one
another depending on the width of the plume.
To determine the influence of the daily variation of meteorological
elements on the scattering of the impurity, the sampling was carried out
at different times of day: in the morning (in the presence of weakly developed
convective exchange), during the day (during the period of maximum turbulent
exchange) and in the evening (when the turbulent exchange had decreased).
Simultaneously with the sampling, a measurement was carried out on the
characteristics of the emission, including the determination of the total
volume of gases discharged and the amount of pollutants discharged for
each ingredient and temperature of the escaping gases.
The collection and analysis of samples in cities with advanced industrial
development have by now been organized. In order to study the pollution of
the air reservoir of cities, stationary sampling points are being organized
[2]. In many cities, samples are collected along a predetermined route by a
specially equipped machine and under the plumes of the main sources.
The sampling points are located on open, well-aerated sites character-
istic of a given district of the city, taking into account the location of
the pollution sources, the wind rose, and as a rule, in places of maximum
pollution. Due consideration was given to the fact that the concentration
maximum is observed at a distance of 20 H from the source (H being the height
of emission of the pollutants). Special booths were set up at the stationary
sampling points (Fig. 2). This made it possible to obtain representative
data and to carry out the sampling at any time of day, an important feature
in determining the relationships between pollution and meteorological conditions.
Simultaneously with sampling at stationary and route points, observations
were made on the speed and direction of the wind, temperature and humidity of
air, and special weather phenomena and special conditions that might influence
the magnitude of surface concentrations.
A general meteorological characterization of an air mass is obtained as a
result of additional meteorological measurements (lapse rate observations,
sounging of the atmosphere with the aid of a surface radiosonde at the nearest
meteorological station).
- 101 -
-------
The data obtained in a survey of one of the thermal power plants
permit the establishment of several dependences on various meteorological
elements [5]. Measurement of the emission characteristics made it possible
to select for analysis cases in which the source operated stably, and thus,
after the surface concentrations have been measured, the source is uniquely
determined in this case by the meteorological conditions of the scattering.
In order to bring out the influence of the daily variation of meteoro-
logical elements on the scattering of pollutants, all the data were divided
into three groups as functions of the sampling time: a morning period from
5 to 9 A.M., a daytime period from 9 A.M. to 4 P.M., and an evening period
from 4 P.M. to 7 P.M. Fig. 3 shows curves enveloping the maximum sulfur
dioxide concentrations at different distances from the source at different
time of day. As is evident from the figure, the lowest concentrations are
observed during the morning periods. On passing to daytime conditions, the
zone of concentrations above the maximum permissible norm is quadrupled, and
from the standpoint of the absolute maximum, more than doubled. During the
evening period, a decrease of the concentrations is observed at all distances.
Fig. 2. Booth for collecting air samples.
Thus the most unfavorable conditions are created during the daylight
hours, when the concentrations are considerably increased and the zone of
pollution is substantially widened.
Such studies were carried out separately for conditions of anticyclonic,
cyclonic and intermediate pressure fields. They gave similar results that
may be explained by a variation in turbulent exchange, which increases from
morning to daytime hours and decreases toward evening. This is in good
- 102 -
-------
agreement with the theoretical treatments developed at the Main Geophysical
[1].
A strong influence on the variation of surface concentrations is exerted
by various meteorological elements. Fig. 4 shows curves enveloping the
maximum concentrations during the winter and summer periods as functions of
cg/ra^
2,0 r 2
toka
Fig. 5. Variation in sulfur dioxide concentrations under
the plume of a high-output source.
1 - morning, 2 - day, 5 - evening
temperature At in the 0.5-2 m layer (for At < 0 and At >0). It is evident
from the course of the curves that in the presence of an isothermal or inversion
mg/m5'
Tig. 4. Variation of sulfur dioxide concentrations in
the region of a major source as a function of the temp-
erature difference At in the 0.5-2 i layer.
Summer: (l) At «$0, (2) A t > 0; wiritert (3) At»$ 0,
CO A t > 0.
- 103 -
-------
distribution of temperature, the surface concentrations of sulfur dioxide
decrease substantially. Thus, when At > 0, they are approximately twice
as large as when At < 0.
On passing from winter to summer conditions, a general increase in
concentration and a 1.5-fold expansion of the range of influence of the
source takes place.
An analogous graph was plotted for the distribution of maximum concen-
trations for different gradations of wind velocity at a height of 2 m above
ground: 0-2 m/sec, 3-6 m/sec, and above 6 m/sec. It follows from this graph
that the minimum sulfur dioxide concentrations are observed at a wind velocity
up to 2 m/sec. There is no distinct maximum, the concentrations do not exceed
the MPC, and are four times less than at wind velocities greater than 2 m/sec.
As the wind velocity increases, the concentration near the ground is seen to
increase. It reaches a maximum value of 1.8 mg/m^ at a wind of 3.0-6.0 m/sec.
As the wind velocity increases further, the absolute value maximum is pre-
served, but it shifts toward the source. This may be explained by the fact
that at such velocities, the effective height of ascent of the plume is
moderate, the plume is nearly horizontal, and as a result of vigorous mixing,
the sulfur dioxide gas is rapidly carried toward the ground to form a concen-
tration maximum near the source.
7X
\
N
*-0.30
Fig. 5. Isolines of equal sulfur dioxide
concentrations on the territory of a city.
1 - city limits, 2 - sampling points,
3 - isolmes of mean monthly concentrations
» 2
- 104 -
-------
H, m
Fig. 6. Height distribution of carbon monoxide concentrations in the plume
of a metallurgical plant for a stable (a) and unstable (b) stratification (t)
at various distances from the source.
- 105 -
-------
A different type of dependence on the wind velocity is observed for
lower and smaller sources located in a city. Maximum sulfur dioxide con-
centrations near the ground are observed in these cases at wind velocities
up to 4 ra/sec. For high sources in the city, as the wind velocity increases,
there is usually an intensification of turbulent mixing, which on the one
hand leads to a more intensive transport of impurities into the lower atmos-
pheric layer, and on the other hand, the atmosphere of the city becomes
cleaner as a result of a stronger aeration of residential sections.
The turbulence coefficient was calculated from lapse-rate observation
data. The data obtained were distributed over three gradations: 0-0.10,
0.11-0.20 and 0.21-0.3 rn^/sec. A graph analogous to the previous ones was
plotted for these gradations. It showed that the magnitude of the concen-
trations in the surface layer strongly depends on the turbulence coefficient:
the lower the latter, the smaller the sulfur dioxide concentration near the
ground. For the first gradation, the concentration maximum is approximately
one-half of the second, and one-third of the third. As the turbulence coef-
ficient increases, so does the zone of high concentrations.
An adequate network of sampling points in a city makes it possible to
use the data obtained to draw lines of equal mean annual sulfur dioxide
concentrations to identify the areas of minimum and maximum pollution of the
air reservoir> the variations in space and time, the influence of meteoro-
logical conditions on the degree of atmospheric pollution of the city, and
also to solve hygienic problems, take active steps to reduce the emission of
pollutants into the atmosphere, etc.
Fig. 5 shows air sampling points and gives isolines of equal concentra-
tions plotted from observational data for a city.
Of great interest is the study of the distribution of noxious substances
with height in plumes of high-output sources and above a city [6]. To obtain
such characteristics, a helicopter equipped for sampling the main ingredients
(dust, CO, SO2, NO2) was used at the Main Geophysical Observatory.
Fig. 6 shows an example of the distribution of carbon monoxide concen-
trations in a plume from a metallurgical complex and presents temperature
profiles based on radiosonde data.
The data obtained for concentrations above a city at heights of 100-400 m
and sometimes up to 1500 m showed that the concentration field is quite spotty
and that extensive data will be needed for a statistical correlation of the
spatial characteristics.
- 106 -
-------
LITERATURE CITED
1. BcpjifiNA M E., OiiHKy.i P. H. H3imecKa)i ocuoBa pacaMO-
poB Tpynu rrO, Dbin. 207, 1968.
6. TopouiKO B. B , 3 a ii n c b B C.HaaapciikoB H Bonpocu .mcto.iiikii
ii pcsj.ibTaTbi ucc.icaoBaiiiiH 3arpn3iieims aTNioccJicpw c noMombio BcpTO.icTa.
Tp\ju rrO, Bun 234, 1968
7. PiiXTcp B B.rii/ibAeiiCKiio.ib.aP C Pacnpcac.ieiinc npii3CMiibi\ koh-
ueiiTpamn'i ccpimcToro rasa n 30.iu b 3011c Tcn.iOBofi a.ieKTpocTaimini TpyAbi
rrO, Bbin 158, 1964
B B GOROSHKO, E N ZASUKHIN
EXPERIMENTAL INVESTIGATION OF AIR POLLUTION
IN INDUSTRIAL REGIONS*
At present, The Main Geophysical Observatory together with other
institutions carries out a great quantity of experimental works on
studying chemical air content under different meteorological condi-
tions. They are performed in two directions, concentration field deter-
mination in the zone of action of the plume from powerful sources
and pollution level estimation in the atmosphere of towns with highly
developed industry The works are carried on in different seasons,
in different climatic zones, taking into account diurnal variations of
meteorological elements
From 1961, 2 or 3 expeditions are organized every year in the
regions of large industrial sources differing in technology of in-
dustry and emission characteristics, respectively. At the fulfillment of
above-mentioned works a broad complex of observations is carried
out which includes.
(a) measurements of temperature, humidity, wind velocity up to
17 m, and thermal balance and observations of wind direction varia-
tions;
(b) aerological observations of meteorological element distribu-
tion with the height using a captive balloon up to 500m and an air-
craft and/or a helicopter up to 1000m,
(c) sampling under the plume simultaneously in about 20 points
at 10—20 km distances from the source;
* Editor's notei The abstract is presented as given in English with the original Russian article.
- 107 -
-------
(d) measurement of emission parameter —a total volume of ef-
fluent gases, the quantity of noxious substances for each ingredient
emitted into the atmosphere, and temperature of effluent gases.
In the system of Hydrometcoservice works on investigation of air
pollution by industrial emissions are performed in moie than
100 towns by means of organizing stationary points for air samp-
ling. Besides, in some towns specially equipped cars arc used which
take air samples along definite route and under the plumes of main
sources of air pollution For this purpose, methodics of concentration
determination are introduced for most prevailing ingredients: SO2,
NO2, H2S, CO, CI I2, soot and dust With tins, observations of meteo-
rological elements determining pollutant dispersion are carried out.
For estimation of pollutant distribution with height over a town and
in pluinc of single sources, samples are taken on helicopters
Obtained materials on air pollution permit to study surface con-
centration field in dependence on variation of meteorological elements
such as wind velocity, temperature distribution with height, exchange
coefficient, etc., and on this basis to examine theoretical calculations
of air pollution dispersion and to find out zones of maximum pollu-
tion.
- 108 -
-------
FIELD STUDIES OF AIR POLLUTION IN THE AREA OF
THE SKAWINA ELECTRIC POWER PLANT
W. Parczewski PPR (Polish People's Republic)
From Glavnoe Upravlenie Gidrometeorologicheskoy Sluzhby Pn Sovete Ministrov SSSR. (Chief Administration of
the Hydrometeorological Service Under the Council of Ministers of the USSR.)_ "lleteorologisheskie Aspekty
Za'gryazneniya Atraosfery". (Meteorological Aspects of Air Pollution.) Sbornik dokladov na mezhdunarodnom
simpoziume v Leningrade - Iyul' 1968 g. (Reports delivered at the International Symposium in Leningrad -
July 1966.) Pod redaktsiey d-ra fiz.-rajt. nauV: U, E. Berlyanda. , (Edited by Prof. U. E. Berlyand.)
GidrometeorolgichesKoe izdatel'stvo, Leningrad, p. 256-240, (1971), (Hydrometeorological Publishing House,
Leningrad, (l97l).) •
In the south of Poland, in the area of the Skavine (Skawina) electric
power plant, which has an output of 400 MW and is a major source of sulfur
dioxide emissions, field studies were carried out as early as 1958-1959.
However, because of the imperfection of the instruments at that time, it
was not possible to obtain data on the sulfur dioxide content near the
plant, and all that was done was to compare the smoke plume shapes for differ-
ent meteorological conditions [1]. In 1961, the field measurements were
resumed. From May to December, six series of complex meteorological measure-
ments, observations of smoke plume shape and sulfur dioxide content were
carried out by using a special method.
Measurements of sulfur dioxide concentrations and inclination of the
smoke plume were made in accordance with Zielinski's project, and the
remaining measurements, according to the author's. Measurements of sulfur
dioxide content in the surface air layer were made by means of instruments
mounted on two trucks. Whenever it was technically possible, the trucks
drove under the smoke band and performed the measurements.
In view of the fact that at a certain distance from the power plant
stack the smoke became invisible, the actual direction of the smoke plume
and also its variation was reported by meteorologists stationed in the vicinity
of the plant, either by radiotelephone or by radio transmitter. The meteorolo-
gists recorded the variations in direction of the smoke band in detail so that
it was possible to represent on diagrams the directions of propagation of the
smoke plume and the distances at which the sulfur dioxide concentration were
measured, and also the time of the measurements. As a result, it was possible
to compare the position of the trucks carrying the measuring instruments and
the wind direction at the level of the stack orifice. Thanks to the use of
trucks, daily measurements of sulfur dioxide were made in the direction of
propagation of the smoke band, and thus not only average, but also instantan-
eous values of sulfur dioxide concentration were obtained.
Measurements of surface concentrations of sulfur dioxide were carried
out at two mobile measuring points by using EEL-SO2METER British analyzers,
which carried out continuous measurements [3]. These analyzers operate on
- 109 -
-------
the photocolorimetric principle by measuring the potential difference at
two photoelectric cells caused by streams of iodine-starch solution flowing
in front of them, one of the streams being partially formed when the solution
comes in contact with atmospheric air containing sulfur dioxide and being
sucked in. The lower sensitivity limit is around 0.03 mg/m3; the time lag
of the method is 2-2.5 min. Let us note that on the diagrams, the sulfur
dioxide concentrations obtained are sometimes negative. This is due to the
influence of oxidants present in air on the solution.
It was established experimentally [3] that air does not contain a suf-
ficient amount of sulfur dioxide that could be measured by the method used
in cases where the point of measurement is located at a certain distance
from the plume axis. For this reason, if it was impossible to drive up in
the truck directly under the plume, no sulfur dioxide observations were made
on such days, so that only about 75% of the days could be used for the
analysis.
The meteorological observations included: pilot balloon measurements of
the wind to a height of about 500 m at a pilot balloon ascent velocity of
100 m/min, measurements of temperature and humidity of air with an aspiration
psychrometer, observations of the state of weather (clouds, visibility,
atmospheric phenomena) and shapes of the smoke plume, chiefly sketches,
photographs and determination of the ascent height of the plume above the
stack and its inclination.
In the final determination of the influence of meteorological conditions
on the spreading of sulfur dioxide in the lower atmospheric layer, not only
the results of observations and measurements in the area of Skawina, but
also the results of meteorological analysis were taken into consideration:
thermal stratification in the 1-2 kilometer layer of air, determined from
radiosonde data at the nearest station, hourly description of the state of
weather at the nearest synoptic station, and weather maps.
Observations of the ascent and inclination of the smoke plume were
made from a position approximately perpendicular to the actual direction of
propagation of smoke and located at a distance of 800 m from the plume.
The measurements were made with Zielinski's instrument [3], consisting of a
mirror photographic camera with a scale of angles and heights marked on the
ground glass. The camera was mounted on a special tripod head that made it
possible to rotate the camera along the vertical axis and also to adjust it
relative to the horizontal by means of micrometer-type screws. A fixed
lower section of the camera simultaneously took pictures of the smoke band.
The angles and heights were read off directly from the image of the plume on
the ground glass by using the scales marked on it. The measurements were
made every 5-10 minutes. At the same time, outlines of the smoke plume were
drawn and photographed at half-hourly average intervals.
- 110 -
-------
Measurements of the actual height at which the spreading of smoke in
the atmosphere took place showed that it was highly variable. In most
cases, from a practical standpoint, there was no natural increase in the
effective stack height. Since the direction of the smoke plume immediately
after it leaves the stack coincides with the direction of the resultant
vertical smoke velocity and wind velocity at the exit from the stack, it
could be postulated that the absence of a smoke ascent was due to signifi-
cant wind velocities at the stack orifice. Table 1 confirms the validity
of these assumptions. It was found that a marked plume ascent was observed
only in certain ranges of wind velocity at the level of the stack orifice.
Table 1
Wind
Velocity at
Stack
Orifice,
n/sec
Number of Cases
Presence
of Ascent
Absence
of Ascent
0
1-2
3-4
5-10
8
6
1
0
0
2
8
6
The use of this simple and conven-
ient method yielded results of theoretical
and practical interest [2]. Moreover,
this method makes it possible to draw by
analogy a number of conclusions regarding
other electric power plants in Poland,
either those already in existence or
those being planned.
LITERATURE CITED
1. Parczewski W. On the spread of plumes of industrial smoke Acta Geoph.
Polon , t VIII, z 2, Warszawa, 1960
2 Parczewski W. Wplyw warunkow metcorologicznjch na rozprzeslrzenianic
sie gaz6w toksycznych w dolnej warstwie atmosfery Wiad Sluzby Hydrol.
i Meteor, z 59, Warszawa, 1965
3. Z i e 11 n s k i E Zasiarczante atmosfery w otoczemu elektrowni duzej niocy.
Instytut Energetyki, Warszawa, 1962
V PARCHEVSKY
FIELD INVESTIGATIONS OF AIR POLLUTION IN THE REGION Or POWER
STATION IN SKAV1NE*
Power station in Skavine with output of 400 MW is a powerful
source of sulphur dioxide emission. Field investigations of emission
distribution were carried out in 1961. Some results of these works
are given. Observation complex included sulphur dioxide measure-
ments. Sampling was carried out with the help of two cars which
were situated directly under the plume at different distances from
the stack during measurement period Observations of plume form
and rise included its photographing with one camera and sketching
the form of the plume. Pilot-balloon observations and meteorologi-
cal observations in surface air layer were carried out simultaneously
in the station area. Analysis of materials allowed to find out wind
velocity effect on the height of plume rise.
* Editor's note: The abstract is presented as given in English with the original Russian article.
- Ill -
-------
EFFECT OF METEOROLOGICAL CONDITIONS ON AIR POLLUTION
IN CITIES OF THE SOVIET UNION
E. Yu. Bezuglaya and L. R. Son'kin (USSR)
From Glnvnoo Uprovlonio Gidromoteorologicheskoy Sluzhby Pri Sovete J.Iinistrov SSSR. (Chief Administration of
the Hydrometeorological Service Under the Council of Ministers of the USSR.) "Meteorologisheskie Aspekty
Zogryazneniya Atmosfery". (Meteorological Aspects of Air Pollution.) Sbornik dokladov na raezhdunarodnoa
simpoziumo v Lcnirgrado - lyul1 1968 g. (Reports delivered at the International Symposium in Leningrad -
July 1968.) Pod rcdaktsioy d-ra fiz.-mat. nauk U. E. Barlyapda. (Edited by Prof. M. E. Berlyand.)
Gidrometeorplgichoskoe lzdatel'stvo, Leningrad, p. 241-252, (1971). (Hydrometeorological Publishing House,
Leningrad, (l97l)0
In studying the influence of meteorological conditions on air pollution
at the A. I. Voyeykov Main Geophysical Observatory (MGO), particular atten-
tion is given to a statistical analysis of data from observations of impurity
concentrations in air carried out in cities. The analysis is performed on
data of regular observations of pollution over a network of points in the
system of the Hydrometeorological Service. These observations are made on
the concentrations of sulfur dioxide, nitrogen oxides, carbon monoxide, dust,
soot and other impurities in air at different times of day.
Data from impurity concentration measurements and meteorological obser-
vations are tabulated. The elaborated form, same for all cities, specifies
subsequent punching of the data directly from the tables, all the columns of
which correspond exactly to the columns on a punched card. A punched card
file has been collected and is being supplemented. The observational material
is used to study the influence of meteorological conditions on air pollution.
The purpose of our study was to find the dependence of pollutant concen-
trations in air on the set of meteorological elements for specific emissions.
In approaching the solution to this problem, it is expedient to determine the
relationships between air pollution and various meteorological elements. This
point has been discussed in several studies [8, 9 etc.]. Our studies broadly
confirm the conclusions reached by other authors, and also contain some new
results.
Depending on the wind velocity, two maxima of urban air pollution are
observed: one in the presence of a relatively weak wind (0-1 m/sec) and one
at 5-6 m/sec (Fig. 1). The first maximum is obviously due to emissions from
low sources, and the second, to high sources, this being in accord with the
conclusions of theoretical studies. At wind velocities of 2-3 m/sec, a rela-
tive decrease of the pollutant concentrations is observed. An increased
pollution of urban air is observed in surface inversions. At the same time,
elevated inversions, which limit the vertical mixing of impurities, also pose
a danger to citie9 (Table 1).
- 112 -
-------
Table 1
Ratio of Average Concentrations in Air in the Presence of
Surface and Elevated Inversions and in Their Absence
Stratification
Surface
Elevated
City
Pollutant
Inversions in
Inver-
Lower
sions
J km Layer
Alma-Ata
Sulfur Dioxide
1,4
1,3
Nitrogen Oxida
1,3
1,1
Dushanbe
Sulfur Dioxide
1,1
1,0
Nitrogen Oxide
1,5
1,4
Moscow
Dust
1,4
—
Sulfur Dioxide
1,1
1,3
Tallin
Dust
1,4
1,3
Carbon Monoxide
1,6
1,5
Tbilisi
Dust
1,3
1,4
Soot
1,9
1.8
The dependence of pollutant concentrations on the wind velocity is
different in surface and elevated inversions (Fig. 2). When surface inver-
sions are present, the basic air pollution maximum is observed at weak wind
velocities (0-1 m/sec), and in the case of elevated inversions, at a wind of
about 5 m/sec. The latter case, however, is due to the influence of elevated
inversions on the pollution of air with high emissions at the dangerous wind
velocity [3].
Precipitation purifies the air by removing the pollutants. After Che
latter have deposited, the urban air is comparatively pure for a period of
time (Table 2).
Table 2
Frequency of Pollutant Concentrations
After Precipitation Relative to Average
Value (Leningrad)
Table 5
Frequency (#) of Increased Pollutant
Concentrations as a Function of Com-
bination of Meteorological Parameters
Pollutants
Time Between End of
Precipitation and
Collection of Samples,
v, m/sec
AT0
0-1
0-12
>12
<0
>1
Dust
0.29
0,77
1,30
<10
27
10
Sulfur Dioxide
0.26
0,85
1,26
>10
21
20
Precipitation decreases the background pollutant concentration in the
city, produced outside the range of influence of industrial emissions.
- 113 -
-------
No purification of air by precipitation is observed when the pollutants
are carried directly from the sources.
3%
30
15
20
ts
10
0123&56V m/seo
Fig. 1. Frequency of increased dust
concentration m air as a function of
wind velocity.
1 - Leningrad, 2 - Moscow
f"m/sec
Fig. 2. Carbon monoxide concentrations in
air as functions of mind velocity.
1 - surface inversion, 2 - elevated inversion
The complex influence of several meteorological factors on air pollution
may be examined with the aid of synoptic situations reflecting the whole
gamut of processes taking place in the atmosphere. However, an objective
physical evaluation of the situations, given their considerable variability,
presents some major difficulties. One can examine only the most distinct
synoptic situations, which are comparatively stable in time. Of particular
importance is the situation of a stationary anticyclone, associated with
heavy air pollution in cities. An effective removal of pollutants from the
air takes place in developed cyclones.
At the present time, a statistical method is being developed for objective
consideration of the influence of a variety of meteorological parameters on
air pollution. The method consists in using the initial meteorological para-
meters to separate the regions in multivariate space where similar magnitudes
of air pollution concentrate. At first, from the group of parameters only
those are selected which are statistically significant. Analysis of data
from observations at one stationary point in Moscow revealed two statistically
significant factors: AT — the temperature difference between the 1 km level
and the ground, and v — the wind speed at the 1 km height. The first attempt
- 114 -
-------
to solve the problem in this manner failed to yield unambiguous results.
However, it was possible to separate regions of different frequencies of
high pollutant concentrations (Table 3).
It is evident from the table that the greatest probability of high
pollutant concentrations is observed in an inversion associated with weak
winds at the 1 km height. It is characteristic that in an unstable strati-
fication, the frequency of high concentrations increases with increasing wind
at the 1 km height. In the presence of a strong wind in the atmospheric
boundary layer, the air pollution is practically independent of the stability.
The problem of a unique determination of the degree of air pollution on
the basis of initial meteorological parameters may evidently be solved by
substantially averaging the pollutant concentration values over space and
time.
Analysis of the observational material showed that there is a tendency
toward a simultaneous variation of air pollution in different parts of the
city. Observational data on concentrations of nitrogen oxides and sulfur
dioxide in air at seven stationary points in Sverdlovsk were used to calculate
the space coefficient of the linear correlation between their values at dif-
ferent points. Fig. 3 shows the field of coefficients of correlation when
the reference point was taken at the center of the city. It is evident from
the figure that the isocorrelates are in the form of concentric circles, and
the correlation gradually decreases with increasing distance from the refer-
ence point.
Based on material for 1967, a special analysis of periods of increased
air pollution and corresponding meteorological conditions was carried out.
An objective method was worked out for separating such periods. It consisted
in the fact that the observational data for the individual stationary points
for each pollutant were used to select a given number of cases with highest
concentrations during a month, a season, or a longer period of time. The
arbitrary criterion taken was a value of the concentration 1.5 times greater
than the average concentration at the given point for a given pollutant. If
in the course of a given period of time (one or several days), m is the number
of cases of increased concentration and n is the total number of observations
of the concentrations of all the measured pollutants at all the stationary
points of the city during the given period of time, the degree of increased
simultaneous air pollution in the city will be given by the quantity P = £1,
which ranges from 0 to 1. n
Values of P were calculated for each day of January and July 1967 from
data of observations of air pollution in a number of cities of the country.
On the average, PftjO.2. For each city, we selected five days with the heaviest
air pollution during a month. During these days, the P values were essentially
0.3-0.5, and climbed to 0.6-0.8 only during certain periods. Periods of
- 115 -
-------
\
0.005 j
/
\
— 1
—2
Fig. 3. Isocorrelates of pollutant concentrations
in the city with the reference point at the center.
1 - nitrogen oxides (numerator), 2 - sulfur dioxide
(denominator)
heaviest air pollution are observed most frequently in connection with the
formation of stationary anticyclones. It is characteristic that a general
increased air pollution was observed simultaneously in several cities
located in the zone of a stationary anticyclone. Meteorological analysis of
the separated periods requires a comparison of the meteorological parameters
with the corresponding characteristics on days with reduced air pollution.
Periods with a general reduction of air pollution are distinguished fairly
easily, and their nonrandomness is demonstrated with the aid of statistical
criteria. For each city, 5 days were selected with the lowest air pollution
(smallest P values) in the course of a month.
In connection with the above, it was found convenient to distinguish
three types of periods with different states of urban air pollution:
Type I - period of heavy general air pollution in the city, P > 0.5.
Type II - period of increased air pollution in the city with P = 0.3-0.5.
Type III - period of reduced air pollution in the city.
Averaged meteorological data during periods of increased and reduced air
pollution are listed in Table 4.
- 116 -
-------
Table b
Average Characteristics of Meteorological Parameters for Different States of
Air Pollution in Cities in 1967.
Types of Period
January
July
Wind Velocity, .
m/sec
Cloudiness,
Points
Difference of
Temperatures Near
the Ground and at
500 m Level in
the Daytime
Wind Velocity,,
ra/seo
Cloudiness,
Points
Difference of
Temperatures Near
the Ground and at
500 m Leyel in
the Daytime
Ground
500'm
Level
Ground
500 m '
Level
I
0.5
5,4
3,5
-3,8
II
2,3
6,2
6,8
-3,2
2,7
5,8
7,5
4,4
111
3,9
11.8
7,1
+0,9
3,1
6,5
7,1
8,8
Analysis of Table 4 leads to the following conclusion: an increased
air pollution in winter over a considerable territory of the city is simul-
taneously observed in the presence of a weak wind near the ground and at a
level of 500 m, with a slight cloudiness and intense inversions persisting
in the daytime. The combination of such conditions is characteristic of a
stagnating air situation.
In the summertime, practically no periods of type I were observed.
In July, there were no appreciable differences in wind velocity during periods
of increased and reduced air pollution. The most distinct difference was
found in the temperature lapse rate of the lower 500 meter layer; for a
reduced air pollution, a pronounced air instability was observed.
On the whole, analysis of air pollution in different cities showed the
presence of two clear-cut dangerous meteorological situations: stagnation
of air, in which a simultaneous increase of pollutant concentrations is
observed in different parts of the city, and a stable wind from the direction
of the chief emission sources.
In most cities, the distribution of the air pollution sources over the
territory is comparatively uniform, and here the greatest danger is posed
by stagnation of air. In areas where a substantial number of industrial
plants are concentrated in a given part of the city, both of the indicated
situations are dangerous. Finally, there are many cities whose industrial
plants are located beyond their boundaries. Obviously, a dangerous air pol-
lution may be produced in this case only when the wind is directed from the
plants toward the city.
In the near future, it will be necessary to develop a procedure for
forecasting heavy air pollution. The usefulness of information on unfavorable
- 117 -
-------
meteorological conditions for the same day is most obvious when the industrial
plants are located beyond the city limits and the increased content of noxious
pollutants in the atmosphere is largely due to the wind direction. Initial
experience is available on taking the weather conditions into consideration
in order to prevent heavy air pollution without decreasing the production.
When unfavorable wind directions are forecast, some industrial plants do not
disconnect or repair their filters, and no forcing of the production is
allowed.
The influence of meteorological conditions on air pollution manifests
itself in the analysis of the annual variation of pollutant concentrations.
Results of a comparison of the annual variation of many pollutants with cor-
responding variations of meteorological elements made it possible to establish
a difference in the character of the influence of meteorological conditions
on individual pollutants. The relationship of the annual variation of carbon
monoxide and nitrogen oxides to the frequency of a wind velocity of 0-1 m/sec
was identified most clearly (Fig. 4). This is due to the fact that an impor-
tant producer of these pollutants is motor transport, which discharges its
emissions almost at ground level. The annual variation of the dust concen-
tration has its own characteristic features, since it is due not only to
industrial emissions, but also to the natural dust content. The latter is an
additional cause of increased dust pollution during the warm half of the year
as compared with the cold half (Fig. 5).
The spring dust pollution maximum noted in [2, 6] is evidently determined
not only by the increased frequency of anticyclonic circulation, but also by
the state of the soil during that season. In many cities, an increase in
sulfur dioxide concentration during the cold half of the year is observed as
a result of increased fuel combustion (Fig. 6a). At the same time, in major
industrial centers there takes place a summer maximum of the sulfur dioxide
content of air in the presence of increased turbulent exchange (Fig. 6b).
In order to identify the regions with the most unfavorable climatic con-
ditions conducive to accumulation of pollutants in air, it is necessary to
•evaluate the air pollution potential. Such a potential represents a combina-
tion of weather conditions determining the possible air pollution level. A
frequent combination of meteorological conditions conducive to accumulation
of pollutants in the surface layer indicates a high air pollution potential.
A low potential exists in areas with a high frequency of meteorological con-
ditions unfavorable to the creation of high surface concentrations.
The chief meteorological factors used for estimating the air pollution
potential are determined by their influence on the process of transport and
scattering of pollutants (velocity and direction of wind, vertical thermal
stratification characteristic, fogs), on the regulation of the amount of
consumed fuel (temperature of air), removal of pollutants from the atmosphere
(precipitation), formation of photochemical smog (solar radiation), etc.
- 118 -
-------
p %
0 L 0
II IV VI VIII X XII
II IV VI I'III X X//
Fig. 4. Annual variation of carbon monoxide concentration (l)
and of the frequency of weak wind velocities (2).
a - Petropavlovsk, b - Lipetsk, o - Magadan, d - Irkutsk,
e - Chelyabinsk, f - Kursk.
a)
p' ' i ' ' '
II IV Yl VIII X XII
b)
' i ' ' i ' J ' ' ¦ i i
II IV Yl Ylll X XII
Fig. 5. Annual variation of the dust content of air.
a - Vil'nyus, b - Magnitogorsk
- 119 -
-------
Studies have now been initiated at the Main Geophysical Observatory
to evaluate the air pollution potential over the territory of the Soviet
Union. The frequency of cases in which the wind velocity did not exceed
1 m/sec (conditions of air stagnation) was taken as one of the first charac-
teristics of the air pollution potential. Data of meteorological observa-
tions for 5 years were used to plot a map of the frequency of days with wind
velocities below 1 m/sec in January, April, July, and October (Fig. 7). As
a result of the analysis of the maps, it was possible to distinguish the
zones where the number of days with wind velocities of 0-1 m/sec in the
course of 24 hours was less than 1 day (IV), from 1 to 5 days (II, III) and
over 5 days (I).
Fig. 7. Regions with different numbers of days with wind velocities of
0-1 m/sec in the course of 2k hours (areas not included in the regional-
ization are crosshatched).
- 120 -
-------
To estimate the pollution potential, it is useful to have data on the
frequency, thickness and intensity of surface inversions, and also on the
intensity of turbulent exchange on the territory of the Soviet Union given
in [1, 2, 4, 5].
The results obtained are still insufficient for a detailed description
of the air pollution potential in the country, but they have led to certain
conclusions, some of which do not flow directly from knowledge of climatic
conditions. The highest potential is observed in Eastern Siberia and the
Far East, excluding the Pacific Ocean coast, and also in central Asia, where
the danger of accumulation of pollutants in the air is compounded by a strong
natural dust pollution. A region of appreciable air pollution potential is
observed in the area of the Urals. It is characteristic that a tendency
toward an increased potential is also noted in the western and northwestern
regions of the European territory of the country, this being due to a high
frequency of weak winds, primarily during the warm part of the year.
Consideration of the potential in the construction of new cities and
industrial plants with the introduction of steps aimed at ensuring the
purity of the atmosphere, and also the prevention of a dangerous buildup of
noxious emissions in the air under unfavorable weather conditions may promote
a decrease in the level of air pollution.
LITERATURE CITED
1 Beayr.iaii 3 KD K onpeae.ieHmo noiciimia.na 3arpn3iieiiiin Bi» rnapoMeTCO»3flaT,/1, 1966
3. Bep/iflHA M. E, OiiHKy/i P H. (})iimieHTc TypSy.ieiiTrforo oSmciib b npnacMiioM c.ioc
zieTOM b AMCBiioe BpeMfl b pa3;ininbix reorpatj^mecKiix pafioiiax CCCP Tpyau
ITO, Bbin 234, 1968
5. K a c b n h H A , O r n e b a T. A, Tepeiosa K M. 3aKonoMepHocin mMciie-
niia TeMnepaiypnoro rpafliieiua b npiraeMHou c^oe B03Ayxa Ha TeppiiTopnir
CCCP Tpyau rrO, Bbin 234, 1968
6. C o h b k ii h JI P, Paj6eracBa E A, T c p e x o b a K. M K Bonpocy o we-
Tcopo.ioriiiecKoii o6yc/ioB.nciMiocTn 3arpH3Heiiiifl B03Ayxa Haa ropoaaMH Tpyabt
rrO. Bun 185. 1966
7. C o h b k ii ii /I. P., M a n h k o b B 06 o6pa6oiKe ii aiia.ni3e Hae.noaennii 33
3arpn3HcnncM B03;iyxa b ropoaax Tpynbi ITO, Bbin 207, 1968
8. Dickson R R Meteorological factors affecting particulate air pollution of a
city BAMS, v 42. No 8. 1961
9 AUCormitk R. A Air Pollution Climatology. Air Poll. Academic Press.
New York, No. I, 1967
- 121 -
-------
E Yu BEZUGLA YA, L R S0NK1N
THE EFFECT OF METEOROLOGICAL FACTORS ON AIR POLLUTION
IN THE CITIES OF THE SOVIET UNION*
The effects of meteorological factors on air pollution are consi-
dered on observation mateirals concerning contents of pollutants in
the air in some towns of the USSR It is shown that depending ori
wind velocity two concentration maxima of pollutants are discove-
red in the air, one at wind velocity of 0—1 m/sec, the other at 5—
6 m/sec For all this, the concentration maximum at wind velocity of
0—1 m/sec is expressed more clearly in the conditions of ground in-
version, the maximum at wind velocity of 5—6 m/sec is seen in other
cases On the average, the air in town is more polluted in the pre-
sence of ground and low-lying elevated inversion, than in the absence
.if inversion in boundary layer.
An analysis of correlation coefficients between concentration va-
lues of air pollutants in different points ol a town showed the pre-
sence of a trend to simultaneous change in air pollution The cases
of air pollution at high level in the whole town are considered. They
are chiefly connected with light winds in boundary layer and stable
stratification
A parallel course of curves of pollution concentration in the air
and liglU wind frequency is found in the annual course. Foi all this,
the above-mentioned connection is expressed more clearly for carbon
and nitrogen oxides which are emitted in large quantities by cars,
that is in the lowest air lajer.
The obtained results allowed to make preliminary conclusions
on air pollution potential distribution on the territory of the USSR.
For this purpose, data on frequency of light winds and stationary an-
ticyclones are used The zones of the highest potential are found in
Eastern Siberia, Middle Asia, and in the regions of Urals.
The results of these studies are proposed to be used for the ela-
boration of air pollution forecast methods.
* Editor's note: The abstract is presented as given in English with the original Russian article.
- 122 -
------- |