JPRS ID: 8998 USSR REPORT ENERGY
Document Type:
Collection:
Document Number (FOIA) /ESDN (CREST):
CIARDP8200850R0002000300394
Release Decision:
RIF
Original Classification:
U
Document Page Count:
183
Document Creation Date:
November 1, 2016
Sequence Number:
39
Case Number:
Content Type:
REPORTS
File:
Attachment  Size 

CIARDP8200850R0002000300394.pdf  9.54 MB 
Body:
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
19 OECEMBER 1979
VKI
ANo
N0. 9. SEPTEMBER
. Ot3Y
1979
i OF 2
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
JPRS L/8823 _
19 December 1979
USSR Report
METEOROLOGY AND HYDROLOGY
No. 9, September 1979
= FB~$ FOREIGN BRG,qDCAST iNFORMATION SERVICE _
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
NOTE
JPRS publications contain infarmation primar4ly from foreign
 newspapers, periodicals and books, but also from news agency
 transmissions and broadcasts. Materials from foreignlanguage
sources are translated; thuse from Englishlanguage sources
, are transcribed or reprinted, with the original phrasing and
other characteristics retained.
Headlines, editorial reports, and material enclosed in brackets
are supplied by JPRS. Processing indicators such as [Text]
_ or [Excerpt] in the first line of each item, or following the
last line of a brief, indicate how the original informa.tion was
processed. Where no processing indicator is given, the infor
mation was summarized or extracted.
Unfamiliar names rendered phonetically or transliterated are
enclosed in parentheses. Words or names preceded by a ques
tion mark and enclosed in parentheses were not clear in the
original but have been supplied as appropriate in context.
Other unattributed parenthetical notes within the body of an
item originate with the source. Times within items are as
given by source.
The contents of this publication in no way represent the poli
cies, views or attitudes of the U.S. Government.
For further information on report content
call (703) 3512938 (economic); 3468
(political, sociological, military); 2726
(life sciences); 2725 (physical. sciences).
COPYRIGHT LAWS AND REGULATIONS GOVERNING OWNERSHIP OF
MATERIALS REPRODUCED HEREIN REQUIRE THAT DISSEMINATION 
OE' THIS PUBLICATION BE RESTRICTED FOR OFFICIAL USE ONLY. =
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE UNLY
JPRS L/8823
19 December 1979
 USSR REPORT
METEOROLOGY AND HYDROLOGY
No. 9, September 1979
Selected articles from the Russianl.anguage journal. METEOROLOGIYA
 I GIDROLOGIYA, MosCOw. _
CONTENTS
PAGE
 ShortRange Forecasting of Air Temperature, Continuous Precipitation and
Wind an the Basis of Prognostic Pressure Charts
_ (A. I. Snitkovskiy)
1
Allowance for Orography in Numerical Weather Forecasting Models
~ (A. I. Romov)
16
_ The Upper Boundary Condition in the Problem of Numerical Forecasting of
_ Meteorological Elements
(V. A. Gordin, B. K. Domatov)
26
Vertical Currents in the Troposphere
(V. A. Shnaydman)
37
Influence of SmallScale Turbulence on Clearing of an Aqueous Aerosol
(S. D. Pinchuk)
48
Computation of Atmospheric Propagation of Effluent of High Industrial
Sources in the Presence of Inversions Aloft
 (F. A. Gisina, S. M. Ponomareva) .e................
54
Economic Effectiveness of Meteorological Support of Civil Aviation
_ (E. I. Monokrovich)
64
Subsurface Salinity Maximum in the Actiti*e Layer of the Ocean and
=
Convective Penetrations
(A. A. Kutalo, Ye. B. Chernyavskiy)
71
~ Variability of Water Salinity in the Coastal Zone of the Sea
77
 a LIII  USSR  33 5& T
FOUO] EL
FOR OFFIC:LAL U5E ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
CONTENTS (Continued)
P age
Computation of Parameters of Resonance Frontal Waves
(G. G. Kuz'minskaya, T. I. Tsareva) 84
Variability in Snow Distribution on Ice in the Arctic Ocean
(A. Ya. Buzuyev, et al.) 90
Checking of Statistical Hypotheses in Computations of Maximum Water ,
Discharges with a Low Guaranteed Probability of Occurrence
(A. V. Khristoforov) .............................o.... 103
Frequency of Recurrence of Dust Storms in the Territory of the USSR
(L. V. Klimenko, L. A. Moskaleva) 112
Modeling of the Process of Forming of the Yield of Winter Wneat
(M. S. Kulik, et al.) 119
Time Validity of Meteorological Information
_ (G. P. Lutsenko, V. D. Nikolayev) .....s 131
Relationship Between the Planetary HighAltitude Frontal Zone and the
Position of the Snow Cover Boundary During the Autumn and Spring
Pariods
(V. B. Afanas'yeva, et al.) 137
Coastal Stationary WaveMeasuring Complex
' (V. B. Vaysband, V. N. Shanin) 142 
Investigation of Correlation Between the Nature of a Radar Signal
j Envelope and the Form of the Sea Surface Reflecting Surface
(I. V. Kireyev, A. V. Svechnikov) .........t.......... 149 
Review of "General Circulation Models of the Atmosphere. Methods in "
Computational Physics." Academic Press, New York  San Francisco 
London, Vol 17, 1977
(S. A. Mashkovich) 156
Review of Monograph by Kh. G. Tooming: SOLNECHNAYA RADIATSIYA I 
FORMIROVANIYE UROZHAYA (Solar Radiation and Yield Formation),
_ Leningrad, Gidrometeoizdat, 1977
(I. A. Shul'gin, I. A. Murey) 160
Sixtieth Birthday of Semen Pavlovich Kosnov 164 
Sixtieth Birthday of Nikolay Yefimovich Zak.harchenko 167
_ Conferences, Meetings and Seminars
(Ts. I. Bobovnikova) 170 
Notes From Abroad
(B. I. Silkin) 174 _
 b 
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
.
PUBZICATION DATA
~ English title : METEOROLOGY AND HYDROLOGY
Russian title : METEOROLOGIYA I G IDRCLOGIYA
Author (s) :
Editor (s) . E. I. Tolstikov
Publishing House : Gidrometeoizdat
Place of Publication � Moscow
Date of Publication : September 1979
Signed to press ~ : 21 Aug 79
Copies : 3870
COPYRIGHT : "Meteorologiya i gidrologiya", 1979
 c 
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
~
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USL ONLY
UDC 551.509.(322+323+324.2)
SHORTRANGE FORECASTING OF AIR TEMPERATURE, CONTINUOUS PRECIPITATION AND _
WIND JN THE BASIS OF PROGNOSTIC PRESSURE CHARTS
Moscow METEOROLOGIYA I GIDROLOGIYA in Russian No 9, Sep 79 pp 515
[Article by Candidate of Geographical Sciences A. I. Snitkovskiy, USSR Hy
drometeorological Scientific Research Center, submitted for publication
3 April 1979]
Abstract: The author investigates the possibil
ities of shortrange forecasting of temperature,
continuous precipitation and wind using regres 
sion analysis in accordance with the PP and MOS
concepts. It is shown that for predicting weather
phenomena and elements for 24 and 36 hours in ad
_ vance it is preferable to write prognostic depen
_ dences on the basis of the MOS concept than on the
 basis of the PP concept, which at the present time 
 is being used in operational synoptic practice. ,
[Text] As is well known, in the operational practice of shortrange forecast
ing of weather phenomena and elements extensive use is made of the prognos
tic pressure and geopotential, on the basis of which advective temperature
values, humidity and wind are determined at the initial points on the trajec
tories. On the basis of these data and observations at the forecasting point
there is computation of virtually all weather phenomena and elements includ
ed in the text of the forecast.
 Due to the absence of sufficiently reliable hydrodynamic models of weather
forecasts of phenomena and elements, in operational weather forerasting use
is made of different empirical dependences between the predicted phenomenon
and tne state of the atmosphere. In the USSR these dependences have long
been based on factual (diagnostic) observational data, proceeding on the
basis of existing physical concepts concerning the development of weather
phenomena. In the case of a real forecast for 24 and 36 hours these rela
tionships use prognostic parameters found from future charts, whereas in
the case of a forecast for the current day, morning observational data are
extrapolated for 1218 hours in advance. Such an approach in the foreign
1
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY .
literature [10] has been given the name "perfect prognosis" (PP). Recently
in the United States, Great Britain and a number of other countries, in
addition to the PP concept, in the prediction of weather phenomena and ele
ments extensive use has been made of the MOS concept (model output statis
tics) [10, 121. The MOS concept makes use of the relationships found only
in prognostic data '1216]. The quality of forecasts of weather phenomena
and elements on the basis of the MOS system is the greater the higher the
quality of the prognostic models and the greater is the completeness and
diversity of the meteorological elements obtained from models of ineteoro
logical elements and the greater is the archives of prognostic hydrodynamic
fields, which makes it possible to write dependences for different fore
casting points and different seasons of the year. In accordance with [10],
the quality of the temperature and precipitation forecasts under the MOS
system in the United States, beginning with an advance period of 24 hours
or more, is higher than for the forecasts prepared by weathermen. As a
rule, for writing the relationships for the MOS system the approach of
multiple linear regression is employed.
In this paper we will examine the possibilities of shortrange forecasting
of minimum and maximum temperatures, continuous precipitation and wind for
24 and 36 hours for lAoscow atld Moskovskaya Oblast, making use of two ap
proaches to forecasting PP and MOS.
At the present time it is difficult to create an MOS system superior in
its indices to that created in the United States. 'I'here are many reasons 
for this and the most important of them are: a lower quality of the prog 
nostic riydrodynamic schemes for predicting pressure and geopotential,
whose relative error for 24 and 36 hours is 0.670.88 [ll]; absence of 
numerical schemes for the prediction of temperature and humidity in the
 troposphere and schemes for predicting meteorological elements in the
~ boundary layer; absence of archives of prognostic fields on machine carriers. _
Accordingly, in the paper an attempt has been made to test the possibilities
of shortrange prediction of weather phenomena and elements on the basis of
the MOS system by those means which were at our disposal. These are the ar
chives of prognostic pressure fields at the earth's surface prepared manual
ly by weathermen, the archives of prognostic maps of geopotential and ver
tical air movements in the troposphere based on the S. L. Belousov hydrody
namic scheme currently used in operational wurk and also a number of algo
rithms for the statistical processing of data.
Statistical Processing of Data
The statistical processing of data included the writing of multiple regres
sion equations, paired correlation matrices and screening of predictors for
inclusion in the prognostic scheme and evaluation of the regression equa
tions using the absolute ) and relative (E ) errors, the correlation
coefficient (r), the N. A. Bagrov reliability test (H) and the A. M. Obu
khov accuracy test [6].
2
FC1R OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
The multiple regression equations were written using the algorithm descriu
ed in [4, 51, having definite differences from the standard programs for
writing linear multiple regression equations employed in the mathematical
support for the electronic computer. The regression employed in this study
. is not dependent an the function of distribution of variables and on the
length of the sample. The basis for the idea was minir.ization of the mean
minimum risk of the regression equation. The regression coefficients are
computed for variables first transformed by means of othogonalization using
the adopted base method [7, 8]. This makes it poss'Lble to reduce the dimen
sionality of initial criterial space. rn order to evaluate the regression
equation "from below" use is made of empirical risk ( (Yer) the mean
square error of the "teaching" equation. In this case there is a comparison
of the predictant sigma ( dy) in the teaching; sample with the empirical risk.
The greater the ratio ay/ aer, the better is the "climatic forecast" for the
particular regression sample. The evaluation "from above" is the mean mini
mum risk ( U'mmr) the theoretical mean square regression error for the
 general set. The greater the o'y/ ommr ratio, the better is the "climatic"
forecast made using the general set regression. As a rule, in wellselected
equations there is satisfaction of the expression
O'er 'Yex'~ dmmr< C"y. (1)
where eX is the mean square error of the equation for the "examination"
sample.
The screening of predictors for inclusion in the prognostic equation was
carried out in two ways. First, by means of comparison of the Ommr values
_ for different groups of predictors. The lesser the o"mmr value, the better
do the.predictors selected in the equation describe the considered weather
phenomenon. Second, having paired correlation coefficients, standard devia
tions and the nonnormalized regression coefficients, in accordance with
_ [lJ it was possible to find.the contribution of the predictors to the dis
persion of the predictant and the contribution of each predictor to the
predictant.
Assume that rpl are the paired correlation coefficients between the pre
dictant and the predictors, UQ is the standard deviation of the predic
tant, 01 is the standard deviation of the predictors, apt are the nonnor
malized regression coefficients. Then the contribution of an individual
predictor will be
, xt  roi ao i,
(2)
where ixol = apei/oO are the coefficients of a normalized regression, and
the fraction of each predictor in the determination of the predictant is
3x, _l R~l. (3)
~ Here R is the multiple regression coefficient and R2 is that contribution
to the dispersion of the predictant which is given by the predictors.
~
~
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/48: CIARDP8200850R0002000344394
FOR OFFICIAL USE ONLY
 In this case L
Ra JAxtI,
and
c
~ Ex,= l.
On the basis of the expression from [1]
oxt>2 R
(4)
(5)
(6)
where (7R = 1 R2/ n 2(here n is the length of the sample) we can as
sert that if for a particular predictor the inequality (6) is satisfied,
the use of this predictor in the multiple regression equacion is feasible.
Initial Data
At our disposal we had prognostic pressure, geopotential and vertical air
movements charts for 19761978 and also the advective temperature and ~
humidity values at the earth and at the standard isobaric surfaces obtain 
ed using these charts. Accordingly, the selection of actual data on tem
perature, the qua.ntity of continuous precipitation and wind was made for
the three mentioned years. The dates of the investi gated cases were
selected from a table of random numbers.
Temperature. As the initial data on the minimum and maximum temgsratures
we used tb.eir values averaged for 7 stations in Moscow and 20 stations
in the Moscow area. For predicting the minimum and maximum temperatures
we used the MOS concept, since it is clear from physical considerations
and in accordance with [2, 9] that the PP concept could not be applied to
the forecasting of temperature. The prognostic predictors were selected
for 24 and 36 hours in advance respectively.
 The prognostic (at the initial points of the trajectories) predictors used
in predicting minimum temperature were: T3 temperature at 0300 hours, in
�C; T850 temperature at the 850mb level; Tmin min.imum temperature;
Tmax maximum temperature (on the I ay before); (T  Td)850 dewpoint
spread at the 850mb level; (T  Td)g50 dewpoitit spread at the 850mb
level at the final point on the trajectory; u and v prognostic components
of uind velocity at the earth, in m/sec; u850 and v850 prognostic compon
 ents of wind velocity at the 850mb level, in m/sec; a is longitude of the
day, in minutes.
The prognostic (at the initial points of the trajectories) predictors for
predicting maximum temperature were: T~ prognostic minimum tempera
ture, obtained using the regression equaPion; Tmax; Tmin; T850; (T  Td)8503
T. fP ; u and v; u850 and v850
min
4
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02148: CIARDP8200850R0402400300394
FOR OFFICIAL USE ONLY
Table 1 gives evaluations of that group of predictors which in the examin
ation sample gave better evaluations of the regression equation.
Table 1
Evaluation of the Predictors Used for Predicting Spring Temperature in Moscow
and Moskovskaya Oblast
flpor
1ioa Teppliropxa i7peAxxT0pW
2
9 TninlMocKea 11 IT& Taw, Tm1n, Tmaa.
V.
MocKOSCxasr12
o6nacrb
10 Tmax MoCNea 11
MocxaBCx42
o6nacrb
KEY :
To ;xe 13
dt Bxnalt npe,~hn 5 6 7 8
3 ~ TOpOB II,R~~ % c9j1 1BK7 QCNii QY
 4
75 25, 29. 23, 2,24 2,21 2,61 6,1'
20. 1,2
73 20, 28. 24, 2.16 2,44 2.77 6.OK
19. 1,2
Tm"in, T9r, Tmin� 7maa. 82 27, 20, 23, 9,48 3,50 5,23 3.03
v, X 21, 8,1
70 )xe 13 77 28. 22. 24, 4,27 3,42 4,98 8.0
22, 2,2
1. Prediction 6. ex
2. Territo.Ly 7. mmr
3. Predictors 8. y
4. ContYibution of predictors to 9. T~n
R2 ~ % 10. Tgax
_ 5. er 11. Moscow
12. Moskovskaya Oblast
13. Same
. An analysis of the matrices of paired correlation between the minimum and
maximum temperatures and the p redictors indicates a high correlation be
~ tween the predictants and predictors reflecting the temperature peculiar
. ities of the boundary layer and also the longitude of the day (r = 0.60
0.85). At the same time there is a weak relationship between temperature
and the dew point spread at the 850mb level at the initial and final
points of the trajectories, which, as is well known, characterizes the
influence of cloud cover on air temperature at the earth, which proved to
, be erroneous at first glance. However, with a further examination of this
fact it was found that since for prediction of T~n and Tmax~ am�ng other
predictors, use is made of the advective values of the minimum and maximum
temperatures at the initial points of the trajectories, in whose values the
influence of cloud cover was also reflected, therefore the dewpoint spread
at the 850mb level in itself exerts no appreciable influence on the prog
nostic values of minimum and maximum temperatures.
k
5
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
It is known from synoptic experience how imgortant it is, in predicting
temperature, to estimate from what direction of the horizon the transfer
of air masses will occur. It was found that the meridional components
of the wind, especially at the earth's surface, correlate better with the
. predicted minimum and maximum temperatures than the zonal wind components.
_ The data on the explicable fraction of dispersion of minimum temperature
(R2) foc Moscow is somewhat greater than for Moskovskaya Oblast, which is
_ indicated by the known fact that there is a greater scatter of minimum tem
_ perature in the oblast than in the city.
The values of the different mean squarz errors (Table 1) for the most part
conform to expression (1), which indicates the legitimacy of using the
selected predictors in the regression equation.
For the purpose of taking into account the climatic and circulation pecul
iarities of the regression equation, for predicting the minimum and maxi
mum temperatures in Moscow and Moskovskaya Oblast for 24 and 36 hours in
advance we made separate determinations for each season of the year. The
archives of initial data for each of the seasons consisted of 138 cases.
Continuous precip itation. The prediction ot continuous precipitation for
24 hours in advance was examined. The initial data were taken fo r the cold
season of the year from October through March. The predictant was the quan
tity of precipitation averaged for 7 stations in Moscow and 20 stations in
Moskovskaya Oblas t for the 'times of day from 2100 to 0900 hours.
A study was made of a total of 456 cases: 330 with precipitation and 126
without precipitation. The distribution of precipitation by gradations
was as follows:
Without
Gradations, mm precipitation 0.00.3 0.43.0 3.1 10.0 >10.1
 Numlier of cases 126(27.6) 183(40.1) 113(24.8) 30(6.6) 4(0.9)
(in parentheses,
Inf.ormation on precipitation of different gradations, as shown above, indi
cates the difficulty of prediction of both considerable continuous precip
itation and the absence of precipitation. These same complexities arose
in predicting the quantity of continuous precipitation in the United States
[12]. ~
The mentioned dis tribution of precipitation by gradations naturally found
its reflectioti in the content of synoptic forecasts of precipitation. For
example, the number of forecasts with precipitation (in the oveYwhelming
majority small) was 2530% greater than was observed, whereas the number
of predictions without precipitat:ton was 2530% fewer than the actual number
of cases of absence of precipitation.
b
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
As is well known, statistical methoda for predicting weather phenomena and
elements give better results when from the complex of different synoptic
situations"the predominanting situations are selected, for each of which a
_ decisive statistical ruJ.e is formulated. If the predominating situation
falls in a wide range of values, then, taking into account the possible
 e rrors in prognostic fields, the forecaster can determine this situation
with a high degree of reliability.
_ An analysis of the selected cases indicated that of the 330 cases with pre
cipitation, in 256 cases the precipitation was observed when the wind direc
tion at the 500mb surface was from 180 to 360�. For this reason fo r the
investigation we selected a situation when the wind direction at the 500mb
level was from the western side of the horizon. As an absence of precipita
tion we used a case when the quantity of precipitation was 0.0 mm and when
 it was nat observed.
The prediction of continuous precipitation by statistical methods is ex
tremely difficult and therefore about 30 predictors were subjected to
screening. It was found that the values of surface pressure, geopotential
at the levels 850, 700 and 500 mb, temperature and dew point at the earth
and at the 850mb level and a number of other predictors have a weak correl
ation with precipitation. As a result of many numerical experiments we
selected 10 predictors whose correlation with precipitation was satisfactory
and good. These were: c500 Wind velocity at the 500mb level, m/sec; w700
ordered vertical air movements at the 700mb level, mb/12 hours; k is a
coefficient equal to (H1000+H500)2H850; 'Z (T  Td) is the total dew point
deficit at the levels 800 and 750 mb; qpW is the quantity of precipitable
water in mm; ~in is the fact of initial precipitation at the point (or in
the region) of the forecast in binary form (coding: (fiin = 3, when precip
 itation > 0.1 mm; S�in = 1 in the remaining cases); q is the quantity of
advective precipitation (the initial points of the trajectories are deter
mined from the wind at the 700mb level) in mm; q5 is the fact of advectivQ
precipitation in binary form (the coding is the same as for 45in)�
It should be noted that a singlestep prediction of the fact and quantity
of continuous precipitation using the regression equations did not give 
satisfactory results. For this reason a number of predictors from among the
selected predictors were represented in a binary form, which gave consider
_ ably better results in the prediction of the fact of precipitation. If the
predicted value of the predictant exceeds a definite threshold (in our case '
 two), then another equation is used in predicting the quantity of precip
itation. 
Table 2 gives estimates of the best predictors selected in the prognostic
scheme. It can be seen that the explicable dispersion of the fact of con tinuous precipitation for Moskovskaya Oblast is greater than for Moscow, _
which is attributab,le to territorial averaging. It should be mentioned that 
allowance for w700 in the regression equation for Moskovskaya Oblast leads
to nonsatisfaction of expression (1), O'mmr < crex' Therefore, in the re
gression equation only the fact of advective precipitation remains as a
prognostic predictor. 7 _
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
Table 2
Explicable Dispersion (R2) of Fact of Precipitation and Contribution of its
Predictors
1
2
,qonx (B 96 )
Konuenutis

[TpeAxKropM
R= %
Ka)xAOro npe
a s R'
KKTO
co,
v
o
y
p
3
A
6
7
12
8 Mocwea
pp 011Cx. cb
23 60, 15, 25
0,93
0,85 I 0.95
1,00

w;,N� ~'�cC� Q~, Q~c,,,, 1
34 47, 12, 20, 21
I
0,33 ~ O,~J1 0,95�
1,00
A10S u'~^''� Oilcc� 0
17 40, 35, 25
0,93 ~ 0,93 I 1,06
1,00
W;,.1, c''�CX, (p, (PC,,,,
23 30, 19, 19, 32
0.93 i 0,93 ~ 1,07
1,00
I
_
CxxonrN
29
0,14 1,02
0,89
1,00
lO �,ecKNA
I
"P
orNOs
I
I
9 1VIocKaecxax o6nacrs
~
PP fi
26 ~ 60, 40
0,80 1,00 1,02
1,00
11 `,Pl�x� 0N J 3
43 I 28, 12 so
0,80 1,00 1,02
1,00
~1~�~C, d~
28 58, 42
0,89 1,00 1,08
1,00

I~tOi 01,c:t, 0� oCNH
41 28, 12 60
0,89 1,00 1,02
1,00
11 CaHOnTx
33
0,85 1,03 0,98
1,00
vecxxR
npoMos
KEY :
1.
Concept
2.
Predictors
3,
Fraction (in of each predictor to RZ
4.
er
_ 5.
ex
6,
mmr
_ 7.
y
8.
Moscow
9.
Moskovskaya Oblast
10.
Synoptic forecast
11.
Synoptic forecast
12.
in
13.
syn
The regr
ession equations, in which only a synoptic forecast of precipitation
" enters as a predictor, show that for Moscow 4ffinr 7 CleX,
whereas for
Moskov
skaya Ob
last 6mmr < deX, th,at is, the
synoptic forecast
of continuous pre
_ cipitat::on for Moskovska.ya OUlast must
not be adequately
successfuZ,
as is
also ind
icated by the estimates of the
probability of precipYtation given
in Table
4.
8
,
FOR OFFICIAL USE ONLY
l
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
Table 3
Mean Probable Success of Forecasts of Minimum and Maximum Temperatures for
Moscow and Moskovskaya Oblast in January, February and Narch 1979
3 AGl'O:IIOTII3fl UWIIIiKJ, i Pl1O. 4 T"miP 5 T"P
nPOC.N0361 1 Mecau2 n max
'MOChOBChan Moch MOCKODCKeA
Mocnna (I o6nacrb ~ W, IoGnacTa ]

g CxuonrHRos
AHB8~b1O
1,6
1,8
2,1
2,2
9 TIO peCpCCCHH
2,0
2,3
1,9
1,8
CirxonTlixos
4)eBpanll
2,5
2,4
1,4
1,8
Iio perpoccxx
2,5
2,8
2,4
2,3 
CxxonTHKOa
12
11,8
1,4
3,0
2,6
MapT
170 ,perpeccxx
0,9
1,2
2,8
2,6
CItHOnrliKOa
CP0AH510
2,0
1,9
2,2
2,2 
I7o ,perpecctiti
1,8
2,1
2.4
2,2
KEY:
' .
l.
Forecasts
2.
Month
3.
Absolute error, degrees
4.
TPr
5.
T$ax
6.
Mo s cow

7.
Moskovskaya Oblast
8.
Weathermen
9.
Using regression equations
_ 10.
January
11.
February
12.
March
13.
Average
_ Wind. A study was made of a wind forecast for 24 and 36 hours in advance on
the basis of the PP and MOS approaches. The period of time from September
through May was investigated. Initially the number of predictors was about
20; then, taking into account the peculiarities of the correlation matrices,
their number was reduced to 7. In the prognostic equations we left those
predictors information concerning which could be read from the prognostic
pressure and geopotential charts. The predictors included: cg and dg the
velocity and direction of the geostrophic wind at the earth's surface in 
m/sec and degrees; C700 and d700 wind velocity and direction at the 700
mb level in m/sec and degrees; u850 and v850 wind velocity components
at the 850mb level in m/sec. In the regression equations for predicting
wind velocity the predictant was the wind velocity itself, averaged for a
twominute interval, whereas for preciicting wind direction the wind velo
city components. Such an approach to the forecasting of wind velocity and
direction is the most feasible [10].
9
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
Table 4
C;valuation of Regression Equations and Forecasts of Weathermen of Fact of
Precipitation
3 Oueuha ypaunenEiN perpeccEie
1 ' I 4  5 ~ 6
K
y ~a.a2mF
KoxqenuHR IIpe,qxxropU o;``2 c o~
r a m ~ �
H
Q Maa mao ~taao
' cvm0 a~~ ~~xx
a C o o. o~
_ loa~1 0~ o~u
PP
MOS
Cxxonrx,qe
7 cxatit nporxoa
8 M
o c
x a
a
0
Wloa O1
. 0 44
0,29
0,40
0,40
70,2
64,5
76,0
Wroo, Oacs, o, Ocan
0.35
0,36
0,37
68,4
64,5
72,4
W7M� O0I:0O
0,31
0.40
0,40
70,2
64,5
76,0
wroo. 0:cz, 01 Oc$a
0,47
6,54
6,54
76,7
74,1
79,3
11
0,44
0,48
0,52
74,7
33,6
65,3
.
9 M oc x o e c x a s
o b n a c r
a
' pp
4'.C:. 0
0.30
0,42
0,42
74,9
71,4
70,3

45acz, 0, 0cae
5,32
0,30
0,30
65,3
71,4
59,2
MOS
O.C:.0
0,34
0,38
0,33
69,1
64,2
74.0
(P*c:. 0, 0c.s
U,34
5,30
U,3U
65,4
64,2
66,6
7
CHxonrxqe
cx~tk ~porxoa ' 0,21 0,24
0,24 61,6 71,4
~
~1,8
KEY
:
1,
Concept

2.
Predictors
,

3.
Evaluation of regression equations
4.
Total probability of success, %
5.
Probability of precipitation, %
6.
Probability of absence of precipitation,
%
7.
Synoptic forecast
8.
Moscow
9.
Moskovskaya Oblast
10.
in
11.
syn
The
init
ial sample of winds in a forecast for 24
hours was 134 cases, and
in
a for
ecast for 36 hours 164 cases. The overwhelming number
of winds
 was
with
a velocity less than 10 m/sec. However,
since the main purpose
 of
the i
nvestigation was a clarification of the p
ossibilities of
predicting
the
wind
in accordance with the MOS concept, such
a formulation of the prob
lem
is e
ntirely legitimate.
10
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
_ Table 5
Evaluation of Regression Equations for Wind Forecast
3a6na
AGco.~tor
.
OueNKx cnopocTx nerpa
xas om~iG
roape

KoNUCn
3
xa.F;anpa4
MeH
~ HOCTb
uEta
anetitis
,
, k 1
2
b
E
r
~ 'aerpa,
zpaB
24
+PP I
2,0
0,36.
( 0,40
56,7
MOS
1,9
0,34
0,54
57,7
36
PP I
6,1
0,13
I 0,13
54,6
MOS
1,9
0,32
i 0,40
56,6
KEY :
1. Advance period, hours
2. Concept
3. Estimates of wind velocity
4. Absolute error in wind direction, degrees
Results
The checking of the derived multiple regression equations for the prediction
of temperature, continuous precipitation and wind was carried out on the
_ basis of an independent examination sample, constituting approximately one
third of the corresponding archives of initial data. This same third of the
cases was used in evaluating the forecasts of weathermen, since at the pres
ent time the quality of the synoptic forecasts of the weather phenomena and
_ elements for 24 and 36 hours is higher than for different computational and
other forecasting methods.
Unfortunately, we were unable to make an evaluation of a synoptic forecast
of the wind. It was also impossible to determine the correlation between
the quantity of continuous precipitation in the forecast made by weathermen
and the quancity of precipitation actually falling due to the fact that
in the synoptic forecasts the quantity of anticipated precipitation has a
qualitative formulation.
Temperature. The mean absolute error in predicting minimum temperature for
24 hours in advance for winter was for Moscow 1.8�C, for Moskovskaya Oblast
2.0�C; the maximum temperature for 36 hours was 1.9 and 1.8�C respective
ly. The forecasts of weathermen for this same sample for minimum temperature
were 1.9 and 2.1�C respectively, and for maximum temperature 2.1 and
2.2�C. The correlation coefficients between the minimum and maximum temper
atures predicted using the regression equations and the actual temperatures
were 0.94 and 0.84.
 Such a relationship of the evaluations of synoptic
forecasts hased on regression equations enabled us
the forecasts of minimum and maximum temperatures
11
FOR OFFICIAL USE ONLY
temperature forecasts and
to carry out testing of
for Moscow and Moskovskaya
.
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
/
FOR OFFICIAL USE ONLY
Oblast ln Tanuary, February and March 1979. The reaults of theae evalua
tions are given in Table 3. These results indicate entirely comparable
evaluations of temperature forecasts with respect to regression equations
and forecasrs of weathermen. As a comparison we note that in the United 
States [14], using the MOS system, the mean absolute error in predicting
minimum temperature for 24 hours in advatce is 2.3�C; the maximum tempera
ture for 36 hours 2.4�C. Accordingly., the evaluations cited in Table 3
for the temperature forecasts of weathermen and forecasts based on regres
sion equations for Moscow are better than the corresponding evaluations of
temperature forecasts in the United States. This indicates that despite
the higher quality of the prognostic charts in the United States, the syn
optic experience and choice of the corresponding predictors on the whole
give better results in the prognostic equations. It must be remembered that
the change in the mean absolute error by 0.20.3�C is approximately 10% of
the success of t'he forecast.
Continuous precipitation. The checking of regression equations for pmQdict
ing the fact of precipitation was carried out for the PP and MOS concepts.
The results of this checking are given in Table 4. We should note one
peculiarity in the analysis of these evaluations. In the forecast of :.he
fact of precipitation for Moscow any of the combinations of predictc,rs 
which we selected in the regression equations gives results which are worse
 than the forecasts of weathermen. Bearing in mind that in the preparation
of operational forecasts weathermen daily use several prognostic schemes
 of forecasts of the quantity of precipitation, we decided to introduce a
synoptic forecast of precipitation (precipitation or absence of precipita _
tion 4 Syn) as an additional predictor in binary form. The results of
 evaluations of thPSe regression equations are given in the denominator of
the corresponding rows in Table 4. It can be seen that the evaluations of
the regression equations were now better and exceeded the evaluations of
the forecasts of weathermen. The pattern of evaluations of the regression
equations for forecasting the fact of precipitation in Moskovskaya Oblast
(Table 4) is different. Here the evaluations of the regression equations
with the predictors which we initially selected were better than the syn _
optic forecasts. The addition of a synoptic forecast of precipitation as
a predictor in the regression equation worsened the results.
It is important to note that precipitation forecasts, in accordance with
the MOS concept for Moscow and Moskovskaya Oblasts, were better than when 
using the PP concept.
The regression equations for predicting the quantity of continuous precip
itation, in accordance with the MOS concept, were written in such a way
that they included predictors having only the quantitative values of dif _
ferent atmospheric characteristics. We note that on the basis of the re
gression equations the quantity of precipitation is not predicted ade _
quately satisfactorily. The correlation coefficient between the quantity of
predicted precipitation and the actual quantity was not more than 0.41;
the mean absolute error was 2.1. These evaluations are low but the avail
able information [3] on the correlation coefficients between the quantity
12 
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
of precipitation determined using different schemes and tii,it actual.ly ob 
served do not exceed 0.38, and the mean absolute error in the gradation
0.04.00 mn is 0.7, and in the gradation 420 mm is 4.04.7. As a compar
ison we note that in the best diagnostic regression equation the correl
_ ation coefficient between the computed quantity of precipitation :ind the
actual quantity is 0.65, and the absolute error is 1.8.
Wind. A comparison of evaluations of the regression equations written in
accordance with the PP and MOS concepts (Table 5) shows that all evaluations
made using the MOS system are higher than the evaluations made using the
PP system. There is a tendency to a worsening of the results iaith an in
crease in the advance time of the forecast, which is directly governed by
the quality of the prognostic charts.
Bearing in mind that the evaluations made relate to wind archives consisting 
for the most part of data on winds with a velocity less than 10 m/sec, we
decided to check the conclusions drawn above using archives in which there
are a sufficient number of cases of a wind with a force greater than 15 m/
sec. For this purpose for a forecast for 24 hours, for Liyepaya station
we prepared an archives of 134 cases, of which 43 were for a case of a
wind greater than 15 m/sec. The results obtained using these data confirmed
 the conclusions drawn. It was found that evaluations for the MOS approach
were better that the corresponding.evaluations made by the PP approach. `
Thus, For example, for the absolute error this difference was 0.25, for the
relative error 0.14, for the correlation coefficient 0.04.
Summary
The results for shortrange (for 24 and 36 hours) forecasting of tempera
ture, continuous precipitation and wind using the multiple regression equa 
tions indicate that at the present time for the writing of statistical de pendences for the prediction of weather phenomena and elements it is better
to use the MOS concept rather than the PP concept.
~ Despite the relatively high relative errors in the prognostic fields of pres
sure and geopotential of existing operational schemes, the use of these
_ charts for the writing of prognostic dependences in accordance with the MOS
concept for the shortrange prediction of weather phenomena and elements
is entirely possible.
The quality of predictions of weather phenomena and elements is dependent
not only on how correctly the predictors in the prognostic equations were 
selected, but also on the correctness of choice of the initial data, which
is exclusively associated with the accuracy of the prognostic fields of
meteorological elements. Therefore, together with attempts to formalize
the shortrange forecas.ting of weather elements it is impossible to ignore
the experience of the weatherman, who like no one else can best evaluate
the nature of development of the synoptic situation and select the initial
data.
13
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
BIBLIOGRAPHY
1. Alekseyev, G. A., OB"YEKTIVNYYE METODY VYRAVPJIVA24IYA I NORMALIZATSII
KORRELYATSION14YKH SVYAZEY (Objective P4ethods for the Smoothing and Nor
malizat~on of Correlations), Leningrad, Gidrameteoizdat, 1971.
2. Bachurina, A. A., "Forecasting of Air Temperature and Dew Point in the
_ Surface Layer of the Troposphere," PROGNOZ PRIZEMNOY TEMPERATURY,
VI,AZHNOSTI VOZDUKHA I DRUGIKH METEOROLOGICHESKIKH ELEMENTOV (Prediction
of Surface Temperature, Air Iiumidity and Other Meteorological Elements),
Leningrad, Gidrometeoizdat, 1970.
3. Bukreyeva, L. A., Veselova, G. K., Results ot Testing of Operational
_ Numerical Scheme for the SynopticHydrodynamic Forecasting of the
Quantity of Precipitation," INF. SB. No 8, REZUL'TATY ISPYTAidIYA RAZ
 LICHNYKH SPOSOBOV I SKHEM KRATKOSROCHNOGO PROGNOSA POGODY (Results of
Testing of Different Methods and Schemes for ShortRange Weather Fore
_ casting), Moscow, Gidrometeoizdat, 1979.
~ 4. Vapnik, V. N., Chervonenkis, A. Ya., Uniform Convergence of the Fre
quencies of Occurrence of Events to Their Probabilities and the Prob
lem of Seeking an Optimum Solution Using Empirical Data," AVTOMATIKA
I TELEMEKHANIKA (Automation and Telemechanics), No 2, 1971.
5. Vapnik, V. N., VOSSTAPIOVLENIYE ZAVISIMOSTEY PO EMPIRICHESKIM DANNYM
(Restoration of Dependences Using Empirical Data), Moscow, Nauka, 1979.
a
6. METODICHESKT.YE UKAZANIYA PO PROVFDENIYU OPERATIVNYKH ISPYTAidIY NOVYKH
METODOV GIDROMETEOROLOGICHESKIKH PROGNOZOV (Systematic Instructions on _
Carrying Out Operational Tests of iJew Methods for Hydrometeorological _
Forecasts), Leningrad, Gidrometeoizdat, 1977. 
7. Pdeymark, Yu. I., Vatalova, Z. S., Vasin, Yu. G., "Image Recognition and _
Medical Diagnosis," TRUDY MEZHDUNARODNOGO SIMPOZIUMA PO TEKHNICHESKIM
I BIOLOGICHESKIM PROBLFI4AM UPRAVLENIYA (Transactions of the International
Symnosium on Technical and Biological Problems in Control), Moscow, Nauka,
1971.
8, Romanov, L. N., Vinogradova, G. M., "Orthogonal Expansions of Synoptic
Situations Using the Caordinates of an Adapted Base," TRUDY ZSRNIGMI
(Transactions of the West Siberian Regional Scientific Research Hydro
meteorological Institute), No 11, 1974.
 9. Snitkovskiy, A. I., Ustinova, G. P., Prediction of Minimum Air Tempera
ture in Moscow With Use of Future Pressure Fields," TRUDY GIDROMETTSENTRA
SSSR (Transactions of the USSR Hydrometeorological Center), No 225, 1979.
10. Snitkovskiy, A. I., Sonechkin, D. M., FuksRabinovich, M. S., Shapoval
ova, N. S., OBZOR. SISTEMA OB"YEKTIVNOGO KRATKOSROCHNOGO PROGNOZA YAV
LENIY I ELEMENTOV POGODY V SShA (Review. System of Objective ShortRange
Forecasting of Weather Phenumena and Elements in the United States), Ob
ninsk, Informatsionnyy Tsentr, 1978. _
14
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR 0FFICIAL USE ONLY
11. Ugryumov, A. I., Veselova, G. K., Chernova, V. F., Ageyeva, A. K.,
Bukreyeva, L. A., "Comparative Evaluation of Regional Schemes for
the Numerical Forecasting of the Pressure Field for 24 and 36 Hours
in Advar`ice," INF. SB. No 6, "REZUL'TATY ISPYTANIYA RAZLICfRYKH SPOSOB
OV I SKHEM KRATKOSRI)CHNOGO PROGNOZA POGODY"), Pioscow, Gidrometeoiadat,
1978. _
12. Bermowitz, R. Y., "An Application of Model Output Statistics to Fore
casting Quantitative Precipitation," MON. WEATHER REV., Vol 103, No
2, 1975. �
13. Hammons, G., Dallvalle, J., "MOS Maximum/Minimum Temperature Forecast
Equations Based on ThreeMonth Seasons," TECHN. PROCEDURE BULL., No
= 155, 1976.
14. Hammon, G., Dallavalle, J., Klein, W. H., "Automat:ed Temperature Guid
ance Based on ThreeMonth Seasons," MON. WEATHER REV., Vol 104, No 12, _
1976. _
15. Klein, W. H., Lewis, F., "Computer Forecast of Maxi;num and Minimum Tem
perature," J. APPL. METEOROL., Vol 9, 1970.
= 16. Klein, W. H., "On the Accuracy of Automatic Max/Min Temperature Fore
cast," J. APPL. METEOROL., Vol 11, No 8, 1972. 
15
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
UDC 551.(5090313:558,21)
ALLOWANCE FOR OROGRAPHY IN NUMERICAL WEATHER FORECASTING MODELS
Moscow METEOROLOGIYA I GIDROLOGIYA in Russian No 9, Sep 79 pp 1624
 [Article by Doctor of Physical and Mathematical Sciences A. I. Romov, Ukrain
ian Scientific Research Hydrometeorological Institute, submitted for public
a;.ion 8 January 1979]
Abstract: The author proposes amodel for numer
ical weather forecasting using fu11 equations
in a(l coordinate system with improved allow
ance for orography. Transformatiun of the equa
tions of motion was carried out. This makes it
possible to eliminate the errors caused by com
putation of small differences in the large val
ues characteristic of the mentioned coordinate
_ system in situations over mountain slopes. As
i a result, there is an increase in the accuracy
of the computations of the pressure gradients,
which here attains the usual accuracy level
for an isobaric coordinate systemo In the pro
Gess of difference solution of the problem the
need disappears for scaling geopotential to non
standard isobaric surfaces. Thus, a solution is
found for the known problem of allowance for
orography in numerical weather forecasting
problems. [Text] In dynamic meteorology diffe;:ent coordinate systems are used, depend
ing on the nature of the studied processes and the conditions unde.r which
they transpire. One of the principal, most commonly used systems at the pres
ent time is a psystem. The system of equations usually used in numerical
weather forecasting schemes, without friction taken into account, in an x,
y, p coordinate system has the follaraing form [1]:
da + u du + du + du dN lv,
ar a vy c g d.r +
IN
16
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
riv dv dv du dH
uc u dX v ~y W ~C g ~y  lu, (2)
ar a7 ar , a7 R7z 7 (3)
dr + u d + v uy r u~ d C = g
du du (4)
dz ~ dp + d C 0'
gC dhl . (5)
K dC
Here u and v are the wind velocity components along the x and y axes; H is
the height of the isobaric surface; C is the ratio of pressure p to the
standard value P= 1000 mb, playing the role of a vertical coordinate; 4j is
the full time derivative (time = t) of the 4 coordinate, representing the
velocity of displacement of air particles relative to the isobaric surfaces;
T is temperature; g is the acceleration of free falling; ~ is the Coriolis
parameter; R is the gas constant; 7'a is the adiabatic temperature gradient.
The soughtfor functions are u, V. H, T, 4J. The forecasting equations in
the form (1)(5) are characterized by relative simplicity and convenience
for a lowland territory. However, it is known that in a psystem it is dif
ficult to describe the conditions at the ].ower boundary of the atmosphere,
which is particularly important in prognostic problems for taking into ac
count the influence of orography on weather.
In 1957 N. Phillips [8] introduced the vertical coordinate o", equal to the
ratio of pressure at a particular point to pressure at the earth's surface
under this point, and proposed a rectifying o'coordinate system whose ad ~
vantages consist in convenience in stipulating the boundary conditions at
the earth and solution of weather forecasting problems using a full system.
This system, the same as the generalized Shuman and Hovermale variant [9],
due to the mentioned advantages is intended for taking into account the in
fluence of orography.
However, we note that up to this time virtually no one has succeeded in
making full use of these important advantages. The fact is that in a dsys
tem, as is well known [1, 2, 6, 10], computations are made of the pressure
gradients which figure in the equations of motion, with appearance of ad
ditional effects of "small differences of large values" of a higher order
of magnitude s:han in the ordinary coordinate systems x, y, z and x, y, p.
This is accompanied by great errors in the difference approximation in
solving the equations of motion and instability of the computatiorLs [6, 10]. 
For a more detailed examination we will write equation's (1)(5) in one of
the modifications of the o'system examined in [3, 4]:
17
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
d~ du du 1 I j, ~ri
dt +u vx +v d, ~ Cup (u~  oW ~ ~s = (6)
dH ? RT d C*
lvg ox Q;"`tll,)Cop ox'
dv +lt dv ~'v dv . ~ (wcw;'~
ar vx ay coP v~
aH Q �r ~c* (7)
_  ~Il  g dy  a~~; + 0 CoP tlyX ,
T
+ ll dX 'U dy 1,~ , ~oP
dt _
RT7a 
g [aC* + (1  ~)CoPI
(8)
* �
w  woP o ~u a qi d
Ux ~
/
dll >op) dV (C* py
~
n� \ dx + dy d:;, ~
(9)
f (1  c) (u d*
~ dX
.1 \d 1 'Inp) vU (;~`;pp) i ' (1.0)
dx t dy , d
a .
+ 'o
;
T= 'p 11
� ~op Q ~
C"
y, t) (tch ddX vh ay l~ (12) 
' :  :op
a _ toP ' . (13)
_ ~*(x, y, t) is the ; value at the earth's
is the individual change in the ~ * value; ~ oro
at the isobaric surface ~oro  const, being the
graphic" air layer in which the mountains are coi
in the limits of which a dcoordinate system is
18
FOR OFFICIAL USE ONLY
li1*
surface (mountains);
is dimensionless pressure
upper boundary of the "oro
npletely submerged and with
used.
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
_ cioro ls the vertical velocity at the surface S oro  const; H is the height
 of the surface d= const above sea level; h is the height of the relief of
the earth's surface; the h indices denote the values ori this surface.
The same as in [3, 41, here we will propose that within the framework of one
scheme use will be made of a combination of both coordinate systems: in the
upper layer (above the surface 4oro = const) use is made of equations in
the form (1)(5), and in the lower layer in the form (6)(12). We note
that in the latter system, in place of the former five, we have seven equa
tions, since here, besides the functions u, v, T, H and rJ , we have addition
ally introduced and investigated the functions c3 * and ~ The initial fields
u, v, T, 4 * rzust be stipulated. The region of solution of the problem is
limited from above by the isobaric surface ~ 0= const (0.1 or 0.2). The
boundary conditions aYe:
w=0 when 10 =60,
 H=h(X, y), when 6=1.
We note that the horizontal derivatives of one and the same function in both
systems, speaking generally, differ greatly in their values. This is attrib
utable to the fact that the isobaric surfaces always remain almost horizontal,
whereas the surfaces o'= const can have different slopes in dependence on the
nature of the relief and the thickness of the layer of submergence of the
mountains oro values).
 These differences are particularly significant for the derivatives of geo
potential (heights H). In each of the equations (6) and (7) the force of 
the pressure gradient is not represented by one term, as in (1) and (2), but _
by the last two terms on the righthand sides, by the sum of these terms.
_ Over the mountain slopes, in the lower air layers, these terms are opposite
in sign and each of them can attain values two orders of magnitude greater
than th e resultant force of the pressure gradient. For example, with a moun
tain slope 102 the order of magnitude of each of the two mentioned terms
 on the earth is 101, at the same time that their sum is 103. (Here and
in the text which follows the estimates are given in the MTS system of
units).
Under such conditions computations of the pressure gradient in the final dif
ferences becomes impossible. This circumstance already over the course of a 
number of years serves as an obstacle for the effective use of a o' coordin
ate system in the practice of numerical weather forecasts using full equations
and in modeling of general circulation of the atmosphere.
In some studies, in order to eliminate the mentioned inconvenience, use is
made of a method in which at each nodal point there is introduction of an
auxiliary isobaric surface, its heights are interpolated by means of scaling
from the surface a'= const and thus the horizontal pressure gradients are
found [7, lO]. Such a method is rather unwieldy, inadequately economical, 
and the error associated with the scaling of geopotential can be comparable
with the value of the pressure gradient itself.
 19
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
u
In a study by S. 0. Krichak [2] the influence of the mentioned shortcoming
is softened by a corresponding choice of pressure at the upper boundary oi
the considered layer of the atmosptiere at which a small horizontal varia
bj:lity of layer thickness is attained. In our case, with combined coordin
ates, the possibilities of such a choice are limited.
In this article we propose a method which makes possible a complete elimin
ation of the additional effects of small differences of large values charac
teristic of a c7'coordinate s}rstem in situations over mountain slopes with
out having recourse to the introduction of auxiliary isobaric surfaces and
scaling of heights. This is attained by a corresponding transformation of
the equations themselves within the framework of a 6coordinate system. In
the layer of submergence of mountains we examine the dimensionless pressure
function 4 , which is represented in the form of the sum of the main 4 value
and the small deviation ~ lz
,
(14)
~
Now we will turn to the last term on the righthand side of equation (6). Us
ing an expression following from (13),
(1  3) roP, (15)
this term can be written as fo'11ows:
� aRT dC" RT dr .
" rt ~  : ) ~oP ~X  v.t Then the sum of both terms in (6) representing the force of the pressure
gradient, after the substitutian (14), neglecting the squares of small val
ues, can be represented in the fc;rm
_ ()H RT dH RT dC RT + (16)
 g d d g ox  ~ a a
~
+ RT C, dX �
[de note that the f irst term on the lefthand side of (16) has the form of
the pressure gradient force in a pcoordinate system and the second term
in a zsystem.
On the righthand side of (16) the first two terms in the mountains have
the order of magnitude 101, tlie next two terms 103. It is natural to
attempt to avoid the first two terms, equating their sum to zero and deter
mining from this equation the expression for e . For this it would be neces
sary to integrate it for x, which cannot be done with accuracy, without know
ing how temperature changes at the surfaces 0'= const. In this case allow
ance for this change is significant, it cannot be neglected, and therefore
 first on the righthand side of (16) it is necessary to carry out the fol
lowing substitution:
20
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007102/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
d d T
R7' a~c` = R X (Tli()  t~'in.
(17)
after which we obtain:
gaN R r a ti aH a
' dx RT dC,
c d g o R a (T1nC) + ~18~
C f Ci dx
f RT a~ + R ln : ar
C' dx dx Here on the righthand side has appeared a term (dependent on )T/a x) with
a value about 103, but the T function now is determined with accuracy from
the condition of a full mutual compensation of the first two terms on the
righthand side of (18), having an identical and the highest order of mag
nitude (101) among the other terms. We will assume
 g az R d (TInC)=0. (19)
Integrating (19) for x and assuming the cordition 1 with H= 0, we obtain
the following expression for 4 , precisaly satisfying equality (19):
Aii
:e Rr.
(20)
Thus, if is computed using formula (20), we obtain the following expres
sions for the components of the pressure gradient force relative to the x
and y coordinates respectively:
(21)
_'dH cRT ~C* RT dC;
g a toP a r+ c, a +
+ kT + R1nC aT
c2 ax aX '
dH a RT d C* R7 d C,
 g ay ~ Qt* + c 1.Q~ top dy C Ct d +
Y
� + RT C, d~ I R In C aT .
[op = oro ] dy dy
Here the order of magnitudes of all the terms on the righthand sides both
over the plains and in the mountains no longer exceeds 103.
It would 'oe convenient to compute the7 fields using initial data and consider
them to be approximately constant with time during the course of the entire
period of the forecast. However, in a numlier of cases, due to the dependence
on temperature, they can change significantly. The T function does not satisfy
21
 FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
 FOR OFFICIAL USE ONLY
the "standardness" condition and therefore it is desirable that it be com
puted in each tirne interval in such a way that there be accurate satisfac
tion of expression (19). The latter is substitution of
expressions following from (20) into equations a
1nC= R~, 
I dC g dH + gH dT (S = x, Y),
C ds RT ds RT'= ds
and the use of these equations in the following form:
du
i + rc tlX + v oy ~Cop (W  o u~~') o' = lv
Rr dC, _gC,eXp IaH _gN ar + (24)
,gHl dx (AH
RT J dx� T oz
exp~� k}+;,
~
F gT C, eXP I Rr j aX '
du
 ~w =  lu 
dl + u dX + v dy + C*  Cop ~ w d�
RT d CL (gN dH gH dT
~ ~H dy  g~i e x p ~ k~.~ dy  T dy
esp;  R%' } +
(25)
I + eXp K AT
T / ( uy .
[op = vro ]
On the righthand sides of equations (24) and (25), as b efore in (6) and
~
(7) , there are terros with c7 H/ r) x and e) Ii/ ay, which after trans f ormation
became approximately two orders of magnitude smaller, since they were mul
tiplied by . In Place of the former terms with x and ds */d Y,
T 1/S
after transformation their analogues, terms with a~1/a x and 3,C1/j Y, aP'"
peared; their values were also reduced by two orders of magnitude. In both
= equations two other terms appeared, these being dependent on the horizontal
temperature gradients; the last (underlined) terms, which contain the fac
tors ~ 1/; have the order of magnitude 105 and they can be neglected.
 Thus, here the equations of motion were transformed to a form convenient _
for taking into account orography in numerical weather forecasts. For this
 pur.pose it is necessary that equations (6), (7) be replaced by equations
(24), (25).
22
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
With respect to the derivatives of x and y in the advective te nns for u, v,
T in equations (6)(10), they also experience certain changes in the O=
_ system. A compensation of the corresponding changes in the advection of
meteorological functions enters into the terms taking into account vertical
transfer, and although these values are not so significant as in a case
with a pressure gradient, they can attain 100% of a11 the advection and 
they cannot be neglected. (The computation of advective terms is also some
what complicated.) One important p ractical conclusion follows from what has
been said: in a p~coordinate system over mountainous regi.ons in the equa
tions of motion it is impossible to neglect the terms repiesenting the ver
_ tical transfer of wind velocity.
The exclusion of derivatives of total pressure and transformation to
the deviations ~ i together with (6) and (7) is also possible in the other
equations. For example, equation (12) can be transformed and used in the
following form; 
= aC0 _ 1 wh
C~ dt~ /
(uh d + vh +
at G" /h dx dy
(121)
 uh aX + vh vy
and accordingly in each time interval the ~ i field is computed.
Now we will discuss the method for computing the initial values of the meteor
ological elements required in forecasting with use of the transformed equa
tions. The surface ,I oro  0�7 can be selected for taking into account oro
graphy over Europe. In our ocoordinate system this corresponds to the
equation O' = 0. The following surfaces are introduced: a) o= 1, coincid 
ing with the lower boundary of the atmosphere and b)Cr = 0.5, lying in the
interval between the earth and the 700mb isobaric surface.
After objective analysis, which is usually carried out in an isobaric coor
dinate system, it is necessary to compute the initial grid fields (values
at the grid points of intersection) of temperature T and pressure ~ on the
two mentioned surfaces and H on the surface O'= 0.5.
First we will discuss computations of initial data at the earth's surface.
If Hk and Hk+l are the heights of two adjacent working isobaric surfaces
at a particular "node vertical" and at the same time h falls in the range
 1`1n+i. 0 and in arder to derive the boundary conditions for total absorp
 tion for the system (D), adhering to [6, 111, using the integral Fourier
transforms F:c~ � and Laplar.e transform Lt~q,we transform the system (D)
to an ordinaly differential equation.
~ With zero initial data from (1') and (2') it follows that
0, . v=i(q9Q) (92I12)1cU.
u=i(t9{1T1) (92i12)1
28
FOR OFFICIAL USE ONLY
IR
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
 Substituting these expressions into equations (3) and (6) and excluding the
function, we obtain
fi T_ x E2
dp' q' + l' (7)
In order to obtain the total absorption conditions for the system (D) it
is necessary, in accordance with [6], to find a fundamental system of solu
tions for (7). In this case x= x(p) and is dependent on the stratifica
tion assumed in the model. The simplest, evidently, is the case
 a) X = const.
A more complex case, but on the other hand, more corresponding to reality,
is
b) 7 = c2p'2, c = const.
Such a stratification was assumed, in particular, in [3].
In both cases the fundamental system of solutions is easily found, and ad
. hering to [6], the total absorption condition for (7) can be obtained in
the form
d
dp (Po)  a(9, 7i) ~(Pa),
where 0' (q, v,) is a function called the symbol of the boundary opera
tor [1],
~ x1/2 Yi2)112 (q' + 1=)112 ~
a (q, E, _ ~ in case a
_ I Ps 1[L+ + rr ~
~ 2 4 92 +in case b).
Using the inverse integral transforms [2, 6, 11], in final form we obtain
the total absorption conditions _
dT r
Q) dp (t, x, Y, f~a) = 4  ~ `[(x  X)2 (Y  }')'11'~2 o [l (t  T)) X
0 1~2 .
~X (dX., I dY~, t(T, X, Y. P.) dXdYdT. .
a~
6) dP (t, x, Pj = 16 cl ~ L(x X, Y Y, t T) X
Kx, y. r .
= XT (T, X, Y, pe) dTdXdY,
where Kx, Y, rr", T) ~(x  X)= +(Y  Y)' < 4 c' (t  T)'`) ~
Yr=~z C ~x` _+y=J
L= f~x~ Y, f( r, Y, 0  u J, (lra) drr,
u
29
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
d= d'' d' ' ~ _
f = ((2 c)_�  or.  vz_  dy_ 1 2 / ct2  xZ
L
y=) I r (.Y= + y=)112 (2 C) I],
Jo, J1 are Bessel functions.
The total absorption boundary condition, thus, is not local; integral oper
ators enter into it. Integration occurs over an infinite region. With in
creasing distance from the point x, y, t the kernels of the integral oper
ators decrease the distant points enter into the boundary condition
with a small weight.
The total absorption condition for a more realistic case (b) is also more
physically sound integration does not occur over the entire halfspace,
but only along the "inverse light cone"; K is a situation characteristic
for hyperbolic problems [5, 7]. This is attributable to the fact that in
such problems perturbations are propagated with a finite velocity.
It follows from the evaluations for the Bessel functions that there is a
powerlaw decrease of the kernels in the convolutions. A close result is
also obtained in a difference case (with replacement of the integrals in
the convolutions by infinite sums). In order to decrease the number of
terms, and accordingly, in order to decrease the volume of the required
memory and to accelerate computations we used the special procedures de
scribed below.
#2. Total Absorption Condition for Difference Problem
A description of an operational hemispherical model is given in [3] and
we will employ the notations given there. As in #1, carrying out linear
ization in the difference system, we obtain
ut  lvr =  (px;
 ~ (P )
^~r �4 lu'   cpy~ .
 (n eA p l ( Iu (P) t ca (P+oP)Jx,, + Iv (P) v (P oP)Iyxl;
I ~ (P + A ~P~(D) it = _ ? [T (p) + ~ cp fi Ap].
ere, as in the model [3], it is assumed that the parameter of increase
H
in a stereographic projection m= 1, and in contrast to the model, for
simplification of the computations the time filter parameter 'V= 0. Apply
ing the Ztw Z transform [8] and the Fourier transform to the system (P),
making computations similar to those in #1, we obtain the system
30
FOR OFFICIAL USE ONLY
~
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OrFICIAL IISE 01TLY
ID(PId p) im(P) +b i(P+ a p)b c(P) = 0;
~ ~P +0P)
where (P) (P .CG p) a ~r (p) 0,
a=x0 p A t(zz1)1, b=h'l (zz~)=+diq(ti+ zj)3]>~ _ X[(zzi)e p etS]1, S2' (]cos 2~ !t cos 2rIh),
which can be rewritten in matrix form
�~b zoa
(ID (P+'A P)l _S (P)1
\T(P~AP)l \ ~(P;.~, 2, a + b
b a'
Assuming ~G = const, we find the ei_genvalues oi' the J3 maLri
~1, s=(a b� 2 l/ ctb ) (b  a) i and when ( z;~~ l I ti,J> 1> 1),: j
The upper boundary condition, in a general case, is written in the form
[B = uP (Per) ]
(At, A 2') (PD) 1 =g (t, X, Y),
(P~J l
(9)
where g is some function, A1p2 are difference operators, whzch like the
operator (8) of the differential problem are not mandatorily local. The
condition on the ShapiroLopatinskiy boundary opezators A1, A2 [1, 11, 16],
guaranteeing the correctness of the mixed boundaxy p.roblem (P)(9) (this
 means that with any initial data and any furict:Lon g a solution of the proU
lem exists and is continuously dependent on the initial data), is a nonortho
gonality of the row of symbols (OZ1, 0(2) to tlle eigenvector of the matrix
B, corresponding to C2. The total absorption condition [47] involves an
orthogonality of the row (�l, aC2) to the eigenvector of the B uiatrix cor
_ responding to cW1 (the first condition obviously follows from the second).
For example, the first row of the matrl.x (B  alJ;) satisfies this secord
condition:
2 ab , 2 ab
)
(a1' a= ba ' ba~
or, proceeding to the normalization az =1, 0~1 _(ab)1/2. As a result,
the total absorption boundary condition has the form
3 B = up (per) ]
Y(Pn) =A t~~ (pu) �
The A1 operator is the difference operator for an infinite number of points;
its symbol c4l is the product of the function of z and the function of ; and
YI. In order to realize the total absorption condition on an electronic
computer it is necessary to approximate the Cle1 operator by a difference
operator with the least possible number of points,
31
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
For an economical inverse Ztransformation [8] of the cofactor in OC1, de
pendent on z, we will use the Pade approximation, which has recently found
application in different fields of mathematics and applied problems [12].
In this case the approximated function is replaced by a rational function
(in our case this point is z"1 = 0) having the same derivatives to the or
 der of magnitude N1 as the approximated function.
The cofactor in a 1, dependent on ~ and q , is approximated by a sum in the
form 2: ski exp (i c klt, + i
k,1 These approximations are described in greater detail in the Appendix. After
these approximations the total absorption condition, in accordance with the
convolution theorem, assumes the form
T (Pa) Q (z ~ (Ps) + P (z') ~ Skr eXp ( i E kJc., i % lh,y ) lb (P. )
o r
ti
=(t, x, Yo R) =2. 9n T(t  I1, t, x, Y� Fe) + (10)
n1 M '
 P� 1 4, 10 (t  nA t, x khx, y lt,,, As)�
n=0 R, /
The total absorption condition was used specifically in such a form in the
model.
Thus, the total absorption boundary condition determines the vertical velo
city value Z at the point of the spatialtemporal grid lying at the upper
computation level (t, x, y, pup) through the'G values at the preceding mo
_ ments in time t n At and the geopotential values at this and preceding
 moments in time at the closest points also lying at the upper computation _
level. The method for computing the corresponding weights qn, Pns skt is
described in the Appendix.
#3. Results of Numerical Experiments and Concluslons
Tests of the described approach were carried out using a regional variant of
the model [3]. As the initial data we used archival data for 0000 hours on
110 April 1977 for the territory including the North Atlantic, Europe, 
_ North Africa, Near East, Central Asia and Western Siberia.
The interval for the horizontal variahles was 300 1m, the time interval was
12 minutes. 
With replacement of the traditional boundary condition at the upper boundary
by the condition of total absorntion of the emerging waves there were ap
preciable.changes.in.the prognostic fields at all level5. In particular,
32
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
due to the ahsence of reflection from the upper houndary there was a decrease
in the kinetic energy, there was in increase in divergence of the wind field,
artificially understated by the traditional condition [3].
Table 1
1 ~ 2 YponeHb, .0
i Z
lOuU I$50 I 700 I 500 I 300 I 100 II 1000 I 850 I700 I500 I 300 I 100
3 K002 3c"}cuxear hoppenaiwx 4 Orttoci+reAr}Ias outifta 5 4 cym U,UO I O,UI I 0,02 I U,U2 I 0,02 I0:04 0,0~I0,051U,U7IO,U7IO,G3I+O,O:i
KEY:
 l. Time of forecast 5. days
2. Level, mb
3. Correlation coefficient
4. Relar_ive error
 Table 1 gives the improvements (tlie correlation coefficient increases, the
relative error usually decreases) due to use of the total absorption condi
tion at the upper boundary in place of the traditional condition in fore
casting for three and four days in advance. All the evaluations, except for
the relative error at 100 mb, are better in case of use of the total absorp
tion condition, which evidently must be considered preferable.
u xM
' x 6CO !x
300 0,21�10'6
 0,e9�10'e
600 300 J00 609 XNM
 3C0
x GGO x
Fig. 2. Overlay and weights (mb�sec/mZ) of operator for horizontal variables
approximating T( operator.
The possible ways to bring about further improvement in approximation of the
total absorption condition (9) are: 1) careful selection of the optimum (for
example, in the statistical sense [4, 5]) coefficients of the boundary oper
Jt0r skt ; 2) allowance for smoothing (Nz# 0) in formulating the boundary 
condition (9); 3) shifting to variant (b) in the case of linearization of _
the model, that is, to that which is used in model [3J. For this it is neces
sary to find a fundamental system of solutions of a seconddegree difference
eqiiation with variable coefficients, which is a complex problem.
_ A more detailed discussion of the mathematical and computational aspects
will be given in a separate publlcation.
33
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007102/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
The authors express appreciation to M. S. FuksRabinovich for constant at
tention to the study and useful consultations.
 Appendix
Method for computin g the coefficients in the total absorption boundary oper
ator for emerging waves for difference baroclinic model
For the inverse Z transformation of the symbol (ab)1/2 it is necessary,
in accordance with [8], to carry out expansion into a Laurent series with
z = L, 0 .
We have (ctb)112 = II (1  z')! I 2 � z z41l 12 _
C)r.
= n ~ 1 = z' k lPk f'k
I kcl
where rj = S,1'h1 (1 A t" 1,)112, ~L 1% t" 12)(1 r .1 t' 11)
Pk is the kth Legendre polynomial. The coefficients of the power series
in even powers of z 1 with F= 0.97 decrease, but extremely slowly. The
number of terms in the Laurent series to which we limit ourselves is equal
to the number of time layers in the total absorption boundary condition
and therefore, due to the limitations on the electronic computer memory
_ resources and speed it cannot be too great. At the same time, limitation
only to the one first term of the expansion was too approximate.
Accordingly, for approximation of the expression in braces we made use of
the Pade approximation j12, 19]
+ ~ Zs k Ipk P't (1~)~ ~ P" (z2)iQ'� (z2),
I k=t 1
where Pn, Qm are po lynomials of the powers n and m, Qm (0) = 1, and the ap
proximate equation, by definition of the P ade approximation, results in a
 coincidence of the Laurent expansions to the number N1 = n+ m inclusive;
in the case m= 0 this is simply a cutoff of an infinite series.
_i As a measure of the approximation error we used I;*
s
2
Lr akt
N,
where ak is the difference of the Laurent coefficiente with the numbers k
of the approximated function and its approximation. In the evaluations we
assumed N2 = 30. A good result was obtained with m= n= 2: E= 0.006.
We note once again that n+ 1 and m+ 1 are the numbers of the time layers
of the H and 'C functions which enter into the boundary condition. If m=
0, that is, the Pade approximation is not used, then it is necessary that
34
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007102/08: CIARDP8200850R0002000300394
FOR OFrICIAL USE ONLY
n7~'5, which would lead to insuperable difficulties in application.
In order to ohtain the coefficients of the polynomials Pn and Qm we used
the program which we wrote and a more general program carried out using
the description in [15] and
F1=1.00001,9925 z2 +0,9997 z4, Q2= 1,00001.9777 z2+0,9997 z4
_ The expansion of the n( A,~) function into a double Fourier series can
be carried out, for example, numerically, using the socalled fast Fourier
transform. Only the coefficients sk� with a sum of the indices a multiple
_ of four are different from zero (this can also be demonstrated analytical
ly). With a cutoff of the series, which is necessary due to restrictions
on the electronic computer resources, the sum of the sk L coefficients, that
is, the value of the cutoff series with :F, _YI = 0, is different from zero,
whereas T((0,0) = 0. Thus, the approximation of the Tt symbol in the zero
harmonic is impaired. It is possible to approximate Tt in a form optimum
in the statistical sense with the limitation that the symbol of the approx
imating operator becomes equal to zero when �_Y? = 0[4, 51.
In the cited experiments we limited ourselves to a very simple overlay of five
points (see Fig. 2); the weights at the lateral points are 1/4 of the weight
of the central point, which was selected empirically and in the final exper
iments was selected equal to 0.84�106.
BIBLIOGRAPHY
1. Agranovich, M. S., "Boundary Problems for Systems of FirstDegree Pseudo
differential I'irstOrder Operators," USPEKHI MATEMATICHESKIKH NAUK (Ad
vances in the Mathematical Sciences), No l, p 24, 1959.
2. Beytmen, G., Erdeyn, A., TABLITSY INTEGRAL'NYKH PREOBRAZOVANIY (Tables
of Integral Transfo rns), Vol I, Moscow, Nauka, 1969.
3. Belousov, S. L., et al., "Operational Model of Numerical Forecasting of
Meteorological Elements for the Northern Hemisphere," TRUDY GIDROMET
TSENTRA SSSR (Transactions of 'te USSR Hydrometeorological Center),
No 212, 1978.
4. Gordin, V. A., "Some Mathematical Problems in Hydrodynamic Forecasting,"
Tezisy DOKLADOV NA II VSESOYUZNOY KONFERENTSII MOLODYKH SPETSIALISTOV
GIDROMETSLUZHBY SSSR (Summaries of Keports at the Second AllUnion Con
ference of Young Specialists of the USSR Hydrometeorological Service),
Obninsk, 1976.
5. Gordin, V. A., "Some Mathematical Problems in Numerical Hydrodynamic Fore
casting," DOKLADY NA II VSESOYUZNOY KONFERENTSII MOLODYKH SPETSIALISTOV
GIDROMETSLUZHBY SSSR (Reports at the Second Al1Union Conference of
Young Specialists of the USSR Hydrometeorological Service), Obninsk,
1977.
35
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOn OFFICIAL USE ONLY
 6. Gordin, V. A., "P'lixed Boundary Prohlem Simulating the Cauchy Problem,"
USPEKHI MATEMATICIiESKIKH NAUK, Vol 33, No 5, 1978.
7. Gordin, V. A., "Mixed Boundary Problem for a Barotropic Atmospheric
_ Model," TRUDY GIBROMF.TTSENTRA SSSR, No 196, 1978.
8. Dech, G., RUKOVODSTVO K PRAKTICHESKOMU PRIMENENIYU PREOBRAZOVANIYA
LAPLASA I ZPREOBRAZOVANIYA (Manual on the Practical Use of the Laplace
Transform and the ZTransform), Moscow, Nauka, 1972.
9. Dikiy, L. A., TEORIYA KOLEBANIY ZEMNOY ATMOSFERY (Theory of Oscillations
of the Earth's Atmosphere), Leningrad, Gidrometeoizdat, 1969.
10. Kibel', I. A., WEDENIYE V GIDRODINAMICHESKIYE METGDY KRATKOSROCHNOGO
PROGNOZA POGODY (Introduction to Hydrodynamic Methods for ShortRange
Weather Forecasting), Moscow, Gostekhizdat, 1957.
11. Shilov, G. Ye., MATEMATICHESKIY ANALIZ. VTOROY SPETSKURS (Mathematical
Analysis. Second Special Course), Moscow, Nauka, 1965.
12. Baker, G. A., Jr., ESSEi1TIALS OF PADE APPROXIMANTS, Acad. Press, New
 York, 1975.
13. Beland, M., Warn, T., "The Radiation for Transient Rossby Waves," J.
_ ATTiOS. SCI., Vol 32, No 10, 1975.
' 14, Bunnet, A. F., "Open Boundary Conditions for Dispersive Waves," J.
ATMOS. SCI., Vol 33, No 2, 1975.
15. Engquist, B., Majda, A., "Boundary Condition," MATH. COMPUT., Vol 31,
1977. .
16. Gustafsson, B., Kreiss, H.0., Sundstrom, A., "Stability Theory of
Differerce Approximations for Mixed Initial Boundary Va1ue Problems,
II," MATH. COMPUT., Vol 26, 1972.
 17. Kirkwood, E., Derome, J., "Some Effects of the Upper Boundary Condition
and Vertical Resolution on Modeling Forced Stationary Planetary Waves,"
MON. WEATHER REV., Vol 105, No 10, 1977.
' 18. Miyakoda, K., Rosati, A., "OneWay Nested Grid Models: the Interface
Conditions and the Numerical Accuracy," MON. WEATHER REV., Vol 105,
No 9, 1977.
_ 19. Starkard, Y., "Subroutine for Calculation of Matrix Pade Approximants,"
COMPUT. PHYS. COMM., Vol 11, 1976.
36
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
UDC 551.558.2
VERTICAL CURRENTS IN THE TROPOSPHERE
Moscow METEOROLOGIYA I GIDROLOGIYA in Russian No 9, Sep 79 pp 3443
[Article by Candidate of Physical and Mathematical Sciences V. A. Shnaydman,
Odessa Hydrometeorological Institute, submitted for publication 12 December 
 1978]
Abstract: Vertical currents in the troposphere
are determined by solution of a differential
equation derived from the vorticity equation,
heat influx equation, continuity equation and 
equation of statics, in which the following are `
taken into account: advection of absolute vortic `
ity, heat advection and frictional currents at
the upper boundary of the planetary boundary
layer. The latter are found from solution of a closed system of equations, including the
equations of motion and the equation for the
balance of kinetic energy of turbulence and
semiempirical closing hypotheses. Standard aero
synoptic data are used as the initial data. The
 results of computations for specific synoptic
situations are given. It is shown that fric
tional vertical currents, computed using the
proposed method, agree better with the precip _
itation field than the currents determined using the
the Dyubyuk formula.
[Text] Prognostic schemes using the full equations of hydrothermodynamics
have now come into broad use. It is known that the main difficulty in ap
plying these schemes is the filtering of ineteorologically insignificant
noise, which to the greatest degree exerts an influence on the values of
Cwodimensional divergence, and accordingly, vertical currents. The employ _
ed filtering methods for smallscale disturbances make it possible to devise
prognostic schemes which are stable in computational respects and to obtain
a reliable description of the geopotential and wind fields. However, use of
these fields for computing differential characteristics of the type of di
vergence is accompanied by considerable errors. Therefore, the idea of
37
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
computing vertical currents using the vorticity equation, heat influx equa
tion, continuity equation and equation of statics from the following differ _
ential equation remains acceptable for practical purposes:
~
 , d'= d A
 (!r) d}~I(lna)2 C.A A,.1:s 9;
dC
(1)
where l= 14,58 � 105 sin y; C 0�~, to = `~C
1 ~ . Qt
mZ = R2 T 7)IgI2+ 2Q =1f �Z = l ( dx  dy
AT= _!u aT +v aT`l, A=_(u +
aQQ a~al
l dx dy / 4 t dx v dY
Here p is pressure; T is temperature; u, v are the horizontal components of
the wind velocity vector in a standard coordina.te system; 'Ya, yare the
dry adiabatic and geometric vertical temperature gradients; 5P is latitude;
R is the gas constant; g is the acceleration of gravity; ~ is reduced pres
_ surep tJ is isotaaric vertical velocity; S2a, SZZ is the vertical component
of absolute and relative vorticity; m is the stability parameter; .Q,is the
Coriolis parameter; ASZ , AT is the horizontal advection of absolute vortic
ity and temperature;,102 cm/sec the influence of smallscale turbulence on the clear
ing of a natural aerosol will be negligible. However, it must be taken into
account for broader beams and in socaZled "stagnation" regions arising dur
ing scanning [12].
In conclusion the author expresses appreciation to L. G. Akul'shin for as
 sistance in carrying out the measurements.
BIBLIOGRAPHY
1. Belyayev, V. P., Volkovitskiy, 0. A., Nerushev, A. F., Nikolayev, V. P.,
Pinchuk, S. D., Skripkin, A. M., "Experimental Investigation of Clearing
of a Fog by Laser Radiation With./l= 10.6�,m," IZVESTIYA AN SSSR, FIZIKA
ATMOSFERY I OKEANA (News of the USSR Academy of Sciences, Physics of the
Atmosphere and Ocean), Vol 11, No 10, 1975.
52
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
2. 13isyarin, V. P., Kolosov, M. A., Pozhidayev, V. N., Sokolov, A. V.,
"Interactton Between Laser Radiation in the UV, Visible and IR Ranges 
and an Aqueous Aerosol," IZVESTIYA WZOV SSSR, FIZIKA (News of Higher
Schools of Education USSR, Physics), No 11, 1977.
3. Volkovitskiy, 0. A., "Experimental Investigation of the Influence of
Radiation of C02 Lasers on a Droplet C7.oud Medium," METEOROLOGIYA I
GIDROLOGIYA (Meteorology and Hydrology), No 9, 1977.
4. Volkovitskiy, 0. A., "Complex of Experimental Apparatus for Geophys
ical Inves.tigations," METEOROLOGIYA I GIDROLOGIYA, No 6, 1965.
 5. Zuyev, V. Ye., Kuzikovskiy, A. V., "Thermal Clearing of Aqueous Aero 
sols by Laser Radiation," IZVESTIYA WZOV SSSR, FIZIKA, No 11, 1977.
6. Kobzev,,,V. V., Petrov, G. D., "Decrease in Attenuation oE Laser Radi
ation in a Fog," TRUDY MOSK. INST. RADTOTEKHNIKI, ELEKTRONIKI I AVTO
MATIKI (Transactions of the Moscow Institute of Radio Engineering, E1
ectronics and Automation), Vo1 4, No 40, 1969.
7. Korotin, A. V., Svetogorov, D. Ye., Sedunov, Yu. S., Semenov, L. P.,
"Formation of Zones of Clearing in Clouds and Fogs," DOKLADY AN SSSR
(Reports of the USSR Academy of Sciences), Vol 220, No 4, 1975.
g. Matveyev, L. T., OSNOVY OBSHCHEY METEOROLOGII. FIZIKA ATMOSFERY (Prin
ciples of General Meteorology. Atmospheric Physics), Leningrad, Gidro=
meteoizdat, 1965.
9. Romanov, G. S., Pustovalov, V. K., "Clearing of the Cloudy Atmosphere
Containing Water Droplets by Intensive Monochromatic Radiation,"
ZHURNAL PRIKLADNOY SPEKTROSKOPII (Journal of Applied Spectroscopy),
Vol XIX, No 2, 1973.
10. Sukhorukov, A. P., Shumilov, E. N., "Clearing of a Polydisperse Fog,"
ZHURNAL TEKHNICHESKOY FIZIKI (Journal of Technical Physics), Va1 43, No
S, 1973.
11. Sukhorukov, A. P., Khokhlov, R. V., Shumilov, E. N., "Dynamics of Clear
_ ing of Clouds by a Laser Beam," PIS'MA V ZHURNAL EKSPERIMENTAL'NOY I
TEORETICHESKOY FIZIKE (Letters to the Journal of Experimental and Theor
' etical Physics), Vol 14, No 4, 1971.
_ 12. Gebhardt, F. G., Smith, D. C., Buser, R. G., Rohde, R. S., "Turbulence
Effects on Thermal Blooming," APPL. OPT., Vol 12, No 8, 1973.
13. Glickler, S. L., "Propagation of a 10.6 Laser Through a Cloud Including
Droplet Vaporization," APPL. OPr., Vol 10, No 3, 1971.
53
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
UDC 551.510.42
COMPUTATION OF ATMOSPHERIC PROPAGATION OF EFFLUENT OF HIGH INDUSTRIAL
 SOURCES IN THE PRESENt;E OF INVERSIONS ALOFT
, Moscow METEOROLOGIYA I GIDROLOGIYA in Russian No 9, Sep 79 pp 4955
[Article by Candidates of Physical and Mathematical Sciences F. A. Gisina
and S. M. Ponomareva, Leningrad Hydrometeorological Institute, submitted
for publication 25 May 1978] .
Abstract: This article gives a brief review
of statistical data on inversions aloft for
different points and existing methods for
determining the concentrations of impurities
under such anomalous conditions. The authors
propose a routine method for computing the
propagation of impurities during inversions
aloft, based on network aerological data.
The influence of inversions of different itr
 tensity and thickness on contamination of the
surface air layer by effluent of high point
sources is considered.
[Text] It is known that in the case of formation of an inversion aloft
there is a considerable increase in air contamination at the earth by ef
fluent of high sources situated in the layer beneath the inversion. Since 
within the inversion turbulent exchange is considerably attenuated, this ~
layer acts as a"lid," preventing the scattering of impurities; at the
same time, in the layer beneath the inversion it is common to observe con
vective conditions favoring the transfer of impurities to the underlying
surface. Many authors consider temperature inversions as one of the main
components of the potential of air contamination [16]. At many places
the frequency of recurrence of such unfavorable conditions is extremely =
high. For Moscow, according to the data in [8], the frequency of recurrence
of inversions aloft in the lower 500m layer during the year is 3334%
(for observation times 0400 and 1600 hours); in the cold halfyear such
inversi.ons are observed still more frequently.
The relative increase in the degree of contamination in comparison with an
inversionfree situation is dependent on the structure of the blocking lay
er its thickness, intensity, altitude of the lower boundary, and also
54
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFrICIAL USE ONLY
the relative positioning of the latter and the level at which the impur
ity escapes. A study of these parameters is especially important for major
indtistrial centers where inversions are characterized by a great thiclcness
and intensity. According to [11], for Leningrad the height of the lower
boundary of inversions aloft falls in the interval 100400 m in 84% of
the cases at nighttime (observation time 0400 hours) and 74% during the
daytime (1600 hours). Such conditions prevent the upward transfer of im
purities even for high stacks.
Until now no adequately reliable methods have been developed for computing =
concentrations under the anomalous conditions of propagation of the inversion
aloft type. Several studies are known which are devoted to a theoretical so
lution of the problem, usually employing numerical experiments. Within the
framework of the socalled statistical approach the effect of inversions 
aloft is taken into account extremely schematically. Source [6] gives two
formulas for computing volume concentration in the presence of an inversion
aloft, but it is noted that the results obtained using tliese formulas differ
substantially and it is impossible to make a choice between them without ex
perimental checking. The indeterminancy increases since there are no recom
mendations on the choice of the dispersions O'y, O'Z entering into these
formulas in the case of formation of a blocking layer.
Within the framework of the diffusion approach the influence of inversions
aloft was examined in studies [2, 6, 131. Heines and Peters [13], for deter
mining the concentration of impurities, used a threedimEnsional turbulent
diffusion equation in which the exchange coefficients in the vertical and ~
transverse directions were considered constant with altitude but proportion
al to distance: Ky = Ayxq, KZ = AZxn. The inversion layer was assumed to be
impermeable for the impurity, that is, it was assumed that Che flux Kz~
a.t the lower boundary of the inversion is equal to zero.
The computations were made for different source altitudes. The wind velocity
was stipulated constant; the coefficient Ay = AZ and n= q. Tt can be seen
from the computed data that the influence of the inversion layer increases
with increasing distance from the source q.id an increase in the ratio of
stack height to the altitude of the lower boundary of the inversion Hu. With
h/Hu = 1 the concentration in the presence oi an inversion layer increases
by a factor of three at the distance x= 10 lan, or by more than an order of 
magnitude with x,> 100 km in comparison with an inversionfree situation.
Monograph [2] gives some results of numexical experiments for investigating
atmospheric contamination in the presence and absence of inversion layers
aloft. For computing the concentration use was made of a thr.eedimensional
= turbulent diffusion equation in which wind velocity was considered variable
with altitude in accordance with a logarithmic law, the horizontal diffusion
coefficient was stipulated proportional to wind velocity Ky = au, and the
 profiles of the vertical diffusion coefficient were postulated taking into
account Che sharp attenuation of exchange in the inversion layer. A study
was made of five characteristic K(z) profiles, for each of which
55 _
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
the ratio of the surface concentration was computed in the presence and
absence of an inversion layer. Unfortunately, for the computed cases
there is not always an indication or' the parameters with which the numerical
estimates of the effect of inversioas (wind velocity, source height) were
obtained. This makes difficult their use and interpretation. It is also not
clear what vertical distribution of the turbulence coefficient corresponds
to an inversionfree case. It dollows from the data cited in [2] that in
dependence on the considered profile the surface concentration in the
preserce of a layer with attenuated turbulence over the stack can increase
at the distance 1 x 10 km by a factor of 11.9.
The objective of this study is a numerical investigation of propagation of
an impurity from high sources in the presence of uplifted inversions, tak
ing their actual structure into account. Computations of the contamination
, 1eve1 in the surface layer corresponding to each type of inversions with
operation of sources of different height were carried out in the following
way. The wind velocity profiles and the turbulence coefficient, as well as
the height of the mixing layer, determining the intensity of exchange, are
found from solution of a closed system of equations for a stationary, hor
izontally homogeneous atmospheric boundary layer, formulated in [5]. The
velocity of the geostrophic wind, roughness parameter and Coriolis force
 are considered known. In addition, on the basis of radiosonde data there
 is stipulation of the ac�tual temperature distribution. By such a method,
using a semiempirical model, it is possible first to find the profiles of
the principal turbulence characteristics in the boundary layer for differ
ent cases of temperature distribution with height, including for different
_ types of uplifted inversions. In such a formulation the problem was for the
first time solved by us in [3].
The distribution of concentrations of a weightless impurity arriving from
a high source is then computed by numerical integration of the twodimen
sional diffusion equation u as = a K as
vx , aZ . = a:
and computations of the volume concentration q(x, y, z) using the formula
[11]
9(x, Y, z) = S (x, z) eXP YZ/V`l r.oy'j/l r.oy .
(2)
In the case of presence of a blocking layer over a source use is made of
an ordinary boundary condition equality of the concentration of impur
ity to zero at the height of the mixing layer, determined in the course
of the system of equations for the boundary layer as the level at which
frictional stress is less than 10% of its surface value. Support for such
a formulation of the boundary condition is given, for example, by the ex
periments of Olsson [15], indicating that the level of penetration of the
impurities can substantially exceed the height of the lower bourtdary of the _
inversion.
56
FOR OFFICIAL USE ON,LY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
Table 1
Ratios of Surface Concentrations in Presence and Absence of Inversion for
Stacks With Height h= 50200 m at Different Distances from Source
X KA!
0,6
]
2
3
5
6
KEY : 10
1. Cases
2� Hmix
_ 1 Cnyaasf
A g .
AH90=300 Ac, DH=100: 450 at,
er= 1,7� c, oT=2,8� c,
2 NQ=426 M Hu 2G8 .4t
50
1 so I
ioo
50
~ so
0,5
0,8
0,3
0,6
0,1
0,4
1,4
2,4
2,(
5,7
1,2
1,4
1,0
1,3
1,1
1.3
3,5
3,9
;
;,, �
0,04
p,3
gXon:~ch
93oaOTOR ; I
33.2
:S3
U
,4 1,1 0,7
1.5 1,4 0,7
1
U,2
0.2
~.1 .
p,.l
O.pg
0,06
,
,0 0,8 O,G
U,2
0,1
0,04
30 ' B paioHe craHUNx frer croxa npecH~tx aoA.
KEY :
_ l. Sea
2. Station
16. Moshchnyy
3. Probability
17. Murmansk
4. Sea of Azov
18. D. Zelenetsk
aya
 5. ~'.ral Sea
19. Kandalaksha
6. :daltic Sea
20. Tyuleniy Isl
and
7. Larents Sea
"
21. Astara
22. Ogui~chinskiy
Island
8
. Wt;ite Sea
23. Okhotsk
9. Cas~;ian Sea
24. Sevastopol'
10. Sea of OkhotG
25. Yalta
11. Black Sea
26. Odessa
12. Sea of Japan
27. Vladivostok
13
. Mysovoye
28. Kholmsk
 14. BarsaKel'mes
29. Zolotoy
15. Klaypeda
30. * In the station
region there
is no runoff
of fresh
water
'1'lie accuracy of an individual salinity determination is t0.10

/00;
with a dis
, creteness of 24 hours this is ensured in
ity values also re
i
8095% of the cases.
Salinity variabil
ma
n considerable
tions to 1 hour. This is indicated
with a discreteness of it
hy Q S
val
i
h
s determina
50, 68 and 95%, cited in Table 2,
1
ues w
t
pr
obab
ilit
ies 1, 5,
79
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
This table shows that the salinity variahility at the 68% probability lev
el, even with a discreteness of determination of 1 hour, considerably ex
ceeds 0.020/0o and for the considered 6 points is 0.071.40/00. An in
crease in the variahility values in hour time intervals in comparison with
24hour intervals (Murmansk, Odessa) is attributable, on the one hand, to
a change in the site of taking of samples., and on the other hand, the con
siderably greater and longterm series used in processing the roadstead
stations. The change in the place of taking the samples also exerts an in
fluence on mean salinity (Murmansk, Odessa). The A Sl values show that with
a considerahle decrease in the discreteness in determining salinity to 1
hour an accuracy of f0.020/0o is excessive.
Table 2
~
8
KEY:
1.
2.
3.
4.
5.
6.
7.
8.
3 AS, o6ecneyet+xocrbto, %
o o
MaPe
PeiraonaR
cTatiuxs
S
�luo
A SI maz
�fw
1 I
5 I
50 I
68 I
95
~ q~
1
2
I
Asoecxce
TeMptoh
10,9
5,3
5,3
I 2,6
0,4
0,3
0,2
648
6apeHUeno
MyptiiaucK
I6,5
8,7
8,7
6,2
1,3
1,4
1,0
1430
I(3ClINHCKOC
OPYP4H7{GICHN
13,5
0,9
0,8
U,:)
0,1
0,07
0,04
997
2 H36epr
11,6
3,8
3,8
1,8
1,3
1,0
0,7
1308
qepxoe
3 Tyance
17,5
1,1
1,0
0,6
0,1
0,07
0,05
1030
ORecca
14,7
8,6
3,6
4,3
1,2
0,9
0,7
1099
Sea
Roadstead station
Probability
Number of observations
Sea of Azov
Barents Sea
Caspian Sea
Black Sea
9. Temryuk
10. Murmansk
11. Ogurchinskiy
12. Izberg
13. Tuapse
14. Odessa
d~A
6
2'
S 10 20
Fig. 2. Probability of values of spatial variability of salinity between hy
drometeorological stations Neftechala and Svinoy Island.
Spatial Variability of Salinity
In order to determine the variability of salinity along the shoreline of thz
coastal zone of the sea we computed the diffQrences (4 S'24) of its synchron
ous determinations at adjacent stations for the 15th of each month during the
80
FOR OFFICIAL ASE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
~
M
ce
0
~
a
H
~
t' C`')
~ p M M O p p ~ CV LV
M
O O O O O O O O O O J O O O O O O
O O O O O O O O O O O O O O O O O
c�~ c"~c c",i c~ 2 ami c`r'o  in ~n In Ln o ~n Ln 0 ln
r c co rn ull w
^ ^ N
o I er CO N l N fN N ^M M M C`1 t0 'C t_
~ O p O O O O O O O C O O O
4m
d
R.
O
O
M
M
CV
tC~
tf~
C?
O
C'l
n
r?
a
GO
O,
~
M
M
'
O
N
4
~
,
cj
N
CO
O
r+
O
O
O
O
O
O
O
O
_
a
w
~
o
~
l
M
~t
tD
N
00
h
tD
O
N
M
O
O
eh
00
^
~
u'1
LO
V
LO
CNj
O
O
O

C)
tf.)
c'7
C'M
~
~
~a
F
V
~
C3.
cti
m
0
'cl J
4
N
cz ~
a '
~
~
~
e?
~n
rl
.s
~
s
~
H
u
M
a
~a
C
.t u'1u1~ %DI~ ~00 00mO~tO O rFi N N cy1M ,t ,t ulul %.O~O I~ I~ CO o0 a%O)00 rl
N NN N NN N N('~kV N M C`'1 MCrI rl M MC'r1 C'') M McnMcn(''1C1 c'7 C'nM rl ,td' ,t
d d Y W.
v ' ~C X 11 1:: 'S +y X ~
X x Cl. GW a CC
U
V a,
u~ oG p,b0a~y u... 0 v0x,~
~ ca Rf ca c~ al m 5 S 00 o~ c.~
c~ p, C, o O~ a o o N h u u o 0 4
~c~uqrt Ct~YSt:~. cvcc ~c~uu~i~~ooc~a 0c~ ~cCSM ycraH
tI ~1 y~~C X O O 4 a c7 m Gi C1 .'ri O K K O O y y y, T~ 7 O O cc
C .
LOOFFvU:r:r'.; c:.lci.lUUY,x txm2$~�= OOCCcn
O
O
O
p
O
O
C~
O
0
p
p'
p
O
p
O
O
tn
o
o
p
LO
~
~
~
tp tr
M p S O O O O d
O
p O G O O O O O
O O O P l13 tCJ v~ O
rn ti ~ ti r Cl CO CO
a o cn cC cl o
~
a
a: r cv ci o ~ Q ip

nj N  CV CV N C
~
^ Cj
a
y
~
O
~
oq CV tA ~
z
N r!' O O ^N N
d
~
O
U
t'J
C
0
w :/J N �i' C'1
O ''D :D C. to O
�J C`t ~r M M e'
a
o
C.,)
~
ai
0
x
u
~s
z
~
M
~
a a ca
0 cd
c7 ta ~O O ILl L7 F,
O 0 Ct U u0 0 c7 ca p
O O p f~ ~ y G
p
0.0
Y
x x c. c.
y CJ
; y
,
�
O
O
w ca cv K U U ~ ~4 Y R
G
~ L..Fn.
~ r~ a. a~ ~ x
~
U V
~ d CCa. J a~ C C 5L CJ u u c6
A~ CJ Cl t~. t7 13
r:3 .j.
U~ V
%pV D0MM O'% 00 rI iINN c+"1C7 T u'1,,0~.p I,,I,000001 O% 00 rIrINN
rlrirlrlrlrlrH rl rirlr1 rIrIrIrirl riNNNNNN
81
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
KEY TO TABLE 3
 1. Pairs of stations 22. Izberg
2. Distance L, km 23. MakhachkaJ..,
3. Sea of Azov 24. Primorskoye
4. Caspian Sea (western shore) 25. Odessa
5. Black Sea 26. Tendra
6, Taganrog 27. Skadovsk
7. 'Lhdanov 28. Khorly
8. Berdyansk 29. Chernomorskoye
9. Mysovoye 30. Yevpatoriya
10. Opasnoye 31. Sevastopol'
11. Temryuk 32. Khersones
12. P. Akhtarsk 33. Yalta
13. Yeysk 34. Feodosiya
14. Taganrog 35. Anapa
15. Astara 36. Novorossiysk
16. Neftechala 37. Gelendzhik
17. 0. Svinoy 38. Tuapse
18. Baku 39. Ochamchire
19. Zhiloy 40. poti
_ 20. KyzylBurun 41. Batumi
21. Derbent
period from 1961 through 1970. The L~S'24 probability values were computed.
As an example, Fig. 2 shows the probability curve for d S'24 between Nefte
chala and 0. Svinoy stations (Caspian Sea), situated at a distance of 70 km
apart. A study of the curve shows that the difference in the salinity values
between these stations at one and the same moments in time attains 7.7~/00
with a probability S%; 4.30/0o with a probability 50% and 2.30/0o with a
probability 95%, that is, on the average in this water body salinity varies
by 0.060/0o in a distance of 1 km. The A S'24 vaiues with a prq~bability 5,
50 and 95% for pairs of stations in the Sea of A2ov, Black Sea and along
the western shore of the Caspian Sea are given in Table 3, from which it
f.ollows that with the existing positioning of the sea stations along the
s.hore in not one of the considered seas is spatial interpolation of salin
ity in the coastal zone of the sea ensured with an accuracy to 0.020/00,
Even in order to compute the salinity field between the stations with an
accuracy to 0.10/0o the disrances between the stations in the Sea of Azov must
average 2 km, in the Caspian Sea 3 km, in tihe Black Sea 10 km.
This investigation makes it possible to draw the following conclusions:
= 1. The temporal variability of salinity in the coastal zone of the seas is
considerable and even with determinations each 1 hour is considerably great
er than 0.020/00. Therefore, witi the discreteness of 24 hours adopted in
the network of sea stations the required accuracy in determining salinity
of f0.020/0o is unjustifiably high.
82
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
2. With the existing spatial positioning of sea coastal stations at consid
 erable distances from one another the interpolation of salinity along the
coast is impossible with an accuracy to f0.020/00.
3. In the network of sea shore stations it is necessary to r.eplace the ar
gentometric method for determining the salinity of water by the areometric 
method, which makes it possible, using the tables in [3], to convert the
specific gravity of water into salinity with an accuracy to 0.10/00 or to
determine salinity by electric salinometers with the very same accurlcy [1].
4. The argentometric method can be retained at several "secular" statl_ons
in the sea located in water areas little subject to the influence of the 
fresh runoff of rivers. �
BIBLIOGRAPHY
l. Ivanov, G. S., Ovsyannikov, A. N., "Variability ot Sea Hydro:logical L:le
ments and Computation of the Discreteness of Obsexvations," METEOROLOG
 IYA I GIDROLOGIYA (Meteorology and Hydrology), ido 11, 1973.
 2. NABLYUDENIYA NA GIDROMETEOROLOGICHESKOY SETI SSSR (Obsexvations i.n the 
Hydrometeorological Networlc USSR), edited by 0. A. Gorodetskiy, Lenin
grad, 1970.
3. OKEANOGRAFICHESKIYE TABLITSY (Uceanographic Tables), Leningrad, Gidro
meteoizdat, 1975.
83
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
UDC 551.466.2
COMPUiAiION OF PARAMETERS OF RESONANCE FRONTAL WAVES
Moscow METEOROLOGIYA I GIDROLOGIYA in Russian No 9, Sep 79 pp 7175
[Article by Candidate of Geographical Sciences G. G. Kuz'minskaya and T. I.
Tsareva, Sochi Wave Research Station, submi;:ted for publication 11 January
1979]
Abstract: Since atmospheric disturbances usual
ly occur on tropospheric fronts, in a case when
pressure gradients over the sea are weak there
 is basis for relating wave development directly
to movement of a front. G. V. Ivankov, G. V.
_ Matushevskiy and G. V. Rzheplinskiy have advanc
ed the hypothesis of formation of such waves
when the sea surface is affected by fluctuations
of atmospheric pressure. The authors of this ari
ticle have made computations using formulas pro
posed by the mentioned authors. The results of
the computations coincided with field observations.
[Text] It is known that atmospheric disturbances in the temperate latitudes
arise for the most part on main tropospheric fronts, that is, on the dis
continuity between tropical and polar or polar and arctic air. Some cy
clones arise under the direct influence of the underlying surface; usually
its influence is an additional factor. Under the conditions prevailing in
the eastern part of the Black Sea an additional front develops, passing
parallel to the line of mountains and caused by their influence.
It is also known that the intensity of waves increases sharply in frontal
zones. It is noted in [1, 46] that high (to hurricane force) wind velocit
ies are associated more frequently with cold fronts primary and secondary,
then with warm fronts, less frequently with occluded fronts. The slope of
 the atmospheric front to the horizon is very small: from 0.001 to 0.01.
Therefore, although a frontal line is plotted on the chart of surface analysis
of synoptic processes, in actuality this is not a line, but a region of sep
aration of air masses. In this region there are strong turbulent air move
ments.
84
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
The aut}iors of [3] analyzed the synoptic conditions causing one of Lhe
cases of waves in the Sochi region in March 1973. In the considered case
the atmospheric pressure gradients over the sea were very small, the winds
were weak and could not cause the waves with a height up to 4.4 m which
were observed in nature. The authors postulated that these waves were
caused by the passage of a cold front. In [2] such waves are caused by res
onance waves because it was assumed that they were generated by fluctua
tions of atmospheric pressure during the passage of a cold froic.
That same study gave a formula for the computation of the frequency O' of
resonance waves
K
o~ J, (1)
where U is the velocity of frontal movement and it was noted thzt spectral
energy density increases with an increase in U; the value of the scaling
factor a according to data for wind waves is
aU", n=46. (2)
Now we will compare our field data on the wave elements during the passage
of atmospheric fronts and weak winds over the sea 5 m/sec) with the results
of computations using these formulas.
The period of the waves for the maximum of spectral density is
2 g 6,24 U, .{rjceK 6,23 U, rc.+tlq km/hour.
U i 'max, CBli 81�3,6 ^ 0,18 U
mat
[cek = sec; km/y = km/hour]
_ Using the dependence of the mean wave period t on GmaX, i= 0.85t maX, we
obtain
sec z0.15U, km/hour. (3)
Assuming the squarcof inean wave height h to be proportional to the maximum 
of the spectral density oi energy, and the spectral width to be inversely
proportional to U, in accordance with (3) we obtain
n1
h.~cU ` , (4) 
c
, lrs% = 1,95 Iz.; cU 2�1,95, (5)
and the coefficient c must be determined using experimental data.
At our disposai we had six cases of the appearance of waves with the pas _
s3ge of a cold front in the region of Sochi and Novorossiysk (approaches
to Sheskharis). One of these cases (the third in Table 1) is described in
_ 85
F'OR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
, [2, 3]. A comparison of the wave elements with the velocity of frontal
movement for these cases is given in Table 1.
 Table 1
Heights and Periods of Resonance Frontal Waves Arising During Movement of a
Cold Front in the Black Sea
1
4
4
5
4
Merro, AaTa 11 apema ttaGnroAeaixa I
U x.w/K 2
I h,u I
hs y.~t I
c cex 3
C04N, 18 III 1970, 18 H 30 .uuK
58
1,50
2,8
8
0
Co4x, 28 Ia 1970, 12 q 30 murc
35
0,60
1,0
,
6
0
uo;nr, OS III 1973, 18 q 00 Afux
65
1,76
3,4
,
4
9
llIecxapiic, 08 \lI 1975, 09 K OO HllH
46
0,50
0
1
,
Covst. 28 IV 1976, 17 K 00 Mux
35
0,46
,
0
9
4
6
Coeu, 27 IV 1977. 07 it 00 uuK
50
'1,50
,
2,4
,
7.4
KEY:
1. Place, date and time of observation
2. km/hour
3. sec
~ 4. Sochi...hours...minutes
5. Sheskharis...hours...minutes
For the third case in Table 1 h was assumed equal to hmaX/2.5, and h5% _
1.95/2.50 hmax when hmax  4.4 m in'accordance with [2]. The mean period
of the waves is assumed equal to 0.85c max with timaX = 10.511.4 sec.
Resonance frontal waves developed in those cases when the atmospheric pres
sure gradients were small or there was a wind from the shore. ,
ns% M r ceK sec r a) r fi, ,
3 9
x x
2 7
x
~ x x $ x x~
Z
m/sec _
0
zo 40 60nlren~:0 40 U 1fM14 lan/hour
Fig. 1. Height (a) and periods (b) of waves arising during the movf'msnt of
a cold front in the eastern part of the Black Sea. 1) according to field
(experimental) data; 2) according to computations using formulas (1) and (6).
86
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
1
` ~ / ,10i 3 ~ 
1 Q) ln075 ` 9005V4,~ ~UG2.s ~,9000~' ~ 3975
9010 ~ \ ~
1 � ~ / _.J`7 i.
I '�'n~C.. ~ ~ \ ` I
, . �,~,y~ ~ ~
~ ti' s ~ � ~ ~ �
H 99lr ~
~ � ~'J (
. ~ ~ ~ ~
\ ~ � `4 _ 991 s ~
'N:D1U I
. ~ ' i ~J'~f:� 1~
: � / .
~
7010 i t. 10OZ,f /
OD.�', J : . 9~ 1003
~ 101 1006
9003 /
11oo2,s
9000 . � .
007~{j~
� `
.
~ y ~
.~~u \ ' ~ \
.
~ 4 ;
. .
� r ~ \j
99', S
,
Fig. 2. [dind field on 18 March 1970 at 1200 (a) ancl 1500 (h) hours. The
figures indicate computed points. H= low
The first case includes the observations of 18 March 1978, wlieii computation
of the waves from the wind, ccmputed using the pressure gradient, gives
h5% = 1.0 m(2.8 m was observed) and the observations of 8 I)ecember 1975
when allowance for the pressure gradient leads to a wind ve).oci_ty of 3 m/
sec (hs% = 1.0 m was :.?oserved).
The remaining observations belong to the second case (w1.nd fr.om the shore).
The data in Table 1 are plotted in Fig. 1, from which ttie erupirical coef
ficients in formula (5) were selected. With these :oefficients taken into
_ account the following formula can be recommended F,rr cocnputing wave height
87
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFiICIAL USE ONLY
hbx M=1,4 � I05 U3, 1cm/hour. (6)
Comparing (4) and (6), we obtain n 1/2 = 3, that is, n= 7, which does not
differ significantly from the n value obtained using formula (2). 
Sometimes the frontal movement is associsted with movement of a small cyclone
on the periphery of a large cyclone. This case is a manifestation of the
Fudziwara effect. 'L'he Fudziwara effect, named after the Japanese physicist
who discovered this phenomenon, invoZves a change in the direction of move
ment of one cyclone under the influence of another cyclone or anticyclone.
If two centers of action in the atmosphere are at such a distance from one
another that the regions of their circulation intersect, each of them act
on rhe other. Here three cases are possible:
a) if the intensity of both cyclones is identical, they begin to rotate one
relative to the other in a counterclockwise direction and the center of ro
tation is situated in the middle between the centers of the cyclones;
b) if the intensity of one cyclone is small in comparison with the inten
sity of the other, the rotation will occur about the center situated near
the large cyclone and accordingly the small cyclone will move more capidly
than the large one;
c) if the cyclone encounters a large anticyclone it will begin to move around
this anticyclone in a clockwise direction.
The field of atmospheric pressure and the wind over the sea for a case de
termined by the influence of the Fudziwara effect.are shown in Fig. 2. Un
der the influence of a large anticyclone situated to the west of the Black
Sea the movement of the small cyclone began. At 1200 hours on 18 March 1970
its center was situated in the Sea of Azov region, and a clearly expressed
atmospheric front passed in the central part of the Black Sea. The gradients
of atmospheric pressure for the sea area were nonuniform: in the neighbor
hood of Kerch Strait 14 mb/degree of ineridian, in the southeastern part
_ of the sea 0.5 mb/degree. After 3 hours the situation had changed. The
center of the cyclone and the region of high pressure gradients were displac
ed eastward from the Black $ea, to the'C'ucasus, and the cold .front moved
into the eastern part of the sea.
In this region the movement of a cyclone with a southerly component is un
usual. It was caused by manifestation of the Fudziwara effect (case "c").
The quite good agreement of the results of computations of waves on the
basis of frontal movement and observations is evidence in support of the
considered computation method (see Fig. 1).
88
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
]3I13LIOGRAPHY
1. Glazova, 0. P., Bel'skaya, N. N., Kuznetsova, S. A., "Some Character
istics of Aerosynoptic Conditions of the llppearance oL I{urricanerorce
 Winds," TRUDY GIDROMETTSENTRA SSSR (Transactions of the USSR Hydr.o
meteorological Center), No 139, 1974.
2. Ivanenkov, G. V., Matushevskiy, G. V., Rzheplinslciy, G. V., "Resonance
Excitation of Surface ldaves in the Sea by Cold Atmospheric Fronts,"
IZVESTIYA AN SSSR, FIZIKA A'TAfOSFERY I OKEANA (News of the USSR Acadeiny
of Sciences, Physics of the Atmosphere and Ocean), Vol XlII, No l,
 1977.
_ 3. Rzheplinskiy, G. V., Matushevskiy, G. V. , Yeslictienko, T.. A. ,"Unusual.
Swell Waves," METEOROLOGIYA I GIDROLOGIYA (Meteorology and IiydroJ.ogy),
No 3, 1975.
4. Sirotov, K. M., "On the Status and Developcnent of Investigations of
Waves in the Ocean," TRUDY GIDROMETTSENTREI SSSR, No 119, 1975.
5. Sirotov, K. M., Sett, L. S., "Curvature of Air rlows and Wind Waves in
the Ocean," TRUDY GIDRCMETTSENTRA SSSR, No 83, 1971.
6. Sirotov, K. Mob Sett, L. S., "Evaluation of rorecasts oE Cdaves in the
 Ocean," TRUDY GIDROMETTSENTRA SSSR, PIo 119, 1975.
7. Sitnikov, I. G., "BETSI," "KAMILLA" I DRUGIYE URAGANY ("Betsy," "Cam
ille," and Other Hurricanes), Leningrad, 1975.
89
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
UDr 551.578.46
VARIABILITY IN SNOW DISTRIBUTION ON ICE IN TEIE ARCTIC OCEAN
Moscow METiOROLOGIYA I GIDROLOGIYA in Ftussian No 9, Sep 79 pp 7685
_ jArticle by Candidates of Geographical Sciences A. Ya. Buzuyev and I. P.
Romanov, V. Ye. Fedyakov, Arctic and Antarctic Scientific Research Inst
itute, subnitted for publication b February 1979]
Ahstract: On the basis of the results of
statistical processing of data from meas
urements of snow depth a study was made
of the variability of this element in de
pendence on hummocking and the age of ice.
[Text] The problem of how great a role is played by snow in formation of the
ice cover has been repeatedly discussed [7, 11, 13].
It is also known that under definite circumstances the snow cover on Arc
~ tic ice exerts an important influence on navigation and the work of avia
tion and auto transportation.
An enormous amount of data has now been collected from measurements of snow
depth on ice. Its generalization made it possible to obtain general ideas
concerning the variation of snow accumulation on ice and to define regions
with high and low snow accumulation [7, 13] and also to formulate procedures
for taking snow depth into account in calculations of ice thickness [5, 10,
11). In the mentioned studies reference is to mean snow depths.
 Considerabiy fewer investigations have been made of the variability of depths
and the peculiarities of the spatial distribution of snow on ice of different
age and hummocking. And this is not accidental. Most of the measurements are
data from standard snow surveys on shore ice, which are usually carried out
at several points without taking into account the scale and nature of spatial
changes of the consiciered element. Only the use of the results of special ob
servations carried out sporadically made it possible to obtain irLformation on
the statistical structure of snow distribution (correlation functions, dis
tribution law) for some age types of ice with different hummocking [1, 21,
to estimate the variability of snow depth on the shore ice in the northeastern
part of the Kara Sea [8].
_ 90
FOR OFFICIAL USE ONLY
I
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02108: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
The peculiarities of snow distribution on the ice in the Arctic Ocean can
be considered in greater detail by using additional materials obtained dur
ing recent years.
It must be stated that measuremEnts of snow depth are made by dif.ferent meth
ods. Therefore, before proceeding to an analysis of the results, we will
briefly examine the sequence for obtaining initial data and their volume.
a) Profile observations on the shore ice of arctic seas. Five measurements
are made each 100 m on two mutually perpendicular seglents each witti a
length of 500 m 19].
The site for the profiles is usually selected at a distance of 100300 m from
the shore.
The analysis was made using data for 34 polar stations (14 in the Kara Sea,
12 in the Laptev Sea, 6 in the E'ast Siberian Sea and 2 in the Chukchi Sea),
a total of more than 5,000 profiles run during the period 19601975.
Profile observations are made during the entire snow accumulation cycle, from
the moment of foru;ing of the ice cover (SeptemberNovember) to the total melt
_ ing of the snow (MayJuly).
The shortcomings of this materiai are as follows:
the number of ineasuring points does not change with time, that is, has no
dependence on the scales of variability of snow cover depth. Therefore, the
problem of the representativeness of changes for this region remains open;
in many cases there is no detailed description of the survey routes and
 accordingly it is frequently impossible to ascertain where the snow depth
was measured: on an even sector, slope of a hummock or its crest (in the
presence of hummocking).
b) Data from special snow surveys are made:
optionally on the "Severnyy Polyus" drifting stations;
in support of operations for delivery of freight to Cape Kharasavey (Feb
. r.uaryApril i9751977);
during expeditions on shore and drifting ice in arctic seas (MarchJune 
19601977).
Special snow surveys as a rule had the objective of obtaining detailed infor =
mation on the distribution of snow on ice, taking into account its age and
hummocking. The measurements were made along routes with an extent from 100
to 10,000 m with intervala between measurements of 1.5 or 10 m and usually
intersected all relief forms characteristic for the ice cover.
The carrying out of such detailed surveys is a timeconsuming task; they have
been'carried out in different seasons and in different regions and do not take
in the entire diversity in the distribution of snow on ice.
91
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
, I NY m I 
19 DECEMeER 1979 N0. 9. SEPTEMBER 1979
2 OF 2
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
r 
t~
I T ~
1~ G
i
i c,
;
~
~ 76, 7J
2
E.
Fig. 1. Regions of the Arctic Ocean within whose limits the generalized char
acteristics of the variability of snow depth were obtained. 1) place and
years of carrying out snowmeasuring surveys on ice in sea; 2) polar sta
tions for which data from profile snow surveys on the ice were used.
v y ~ 40 Q) a b ;
14 ~ 20
u o
a) ~
4+ o 0 20 10 0 10 20 30 3 20 90 0 90 20
.0 Fi
4c~
. ~ C O) c 2)
 v 20
~
= cr
~
Q
4' ZO 10 010 ZO JO 40� �20 l0 0 90dhcM
 Fig. 2. Histograms of distribution of deviations of snow depth Ah from their
mean value characteristic for twoyear (d), oneyear thin (c), oneyear aver '
age thickness (b) and thin (a) ice in DecemberFebruary. Along yaxis: Fre
quency of recurrence ,
c) Data from optional observations carried out at landing points on the ice on the "Sever" expedition (MarchMay 19741978). The method for these observ ~
ations was developed by I. P. Romanov and involved the following.
~ The peculiarities of snow distribution are evaluated first from aboard an air
craft by visual means in the region of the proposed landing and the charac
_ teristic sectors in which measurements must be made are noted. After the
landing of the aircraft 1020 measurements of snow depth are made in each
sector at arbitraril.y selected points on different relief forms (even sector,
freezing snowdrifts, ridges of hummocks, etc.).
92
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
~ The depth of snow in snow banks is measured at the base, in the middle and
 final part5; usually the measurements invo'lved 1012 snow banks.
 The information, averaged using data from the enumerated measurements, serv
_ ed as the initial data for analysis.
_ The prccessing of observational data was preceded by the breakdown of the
entire area of the Arctic Ocean into regions (Fig. 1). �
, In order to characterize the variability of snow depths in regions IVI we _
 used primarily materials from profile measurements; also given are data
= from numerous observations on the drif ting ice.
Due to the iow information content of individual series of profi_le measure _
ments first for each series we computed the values
n
I ht
h_ r= n ;A hl  1' J
when a= h= b,
(1)
o
L
where C is a proportionality factor, ah =00 (we recall that we are ,
considering the A h value).
~
96
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
Table 4
Values of Empirical Coefficients A, B
B03pdCTH01i
eiIA nbAa
,
4 OAxonersifsn
7
 TUHKHN
5 O,q110nerxxH
cpe,~HeH
8
TOlI=N6[
6 O,qIlOAC'IHHIi
8
TOJlCT6I}1
6 OANOrcerxxEi
TOl(CTbIFi
9
iiepjio.� +n"lu
ua6ntoA. cepxN
lHa6moq.
2 3
Cexrxbpb
xos6pb
Reha6pb
4~eapanb
Aetia6pb
(Peepanb
niapr
M311
199
349
" I
2: h;
.4 6 h~n;n j~l71JX
il c,Ir
n
c.v
,422 0,835 I IU S I2.1.1
1489 0,86 ~ i;;,p 1,8 ~ 63,2
315 0,687 526 10,869 10,7430,65ti I
1.5,
~~6,2
1.3 ,la, l
1.5 70,1
KEY:
1. Age type of ice 7. SeptemberNovember.
" 2. Observation period 8. DecemberP'ebruary
3. Number of observation series 9. MarchMay
4. Oneyear thin
5. Oneyear, average thickness
6. Oneyear, thick
According to [12], if the general mean h0 is greater than doubl.e the gen
= eral standard deviation a"o, the coefficient C is close to 1 and the math
ematical expectation M(h) and the standard deviation of the truncated nor
mal distribution and the corresponding general values of the normal distrib
ution are close in their values.
In the course of processing the material we obtained a dependence in the
form (y = f (h) .
The approximation was made by the least squares method for the function 6=
AhB (Table 4). The results of the computations agree coell taith the initial
data, and since in virtually all cases it was found that h> 2 a"h, as the
general characteristics in computations of the distribution of snow depths
on oneyear ice it is possib.le to use h, O'k assuming C= l,
The results of the computations (Fig. 3) agree fairly well with the general
ized observational data.
Thus, having information on the mean snow depths it is possi.ble to compute
the peculiarifie5 of its distribution on even iee by using the expressions
cited above. The situation is more complex with an evaluation of the in
fluence of hummock formations on snow distribution. Up to the present time
the method for carrying out snow surveys on ice with different degrees of :
hummocking has essentially remained undeveloped. This is attributable primarily to the complexity of ineasuring great snow depths in hummocked areas.
97
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
A generalization of the available measurements, made directly in hummocks
_ (all the measurements were made at the end of the snow accumulation per
iod) made it possible to ohtain the dependencP between the height Hhum of
the hummocks and snow depth h. The dependence is fairly well approximated
by the expression .
h  0,11 HT,30, (2)
[HT = Hhum] and the error in estimating the depth of the snow amidst tlie
hummocks is 20%. The checking of dependence (2) using data from snow sur
veys on the shore ice at Cape Kharasavey in 1477 (these data were furnish
ed through the courtesy of V. M. Klimovich) confirmed the correctness of
this empirical expression.
The cited results agree with data from extremely representative observations
of snow distribution in the vicinity of snow fences [3]. These observations,
in particular, established that in regions with intensive blizzards snow
fences cease to serve their purpose if the snow depth attains 2/3 of the
height of the structure. As noted in j41, a great influence is exerted on
snow accumulation by the "fetch" of the blizzard (in our case, the distance
between the rows of hummocks) and the width of the snowprotecting struc
 tures (that is, the width of the hummock ridges). On the basis of theoret
ical investigations, field observations and modeling it was possible to ob
tain expressions making it possible to calculate the height of structures
of a stipulated profile which will not be drifted over by snow, the width of
snow accumulation zones. The use of these expressions is also justified in
computations of the peculiarities of snow distribution on ice. In actuality,
the computed relative "excess" of the structures (having a section similar
to the section of a h.ummock ridge) above the snow surface is in the range
0.50.3 m. The "excesses" of the hummock ridges above the level of ttie snow
surface vary precisely in such a range.
The influence of a barrier with the height Hbar on snow accumulation ceases
= to be felt, according to calculations, at a distance of about 2030 Hbar�
The observational data indicate that the influence of hummock ridges on
snow accummulation ceases to be felt precisely within these limits.
In zones of breaks in snow barrier structures there is formation of banks
whose calculated height is appraximately 2/3 Hbar and the predominant length
is 2030 Hbar�
_ According to the observations of I. P. Rosanov, the snow depth in these banks
attains 100 cm and their extent is about 14 m, which in general agrees with
the data cited ahove. However, it must be taken into account that the broa.d
use of the results of investigations [3, 4] in computations of the distribu
_ tion of snow on ice is limited by the predominance of a random positioning
of the hummock formations and rotation of the ice fields. These circumstances
 98
FOR OFFICIAL USF ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
do not make it possible to introduce into the computations such an impor
tant factor as the predominant direction of transfer. Accordingly, the
peculiarities of snow distribution on ice with different degrees of hum
mocking are considered below as a random process.
On the basis of a generaZization of the results of ineasurements o:E tlle
heights of hummocks it was demonstrated that the collected data can be in
terpreted by a Maxwell distribution [6]. The probability density is cleter
mined by the expression x,
1
 `p (X) Q3 ~ n e 2Q7 , (3)
T
where the C"hum value is a parameter of the distribution and is expressecl
_ through the mathematical expectation x or the standard deviation d
x a
1, 596  0 , fi73 '
Substituting expression (2) into (3), after transformations we ootai.n an
expressiun making it possible, in general foxm, to evaluate the dist:ribu 
tion of snow in zones of kummock formations
 28,94 h
~ Y(h) = 119,81 hl'31 2 e 2 pT ~4~ `
aT 3 I" n
Thus, for regions with different hummocking the distributi.on of snow dept' "
on oneyear ice is determined by the dependence
f(h)=*(h)STIcP(h) (1ST), (5)
where Shum is the area occupied by the hummocks; 5P(h) is the snow distrib
ution on even sectors.
After knowing the general pattern of distribution of snow deptlls, it is
possible to solve the problen of the necessary number of ineasurements n
with which there is assurance of a stipulated accuracy in determining h. 
In conclusion, we wi11 examirie the variability of snow depths in the prin
cipal forms of snow relief; uniform cover on even oneyear ice in frozen
leads; zastrugi, banks, drifts on ice of different age; amidst hummock for
mations, etc.
The observational data used were the results of observations carried out
on the "Sever" expeditions. The initial data were generalized for regions
VIIXII (Fig. 1). It should be said at once that observations were made sys
 tematically for a total of five years, but for some elenents, no more than
two years.
99
 FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
Table 5
Mean Values and I,imits of Changes in Snoia Depths (cm) on Even Tce Sectors
(heven), in Hummocks (hhum) and L'or Zastrugi (hzas)
nP L hT 3 ft3
Pai~o i ~ I v Q, I a. I c
G
C U
i
~ 8 Tonruxita 6 Aa 3070 6` 5
VII
VIII
IX
X
XI
XII
Cpeltxee
9 no scem
paHOHaM
A
VII
VIII
. IX
X
XI
XII
Gpeltxee
9 ao acesc
vafiottaM
5
38
79
50105
29
273(
6
39
50
3473
15
102~
5
39
129
83 I70
29
1041
4
3
24
45
117$
11
32C
5
 (
76

21

Toa
vlxxa n
bAa
120160 c
M
6
411
58
3676
23
1930
6
58
64
58 75
20
1432
7
315
102
60136
22
1133
4
37
64
3590
13
1214
5
37
73
6389
21
1325
4
27
74
43115
22
1734
5 I 1 72 1
201 
clip
J
hr
4 h}
n.
I c
ci, I
C
o'
c'
I rCJ
c
u
~
u
.
8 TonucHtta .nbAa 70120 cx
7
4 10
63
4684
17
3* 25
39
57
4774
16.
9 23
4
26
125
65167
27
845
4
35
62
4280
17 `
730
3
24
64
28101
` 12:

5
74

18
TonMxxa n
bAa 160200 cx '
12
330
89
53170
41
2085
9
614
85
55150
30
1960
S
215
118
93143
19
830
G
49
37
93100
24
1933
6
38
73
60102
22
1530
6
47
100
70127
22
2122
g

92

26

KEY:
1. Region 7. mean
2� heven 8. Ice thickness...cm
3. hhum 9. Mean for all regions
4. hzas
5. mean
6, limits
The results of generalization of the data (Table 5) indicate that the mean
value and amplitude in snow dep.ths on even oneyear ice of different thick
ness vary in a relatively narrow range through all the considered regions.
One can only note a tendency to an increase in the mean depths witli an in
crease in ice thickness. The mean values and variability of snow depths in
zastrugi and amidst hutmnocked formations are considerably greater (Table 5).
The areas occupied by zastrugi and drifts vary in a wide range (Table 6),
_ but due to the limited amount of initial data it was not possible to give
 a reliable characterization of this variability.
Nevertheles.s, the results given in Tahles 5, b give a graphic idea concern
ing the principal peculiarities of the variability in the distribution of
the snow cover on ice in the Central Aretic. These peculiariti(:s supplement
= the quantitative characterization of the variability of snow depths which
was formulated on the basis of the results of profile measurements and
special snow surveys on ice of different age and hummocking.
100 '
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
Table 6
Generalized Characteristics for Zastrugi (Drifts) and Snow Banks in Frozen
Lead
10
2 17.701lI3,46, 33}[A738 3dCTPYC3NIi
3(apaKTCpHCTHK2 CHC?KHNX KOC:
,
(HB,QyB21NH)
PBROH
8 3d![EP3 4
H3 OCNOB 5
6
axcora ctiera7 B
1
W0![ P2360,qbP
HOM A61(y
An11N2, A(
hpcgX, GK
I
cpe.a
.
8
rrpelle
nu 9
cpe;tu.
npetce
~Peltx8
npeue
9
cpen~~'
IripeAe
nm
_
VII
28
080
24
080
16
550
50
25120
VIII
28
070
22
085
9
420
41
17100
IX
24
060
19
040
15
655
50
35100
X
32
040
18
060
12

50

XI
25
050
18
080
20
780
50
2080
XII
24
940
20
060
12
720
46
2075
CpeA
27

20
14

48

Hee no
ecem
paAo
xam
KEY :
1. Region
= 2. Area occupied by zastrugi (drifts)
3. Characteristics of snow banks
4. In frozen lead
5. On main ice
6. Length, m
7. Snow depth in banks, cm
8. Mean
9. Limits
10. Mean for all regions
_ BIBLIOGRAPHY
l. Buzuyev, A. Ya., "Statistical Evaluation of the Spatial Distribution
of the Principal Parameters of the Ice Cover," TRUDY AANII (Transac
tions of the.Arctic and Antarctic Scientific Research Institute), Vol
326, 1975.
2. Buzuyev, A. Ya., Dubovtsev, V. F., "Statistical Characteristics of Some
~ Ice Cover Parameters in the Arctic," TRUDY AANII, Vol 303, 1971.
3. Byalobzheskiy, G. V., et al., SNEGOZASHCHITNYYE SHCHITY I ZABORY (Snow
Protection Panels and Fences), Moscow, Avtotransizdat, 1961.
101
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
4. Dyunin, A. K., MEKHANIKA METELEY (Mechanics of Blizzards), Novosibirsk,
' SO AN SSSR, 1463.
5. Kirillov, V. A., Spichkin, V. A., Belen`kaya, S. S., "Prediction of the
Thickness of Shore Ice During Spring," TRUDY AANII, Vol 320, 1976.
6. Klimovich, V. M., "Characteristics of Hwmnocks on Shore Ice," PETEORO
LOGIYA I GIDROLOGIYA (Meteorology and Hydrology), No 5, 1972.
7. Loshchilov, V. S., "Snow Cover on Ice in the Central Arctic," PROBLEMY
ARKTIKI I ANTARKTIKI (Problems of the Arctic and Antarctic), No 17,
1964.
8. Nazintsev, Yu. L., "Snow Accumulation on Ice in t.he Kara Sea," TRUDY
AANII, Vol 303, 1971.
9. NASTAVLENIYE GIDROMETEOROLOGICiiESKIM STANTSIYAM 7 POSTAM (Instructions
for Hydrometeorological Stations and Posts), No 9, Part I, Gidrometeo
 izdat, 1968.
10. Sabinin, K. D., "On the Problem of the Influence of the Snow Cover and
Water Heat on Ice Growth," OKEANCILOGIYA (Oceanology), Vol 3, No 1, 1963.
11. Tsurikov, V. L., "Analysis of Growth of Sea Ice Under a Snow Cover,"
OKEANOLOGIYA, Vol 3, No 3, 1963.
12. Shor, Ya. B., STATISTICHESKIYE METODY ANALIZA I KOidTROLYA KACHESTVA I
NADEZHNOSTI (Statistical Methods for Analysis and Monitoring of Quality
and Reliability), Moscow, Sovetskoye Radio, 1962.
13. Yakovlev, G. I., "Snow Cover on the Drifting Ice of the Central Arctic,"
 PROBLEMY ARKTIKI I ANTARKTIKI, No 3, 1960.
Im
102
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL tiSE ONLY
UDC 556.048
CHECKING OF STATISTICAL HYPOTHESES IN COMPUTATIONS Or MAXIMUM WATER
 LISCHARGES WITH A LOW GUARANTEED PROBABILITY OF OCCURRENCE
Pioscow METEOROLOGIYA I GIDROLOGIYA in Russian No 9, Sep 79 pp 8692
[Article by Candidate of Geographical Sciences A. V. Khristoforov, Moscow
State University, submitted for publication 14 December 19781
 Abstract: The computation of maximum discharges
is usually reduced to extrapolation of the guar
1 anteed probability curves. The hypotheses on the
form of the law of distribution of probabilities
used in this case can serve as additional sources
 of computations. It is shown in this article that
even great deviations of the true guaranteed prob
_ ability curves from the hypothetical curves if
they occur only in the region of extremal values, can
not be reliably detected by any statistical means.
~ The author determined the range of the guaranteed
probability values within whose limits the obtain
 ing of any reliable evaluations is impossible with
out drawing upon additional information. The con
_ clusions clrawn are applied to a case when spatial
temporal generalizations are used in the computa
tions
[Text] The determination of maximum water discharges with an annual excess
probability from 0.1 to 0.0001, necessary for different water management
computations, is usually reduced to extrapolation ot the empirical guaran
teed p robability curves by the methods of the theory of random values and
mathematical statistics [l, 3, 6].
The general scheme consists of three stages:
1. Adoption of thF: Hp hypothesis concerning the form of the distribution law.
2. Evaluation of t:he unknown parameters determining the hypothetical distrib
ution in the series of observations.
3. Checking of the correspondence between the adopted hypothesis and the
sample.
103
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
_ FOR OFFICIAL USE ONLY
The adopted hypothesis stipulates the approxi.mating parametric family of
distributions (Pearson family, threeparameter gama distribution, etc.)
or determines the rule of use of spatialtemporal generalizations, for ex
ample, by means of generalized guaranteed prohability curves jl, 2].
Lach drainage basin and each hydrological station is characterized by its
own laws of fluctuations of runoff, dependent on the conditions for water
supply to the river and movement of the flow along the channel; therefore,
the deviation of the true distribution law from the hypothetical law al
ways exists.
 As a result, the total error in computing the maximum discharges with a
low guaranteed probability consists of the errors caused by instability of
the sample evaluations of the distribution parame*ers and the errors as a
result of deviation of the truedistribution from the hypothetical value
[3]. The problems involved in the accuracy of determination of the para
meters on the assumption of correctness of the adopted hypothesis at the
present time have been investigated in consiuerable detail [1, 8]. However,
the results of the erroneousness of the hypothesis, which can be particular
ly significant in computations of values with a rare frequency of occurr
ence, for the time being have been studied poorly. This article is dEVOted
to an evaluation of the possibilities of checking the results from the
adoption of hypotheses in computations of maximum runoff.
The principal difficulty in studying the total error in determining the
guaranteed probability curves for runoff in the range of rare frequency
of occurrence is as follows: even a very large deviation of the true
guaranteed probability curve from the hypothetical curve, if it occurs only
in the region of extremal values, cannot be reliably detected by any stat
istical means.
As a confirmation we will examine first the case of checking of the hypo
thesis on the particular form of distribution law obtained using an indi
vidual sample.
Assume that the hypothesis Ho, lying at the basis of the computation scheme,
stipulates the guaranteed probability curve in the form xp(p; 8 1,..., (9S),
where 8 1,...,..., e S are unknown parameters subject to evaluation using
the sample x1,...,xN. With an}s vaiues of the parameters the function xp(p;
81,..., (9S) decreases with an increase in p and its curve has a convexity
downwards with small values of the argument.
We will represent the true guaranteed probability curve in the form
_ x (P; E� . . , 9s, e:+11 . . . , 9s+r) = xo (P; 9� . . . , HS) (1)
,
+ (P), ,
104
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
where the functions 91(p),..., ~Pr(p) are equal to zero with p> po and are
monotonically decreasing convex downward in the segment [0, po]. The great
_ er the r value, the more possiliilities there are for describing different
deviations from the Ho hypothesis. In this particular case the Hp hypothesis
can be represented as
Ho : Eis+, _ . . . = 95+,  0. (2)
The deviation of the true distribution from the hypothetical distribution
occurs only in the region of small guaranteed probabilities p and will be
the greater the greater are the parameters 8 s+l,���, 0 s+r Wliich are not
taken into account.
The question arises as to the maximum admissible deviations of the true dis
tribution from the Hp hypothesis with which the methods of mathematical
~ statistics still cannot exhibit a contradiction Ustween the hypot}lesis
and the true distribution of elements of the sample x1,,..,xN.
Extremely widespread and with optimality properties is the criterion of
the probability ratio [5, 7], which refutes Hp in the case of high values
 of the statistics 'f :
N
rI f (zl; 9j*, . . . , AS*; AS+1 = 0, . . . , 8.5}r = 01 (3)
T=2ln t'
,
11 f(Xl; e7,
tl
where f(x; 81,..., e s+r~ is the probability density function, correspond
ing to the guaranteed probability curve (1), e i,..., e s+r are evaluations
of the parameters of the true distribution by the maximum probability method.
e'~S are evaluations of the parameters by the maximum probability
method on assumptian (2) of the Hp hypothesis.
If the Hp hypothesis 'is correct, the statistics'G has an asymptotically x2
distribution with r degrees of freedom. Otherwise the 2 statistics has an
approximately noncentral ~C2 distribution with r degrees of freedom and the
noncentrality parameter & .
r
e N ~ 9S+! /~l ) d I n f d l n j
,
j, 1=1 l d9sf t d d s+/ ~
where M is the mathematical expectation symbol [5].
The rollowing ratio follows from formulas (1), (4)
pn
: V i 1 I
u
xo (P: A,. . . . . As) ,s
~Fi (P)
~=t
105
FOR OFFICIAL USE ONLY
dp< p, N,'
(4)
(5)
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
a
 since the derivatives xp', r are negative in the segment [0,
, p0].
 Table 1
Maximum Values of Probability
Ratio Criterion
I J�Ipos('Hb 3H841IMOCT31 Q
r
0.01
I O,US I
0,10 ~
0,20
 1
0,03
0,11
0,18
0,30
16
0,03
0,19
0,30
0,45'
 .5o ~
1,05
0,16
0,27
0,42
KEY� 500
.
I0,02
0,06
U,11
0,21
1. S ignificance level
If the random value x2(r, F_ ) has a noncentral X 2 distribution with the
paramei:ers (r, F, the random value r+F_/r + 26 x2(r, E) has an approx
imately x 2 distribution with r+ E 2/r + 2� degrees of freedom [5]. With
& values not exceeding V/"r, the validity of this � criterion differs little
from the significance level ol, that is, the probability of refuting the
incorrect hypothesis Hp is only a little greater than the probability of re
futing Hp when it is correct. .
Table 1 gives the maximum N.(oL) values for the probability ratio criterion
correspond:Lng to the case 6 _ v/r.
Accordingly, if the true~distribution function differs from the hypothetical
function only in the field of extremely high values, whose guaranteed prob
abilities do not exceed po, and the critical pp value s atisfies the inequal
ity
Po=W; r(6)
then this criterion, despite its optimum properties, is unable to reveal even
the largest possible errors in computing the guaranteed probability curve
in the zone [0, pp].
The comp lex Ho hypothesis can be checked using the Pearson (X2) test [5, 7].
In this case there is stipulation of k> s+ 1 intervals A1,...,Ak and from
*
the sample it is possible to determine the frequencies pi,...,pk o* entry
into the corresponding intervals; evaluations of the parameters e 1,���, e s
are found f.rom the minimum of the function /G2( el, es) �
~ , .
[pi po 1 (e1, . . . , HS)]' (7)
S~ Po1 (di,
t_l
106
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
where poi (9s) are the hypotheCical probabilitie5 of falling into
_ the intervals Ai; i= 1,...,k.
The Hp hypothesis is refuted in the case of high values of the statistics 
X2( e i,�.�, e S)� If the checked hypotbesis is true, the statistics of the
 criterion has an asymptotically X2 distribution with ks1 de grees of free
dom. Otherwise the )t2( B i,..., e S) value has an approximately noncentral 
X2 distriliution with ks1 degrees of freedom and the noncentrality para
i meter 6 [5, 71, i
~r[PtP.~t 9s11'
o
.
j_1 Put ldl, . ~ Hs) (g)
where e 1,..., 6) S are the true values of the first s parameters in formula
_ ('1) , and
P1= pi
are the true probabilities of falling into Ai; i= 1,...,k, which coincide
_ with POi � l,���, (9 s) for all the intervals lying to the left of the
point xo, the guaranteed probability po.
The parameter S o will be maximum when the end of one of the intervals co
incides with the point xp.
The ()62) criterion is oriented on detection of the discrepancies between the
probability density functions and not the guaranteed probability curves. If ~ we limit ourselves only to those breakdowns A1,...,Ak for which large devi
ations of the hypothetical guaranteed probability curve from the true curve
correspond to a high significance of the test, that is, for which the non
~ centrality S is an increasing function of the parameters 8 s+l~���3, (9 s+r 
not taken into account, the significance of this criterion will be maximum
with a tendency of the parameters Bs+l,���, es+r to infinity. In this case
the true probabilities of falling into intervals lying to the right of xp tend to zero and accordingly, in this class of breatcdowns Al,...,Ak the non�
centrality parameter satisfies the inequality
a =0n =Po N. ~9~ 
Thus, with r= k s 1 formula (6), as for the Pearson ( ?C2) test, en 
sures the impossibility of checking the airs in computing the guaranteed 
 probability curve in the region of high values.
' As an illustration we will examine a three�parameter gamma distribution, de
termined by the mathematical expectation X, the variation coefficients Cv
and the asymmetry coefficients CS. We will denote the values X; Cv; Cs/CV _
2 by 81; 612; e3. The guaranteed probability curves, corresponding to
the values of the parameter e3 = 0; 1; 2 differ substantially from one
_ another in the region of low guaranteed probabilities only in the segment
j0, po], with p0 equal to 0.02. Thus, this case from the practice of hydro
logical computations can lae conditionally assigned to the situation describ
ed by the theoretical example (1), with s equal to 2 and r equal to 1.
107
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
 The hypothesis of an ordinary gamma distribution (CS/Cv = 2) corresponds
to the HQ hypothesis of equality of the e 3 parameter to zero.
In 14] tha statistical tests methoci was used in determining the validity
of the (x 2) test for checking the Ho hypothesis against the alternative
e3 = 7.;2.
It was found that with N 10 in order for the significance of the _
) criterion to exceed the significance level a by not more than 0.35
it is sufficient that the limit po of the region of discrepancy of the hypo
thetical and true guaranteed probability curves satisfy the inequality
. 1
Po ~ N+ (m, a) � (1$)
The values of the Y(m,aG) function are given in Table 3.
The function q(m,oL) was close to vrm and accordingly it can be concluded
_ that expression (12) also for the "guaranteed probahility of guaranteed
 probabilities" method leads to an inadequately great difference between
the significance of this criterion and the significance level. In a case
when the hypothesis Hp is complex the necessary statistical evaluation of
the unknown parameters makes the significance of the criterion still less
. [2, 41.
It must be noted that the presence of a correlation between rhe samples,
possible deviations from stationarity and other deviations from the clas
sical formulation of the problem favor a decrease in significance of the 
110
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
criteria [4] and accordingly broaden the limits of the zone of total inde
terminancy 10, po].
7f the critical value of the guaranteed probability pp satisfies inequality
(12), even the use of spatialtemporal generalizations does not malce it
possible to check the results of adoption of hypotheses on the form of the
distribution using standard statistical means.
Thus, we have found the limits of the zone of guaranteed probability wittiin
which there can be same reliable extrapolation of the guaranteed probability
IFcurves for computations of maximum runoff.
.
BIBLIOGRAPHy
1. Blokhinov, Ye. G., RASPREDELENIYE VEROYATNOSTEY VELICHIN RECHNOGO STOKA
(Distribution of Probabilities of River Runoff Values), Moscow, Nauka,
1974. 
2� Yevstigneyev, V. M., Zhuk, V. A., Kalinin, G. P.
RASCHET RECHNOGO STOKA PO OBOBSHCHENNYM KRIVYM pBESPECHENNOSTI (Com~
putation of River Runoff Using Generalized Guaranteed Probability
Curves), Moscow, Izdvo MGU, 1975.
3� Yevstigneyev, V. M., Zhuk, V. A., Khristoforov, A. V., "Accuracy of Com
putations of Maximum Water Discharges With a Rare Frequency of Recurr
ence," EKSPRESSINFORMATSIYA VNIIGMI_MTsD (Express Information,
VNIIGMIMTsD (AllUnion Scientific Research Institute of Hydrometeor
ological InformationWorld Data Center)), Pdo 3(47), 1976.
4. Zhuk, V. A., Yevstigneyev, V. M., Chutkina, L. P., "Peculiarities of
Use of Agreement Criteria in the Checking of Hypotheses on the Laws
of Distribution of Characteristic Runoff Values," PROBLEMY GIDROLOGII
(Problems in Hydrology), Moscow, Nauka, 1978.
5. Kendall, M. J.$ tewart, A,, STATISTICHESKIYE VYVODY I SVYAZI (Statis
tical Conclusions and Correlations), Moscow, Nauka, 1973.
6. Kritskiy, S. N., Menkel', M. R., "Evaluation of Probabilities of Fre
quency of Recurrence of Rarely Observed Hydrological Phenomena," PROB
_ LEMY REGULIROVANIyA RECHNOGO STOKA (Problems in Regulation of River
Runoff), No 6, Moscow, Izdvo AN SSSR, 1956.
7. Leman, E., PROVERKA STATISTICHESKIKH GIPOTEZ (Checking of Statistical
Hypotheses), Moscow, Nauka, 1964.
_ 8. Rozhdestvenskaya, A. V., OTSENKA TOCHNOSTI KRIVYKH RASPREDELENIYA
GIDROLOGICHESKIKH KHARAKTERISTIK (Evaluation of Accuracy of Distribu
tion Curves for Hydrological Characteristics), Leningrad, Gidrometeo
izdat, 1977,
111
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
UDC 551.515.3(47+57)
FREQUENCY OF RECURRENCE OF DUST STORMS IN THE TERRITORY OF THE USSR
Mosr_ow METEOROLOGIYA I i.IDROLOGIYA in Russian No 9, Sep 79 pp 9397
 [Article by Candidate of Geographical Sciences L. V. Klimenko and L. A. Mos
. kalEVa, Moscow State University, submitted for publication 15 February 19791
A'istract: The distribution and frequency of re
currence of dust storms in the territory of the
USSR are discussed. The article gives a map and
description of the regional peculiarities of
the appearance cf dust storms and defines the
role of economic activity in their appearance.
The annual variation of dust storms in the
places with their greatest frequency of recur
rence is discussed.
[Text] The literature has repeatedly touched upon the problem of the reasons
and conditions for the appeararice of dust storms, their geographical dis
tribution [2] and measures for contending with them; descriptions have been
' given for individual catastrophic cases and territories have been region
alized on the basis of the degree to which they are subject to dust storms
[1]. Some studies give an analysis of the frequency of recurrence of this
phenomenon for individual regions [3, 4, 6].
For example, a study by N. N. Romanov [4] contains statistical material on
the frequency of recurrence of dust storms over the territory of Central
Asia for a fiveyear period (19511955), maps were prepared showing total
fiveyear values, and the annual and diurnal variation of dust storms; their
genesis in dependence on different synoptic situations is analyzed. Yu. I.
Chirkov and M. M. Agafonova [6] prepared a map of the mean longterm number
of days with dust storms in the territory of the main agricultural regions
of the USSR (except for Central Asia).
 However, the available literature on dust storms does not give a sufficiently
complete idea concerning their frequency of recurrence and distribution over
the entire territory of the USSR where they are observed. Accordingly, an
attempt has been made to bring together the available materials for the
purpose of compiling a generalized map of the frequency of recurrence of
112
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
dust storms over the territory of the USSR (Fig. 1). The map was compiled
using data from the HANDBOOK ON USSR CLIMATE [5] on the mean annual number
of days with dust storms during a 25year period, for the most part from
1936 througli the beginning of the 1960's. For revealing greater detail con cerning the phenomenon we initially prepared maps of the frequency of re
currence of the dust storms in individual regions for all meteorological
stations listed in the corresponding handhook. Then the regional maps were ~
reduced to a common scale of 1:10,000,000 on which we drew isolines with _
the values of the mean number of days with dust storms 1, 5, 10, 20 or
more days. To the north, where the dust storms do not occur annually, their
frequency of recurrence is small and irregular, it was not possible to draw
continuous isolines of their mean annual number.
/
~ � MOCKB t ^
5 ~ _ 1' �
ra40 a~
, > > ~r:,_ ~ za ~ ~I '1� ~f 10 �'q
20
.Y:~.~S
30 i %~dJ r 1^ ~ 0 ? I S 10~ 70
20 4
r~ 5 rz
;,7
JO J fo � 1J
~v ~ Q s to
J, r o eS 5� 1~ 10
10 p 10 lj 3020 qD~ 10 ~
/ 33. 10
~
~010 020 1 s zo,,0zd0
~ , fo
 Fig. l. Mean annual number of dust storms in the territory of the Soviet Union.
In the European part of the USSR, within the limits of the steppe zone, it is
possible to discriminate two regions with an increased frequency of recurrence
and duration of dust storms southern Ukraine and southeastern European USSR.
In the Ukrainian SSR there are three clearly expressed maxima of the frequency
of recurrence of dust storms: first, in the KhersonKakhovka region, where
the number of days with dust storms attains an annua.l average of 17.3 (Nizh
niye Serogozy), second, in the Voroshilovgrad region and third, in the south
western part of Odesskaya Oblast with its center in the neighborhood of Sar
ata. 113
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
In the southeastern European USSR an increase in the number of dust storms
is observed within the KtunoManychskaya depression and the Sal's;ciye steppes,
where on the average there are up to 2023 days annually with dust storms
(populated places Gigant, Zavetnoye). 
In the delta and floodplain of the Volga strong southeasterly and southerly
ainds cause salt storms in the spring and autumn when there is dessication
of the lake: in spring to high water, and in autumn after the waters
drop off [6].
In the remaining territory of the USSR dust storms are propagated in the
southern part of Western Siberia, southwestern Central Siberia, in Kazakh
stan and Central Asia.
In the southern part of Western Siberia dust storms occur in Tyumenskaya,
Omskaya (Zshimskaya Steppe) and Novosibirskaya (Barabinskaya Steppe) Ob
lasts and Altayskiy Kray (Kulundinskaya Steppe). 
In the southern part of T'yumenskaya Oblast the mean annual number of days ~
with dust storms varies from 0.2 to 4, and in Omskaya Oblast from 0.3 to
1213 days, gradually increasing southward. In the example of this region
one can clearly see the influence of man's economi.c activity on the envir
onment and the processes transpiring in it. After being plowed in the _
1950's in the southern part of Omskaya Oblast the frequency of occurrence
of dust storms on the average increased by a factor of 2.5, and according 
to data from Bol'sherech'ye and PokrovoIrtyshskoye stations by a factor
of 56. In Novosibirskaya Oblast and Altayskiy Kray the mean annual nimmber
of days with dust storms reaches 2028. 
In the southern part of Krasnoyarskiy Kray and in the Tuvinskaya ASSR dust
storms are observed most frequently in MayJune, but also occur in winter
when there is a small snow cover or when there is none. The greatest mean
annual number of days with dust storms (up to 1012) is abserved in the
Minusinskaya Basin and in the northeastern part of the Khakasskaya Auton
omous Oblast.
Kazakhstan and Central Asia, with their extensive expanses of dry steppes,
semideserts and deserts, are characterized by the greatest frequency of
' recurrence of dust storms.
The distribution of the mean annual number of days with dust storms in
the e~ttensive territory of Kazakhstan is nonuniform and to a considerable
degree is local, which is associated not only with the wind velocity, but
also the churacter of the underlying surface. In steppe regions with light
"sandy soils as an average for the year .*_here are 2030 days with dust
.storms. In the regions ~f sandy deserts of the Bol'shiye Barsuki sands
ann *o the south. of Balkhash, i`,he SaryIshikotrau sands, the number of
sand storms increases from 30~;0 to 5060 days. In territories with a
heavy mechanical composition ol,~; the soils, and also in mountainous regions,
114
FOR OFFICIA;.: USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
_ the number of days with dust sto nns decreases to 2 days per year [5].
In Central Asia the gxeatest frequency of recurrence of dust storms is not
ed in Turkmenia; this is attrihutalile to the broad occurrence of poorly
hound sands, gray desert soils and an exceedingl,y small quantity of pre
_ cipitation (100200 mm annually). Under such conditions dust sto nns can
 arise not only when there are considerahle wind velocities (710 m/sec),
but also when there are moderate winds (46 m/sec).
In the Tsentraltnyye and Zaunguzskiye Karakum and in the southwestern part
 of the republic in summer dust storms are observed on the average during
the course of 68 days per month. Due to the great dryness of the soil and
_ the absence of a snow cover dust storms also occur in Turkmenia in winter
when their numher attains up to 46 days per month [5].
Dust storms are also observed in the mountainous regions of Central Asia
where their distribution, to a still greater degree than in the lowland _
territories, has a nonuniform, local character. In Kirgizia the greatest _
' mean annual numbers of days with dust storms (1416 days) are noted in the
western part of the IssykKul'skaya Basin and.in the lowland part of the r
_ Chuyskaya Valley; in the Kochkorskaya Valley and in the neighborhood of i
Novorossiysk 1213 days; in DarautKurgan 10 days. At the remaining
_ stations from 4 to 8 days per year. In Tadzhikistan the greatest fre
quency of recurrence of dust stbrms, from 20 days or more, is noted in
the southern part of the republic, and in the Eastern and Southern Pamir
on the average 814 days.
Thus, the map shows that the number of days with dust storms increases from
the zone of the wooded steppes and steppes to the ueserts of Central Asia.
The focus of the maximiun frequency of occurrence of dust storms for the
USSR is situated in Turkmenia, where they are observed on the average not
_ less than 4050 days per year, with a maximum in individual places up to 80
or more days. The general pattern of frequency of recurrence of dust stoxms
conforms to the geographic zonality. However, against this background there
is a clearly expressed focal nature of the distribution of dust storms,
~ which is caused both by the local natural peculiarities of the territories
meteorological conditions, character of relief, soilvegetation cover,
mechanical composition of the ground and also man's economic activity.
For example, plowing in regions with a relatively regular appearance of
dust storms leads to the formation of their new centers with a high fre
 quency of recurrence.
In addition to the mean annual number of days with dust storms, Table 1 gives
the annual variation of the frequency of recurrence of dust storms for the
regions most sub.ject to these processes. In the regions of occurre.nce of
sandy deserts dust storms can tie observed throughout the year with a small
 variability in their frequency of recurr.ence from month to month and with
~ a maximum falling in the hottest summer months (Repetek).
115
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
r{
41
rI
.n
cd
H
is
 o
41
rn
JW
_ N
~
A
.C
JJ
rl
3
~
~
co
A
44
O
N
~
~
z
44
O
~
\
v
N
U
t'+
N
N
~
U
aJ
~
44
0
~
U
N
~
 a
 a,
w
60, , 7
~ x U~ x
s ~
OTO_ ~ M
cD
a0 CO O? N t
N 00
o
O1
~
O
~
(D
~ p.W
~ N
N
~ M N N N
N ~
t
0
tD
U �
~t
cn tt
o
o 0 0 o
0 0 0 0
O
0
o 0 0 0 0
0 00 0 0
0 o
0 o
p
p
$ o
~ q
g
~
a
C,
~
I
>G
O c'6 I I
~
I I I O i
N I
O
O
M
4
CV t_
O~
tl er o0
a0 M
't
00 t"
q
00
X
N
CV
O I I ^ O
O GN
O
O CV
M
x
tl N
~
6
er
i
aC cD 00
cD N Ci c0
c0 tq
h ~
pp
~
d'
Qi t
oq
O~
M
.p .
cp
u
v
v
�
~
ip
O
~ M 00 l`
T N
Of
GV M
N
T
a; t,: ,6
~
L6
~ I w ^ ^
L6 t.6
m
m
,
O M
O
c 00 t~ cD l
Q~ O
N
~
'v t
t
cl
CD to n io ^
M h
0
^ W
w

O
. :
cj
V
a
O  M
O
cD Nr O
O L'7
h
cl
"7
h
~
h t~ O~
~
:7~ c'~ a ~
N
c0
c0 t~
O
M
N
t~ O d' t~
tD
N 0 M. ~ M
00 (7)
G~l
00 a0
CV
tD
~
00 N N cD
~ r+
M
00 O o0 c0 00
r y
~ M
0
r.
00 c0
~
O
ll
00
d'
~
t~ ~ JO
:J
tD to N ~ C~V
=
m
O
r+ ~ C N
O
d' r~  O oO
~ cD
d'
C1 61
00
d'
op CV 00
to,
h Ie)" c0 cD
0 00
V
V ~
00
O
d M O O
M
I II O j
~ I
I
I N
Go

c'!
r:
~
O o M
N
m
0
I
I
I I! I I
�
I
I
I
Ct
;
t i I

o

a "
~
m
c
x
z
x
u
o " x
a
o. yd c. m
'a
0
O' t~.
d
o o
t0
c~
F ~ ~
rn ~s
z
�
~
A
9
~
~ c
~
o~
m
a
e
.
y<
~
a
,
�
o=
o`0r419 u
`
�
�
o
i,Ysac~
G
u~
.a
,
n.~`''C a.
c, a o u.~
~
u
o
a$~ c'
m~ otn ~
O
cs 0 >,0.
~
�
~a a p
U:t n c
~
~Y n
v
LnGM
_ O I
~ Q
a m
cJ
U
.
F�
v
~ ~
.o
~ O 0�1 ~ F
F
st
F
o x
W
_ c0 ^ Q'i
J 0 p Y I Cj V c; O
~ O
a~
m O...
F.. U
a
l� U
~
w
,
`
`
~
~
U~c.mXo
x~C cC0 em o
~
cax
to~A0Oo)0
t~ 0`..4 xC~E4~C. x~ v
Y
C( c,
YR~
L~ uW o
~x
u
~x
0. u
00 O~ 0 ri clrn ri rl ri ~i ~I rI N N
116
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
KEY TO TABLE 1
1. Meteorological station
2. Month
3. Year, %
4. Mean annual numher of days with dust storms
5. Nizhniye Serogozy (Khersons?caya Oblast)
6. Pavlograd (Dnepropetrovskaya Oblast)
7. Verkhnedneprovsk (Dnepropetrovskaya Oblast)
8. Zavetnoye (Rostovskaya Oblast)
9. Gigant (Rostovskaya Oblast)
10. Kapustin Yar (Astrakhanskaya Oblast)
11. PokrovoIr.tyshskoye (Omskaya Oblast) 19361950
12. Same, 19511962
13. Aleyskaya Railroad Station (Altayskaya Kray)
14. Rubtsovsk (Altayskiy Kray)
15. Budenovskaya (Khakasskaya Autonomous Oblast)
16. Kyzyl (Tuvinskaya ASSR)
17. Dzhambeyty (Ural'skaya Oblast)
18. Dzhaltyr (Tselinogradskaya Oblast)
19. Bakanas (AlmaAtinskaya Oblast)
20. NebitDag (Turlmenskaya SSR) 21. Repetek (Turkmenskaya SSR)
In the more northerly, steppe regions dust storms arise for the most part in
the warm season of the year. Frequently there are two maxima of their frequency
of occurrence: in spring, when the fields are plowed, the vegetation cover
is absent and strong winds blow, and at the end of summerearly autumn when
the soils are dessicated and the fields are again plowed.
The plowing of some virgin lands in our country in the 1950's resulted not only
in an increase in the number of dust stornis, but also a change in the annual
variation of their frequency of recurrence. Thus, in the example of the data
for PokrovoIrtyshskoye station it can be seen clearly that after plowinp
there was a marked increase in the mean annual number of days with dust storms
(Table 1). The dust storm maximum in this case does not fall in any one month
(May), as was the case before plowing, but there is a smoother variation of
the frequency of occurrence of dust storms in the course of the warm season
of the year with a decrease in the spring maximum, an increase in the number
of dust storms in the remaining months and the appearance of a second, autumn
maximum approaching that of spring.
B,IBLIOGRAPHY
1. Buchinskiy, I. Ye., ZASUKHI I SUKHOVEI (Droughts and Searing Winds), Len
ingrad, Gidrometeoizdat, 1976.
117
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
2, Zakharov, P. S., PYL'NYYE BURI (Dust Storms), Leningrad, Gidrometeoizdat,
1965.
3. PYL'NYYE BURI I IKH PREDOTVRASHCHENIYE (Dust Storms and Their Prevention),
Moscow, Izdvo AN SSSR, 1963.
4. Romanov, N. N., "Dust Storms in Ceatral Asia," TRUDY SAGU (TrN s174iOFIZ
f
Central Asian State University), NOVAYA SERIYA (New Series), ,
ICHESKIYE NAUKI (Physical Sciences), Kn 20, 1960.
5. SPRAVOCHNIK rO KLIMATU SSSR (Handbook of USSR Climate), No 10, Part 5,
1969; No 13, 17, 18, Part 3, 1967; No 19, 20, Part 3, 1966; No 21, Part
3, 1967; No 30, 31, Part 3, 1968; No 32, Part 3, 1967.
6. Chirkov, Yu. I., "Frequency of Recurrence of Dust Stornas in the Territory
of the USSR and the Possibility of Predicting Their Formation," TRUDY
GIDROMETEOROL. NAUCHNOISSLED. TSENTRA SSSR (Transactions of the USSR Hydro
meteorological Scientific Research Center), No 69, 1970.
118
FOR OFFICIAL USE QNLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFlcIAL USE ONLY
UDC 551. 50 : 6 33.11 "
MODELING OF THE PROCESS OF FORMING OF TIiE YIELD OF WINTER WHEAT
Moscow METEOROLOGIYA I GIDROLOGIYA in Russian No 9, 1979 pp 98106
[Article by Candidates of Geographical Sciences M. S. Kulilc, A. N. Polevoy
and I. Ye. Vol'vach, AllUnion Scientific Research Institute of Agricultural
= Metf:orology, submitted for publication 3 April 1979]
Abstract: This article gives a dynamic model
of formation of the yield of winter wheat de
scribing the principal processes of vital ac
 tivity of a plant (respiration, photosynthesis,
growth of individual organs) and the influence
of environmental factors on their intensity.
[Text] Yield formation is the complex totality of a whole series of physiolog
ical processes whose intensity is determined by the biological peculiarities
of the plants, environmental factors and the interrelationship among the pro
cesses themselves.
 However, the principal role in yield formation is played by photosynthesis.
In grain ear crops the photosynthesis process can take place not only in
the leaves, but also in the other aboveground organs leaf sheaths, stems
 and ears. The data given in [15] show that different plant organs assimilate
with different intensities.
The ears have a relatively high intensity of photosynthesis. At the ons eC
_ of the milky ripeness phase the intensity of photosynthesis of the awns,
spikelet and flowering glumes of the ear is as great as for the leaves of
= the two upper levels and is considerably greater than for the leaves of the
third 1eve1 from above [11].
 The photosynthesis of each of the photosynthesizing organs (leaves
stems s, ears p) can be represented by the formula [19, 22]
119
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02148: CIARDP8200850R0402400300394
FOR OFFICIAL USE ONLY
Omex 16 11
[(b IJ)' f 02 ~~1 '
_ mexi
where !ki is the intensity of photosynthesis of the ith organ under opti
mum conditions of heat and moisture supply and real illumination conditions
(mg C02/(dm2�hour)); Omax i is tlie intensity of photosynthesis of the ith
organ in the case of light saturation and a normal C02 concentration (mg C02/
 (dm2�hour)) ; b is the initial slope of the light curve of photosynthesis
(mg C02�dm 2hour'2/(cal�cm 2�miri 1)); I is the intensity of photosynthet
ically active radiation within the sown area (cal/(cm2�min)). When deter
mining the intensity of photosynthesis of ears I is assumed equal to the
intensity of photosynthetically active radiation at the upper boundary of
the sown area; j is the number of days in the computation period.
_ In ontogenesis the photosynthetic activity of the leaves and other photosyn
thesizing organs is determined by their age and change in internal structure.
In the sprouting phase in spring crops and at the time of renewal of the
 growing season of winter crops only the forming leaves are not characteriz
ed by a high intensity of photosynthesis. Such a structure of the leaf
(especially middlelevel leaves) is formed by the "earingflowering" phase
which is optimum for their photosynthetic activity.
For computing the photosynthesis in ontogenesis under real environmental
conditions different from the biological optimum, we write the expression
IDi  (Do a~ ~'p~ 7~
(2)
1115= ph] where Ot is the intensity of photosynthesis under real environ
mental conditions (mg C02/(dm2�ho ur)), olPh is the ontogenetic photosyn
thesis curve, VPh and Yph are functions of the effect of environmental
factors (air temperature and soil moisture content), representing sing].e
peak curves approximated by the expressions
TA Txn 16( 1(3)
and ~'2 TA onT  Tnn Ta onr ~~nn I J~
_ w~ Wf
A' exp l f~g ~/xne )A3 exp (A4 Wxns (4)
~
[A = d; Tf = thr; O?TT = op(timum) ; Hri'B = mfmc = minimimm field moisture
capacity] where Td is the mean daily air temperature, Tthr and Td opt are
the threshold and optimum air temperatures for photosynthesis, W ara the
reserves of productive moisture in the halfmeter sail layer (mn), Wmfmc
is the minimum field moisture capacity of the halfmeter soil layer (mm).
The approximation of the temperature curve for photosynthesis in the form
of expression (3) is given in [14].
The curve characterizing the influence of soil moisture content on photo
synthesis was constructed, adhering to [16], separately for sandy loam and
clayey loam soils.
120
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
The functions � Ph, V ph and YPh, entering into expression (2), were nor
malized and vary from 0 to 1.
The photosynthesis of each organ during the light time cf day can be comput 
ed using the formula
_ (Di = E(D! Li Tl ,
r (5)
where Spi is the daily photosynthesis of the ith organ per unit area (g/(m2�
day)), E is the coefficient of efficiency of photosynthesis, Li is the area
of the assimilating surface of the ith organ (mZ/mz), Gd is length of day. '
The photosynthesis of the sown area during the 24 hours is 
l, s. p
(D;. (6)
I
In contrast to the photosynthesis process all plant organs have the capacity
for respiratory gas exchange.
= r
Expenditures on respiration are subdivided into respiration associated with =
maintenance of the structural organization of the tissues and respiration
associated with movement of substances, photosynthesis and creation of new
structural units [12, 18, 20].
The increment of biomass of the ith plant organ during the day is determin
ed by the difference between the receipt of "fresh" products of photosynthe
sis in the organ, respiration and redistribution of "old" assimilates.
Taking into account the equations of growth proposed in [8], their modific
ation as in [4] and the
inclusion of the diurnal
gas exchange
of the sown
area [I2] in them, we wi11 write
the biomass balance equation
of the ith 
 organ in the form ~
~ m
i
l
'
pi (DJ _ 1
at c' ~l + vi) ml; 
R ~ i
( R
. A i
1 aRt
cj ` 1t aR!
l
Cq
p
=
ap ~I  ~
aR Cl
i, s, r
yi Mi
A t
~
1 ak
C~ 1i aR
Cs p
i
P
P
l
.
where mi is the total dry biomass of individual i Ei, s, r(r are the �
roots) of the organs (g/m2), mi is the living biomass of the ith organ,
?Bi is the redistribution function for "fresh" assimilates, vi is the re
distribution function for "old" assimilates, � Ri is the ontogenetic respir
ation curve, ci is a coefficient characterizing the expenditures on main
taining the structure, c2 is a coefficient characterizing the expenditures
associated with movement of substances, photosynthesis and creation of new
structural units, 95R is the temperature coefficient of respiration, which, `
in accordance with 118], is determined using the formula
121
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
A  To)
~PR=2 0. 1 (T ,
(8)
where Tp is the cardinal temperature of respiration.
The increment in the total hiomass of the sown area is e Mw' t' s' r' p Am j+1 
. ~ .
where M is the total biomass of the sown area (g/m~').
With LS mi > 0 tlie increment of ].iving biomass of tlle ith organ then
 will be equal to the increment� of its total biomass.
9 mi L c1 mi (10)
 pt  At ' 
If it is assumed that the decrease in the living bioma.ss of the vegetative
organs can proceed only to an entirely definite critical value, in such
a case when A mi/ 0 t< 0 the decrease in the total biomass of the v~ege
~ tative organ by some value results in the dying out of a considerably
greater part of the living biomass, which will be determined by the ex 
_ pression '
ami a mi 1 (11)
where Kc is a parameter characterizing the critical value of decrease in
the living biomass of the ith organ with which its dyingout begins.
 The equations for describing the growth of living and general biomass of the
 ith plant organ will have the form �
~ i+l ~ ml + e mJt
m
p (12)
mj+l _ mj + ei{
~r
t E p
In equation (7) the expression a mi/At is the increment of total dry
biomass of the entire reproductivePorgan (stem of ear, spikelets and
grains).
The dynamics of accumulation of dry matter of grain can be determined us
ing a formula in the form
122
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
1 1
m1+1= mi + 1 mR R,ex 'nn
S 9 1 A mI ~mi ~1 ~t
fC ~tiax , j (13)
m
where mg is the dry hiomass of the grain (g/m2), Km is the initial slope
of the curve, characterizing the rate of growth of the grain in dependence
on the level of inflow of the assimilates into tlie ear, mg max is the max 
_ imum possible increment of grain weight (g/m2) with the real relationship 
of the elements determining the productive plant stand per unit area aiid
the size of the ear, which can be obtained using the formula
A ntK max  2.3 A,, mR maz � 10~1,. }1
_ At (I F IOAu1r p (IZI)
where mg max is the maximum possible weight of the grain (g/m2).
For its estimate we use the expression
mg max = NS Cp rip n g mgr�l (15)
where NS is the density of plant stands in different development phases;
Sp is the productive bushiness, n is the number of spikelets in an ear,
ng is the number of grains in a spikelet, mgv is the maximum possible
weight of 1,000 grains characterizing this variety.
Computations of the surface area of the assimilating organs in the case of
a positive balance of their biomass can be accomplished using the formulas =
proposed in [4, 12]
L~+~ _ L~ + ~,,l 1
1 i , t n1~ '
1
when o t = 0, 1 E 1, s, p, (16)
where mt is the specific surface density of the ith organ (g/m2).
With a negative incremant in the biomass of leaves and stems for description
of the increase in their assimilating surface we use an expression in the
" form
5 ai 1
LT+'=L~ et m '
tt
123
FOR QFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
 FOR OFFICIAL USE ONLY
Am` ~t
e Mi+1 oMI
if ~t < ot '
(18)
then the increment of the aboveground biomass is determined from the ex
pression
12.4
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
I
FOR OFFICIAL USE ONLY
A m1 A~>>+1 
mr+l
ar ~ if st < ot
o t 0, if A 'n~+l' A MJ~'t
(19)
at r at '
where mabove is the biomass of the aboveground orgar,s (g/m2).
The growth functions of the aboveground plant organs are determined similar
ly [5].
The dynamics of increments of total biomass of the sown area is examined
here as an internal "switch" for the adahktive increase in the fraction
of underground organs when the plant is exposed to unfavorable conditions.
In order to compute photosynthesis (1) it is necessary to have the intensit
ies of photosynthetically active radiation (PAR) in the sown crop. This val
ue is computed using the formula fl (20)
_ 1} cLj'
where IQ is the intensity of PAR at the upper boundary of the sown area
(cal/(cm2�min)), c is an empirical constant.
For computing the flux of PAR at rhe upper boundary of the sown area we use
the formula
16 = 4,5 q~ ,
�tx
(21)
[II H= day] where Q is total radiation (cal/(cm2�day)), 'Cday is the length 
of day.
The total radiation is computed using the Sivkov formula [10]
[Tr = mid(day) ] Qj = 12,66 ISSj)1'31 I 315 (sin h")2'1 , (22)
where SS is the duration of sunshine during the day (hours), where hmid
is midday solar altitude, which is determined using the formula
 si[t hn  A + B, (23)
where A= sin ~ sin B= cos 91 cos S, 5P is geographic latitude, S is solar
decl.ination.
The table of solar declinations for the spring and autumn months was approx
imated by the polynomial [1]:
[0,4i 3 (to ~1I) 0,196 � 102 (to IJ)'  (24) 
 0,407 � 105 (to j)'  0,6161 � 0,017453,
125
FOK OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
where to is the number of the day from the beginning of the computation per '
iod, reckoned from 20 March, j is the number of the day of the computation 
period (j = 0, 1,...,N). _
In the case of absence of observational data on sunshine, total radiation
_ can be computed using information on lower and total cloud cover using the
formula 12)
Q = (lo P  C� Cco (n  ia�)1. (25)
[H = low(er); CB = m1 = middlelower] where Qp is the total radiation in
the clear sky (cal/(cm2day)), Clow is the coefficient for lowerlevel
 clouds, C,1_1 is the coefficient for clouds of the middle and upper levels,
nloW, n is the quantity of clouds in fractions of unity of the lower level
 and total cloud cover.
The length of day is determined using the formula
'C! i1t~
~ 9 tlp
[L( = d; 3= ss = sunset; B= sr = sunrise]
(26)
 where the time of sunrise ('~Sr) and sunset ('CSS) is determined using the
_ formulas
~ 23.= 12 4 12 arc cos ~(27)
~
~'8=~4  .3. .
Formula (4) includes the reserves of productive soil moisture in the half
meter soil layer; change in moisture supplies during the course of the 10
day period (between the times of observations) can be determined using the
expression
W1+1 = W> 61 + Vi  EJ  CJ,
(28)
where e is the precipitation sum in 24 hours (mm), V is the movement of
ground water into the aeration zone per day (mm), E is total evaporation
per day (mm), C is the infiltration of atmuspheric precipitation (mm).
The flow of ground water into the aeration zone is computed using the for
mula
E~
. vi (29)
where Eo is evaporability per day (mm), computed using the mean daily dew
point spread, f is a parameter dependent on the hydrophysical properties
of the soil, H is the depth of the ground water (m).
126
FOR OFFICIAL USE ONLY
~
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
The diurnal evaporation values can be obtained using the formula proposed
 in [13]
Ef = 2 WI + QJ + VJ
 1 I 2 Wilno ' (30)
[HTiB = mfmc = minimum field moisture capacity] is the minimum field mois
ture capacity in the halfmeter soil layer (mm), Yt is a parameter dependent
on the type and phase of development of the plants.
For computing the infiltration of atmospheric precipitation we use the ex
pression
CJ  W! AJ  Er  WHOB� (31)
[HTT B = mfmc]
The ground water level between observation times entering into formula (29)
is found using the formula
H1+' =Hj+oHJ, (32)
and the change in the ground water level during a day (m) using the
formulas Cl I
~ H1 100 when C> 0;
(33)
A Hi  o When C< 0,
where ~x is the coefficient of ground water yield.
_ In order to carry out computations using the proposed model of formation
of the yield of winter wheat during the springsummer period it is necessary
to have the following agrometeorological information:
maximum air temperature;
mean daily air temperature;
mean daily dewpoint spread;
number of hours of sunshine daily (or information on cloud cover);
sum of precipitation per 24 hours;
reserves of productive moisture in the halfmeter soil layer;
ground water level.
The dynamics of biomass of individual organs of winter wheat and the grain
yield is computed from the moment of renewal of the growing season (j = 0).
The described method also makes it possible to evaluate the influence of en
viranmental factors on the increment of the vegetation mass for different
combinations of different parameters characterizing the agrometeorological
conditions.
127
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
B
_ Fig. 1. Light curves of productivity of winter wheat against a background
of different soil moistening. 1) W/Wmfmc = 0.8; 2) W/Wmfmc = 0�2�
KEY:
A) g/m2
B) cal/ (cm2� day)
Some results of such numerical experiments are given in Fig. L. It can be
seen that the productivity light curve obtained against a background of
different moistening has some peculiarities. With an increase in the den
sity of the light flux in the entire considered range the plant produc
tivity increases, but the biomass increment increases rapidly and attains
considerable values only against a background of optimum moistening (80%
of the maximum soil moisture content)..In the case of.an inadequacy of
moisture (20% of the maximum soil moisture content) the productivity max
imum is almost four times less than under good conditions. We also note
that deterioration of moisture suppl.y conditions leads to some displace
ment of the compensation point in the direction of high light flux density
values.
One of the possible ways to apply a model of the production process for
winter wheat in actual practice can be the development, on its basis, of
a method for a quantitative evaluation of the conditions for growth of
the crop during the springsummer period.
The numerous reviews published by operational agencies of the State Com
mittee on Hydrometeorology are based on a detailed evaluation of the ex
isting and anticipated weather conditions. The latter is carried out using
relatively simple and accessible criteria which make it possible to eval
uate the degree of favorability of existing agrometeorological conditions
for the growth, development and formation of the~yield of a crop in compar
ison with the optimum mean longterm conditions or the conditions for auy
 year taken as a standard [9].
An evaluation of the conditions is carried out for 10day periods, months ~
and periods during the growing cycle; the values of the meteorological el ~
ements used in the evaluation are averaged for the corresponding time in '
 terval. Naturally, in this case the evaluations themselves are smoothed. In '
[3] K. S. Veselovskiy already wrote that "in the mean values the extremes
128
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
disappear, hut the higher and lower temperatures characterize climate
better in that these exert a special infuence on organic life."
The computations of such evaluations using a model make it possible to
make them more complete and thorough; in evaluating each successive per
iod it is possilile to take the prehistory into account, that is, the con
sequence of the preceding period, and also take into account the extremal
values of the environmental elements during the course of any evaluated
period.
BIBLIOGRAPHY
1. Abashina, Ye. V., Prosvirkina, A. G., Sirotenko, 0. D., "Simplified Dy
np.':aic Model of Formation of the Yield of Spring Barley," TRUDY IEM
(Transactions of the Institute of Experimental Pieteorology), No 8(67),
1977.
2. Berlyand, T. G., RASPREDELENIYE SOLNECHNOY RADIATSII NA KONTINENTAKH
(Distribution of Solar Radiation on the Continents), Leningrad, Gidro
meteoizdat, 1961.
3. Veselovskiy, K. S., 0 KLIMATE ROSSII (Climate of Russia), St. Peter
b urg, Izdvo AN, 1857.
4. Galyamin, Ye. P., "Construction of a Dynamic Model of Formation of
Yields of Agrocoenoses," BIOLOGICHESKIYE SISTEMY V ZEMLEDELII I LESO
VODSTVE (Biological Systems in Agriculture and Forestry), Moscow,
Nauka, 1974.
5. Galyamin, Ye. P., Siptits, S. 0., Milyutin, N. N., "Mode1 of Formation
of the Yield of an Agrobiocoenosis and its Identification," MODELIRO
VANIYE PRODUKTSIONNYKH PROTSESSOV V AGROEKOSISTEMAICH (Modeling of Pro
duction Processes in Agroecosystems), Moscow, Nauka, 1976.
6. Kupe rnan, I. A., Khitrovo, Ye. V., "Dynamics of Some Components of Mass
Exchange of Wheat During a Progressive Soil Drought," FIZIOLOGIYA PRI
SPOSOBLENIYA RASTENIY K POCHVENNYM USLOVIYAM (Physiology of Adaptation
of Plants to Soil Conditions), Novosibirsk, Nauka, 1973.
7. Moldau, "Influence of Water Deficit on Plant Increment," IZV. AN
ESSR (News of the Estonian Academy of Sciences), Biologiya (Biology), 
Vol 23, No 4, 1974. 
8. Ross, Yu., "System of Equations for the Quantitative Growth of Plants
FITOAKTINOMETRICHESKIYE ISSLEDOVANIYA RASTITEL'NOGO POKROVA (Phytoac
tinometric Investigations of the Plant Cover), Tallin, Valgus, 1971.
9. RUKOVODSTVO PO SOSTAVLENIYU AGROMETEOROLOGICHESKIKH PROGNOZOV (Manual
on Compilation of Agrometeorological Forecasts), Leningrad, Gidro
meteoizdat, 1962.
129
_ FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
10. Sivkov, S. I., METODY RASCHETA KHARAKTERISTIK SOLNECHNOY RADIATSII
(Methods for Computing the Characteristics of Solar Radiation), Len
ingrad, Gidrometeoizdat, 1968.
11. Tarchevs.kiy, I. A., FOTOSINTEZ PSHENITSY. FIZIOLOGIYA SEL'SKOKHOZYAY
 STVENNYKH RASTENIY (Photosynthesis of Wheat. PhysioYogy of Agricul
tural Plants), Moscow, Tzdvo MGU, Vol IV, 1969.
12. Tooming, Kh. G., SOLNECHNAYA RADIATSIYA I FORMIROVANIYE UROZHAYA (Sol
ar Radiation and Yield Formation), Leningrad, Gidrometeoizdat, 1977.
13. Kharchenko, S. I., GIDROLOGIYA OROSHAYEMYKH ZEMEL' (Hydrology of Irri
gated Lands), Leningrad, Gidrometeoizdat, 1975.
14. Chmora, S. N., Oya, V. M., "Study of the Temperature Dependence of
Leaf Photosynthesis," FIZIOLOGIYA RASTENIY (Plant Physiology), 14,
No 4, 1967.
15. Shatilov, I. S., Vaulin, A. V., "Dynamics of an Assimilating Surface
and the Role of Individual Plant Organs in Formation of Barley Yield,"
IZVESTIYA TSKhA (News of the Academy of Agricultural Sciences), No 1,
1972.
16. Baker, D. N., Musgrave, R. B., "The Effect of LowLevel Moisture
_ Stresses on the Rate of Apparent Photosynthesis in Corn," CROP SCI
ENCE, Vol 4, No 3, 1964.
17. Brouwer, R., Morphological and Physiological Adaptations to External
Conditions," ACTA BOT. NEDERL., Vol 17, No 1, 1968.,
18. Curry, R. B., "Dynamic Simulation of Plant Growth. Development of a
Model," TRANS. ASAE, Vol 14, No 5, 1971.
 19. Horie, T. "Simulation of Sunflower Growth. I. Formulation and Para
meterization of Dry Matter Production, Leaf Photosynthesls, Respira
tion and Partitioning of Photosynthesis," BULL. NAT. INST. AGRIC. SCI.,
 JAPAN, SER. A24, 1977.
20. McCree, K. J., An Equation for the Rate of Respiration of White Clover
Plants Grown Under Controlled Conditions," PREDICTION AND MEASUREMENT
OF PHOTOSYNTHETIC PRODUCTIVITY, Wageningen, Pudoc, 1970.
21. Struik, G. J., Bray, J. R., "RootShoot Ratios of Native Forest Herbs
and Zea mays at Different SoilMoisture Levels,�t ECOLOGY, Vol 51, No
5, 1970.
22. Thornley, J. H. M., MATHEMATICAL MODELS IN PLANT PHYSIOLOGY, Acad.
Press, London, New York, 1976.
130
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
UDC 551.(524+55)
TIME VALIDITY OF METEOROLOGICAL INFORMATION
Moscow METEOROLOGIYA I GIDROLOGIYA in Russian No 9; Sep 79 pp 107109
[Article by G. P. Lutsenko and Candidate of Technical Sciences V. D. Nikol
ayev, submitted for publication 24 November 1978]
Abstract: The authors carried out investigations
and an analysis of the temporal variability of
the wind and temperature fields under conditions
of different thermodynamic and circulation states
of the atmosphere. The article gives quantitative
estimates of the time validity of data on wind and
air temperature in cyclones and anticyclones, for
different wind velocities, in stable and unstable
air mas.ses.
[Text] By the term "time validity" of ineteorological information is meant
the time interval during which this information can be used in solving
different classes of problems. In a general case, the time validity of
meteorological data is determined by the following principal factors:
the scale of atmospheric processes elucidated in time;
the required degree of informativeness (detail);
the nature of the specific problems providing for the usz of ineteorolog,
ical data.
The time validity can differ substantially in dependence on each of these .
factors. In actuality, in solving the problem of the time interval during
which, for example, a definite type of atmospheric circulation will persist
in some region and the required informativeness is reduced only to a quan
titative evaluation (zonal circulation, arctic intrusion, etc.), it is only
possible to deal with 24hour intervals [2] (Table 1).
From the point of view of implementation.of specific tasks, the time valid
ities of ineteorological information are usually more rigorous. In i:he systeM
of the USSR Hydrometeorological Service such time periods are intradiurnal
intervals and in particular, 6hour intervals.
These times have been estahlished on the liasis of an analysis of definite re
lationships between the mean square error in meteorological data, which leads
to errors in prediction above the norm o"d,and the temporal variability of
meteorological elements.
131
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
Fundamentally the analytical expression for the validity rimes for meteor
 ological information tk in a general case has the form [6]
r Q '
1
lk6la(6K) '
~
(1)
[ .A = d] where o'(6 hours) is the mean square variability of the meteorolog
ical element in the course of 6 hours.
Investigations of the statistical variability of ineteorological elements
in the intradiurnal interval indicated that the mean square differences in
the values of the meteorological elements increase in conformity to the
"firstorder law" [6]
Q(t)=J~lYt.
where (71 = 0.4 0' (6 hours).
(2)
However, it must be noted that this law was established (theoretically and
experimentally) for the entire diversity of synoptic processes without al
lowance for specific weather conditions.
At the same time, already from purely physical considerations it can be pos
tulated that U(t) for different thermodynamic and circulation states of
the atmosphere are substantially different, which, in turn, leac'.s to a dif
ference in the validity times for meteorological information under these
conditions from those existing at the present time.
In actuality, the physical nature of the variability of atmospheric char
acteristics is such that in the process of turbulent mixing of air particles
in the atmosphere there is transport of the main substances of these par
ticles. The intensity of this transport (in the last analysis, the intens
ity of variability) is usually characterized, as is well known, by the Rich
ardson number Re, which is a function of the velocity v of movement of air
particles v:
_ Re  vy (3)
where L is the characteristic scale of the flow; V is the kinematic viscos
ity of air.
Naturally, with small v values Re itsElf is also insignificant, that is,
the flow is laminar in nature, which also predetermines the relatively low
variability of the meteorological elements.
With respect to the relationship between the variability of atmospheric char
acteristics and the nature of the pressure field, here it is necessary to
take into account the limitation of the pressure gradient from above in
anticyclonic formations [4].
132
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
Table 1
Duration of Types of Atmospheric Circulation for European USSR
IlpoAOnx:xrenb
HOCTb, K ' 2
TIIII 1(NpKYJIRI(1(1( C:;
3 x 4 I 5
~
1 ~ = s
~
~cl
6 :3otianbxag 43 ~ ?09 21a
] I1E1xnottltvecxas ,qeR ;
Tenbftocrb {S I 105 102
� 8 API(THqCCKOC DTOP}f(C
xue 48 101 310
KEY:
1. Type of circulation 6. Zonal
2. Duration, hours 7. Cyclonic activity
3. Minimum 8. Arctic intrusion
4. Mean
5. Maximum
1 Ar1
10
, f
0
4
8 6.. f
2 j 6r
' : F
Fig. 1. Mean square variability of wind velocity and air temperature as a
function of thermodynamic and circulation states of atmosphere. 1) v= 03
m/sec; 2) v= 38 m/sec; 3) v= 816 m/sec; 4) v= 1624 m/sec; 5) stable
air mass; 6) unstable air mass; 7) cyclone with wind velocit:Les 03 m/sec;
8) cyclone with wind velocities 1624 m/sec; 9) anticyclone with wind velo
cities 03 m/sec; 10) anticyclune with wind velocities 1624 m/sec; 11) cy
clone; 12) anticyclone.
133
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02148: CIARDP8200850R0402400300394
FOR OFFICIAL USE ONLY
The value of the Re parameter, all other conditions being equ.sl, in cy
 clones on the average is greater than in anticyclones because the varia
bility of ineteoralogical elements in cyclones is greater than in anti
cyclones.
In this study we investigated the variability of the wind characteristics
_ (direction ocv and velocity v) and air temperature T for specific states
of the atmosphere and conclusions were drawn concerning the v alidity times
for data on wind and temperature under these conditions. It was taken into
account that the time differences of wind and tempersture adhere to a nor
mal distribution law [5].
As the thermodynamic and circulation characteristics of the state of the
= atmosphere, ir. the investigation we used: the deviations of atmospheric
pressure f rom the norm (as the no rm within the limits of the European USSR
we used p= 1005 mb); wind velocity (by gradations); degree of thermal
stability of air masses (y< 0.7�C/100 m stable air mass SAM; T50.7�
C/100 m unstable air mass UAM, pressure trend, nature of pressure
field (cycZone  anticyclone), and also the wind regime of pressure systems.
 In addition, the seasonal states of the atmosphere were taken into account.
Computations of the characteristics of variability were made for several
atmospheric sounding stations within the limits of the E,:ropean USSR. The
data are given to the altitudes of the middle troposphere.
The principal results of the investigations can be summarized as follows
(Fig. 1).
The variability of wind velocity and air temperature to a considerable de
gree is dependent on the atmospheric wind regime. For example, on the av
erage, within the limits of the troposphere when there is a weak wind (not
more than 68 m/sec) the variability of wind velocity and temperature is
22.5 times less than when there are considerable wind velocities (more
than 1520 m/sec).
~ The dependence of the variability of wind velocity on the thermal state of
air masses is somewhat smoothed, although here aa.so in stabl,e air riiasses
the vali.ability of wind velocity is 1.31.5 times greater than in unstable
air masses.
The dependence of air temperature variability both on wind velocity and
= on the nature of the thermal stability of air masses is smoothed. However,
it is easy to note that in unstable air masses and when there is a strong
wind ttie variability of temperature is 1.31.5 times greater than in stable
air masses and when there is a weak wind.
The dependence of the varialiility of wind velocity and air temperature on
the nature of the pressure field is considerable; this dependence is mani
 fested most clearly under conditions of a different wind regime of the
134 .
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
pressure systems. For example, the variahility of wind veloctty in the
 case nf its high values in cyclonic eddies is three times greater than
in an anticyclone with a weak wtnd. The variability of aix temperature
in a cyclone is also almost twice as great as in an anticyclone.
_ With respect to the variability of wind direction, with one and the same
_ wind velocities the variability of wind in the cyclone is 1.52 times _
_ greater than in an anticyclone. With different wind velocities the vari
ability of its direction conforms to the general patterns obtained by
, different authors [1, 3, 6]: it decreases with an increase in wind velo
city and vice versa (Table 2).
Table 2
Mean Square Variability of Wind Direction (Degrees) in Six Hours 5or
Different Velocities*
 ' ~ v M/cerc
zKX
03 I 38 i 816 ~!624
I
0 51 3G 28 
, 1,5 83 48 26 
7 82 411 32 24
9 97 til 40 3U
Z* 3H840fIF[S 6Q Z, paCCYHT61B3JI1iCb NO
'
repeAxHy xaWAoR rpa,zawr!.
KEY:
v m/sec
_ 2. * The 0a v values were computed in the middle of each gradation
It was impossible to establish any significant correlation between wind
_ variability and air temperature with the deviation of atmospheric pressure
 and the pressure tendency from the norm in the process of tYie investiga
tions.
Thus, our investigations demonstrated that the variability of the wind char
 acteristics and air temperature is essentially dependent on the specific
= state of the atmosphere. This, in turn, leads to the following: the "valid
ity times" for the mentionPd meteorological elements with different states
of the atmosphere are also different. Already on the basis of preliminary
evaluations it can he concluded that the "validity times" for data on the
wind and air temperature, and at the same time, the regime of wind and tem
perature sounding of the atmosphere in cyclones and anticyclones with high
and low wind velocities within the limits of the troposphere, dif f er fram
those now existing by a factor of 22.5.
Further investigations of the temporal structure of the fields of ineteoro
logical elements as a function of thermodynamic and circulation states of
the atmosphere make possible a more specific solution of problpms relating
135 .
FOR OFFICIAL USE ONLY 
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
_ to the "validity times" for meteorological information, atmospheric sound
ing regimes and the principles for control of operation of atmospheric
sounding stations under these conditions.
BIBLIOGRAPHY
1. Garifulin, K. K., IZMENCHIVOST' VETRA V SVOBODNOY ATMOSFERE (Wind Var
iability in the Free Atmosphere), Leningrad, Gidrometeoizdat, 1967.
2. Dzerdzeyevskiy, B. L., Kurganskaya, V. M., Vitvitskaya, Z. M., "Clas 
sification of Circulation Mechanisms in the Northern Hemisphere and
Characteristics of Synoptic Seasons," TRUDY NIU GUGMS (Transactions 
of Scientific Research Institutes of the Main Administration of the ,
Hydrometeorological Service), Ser 2, No 21, 1946.
3. Kovalenko, V. V., Zelenoy, I. K., "Wind and Temperature Variability
in the Atmosphere With Time and Distance," SBORNIK RABOT LENINGRADSKOY
 GIDROMETEOROLOGICHESKOY OBSERVATORII (Collection of Papers of the Len
ingrad Hydrometeorological Observatory), 1962.
4. Laykhtman, D. L., DINAMICHESKAYA METEOROLOGIYA (Dynamic Meteorology),
Leningrad, Gidrometeoizdat, 1976.
. 5. Lutsenko, G. P., Nikolayev, V. D., "Variahility of the Temperature and
Wind Fields in Different Forms of Atmospheric Circulation," METEOROLOG
IYA I GIDROLOGIYA (Meteorology and Hydrology), No 11, 1976.
6. Reshetov, V. D., IZMENICHIVOST' METEOROLOGICHESKIKH ELEMENTOV V ATMO
SFERE (Variability of Meteorological Elements in the Atmosphere), Len
 ingrad, Gidrometeoizdat, 1973.
136
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
UDC 551.(557.5:578.42:509.323)
 RELATIONSHIP BETWEEN THE PLANETARY HIGHALTITUDE FRONTAL ZONE AND THE POSITION =
OF TEIE SNOW COVER BOUNDARY DURING THE AUTUIrIlV AND SPRING PERIODS
Moscow METEOROLOGIYA I GIDROLOGIYA in Russian No 9, Sep 79, pp 110112
[Article by Candidate of Geographical Sciences V. B. Afanas'yeva, Candidate
of Physical and Mathematical Sciences N. P. Yesakova and R. V. Klimentova,
Main Geophysical Observatory, submitted for publication 29 September 19781
Abstract: A study was made of the relation
shig between the planetary highaltitude fron
tal zone (PHAFZ) and the position of the snow
cover boundary in spring and autumn. It was es
tablished that both in spring and autumn in the
territory of Eurasia the PHAFZ has a definite
relationship to the position of the snow cover
boundary. The correlation coefficients are given. 
The results obtained make it possible to use data
on tfie pos.ition of the PHAI'Z for determining the
 houndary of the snow cover, being an important
predictor in the prediction of inean 10day tem
 perature.
[Text] The snow cover, having singular radi.ation and thermal properties,
forms a surface differing sharply from the snowless underlying surface. 
This circumstance exerts a strong influence on climate and must be taken
into account when preparing longrange weather forecastso
In the Dynamic Meteorology Section of the Main Geophysical Observatory, in
 the development of the hydrodynamicstatistical method for predicting mean
10day temperature, in addition to the circulation factors presently used
 in the forecasts, an allowance was made for cloud cover, ice content of
northern seas and snow cover. At tlie same time it was established that the
_ snow cover is one of the principal predictors in temperature forecasting.
fIowever, the collection of data on the position of its boundary frequently
involves definite difficulties. The presently accumulated extensive observ
ational data made it possible to construct mean 10day maps of these ele
ments, in particular, of the`snow cover, on which were plotted isolines
137
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
F
FOR OFFICIAL USE ONLY
0
Q) 10 0
a
D
~
10
MOCKI
'0
Ko
OMOCKBA
~i 10 '
5
 c
Ktd~B MBCKBA
S`
'10
~
0
V61c,
�MNHCK
OMOCKBA
Y '^S I
40 �oD
e al
s 10MyPMAH ~
p
~
MDC~KBA p
f e~ S ~ � �oo
0
~
l 10
s
~i
APXAHfEl16CF~ ~
0
t
Fig, 1. Maps of distribution of the snow cover, 1956, second 10day period.
a) October, b) November, c) December, d) March, e) April, f) May
of 0, 5 and 10 days with preser_ce of a snow cover in a 10day period during
the autumn and spring periods (OctoberDecember, MarchMay). The method for
preparing these maps has been described in considerable detail in published
studies. An analysis of the maps indicated that the position of the snow
cover boundary varies in different years in a considezable range [1].
In order to take into account the position of ttte snow cover boundary in the
preparation of statistical forecasts it was necessary to establish its stat
istical correlations with different meteorological elements. In this study
an attempt is made to establish statistical correlations between the posi
tion of the snow cover and the planetary highaltitude frontal zone (PHAFZ),
and if a sufficiently close correlation is obtained, use the position of the
PHAFZ for determining the boundary of the snow cover, since in itself a de
termination of the position of the PHAFZ does not constitute significant
difficulties.
138
FOR OFFICIAL USE ONLY p,
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICYAL USE ONLY
In studies by V. A. Bugayev and V. A. Dzhordzhio [2], V. I. Vorob'yev [4],
and V. A. Bugayev 131 it is pointed out that the PHAFZ is most sharply ex
pressed at the,surface 300200 mb. However, it is also quite clearly ex
pressed on AT500 Pressure pattern charts. In our study the position of the
PHAFZ was determined extremely approximately as the position of the isohypse
536 gp dam on AT500 charts. The position of the 536 gp dam isohypse, aver
aged for a 10day period, was plotted on mean 10day maps of the snow cover
for the autumn and spring periods (OctoberDecember and MarchMay). It was
possible to construct a great ntunber of maps for the period from 1949 through
1968. As an example we have given the maps for 1956 (Fig. 1). It should be 
noted that for the autumn period such work had been done earlier [S], and
then supplemented for spring.
As a result of examination of the maps it became obvious that iri autumii the
position of the snow cover boundary has a definite relationship to the
PHAFZ (Fig. 1).
In October, when the snow cover boundary runs along the northern part of
the continent, the position of the PHAFZ (dashed line) coincides quite
well with the isoline 0 days with snow in the 10day period (Fig. la),
 which separates the territory free of a snow cover. In midNovember, when
the snow cover boundary is displaced to the south, the PHAFZ (Fig. lb) is
displaced in the direction of an increase in the values of the isolines
 representing the distribution of the snow cover and in the second 10day
period approaches the isoline for 5 days with snow.
In midDecember, when the snow cover boundary is situated in the southern
European USSR the PHAFZ is displaced still more southward and in the third
_ 10day period in December (Fig. lc) approaches the isoline 10 days with
snow. A more detailed examination of the correlation between the PHAFZ and
the snow cover boundary in the autumn is given in [5].
During the spring period the relationship between the position of the snow
cover boundary and the PHAFZ is also obvious. As might be expected, in mid
March, when the snow cover boundary passes through the southern part of the
European USSR (Fig. le), the PHAFZ is situated closer to the isoline for
10 days with snow during the 10day period. By midApril, when the snow
cover boundary (Fig. lb) is displaced northward, the PHAFZ is shifted in _
the direction of a decrease in the values of the isolines for the distrib
ution of snow cover and is situated closest to the 5day isoline, and fin
ally, in midMay, when the snow cover boundary passes through the northern

part of the European USSR, the PHAFZ approaches the isoline for O.days with
snow (Fig. lf).
Thus, the PHAFZ is displaced in the same direction as the snow cover boun
dary, lagging somewhat within the limits of the region with an unstable
snow cover, which is situated between the isolines for 0 and 10 days with
snow in the 10day period, which is confirmed below by computations.
139
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
Table 1
Correlation Coefficients Between PHAFZ and Position of Isolines for 0, 5
and 10 Days With Snow in 10Day Period
0 1 5 1 10 11 0 1 5 I 10 II. 0 I 5 I 10
1 Ocexb
3
OKTHGPb
4 Noa6pb
II
5 Aci;a6pb
0,78
1 0,55 1 0,56 ~I
0,43 1 0,63 1 0,58
0,52 1 0,53 10,73
2 Becxa
6
MapT
7 Anpenb
II
8 Mati
41.4
I 50,4 I 66,0 II
62,0 1 65,2 1 6411
67,0 1 52,1 156,2
KEY :
1. Autumn 5. December
2. Spring 6. March
3. October 7. April
4. November 8� May
Table 1 gives the results of the correlation coefficients between the PHAFZ
and the position of the isolines for 0, 5 and 10 days with snow in the 10
day period for autumn and spring, averaged by months for 8 years (19491956).
The table shows that for the autumn the highest correlation coefficients in
October (0.78) are observed between the PHAFZ and the isoline for 0 days
with snow, in November with the isoline for 5 days and in December
with the isoline for 10 days with snow, as is indicated graphically in the.
figures. In spring the reverse picture is observed. In March the highest
correlation coefficient is observed between the PHAFZ and the isoline for 10
days with snow, in April with the isoline for 5 days and in May with
the isoline for 0 days with snow.
The physical relationship between the PHAFZ and the.snow cover boundary is
entirely explicable since on the snow cover boundary there is a break be�
tween albedo and heat flows and a zone of maximum contrasts arises near ita
The ascertained dependence can evidently be of prognostic importance since,
on the basis of the position of the PHAFZ to a certain degree it is possible
to judge the position of the snow cover boundary.
In the future plans ca11 for checking the computations of the correlation
coefficients for a longer series of years.
140
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONL�
BIBLIOGRAPHY
1. Afanas'yeva, V. B., Yesakova, N. P., "Correlation Between the Planetary
HighAltitude Frontal Zone and Position of the Snow Cover Boundary,"
TRUDY GGO (Transactions of the Main Geophysical Ob.servatory), No 236,
1960.
2. Bugayev, V. A., Dzhordzhio, V. A., "Planetary HighAltitude Frontal
 Zone," TRUDY TsIP (Transacti,ns of the Central Institute of Forecasts),
No 25, 1951.
3. Bugayev, V. A., "Planetary HighAltitude Frontal Zone a.nd Cyclogenesis,"
METEOROLOGIYA I GIDROLOGIYA V UZBEKISTAN (Meteorology and Hydrology in
Uzbekistan), Izdvo AN Uzb. SSR, 1955.
4. Vorob'ye, V. I., STRUYNYYE TECHENIYE V VYSOKIKH I UMERENNYKH SHIROTAKH
(Jet Streams in the High and Middle Latitudes), Leningrad, Gidrometeo
izdat, 1960.
5. Yesakova, N. P., Afanas'yeva, V. B., "Methods for Characterizing Anomal
ies of Cloud Cover, Snow Cover and Radiation Fluxes," TRUDY GGO, No 143,
1962.
141
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAT, USE ONLY
UDC 551.46.086
COASTAL STATIONARY WAVEMEASURING COMPLEX
Moscow METEOROLOGIYA I GIDROLOGIYA in Russian No 9, Sep 79 pp 113116
[Article by Candidate of Technical Sciences V. B. Vaysband and V. N. Shanin,
Institute of Oceanology USSR Academy of Sciences, submitted for publication
18 September 1978]
Abstract: The article describes an operating
wavemeasuring complex located in the coastal
zone of the Black Sea. The authors give the
technical specifications of a base of rigid
construction on which the wave sensors are
placed. There is a description of these sen
sors and also the circuit diagram of a three
string wave recorder, making it possible to
obtain all the hasic wave elements. The experi
ence of prolonged operation of the complex
demonstrated its effectiveness when making
longterm wave measurements. [Text] At the present time in our country and abroad there has been.a con
siderable increase in interest in longterm instrumental wave observations.
The organization of such observations is complicated by the fact that they
are impossible without a specially fabricated and technically outfitted
wavemeasuring complex.
There is a substantial increase in the requirements on the reliability of
operation of the measuring apparatus and on the stable accuracy of ineasure
ment of wave parameters. However, the satisfaction of these requirements
and conditions when creating a wavemeasuring complex considerably increases
material expenditures. And nevertheless, great material and technical ex
penditures in the organization of longterm wave measurements in the long
run are justified.
A stationary wavemeasuring complex is now operating in the coastal zone in
the Southern Division of the Institute of Oceanology imeni P. P. Shirshov
USSR Academy of Sciences. The afiovementioned requirements were taken into
account in its creation. The wavemeasuring complex is situated in the
142
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
coastal zone and consists of two instruments, one of which is installed in
a sector with a~depth of 5 m, whereas the other is in a sector with a depth
of 7.5 m. Tlie first wave meter measuses the waves at one point, whereas the
other is a threepoint wave meter by means of which it is possilile to deter
mine all the principal wave elements: height, period, velocity, length and
direction of their movement. Tfie spatial positioning of the instruments makes
it possible to take into account the transformation of a number of wave ele
ments.
The wavemeasuring complex consists of the following technological elements:
1) A system for the placement (carrier) of the measurement instrumentation,
2) Sensors of the wave process, 3) Underwater lines for supplying electric
current to the sensors and for transmitting information tothe recorder, 4)
Recording instrumentation, 5) Measurement circuit.
The site of placement of these wave meters is characterized by an intensive
hydrodynamic process. This necessitated the creation and development of some
solid and stable systems for the placement of sensors capable of withstand
ing, over a long period, of high.wave loads and active corrosion processes.
In addition, these carriers must be easily transportable and be installed
in the coastal zone by means of unspecialized ships. The bases constructed
and installed with these requirements taken into account are identical in
design, although with respect to size they differ somewhat from one another.
As an example we will consider the construction of the base of a threepoint
wave recorder (Fig. la).
A vertical steel pipe 2 with a diameter of 150 mm and a height of 6.0 m is
rigidly mounted on a metal platform measuring 2.5 x 2.5 m. Four cast i,ron
weights are placed on the platform in special nests. Each weight has a mass
of 250 kg. In order to impart rigidity the central pipe is connected to the
platform with steel tie rods 4. Attached within the pipe is a telescopic hol
low shaft 5 with a diameter of 85 mm and a height of 6.0 m, which by means
of the collar 6 and the hel.ical stopper 7 can be rigidly fixed at any level
within the limits of water body depths from 7.0 to 8.5 m. Three horizontal
supports 8 with a length of 1.0 m are welded on a telescopic shaft at two
levels at an angle of 120� from one another. Metal braces are attached for
making the supports rigid. At the ends the lower supports have metal funnels
9 dasigned for attachment of the base for the wave sensor body 10. The ends
of the upper supports have tightening collars for attachment of the body of
the urave sensors 11.
The base of the wave recorder, except for the platform, is fabricated from
stainless steel. The total mass of the entire base with the weights is about
1.5 tons. The base of the singlepoint wave recorder is somewhat less in
its dimensions and accordingly also in mass. As the sensors in hoth wave
recorders use is made of a calibrated Nichrome string supplied ac current
with a sonic frequency. Experience in use of sensors of such a type in sea
 water has indicated that thair metrological characteristics are sufficient
ly precise and sta&le with time. However, the use of s.trings unprotected
143
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
from the external mechanical effect of the medium leads to their frequent
damage, and accordingly, cessation of ineasurements. Yn order to preclude
such damage tfie strings were enclosed in light perforated tubes (Fig. lb).
a)
11
f.:
6)
Fig. 1. General view of threestring wave recorder (a) and external appear
ance of body of string sensor of wave height (b).
In the upper part of the tuhe the string 1 was attached in a screw clamp 2,
to which electric current was carried through an underwater cable. At the
bottom of the tuhe the string was attached on the lower currentinsulating
sleeve. The latter was simultaneously the base of the sensor body, entering
into the nest of the guide funnel on the lower support. The total area of
the perforated openings in the s.ensor body was approximately half the total
area of the entire body, whi.ch ensures free escape of water from the tube or
_ its filling during oscillation of the wavecovered water surface, whose
level within and outside the tw dy is virtually identical. Such a design
solution for the sensor ens.ures not only protection of the string against
damage, but also its rapid replacement, which in the practice of wave meas
urements is a circums.tance of more than a little importance. The body of the
144
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
sensors was fabricated from stainless steel. The standard dimensions of
the sensor bodies ensured their interchangeability. A multistrand cable of
 the KVD type with internal shielding is used for supplying electric current
_ to the sensors. This type of cable has not only a high mPChanical strength
but also a high internal waterinsulating capacity.
By virtue of intensive movements of entrained sediments and their mechan
ical effect most often cases of damage to cables occurs in those zones of
� a water body affected by surf or near the water line. In order to exclude
such an effect all the cahle lines were enclosed in metal tubes intersect
ing the dangerous zone along the bottom of the water body. A significant
circumstance having great importance for the preservation of the tubes
themselves is that they are fabricated from stainless steel. The tubes,
made of carbon steel, fail after 12 years of operation. The removal of
the damaged tubes, the extraction of the cables (if they have remained in
tact) and the laying of new tubes each 12 years is far more expensive
than laying of stainless steel tubes.
n, 
R1 L~ CS flg I Cb C9 R1B
Cz I~ TP2 AS L3 + R
21 7 flumeHUe14v
C, f a R6 Rro R~ T Ifanuapamo
�L n? J QomvuK
4 CoHOnuce
_ T1 A~r C7 Clo Riy S i QomJuK
Tz � T3 Tp3 + B Cvriv'nuceu
L % mvuri
R2 Ry R7 R8 R12 A6
4 R16 2 RZ2 B Carionuceu
Tp1 s Iropn'c
7 i
LZ Rs R,J Ce C'f RTo
T 4 ~
CJ  A1 A4+ Tp *
I+ Cy 1 R14 R7 LR� I R23
.L
Fig. 2. Circuit diagram of threestring wave recorder.
KEY;
1.
Current  24 V
6.
Automatic recorder
 2.
Calibrator
7.
Sensor
3.
Sensor
8.
Automatic recorder
4.
Automatic recorder
9.
Body
5.
Sensor
As already noted, the sensors used in the wave recorder are strings with a
high resistance. This principle for measuring waves is relatively success
fully embodied inthe GM61 instrument jl]. However, for a number of tech
nical reasons we could not use the electric power units from the GM61
145
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
instrument for a multipoint wave recorder.
construct a special unit differing somewhat
wave recorder.
sensor 1
sensor 2
sensor 3
It was necessary to develop and
from the unit for a standard
Fig. 3. Fragment Qf synchronous record of wave profiles on tape of N 327/
3 recorder.
The special measuring unit consists of the following main parts: one GNCh
lowfrequency generator with a power of 5 W, three decoupled identical IS
measurement circuits and an AC autocompensator.
The lowfrequency generator ensures the feeding of current to the wave re
corder strings through a decoupling circuit with a frequency of 4 KHz. It
includes three transistors, in the collector of one of them, T3KT802A, four
transformers Tr 1, Tr 2, Tr 3, Tr 4 are cut in inparallel as decoupling
circuits. These ensure supply of electric current to the detector strings
and to the autocompensator (Fig. 2). Such a decoupling method eliminates
harmful potential connections between the sensors and the measuring circuits
and also their mutual influence.
The resistors Rg Rlp, R11 R12 and R13 R14 are for regulating the working
current and voltage across the sensors. The condensers C(, C7, Cg prevent
the penetration of constant current interference from the string to the
measuring circuits. The latter can be created from polarization of the
strings, from stray earth.currents and network interference. The measured
current, by means of the diodes D5_7, forms a positive voltage across the
load resistors; this consists of the constant component Vconst and the
_ useful signal Vuse� At the same time, a negative voltage of the same
strength as Vconst is fed from the autocompensator through the resistors
R21_23 to the resistors R1517� This ensures in R1517 the discrimination
of a useful signal from the change in current as a result of wave oscilla
tions.
146 FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
Since the autocompensator is fed from the lowfrequency generator (ac
current) tnrough the decoupling transformer Tr 1, because of this connec
tion the change in the output voltage of the lowfrequency generator leads
to a proportional change in the current in the measuring and compensating
circuits. In this case the zero level of the current at the output of the
measuring circuit does not change. Such an autocompensation circuit tnakes
it possible to couple the signal zero level with the recorder zero and to
the static level of the sea, which in case of necessity makes it possible
to restore the latter on the automatic recorder tape.
The instrument measuring circuit provides for the possibility of checking !
and routinely changing the scale of the record of wave heights in the process
of ineasuring them. Usually in the course of ineasuring wave heights the scale of
their registry on the recorder tape is determined in the process of calibrat 
ing the sensor at sea. In this case a relationship is established between
the depth of submergence of the sensor in the water Q H and the signal lev
el A S on the recorder tape.
If for any reasons in the course of instrument operation there is a change
in the characteristics of the measuring circuit, it is possible to discover
this deviation only with subsequent calibration. In order to exclude the
uncertainty element, in the circuit of the threestring wave recorder pro
vision must be made for the possibility of current checking. This can be
done because the linear calibration of the sensors is related not only to
the registry of the signal on the recorder tape, but also to the change in
resistance A I of the string. The operation of determining this relation
" ship is accomplished using a precise resistance box which is connected
_ through the switch 7'(1 Tr3 to the measuring circuit in place of the s trin�
sensors. Then from the recorder tape, using the resistance box, a resistance
is selected which causes an amplitude of the recorder p.en identical with
A S.
' After determining 6 I in the entire range !S S a calibration curve in the
form I= f(AH, 0 S) is constructed which in the process of operation of
the wave recorder makes it possible to detect the presence of distortions at
the scale of the wave record. In addition, the presence of such a dependence
makes it possible to change the scale of the wave record in dependence on
wave strength. This was particularly important for us because the channels
of the N 327/3 recorder have a relatively narrow dynamic range (Fig. 3).
 The width of the working part of each record channel is 40 mm. It must be
_ mentioned that in this recorder the record of the readings is by pen on
a tape in a curvilinear coordinate system. However, with a rate of movement
of the recorder tape 10 mm/sec or more the error in determining the wave
profile as a result of the curvilinear form of the record is practically
minimum. At the same time the N 327/3 recorder has a number of technical
qualities which_distinguish it advantageously from other types of recorders
with a linear form of registry.
147
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02148: CIARDP8200850R0402400300394
FOR OFFICIAL USE ONLY 
The N 327/3 recorder is an instrument making it possihle with a high accur 
 acy to register rapidly transpiring processes in the range trom 0 to 50 Hz.
LJith adlierence to normal operating conditions the maximum error in qignal
registry does not exceed the limits 5%. A great convenience in making the
 measurements is that each.measuring channel in the recorder is in the form
= of an individual measurement uni.t, including a measurement mechanism, record
ing device and amplifier.
. i
The experience in operation of the wavemeasuring cowplex over a petiod of
= years has shown that its design characteristics and technical elements cor ~
 respond to the requirements of longterm wave observations in the.coastal ~
_ zone of the sea.
BIBLIOGRAPHY ~
 1. PRIBREZHNYY VOLNOGRAF GM61: METODICHESKIYE UKAZANIYA (GM61 Coastal Wave ~
Recorder: Systematic Instructions), No 38, GUGMS, GOIN, 1974. ~
i
i_
~
148
_ FOR OFFICIAL USE ONLY
~
i
~
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
UDC 551.46:621.396.969
INVESTIGATION OF CORRELATION BETWEEN THE NATURE OF A RAD'AR SIGNAL ENVrLOPE
AND THE FORM OF THE SEA SURFACE REFLECTING SURFACE _
Moscow METEOROLOGIYA I GIDROLOGIYA in Russian No 9, Sep 79 pp 116119
 [Article by Candidate of Technical Sciences I. V. Kire;av and A. V. Svech _
nikov, State Oceanographic Institute, submitted for publication 28 August
1978] 
Abstract: It is demonstrated that anomalous reflec
_ tion occurs from the crests of waves which are in
a stage prior to collapse; the principal contribu
tion to the reflected signal in the case of wind
waves is from sectors of the slope situated clos
er to the crest, whereas for swell those situ
ated closer to its middle; the surface sectors
covered with foam reflect more weakly; on the
basis of the nature of the envelope of the reflect _
_ ed signal it is possible to judge the state of' indi
 vidual wave elements. I
[Text] In developing radar methods for measuring sea waves the emphasis is
usually on determining the correlations between the statistical character
_ istics of sea waves and the radar signal. To a lesser degree the correla
 tions of their fine structures were investigated. An erample of determina
_ tion of such correlatians is source [3], in which it was shown that wave
 crests prior to collapse cause stable bursts of a horizontally polarized 
radar signal which fluctuate little. The determination of such correlations 
already makes it possible to study not only the statistical characteristics
but also i}idividual phenomena.
The study of the correlation of fine structures of waves and the radar sig
_ nal is, indeed, the only direction in study and investigation of the devia
tions, observed in a number of cases, of the quantitative char.acteristics
 of the reflected signal from the characteristics determined by the principal
backscattering patterns. For example, ir. the centimeter range of radio waves
there was found to be an excess of the ~pecific effective scattering area
in the case of horizontal polarization (orflh) over the LSA for vertical pol _
arization (o'~) in the case of waves higher than class 23 and irradiation 149
, FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
of sea waves from the leeward side in the case of small glancing angles.
In the case of irradiation from the windward side this anomalous phenomenon
is not observed. This effect can he explained only on the basis of a study
of the differences in reflectivities of the leeward and windward slopes of
a sea wave.
Research Method
An experiment, whose essence is described below, was carried out on the 18th
voyage of the scientific research ship "Akademik Korolev."
A buoy with corner reflectors was lowered on a cable into the water from the
drifting vessel. Thr. cable was let out to a length which ensured that the
buoy would move 150200 m away from the ship. A weight with a mass of 25
kg was suspended on a cable 50 m in length.to the end of a pipe which passed
through the center of a foam plastic circle in order to stabilize the buoy.
Buoy position was determined from the circular scan display of a"Don" radar
on a scale of 0.8 mile. Then the radar antenna was directed toward the buoy.
The signal received from the reflector was fed from the radar receiver out
put to the input of a wavemeasuring device created at the Radioelectronics
Institute Ukrainian Academy of Sciences. It operated in a moving strobe re
gime. The signal from the output of the peak detector is fed to an HO41
lightray oscillograph.
First only the signal from a corner reflector was recorded, and then, wher
the buoy moved beyond the range of the radar ray, during approach of a
wave of any type to the buoy, a command was fed for registry of the signal
on the oscillograph and there was simultaneous photographing of the sea
surface by a camera with a teleohjective of the "MTO500 A" type. By compar
ing the results of registry of the radar signal envelope on a phototape
with a photograph of the sea surface in the neighborhood of the buoy and
with its visual description, prepared at the same time, a correlation is
established between the nature of the envelope and the form of the reflect
ing sector of the sea surface.
In the experiment it was possible to define the following types of surfaces:
waves without foaming crests.,. waves with foaming crests, wave slopes covered
with foam.
It is known [4] that at one and the same moment in time the input of the
radar receiver receives signals from the reflectors, which are situated
in an area with a length 0.5 c t. This value determines the radar range
resolution.
Since range selection was accomplished by strobing the input of the device,
during the time interval equal to the strob.e pulse duration ('Csec) the
radio pulses moVed along the direction of propagation of electromagnetic
energy by a finite value. As a result, a signal passes through the device
150
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
which will be dependent on the nature and parameters of the reflectors,
situated in an area with the length 0.5 (1, +"6sed c. The peak detector
discriminates the maximum of the signals arriving from this area.
ct~
2
/ 6c  ~sec
Fig. l. Diagram explaining the nature of the signal envelope.
0,02CeK sec
M ~
Fig. 2. Oscillogram of envelope of signal reflected from corner reflector.
1) voltage of range scanning; 2) oscillogram
. ~ ~oOpcex sec
~ Fig. 3. Oscillogram of envelope of signal reflected from sea surface. 1) dur
ing firct perio3 of range scanning; during subsequent period
Now we will proceed to an examination of the process of reflection of radio
pulses from the corner reflector mounted on the buoy and the nature of the
oscillograms of the envelope of a radar signal with a moving strobe, that
is, when the time during which the input of the wavemeasuring device is
"open" is constant and equal to 'G sec and the distance to the area from
which the signal is received continuously changes.
151
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
The radar operating regime during this experiment was as follows: duration
of radio pulse 0.12 � sec, pulse repetition rate 3200 pulses/sec,
strobe duration 0.03 �.sec, intermediate frequency amplification 18
MHz, rate of movement of stroh.ing sector of sea surface 600 m/sec.
The reader should note the absence of a dependence of readings of the wave
measuring device on the shape of the pulse passing through the radar inter
mediate frequency amplifier. The fact is that in the device there is dis
crimination of the envelope of video pulses and for its normal operation
it is only necessary that no pulses be present in the receiving channel.
Special investigations carried out by Siforov [1] demonstrated that the
optimum passband of the receiving channel at the 0.7level falls in the
range (11.3)/,(, , since in this band the voltage amplitude of the radio
pulse at the detector input attains values close to the steady value.
In the considered case the optimtmm band is 8.511 MHz, and accordingly the
amplitude of pulse voltage does not encumber the radar receiving channel.
The initial prerequisites f4r explaining the oscillograms will be examined
using Fig. l. We will assume that the distance to the buoy with the reflector
is equal to Rp. A signal begins to arrive at the input of the wavemeasuring
device from the reflector at the time when the strobe delay is equal to
2 Rn  2 1c ,
l
[c = sec] that is, when by the end of strobe action there is arrival of a
signal caused by approach of the leading edge of the radio pulse to the re _
flector. The distance to the area from which the signal arrives at this time
will be equal to
Ro0,5 lT + 'cc) c
[c = sec]
 Since the leading edge of the radio pulse has only approached the reflector,
the nature of the signal will be determined by the parameters of the sea
surface situated in front of the buoy. With advance of the strobe an ever
greater part of the energy reradiated by the reflector will be fed to the
input of the radar receiver, and accordingly, also the attachment. Some of
the electromagnetic energy scatCered by the nonideal reflector in the direc
 tion of the sea surface is reradiated by the latter and is partially fed to 
the radar input, causing an amplification of fluctuations of the signal en _ velope (Fig. 2). [An inaccuracy in maintaining a right angle by only 1�
leads to a decrease in the maximum value 0"of the corner reflector by a _
factor of 25 due to the scattering of el.ectromagnetic energy.] However,
it can be postulated that the corner reflector is also reached by electro
magnetic waves reflected from the sea surface, which, being scattered by �
the reflector, are fed to the radar input or are reradiated in the direc .
tion of arrival of electromagnetic waves from the sea surface. In the latter
_ case the maximum difference in the path in one direction must be half as
great as in the first (22 and 11 m). 
152
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
The maximum of the envelope of the reflected signal appears with movement
of the strobe in the time interval 0.25G and with a shift into the inter _
val 0.5 (f,+ tSeC) the arrival of the signal from iche reflector in that
time segment when the input is open will end. _
Since after shifting of the strobe by p.25G or more the signal scattered by
the reflector and reradiated by the sea surface no longer arrives at the 
input, the envelope fluctuations decrease considerably (the tr.ailing edge
of the pulse on the oscillogram). _
 The total length of the surface participating in the shaping of the signal,
together with the corner reflector, is equal to c(G +'rsec) �
Figure 3 shows oscillograms of the reflected signal envelope. In the course
of the first period of spatial scanning there is reflection f rom two waves
_ with a length of about 85 m and three shorter waves. In its form the conyid
ered oscillogram is very similar to the oscillogram received from the corner
reflector (Fig. 2). This gives basis for postulating that on ttie front slope
of sea waves there is some intensively reflecting sector whose extent under
experimental conditions was considerably less than the length of the slope
of the sea wave. The reflections from other sectors of the slope lead only
to fluctuations of the reflected signal envelope.
As a result of carrying out a series of experiments it was established that
rather intensive radar signals are regietered from the wind waves, even in
the case of small glancing angles (less Chan 1�), whereas from swe11 we
registered a reflection only in the case of glancing angles greater than
5�, and then only from the slope of the one nearestlying wave. According
ly, the strongly reflecting sector of wind waves is situated at the crest
of the wave, whereas in the case of surf closer to the middle of the
slope. 
This same phenomenon was noted in [3] in measurements with a fixed strobe,
that is, in a study of the temporal invariability of the reflected signal.
As we see, the two methods ("temporal" and "spatial") gave one and the
same result onthe position of the sector of the sea wave slope determining
the strength and nature of the reflected signal.
_ At the peak of the envelope for the signal caused by the first wave and in
itially in the envelope characterizing the reflection from the second wave
it is possible to see sharp dips with an extent of 36 m. By comparing the
nature of the signal with the corresponding photograph we conclude that
they arise with reflection from sectors of waves covered with foam.
The region of the maximum of the oscillogram for the second wave differs
substantially from the similar region of the preceding wave, since it al
most does not fluctuate. As indicated by an analysis of the oscillogram
shown in Fig. 3, and the correspanding photograph of the sea surface, such
a signal arises with reflection from waves near the collapse.stage. The
L9
153
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
characteris.tic dimension of such a sector of the sea surface is 45 m.
On the oscillogram corresponding to the second spatial period (Fig. 3), each
sector of which has a time shift of 0.5 sec from the preceding one, the sig
nal from the first wave zlmost retained its characteristic form and was dis
placed to the lzft by one interval between the time marks (approximately
6 m in space). However, the shape of the envelope associated with the sec
ond wave changed sharply. It was established thst such an envelope shape is
typical for reflections from a collapsing wave with considerable surface sec
_ tors covered with foam.
Thus, confirmation is obtained for observations' of bursts of a signal [3]
when working with a fixed strobe. Allowance for the characteristics of bursts
 can increase the accuracy in measurement of the principal wave parameters
[2]�
The established correlation of the spatial oscillograms of a radiu signal
reflected from the wavecovered sea surface makes it possibl.e to develop a
method for determining the degree of coverage of the sea surface with foam
and also study the physical processes transpiring on Lhe wavecovered sea
surface.
S umma ry
1. It has been established that the spatial positioning of the bursts coin
cides with the crests of waves in the stage prior to collapse. This confirms
the qualitative conclusions drawn in [3] on the nature of these bursts. ,
2. Spatial observations of the signal indicated that the maximum reflection
for wind waves is observed from the sector of the slope situated closer to
the crest, whereas with reflection from swell covered w3.th ripples these
sectors are situated closer to the middle of the slope.
3. The level of the signal reflected from surface sectors covered with foam
is below the level of the signals reflected from the slopes of noncollapsing
waves.
BIBLIOGRAPHY
1. Gutkin, L. S., Lebedev, V. L., Siforov, V. I., RADIOPRIYEMNYYE USTROYSTVA
(Radio Receivers), Part II, Moscow, Sovetskoye Radio, 1963.
2. Zamarayev, B. D., Kalmykov, A. I., Kireyev, I. V., et al., "Methods for
Determining the Characteristics of Waves by the Radar Method," NEKONTAKT
NYYE METODY IZMERENIYA OKEANOGRAFICH.FSKIKH PARAMETROV (Noncontact Methods
for Measuring Oceanographic Parameters), Moscow, Gidrometeoizdat, 1975.
_ 3. Kalmykov, A. I., Kurekin, A. S., Lementa, Yu. A., et al., "Some Peculiar
. ities of Backscattering of Radio Waves in the Superhigh Frequency Range
at the Sea Surface in the Case af Small Glancing Angles," Preprint.Uk
rainian Academy of Sciences, Institute of Radioelectronics, No 40, Khar'
kov, 1974.
154
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
4. SOVREMENNA.YA RADIOLOKATSIYA. ANALIZ, RASCHET I PROYEKTIROVANIYE SISTEM
 (Modern Radar. Analysis, Computation and Designing of Systems), trans
 lated from English, edi.ted by Yu. B. Kohzarev, Moscow, Sovetskoye Radio,
 1969.
~
155
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
REVIEW UF "GENERAL CIRCULATION MODELS OF THE ATMOSPHERE. METHODS IN
COMPUTATIONAL PHYSICS." ACADEMIC PRESS, NEW YORK  SAN FRANCISCO 
LONDON, VOL 17, 1977
Moscow METEOROLOGIYA I GIDROLOGIYA in Russian No 9, Sep 79 pp 120121
jReview by Professor S. A. Mashkovich]
[Text] A volume devoted to models of general circulation of the atmosphere
has been published in the series METHODS IN COMPUTATIONAL PHYSICS. The pre
ceding volumes of this publication were devoted to the status of investig
ations and applications in different fields of physics (statistical physics,
quantum mechanics, hydrodynamics, nuclear physics, astrophysics, seismology,
radioastronomy, etc.).
The fact that the publishers of the mentioned series have turned to the prob
lem of modeling cf general circulation of the atmosphere is entirely legit
imate. In actuality, numerical modeling of general circulation of the atmo
sphere is one of the complex, important and interesting problems of physics.
For its solution it is necessary to use the most powerful electronic comput
ers and the most effective numerical methods. Moreover, further progress in
the modeling of general circulation of the atmosphere (improvement in spatial
resolution in models, detailed allowance for different physical processes,
broadening the sphere of applications of such models, etc.) is essentially
 dependent on the creation of still more productive electronic computers. In
the investigations of this problem there has been graphic manifestation of
hoth the potential and the limitations associated with computational appar
atus. Therefore, the experience in modeling of general circulation of the
atmosphere can be used by specialists working in other fields. For example,
 as mentioned in the foreword to the book, it can be interesting for everyone
carrying out computations with threedimensional models of complex physical
systems.
However, the significance of the book goes beyond this. Recently models of
 general circulation of the atmosphere have heen used more an~d more extensive
ly for solution of a number of practical prohlems relating to the environment.
We can mention their use for study and prediction of changes in climate, for
analysis of anthropogenic influences on climate, for the development of
156
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
numerical methods for longrange weather forecasting, etc. The familiar
ization of.a wide range of readers with approaches to solution of such
problems seems useful. Finally, this hook is also important for meteoral
ogists. The concise exposition of the range of problems associated with
the formulation of numerical models of general circulation of the atmosphere,
the description of specific models of general circulation of the atmosphere
and some results obtained on their basis, all this will unquestionably at
tract the attention of both apecialiats directly concerned with this prob
lem and those desiring to familiarize themselves with it.
The book consists of five articles.
The first of these (A. Kasahara "Computational Aspects of Numerical Models
for Weather Forecasting and the Reproduction of Climate") gives general in
formation on the modeling of general circulation of the atmnsphere (funda
mental equations, choice of schemes for taking,.into..account different phys
ical processes, numerical solution methods, etc.). This is far from a com
plete list of the problems dealt with in the article: radiation processes,
_ prediction of cloud cover and precipitation, atmospheric boundary layer,
~ influence of the oceans, parameterization of movements of a subgrid scale,
_ boundary condition effect at the upper boundary of the atmosphere, differ
ence approximation, scheme with retention of quadratic invariants, implicit
integration methods, nonlinear instability, initial conditions, initializa
tion, fourdimensional analysis, atmospheric predictability. However, it
should be noted that some subjects are dealt with extremely briefly.
The three subsequent articles are devoted to a description of specific fin
itedifference models: fivelevel model of the British Meteorological Ser
vice (G. A. Corby, A. Gilchrist, P. R. Roventree), the model of the National
Atmospheric Research Center in the United States (W. Washington, D. William
son) and a 12layer model of the University of California at Los Angeles
(A. Arakawa and W. Lamb). The first two models differ appreciably with re
spect to the degree of detail in taking physical processes into account.
The British model is more economical, whereas the more detailed allowance
for physical processes in the WashingtonWilliams model makes it more flex
ible. Here there is a graphic manifestation of the dependence of the prac
tical realization of the model on the compromise between the detail of the
physical description and the reasonable choice of duration of calculations.
This can be illustrated by the following figures. The integration of the
British model for 24 hours requires 10 minutes time with an "IBM 360/195"
electronic computer. The National Atmospheric Research Center model (six
layer variant, latitudelongtrude grid with an interval 2.5�) requires 2
hours�time urith a"Suveg 7600" electronic computer and a doubling of the 
 horizontal resolution leads to an eightfold increase in expenditures of
computer time.
The article by Corby, et al. described in detail a fivelevel variant of a
model with a Kurihara grid (4626 points on a sghere) in which allowance for
radiation has been parametertzed and the scheme for the computation of
157
FOR OFFICIAL USE ONItY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
exctiange at the surface operates with a restriction on the number of types
of underlying surface. Nevertheless, the model gives realistic results for
different seasons, as is illustrated hy data on the reproduction of ineteor
ological fields for January and July. More complex variants of the model
are a.lso mentioned: a 13layer"variant (with 8 levels in the stratosphere)
for computing changes in the ozone layer; a variant with a detailed de
scription of radiation taking into account the temporal changes in cloud
cover, computed witfiin tiie framework of the model, intended for evalua
tion of the sensitivity of climatic computations to the cloud cover para
meters.
The description of the National Atmospheric Research Center model contains'
information on the system of equations, boundary conditions and physical
processes taken into account in solving the prob.lem and nunerical approxi
mation of the equations. The article enumerates various applications of the
Model (reproduction of climate and general circulation of the atmosphere,
numerical forecasting, initialization, fourdimensional analysis, etc.).
The article by Arakawa and Lamb devotes much attention to a thorough anal
ysis of the computational aspects of the problem. There is a detailed de
scription of the problem of such spatialfinitedifference schemes which
would ensure maintenance of discrete analogues of some physically important
integral relationships for the continuous atmosphere. The extremely rich
physical content of the model is set forth very concisely.
The last article is devated to modeling of general circulation of the atmo
sphere and numerical forecasting with the use of the spectral method. The
last decade has been characterized by intensive development and use of
nonlinear spectral models of the atmosphere. Effective methods for numer
ical solution of prognosttc equations have been developed and tested. These
make possible a considerable reduction in the volume of computat{ons. The
testing of spectral models demonstrated not only their capacity for compet
ing with finitedifference models, but also definite advantages in compar
ison with the latter. In this connection, at a number of prognostic and re
search centers at the present time spectral models of forecasting, general
circulation of the atmosphere and climate are either used together with
traditional finitedifference models or have replaced them. Therefore,
there is complete justification for including in the book the timely ar
ticle by W. Burke, B. Macavaney, and others on the modeling of.global atmo
spheric currents by spectral methods.
The article describes a variant of a spectral model developed by W. Burke
which tias come into rather broad use abroad. The solution of the problem
is sought in the form of series o.F spherical functions. For computing the
nonlinear terms in the equations use is made of the spectralgrid transform
ation method. Time integration is carried out using a"semiimplicit" scheme,
especially effective in combination with.the spectral method. Also described
 are schemes for including topography ef the earth's surface, radiation pro
cesses, condensation, horizontal and vertical diffusion, surface friction in
158
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
the model, and also a method for stipulating temperature at the ocean sur
face. The article gives the results of operational use of the model for
computing twoday forecasts for the earth's southern hemisphere, and also
preliminary information on computation of glohal circulation of the atmo
sphere (reproduction of January circulation). Thirteen minutes of operation
of the "Suveg 7600" electronic computer are expended for calculating global
circulation.for 24 hours.
The content of the article illustrates well the great possibilities of the
 spectral models. However, one must view critically certain statements made
by the author: "The development of spectral methods for numerical integra
tion of the equations of atmospheric motion can be traced from the study by
 Silberman (1954)," "the pioneering application of spectral methods by Sil
berman in 1954." It is fitting to recall that the spectral approach to solu
tion of prognostic problems and problems in the theory of general circula
tion of the atmosphere and climate was proposed by Ye. N. Blinova (1943),
 and on this basis a number of linear prognostic schemes and models of gen
eral circulation of the atmosphere and climate were created. The successive
approximations method is used at the Central Institute of Forecasts for solu
tion of nonlinear spectral equations. A.solution was obtained in two vari
ants. In the first (19521953) it was.expressed through definite integrals
of the product of Legendre polynomials and their derivatives (these inte
grals are called interaction coefficients by Silberman). In the second vari
ant (1954) the nonlinear terms are computed by conversion from a spectral
representation of ineteorological fields to the values at the points of in
tersection of a regular grid.
Thus, the method for use of spectral methods for solution of prognostic
_ problems must be sought earlier than indicated by the authors of the ar
ticle.
Other comments can be made concerning the book's contents. For example, it
is hard to understand why the book does not include a description of the
model of general circulation of the atmosphere created in the United States
Laboratory of Geophysical Hydrodynamics, which is best developed and which is
widely used for different purposes. Also left out are the problems relating
to modeling of joint circulation of the atmosphere and ocean.
However, these comments do not detract from the importance of
tion. In conclusion it must he pointed out that the book has
bibliography on numerical modeling of general circulation of
and climate.
159
FOR OFFICIAL USE ONLY
this publica
an extensive
the atmosphere
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
;
(
~II REVIEW OF MONOGRAPH BY KH. G. TOOMING: SOLNECHNAYA RADIATSIYA I FORMIROVANIYE
UROZHAYA (SOLAR RADIATION AND YIELD FORMATION), LENINGRAD, GIDROMETEOIZDAT,
1977
Moscow METEOROLOGIYA I GIDROLOGIYA in Russian No 9, Sep 79 pp 122123
[Review by I. A. Shul'gin and I. A. Murey]
[Text] In a series of books on agrometeorology which have been published
 during recent years in the USSR and abroad special and proper attention
has been given to a monograph by Kh. G. Tooming, published by the Gidro
meteoizdat and entitled SOLNECHNAYA RADIATSIYA I FORMIROVANIYE UROZHAYA
(Solar Radiation and Yield Formation). The book is devoted to the theory
of photosynthetic productivity of plants and quantitative determination
of hydrometeorological factors in the processes of biological and economic
productivity of sown crops.
The monograph examines an exceedingly important problem: the possibility
 � of most effective use of the highly important factor of productivity and
the main energy factor, solar radiation, in the process of forir~ation
of biological production with the participation of diverse processes trans
piring in the plant itself.
This problem, by virtue of its exceptional theoretical and practical signif
icance, has attracted the attention of many scientific workers in their
examination of different aspects of activity of the plant cover, natural
and artificial phytocoenoses and the production process itself.
 The intensive development of the theory of photosynthetic activity of coen
oses has led to the formulation of ideas concerning a sown crop as a com
plex, but at the same time, an integrated opticobiological system with
characteristic regularities of formation, distiribution of photosynthetical
ly active radiation, photosynthesis, respiration, growth of the assimila
tion surface, and redistribution of the products in the storing or repro
ductive organs.
Within the framework of this theory there was validation of the position of
 the practical possibility of effective use of solar radiation by means of
formation of a sown area with definite architectonics and leaf area, on the
one hand, and the harmonious nutrition of plants with mineral elements and
160
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
moisture supply in accordance with the receipt of radiati.on, on the other
hand.
In his monograph Kh. G. Tooming took an extremely original approach to solu
tion of the problem. It is possible to agree with the editor of the book,
Yu. K. Ross, who notes in the foreword that the author, being a Candidate
of Physical and Mathematical Sciences and Doctor of Biological Sciences,
heing a master of tfie approaches and methods of the precise sciences and
having a good knowledge of the physiology of the production process, has
 aptly selected the mathematical modeling and numerical experiments approach
for solution of this problem.
Kh. G. Tooming has mathematically formulated the concept that during the
period of growth of the vegetative organs the sown area "tends" to max
imize its. productivity (C02 gas exchange), and accordingly, efficiency as
well. Applying this principle relative to the radiation regime and the
architectonics of the sown area and solving the corresponding variational
problem, he established the quantitative interrelationships between photo
synthesis, respiration and architectonics of plants, on the one hand, and
the radiation regime, on the other. This enabled the author to evaluate the
role of the considered parameters in maximizing productivity, to formulate a
number of general principles useful for the practical work of seedselection
stations.
The book consists of an introuuction, five chapters and a conclusion. The
bibliography contains 501 items, among them 264 in the Russian language.
It is emphasized in the introduction that the main task of agrometeorology,
in connection with the programming of yield, is determination of the upper
 limit of productivity of the principal agricultural crops in different geo
graphical regions on the basis of the receipts of photosynthetically active
radiation and efficiency and development of the agrometeorological prin
ciples for lessening the gap between the theoretically possible and real
yields. This can be achieved: 1) by creating, by means of inelioration and
agrotechnology, of those environmental conditions which would best correspond
" to the needs of the plants in sown areas; 2) by optimum regionalization of
existing varieties in accordance with the climate and microclimate; 3) by
producing and using varieties best corresponding to the environmental con
ditions in this region.
The first chapter characterizes the patterns of receipt of solar energy,
including photosyntlietically active radiation, per unit of horizontal
surface.
The second, and in essence, the main theoretical chapter, is devoted to an
exposition of the mathematical model of the dependence of tlie production
process and yield of phytocoenoses on the regime of photosynthetically ac
tive radiation. Mathematical models of the radiation regime in sown crops
 and photosyntheais as the principal component in the production process and
in the growth of plants are described in a concise form, but in a form which
is clear and accessible for agrometeorologists. Also given are the principal
161
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FJR OFFICIAL USE ONLY
tlicoretical concepts of the author and their mathematical realization. As
we have already noted, the principal scientific merit of the author is the
formulation (mathematically in the form of the variational p roblem) of
the concept of maximum productivity of a phytocoenosis, the solution of
which made it possible to determine the correlations between the radiation
= regime, light curves of photosynthesis and respiration (in the darkness),
under given environmental conditions resulting in maximum p roductivity and
revealing the nature of adaptation of the phytocoenosis to different regimes
of photosynthetically active radiation at its different levels. Although
the physiological mechanisms of such light adaptation are not examined
by the author, however the conclusions themselves unambiguously indicate
the fundamentally important role of the studied processes in the effective
' use of photosynthetically active radiation and the conclusions themselves
agree with the existing concepts. It is important that the Kh. G. Tooming
 concept makes it possible, from unified points. of view, to explain much
_ diverse experimental data on adaptation of plants to light and on the in
terrelationship between photosynthesis and the azchitectonics of plants.
True, the book does not adequately explain (for the most part due to the
lac:k of experimental data at the time of writing of the book) relaxation
= times and the rates of adaptation o'L plants to the regime of photosynthet
_ ically active radiation, which we did only in 19771978.
This same chapter also presents the author's own results on study of the
radiation regime of phytocoenoses. Tfia new concept of "intensity of adap
tation radiation" (IAR) is introduced. ThLs determines the intensity of
photosynthetically active radiation with which this phytocoenosis or its
layer have a maximum efficiency of PAR in the production process. It mdst =
be emphasized that the IAR value corresponds under plant habitat conditions 
_ to the diurnal dose of PAR (Shul'gin, 1973). The relationships between IAR `
and rhe anatomical criteria of leaves following from the computations are
interesting and promising.
The third chapter gives the results of numerical experiments for the pur
poses of explaining the dependence between the productivity of phytocoen
oses and their geometrical structure and the regime of photosynthetically
active radiation, the ideal geometrical structure of the phytocoenosis
and the geographical change in productivity and efficiency of a phytocoen
_ osis in relation to the change in receipts of PAR. The results make a valu _
able contribution to theoretical ecology, make possible a more detailed and
_ deeper understanding of the complex interrelationships between the radia
tion regime, architectonics and productivity of plants under different en
vironmental conditions. These investigations emphasize still more that ~
che role of radiation in the production process is more complex and multi
sided than was assumed earlier by agrometeorolugists,< Radiation is not only =
an energy source for photosynthesis as such (substrate role), but also ex 
 erts an influence on pho tomorpho genesis of a coenosis or plant and their 
architectonics. On the other fiand, tfie intactness of a plant as a system, ~
with its numerous. and interrelated processes, is manifested more clearly.
162 
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
The fourth chapter, which is brief, is devoted to an analysis of Che ef
fectiveness of use of photosynthetically active radiation by phytocoen
oses (efficiency).
 The last, f ifth chapter has the most practical character and is closely re
lated to modern agrometeorology and the theory of programming oF high 
yields, the development of which in our country is related to the studies
of I. S. Shatilov. Proceeding on the fiasis of his theoretical concepts and
 the results of numerical experiments, the author in quantitative form for
mula.tes a series of principles and requirements for the optimization oE
the growth of leaf area in a sown ~rop, optimization of the light curves .for
productivity and growth of plants, develops a series of recommendaLions
for seed selection specialists for obtaining new varieties and maktng more
 effective use of PAR in.a coenosis. It should be noted that some of these
principles were formulated earlier only qualitatively. Among the author`s "
conclusions the most important, possibly, is the conclusion that a change
in any of the plant parameters can lead to an increase in yield only in
a case if there is retention of the optimum extent of the leaf surface in
the sown crop and the optimum area of leaves is not constant. A new varir:otv
with modified indices of photosynthetic activity can give a higher yield �
only in a case if it is possible to create a highly productive (that is,
wellorganized) sown crop. 
_ To be sure, models of the Tooming production process to a certain degree
are characterized by the shortcomings of all such models. This is well
recognized by the author himself, emphasizing that on the basis of his com
putations he formulates only the general principles of processes and evalu
ates the relative role of any factor in the process of yield formatian.
But today this is more important tfian dozens of scattered experimental
 studies. At the same timP, on the basis of many investigations, including
also our 4nvestigations under "factorostatic" conditions for study of the
 photosynthetic activity of coenoses of different density, it can be assert
ed that the Kh. G. Tooming models fundamentally correctly take into account
the role of the radiation regime and architectonics in the production pro
cess.
In general, the book by Kh. G. Tooming has been written on a high theoret
 ical level. The author has good intuition as a natural scientist. Being 
current with the entire world literature in the field of the production
process of phytocoenoses and having the mathematical techniques at his
 command, he was able to write, in compressed form, a book with a very
_ substantial content which is valuable for theoretical ecology and physiol 
_ ogy of plants on the basis of an analysis of the interrelationships between
the receipts of photosynthetically active radiation and processes in phyto
_ coenoses, on the one hand, and for agrometeorology and seed selection for
increasing the yield of agricultural sown crops, on the other.
The book has highquality printing and binding.
The publicatian of this book by Kh. G. Tooming, "Solar Radiation and Yield
_ Formation," is an event in the entire world literature on biogeophysics and
agrometeorology.
163
FOR OFFICI.4L USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONI,Y
SIXTIETH BTRTHDAY OF SEMEN PAVLOVICH KOZNOV
Moscow METEOROLOGIYA I GIDROLOGIYA in Russian No 9, Sep 79 pp 124125
[Article by the Board of the USSR State Committee on Hydrometeorology and
 Environmental Monitoring]
[Text] Semen Pavlovich Koznov, head of the Northwestern Territorial Admin
istration on Hydrometeorology and Environmental Monitoring, marked his 60th
birthday on 19 September 1479.
Semen Pavlovich began his work activity in the liydrometeoro~ogical Service
in 1940 after graduating from the Moscow Hydrometeorological Technical
School. During the years of the Great Fatherland War he was in the active
army on the Leningrad front. He.was presented government awards for partic
ipation in battles with the GermanFascist invaders. Semen Pavlovich entered
the ranks of the CPSU in 1945.
After demobilization, from January 1946 through September 1947, S. P. Koznov
worked in the hydrometeorological service of the Baltic fleet, and from Octo
ber 1947 to the present time all his activity has been associated with the
Northwestern Territorial Administration on Hydrometeorology and Environmental
Monitoring. 164
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFTCIAL USE ONLY
During the period from 1947 through 1949 S. P. Koznov worked as a weather
forecasting engineer. Then he was head of the aviation meteorological sta
tion at the Leningrad airport. During the years 19511959 he was head of the
section on the servicing of aviation and the national economy and deputy head
of the forecasting service. During these years he devoted many efforts to the 
organization of ineteorological support of aviation. During 19491951, Uy di
rection of the Main Administration of the Hydrometeorological Service, he
worked as head of the Forecasting Division of the HydromeLeorological lsureziu
in the Mongolian People's Republic. In 1956, while continuing on the job, he
graduated from the university.
In the years which followed, while working as head of the forecasting service
and head of the Leningrad Weather Bureau, he devoted great attention to study
of the specifics of productive activity of national economic organizations,
the organization of specialized hydrometeorological support, an increase j.n
the qualit; of forecasting of weather and hydrological conditi.ons, investi
gation, development and introduction of new forecasting and information
methods.
Since May 1963 S. P. Koznov has headed one of the largest territorial admin
istrations of the State Committee on Hydrometeorology and Environmental Mon
itoring.
Being the director of a major body of specialists, having rich work experi
ence, high skills and good organizational capabilities, Semen Pavlovich
carries out much work for strengthening operational and observational agen
cies, the organization of a higher quality, specialized hydrometeorological
support of the most important branches of the national economy, such as the
sea and river fleet, civil aviation, and especially agriculture and water
= management.
� Being a member of the Lengorispolkom comnission for contending with disasters,
S. P. Koznov directly participates in the support of Party and Soviet agencies
for forecasts and warnings of Leningrad floods. He is working on study and in
troduction of new forecasting methods.
= He has done much for shifting the processing of the results of hydrometeor
ological observations to computer, for the introduction of new technology
in the network of stations, the compilation and publication of regime and
_ reference materials.
Semen Pavlovich has been repeatedly commended by the directors of the State 
Committee on Hydrometeorology and the Central Committee of the Trade Union
of Aviation Workers for his initiative and practical introduction of new
observational methods, increasing the effectiveness of hydrometeorological
support of the national economy. In 1957 he was awarded the emblem "Dis
tinguished Worker of the USSR Hydrometeorological Service" and in 1970, the
silver medal of the USSR Exhihi.tion of Achievements in the National Economy.
165
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
His activity has been directed to carrying out the new tasks assigned by _
resolutions of the Party and government to the State Coffiittee on Hydro
meteorology. Under his direction and with his direct participation there
has been successful organization of a.. national service for observations and
monitoring of preservation of the environment and automated systems for mon ,
itoring the quality of atmospheric air and surface waters in Leningrad are !
being introduced.
Semen Pavlovich is the chairman of the liasin section "Atlantic Ocean and
Baltic Sea" of the USSR State Committee on Science and Technology. He coor
dinates hydrometeorological investigations of the Baltic Sea.
S. P. Koznov, with great operationalproductive and organizational activity,
is actively participating;in public life. The businesslike and personal qual
ities of Semen Pavlovich, his demanding and sensitive relationship to the
workers, have earned him merited authority in the organizations of the 
State Committee on Hydrometeorology, among the chiefs of its subdivisions and
rankandfile workers.
Semen Pavlovich Koznov meets his 60th birthday full of creative forces. Con
gratulating him on his birthday, we wish him good health, personal happi
ness and new creative successes. "
166
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
SIXTIETH BIRTHDAY OF NIKOLAY YEFIMOVICH ZAKHARCHENKO
Moscow METEOROLOGIYA I GIDROLOGIYA in Russian No 9, Sep 79 pp 125126
[Article by the Board of the USSR State Committee on Hydrometeorology and
Environmental Monitoring]
~
[Text] Nikolay Yefimovich Zakharchenko, head of the Latvian Republic Adminis
tration on Hydrometeorology and Environmental Monitoring, Meritorious Ldorker
in Preservation of Nature Latvian SSR, noted his 60th birthday and 40 years
of productivescientific activity on 24 September 1979.
Pdikolay Yefimovich laegan his work activity in the mid 1930's as an apprentice
metal craftsman in a factory and as a kolkhoz shepherd. After graduating
from the Feodosiya Hydrometeorological Technical School Nikolay Yefimovich
Zakharchenko was designated head of the hydrometeorological station Ostrov
Svinoy of the Azerhaydzhan Administration of the Hydrometeorological Service. _
During the years of the Great Fatherland War tdikolay Yefimovich served in
_ tank units, and after the War, in the Administration of the Hydrometeorolog
ical Service of the Black Sea Fleet. Advancing from senior technician to
167
 FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
head of the technical inspection section, in 1950 he was sent to work at
the Administration of the Hydrometeorological Service Estonian SSR as head
of the section on study of the hydrometeorological regime of the sea.
In February 1453 he was designated deputy head of the Administration of the
llydrometeorological Service Latvian SSR. Being in this responsible post,
Nikolay Yefimovich was cor.stantly studying and in 1963 graduated from the
Leningrad Hydrometeorological Institute in the field of specialization
_ "oceanology." During this period he carried out a number of research stud
ies, writing, in particular, "Transparency of Waters in the Gulf of Riga,"
"Currents in the Gulf of Riga and Methods for Their Investigation," "Hori
zontal Turbulent Exchange and its Relationship to the Resultant Current
Direction," "Heat Balance of the Gulf of Riga and the Kole of Convection"
and others.
With the designation of N. Ye. Zakharchenko to the post of head of this ad
riinistration he displayed his organizational capabilities. He is devoting
considerable attention to improving the forms and methods for hydrometeor
ological support of Party and soviet agencies in LaCvia, which is consider
ably i_ncreasing the effectiveness of use of service data in the national
economy. During recent years Nikolay Yefimovich has been doing much work
in the field of preservation of nature. Under his direction and with his
direct participation a service has been established in the republic for ob
serving and monitoring contamination of the environment.
Taking into account the exceptional timeliness and to a large extent the
newness of the problem of preserving the environment, Nikolay Yefimovich
 is personallv directing research work in this field. For example, studies
have been carried out for evaluating contamination of t.he southern part
of the Gulf of Riga for the planning of purification structures, evalua
tion of contamination of the Liyelupe River for the purpose of issuance of
recommendations on further use of the river and its preservation as a nat
ural water body. A multisided program has been drawn up for the preserva
tion of nature and effective use of the natural resou::ces of the Latvian SSR
during 1976 1990.
A fundamental scheme of interdepartmental information on environmental con
tamination in the Latvian SSR was drawn up with his direct participation
and has been adopted. This work, which was awarded'a gold medal and a
firstdegree diploma of the USSR Exhibition of Achievements in the National
Economy, has been recommended by the USSR State Committee on Hydrometeorology
and Environmental Monitoring for introduction by other administrations in
the country.
Nikolay Yefimovich Zakharchenko is taking an active part in the public life
of the city and republic, being a deputy of the Oktyabr'skiy Rayon Soviet
of Peoples Deputies in Riga where he heads the commission on the preserva
tion of nature. He is a memlier of the expert commission in the Gosplan
168
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
Lithuanian SSR on the problems involved in the dumping of waste water. He
is a member of the scientific council of the USSR State Committee on Sci
~ ence and Technology on the problem "Study of the Oceans and Seas and Use
of Their Resources" and also a memlier of the basin section "Atlantic Ocean
and the Baltic Sea."
From the day of organization of the AllUnion Society of Inventors_ and
Rationalizers he for many years headed the branch council of the AllUnion
Society of Inventors and Rationalizers of Aviation Workers of Latvia; for
more than 20 years he has been a propagandist, director of political and
economic education of the higher echelon of the administration.
For his merits in developing the hydrometeorological service and active
participation in public life he has a number of government awards and twice
has been given diplomas of honor by the Presidium Supreme Soviet Latvian
SSR.
Nikolay Yefimovich is reaching his 60th birthday at the height of his
creative forces.
Warmly congratulating this veteran of the Hydrometeorological Service on
his noteworthy anniversary, we wish him long and productive years of life,
excellent health and new successes in public activity.
169
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
CONFERENCES, MEETINGS AND SEMINARS
Moscow METEOROLOGIYA I GIDROLOGIYA in Russian No 9, Sep 79 pp 126127
[Article by Ts. I. Bobovnikova]
 [Text] The Second AllUnion Conference "Migration of Contaminating Substances
in the Soil, in SoilWater, SoilPlant Systems" was held at the Institute
of Experimental Meteorology. The work of the corference was attended by
80 specialists from institutes of the USSR State Committee on Hydrometeor
ology and Environmental Monitoring, USSR Agriculture Ministry, USSR Health
Ministry, USSR Academy of Sciences, USSR Geology Ministry, USSR Ministry
of Higher Education, USSR Water Management Ministry and others. Fifty
seven reports were presented and discussed, covering the following scien
~
tific directions: physicomathematical model3ng; migration and transformation
of contaminating substances in soils and on the boundaries of the soil with
contiguvus media; study of the migration of pesticides, metals, carcinogenic
 substances, petroleum products and biogenous substances in soils and in con
 tiguous media; hygienic evaluation of soil contamination; influence of an
thropogenic contaminating substances on the physicochemical and biological
indices of soils.
A report by Ya. I. Gaziyev (Institute of Experimental Meteorology) examin
ed the problems involved in formulating prognostic models of soil contamin
ation by the smoke effluent of state regional electric power stations and
gave some evaluations obtained using a model, developed by the author, of
fractional transport in the atmosphere and fallout onto the earth's surface
of the smoke effluent of state regional electric power stations.
A report by Ye. I. Spyn and R. Ye. Sova (VNIIGINTOKS, Kiev) was devoted to
new approaches to the principles for normsetting for pesticides in the
soil. The authors proposed a method .for the validation of norms by computa ;
tions the approximate admissible concentration (AAC) of pesticides in
the soil. _
A report by V. A. Borzilov and N. B. Senilov was a review of work by the
Institute of Experimental Meteorology in creating physicomathematical
models of the behavior of contaminating substances in the environment,
170
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
in particular, their falling onto the soil surface, behavior in the water
and entry into plants.
V. V. Yermakov (VNIIVS, Moscow), in his communication, entitled "Biogenous
Migration of Mercury Under Conditions of Technogenesis in the Biosphere,"
gave particular attention to the detection of inercury allcyl in different
objects in the liiosphere, to tfie transformation of inercury compounds in or
ganisms and the peculiarities of migration of this toxic element in the bio
sphere.
N. G. Zyrin, et al. (Moscow State University) examined the influence of a
major lead and zinc combine on the contamination of soils, defined zones
with different degrees of contamination and demonstrated that the compounds
of heavy metals, entering the soil due to atmospheric contamination, have
a greater soluhility and are more accessible to plants than compounds of
metals from the soils of natural landscapes.
N. F. Beloborodov, L. V. Gavrilov and I. A. Orlik (Central Asian Regional
Scientific Research Hydrometeorological Institute) obtained interesting re
 sults from study of scrption processes of pesticides, regarding the soil
layer as a chromatographic column. The investigations were carried out on
an apparatus developed at this institute.
Ttao sections operated at the conference: "Behavior and Migration of Pesti
cides in the Soil and Adjacent Media" and "Behavior and Migration of Tech
nogenic Contaminants in the Soil and Adjacent Media."
A considerable part of the reports in the first section were devoted to the
behavior of pesticides in the soil and their migration in plants.
A report by F. I. Vayntraub, et al. (AllUnion Scientific Research Insti _
tute of Biological Methods for Protecting Crops, Kishinev) examined the
_ lifetime of some organophosphorus pesticides in the soil and it was demon `
strated that such organophosphorus substances as phosalone, metathione,
phthalophos and siphos can persist in the soil for not more than one grow�
 ing season.
_ V. V. Ivanchenko (Saratov Agricultural Institute), who studied the behavior
of phosalone in the soil, demonstrated that among the many factors, such as
type of soil, humus content, pH, temperature and others, soil moisture plays
a special role in the retention and migration of pesticide.
A communication b.y T. M. Petrova and K. V. Novozhilov (AllUnion Research
Institute of Plant Protection) was devoted to an examination of the factors
favoring the destruction of 12 insecticides in the soil and plants, their
distribution in agricultural crops and the soil for different methods of
use, and also movement of pesticides from the soil into plants.
The reports of F. I. Kopytova (Al1Union Research Institute of Plant Protec
tion) and A. N. Stroy (VNIIGINTOKS), Ye. S. Kovaleva and G. A. Talanov
(VNIIVS) examined problems relating to the migration of organophosphorus
171
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
pesticides from the soil into plants and the duration of their retention
in plants.
 A number of interesting communications on the migration of toxic metals in
soils and in plants were presented by specialists in the Department of Soil
Science and Geography at Moscow State University in the second section.
By investigating the migration of Zn, Cd and Pb in the zone affected by
a zinc smelting combine, R. A. Makarevich, A. I. Obukhov and N. G. Zyrin
_ (Moscow State University) demonstrated that toxic metals, entering onto
the soil surface with the effluent from the enterprises, are concentrat
ed in the upper humus layer and little subject to migration and redistrib
ution in the soil layer. The authors discovered a rather close dependence
between the humus content in the upper soil layer and the concentration
of heavy metals in it.
R. I. Pervunina (Institute of Experimental Meteorology), in collaboration
with N. G. Zyrin (Moscow State University), made preliminary computations
of t}ie balance of cadmium distribution in the soilplant system. These
 demonstrated that the loss of this element from the rootoccupied layer,
as a result of migration through the soil profile is 4058%, whereas the
loss with the crop yield is 0.030.24%. ,
At the conference much interest was shown in a zeport by N. P. Solntseva
 and Yu. I. Pikovskiy (Moscow State University). In the example of soddy
podzolic and soddygleyey soils in southern taiga landscapes he demonstrat
ed the influence of petroleum contamination and contamination of mineral
ized waters on soil transformation. They detected a change in the reaction
of the medium, an intensiflcation of reduction processes, soditmm chloride
salinization and other disruptions of soil processes under the influence
of these contaminants.
The results of an investigation of the influence of contaminating substances
(fluorine, benzapyrene) on the microbiological indices of the soil were pre
sented in communications by specialists of the Institute of Experimental
Meteorology.
E. I. Gaponyuk, T. N. Morshina and N. P. Kremlenkova carried out an investi
gation of the influence of fluorine compounds on some physicochemical and
biological properties of two types of soils soddypodzolic and gray soils.
They demonstrated that the nature of the effect of fluorine is different in
gray soils and in soddywoody soil and that there is a decrease in the ac
tivity of dehydrogenase, phosphatase and urease under the influence of
fluorine when its content is 1000 mg/kg or abovE.
A similar study of the change in the activity of dehydrogenase, phosphatase
and soil respiration under the influence of benzapyrene was carried out by
Ye. M. Vishen?wva, L. V. Vaneyeva, E. I. Gaponyuk, N. P. Kremlenkova and
A. I. Shilina. The most sensitive microbiological indices of soil contam
ination by benzapyrene were found to be the actual dehydrogenase activity
172
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
and the soil respiration process. Actual dehydrogenase zctivity is suppressed
under the influence of benzapprsne.
A report by A. S. Demchenko, L. G. Korotova, M. N. Tarasova and V. S. Zolotar
eva (Geochemical Institute, Novocherkassk) dealt with *_he problem of trans
port of hexachlorocyclohexane hy surface runoff waters. The authors demon
strated that the transport of hexachlorocyclohexane from unirrigated basins
is accomplished by sur.face runoff waters, primarily during the period of
spring high water up to 80% of its surface runoff. In addition, the great
est transport is observed during the first days of the high water.
A communication by Ts. I. Bobovnikova (Institute of Experimental Meteorology)
dealt with the role of atmospheric precipitation (snow and rain) and contam
ination of rivers by stable chlororganic pesticides. An example of this
is the observations made by the author in the Moskva River basin. The quan
tity of DDT and '(hexachlorocyclohexane during the period of the spring high
water was almost equal to the quantity of these pesticides accumulating in
the snow in the basin.
Such results were obtained by V. N. Bashkin and A. Yu. Kudeyarova (IAP [ex
pansion unknown], Pushchino), who studied contamination of the Skniga River
(a tributary of the Oka) by compounds of nitrogen and phosphorus. They dem
onstrated that the contamination of water by nitrogen compounds in a period
of high water is completely attributable to its reserves in the snow cover
in the area of the basin. However, the quantity of phosphorus transported
by the river is more than three times greater than its reserves in Che snow.
A resolution was passed calling for the next conference to be held in 1980.
173
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
NOTES FROM ABROAD
Moscow METEOROLOGIYA I GIDROLOGIYA in Russian No 4, Sep 79 p 128
[Article by B. I. Silkin]
[Text] As reported in SCIENCE NEWS, Vol 114, No 7, p 105, 1978, a scientific
 specialist at the Scripps Oceanographic Institute, J. Namias, at a symposium
"SolarTerrestrial Relationships and Their Influence on Weather and Climate"
 organized by Ohio State University at Columbus, presented a report devoted
to the characteristic meteorological conditions accompanying a drought.
He made an analysis of the meteorological conditions accompanying such out
standing phenomena as the droughts which during 19761977 affected the .
Pacific Ocean Coast of the United States, in 1972 eastern Europe, includ
ing the territory of the USSR, in 1976 the British Isles, in 19521954
the southwestern United States and the intense drought of 1930, entering
into the history of the United States West as the "year of dust storms."
The researcher established that all these catastrophes accompanied:
" 1) the appearance of descending air flows, sometimes attaining 700 m/day,
which as a result of an increase in pressure increases temperature and de
creases atmospheric humidity, prevents the rising of individual air masses
 and the condensation of moisture in them, capable of causing precipitation;
2) formation Qf cells with unusually high pressure directly over the region
of the drought;
 3) formation of unusually powerful "accompanying" highpressure cells along
both sides of the continental high pressure, usually existing only over the
ocean. These "accompanying" cells maintain the continental pressure maximum
proper.
In the researcher's opinion, the formation of "accompanying" cells and the
appearance of repeating droughts can be favored by unusual air temperatures
over the sea surface. The anomalous heating or cooling of water, occurring
for different reasons, can cause an unusual increase in pressure over the
ocean. Since the ocean retzins thermal energy a longer time than the atmo
a sphere, these highpressure regions are capable of creatingconditions in
the atmosphere causing a drought even in Che years which follow. The pro
longing of droughts can also be favored by dust particles in the atmosphere
preventing the formation of a cloud cover.
174
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
 FOR OFFICIAL USE ONLY
Without completely denying the probability of a solar influence on droughts,
the speaker nevertheless emphasized that there must be a careful study of
the global and local "terrestrial" factors also giving rise to catastrophic
phenomena.
As reported in the NEW SCIENTIST, Vol 80, No 1130, p 605, 1978, the meteor
 ologist R. Carrie (United States) has completed a detailed analysis of a
long series of observations carried out at 100 meteorological stations in
North America and their comparison with astrophysical data characterizin g
solar activity. He established a definite correlation between the air tem
perature registered at 55 of these stations and 11year cyclicity of spot
forming activity un the sun. Such a correlation is particularly conspicuous
in the meteorological data collected at stations which are situated in the
_ northeastern part of the American continent where there are 51 of these
meteorological stations.
According to the data f rom this analysis, the mean cyclicity in air temper
ature variations has a period of 10.7 years (f0.4 year) and the amplitude
of its.variations is 0.290C (t0.15�C). In the data it is also possible to
 see evidence of the existence of still another, lesser cycle with an ampli
tude of 0.06�C and a period of 18.6 years, that is, equal to the lunar
nodal period.
As reported in THE SCIENCES, Vol 18, No 9, p 4, 1978, auroras constitute a
powerful electric discharge transpiring at altitudes of approximately 100
to"500 kilometers above the earth's surface in the atmosphere, most fre
quently over the Far North or Far South. The energy source for these aur
oras is the magnetosphere a natural "generator" which can produce ap
proximately 100 billion watts, that is, two orders of magnitude greater
than the quantity of electric power that is used annually in the United
States.
This generator arises as a result of interaction between the earth's intrin
sic magnetic field and the solar wind the flow of heated ionized hydro
gen and helium flowing from the sun.
In many cases this phenomenon is accompanied by serious disrupt.ions of
radio communication and it leads to interruptions in the activity of radar
systems and highvoltage transmiss3.on lines. For this reason a national
economic need arises for the prediction of the appearance of auroras and
their intensity. Up to the present time such attempts have been fruitless.
 Recently a group of specialists headed by SyunIchi Akasofu (Geophysical
Institute, University of Alaska, Fairbanks) completed the processing of
data oIatained as a result of the launching of an artificial satellite
' (International Satellite "Explorer""SunEarth"). The launching of this
 satellite, which took place in the summer of 1978, by NASA in the United
 States, is part of the program of international investigations of the
magnetosphere.
175
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394
FOR OFFICIAL USE ONLY
' An analysis of the information received from trts satellite enabled Akasofu
and his colleagues, taking into account the new data on velocity with which
_ the solar wind is propagated, on the intensity and orientation of the mag 
netic field which it carries along, to formulate a forecast of an aurora. _
 The appearance of this phenomenon and its future intensity can now be pre
dicted for approximately two hours in advance. .
COPYRIGHT: t�Meteorologiya i gidrolo giya," 1979
[15303]
r
5303
 CSO: 1864
m
END
Z76
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/08: CIARDP8200850R0002000300394