MONTHLY, WEATHER REVIEW Editor, EDGAR W. WOOLARD
VOL. 68 No. 7
W. B. do. 1300 JULY 1940 CLOSED SEPTEMBER 4,1940
IESUED OCTOBER 17, 1940
NIGHT RADIATION AND UNUSUAL MINIMUM TEMPERATURES NEAR
NEW ORLEANS, LA.
By W. F. MCDONALD
[U. 8. Weather Bureau. New Orleens, La., April 19391
On the morning of November 30, 1938, aft.er a clear
calm night attending a cent,er of moderately high pressure
over the middle Gulf coast,, a minimum temperature of
43' F. was recorded, 76 feet above ground, at t,he New
Orleans Weather Bureau office. On the same morning at Belle Chasse substation, 5% miles southeast of t'he Weather Bureau office, the minimum temperatmure at 5
feet was lao, a difference of 25' between the two stations. The difference on the prec.eding morning was 24O, with
40' at the Weather Bureau and 16' at Belle Chasse. The Belle Chasse weather station was esta.blished in
the summer of 1935 in c,onnection with an int,ensive
experiment in artificial heating for protection of citrus and high-value winter veget'able crops. Fourt,een ant,hen-
tic cases in less than 4 years since the beginning of these
records show minimum t,emperat,ures at Belle Chasse
20' or more below t.hose re.corded on the same dates at
the We,ather Bureau office. The difference in sea-level
elevation is about 80 feet. Records are obtained from standard Weather Bureau instruments, with three sets
of maximum and minimum thermometers and two thermo- graphs in use at Belle Chasse in standard Weather Bureau
shehers. The area in which this station is located lies within a
huge loop of the Mississippi River, about 7 miles long and of similar breadth. The main river levees surround
the tract on three sides; the fourth is bounded by a drain- age levee. Land slopes are very slight. The point of observation is near the middle of the mea and the ground
there, is approximately 1 foot above sea level with a slight
slope upward toward the river, where elevations are
somewhst more than 10 feet in places, but the top of the main levee line is about 25 feet a.bove sea level. The
drainage lcvee crossing the open side of the river bend is
lower, it,s top being about 12 feet above sea level. The soil throughout the tract consists of the usual heavy black delta silt, known locally as "gumbo." No great fraction of the area is under cultivation; much of
i t is covcred by dense thicket, or forest but the whole area
is artificially drained. The flat basin enclosed by levees
is undoubtedly an ideal place within which to collect a
shallow pool of cold air as the result of loss of heat b radiation when very still, c1ea.r night conditions preva.it
The extraordinary difference of 25' between minima
in the two exposures separated by less than 6 miles of horizonta.1 distance, and only 80 feet in sea-level elevation,
is (so far as the present writer can discover in the avail- able literature) unparalleled in meteorological . records.
The case therefore deserves a full report and discussion which it is the purpose of this paper to present. Records
of several other grouncl-level stations in the flat, densely
overgrown Mississippi delta will be added, for the light
p59231--4(!
they throw on the factors operating to produce the con- ditions described.
Before proceeding to more detailed discussion of the
Belle Chasse temperature records, it may be well to survey briefly some reports of other investigations in the same
field. Cox (2) studied records from the cranberry bogs of
Wisconsin and found extreme differences of 14' to 16'
in the minimum temperatures observed at stations sep-
arated by only about 700 feet. Careful examination of his data shows, however, that the higher of these tempera-
tures occurred within a standard shelter 5 feet above sandy upland, while the lower readings were obtained
from an unshielded minimum thermometer exposed a
very short distance above moss in the bog. Lack of shielding introduces into those records from
3' to 8' of incompatibility and the reported differences in temperatures must be reduced accordingly. Further-
more, i t has more than once been shown (6, 7) that a sharp inversion of temperature amounting to as much as
6' to 8' often exists within 5 feet of the ground surface under active night radiation. This effect is much more
strongly represented in the cranberry-bog records cited
by Cox than in the Belle Chasse records under discussion inasmuch as the lower bog station was in the air layer
within which this ground surface inversion exists whereas
Belle Chasse observations are made at about 5 feet eleva- tion and thus avoid most of the effect of the low surface inversion.
In another paper (3) Cox discusses thermal belts in the Carolina highlands and he there reports a maximum
difference in night temperatures amounting to 31' F., but the stations under comparison differed by 1,000 feet in elevation. Air drainage rather than simple cooling by
radiation enters strongly into tbese highland situations, but this factor must be almost completely absent from
Louisiana delta conditions due to the lack of topographic relief.
Young (8) reports a variation of 38' in adjacent rec-
ords of minimum temperature at stations separated by only a half mile in the Pacific-coast re ion, but there
points involved. Here also, clearly, there was pro-
nounced opportunity for air drainage to affect the
si tuntion. A few years ago Dyke (4) studied local variations in
minimum teniperatures observed within the city of New Orleans, comparing the records obtained a t the Weather
Bureau office with those taken in Audubon Park, 5 feet
above ground. He found the greatest difference in minima to be 16'. The same author examined records
from the Weather Bureau office in Houston, Tex. (almost
was it difference of 225 feet in ground e T evation a t the
181
182 MONTHLY WEATHER REVIEW JULY 1940
300 feet above ground), and those from a station having
standard grouncl exposure in open country at Harrisburg,
Tex., about 35 miles away, but found no difference greater
than 17' in minimum temperatures during the period studied.
A number of situations can be cited in which variations
of 10' to 18' can be found between closely adjacent situn-
tions, especially between roof exposures a t mid-city Weather Bureau stations as compared with nearby subur-
ban records obtained from ground exposures, but no other
case has been found with so much difference where air drainage is excluded.
Table 1 contains details for 14 dates on each of which the minimum temperature a t Belle Chasse was 20' or
more below that a t the Weather Bureau office in New Orleans on the same morning. The average difference for
these 14 dates is 22'. Shown also in this table are records for two additional ground level stations, De!ta Farms and
Houma, both located in the coastal delta region. Climatic
and topographic conditions are much alike a t all stations
named. Delta Farms, like Belle Chasse, has a ground
0.oy ...
FIGURE 1.-Composite isobars tor dates on which di8emoeso22O0 or more were developed between minimum temperatures at Belle Cham and the Weather Bureau office, New Orlesns, La.
surface practically a t sea level, and is on a tract under
art;ficid drainage. The Houma station is located on ground about 10 feet above sea level.
The reliability of the Delta Farms record was some- what dubious until the middle of 1938, but after that time
there was opportunity for satisfactory comparison on 9 of
the 14 dates. These give an average 17' below the mininin
for the same dates a t New Orleans; this is 5' warmer than
the average for Belle Chasse. The readings for Hounia are taken on a cane plantation, and the soil of the area
is a sandy loam. Good records at this station extend over
the entire period covered by table 1, and the average depression of these minima in comparison with New Orleans
is 16', a difference almost identical with that shown for Delta Farms.
Houma lies 45 miles southwest of New Orleans, and Delta Farms is about halfway between that point and New
Orleans. Both stations are located where there is no
forest influence in the immediate surroundings. While unusually cold conditions are shown to develop with some
frequency over a rather wide area near New Orleans, the most intense effects occur a t Belle Chasse, which on some
of these occasions had the lowest temperature officially
recorded in the entire State of Louisiana on the given morning.
In order to have a still more complete setting for these
selected occasions of unusually large abnormality in
temperature, the whole period of the Belle Chasse records
(44 months) was very carefully surveyed and the daily
differences in minimum temperature as related to New Orleans were computed and tabulated.
Results are given in table 2, which shows Belle Chasse minima nearly 8' below those for the Weather Bureau
office, for the year as a whole. The average monthly
differences vary from 4' in Janua.ry and February to 9'
in 0ct.ober. Belle Cha.sse temperatures are 10' or more
below the Weather Bureau office readings on 30 percent
of all the days of record; and 8 perc.ent of the time the difference is 15' or more. The occasions when Belle
Chasse is 15' or more colder than New Orlea,ns are strongly grouped in the last 3 calendar months, occurring
about 1 day out of 5 in the period from October to Deceni-
ber, inclusive. Ten of the fourteen cases of 20' differences
listed in ta.ble 1 occurred in 2 months, October and
November; all lie between October 19 and March 19.
Table 2 reveals a double seasonal arrangement, how-
ever, with lower values a t niidwin ter and midsummer.
(See fig. 3.) Higher values occur in 2 periods of 4 or 5
months each, centered roughly on spring and autumn. This is particularly evident, in the columns showing the
pexentage of cases with 10' and 15' of depression in the
Belle Chnsse da.ily minima. The ge.nera1 background for the more pronounced ca.ses
of cooling a t Belle Cha.sse (listed by dates in table 1)
can be best indicated by composite isobars from daily weather maps attending these occurrences. This com- posite is represented by figure 1, which shows the sig-
nificant tFpe condition, namely, a high-pressure area
centered over southern Louisiana. The individual weather charts from which figure 1 is generalized are more often characterized by a high-pres-
sure ridge thn.n by the localized center shown on the
composite, but in nearly all cases the axis of the ridge lay east-west or northeast-southwest with the center line passing through Louisiana. It goes without saying that
the individual high-pressure areas involved in these situa-
tions are of t.he continental and not the marine type.
The most brilliantly clear skies a t New Orleans occur with the advent of large masses of Pc air and winds of
high velocity in the free air from a direction definitely north of west. Another feature of the general weather situation should
be mentioned. The extreme development of differences in night temperature a t Belle Chasse as compared with New Orleans does not occur immediately upon establish-
ment of true cold wave conditions, but is usual1 found
when temperatures a t New Orleans have passed the lowest
point. This of course is due to the part played by low wind movement in producing these local differences of
temperature. It is only the calm conditions attending
the central area of the anticyclonic formation that favor strstiflcation of cold air a t the eart.h's surface under clear
night skies, which is necessary to establish the strong inversion of temperature involved in the situation.
To illustrate how completely the movement of wind a t t,he low-level station ent,ers as a control on radiation
minima, two composite thermograph traces are shown in
figure 2. Thcse are somewhat idealized and simplified, but in character well represent the march of temperature
on clear nights. The first section of this figure depicts the
niasimuni effect of undisturbed radiat>ion wit,h very low wind movement and shows how under such conditions the
temperature curve for Belle Chasse practically doubles the
range of that for the Weather Bureau oilice. The lower
on the second or even the third night of the cod 9 spell,
JULY 1940 MONTHLY WEATHER REVIEW 183
section of this figure shows a t first the regular down-curve
that is typical of the simple cumulative effect of outgoing
radiation as it increases from early afternoon. The differ- ence in temperature between thermognms a t the two stations increases steadily until an increase in wind
velocity to 4 or 5 miles per hour occurs a t Belle Chasse;
when this happens the tempera tiire irnmediat,ely rises there and the extreme dift'erence in minima cannot thereafter be
established, even though wind movemmt should again
TEMPERATURE
CURVE FOR
MAXIMUM
RADIATION
J
L L
35.-
BELLE CHASSE
W I N D V E L .
'4
' a
2 0
2 Et\ /
0 - *-a- 4- 4-8-1--1- I-w -a
50'-
APPRECIABLE
B E L L E CRASSE
35. -
IO ='
5 : /+
- 20. 0 =
FIGURE 2.-TypimI smoothed thermograms for New Orleans Weather Bureau office (dotted lines) and Belle Chasse substation (solid lines) representing in the upper half the extreme inversion produced under verp calnl night conditions. and in the lower half the effect of an increase in wind velocity to more th!n 4-5 miles prr hour, on the march of night temperatures at Belle Chasse, other condltions being equal.
decrease and permit the Belle Chasse thermogram to re-
sume the normal curvature for night radiation. Both curves show a sharply increased rate of rise in the
forenoon as compared with rate of cooling in the e,arly night hours. This indicates how the night inversion of
temperature (which, under extreme conditions is 20' 60
25' in the 80 feet of difference in elevation of t,he two sta- tions) breaks down with the first inception of morning
turbulence, and warning proceeds by mixing conibine.cl
with direct insolation. The new maximum is thus rea.ched in about half the time required to estd>lish the pre,vioiis
minimum temperature, indicating that mixing, under
extreme conditions, can be as effective as insolation in the warming process. These.
questions are: (a ) Why are radiation minima at Belle
Chssse 5' to 6' lower in the average t.han those a t tho similarly situated stations, Delta Farms and Houma;
and (b ) What is the esplanation for the double seasonal
period in radiational influence revealed by the monthly survey of difference in table 2?
I n considering the first, we note that the high levee
almost surrounding the area in which Belle Chasse is
located has no counterpart a t either of the other ground-
Two questions a.re raised by these observations.
level stations. This 1eve.e is certainly a significant factor
in the observed localization of low-temperature effects
near Belle Chasse, acting doubtless to conserve a pool of cold air. There arc., however, other physical differences in the
environnient of the t h e stations compared, that may be
equally or perhaps even more significant. Belle Chasse stands in a locality that is, in the main, overgrown with
high vege tation including inucli low forest of almost jmigle
density. In cont,rast, Delt,n Farms is surrounded by low- growing marsh vegetation, n.nd at Houma the condition of
t.he adjacent ca.ne. fields ra.nges from a bare cultivated
surface in tlie early part of the yea.r to the dense 10-foot growth of mature cane prior to harvest, near the end of the
year. Several investig:ttors (1, 5 ) have ca.lled attention to the
part played by different types of vegetative cover in pro- ducing variable effects upon night radiation and minimum
air tempera.tures, but tlie role pla.yecl by cover as distinct
from type of soil has seldom been given a.np speciad
emphasis. Some unpublished temperatmure observat,ions made by Arceneaux and Lauritzen st the United States Cane
Esperiment Sta,tion, Houmn,, La.., which bhe present writer has been permitted to examine, indicate very
sbrongly that night ra.diation a t the level of the tops in full-grown stands of suga.rcane produces on very still clear nights a peculiar stratification of the air, such that
the temperature a t the upper level is lower than t,hat a t the ground surface beneath. Later in the season, when
the cane leaves have been killed by frost, this effect is no longer observable; a t t8hat time the lowest temperature
within the same stand of cane occurs, not a t the top but a t the base of the plant.
Cornford (1) cites data (in his study of night tempera- tures in Britain) that directsly confirm these observations
by Arceneaus and Lauritzen. He states, for example,
with reference to a stand of wheat, that "at 3 feet high it
J F M A M J J A S O N D
1 1 1 1 1 1 1 1 1 1 1 1
F. x
SCALE FOR
CURVElI
SCALE
CURVES,I,IU,,IP
- 60%
- 50%
- 40%
- 30%
- 20%
- IO%
-0
FIGURE 3.-Curve I: Composite of (a) the percentage of observations at New Orleans La. showing depression of the wet bulb nmounting to 1 2 O or more. with (b ) monthl; per! centage of total hours between sunset and sunrise (night hours) based on tot31 hours in t i e month. Curie 11: Mean daily difference between miuimum temperatures at New Orleans and Belle Chasse. Curves 111 and IV: Percontnge of days with mini- mum 10' or more colder (111). and 15' or more colder (IV) at Belle Chssse as com- pared with New Orleans Weather Bureau office records.
is colder over the wheat than over the (adjacent) bare
soil. Between the stems of wheat the air is relatively
WC'N~IZL." The temperature differences shown by his de-
tailed data range from 0.7' to 1.7'. The practically uninterrupted surface of dense green
vegntation may therefore be assumed to act as the plane
of maximum night cooling. Such green surfaces probably
approach in effectiveness the rate of black body radiation. At Belle Chasse dense vegetation closely surrounds the
184 MONTHLY WEATHER REVIEW JULY 1940
small acreage of cleared land on which the observing sta-
tion is located, with an average top height estimated a t
15 to 35 feet above ground level. If these tops become
the coldest surfaces under night, radiation there must be
at least a slight tendency for the air from those levels to
settle down into the adjacent clearing in the manner de- scribed in Cornford’s studies (1). With very light air
movement the coldest air should thus accuniulate near
the point of observation, located on agricultural land having relatively low cover. Such effects coulcl not be
expected in equal degree a t Houma or Delta Farms due to lark of similar contrast in level of the vegetative sur-
faces from which nocturnal radiation proceeds. The Louisiana delta region supports veget,ation in re- markable profusion. Native plants are only partially
deciduous and the rest period for annuals and deciduous
perennials is quite short, confined nininly to the 2 months, January and February. When greenery is fully estah-
lishetl tliert. is hardly a square foot of overgrown area
through which Indiation can proceed directly to or from the soil surface. The usual influence of soil type and soil
moisture as affecting night temperatures is thus lacking and there is instead the far more uniform and in general more effective radiation from an unbroken expense of
green leaf surfaces acting somewhat like well insulated
black to bring abqiit nocturnal temperatures lower than similar weather situations can produce in less fertile
regions. It is interesting to note that the thermograms from Belle
Chasse frequently show a slight dip in temperature just about sunrise, coincident with the first increase in air
movement following a calm night. This drop in tempera-
ture appears to result from mixture of the air a t the level
of the recording thermometers (about 5 feet) with a colder stratum from some adjacent level, but whether from a
lower or a higher source i t is impossible with the data in
hand to determine.
In seeking for the solution of the second problem-that
of the double seasonal curve in frequency of larger teniper- atsure differences at Belle Chasse shown by table 2-proba- ble explanations must be more tentatively advanced.
The degree to which minimum temperatures are depressed depends not only upon the character of the radiating surface, heretofore discussed, but also upon the relative
proportions of night and day. The longer nights of au- tumn and winter offer opportunity for a larger cumulative
loss of heat by radhtion than will be possible in the shorter
nights of spring and summer. Hence, if this were the only influence at work, the major inversions in tempera- ture (which really govern the observed differences between
records a t Belle Chasse and in New Orleans) should be commonly noted in autumn and winter, when there should
be a peak in frequency, with decreased frequency in spring and summer. Instead we h d the principal minimum of
fre uency in midwinter and a secondary peak in the spring
This might be partially attributed t,o loss of green coyer by winter-killing during the coldest time of the year, with
a rapid recovery in spring. The greater differential cooling at Belle Chasse in suninier (n-hen the averagc clifTercncc in
daily minimum as compared with New Orleans ~rnounts
an 1 early summer.
to 6’ or 7’ in contrast with the value of 4’ in midwinter) argues for the effectiveness of green vegetation in produc-
ing this midsummer excess. However, this line of reason- ing does not explain the spring peak, 11s there is no peak
in vegetative cover at that season. Some additional factor or factors must therefore be
sought having a variability in the year similar to that of the data under examination. Recalling the fact that the
major temperature cliff erences were recorded when the region was under control of air masses of polar continental
origin, which are characterized by low specific humidity, it seemed logical to search for a practical index to the
incidence of dry air masses at the various seasons of the
year. Fortunately, the depression of the wet bulb, as recorded in 35 years of observations a t the New Orleans Weather
Bureau office, had already been tabulated with the yer- centage of cases with depression of 5’ or more, 8‘ or more,
1 2 O or more, etc., worked out by months. Study of these
diita shows a double periodicity in the seasonal march of larger values, and pronounced spring and autumn maxima. A curve mas finally developed having a reasonable basis in probable causal relationship and showing a seasonal
distribution of magnitudes significantly resembling those
given for cooling a t Belle Chasse (table 2). It was ob- tained by an empirical combination of two percentage figures. The first element, which provides a numerical
index for proportional length of night compared to day, is the percent of total hours in each month that lie between
sunset and sunrise. (This is simply the coniplement of the “total possible hours of sunshine” divided by the “total hours in the month.”) The second element is that
already described, namely, the percent of observations at
New Orleans with depression of the dewpoint amounting to 12’ or more. The simple means of these two monthly
percentage values have been plotted, together with the monthly items from several columns of table 2, to produce
figure 3, where the similarity in the various curves may
be tested by inspection.
It appears that dryness of the atmosphere is quite as important as length of night in lowering noctural tem- peratures near the earth’s surface.
LITERATURE CITED
Cornford, C . E. Katabatic winds and the prevention of frost damage. v. 64. Quarterly Journal of the Royal Meteoro- logical Society. October 1938. Cox, Henry J. Frost and temperature conditions in the cran- berry marshes of Wisconsin. Bulletin T. U. S. Weather Bureau. 1910. Cox, Henry J. Thermal belts and inversions of temperature in the North Carolina mountain region. (Abstract.) v. 47.
Mo. WEA. REV. December 1919. Dyke, Ray A. Nocturnal temperature inversions near the Gulf coast. V. 57. Mo. WEA. REV. December 1929, Haines, Ernest H. Influence of varying soil conditions on night air temperatures. v. 50. Mo. WEA. REV. July 1922. Rellmann, Gustav. (Abstract by R. Corless.) On the cooling of air near the ground a t night. Science Abstracts, Ser. A. December 31, 1918. Tamura, S. Tetsu. Mathematical theory of the nocturnal cooling of the atmosphere. Vol. 33. Mo. WEA. REV. April 1905. Youne, Floyd D. Nocturnal temperature inversions in Oregon and California. v. 49. Mo. WEA. REV. March 1921.
JULY 1940 MONTHLY WEATHER REVIEW 185
TABLE I.-Comparison of mininiu tn temperature observations at Belle Chasse, La., and 2 additional ground-level stations, with the minimum at New Orleans Weather Bureau o&e as the basis for comparison, on 14 dates when Belle Chasse was 20’ or more below
TABLE 2.-Tabulation of daily differences in minimum temperature at Belle Chasse compared with those at the Weather Bureau Ofice in New Orleans. (All temperatures at Belle Chasse are lower than those with which they are compared.) Based on 44 months of record; New Orleans 1936-39
monthly depression of minima at Belle Chasse
OF. 4 4 7 7 7 8 6
7 7 9
8 7
7.6
I I I I
-
10‘ or more 15O or more 2 0 O or more below New below New below New Orleans Orleans Orleans
Percent Permt
-______
16 5 17 0 35 16 37 8 45 5 33 1 12 0 18 0 28 5 42 I9 37 22 34 18
1
Date
Nov. 30, 1935 ___._......_.
Dec. 23, 1936 .____._._____ Mar. 19, 1938 _..__________
Oct. 19. 1936 .____._..._..
Oct. 25, 1938 _.___________
Oct. 26, 1938 ..___________
Oct. 30. 1938.
~
________ Nov. 10, 1938 .____..______ Nov. 21, 1938 ____________.
Nov. 29, I938 ____.__
~
_____ Nov. 30, 1938. ______._____
Dec. 6, 1938 ._______._____
Jan. 20, 1939 _____.________
Average depression of minimum be- low New Or- leans _.___________
Oct. B, 1935 ..___________
--___________
0 0 0 0 0 0
? .__..___ 36 49 29 20 64 43 21 ? ..._.--_ 49
50 26 24 ? .___---_ 33
55 32 23 36 19 39 55 34 21 36 19 40 59 39 20 44 15 43 60 38 22 44 16 42 48 26 22 34 14 32 62 31 21 36 16 36 40 16 24 23 17 24 43 18 25 24 19 27 47 23 24 29 18 32 4 7 2 7 2 0 ? _-_-____ 29
68 45 23 r ____.... 57
-------
__________ ________ 22 ________ 17 ________
RADIATIVE COOLING
Differ- Month
ence
I
Percentage of daily observations with the minimum temperature Average
I
at Belle Chasse-
IN THE LOWER ATMOSPHERE
By WALTER M. ELSASSER
[California Institute of Technology and U. 8. Weather Bureau, August 19401
The writer has recent1 developed a graphical method
for the determination o 9 radiative heat transfer in the
atmosphere (1). This is a modification of the graphical method introduced some years ago by Mugge and Moller
(2). In this method moisture and temperature values of a given atmosphere are plotted on a printed diagram (later
referred to as Radiation Chart) and the radiative flux at any level can be obtained by evaluating an area on the
chart. The results given below represent the first prac- tical tests of our chart. A comprehensive the theory of the chart has just been pub ished (1) and
we shall therefore omit references to the theoretical founda- tions of this work and confine ourselves to a communica-
tion of the results.*
paper
I. FREE AIR COOLING
We used airplane observations of free air moistures and
temperatures. The stations selected (with the exception of Fort Smith, Northwest Territory) are located in two
north-south cross sections over the United States. The
mean values of February 1937 and of August 1937 served as basis for these calculations. The cooling calculated represents the mean cooling in layers 1 kilometer thick due
to the long-wave radiation of water vapor (the cooling due
to carbon-dioxide radiation is found negligible). The
procedure of evaluating the cooling was as follows. First, specific humidity (with a pressure correction applied, see
below) was plotted against pressure. The points were joined by a curve and the total amount of moisture between
successive levels, 1 kilometer distant, was determined by means of a planimeter. These values of total moisture
were then plotted against temperature on the radiation
chart. It is usually possible to plot, on the same chart, curves corresponding to several or to all levels of one
station. The area contained between curves representing
successive levels measures the heat loss of the layer be- tween them; this loss divided by the heat capacity of the
layer gives the net cooling.
*Part of the calculations was carried out by A. C. Gibson of the U, S. Weather Bureau, now at Jacksonville. Fla.
There is still a certain doubt about the manner in which
the air pressure affects the radiative properties of water vapor. According to a theoretical formula (3) the
absorption should be proportional to the pressure, while
F. Schmidt (4) derives from measurements of G. Falcken-
berg (5 ) the result that the absorption is proportional to
the square root of the air pressure. The latter view is
sustained by other, yet unpublished, ex eriments carried
nology. We therefore used the square root pressure
correction in our computations. The figures in table 1 represent mean values of the
cooling in layers 1 kilometer thick. I t is to be under- stood that these layers have nothing to do with the
division of the atmosphere in layers in the manner of Simpson (6). The latter division originates from a
method of approximation where differentials are replaced
by finite differences. Our figures, on the other hand,
represent rigorous solutions of the differential equations
of radiative transfer, once the absorption coefficients of
water vapor are given. It would be possible to calculate
the “local” cooling at any given level, but the determina- tion of the mean cooling of a layer of reasonable thickness
is less laborious and also much more accurate. The
values given in table 1 are in degrees centigrade per day.
All the cooling values contained in table 1 are plotted in figure 1 with the decadic logarithm of the specific
humidity as abscissa. The oblique line represents the
empirical relation
out by John Strong a t the California f nstitute of Tech-
(AT)doy=1+2 log10 w (1)
The two dashed lines are set off from the main line by 0.4O
on each side. I t is seen that the large majority of the
points falls within these boundaries. The major devia-
tions seem to occur in the lowest kilometer; the points
representing these layers are indicated by rings in figure
1. The cause of this decrease in cooling is presumably to be found in the relatively lower mean temperature of
tbe lowest kilometer due to the influence of the nocturnal