EPA-650/4-75-017
      May  1975
Environmental  Monitoring  Series
X'X'X'X^'X'X'X'X'X'X'X'X'X'X'X'X'X'X'X'X'X'XvX'XvX'X'X'X'X'XvX'X'Xv"
X'X'X'X • X*X*X"XvX*X.X"XvX*XvX"X • X'XvXvX'X vXvX.X • •"XvX"X*XvX






-------
                                     ERRATA
Page 6, line 7 should read "is the Stefan-Boltzman constant and T is the
     absolute temperature measured"

Page 15 (Figure 3-1) and page 22 (Figure 3-2) "(/*)" should be deleted
     from the abscissa labels

Page 26, line 2 of Table 3-2 legend should read "READINGS (T/km) AND
     TURBIDITY (I/optical air mass)"

-------
                                   EPA-650/4-75-017
EFFECTS  OF  ATMOSPHERIC AEROSOLS
       ON  INFRARED  IRRADIANCE
       AT  THE EARTH'S SURFACE
   IN A NONURBAN ENVIRONMENT
                      by
                  Michael R. Riches

               Department of Geosciences
             North Carolina State University
                Raleigh, North Carolina

           James T. Peterson and Edwin C. Flowers
                Meteorology Laboratory
                Program Element No. 1AA009
                    ROAP No. 26AAS
                   Prepared for
            ENVIRONMENTAL PROTECTION AGENCY
           Office of Research and Development
          National Environmental Research Center
           Research Triangle Park, N. C. 27711

                   May 1975

-------
James T. Peterson and Edwin C. Flowers are on assignment from the
National Oceanic and Atmospheric Administration, U.S . Department
of Commerce.
                        EPA REVIEW NOTICE

This report has been reviewed by the National Environmental Research
Center - Research Triangle Park, Office of Research and Development,
EPA, and approved for publication.  Approval does not signify that the
contents necessarily reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
                    RESEARCH REPORTING SERIES

Research reports of the Office of Research and Development, U.S. Environ-
mental Protection Agency, have been grouped into series.  These broad
categories were established to facilitate further development and applica-
tion of environmental technology.  Elimination of traditional grouping was
consciously planned to foster technology transfer and maximum interface
in related fields. These series are:

          1.  ENVIRONMENTAL HEALTH EFFECTS RESEARCH

          2.  ENVIRONMENTAL PROTECTION TECHNOLOGY

          3.  ECOLOGICAL RESEARCH

          4.  ENVIRONMENTAL MONITORING

          5.  SOCIOECONOMIC ENVIRONMENTAL STUDIES

          6.  SCIENTIFIC AND TECHNICAL ASSESSMENT REPORTS

          9.  MISCELLANEOUS

This report has been assigned to the ENVIRONMENTAL MONITORING
series. This series describes  research conducted to develop new or
improved methods and instrumentation for the identification and quanti-
fication of environmental pollutants at the lowest conceivably significant
concentrations. It also includes studies to determine the ambient concen-
trations of pollutants in the environment and/or the variance  of pollutants
as a function of time or meteorological factors.
This document is available to the public for sale through the National
Technical Information Service, Springfield, Virginia 22161.

                 Publication No. EPA-650/4-75-017

-------
                           CONTENTS
                                                                      Page
LIST OF FIGURES	iv
LIST OF TABLES	iv
ABSTRACT 	   v
1.  INTRODUCTION AND LITERATURE REVIEW 	   1
      INTRODUCTION     	   1
      LITERATURE REVIEW	   1
2.  EXPERIMENTAL DESIGN	   5
      INTRODUCTION 	   5
      RADIATION INSTRUMENTS	   5
      CALIBRATION OF RADIATION SENSORS 	   6
      MEASUREMENT OF AEROSOLS	   7
      CALCULATION OF THE DOWNWARD-DIRECTED  INFRARED  IRRADIANCE ....   8
      DATA COLLECTION AND SITE LOCATION	11
3.  RESULTS	13
      OBSERVED-MINUS-CALCULATED IRRADIANCE AS A FUNCTION OF
        NEPHELOMETER READINGS AND  TURBIDITY	13
      RELATIVE HUMIDITY RESULTS	  24
4.  DISCUSSION	30
5.  LIST OF REFERENCES	33
TECHNICAL REPORT DATA SHEET	36
                                 ni

-------
                        LIST OF FIGURES

Figure                                                              Page
 3-1      Scatter Diagram for  Observed-Minus-Calculated Irradiance
         Versus  Nephelometer  Readings  (Extinction Coefficient)
         for all Data  (Calculated Irra,diance from the
         Yamamoto Chart)	15

 3-2      Scatter Diagram for  Observed-Minus-Calculated Irradiance
         Versus  Nephelometer  Readings  (Extinction Coefficient)
         for the Hazy  Season  (calculated  irradiance from the
         Yamamoto Chart)	22
 3-3      Scatter Diagram for  Observed-Minus-Calculated Irradiance
         Versus  Turbidity for the Hazy Season  (Calculated
         Irradiance from the  Yamamoto  Chart)	23

 3-4      Scatter Diagram for  Observed-Minus-Calculated Irradiance
         Versus  Relative Humidity for  all  Data (Calculated
         Irradiance from the  Yamamoto  Chart)	28
                        LIST  OF TABLES
Table                                                               Table
 3-1     Regression Analysis  and Analysis  of Variance  for Observed-
         Minus-Calculated Irradiances  (l.y/nrin)  Versus  Nephelometer
         Readings (I/km x 10)  and Turbidity  (1/optical air mass)
         for all Data and for Clean and Hazy Seasons	17
 3-2     Regression Analysis  and Analysis  of Variance  for
         Nephelometer Readings (I/km x 10) and  Turbidity  (I/optical
         air mass) Versus Relative  Humidity  (percent)  for the
         Three Stratifications	26

 3-3     Regression Analysis  and Analysis  of Variance  for Observed-
         Minus-Calculated Irradiances  (ly/min)  Versus  Relative
         Humidity (percent) for all Data and for Clean and
         Hazy Seasons	27
                                IV

-------
                             ABSTRACT
     Atmospheric aerosols can affect the surface radiative  energy
budget through their effects  on  solar (0.3  to  3.0 micrometers) and
infrared (3 to 50 micrometers) radiative transfer.  While many studies
have focused on the relation  between aerosols  and observed  solar radia-
tion, very few in situ measurements  have been  simultaneously made of
aerosol amounts and infrared  radiation.   This  report  describes a study
designed to measure hemispheric  infrared downward-directed  irradiance
at the earth's surface and ambient aerosol  concentrations at Research
Triangle Park, North Carolina.   A Funk type net radiometer  (with a
blackened cavity on the underside) was used to measure the  incident
all-wave energy.  From the value obtained,  the observed solar radiation
was subtracted to determine the  infrared component.  The expected
incident infrared irradiance was calculated from prevailing atmospheric
conditions.  Six methods were used for these calculations:   four
empirical equations based on surface conditions, the Yamamoto  chart,
and a radiative transfer program using vertical profiles of temperature
and moisture.
     The observed-minus-calculated downwelling irradiances  were  then
compared to concurrent measurements of the  turbidity obtained with  a
Volz sunphotometer, nephelometer-indicated  atmospheric extinction
coefficient, and relative humidity.  These  measurements were analyzed
by least-squares regression to determine the extent to which incident
hemispheric infrared radiation is affected  by varying amounts  of

-------
atmospheric aerosols and relative humidity.  The results suggested
that for a typical hazy summer afternoon with 00250 turbidity and
0.2 km~  extinction coefficient the excess downward-directed irradiance
was approximately 0.03 langley per minute, which is some 6 percent of
typical downward infrared irradiances.  A nonlinear dependence of excess
downwelling irradiance on relative humidity was also suggested.
                                   VI

-------
 EFFECTS OF  ATMOSPHERIC  AEROSOLS

         ON  INFRARED IRRADIANCE
         AT  THE EARTH'S  SURFACE
    IN  A  NONURBAN  ENVIRONMENT

   SECTION 1.   INTRODUCTION AND LITERATURE REVIEW

INTRODUCTION
    Atmospheric aerosols can influence the surface radiative energy
budget  through their effects on solar (003 to 3»0 micrometers (ym)
wavelength) and terrestrial infrared (3 to 50 ym wavelength) radiative
transfer.  Many studies have focused on the relation between aerosol
concentrations and observed solar radiation; for example,  Robinson  (1962),
Flowers and Viebrock (1965), McCornrick and Ludwig (1967),  Paltridge and
Platt (1972), and Idso (1972b).  Investigators have, however, made very
few in  situ measurements simultaneously of aerosol concentrations and
infrared radiation. The study covered in this report was  designed to
provide insight into the relation between ambient aerosol  concentrations
and hemispheric infrared irradiance incident at the earth's surface.
    Measurements of hemispheric infrared downward-directed radiation
(HIDR)  at the earth's surface, along with turbidity and nephelometer
readings (indicators of atmospheric aerosol concentrations), were taken
during  cloud-free conditions. The expected HIDR for an aerosol-free
                           1

-------
atmosphere was also calculated at each observation time.   On the basis
of a comparison of the observed-minus-calculated irradiances with the
turbidity and nephelometer readings,  the interdependence  of these
quantities was studied.
LITERATURE REVIEW
     One of the first references to the effect of aerosols on the infrared
irradiance may be found in the literature of more than 20 years ago, as
Robinson (1950) noted a variable component of up to 10 percent in his
measurements of HIDR.  He ascribed this variation to atmospheric
aerosols.  Increased aerosol concentrations were associated with increased
HIDR.  Sheppard (1958) calculated the expected HIDR for an aerosol-
free atmosphere and compared these calculated values to measured
infrared irradiances.  His calculations showed about a 10 percent
excess in the HIDR as compared to an aerosol-free atmosphere, which
agreed with Robinson's conclusions.
     More recently, measurements and theoretical studies of the
influence of atmospheric dust on infrared radiation over Northwest
India (Peterson and Bryson, 1968; Sargent and Beckman, 1973; and
Lai, 1973) have shown an increase in infrared irradiance at the
earth's surface apparently due to the dust.  Using dust profiles and
meteorological data obtained over the Rajasthan Desert, Sargent and
Beckman found as much as a 20 percent increase in HIDR as compared
to that calculated for a dust-free atmosphere.  They also found that
the  increase shown by the model was highly dependent on the amounts
and  vertical distribution of  the aerosol.  The basic results of all
three studies indicated  that  the three  primary effects of the  aerosol
                                  2

-------
were a decrease in the upward infrared flux, an increase in downward
infrared flux, and a decrease in the net infrared flux, compared to
a dust-free atmosphere.
     Other studies on wind-blown dust (Idso, 1972a; 1973) at Phoenix,
Arizona, also have shown a significant increase in HIDR apparently due
to the dust.  Idso found a 12 percent increase for a winter dust storm
and a 4.3 percent increase for a summer dust storm, reemphasizing
Robinson's earlier conclusions.  Staley and Jurica (1972) computed
the effective atmospheric emissivity for an aerosol-free atmosphere.
Measurements made in conjunction with their study suggested additional
HIDR from aerosols.  From aircraft spectral measurements at 8.5 ym to
16 ym over desert terrain, Hovis et al. (1968) also determined that
aerosols have a significant effect on the emissivity of the atmos-
phere.
     Not all researchers agree that increased concentrations of
atmospheric aerosols result in significantly increased HIDR.  Primarily,
the lack of data on aerosol absorption and scattering coefficients in
the infrared region have handicapped modelers and forced assumptions
that may not be totally realistic.  Rasool and Schneider (1971)
included aerosols in their climatic model, but found little effect
on the infrared flux.  Other climatic modelers (e.g., Mitchell, 1971)
have chosen to assume that aerosols have little or no effect on
infrared radiative transfer.  In a dense haze (continental  origin)
over the sea at Coff's Harbour, Australia, Paltridge and Platt (1972)
found no increase in the HIDR as compared to clear-sky data taken on
an earlier expedition (Platt, 1972).  Their computations also verified
this result.

-------
     Recently, two studies of urban-rural  variations  in HIDR have
been reported in the literature.   Oke and Fuggle (1972) measured this
parameter at night, but ascribed the excess HIDR in the urban environ-
ment to warmer atmospheric temperatures, instead of to aerosols.
At Hamilton, Ontario, Canada, Rouse et al. (1973) measured the HIDR
at a rural and an urban site over a 3-year period.  Their data
indicated a significant increase in the HIDR during the day at the
urban site as compared to the rural site.  Little difference was
found at night.  The incident all-wave  (solar plus infrared) radiation
was about the same for both stations.  The authors ascribed these
observations  to higher relative emissivities because of heavy particulate
loading in the urban atmosphere.
     The absorption and emission of infrared radiation by different
substances are wavelength dependent.  This wavelength dependence
results in the so-called atmospheric window (8 ym to 12 jam).  In this
region the two major absorbers of infrared radiation, water vapor and
carbon dioxide, do not influence infrared radiation,,  Thus, it  is
in the atmospheric window that aerosols can most readily affect
infrared radiative transfer.  On the basis of their spectral absorption
characteristics,  there is reason to expect that  aerosols will influence
infrared transfer.   For example, clay minerals  (Flanigan and Delong,
1970) and silica  (Peterson and Weinman, 1969) are natural substances
that have absorption bands in the atmospheric window,,  Man-made
substances   such  as  ammonium sulfate  (Neumann,  1972) and carbonaceous
materials (Twitty and Weinman, 1971) also  have  absorption bands in  the
window region.  For  aerosols collected  mainly from precipitation
samples, containing  natural and man-made  substances, Volz (1972a;  1972b)
found the maximum absorption to be  in the  infrared at  9 ym.
                                  4

-------
              SECTION  2.   EXPERIMENTAL  DESIGN
INTRODUCTION
     The basic experiment consisted of coincident measurements, during
cloud-free conditions,  of hemispheric infrared (3 to 50 pm) downward-
directed radiation (HIDR) at the  earth's surface, and of measurements of
indicated atmospheric aerosol concentrations from turbidity and nephelometer
readings.  At each observation  time, the expected HIDR for an aerosol-free
atmosphere was also calculated  from surface temperature and dew point or
radiosonde data as appropriate  for the calculation scheme.  The observed
turbidity and nephelometer readings were then compared to the observed-
minus-calculated irradiances, and thus the interdependence of these
quantities was studied.
RADIATION INSTRUMENTS
     To determine the HIDR at the earth's surface, two types of instruments
were used.  The net all-wavelength radiation was measured by the Funk (CSIRO)
type net all-wavelength radiometer (Funk, 1959), with the polyethylene dome
from the bottom side replaced by  a blackened cavity.  The inside temperature
of the cavity was continuously  monitored with a thermocouple so that the
instrument effectively  measured the unidirectional irradiance.   Three such
instruments were used during the  course of study„  The downward-directed
solar irradiance was measured independently by an Eppley Precision
Spectral Pyranometer with a  WG7 clear-glass dome, transparent from a
wavelength of about 0.3 ym to 3.0 pm.  The downward-directed infrared

-------
component was then determined from the following relation:
     NET = SWi - SW+ + LWi - TC                        (2.1)
where NET is the reading from the Funk instrument, Sw> is the downward-
directed solar irradiance from the Eppley instrument, and LWi is the
HIDR to be determined.  TC is the upward-directed infrared irradiance
measured by the thermocouple in the blackened cavity (i.e., aT , where a
is the Stefan-Bo!tzman constant and absolute T is the temperature measured
by the thermocouple), and SWi is the upward-directed solar radiation, which
is zero (cavity covering lower sensor).  Thus, the HIDR was calculated
as follows:
     LW4- = NET - SW+ + TC                              (2.2)
CALIBRATION OF RADIATION SENSORS
     To field check the calibration of the Eppley Pyranometer, the direct
component of  the solar beam was shaded with a disc designed to shade out a
solid angle of 5° 43'.  The change in output of the pyranometer  (millivolts)
was then divided by the output of the Eppley Normal Incidence Pyrheliometer
       2    -1
(Cal/cm -min~ ) with a similar aperture.
     The sun  shade method was also used to calibrate the CSIRO radiometers
(Funk, 1961;  and Latimer, 1963).  The error associated with the  calibration
constant thus determined was on the order of ±5 percent.  This value was
estimated  to  include ±1.5 percent error for non-cosine instrument response*
     *An instrument is said to have perfect cosine response if  the energy
from the direct solar beam falling on the horizontal sensing surface of
area A is equal to the energy in  the area normal  to the direct  solar beam
multiplied by cos 0.  The angle 0 is the solar  zenith angle.  If  the
instrument cosine response is perfect, this relation will hold  for all
0 and for all possible solar paths across the sensing surface of  the
instrument.

-------
(Funk, 1959), ±1  percent error for reading stripcharts,  and  a  cumulative  ±2.5



percent error for the shade technique (such as  instrument overshoot  and



lag errors (Latimer,  1963), errors in the Pyrheliometer  and  errors in  the



mechanics of shading  the instrument).  A complete discussion of calibration



techniques and results was presented by Riches  (1974).




MEASUREMENT OF AEROSOLS



     The atmospheric  aerosol content was estimated in two ways:  by



measurements with a sunphotometer (Flowers et al0, 1969) and an integrat-



ing nephelometer (Charlson et a!., 1969).  Both devices  are  based on light-



scattering principles, and have an effective wavelength  of approximately



0.5 ym for the sunphotometer and the nephelometer.  If the ambient aerosol



followed a Junge (1955) size distribution, few  particles outside the



range OJ to 1.0 ym in radius would affect the  sunphotometer or nephelo-



meter measurements.  Even though the most efficient particles  for infrared



emission would likely be somewhat larger than this effective size range



because of the longer wavelengths of the emitted energy, these aerosol



monitoring techniques were selected because of  their ease of operation and



instantaneous output.



     The sunphotometer measures the solar intensity at Oo5 ym  to yield the



atmospheric turbidity, or aerosol extinction coefficient, through the  relation


     T    T     IA~(T , + T0, + B. ) m                  /o 0\
     I, = I   .  10 v  rx    3x    A                     (2.3]
      A    0 A


where I  is the irradiance at wavelength, A, at the observing  point; I
       A                                                              OA


is the extraterrestrial irradiance at wavelength, A; T , is  the (known)
                                                      r A


scattering coefficient for air molecules; TO, is the (known) absorption



coefficient for ozone; B..  is the turbidity coefficient (to be  determined);
                        A

-------
and m is the optical  air mass (path length of the direct solar  beam in
the atmosphere) adjusted for atmospheric pressure at the observer's
location.  The atmospheric turbidity thus obtained is representative of
the entire vertical  extent of the atmosphere above the observer.
     The nephelometer measures the aerosol extinction coefficient by
continuously drawing ambient air into a chamber where the extinction
coefficient is determined by light scatter.  Thus, this measurement
is representative of very local conditions.  Actually, the extinction
coefficient is dependent on scattering and absorption by both gases and
aerosols.  At 0.55 vim (the effective wavelength of the nephelometer),
however, aerosol scattering is almost always the dominant factor affecting
the extinction coefficient.
CALCULATION OF THE DOWNWARD-DIRECTED INFRARED  IRRADIANCE
     There are basically three ways to calculate the expected HIDR.  The
first method is to use an empirical equation obtained from a regression
analysis of actual incoming infrared flux measurements.  Usually tempera-
ture and water vapor pressure  (e.g., Brunt, 1932) or temperature alone
(e.g., Swinbank, 1963) are used as independent variables in the regression
analysis.  Temperature is suggested as a variable by the Stefan-Boltzmann
law, which states that a perfect absorber-emitter should emit a radiant
flux proportional to the fourth power of  its temperature.  The exact relation
     B = EaT4                                     (2.4)
                                                                             /
where:   B  is  the  radiant flux,  E  is  the emissivity,  a  is the Stefan-
Boltzmann  constant, and T is  the  temperature of the  emitter; for a = 0.817
x  10"    ly/(min - °K ), where  T is in degrees  Kelvin.  Water vapor pressure

-------
is also important since water vapor is the major absorber and emitter of
atmospheric infrared radiation.
     The second and third methods involve the solution of the equations
of radiative transfer either by  a chart (e.g., Elsasser, 1942; Yamamoto,
1952) or by computer (e.g., Atwater, 1966).  The derivation of the equa-
tions of radiative transfer is available from several  sources (e.g.,
Kondratyev, 1969).  These methods usually treat absorption by water
vapor, carbon dioxide,  and possibly ozone.  The vertical distribution
of temperature and water vapor is obtained from radiosonde data,  and
carbon dioxide is usually considered well-mixed throughout the entire
atmospheric column.  The resultant infrared irradiances are for an
aerosol-free atmosphere.
     For this study, six schemes were used to calculate the expected
HIDR, since there is no standard method.  Four of the methods were
semi-empirical equations of Swinbank (1963), Idso and Jackson (1969),
Brunt (1932), and Geiger (1965).  The equations of Swinbank and of Idso
and Jackson are based on surface temperature only, while those of Brunt
and Geiger have both surface temperature and water vapor pressure as the
independent variables.   The original derivations of these four equations
were based on best-fit regression curves to actual infrared measurements
and thus were representative of an atmosphere containing some aerosols.
The equations used with the appropriate constants (Morgan et al., 1971)
are given below:
Geiger:  R = EaT4 [a - b • exp (-2.3 ce2)]                   (2.5)
         a = 0.82 b = 0.25 c = 0.094

-------
Brunt:   R = EaT4 (a + b vi^)                             ,    (2.6)
         a = 0.605 b = 0.048
Swinbank:  R = Ea a T6                                       (2.7)
           a = 9.35 x 10"6
Idso and Jackson:  R = EaT4 [1  - c •  exp (-d (273.16-T)2)]    (2.8)
                   c = 0.261 d  = 0.000777
Where:
E = longwave emissivity assumed to be 1.0
a = Stefan-Boltzmann Constant = 0.817 x 10~10ly/(min-°K4)
T = air temperature (°K) at 2 m
e2 = water vapor pressure (mb)  at 2 m
     The fifth and sixth techniques were based on the equations of
radiative transfer.  In one method, the investigators used the algorithms
of Sasamori (1968), which are based on the Yamamoto chart (Yamamoto, 1952).
The algorithms included transmission functions for water vapor, carbon
dioxide, and ozone.  With the inclusion of ozone, it was believed that
the Yamamoto chart was the most theoretically correct of the chart
solutions available.  This belief was the basis for its selection as one
of the six schemes for calculating the expected HIDR.  The final method
involved a computer solution of the radiative transfer equation (Atwater,
1966), based on  the water vapor and carbon dioxide transmission functions
of Davis and Viezee  (1964).  Standard atmospheric distributions of ozone
and carbon dioxide were used as input for these last two computational schemes,
                                    10

-------
DATA COLLECTION AND SITE  LOCATION
     The infrared irradiance measurements were made continuously from
April 1972 through August 1973.  Data were reduced for this study,  however,
only when the sky was cloud-free or when thin cirrus did not extend
beyond 15 degrees above the horizon.  Moreover, the data were evaluated
only when turbidity observations were available, which was usually  once
an hour, 5 days a week.  The nephelometer ran continuously from April
through July 1972 and from October 1972 through August 1973.  The
nephelometer data were evaluated only when both turbidity and infrared
irradiance measurements were available.  At the observation times,  surface
temperature, dew point, and pressure were taken locally or, when local
equipment failed, were estimated from the National Weather Service  hourly
observations at Raleigh-Durham airport about 5 miles to the east over
mostly rural countryside.  The upper air profiles of pressure, tempera-
ture, and dew point (for input to the fifth and sixth computation schemes)
at Research Triangle Park, North Carolina, were estimated for each  observa-
tion time from the 1200Z and OOOOZ radiosondes taken at Greensboro,
North Carolina, which is about 60 miles to the west-northwest and is the
nearest radiosonde station to Research Triangle Park.  Since the radiation
measurements were reduced on "clear" days only, at each measurement time the
1200Z radiosonde data were modified in the lower layers by assuming an
adiabatic lapse rate and a constant water vapor mixing ratio through the
mixing layer.  The absolute values of these profiles were determined from
observed local surface pressure, temperature, and dew point observations.
Unless otherwise indicated by the synoptic situation, a linear time
                                    11

-------
interpolation for temperature and dew point was used between the 1200Z
and OOOOZ data above the mixing height.
     The platform from which the radiation measurements were made is
approximately 23 feet above a mowed grass surface in a nonurban environ-
ment.  There are few obstructions (trees only) above 5 degrees from
the horizon, and there are no obstructions above 10 degrees from the
horizon.  The temperature, dew point, pressure, nephelometer, and
turbidity observations were made 200 feet west of the platform from
which the radiation measurements were taken.
                                   12

-------
                        SECTION  3.  RESULTS
OBSERVED-MINUS-CALCULATED IRRADIANCE AS A FUNCTION
OF NEPHELOMETER READINGS AND TURBIDITY
     The observed-nit nus-calculated irradiance was analyzed as a function
of both the sunphotometer- and nephelometer-indicated atmospheric aerosol
content for all data collected from April 1972 through August 1973.
There were 296 data points for the turbidity comparisons and 231 data
points for the nephelometer comparisons.  The data were stratified three
ways for each type of aerosol measurement.  First, all data were considered.
Then the data were printed out chronologically by the computer, along with
the reported sky condition at the time of observation.  The May-through-
September period for both 1972 and 1973 contained the majority of reported
hazy sky conditions.  On this basis, May through September was designated
the "hazy" season (177 data points) and October through April was designated
the "clean" season (119 data points).  For each of the three stratifications
and the six calculation schemes, the observed-minus-calculated irradiance
was compared to the surface relative humidity, turbidity, and nephelometer
readings.
     Before the results were analyzed in detail, a cursory study was made
of the interrelationship between the values of the infrared irradiance
determined by the six empirical or theoretical schemes and the observed
data.  Infrared irradiance determined by the two formulae based only on
surface temperature data (Swinbank and Idso) showed large diurnal variation
and highest values during midday.  This problem of excessive daytime
                                     13

-------
estimates has  been discussed previously by Paltridge  (1970).   In  general,
the computations  from the two equations based on  surface  temperature  and
vapor pressure and the two techniques  using upper-air data  were consistent.
All Data
     Considering  all  data, investigators found that positive  correlations
resulted between  nephelometer readings and all six schemes  of observed-
minus-calculated  irradiances.  This fact is exemplified by  the scatter
diagram shown in  Figure 3.1, which is  a plot of the observed-minus-calculated
irradiance versus nephelometer readings.  The calculated values were  obtained
from the Yamamoto chart.
     Two points should be made about the data presented in  Figure 3.1.
First, 80 percent of the measured extinction coefficients are less than
0.2 per kilometer.*  This bias results largely from the nonurban
environment of the study site.  Second, the data  show considerable
scatter about the regression line.  This scatter indicates  that the
nephelometer readings are not necessarily a reliable measure of the
extinction coefficient for infrared radiation.  Since nephelometer read-
ings are a measure of the light scatter or extinction coefficient in
the visible range, it is possible for the particle size distribution
to include excess large (small) particles that significantly affect the
infrared (visible) but not the visible  (infrared) radiation.
     *With the use of the nephelometer factory calibrations, an extinction
coefficient of 0.2 knH is equivalent to a visibility of 23 km and a
mass loading of 87 yg m~3.
                                     14

-------
            M M I-H II 1^1 I I 1^1 I  l
'(Qiivnn3iv3-a3Auasao)
            15

-------
     Other contributions  to the scatter shown  in  Figure  3.1  (and  all  other
diagrams presented herein)  are made from several  sources.   The  accuracy
of the instruments and the  errors  introduced during  the  reduction of the
data from these instruments contribute through Equation  2.1.  The
calculated HIDRs are in error to the extent that  the meteorological  input
data are in error.  The formulae themselves are also a source of  scatter
since they are not perfect  predictors of the irradiance, as indicated by
the fact there is no standard method for calculating the expected downward
irradiance.  Lastly, the aerosol measurement techniques  contain instru-
ment and data reduction errors that can also contribute to the  scatter.
     A third point of interest in Figure 3,1 is the  negative intercept
of the regression line.  In Table 3.1 the regression analysis shows that
all but the Atwater program have negative intercepts for the linear
regression.  The negative intercepts are the largest for the two
equations having temperature as the sole independent variable (Swinbank
and Idso formulae).  Next largest are for Brunt and  Geiger formulae,
which have temperature and water vapor for independent variables.  The
intercepts for the theoretical schemes are the smallest, with the regres-
sion for the Yamamoto chart being negative and the Atwater scheme being
positive.  Two possible explanations for the negative intercepts  follow.
First, for summer afternoons, Paltridge (1970) proposed a negative
correction of Oo043 ly/min for the Swinbank formula, based on comparisons
between that formula and observed data.  Paltridge's hypothesis was that
the correction is needed because the original formula was obtained from
measurements made at night.  During daytime, with low-level temperature
lapse conditions, the surface  temperature  is an overestimate of  the
                                    16

-------
Table 3-1.  REGRESSION ANALYSIS AND ANALYSIS  OF VARIANCE FOR OBSERVED-MINUS-
        CALCULATED  IRRADIANCES  (ly/min)  VERSUS  NEPHELOMETER  READINGS
         (I/km  x  10)  AND TURBIDITY  (I/optical air mass)  FOR  ALL
                 DATA AND  FOR  CLEAN AND HAZY SEASONS
Scheme
Irradiance
Idso
Swinbank
Brunt
Geiger
Yamamoto
Atwater
Irradiance
Idso
Swinbank
Brunt
Geiger
Yamamoto
Atwater
Irradiance
Idso
Swinbank
Brunt
Geiger
Yamamoto
Atwater
Irradiance
Idso
Swinbank
Brunt
Geiger
Yamamoto
Atwater
Slope
vs. neph
0.0179
0.0105
0.0058
0.0089
0.0079
0.0048
vs. turb
0.0566
0.0508
0.0057
0.0552
0.0498
0.0212
Intercept
jlometer fo
-0.0676
-0.0614
-0.0384
-0.0528
-0.0283
0.0030
Std error
of slope
r all data
0.0381
0.0385
0.0331
0.0340
0.0309
0.0272
idity for all data
-0.0583
-0.0516
-0.0311
-0.0466
-0.0230
0.0067
vs. nephelometer fo
0.0296
0.0353
0.0122
0.0145
0.0135
-0.0010
vs. turb
-0.2664
-0.3181
-0.3270
-0.3762
-0.3165
-0.3329
-0.0664
-0.0639
-0.0255
-0.0403
-0.0196
0.0178
0.5722
0.5768
0.4982
0.5147
0.4705
0.4169
r clean sea
0.1890
0.1943
0.1631
0.1735
0.1536
0.1440
idity for clean season
-0.0190
-0.0077
0.0150
0.0013
0.0162
0.0422
1 . 5440
1.6020
1.2947
1.3779
1.2073
1.0893
Correlation
coefficient

0.27
0.26
0.17
0.25
0.25
0.17

0.10
0.09
0.01
0.11
0.11
0.05
son
0.15
0.18
0.07
0.08
0.09
0.01

0.17
0.19
0.24
0.26
0.25
0.29
Std error of
estimate

0.0646
0.0653
0.0561
0.0577
0.0523
0.0462

0.0745
0.0751
0.0648
0.0670
0.0612
0.0543

0.0548
0.0563
0.0473
0.0503
0.0445
0.0417

0.0559
0.0580
0.0469
0.0499
0.0437
0.0395
F test

18. 4a
17. Oa
6.9b
15. 7a
14. 8a
7.lb

2.9C
2.3
<1
3.4C
3.3C
<1

2.6
3.5C
<1
<1

-------
  Table 3-1.(continued).   REGRESSION ANALYSIS AND ANALYSIS OF VARIANCE  FOR
    OBSERVED-MINUS-CALCULATED  IRRADIANCES  (ly/min)  VERSUS NEPHELOMETER
     READINGS (I/km x 10) AND TURBIDITY (1/optical  air mass)  FOR ALL
                  DATA  AND FOR CLEAN AND HAZY SEASONS
Scheme
Irradiance
Idso
Swinbank
Brunt
Geiger
Yamamoto
Atwater
Slope
vs. neph
0.0154
0.0154
0.0110
0.0138
0.0116
0.0078
Intercept
jlometer fo
-0.0926
-0.0882
-0.0660
-0.0788
-0.0484
-0.0123
Std error
of slope
r hazy seas
0.0328
0.0325
0.0276
0.0266
0.0266
0.0230
Irradiance vs. turbidity for hazy season
Idso
Swinbank
Brunt
Geiger
Yamamoto
Atwater
0.1464
0.1490
0.1038
0.1535
0.1263
0.0838
-0.0879
-0.0839
-0.0636
-0.0785
-0.0476
-0.0131
0.5814
0.5762
0.5004
0.5128
0.4823
0.4269
Correlation
coefficient
Dn
0.42
0.43
0.37
0.44
0.40
0.32

0.24
0.25
0.20
0.29
0.25
0.19
Std error of
estimate

0.0683
0.0678
0.0576
0.0590
0.0554
0.0480

0.0818
0.0811
0.0704
0.0721
0.0679
0.0601
F test

26. 9a
27. 3a
19. 2a
29. la
23. 4a
14. Oa

ll.la
11. 7a
7.5^
15. 7a
12. Oa
6.7b
Significant at 99.5 percent confidence level.
Significant at 95 percent confidence level.
GSignificant at 90 percent confidence level.
effective radiating temperature of the atmosphere.   He also suggests
applying the correction to other formulae calibrated with night measure-
ments of infrared radiation.  Although this correction can vary with time
of day and with season, the magnitude of the proposed correction is large
enough to decrease significantly or change the sign on the intercepts
presented.
     The second possible explanation stems from the empirical nature of
the first four computational schemes.  Since they were derived from
HIDR observations, these schemes represent an atmosphere in which the
                                    18

-------
aerosol  content is not zero,  but some higher  value.  Thus,  one would  not
expect the observed-minus-calculated differences  to  be  zero at zero aerosol
concentration, but at some higher value.   In  other words, a negative
intercept for the empirical schemes is realistic  if  aerosols do,  in fact,
significantly alter the HIDR.
    The slopes, or coefficients of the independent variable, for  all
six linear regressions presented in Table 3.1 (irradiance vs. nephelometer
for all  data) are positive and statistically  significant at least at  the
95 percent confidence level as determined by  a standard F-test.   Four
of the schemes (Idso, Swinbank, Geiger, and Yamamoto)  are significant at
the 99.5 percent confidence level.  This  fact suggests  an excess  of HIDR
over that expected for an aerosol-free atmosphere.,  The correlation
coefficients range from 0.17 for the Atwater  and  Brunt schemes  to 0.27 for
the Idso scheme.  Note also that in each  case the standard  error  of  the slope
and the standard error of estimate are large.  This  is  a reflection of the
scatter of the data, as pointed out earlier in the discussion.
     Similar regression statistics for the observed-minus-calculated
irradiances as a function of turbidity are presented in Table 3.1
(irradiance vs. turbidity'for all data).   The positive slopes of  the
regression lines are significantly different from zero at  the 90  percent
level for only three of the schemes, which are the formulae of Idso  and
Geiger and the Yamamoto chart.  Moreover, all the schemes  have low correla-
tion coefficients and large standard errors of slope.   The  large  scatter
results in part from the fact that turbidity, like the nephelometer
scattering coefficient, is not necessarily a reliable measure of  the
extinction coefficient in the infrared and in part from the calculation
                                   19

-------
and observational  errors  discussed previously.   The negative intercepts
(Table 3.1  - irradiance vs.  turbidity for all  data) again suggest an over-
estimate for the calculated  daytime HIDR.
     Another reason for the  large data scatter is  that turbidity is  a measure
of the extinction in the entire atmospheric column, whereas  the nephelometer
readings represent the local ground-level extinction coefficient.  Since
about 75 percent of the HIDR typically originates  in the lowest 400  meters  of
the atmosphere (Sellers,  1965), nephelometer readings might  be expected to
be better correlated to the  excess irradiance than turbidity is.
Clean Season Data
     Data for the clean season (October through April) were  generally
inconclusive.  The regression analyses for the observed-minus-calculated
irradiances as a function of extinction coefficient and turbidity are
presented in Table 3.1 (irradiance vs. nephelometer for the  clean season
and irradiance vs. turbidity for the clean season, respectively).  The
data  had a  strong bias toward  low  values of turbidity and extinction
coefficient and large  scatter  about  the  regression  lines.  The  generally
poor  results were likely because  the  data  clustered about low aerosol
values  resulting  from  the generally  clean, dry  atmospheric conditions.
Moreover,  the  large  scatter also  resulted  from  (1)  the  use of turbidity
and  the extinction coefficient in  the visible range to  represent the
infrared extinction  coefficient;  (2)  the fact that turbidity  is a weak
function of the lower  atmospheric  layer  responsible for the majority  of
the  HIDR,  as discussed earlier;  and  (3)  the observational and calculation
errors.
                                   20

-------
Hazy Season Data
     The regression analysis and the analysis of variance for the
observed-calculated irradiance as a function of nephelometer readings
for the hazy season (May through September) are presented in Table 3.1
(irradiance vs.  nephelometer for the hazy season).   Note that the positive
slopes for all  six schemes are significantly different from zero at the
99.5 percent confidence level and that the standard error of the slope
is smaller than that for the slope in the cases discussed previously.
Correlation coefficients are about three times larger than those obtained
for the clean season data, and twice that obtained for the total data
set.  The more significant statistical results from the hazy season suggest
that the narrow range of aerosol concentrations during the clean period
was not sufficient to elucidate an aerosol-irradiance relation.
     The intercept for all six regression equations is negative (see
Table 3.1 - irradiance vs. nephelometer for the hazy season).  As discussed
previously, this fact is not unrealistic for the empirical schemes, but
could be because of an overestimate of the infrared irradiance by the
schemes for daytime observations.  The scatter of  the data about the
regression as shown in Figure 3.2 is also large for these hazy-season
measurements.
      In contrast to the statistics presented for the clean season, the
statistics for the irradiance differences as a function of turbidity
during the hazy months (Table 3.1 - irradiance vs. turbidity for the hazy
season) show a positive slope that is significantly different from zero
at least at the 95 percent confidence level for all six schemes„  The
standard error of the slope is the smallest yet found for turbidity
regressions.  The wider and more evenly distributed range of turbidity
                                  21

-------
:j  <=*.
         LU
         DC
         CC


         £

         o
- -       UJ
                cn
                c
               JE ra
                P-E
         j
         C9
         Z
- -  o    o     J5
                !D QJ
               ^; o
-«t5

 91
 en—-
               "O e/>
                
-------
values (see example  scatter diagram in Figure 3.3)  allows  more physical


significance to be given to the positive slopes found  for  the hazy


season, even though  considerable scatter about the  regression line is


still present.  Similar to the regression on nephelometer  readings for


the hazy season,  the correlation coefficients are about twice as large


as those found for the total data set, averaging about 0.24.   They are


smaller, however, than those found for the nephelometer readings.  These


facts show that turbidity is more indicative of the lower-atmospheric


aerosol concentration during the hazy season than during the  clean season,


but that it is not as reliable a measure as the nephelometer readings.


     The wider and more evenly distributed range of turbidity values and


nephelometer readings found for the hazy-season data allow a  discussion


of the physical meaning of the regression equations derived earlier.   If


the six results of the observed-minus-calculated irradiance are used as


          0.2
      a
      UJ
0.1
      cc
      LLJ
      <
      a
      <
0.1
                0.2
                                  0.3     0.4      0.5


                                 TURBIDITY, I/optical air mass
0.6
0.7
0.8
       Figure 3-3. Scatter diagram for observed-minus-calculated irradiance versus
       turbidity for the hazy season (calculated irradiance from the Yamamoto chart).

                                     23

-------
a function of nephelometer readings  (extinction coefficient [knf  ])
during the hazy season,  the average  slope of the regression line  is
0.125 (ly/km-min).   Thus, the results  presented herein indicate that
an ambient low-level atmospheric aerosol  concentration with extinction
coefficient of 0.1  km   would increase the HIDR by 0.0125 ly/min.  For a
typical  summer afternoon with an extinction coefficient of 0.2 km" , the
excess irradiance is 0.025 ly/min.   The average slope from the six schemes
for the data of all seasons (with nephelometer readings considered)  is
0.093 ly/km-min.
     The average slope of the linear regression line between the  irradiance
difference during the hazy season and the turbidity coefficient is 0.127
ly/optical air mass-min.  Thus, for  a typical summer afternoon in North
Carolina with a turbidity of 0.250,  which is similar to a nephelometer
extinction coefficient of 0.2 km" ,  the excess downward irradiance would
be 0.032  ly/min.  This value (0.032 ly/min) represents about 6 percent
of the typical HIDR total.
RELATIVE  HUMIDITY RESULTS
      It  has  been established that relative humidity is an important
parameter  in aerosol growth.  For example, Covert et al.  (1972)  and
Winkler  (1973) have shown  that ambient aerosols grow as relative  humidity
increases.   Since  the absorption coefficient for aerosols depends
directly  on  its radius  (Deirmendjian,  1969), the increased  particle size
should increase the particle absorption  (or emission)  in  the  infrared.
The  adsorbed or absorbed water should  also increase the particle  absorp-
tion  (or emission)  since water is an effective  absorber in  the infrared
 (Kondratyev, 1969).  Thus,  relative humidity should correlate with  the
observed-minus-calculated  HIDR.  Since scatter  by aerosols  is proportional
                                 24

-------
to particle size (Kondratyev, 1969), nephelometer readings and turbidity
should also correlate with the surface relative humidity.   To test these
hypotheses, the observed-minus-calculated HIDR, turbidity, and nephelometer
data were analyzed by least-square linear regression as a  function of
surface relative humidity.
     In Table 3.2, the regression analyses and analyses of variance are pre--
sented for all  data, and for clean and hazy stratifications for nephelometer
readings and turbidity as a function of relative humidity.  The results
are as suggested above.  The nephelometer readings are positively
correlated with relative humidity for all three stratifications, with
correlation coefficients of about 0.55; and the positive slopes are
statistically significantly different from zero at the 99.5 percent
confidence level, with relatively small standard error of  the slopes.
Note that the slopes for the clean season are about four times larger
than those for the hazy season.  This fact may be the result of the bias
toward low nephelometer readings during the clean months as discussed
previously.  This large change in slope is most likely not physically
significant.  The results for the "hazy" data set are similar to those
obtained for all the data.
     The results for turbidity as a function of relative humidity are
similar to those presented for nephelometer readings.  The slopes are
positive and significantly different from zero, at least at the 95 percent
confidence level; but the correlation coefficients are lower, and the
standard errors of slope are larger than the corresponding statistics
for the nephelometer readings.  Again, the slope for the clean season
differs from those obtained for the two other stratifications, and this
difference is attributed to the bias toward low turbidity  during the
clean season.
                                   25

-------
   Table 3-2.   REGRESSION ANALYSIS  AND  ANALYSIS  OF  VARIANCE  FOR  NEPHELOMETER
           READINGS (I/km x 10)  AND TURBIDITY  (I/optical  air mass)
      VERSUS RELATIVE HUMIDITY  (percent)  FOR THE THREE  STRATIFICATIONS
Scheme
Nephelometer
All data
Clean season
Hazy season
Turbidity
All data
Clean season
Hazy season
Slope

5.81
28.90
4.54

59.29
97.45
42.73
Intercept

44.6
22.3
50.2

47.4
41.5
53.0
Std error
of slope

8.78
43.55
7.08

125.5
458.0
111.6
Correlation
coefficient

0.55
0.55
0.54

0.43
0.21
0.36
Std error of
estimate

14.9
12.6
14.8

16.3
16.6
15.7
F test

100.3a
46. 2a
50. 3a

65. 6a
5.3b
25. 6a
Significant at 99.5 percent confidence level.
^Significant at 95 percent confidence level.

     The regression analyses  and analyses  of  variance for the  observed-
minus-calculated HIDR as a function of relative humidity for all  six
schemes are presented in Table 3.3 for the total  data set, clean  season,
dnd hazy season, respectively.  All 18 slopes are positive and
significantly different from zero at the 99.5 percent confidence  level.
The correlation coefficients are highest for  the Idso and Swinbank
schemes and lowest for the Atwater scheme, with the other three schemes
in the middle.   Since Idso and Swinbank methods do not contain water
<  r/or as an independent variable, the higher  correlations may  reflect a
need to include water vapor as a parameter, rather than a better  correlation
with particle growth.  The seasonal (clean versus hazy) difference in
correlation coefficients is the result of the dryer air dominating the
clean season.  In other words, during the clean season, the absolute
                                    26

-------
 humidity  is  less  than  that  found during  the hazy season, which  is fre-

 quently influenced by  a warm, moist air  mass of maritime origin.  Thus,

 high  relative  humidities  during the clean season may not be associated

 with  enough  water vapor for significant  particle growth.


  Table 3-3.  REGRESSION ANALYSIS AND ANALYSIS OF VARIANCE FOR OBSERVED-
     MINUS-CALCULATED IRRADIANCES (ly/min)  VERSUS RELATIVE HUMIDITY
          (percent) FOR ALL DATA AND FOR CLEAN AND HAZY SEASONS
Scheme
All data
Idso
Swinbank
Brunt
Geiger
Yamamoto
Atwater
Clean seas
Idso
Swinbank
Brunt
Geiger
Yamamoto
Atwater
Hazy seaso
Idso
Swinbank
Brunt
Geiger
Yamamoto
Atwater
Slope

0.0025
0.0025
0.0017
0.0020
0.0017
0.0012
Dn
0.0019
0.0020
0.0012
0.0012
0.0010
0.0006
1
0.0039
0.0039
0.0029
0.0033
0.0030
0.0022
Intercept

-0.1922
-0.1863
-0.1240
-0.1485
-0.1139
-0.0599

-0.1319
-0.1300
-0.0720
-0.0856
-0.0547
-0.0132

-0.3014
-0.2953
-0.2240
-0.2534
-0.2062
-0.1331
Std error
of slope

0.0033
0.0033
0.0032
0.0032
0.0029
0.0027

0.0027
0.0028
0.0026
0.0028
0.0025
0.0023

0.0031
0.0031
0.0031
0.0030
0.0029
0.0029
Correlation
coefficient

0.61
0.60
0.46
0.52
0.51
0.41

0.57
0.59
0.43
0.40
0.37
0.27

0.79
0.79
0.69
0.75
0.71
0.61
Std error of
estimate

0.0594
0.0601
0.0575
0.0574
0.0529
0.0495

0.0464
0.0480
0.0438
0.0474
0.0420
0.0398

0.0521
0.0517
0.0522
0.0501
0.0490
0.0485
F test9

173.5
169.1
79.8
111.2
104.4
59.9

58.1
61.2
25.8
22.6
18.7
8.9

283.5
284.6
157.4
220.5
184.4
104.0
aAll data in this column significant at 99.5 percent confidence level.

                                 27

-------
      Secondly, notice  the large negative intercept for all  18 cases.
 This can  be accounted  for in part  by the tendency of the calculation
 schemes to overestimate  the irradiance for daytime observations.
 The negative intercept may also result from the  mechanics of  particle
 growth.   Covert et  al . (1972) show that particle growth for relative
 humidities  below 60 percent is very  small.   For  relative humidities
 greater than 60 percent,  particle  growth increases  very raoidly with
 relative  humidity.  Figure 3.4 shows the positive corre'iati! i between
 observed-minus-calculated irradiance and relative hum'uity, which
 presumably results  from  aerosol growth dependence on relative humidity.
 Because the aerosol growth is strongly nonlinearly dependent  on  humidity,
 however,  a nonlinear curve should  be used to fit the data of  Figure 3.4.
 These data do suggest  such a curve with wide scatter and little  apparent
 slope at  relative humidities less  than about 60  percent.
     0.24-
c
1
"
     0.14-
>
cc
CO
00
cc
cc
             iiNiiHniMiiiiiiiiintMnniiiiiiiiiiiHlniiiiuiMiiiiitMiiiMiiiiniiiiuiiiiiiniiiinHMiHiiiiniiniiniiiiMiuiiiiiiiiiitiiiiiiiiiiiiiiiiiniiiiMiiiiiiiiiitiitiiiiiiiiii
       0      10.0     20.0    30.0    40.0     50.0    60.0    70.0     80.0     90.0    100.0
                                  RELATIVE HUMIDITY, percent
   Figure 3-4. Scatter diagram for observed-minus-calculated irradiance versus relative humidity
   for all data (calculated irradiance from the Yamamoto chart).
                                      28

-------
                     SECTION  4.  DISCUSSION
     Analysis of the data suggests  that for  a  nonurban atmosphere,
aerosols can contribute a significant excess of  hemispheric infrared
downward-directed radiation,  particularly in a hazy atmosphere.  The
results of averaging all  six  computational schemes together indicated
that 0.013 ly/min is produced for  each 10th  of turbidity, and 0.0125
ly/min is produced for an aerosol  concentration  equivalent to a nephelo-
meter-indicated extinction coefficient of 0.1  km"  .   For a typical
summer day with 0.250 turbidity  and 0.2 km"  extinction coefficient, the
excess irradiance suggested is 0.033 and 0.025,  respectively, or approxim-
ately 0.03 ly/min on the average.   During several  summer days in St. Louis,
Missouri, rviciaun ana Mowers (1974) measured urban-rural differences  in
turbidity of about 0.05,  which corresponds to  a  0.007 ly/min urban-rural
difference in down-welling infrared irradiance.   In Los Angeles, however,
with considerably higher urban pollution concentrations, they measured
turbidity differences through the  lowest 1700  meters  of the atmosphere
of more than 0.2, which corresponds to 0.026 ly/min of excess irradiance.
     For the significance of  these results to  be interpreted in terms
of surface energy budgets, they  have to be put in  proper perspective.
During midday in summer,  with cloudless conditions in the central U.S.,
for example, the incident solar  irradiance would approximate 1.3 ly/min.
Obviously, an excess infrared irradiance of  up to  0.03 ly/min would have
minor significance during these  hours.  During nighttime, however, a much
different picture emerges.  Then,  with no solar  component, the net radiative
                                   29

-------
flux at the surface would typically be about -0.1  ly/min (a negative value
means a net flux away- from the earth.}.  An excess  infrared flux resulting from
aerosols of 0.03, or even 0.007, ly/min now becomes an important factor.
     In terms of a daily radiative energy budget,  the following example
for the central U. S. in July is illuminating.   The average daily
receipt of solar radiation is about 600 ly/day.  With an albedo of
15 percent, 510 ly/day are absorbed by the ground.  A typical net
infrared flux of -0.1 ly/min (144 ly/day} would yield a net all-wave
radiation receipt of 366 ly/day.  The data from this study suggest that
an aerosol loading equivalent to a turbidity of 0.2 would cause an
additional downward infrared flux of 37 ly/day.  A turbidity of 0.05
corresponds to 9 ly/day.  Thus, for a clean atmosphere or small urban-
rural turbidity difference, on a daily basis, the excess infrared flux
would have minor significance.  In contrast, higher turbidities, as in
the example above, could alter the net radiative flux by some 10 percent.
     The data presented herein showed a statistically significant
correlation between relative humidity and atmospheric aerosol concentra-
tions as indicated by nephelometer and sunphotometer.  This finding presumably
reflected the effect of humidity on aerosol growth, especially at higher
Humidities.  The data also showed a significant correlation between
observed-calculated HIDR and relative humidity.  A nonlinear dependence
was suggested by the results.  Large scatter and little trend were
evident at humidities less than 60 percent; at higher humidities, the
excess HIDR increased noticeably as a function of humidity.  Thus, relative
numidity could also be  used as an indicator of excess infrared irradiance.
                                     30

-------
It is possible, however, that the equations used to calculate the HIDR
underestimate the flux at high humidities and the derived relation:;
should be checked further during these conditions.
     In summary, the results of this study do indicate that atmospheric
aerosols can measurably influence downward-directed infrared radiation.
Thus, their infrared radiative effects should be included in certain
surface energy budget studies, especially in areas with high aerosol
concentrations.
                                     31

-------
                    5.   LIST OF  REFERENCES
Angstrom,  A.,  1929;   On the atmospheric  transmission of sun radiation
   and on  dust in the air.   Geograph.  Ann., Vol. 11, 156-166.

Atwater, M. A., 1966:  Comparison  of  numerical methods for computing
   radiative temperature changes in the  atmospheric boundary  layer.
   J. Appl. Meteor.,  Vol.  5, No. 6, 824-831.

Brunt, D., 1932:   Notes on  radiation  in  the atmosphere.  Quart. J. Roy.
   Meteor. Soc.,  Vol. 58,  No.  247, 389-420.

Charlson,  R. J.,  Ahlquist,  N.  C.,  Selvidge, H.,  and MacCready, P.  B.,
   1969:  Monitoring  of atmospheric aerosol parameters with the
   integrating nephelometer.  J. APCA.,  Vol. 19,  No. 12, 937-942.

Covert, D. S., Charlson, R. J.,  and Ahlquist,  N.  C., 1972:  A study of
   the relationship  of chemical  composition and  humidity to light
   scattering by aerosols.   J. Appl.  Meteor.,  Vol. 11, No. 6, 968-976.

Davis, P.  A.,  and W.  Viezee, 1964:  A model for  computing infrared
   transmission through atmospheric water vapor  and carbon dioxide.
   J. Geophys. Res.,  Vol.  69,  No.  18, 3785-3794.

Deirmendjian, D., 1969:  Electromagnetic Scattering on Spherical
   Polydispersions.   Elsevier Publ. Co., N. Y.

Elsasser,  U. M.,  1942:  Heat Transfer by Infrared Radiation in the
   Atmosphere.  Harvard Meteor.  Studies, No. 6,  Harvard Univ. Press,
   Cambridge, Mass.

Flanigan,  D. R.,  and DeLong, H.  P.,  1970:  Spectral absorption
   characteristics of the major  components of  dust clouds.  Edgewood
   Arsenal technical  report 4430,  Dept.  of the Army, Edgewood Arsenal,
   Maryland.   (AD 712989).

Flowers, E. C., McCormick, R.  A.,  and Kurfis,  K.  R., 1969:  Atmospheric
   turbidity over the United States,  1961-1966.   J. Appl. Meteor., Vol.  8,
   No. 6,  955-962.

Flowers, E. C., and Viebrock, H.  J.,  1965: Solar radiation:  an  anomalous
   decrease of direct solar radiation.  Science, Vol.  148, No. 3669,  493-494.

Funk, J. P., 1959:   Improved polythene-shielded  net  radiometer.   J. Sci.
   Instruments, Vol.  36, 267-270.
                                    32

-------
Funk, J. P., 1961:  A note on the long-wave calibration of convectively
   shielded net radiometers.  Arch. Meteor. Geoph. Biokl., Ser.  B,
   Vol. 11, No. 1, 70-74.

Geiger, R., 1965:  The climate near the ground.  Harvard Univ. Press,
   Cambridge, Mass.

Hovis, to. A., Blaine, L. R., and Callahan, W. R., 1968:  Infrared
   aircraft spectra over desert terrain 8.5 p to 16 p.  Appl.  Optics,
   Vol. 7, No. 6, 1137-1140.

Idso, S. B., 1972a:  Radiation fluxes during a dust storm.  Weather,
   Vol. 27, No. 5, 204-208.

Idso, S. B., 1972b:  Solar radiation measurements augment air  pollution
   studies.  J. APCA., Vol. 22, No. 5, 364-368.

Idso, S. B., 1973:  Thermal radiation from a tropospheric dust suspen-
   sion.  Nature, Vol. 241, No. 5390, 448-449.

Idso, S. B., and Jackson, R. D., 1969:  Thermal radiation from the
   atmosphere.  J. Geophys. Res., Vol. 74, No. 23, 5397-5403.

Junge, C., 1955:  The size distribution and aging of natural aerosols
   as determined from electrical or optical data on the atmosphere.
   J. Meteor., Vol. 12, No. 1, 13-25.

Kondratyev, K. Ya., 1969:  Radiation in the Atmosphere.  Academic Press,
   New York.

Lai, M., 1973:  Terrestrial radiation balance in the atmosphere  over
   NW-India.  Arch. Meteor. Geoph. Biokl., Ser. B, Vol. 21, No.  2-3,
   233-242.

Latimer, J. R., 1963:  The accuracy of total radiometers.  Symposium on
   the heat exchange at snow and ice surfaces, 26 October 1962.   National
   Research Council of Canada, Associate Committee on Soil and Snow
   Mechanics, Tech. Memo, No.  78.

McCormick, R. A., and Ludwig,  J. H., 1967:  Climate modification by
   atmospheric aerosols.  Science, Vol. 156, No. 3780, 1358-1359.

Mitchell, J. M., 1971:  The effect of atmospheric aerosols on  climate
   with special reference to temperature near the Earth's surface.
   J. Appl. Meteor.,  Vol. 10,  No. 4, 703-714.

Morgan, D. L., Pruitt, W. 0.,  and Lourence, F. J., 1971:  Estimations
   of atmospheric radiation.  J. Appl. Meteor., Vol. 10, No. 3,  463-468.
                                    33

-------
Neumann,  J., 1972:   Radiation absorption  by  droplets  of  sulfuric acid
   water  solutions  and by ammonium sulfate particles.  Department of
   Atmospheric Sciences.   The Hebrew University  of  Jerusalem,  Israel.

Oke, T. R., and Fuggle, R. F., 1972:  Comparison of urban/rural counter
   and net radiation at night.  Bound.-Layer Meteor.,  Vol.  2,  No. 3,
   290-308.

Paltridge, G.  W., 1970:  Day-time long-wave  radiation from  the sky.
   Quart. J. Roy. Meteor. Soc., Vol. 96,  No. 410, 645-653.

Paltridge, G.  W., and Platt, C. M. R.,  1972:  Absorption and  scatter
   of radiation by an aerosol layer in the free  atmosphere,,   J. Atmos.
   Sci.,  Vol.  30, No. 4,  734-737.

Peterson, J. T., and Bryson, R. A., 1968: The influence of atmospheric
   particulates on the infrared radiation balance of Northwest India.
   Proc.  1st.  National Conf. on Weather Modification, Albany,  New York.

Peterson, J. T., and Flowers, E. C., 1974:   Urban-rural  solar radiation
   and aerosol measurements in St. Louis and Los Angeles.  Preprints,
   A.M.S. Symposium on Atmospheric Diffusion and Air Pollution.  Santa
   Barbara, Calif., 129-132.

Peterson, J. T., and Weinman, J. A., 1969:   Optical properties of quartz
   dust particles at infrared wavelengths.   J. Geophys.  Res., Vol.  74,
   No. 28, 6947-6950.

Platt, C. M. R., 1972:  Airborne infrared radiance measurements (10 to
   12 micron wavelength) off tropical East-Coast Australia.  J. Geophys.
   Res.,  Vol.  77, No. 9, 1597-1609.

Rasool, S. I., and Schneider, S. H., 1971:   Atmospheric  carbon dioxide
   and aerosols:  effects of large increases on global climate.  Sci.,
   Vol. 173, No. 3992, 138-141.

Riches, M. R., 1974:  A study of the effect  of atmospheric  aerosols on
   infrared irradiance at the earth's surface in a non-urban  environment.
   M. S.  Thesis, N. C. State Univ., Raleigh, N.  C., 54 p.

Robinson, G. D., 1950:  Notes on the measurement and estimation of
   atmospheric radiation.  Quart. J. Roy. Meteor. Soc.,  Vol.  76,
   No. 327, 37-51.

Robinson, G. D., 1962:  Absorption of solar radiation by atmospheric
   aerosol, as revealed by measurements at the ground.  Arch. Meteor.
   Geoph. Biokl., Ser. B, Vol. 12, No. 1, 19-40»
                                     34

-------
Rouse, W.  R.,  Noad,  D.,  and McCutcheon,  J.,  1973:   Radiation,  temperature
   and atmosphere emissivities  in  a  polluted urban  atmosphere  at
   Hamilton,  Ontario.   J.  Appl.  Meteor., Vol.  12, No.  5,  798-807.

Sargent, S.  L.,  and  Beckman, W.  A.,  1973:  A numerical model of thermal
   radiation in  a dusty atmosphere.   J.  Atmos.  Sci.,  Vol.  30,  No.  1,
   88-94.

Sasamori,  T.,  1968:   The radiative cooling calculation for application
   to general  circulation  experiments.   J. Appl. Meteor.,  Vol. 7,  No. 5,
   721-729.

Sellers, W.  D.,  1965:   Physical  Climatology,,  Univ.  of Chicago Press,
   Chicago,  Illinois.

Sheppard,  P0  A., 1958:   The effect of pollution on  radiation in the
   atmosphere.  Int.  J.  Air Poll., Vol.  1, No.  1, 31-43.

Staley, D. 0., and Jurica, G. M.,  1972:   Effective  atmospheric emissivity
   under clear skies.   J.  Appl.  Meteor., Vol.  11, No.  2,  349-356.

Swinbank,  W.  C., 1963:   Long-wave  radiation  from clear skies.  Quart.
   J. Roy. Meteor. Soc., Vol. 89,  No. 381, 339-3480

Twitty, J. T., and Weinman, J.  A., 1971:  Radiative properties of
   carbonaceous  aerosols.   J. Appl.  Meteor., Vol. 10,  No.  4, 725-731.

Volz, F. E.,  1972a:   Infrared absorption by  atmospheric aerosol
   substances.  J. Geophys. Res.,  Vol.  77, No.  6, 1017-1031 „

Volz, F. E.,  1972b:   Infrared refractive index of atmospheric  aerosol
   substances.  Appl.  Optics, Vol. 11,  No. 4,  755-759.

winkler, P.,  1973:  The growth  of  atmospheric  particles as a function
   of relative humidity -  an improved concept of mixed nuclei.  J.
   Aerosol Sci., Vol.  4, No. 5,  373-387.

Yamamoto,  G.,  1952:   On a  radiation chart.   The Sci.  Rep.  of the Tokyo
   Univ.,  Ser. 5, Geophy., Vol.  4, No.  1,  9-23.
                                     35

-------
                                TECHNICAL REPORT DATA
                         (Please read Instructions on the reverse before completing]
1. REPORT NO.
  EPA-650/4--75-017
                                                      3. RECIPIENT'S >CCESSIOWNO.
4. TITLE AND SUBTITLE
    Effects  of  Atmospheric Aerosols  on Infrared
    Irradiance at the Earth's  Surface in a
    Nonurban  EnvironrrBnt
5. REPORT DATE
  May 1975
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
                                                      8. PERFORMING ORGANIZATION REPORT NO.
  M.R.  Riches,  J.T. Peterson,  and E.G. Flowers
9. P£RFORMING ORGANIZATION NAME AND ADDRESS
  Meteorology Laboratory
  Environmental Protection Agency
  Research  Triangle Park,  N.C.  27711
10. PROGRAM ELEMENT NO. 1AA009
 ROAP No. 26AAS
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
                                                       13. TYPE OF REPORT AND PERIOD COVERED
  Same
                                                       14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
 This  report describes a  study designed  to  measure hemispheric  infrared
 downward-directed  irradiance at the earth's surface and  ambient aerosol
 concentrations  at  Research Triangle Park,  North  Carolina.  A  Funk-type
 net radiometer (with  a blackened  cavity  on the underside) was  used to
 measure  the  incident  all-wave energy.  From  the value obtained, the observed
 solar radiation was subtracted to  determine the  infrared component.  The ex-
 pected  incident infrared irradiance was calculated from prevailing atmospheric
 conditions. Six  methods were used for  these calculations:  four empirical
 equations based on  surface conditions, the  Yamamoto chart,  and a  radiative
 transfer program  using vertical profiles  of  temperature and  moisture.  The
 observed-minus-calculated downwelling  irradiances were  then compared  to
 concurrent measurements of  turbidity obtained with a Volz sunphotometer,
 nephelometer-indicated atmospheric  extinction coefficient, and relative humidity.
 These  measurements were analyzed by  least-squares  regression to determine
 the extent to  which  incident hemispheric  infrared radiation is affected by  vary-
 ing amounts of atmospheric aerosols and relative humidity.
17.
                             KEY WORDS AND DOCUMENT ANALYSIS
                DESCRIPTORS
                                          b.IDENTIFIERS/OPEN ENDED TERMS
             c. COSATI Field/Group
 Infrared irradiance
 aerosols
 turbidity
 solar radiation
 atmospheric extinction  coefficient
 relative humidity
13. DISTRIBUTION STATEMENT
  Release unlimited
                                           19. SECURITY CLASS (ThisReport)
                                               none
             21. NO. OF PAGES
                 42
                                          20. SECURITY CLASS (This page)
                                               none
                                                                   22. PRICE
EPA Form 2220-1 (9-73)
                                         36

-------
-o


CD
«.
Un
 •

O
                                                                           o>
                                                                        O>  rt)
                                                                        Q.  —
                                                                        Q.  _
                                                                           c

                                                                        in  u.

                                                                        3  °

                                                                        *  2
                                                                        fo  o
                                                                        Q.  U
                                                                        0)  H
                                                                        Q.
                                                                        Q.  (5
                                                                        _  n
                                                                        n  (B

                                                                        51  <
                                                                        ro  
   o  ^
   n  Q)
                                                                                                                  5°
                                                                                                                  2 vi
                                                                                                                   • O
                                                                                                               Jj 0°
                                                                                                               t/» "^ T*
                                                                                                                  rn m

                                                                                                                  O m
                                                                                                                  -* 
-------