LONG-TERM EFFECTS OF AIR POLLUTION—
A SURVEY
June 1970 G. D. Robinson
CEM 4029-400 Principal Investigator
The Center for the Environment and Man, Inc.
250 CONSTITUTION PLAZA • HARTFORD, CONNECTICUT 06103 • 203 277-0133
-------
LONG-TERM EFFECTS OF Am POLLUTION—
A SURVEY
G. D. Robinson
Principal Investigator
June 1970
GEM 4029-400
THE CENTER FOR THE ENVIRONMENT AND MAN, INC.
250 Constitution Plaza Hartford, Connecticut 06103
-------
PREFACE
This report stems from the desire of the National Air Pollution Control Admin-
istration to assess the need for research into the long-term geophysical and biological
effects of air pollution. It takes the form of a survey of the problem areas and a
broad delineation of the lines on which useful research could be pursued.
The report was prepared within The Travelers Research Corporation by
Dr. G. D. Robinson with assistance from Miss Marcella Czarnecki and Mr. Marshall
A. Atwater. Mr. Atwater's substantial contributions to Section 5 include original
work on the radiative effects of aerosol, part of an investigation which I hope will be
published shortly. The subject area of precipitation physics was dealt with, under a
sub-contract, by a team at Meteorology Research, Inc., Altadena, California, headed
by Dr. Theodore B. Smith. Their report has been condensed and paraphrased in the
main body of this document, which will have a wide distribution. Their full report
is an appendix to the document submitted to the sponsor. Dr. William H. Smith, of
the School of Forestry, Yale University, acted as consultant on biological matters.
His advice is quoted and paraphrased in the appropriate sections and his reports are
attached as appendices for the sponsor.
Many other colleagues have been questioned and c onsulted informally during
preparation of the report, and some of them may recognize their unacknowledged
contributions (though appropriate reference has been made when original scientific
work is concerned).
The authors and sponsor hope that readers of this document will not hesitate to
offer critical comment.
ii
-------
TABLE OF CONTENTS
Section Title
1.0 SCOPE OF THE STUDY 1
1.1 Geophysical Effects 1
1.2 Biological Effects 1
1.3 Areas of Research 1
2.0 THE MAJOR POLLUTANTS: THEIR EMISSION AND ROLE IN THE 3
ATMOSPHERE
2.1 Carbon Dioxide 3
2.2 Particulate Content of the Atmosphere 3
2.3 Carbon Monoxide 5
2.4 Sulphur Dioxide 6
2.5 Nitrogen Oxides 7
2.6 Ozone 7
2.7 Water and Heat as Air Pollutants 8
3.0 CLIMATE, CLIMATIC CHANGE, AND THEORIES OF CLIMATE 10
3.1 Climate and Climate Change 10
3.2 Simple Climatic Change Models 12
3.3 Detailed Climatic Models 14
3.4 Future Development of Climatic Models 16
4.0 TRANSPORT OF POLLUTANTS IN THE CONTEXT OF LONG-TERM 17
EFFECTS
5.0 POLLUTANTS AND RADIATIVE PROCESSES 19
5.1 Radiative Effects of CO 19
5.2 Radiative Effects of Atmospheric Aerosol 21
6.0 POLLUTANTS AND CLOUD CONDENSATION PROCESSES 25
6.1 General Nature of the Nucleation Process 25
6.2 Nature and Origins of Cloud Nuclei 26
6.3 Pollution and Nucleation—Condensation Nuclei 26
6.4 Pollution and Nucleation—Ice Nuclei 26
6.5 Observational Evidence of Precipitation Changes 28
6.5.1 Local Effects 28
6.5.2 Regional Effects 30
7.0 THE BIOSPHERE AND LONG-TERM EFFECTS OF AIR POLLUTION 32
8.0 MONITORING 34
8.1 Rationale of Monitoring Long-term and Large-scale Phenomena 34
8.2 Artificial Earth Satellites and the Monitoring of Long-term Trends 35
9.0 CONCLUSIONS—AREAS OF UNCERTAINTY 37
9.1 The Major Pollutants 37
9.1.1 Carbon Dioxide—Concentration and Life Cycle 37
2.1.2 Aerosol—Concentration, Nature and Life Cycle 37
iii
-------
Section Title Pagt
9.1.3 Sulphur Dioxide—Transport and Life Cycle 37
9.1.4 Nitrogen Oxides—Transport and Life Cycle 37
9.1.5 HO Concentration 38
9.1.6 Carbon Monoxide—Concentration, Transport and Life Cycle 38
9.2 Long-term Climatic Effects 38
9.3 Other Long-term Geophysical Effects 38
9.4 Solar Radiation and Radiative Transfer (Other than as a Facet of 38
Climatic Models)
9.5 Nucleation and Precipitation 39
9.5.1 Nuclei Concentrations 39
9.5.2 Observational Studies 39
9.5.3 Model Studies 40
9.5.4 Statistical Studies 40
9.5.5 Laboratory Studies 40
6.6 Applications of Artificial Earth Satellites 40
10.0 PROJECTS FOR EARLY ATTENTION 41
11.0 REFERENCES 42
iv
-------
1.0 SCOPE OF THE STUDY
We cannot begin the task of formulating a research program into the long-term
geophysical and biological effects of air pollution with a clear and generally acceptable
specification of what kind of effect we are concerned with. There might be less need
for a research program if we could. The subject matter of such a program can
perhaps better be indicated at the present time by example than by an attempt at
formal definition.
1.1 Geophysical Effects
If an effect stemming from localized pollution emissions is global in nature, we
are clearly concerned with it. The obvious example is the global climatic (geophysical)
change which might follow an increase in the atmospheric CC>2 content. But there
are other more direct and less widespread climatic or geophysical effects of pollu-
tion which are long-term. Examples are the reduction of sunlight in major cities
and increase of rainfall downwind of certain industrial complexes. These "regional"
and "localized" long-term effects are included in the scope of the study, both because
of their immediate impact on public welfare and because we cannot, in the present
state of knowledge, be quite sure that some of them do not have a reaction on global
conditions.
1.2 Biological Effects
The effects of air pollutants on plants and animals may be classified into three
rather broad categories. The first group contains those effects which result from
relatively short-term exposures to high levels of pollutants. The consequences of
these exposures are commonly termed acute effects. The second group encompasses
those effects which occur as a result of exposure to comparatively low levels of
pollutants over more extended time periods. These phenomena are typically labelled
chronic effects. The third category of effect includes those abnormal ecological and
physiological alterations in living systems which may accrue after long-term
coexistence of ecosystems and air contaminants. For lack of a better term, we might
classify these latter abnormalities long-term effects. Long-term effects differ from
chronic effects in several regards. In the case of chronic effects, the time horizon
may be measured in years whereas the long-term effects may have a time horizon
of decades, centuries or greater. In the instance of gaseous air pollutants, chronic
effects are presumably caused by the gas or gases acting directly on the plant or
animal. With long-term effects, however, the significant agent may be a by-product
of the gas, rather than the gas itself (e.g., nitrate in the case of oxides of nitrogen)
and it may influence the plant or animal indirectly rather than directly. Finally,
and perhaps more importantly, our knowledge concerning the ramifications of the
influences of air pollutants are greatest in the instance of accute effects, much less
in the case of chronic effects, and almost nil in regard to long-term effects.
1.3 Areas of Research
The study of air pollution involves study of the emission of pollutants, their
transport and transformation in the atmosphere, their deposition on land and ocean,
their interaction with the biosphere, and the ways in which they modify the physics
-------
and chemistry of the unpolluted atmosphere. All these study areas are involved
when we investigate long-term effects. We are also concerned with monitoring,
which is taken to mean the purposeful, controlled, continuing observation on a global
or local scale of a pollutant, or of an established or suspected effect of a pollutant.
It is convenient at this point to stress the need for continuing study of likely
future emissions, in the light of population trends, changing patterns of life, economic
and natural resources (for example, availability of fossil fuel), available technology,
and available instruments for abatement on an international scale. Actual and pro-
jected emission data will at some point be needed as an input into quantitative research
on long-term effects. Equally, since there is a possibility (and it cannot be described
at this time, as more than a possibility) that physical, chemical, mathematical or
biological studies might at some stage indicate disastrous consequences if emission
of some pollutant were to continue, these studies should be paralleled by technological-
economic studies on if and how the world's increasing energy conversion needs could
be met with varying constraints on pollution emissions. No further reference will be
made to this subject.
-------
2.0 THE MAJOR POLLUTANTS: THEIR EMISSION AND ROLE IN THE ATMOSPHERE
A recent report by E. Robinson and E.G. Robbins [1] gives an excellent summary
of the incidence, chemistry and physics of the major pollutant gases. This is the main
source of the following brief summary of some of the facts which seem particularly
relevant in planning an investigation of long-term effects.
2.1 Carbon Dioxide
This is not the place to discuss in depth the literature on the carbon dioxide
content of the atmosphere. It has rather cursory treatment in Robinson and Robbin's
report. The comprehensive work of Bolin and Keeling [2] has been updated in a
publication by Bolin and Bischof which became available after completion of the text
of this report. The Bolin and Keeling paper, with others which are referenced in it,
covers the major factors of interest without perhaps doing full justice to the role of
the biosphere in the CO2 turnover. The broad facts which appear to be established
are that there is an increasing annual production of CO2 from fossil fuel combustion
of about 1 x lO1^ tons per year at present, and that there has been for many years a
steady increase in the atmospheric CO2 concentration which in the mid-1960's was
around 7 x 10~7 per year, accounting for about half the annual production of CO2 by
combustion. There is recent evidence (Machta, personal communication) that this
factor may have fallen to about one-third the annual production. There is an exchange
of CC>2 between atmosphere and biosphere—continental and oceanic—involving a
quantity of CO2 greater than that produced by fossil fuel combustion. There is a very
large oceanic reservoir of dissolved CO2 and CO2 "fixed" as carbonate, with a
chemistry, involving silicates and phosphates as well as carbonates, which has been
studied in some detail but which is not yet able to account quantitatively for the
ocean-atmosphere exchange of CO2 and its geographical variation or for the fate of
that fraction of the CO2 produced by combustion which does not appear to be retained
in the atmosphere.
Estimates of the quantity of CO2 fixed in the biosphere are about 25 percent
of the total atmospheric content. There is a detectable seasonal effect of the biosphere
on the atmospheric content, but nothing is known about secular change.
A recently reported observation (Seiler and Junge [ 3], quoting Georgii) intro-
duces a complication, and if confirmed, may require a re-examination of the mechanism
of the atmospheric transport of the CO2 produced or exchanged at the surface. This
is an apparent change in the CO2 content of about 0.2 percent, "a small but distinct
difference," as the tropopause is crossed—the stratospheric content being lower.
Apart from this, the relatively small systematic temporal and spatial variations in
CO2 content are so satisfactorily explained by Bolin and Keeling's simple diffusive
transport model that we might expect a comprehensive general circulation model to
account for them quantitatively.
2.2 Particulate Content of the Atmosphere
There are recent summaries and bibliographies by Horak [ 4] Shah [ 5] and
Twomey and Wojciechowski [6] dealing with the particulate content of the atmosphere.
-------
A few notes concerning the role of pollutant aerosol are added here. Radiative effects
are discussed in Section 5 and effects on precipitation processes in Section 6.
We cannot entirely separate "aerosol pollution" from "natural aerosol" even by
definition— the New Zealand meteorologist who remarked that he had never seen clean
air north of 40 °S did not necessarily mean unpolluted air. Avoiding pedantic class-
ification, we can recognize three broad types of natural aerosol— volcanic dust, wind
raised dust and water-soluble nuclei. The first two categories are clearly enough
definable but soluble nuclei are produced by all combustion processes, as well as by
seaspray and by chemical reactions in an "unpolluted" atmosphere, and the extent to
which any population is "natural" must always be in some doubt.
There is evidence of a fairly homogeneous population of soluble nuclei in air
near the surface in the world's remaining empty spaces, and growing evidence of a
similar state of affairs in the stratosphere. Some of the evidence is indirect— it
comprises the measurements of atmospheric electrical conductivity and the earth's
potential gradient which were commonly made on scientific expeditions in the early
years of the century. Conductivity decreases, and the potential gradient increases,
with the nucleus and particle content of the air because the particles capture small
ions and reduce ionic mobility. Atmospheric potential gradient, indicating the concen-
tration of condensation nuclei, is an excellent man-detector or rather combustion-
detector. (Radioactive fallout has confused this issue since World War II.) There
are some more direct observations of particulate content of air over the oceans but
realization of the importance of size distribution and the practicability of determining
it readily are developments of the last 20 years, and we may in time find the electrical
observations made by early Antarctic explorers and the cruises of the "Carnegie" a
useful indication of the particulate content of air in the earliest days of the population
explosion. There is evidence in recent NCAR observations in Panama and over the
adjacent sea that the smaller "natural" soluble nuclei, Aitken nuclei of radius <0.1 Jim,
are largely (NH.)2 SO. particles formed by oxidation of F^S or SC^ in the presence
of an excess (around 10 ppb) of free
The larger soluble nuclei act as cloud condensation nuclei (Section 6). They
appear to have a "natural" concentration over the oceans around 100 cm~3, but their
number is much more variable in populated continental regions— from several hundreds
to several thousands per cm^. Nevertheless, according to Squires [7] , man's contribu-
tion to the number of cloud condensation nuclei is only a few percent.
It is very difficult to estimate the life-time of soluble particles in the atmosphere.
Twomey [8] has estimated that air masses moving from continent to ocean achieve the
typical maritime concentration in about three days. A similar time seems to be avail-
able for the processes, whatever they may be, which produce the extreme clarity of
Arctic air. There is no entirely satisfying explanation of this cleansing process.
Wind-raised and wind-blown dust, the other particulate component of the tropo-
sphere not the result of industrial processes, may be natural or artificial, in the
sense of resulting from cultivation. Wind-blown material from desert land and semi-
arid or temporarily arid cultivation seems to be more pervasive than was once
suspected: indeed it is suggested in Appendix 1 (on the basis of an unpublished survey)
that only a few percent of the total mass of atmospheric particulate is man-made. It
-------
is often detected by its effects on solar radiation and it appears to be a major source
of ice nuclei. Together with stratospheric fall-out and industrial particulates, it is to
some extent preserved in the annual layers of permanent snowfields so that there is
a possibility of investigating its secular changes in the past. Russian workers claim
to have traced the industrial development of the USSR in the snows of the Caucasus.
A more recent development, potentially of considerable biological importance,
is the long-distance transport of persistent insecticides, either on raised dust or
directly on the base used in crop dusting operations. Much pollutant insecticide is
water-borne but it seems that a proportion, not negligible in some localities, has been
an air pollutant at one stage of its unfortunately long life.
The modification of the earth's albedo by volcanic dust is one of the mechanisms
postulated to cause climatic change. A recent incident—the stratospheric dust cloud
following an eruption on Bali in 1963—has been well documented and is the first such
occurrence which has been extensively investigated quantitatively. Dust content of
the middle stratosphere appears to have temporarily increased by an order of magni-
tude, with a half-life of a year or two. The extent to which the sulphate aerosol
discovered in the stratosphere by Junge is of volcanic origin is not clear. This aerosol,
which appears to have a maximum concentration in the height range 15 km to 25 km
has been directly sampled and chemically analyzed and has been detected optically by
several techniques—searchlights, lasers, and scattering from the solar beam at
twilight. There is as yet no convincing explanation for this concentration, and although
it seems unlikely, pollutant SC>2 may have some part in its formation. For this reason,
it seems advisable that it should be extensively observed for a number of years to
determine any secular trend.
2.3 Carbon Monoxide
There is no reason to expect CO to produce any direct geophysical changes at
the volume concentration of about 10~? which appears to be its present atmospheric
level in regions remote from pollution sources. Furthermore, there is no evidence
that this concentration is increasing. Atwater has confirmed that at this concentration
possible radiative effects are far less than our uncertainty of the radiative effects of
H2O, CO2, cloud and aerosol. CO cannot be ignored in the context of long-term changes,
if only because we do not know how it is removed from the atmosphere, or whether or
not there is a natural source comparable in magnitude with the very considerable
pollution source. Robinson and Robbins [1] summarize the data on pollutant emission
and distribution in the atmosphere—we see from these data that CO production by
combustion processes, about 2 x 10^ tons per year, is 1 to 2 percent of CO2 production
so that if CO is removed from the atmosphere by a process leading ultimately to its
oxidation to CO2, it will add negligibly to the CO2 pollution load.
There is a possibility of biological effects connected with CO. It has low solu-
bility, and there are some reports of saturated or even super-saturated ocean water
and large concentrations associated with certain aquatic plants, and there has been
speculation on the possibility of biological sources, as well as of biological sinks of
the gas.
-------
Recent observations (Seller and Junge [3]) and theories (Hesstvedt [9]) (Pressman
and Warneck[10]) may, if confirmed, remove at least some of the uncertainties. The
observations, made on commercial aircraft, indicate a sudden decrease of CO content
on passing from the troposphere to the stratosphere; the theory explains this as the
result of a series of photochemical reactions in which CO is involved in the Oq-I^O
photochemistry in the lower stratosphere and ultimately oxidized to CO2 with destruc-
tion of Og. The theory has only been briefly reported and there are some puzzling
features which may be resolved in a fuller publication. If the theory is quantitatively
sound detailed observations of CO concentrations in the region of the troposphere
will be of considerable meteorological interest, but not in the context of long-term
geophysical change. (If CO were removed by a different photochemical oxidation in
the lower ionosphere there might be cause for concern, since in this region photo-
chemistry and ionization are interrelated.)
Whilst the fate of CO remains unexplained, it is a reasonable subject for research
in the context of long-term effects of pollution, if only because of the magnitude of CO
emissions and the social and economic implications of substantial control. At present
it seems reasonable to wait for clarification of the suggested photochemical removal
process, though this may well require future measurements in the tropopause region.
2.4 Sulphur Dioxide
There is no reason to expect geophysical influence of pollutant SO2 in its gaseous
form. It now seems well established that the volume concentration of 862 in the upper
atmosphere and over the oceans and remote land areas is as low as 2 x 10~*0
(Robinson and Robbins [1]); Pate (private communication). At this concentration its
radiative influence is negligible. As with CO, interest from the point of view of long-
term effects centers on the methods by which it is removed from the atmosphere, but
unlike the situation in respect of CO, the difficulty is not to find plausible removal
mechanisms but to decide which among many possibilities are significant. From the
geophysical point of view, the interest is in the formation and persistence of sulphate
aerosol; from the biological point of view, in the accumulation in soil and surface
waters of sulphuric and sulphurous acids and their salts following rain and fallout.
£}
SO2 is a short-lived atmospheric constituent. Pollutant production of 1.5 x 10
tons per year and a concentration of 2 x 10"^" require a mean residence time (as 802)
of about 2.5 days. Indirect estimates of the life-time of SO2 in highly polluted regions
are much shorter, ranging from Meetham's 12 hours (southern England 1950), to a
half-life of 1 to 3 hours estimated for the State of Connecticut (Hilst [11]). There
have been reports of half-lives as short as 20 minutes observed by European investigatore,
but details are not yet generally available. There is little doubt that variations in
pollutant aerosol concentration have some influence on the variation of these figures,
which suggest that acidic rain-out and fallout could be a problem at distances varying
from a few tens to a few hundreds of miles downwind from major pollution sources
(i.e., about 24 hours travel).
Persistent geophysical effects are more likely to be associated with the more
persistent sulphate aerosol than with the gas. There is some evidence that over some
seas and tropical land areas, the condensation nuclei are largely (NH^SO, associated
-------
with a free NH3 content of about 10 ppb (Pate, personal communication), and Junge
identified (NH4)2SO4 as an important constituent of the stratospheric aerosol. It is
not clear to what extent pollutant SO2 contributes to the "background" 0.2 ppb of SC>2
in remote areas, or to the formation of the 864 ion in these areas and in the upper
atmosphere. There is a sufficient biological source of H2S, and plausible oxidation
mechanisms, to account for current levels of the "background" sulphate aerosol without
the intervention of pollutant 862. The residence time of sulphur in the form of sulphate
aerosol is currently estimated as being of order tens of days in the troposphere, and
hundreds of days in the stratosphere. A question of major geophysical importance in
the very long term is whether there is a mechanism for net transport of pollutant SO2,
even in very small proportion of the total output, into the stratosphere. Alternatively,
or in addition, is there a mechanism for transport of sulphate aerosol from troposphere
to stratosphere in amounts even very slightly greater than the loss by fallout or mixing?
2.5 Nitrogen Oxides
N2O is a long-lived, fully-mixed constituted of the atmosphere. It is not sig-
nificantly an industrial pollutant. NO is a combustion product, particularly of com-
bustion at high temperatures, and is readily oxidized in the atmosphere to NO2 by
reaction with Og—a reaction so fast that even at the low (order 10~8) concentrations
of Og in the lower atmosphere the NO2 concentration is significant. NO2 is radiatively
important—it absorbs solar radiation in the visible (with photochemical decomposition).
It may, in fact, be responsible for a considerable part of the absorption of solar radia-
tion in urban atmospheres elsewhere attributed to aerosol (Robinson [12]). It is a
key consistituent of photochemical smog. Because of its high reactivity, it does not
seem to present problems on the global scale, but it could have significance in chronic
and long-term biological effects in the vicinity of large cities.
Photochemical reactions involving NO are of prime importance in the very high
atmosphere, and touch significantly on human activity (communications) in their
control of ionization in the lower region of the ionosphere. Again, because of high
reactivity, it seems unlikely that pollutant NO or NO« from the surface could be trans-
ported to the ionosphere.
2.6 Ozone
Ozone is a constituent of the unpolluted atmosphere, being formed and destroyed
photochemically. It is a highly reactive gas, particularly liable to destruction on
surfaces. It is considered that the normal tropospheric concentration, of order 10 ,
represents an equilibrium between transport from a predominantly stratospheric
source and destruction at the surface. One very fast homogeneous reaction
O3 + NO -* O2 + NO2 may be important even in air which might otherwise be considered
unpolluted.
Ozone is also a pollutant of major importance and is unusual in being a secondary
pollutant formed in a photochemically initiated chain reaction—the "photochemical
smog" process. In smog conditions ozone concentrations may be as high as 5 x 10~7.
Smog does not seem to be a significant factor on the global scale, but it certainly
affects city climates and appears to be responsible for widespread chronic and acute
-------
damage to plants in the areas which it affects, so that the possibility of secondary
"long-term" biological effects must be considered. The associated aerosol—poly-
merized organic nitrates ?—appears also to be removed locally, but the possibility
that a very small proportion of it is long-lived and mixed on the global scale cannot
be dismissed.
Speculations concerning artificial climate modification have included some on
possible manipulation of the stratospheric ozone layer—"punching a hole in the ozone"
is a phrase which has been used. Since the layer is a most effective u-v filter, its
removal would be biologically disastrous. The proposed mechanism is by introduction
of a substance very reactive with Og (e.g., NO). Published discussion follows the
consequences of such an introduction on the stratospheric photochemistry only to the
first step of the chain and it is difficult to see how anything more than a change in the
distribution of the Oo with height could result. Normal pollutant materials to be
considered in this context are sulphate and other aerosol, and oxides of nitrogen, and
the possibility now arises that the development of supersonic transport might intro-
duce both NO and hydrocarbons into the lower section of the ozone layer The first
sign of a "threat" of modification of the Oo distribution might be an increase in the
stratospheric content of these materials.
2.7 Water and Heat as Air Pollutants
"Waste" heat is an inevitable consequence of the utilization of the energy stored
in fossil fuel or the atom, and it dissipates in the atmosphere, surface layers of the
earth, and surface waters. An estimate of the level at which it might induce global
climate change is not difficult to make. The average rate of absorption of solar
energy is about 250 watts m~^ and simple climatic models suggest that variations of
1 percent of this might have serious consequences. This is only a few watts per
square meter but is the equivalent of about half a million 1000 MW generating plants
distributed throughout the world and more than 100 times man's current level of
energy conversion. We, therefore, need not yet consider "heat pollution" as a global
problem. It is, however, already significant locally—the air temperature of large
cities is notably modified, by direct heating, from that of the surrounding country-
side, individual smoke-stacks frequently generate cumulus clouds, and precipitation
patterns may be modified. It is conceivable that, if some current predictions of
population trends are sound, "thermal air pollution" might have more than local
climatic effect through modification of such significant features as the sharp temper-
ature gradient near the eastern seaboard of North America in winter. Current
atmospheric models could be used to investigate these possibilities by introduction
of an artificial surface heat source. From the geophysical point of view, the exact
mode of disposal of the waste heat is probably not very important, but persistent
local climatic and biological effects could be quite different from dry cooling towers,
wet towers, or cooling to bodies of water. "Water pollution" is already producing
unpleasant modification of fog and drizzle frequencies near badly sited wet cooling
towers, and "thermal pollution" of streams is leading to biological changes.
Some concern has been expressed over the introduction of t^O into the high
troposphere and lower stratosphere by aircraft. There have been suggestions of a
significant increase in cirrus cloud amounts in some localities, resulting from aircraft
8
-------
condensation trails and some confirmation of this has recently been found by search
of standard meteorological records (Machta, private communication). It has been
argued that circumstances in which condensation trials persist or grow, and in which
cloud would not otherwise form, are meteorological rarities since they require high
relative humidity and absence of appreciable vertical motion, either up or down.
In the lower stratosphere, where supersonic transport aircraft will operate,
the existing H2O concentration is low—2 x 10". This is almost certainly a level set
by atmospheric dynamics—it corresponds to saturation at the coldest tropospheric
temperature. Residence time of water in the lower stratosphere is probably about
one year. With this residence time computation shows that commercial operation of
one large transport aircraft in the lower stratosphere would increase the water content
by a factor 10 . Some traffic projections have suggested that in 20 years time there
may be several hundred such aircraft in operation so that the HoO concentration could
be changed by several percent. It would, therefore, be prudent to monitor the water
content of the lower stratosphere, and to investigate theoretically the radiative and
other effects of a few percent increase in concentration.
-------
3.0 CLIMATE, CLIMATIC CHANGE. AND THEORIES OF CLIMATE
Considerable publicity has been given to some predictions of major changes in
world climate consequent on continuing emission of pollutants. In this section, the
nature of climatic change, the possible mechanisms of change, and the mathematical
models which are tools for the investigation of climate are examined briefly.
3.1 Climate and Climate Change
When we come to investigate climatic change, we are faced with an initial
difficulty of definition which has far-reaching consequences. The climate of a
locality may be described by statistics of certain weather elements. Let us take air
temperature over central and sountern England, as an example (because observations
have been made there for a long time). Craddock [13] has examined about 100 years
of the instrumental record of daily mean temperature at Kew (London) and Manley's
series of about 250 years of annual mean temperatures of central England, using
numerical filters which isolate the variance connected with various periodicities.
Eliminating diurnal and annual periodicities and their subharmonics, Fig. 1 illustrates
his findings. There is significant variance at the longest period plotted. The 250-year-
long series of observed temperatures is not stationary . If we define climate, as it
has been conventionally defined,by the mean and variance of a 30-year series of
observations, then the climate of central England has been subject to continuous
change for the past 250 years; if we look before the instrumental record we find clear
historical and geological evidence that climate change is a global phenomenon, and
was so before man appeared. Early in the present millenium a pastoral community
was able to maintain itself in southern Greenland. Twenty thousand years ago ice
covered much of Europe and Canada. Climatic change is obviously not necessarily
dependent on man's activity. This is not to say that man cannot change, or indeed
is not changing climate, but we have insufficient statistics of "natural" changes to
allow us to recognize artificial changes on the global or continental scale by statistical
techniques, and it is the e&eence of the problem that we will now never have them.
We, therefore, turn from exclusively statistical to physical methods, though if
climate is statistically defined we cannot completely exclude statistical considerations.
If we understand in detail the dynamical and thermodynamical equations of the atmos-
phere and can solve them, we can investigate the consequences of the sort of perturba-
tion which man's activity might introduce into the initial conditions. We have made
enormous progress in this direction, best exemplified in the work of Smagorinsky's
laboratory (e.g., Miyakoda, et al. [15]), but there is a fundamental difficulty which has
been most clearly exposed by Lorenz2 [ 16]. The behavior of the atmosphere is
continuous line in Fig. 1 is the spectrum of a stationary series with a long
time scale proposed by Charnock and Robinson [14] as an empirical approximation to
many meteorological time-series. It is fitted to Craddock's spectrum at a period of
30 days.
2I first encountered the idea many years ago in an unpublished review by R. C.
Sutcliffe and, no doubt, it has been intuitively recognized by most meteorologists; but
it is, surprisingly, not a prominent feature of meteorological literature.
10
-------
11
— Analysis of Craddock
Spectrum of a stationary series
(Charnock, Robinson)
O Central England temperatures (250 yrs)
X Kew temperatures (100 yrs)
\
-4 -3 -2 -1 0
log frequency (days"*)
Fig. 1. Spectrum of temperature variations.
expressed as a closed set of equations. This set of equations has numerous solutions,
each representing a possible state of the atmosphere. Climate can be mathematically
defined as the statistics of the solutions of these equations. If a unique (stationary)
set of such statistics exists, the system governed by the equations is said to be transi-
tive. The equations governing the state of the atmosphere certainly include non-linear
differential equations, and it is known that the uniqueness of long-term statistics of
solutions of sets of such equations is not assured. Statistically different sets of
solutions might, for example, develop from different initial conditions. If a unique
set does not exist, the system is said to be intransitive. Lorenz introduces the
concept of the "almost-intransitive" system, one for which an infinite set of solutions
exists which is independent of the initial conditions but for which large but finite
sub-sets exist which are very dependent on initial conditions. Lorenz postulates that
the atmosphere may be an almost-intransitive system. The consequences for the
student of climatic change are best stated in his own words—"For one thing, the mere
existence of long-term climatic changes cannot by itself be taken as proof of environ-
mental change; alternative explanations are now available. Finally, what about the
not unlikely possibility that the atmosphere would be almost-intransitive if the environ-
mental influences were constant, while at the same time external environmental
changes actually are taking place ? The effect of these changes will then be harder to
detect, and causative connection will be more difficult to establish. For example, an
environmental change which ought to bring about a 2°C temperature rise might occur
just at the time when the temperature was in the process of falling 2°C as a result of
almost-intransitivity. The environmental change might then go unnoticed simply
because no one would see any reason to look for it."
With these considerations in mind, let us take a superficial look at climatic
trends in the present century (we will need to look a little more closely later). Over
that part of the world where air temperature is comprehensively observed, it is
reasonably well established that temperatures increased over a 30- or 40-year
11
-------
period, ending some time between 1940 and 1955. It is also reasonably well established
that the CC^ content of the atmosphere increased during the same period, by an
amount not inconsistent with the rate of production by consumption of fossil fuels. The
two trends were associated, and computation using the radiative transfer equations
showed that quantitatively the COo increase could explain the temperature rise.
Many scientists did not hesitate to say that it did explain the temperature rise. A
few did not hesitate to extrapolate the trend and predict a man-made climatic change
which would melt the polar ice and drown many of the world's major cities. More
cautious meteorologists recognized the inadvisability of the extrapolation but had to
agree that the argument connecting the two trends was physically sound. But by 1960
it was becoming clear that the temperature was falling, and somewhat later it was
noted that atmospheric turbidity was increasing, not only in the immediate vicinity of
cities but over quite wide areas of Eurasia and North America. The increased
turbidity correlated well, geographically, with the emission of pollutant aerosols, and
the decreasing temperature was tentatively ascribed to an increased reflection to
space of solar radiation by man-made pollution. The argument is plausible but,
basically because of lack of knowledge of certain physical properties of the aerosol,
it cannot be developed as precisely as was that concerning CC>2, and perhaps for this
reason there has been less talk of a man-made ice-age than there was of a man-made
deluge.
The relation of these recent events and the conclusions of Lorenz quoted above
is obvious. Man has detectably changed the constitution of the atmosphere; in one
respect globally, in another at least on a sub-continental scale. At the same time,
there have been minor variations of climate. Our present knowledge of atmospheric
processes suggests that these changes are what would be expected to follow from
man's interference with the atmosphere. But they are also quite compatible with what
we know of the statistics appropriate to an atmosphere of undisturbed constitution
and they are also compatible with the possible behavior of a dynamical system as
complex as the atmosphere. Climate has changed and will change, and man may never
know to what extent he has contributed to or inhibited the change. Recognition of this
has quite important consequences in the planning of research on the climatic effects
of pollution.
3.2 Simple Climatic Change Models
There have been some attempts to estimate climatic effects of changes in the
input of solar radiation (either by a change of solar constant or of global albedo) by
very simple methods. They must be considered, if only because of the nature of
some of the conclusions which have been drawn from them. The prototype of such
models appears to be that of Sawyer [17 and 18] who treated the problem "very
crudely as though the atmosphere consisted of two blocks of uniform temperature
between which heat was transferred at a rate depending on the temperature difference."
On this model a one percent decrease in absorbed solar radiation caused a decrease
of temperature in the equatorial section of about 0.75C and in the polar section of
about 0.6C. In a second application of the model, Sawyer reduced the mean annual
solar input by one percent in regions poleward of 50° latitude and computed a tempera-
ture change in this sector of about 0.2C.
12
-------
This model is illustrated in Fig. 2 in which have been inserted values of radia-
tive input consistent with the most modern estimates of the solar constant, the global
albedo as measured by satellites, the radiative temperatures measured by satellites,
and the estimates of heat-transfer by ocean and atmosphere across latitude 50°. This
can be used as an educational toy for simple numerical experiments. The basic
assumption of the models under discussion is the relation between advective transport
and temperature difference. The toy can be used to illustrate the difficulty of
preserving a balance for arbitrary variations of the heat input given a simple, e.g.,
linear, relation between heat transport and temperature difference. It gives some
appreciation of the potentialities and pitfalls of Sawyer's model and its elaborations.
Solar Input
—1 —22
Albedo cal. year x 10
0.53
0.275
7.5
38.5
Air-ocean transport
-1 -22
cal. year x 10
Radiative output
-1 -22
Radiative temp. cal. year x 10
246K
256K
10
36
50°N
21 —1 —1
Transfer coefficient 2.5 x 10 cal. year deg.
Whole earth albedo
Suggested exercise
0.30
mean radiative temp. 252.5K
Hold solar constant and transfer coefficient fixed.
Specify a relation (or relations) between albedo and
radiative temperature. Arbitrarily change one albedo.
Proceed by trial and error to a new balance.
Fig. 2. Basis of the simplest climate model.
Two of the elaborations, described by Sellers [ 19] and by Rakipova [ 20] *, are
directly applicable to the problem of the climatic consequence of a change of albedo.
The simpler model (Sellers) carries only sea-level temperature as an (independent)
indicator of climate. The type of conclusion which he draws is typified by: "If all
other variables are held constant, a decrease in the solar constant by about 2 percent
would be sufficient to create another ice age with the ice-caps extending equatorward
to 50°." Rakipova's model is more complex in detail (though not in principle) than
Sellers', and carries temperature at various heights in the atmosphere. One of her
conclusions is that a one percent decrease in the solar input would result in a tempera-
ture decrease varying between 0.3C at the equator to 1.4C at the poles. The
apparent difference between this and Sellers' more startling conclusion is caused by
a difference in the treatment of the relation between temperature and albedo. This
relation is a key factor in the models. According to Sellers, the albedo in all latitudes
* Rakipova's work exemplified the approach to this problem by the Leningrad
group under M. I. Budyko.
13
-------
would be expected to increase as temperature decreases. This is a destabilizing
mechanism—a drop in temperature reduces the heat input. The relation proposed is
empirical; the physical cause appears to be the high albedo of a snow- and ice-covered
surface. The weakness of the relation is that it does not seem adequate to cover the
effect of the cloud. Lower temperatures lead to a lower atmospheric water content
which might mean less cloud and lower albedo. This type of control of albedo produces
a stabilizing mechanism, lower temperatures leading to greater heat input, and the
possibility that the planet earth is behaving as an inefficiently stirred thermostat.
In Rakipova's computation, cited above, the albedo is not changed, but in the text of
her paper she quotes a "warm season" global albedo of 0.410 and a "cold season"
global albedo of 0.384. Both figures may be too high, but the overall relation proposed
is a stabilizing relation.
The temperature-albedo relation is the sensitive problem in the Sawyer-Sellers-
Rakipova type models, and there is some prospect of an empirical solution to it when
many more data have been collected from meteorological satellites, but the models
have also a major difficulty in principle. Raklpova states this clearly early in her
text and does not allow it to inhibit her further. She states, "We are analyzing a purely
zonal situation for which (the mean meridional and vertical velocities ) =0." One
cannot accept this "purely zonal situation" as a reasonable model for investigation of
the effects of heat transfer between latitudes. Sellers apparently shares this view but
does not appear to circumvent the problem. He states "the inclusion of the mean
meridional motion is necessary in order to avoid having to deal with negative diffusivi-
ties," but the mean meridional motions he postulates do not satisfy mass continuity
and he does not discuss the transports of potential energy consequent on realistic mean
meridional circulations.
3.3 Detailed Climatic Models
The basis of modern meteorology is mathematical simulation of the atmosphere
and the earth's surface as a nearly-closed thermodynamic system, and the current
trend in such simulation is toward increasingly detailed realism, at the expense of
computation loads which stretch the limits of foreseen technology. This type of model
was initially conceived as a tool for objective weather forecasting and is often justified
economically in this context, but its value in climatology was recognized at an early
stage. Reservations concerning the validity of this approach to the detailed forecasting
problem (Robinson [21], Lorenz [22]), even if sound, are not related to the use of the
models as generators of climatic statistics. For the present purpose, a minimal
description of the methods must suffice. Several groups are engaged in the develop-
ment and use of comprehensive atmospheric models. Probably the most advanced and
best documented model is that of Smagorinsky's group at Princeton. The following
comments, superficial in the sense that they include only aspects relevant to the
present task, are based'on a perhaps partial understanding of this model and of that
described by Bushby and Timpson [23].
The mathematical basis is the conservation equations for momentum, energy
and matter—partial differential equations, some of which are non-linear—the integral
radiative transfer equation, the perfect gas equation of state and the thermodynamic
equations describing the phase changes of H2O. Mathematically, solution is only
14
-------
possible by numerical finite difference methods; physically, initial and boundary
conditions can only be supplied as averages in time and space. The magnitude of the
computation load and the resolution of observations set lower limits to feasible and
available spatial definition; when the space scale has been chosen an upper limit to
the finite time step which can be used in solution of the momentum equations follows.
A spatial resolution greater than about 500 km is probably Insufficient in a model
which is expected, for example, to indicate changes in the ice and snow covered areas
of the globe. The corresponding maximum allowable time-step is of order 10 minutes.
To seek to use such a model in straightforward continuous simulation to examine a
period even as short as the 250 years for which we have some instrumental record,
cannot at present be contemplated. Perhaps this will not still be true 50 years from
now but we are concerned with the next few years. In this period the use of models to
investigate climate modification will probably be by changing initial and boundary
conditions—surface albedo, solar input, etc., or coefficients in the radiative transfer
equations (i.e., atmospheric composition) and following the consequences to a quasi-
steady state (or to breakdown of the computational scheme).
For the investigation of the effects of pollution, current comprehensive models
have one very serious defect—they do not develop and transport cloud systems.
Climatological averages of cloud distribution are used in application of the radiative
transfer equations and the water-cycle is handled by assuming immediate precipita-
tion of water in excess of saturation—the models simulate rain and snow but not cloud.
There is no major difficulty of principle in modifying models of large-scale processes
to incorporate generation, transport, and dissolution of cloud, but this would increase
very considerably the complexity of a computation which is already probably the
most extensive ever undertaken on a continuing basis. Lack of this modification is
not of prime importance in the weather prediction application of large-scale models,
but it is a most serious short-coming when they are used to investigate climatic
change, because of the high sensitivity of climatic statistics to albedo, the sensitive
dependence of albedo on cloud amount and type, and the critical relation of cloud
amount and type to atmospheric motion and stability. We see that the major defect
of the existing comprehensive models of the atmosphere, in their application to
study of climate, is the same as that we noted in the simple models—inadequacy in
simulation of the relation of albedo to other parameters of climate.
It is probably safe to assume that on the global and climatic scale, though not
on the local scale, the amount and gross radiative properties of cloud do not depend
on space and time variations in the nucleating properties of the atmosphere. If
this is so, the cloud problem can, at least in principle, be resolved within the logical
framework of the model without further empiricism—the dynamic and thermodynamic
processes generate and dissolve the cloud; the condensation and evaporation react
grossly on atmospheric temperature and motion. The situation in respect of aerosol
is rather different. There are both surface and internal sources of both man-made
and natural aerosols. These aerosols are transported by atmospheric motions
but there is no first-order feedback between aerosol content, and motion—the inter-
action is slow and through the radiative terms. Transport can be handled by the
models—on the small scale this is done in "air pollution models" and on a scale
of hundreds of kilometers, R. J. Murgatroyd has recently applied a detailed fore-
15
-------
casting model (Bushby-Timpson, 10 layer 80 km grid square) to the problem of
three-dimensional transport of pollutants. At the time of writing only abstracts of
Murgatroyd's work have been seen: the outstanding problem may be the handling,
with a tolerable number of layers in the vertical, of the quasi-isentropic nature of the
three dimensional motion and the resultant tendency for locally injected pollutant to
be confined, in a stable atmosphere, to gently-sloping layers of very limited depth.
An approach other than empirical to the specification of a source terms for the
natural aerosol (e.g., raising of surface dust by the model-generated wind) is a
development for the distant future. Man-made sources must always be empirically
specified.
3.4 Future Development of Climatic Models
The brief overview of simple climatic models did not lead us to any very
encouraging conclusions. It is not easy to formulate recommendations for further
work with them because one cannot have real confidence, even qualitatively, in the
results so far announced. On the other hand, there is now general agreement that
modeling of the whole surface-atmosphere complex is essential to an understanding
of climatic change consequent on any local or global modification of the atmosphere or
surface and inactivity will not advance the cause. A possible approach would be use
of the comprehensive global weather models of the type developed by Smagorinsky's
group in something like a "Monte Carlo technique" mode to accumulate statistics on
the climates of atmospheres of differing constitution, but this expensive undertaking is
not likely to commend itself to the very few institutions at present able to contemplate
it. The expense and elaboration are sufficient to make this a very inefficient approach,
so long as the models do not carry cloud cover as a dependent variable.
It would perhaps be useful to continue development of the simple models as
"educational toys"—for example: to attempt to improve the treatment of mean
meridional circulations in Sellers' model and then examine the effects of imposing
different relations between albedo and temperature. This relation is one of the keys
to the modeling of long-term climatic change, and it might be that a new approach
with more parameterization of dynamical aspects and more detailed treatment of
radiation and cloud physics could be explored. The simple models carry parametri-
zation of all processes to the extreme.
16
-------
4.0 TRANSPORT OF POLLUTANTS IN THE CONTEXT OF LONG-TERM EFFECTS
Short- and medium-range transport of pollutants is treated in the context of
urban and regional air pollution models—see for example, the report of a seminar on
Urban Air Pollution Models recently organized by NAPCA [24]. Mixing of a long-
lived pollutant on a global scale is exemplified in the study of CO2 by Bolin and
Keeling [2]; and work of this kind can be, indeed has been refined, but its impact would
be more on general dynamic meteorology than on the study of pollution and its effects.
(The pollutant might be used as a tracer of atmospheric motion and as a lead to the
parameterization of various scales of motion.) Between the local and the global scales,
there is a scale where the transport of pollutants, averaged in the long-term, is of
biological interest and could become of political importance. The problem is that of
deposition of pollutants and the products of their transformations downwind of major
extended sources. Regional or sub-continental pollution models are required for this
investigation, and they are not yet developed and tested. The problem is referred
to in Section 3 where it is suggested that such a model might be constructed from
the more detailed meteorological models now being developed by Weather Services.
The most obvious possibility is to include the continuity equation for a pollutant,
with appropriate boundary and internal source terms, within a "primitive equation"
meteorological model; but difficulties as well as possibilities are obvious and another
approach might be more profitable at least initially. The fact that pollution on this
scale is an international problem is already recognized: international negotations
may at some stage be called for, and it is not too early to try to establish a sound
basis for them. "Specification of the internal source term" takes us back, of course,
to little known areas of air chemistry so there is no reason to expect easy success.
Empirical testing of a model on this scale might well throw some light on the
chemical problem. The requirement is to relate the type of fact summarized in
Fig. 3—the amount of sulphur deposited in precipitation over the United States—to
the natural and artificial sources.
17
-------
00
Fig. 3. Sulfur in precipitation over the U.S.
-2 -1
Unit—gm yr . Isopleths from Ericksson, quoted by Robinson and Bobbins.
-------
5.0 POLLUTANTS AND RADIATIVE PROCESSES
The most obvious way in which a pollutant can affect climate in the long term is
by interfering with the radiative processes in the atmosphere. We have a fairly sound
knowledge of the theory of radiation in the atmosphere and can compute the first stage
in the chain of effects which would follow the addition of a pollutant of known optical
properties. In this section we examine the relevance of this knowledge to our problem
of predicting the long-term effects of pollution. We find major deficiencies in mete-
orological modeling and in our knowledge of the optical properties of aerosol.
5.1 Radiative Effects of COo
CO2 absorbs solar radiation in the near infra-red in regions where water also
absorbs. This is not an important source of atmospheric heating and can be neglected
when the effects of change of CO2 content are considered. Unlike areosol pollution,
CC>2 pollution does not change the heat supply immediately available to the planet.
There is no reason to expect change of atmospheric CC>2 content to affect the mean
temperature of the planet since the outgoing radiation is fully black at some tempera-
ture already existing in the system and the total outgoing energy is controlled by the
solar input. An increase of CC^ should not, on this qualitative argument, change the
average net terrestrial radiation at either the top of the atmosphere or the earth's
surface, and, since an increase of CC>2 could only increase the emissivity of air, a
reduced temperature should result high in the atmosphere, compensated by an increased
temperature near the earth's surface—the "greenhouse effect."
Quantitative attacks on the problem confirm the qualitative deductions, but over
the years there have been varied estimates of the magnitude of the rise of surface
temperature which might accompany an increase in CC^ content. Callendar, before
World War II, and Plass in the early 1950's, computed the effect of CO2 alone, and
Plass concluded that doubling the CO2 content would result in an average surface
temperature increase of 3C to 4C. Kaplan argued that the presence of cloud (in a
fixed amount equal to the present global average) would reduce this figure by about
60 percent. Kondratiev pointed out the importance of considering the H2O absorption
region. Using a fixed water content and ignoring cloud, he computed surface tempera-
ture changes only 15 to 20 percent of those suggested by Plass. All these computa-
tions were of the primary reaction of the atmosphere on the supposition that CO2
content increased and nothing else changed except air temperature. Mbller [26]
was the first to study a secondary effect influencing both solar and terrestrial radia-
tion. (His paper also contained a succinct summary of earlier work.) He pointed out
that increased temperature of the lower atmosphere inevitably implied increased
water content. He found, assuming unchanged relative humidity, that this greatly
affected the computed temperature change and he quotes a rise in surface tempera-
ture of IOC for a doubling of CC>2 content. We are, at this level of complication,
concerned with a destabilizing situation—the "greenhouse effect" induces a rise in
surface temperature, which evaporates more water, which in turn increases the
"greenhouse effect." Mb'ller also, however, pointed out the existence of further
secondary effects, particularly those dependent on cloud. His computations, like all
previous and subsequent computations, held the amount of cloud constant. Mbller
19
-------
attempted a semi-quantitative estimate of the effects of cloud on the COg problem.
He concludes:
"It is not difficult to infer from these numbers that the variation
in the radiation budget from a changed CC>2 concentration can be
compensated for completely without any variation in the surface
temperature when the cloudiness is increased by—1 percent of
its value or—the water vapor content (decreased) by 3 percent of
its value. No meteorologist or climatologist would dare to deter-
mine the mean cloudiness or the mean water vapor content of the
atmosphere with such accuracy, much less can a change of this
order in magnitude be proved or its existence be denied."
The studies which culminated in Mb'ller's work preceded by computation of the down-
ward terrestrial radiation at the earth's surface and related this to the change in
surface temperature by consideration of the surface energy budget. Manabe and
Wetherald [27] extended the scope of the computations to cover the temperature
structure of a whole atmosphere in convective-radiative equilibrium. They considered
all radiating gases, with a fixed "climatological" cloud distribution. They specified
water content by a given relative humidity, and, as did Moller, included the effect of
changed water content on solar radiation. They have, therefore, the same "destabiliz-
ing" condition as Moller, but its effects on surface temperature are tempered by
convective readjustment and their final result for a 100 percent increase in CC>2
content, with (existing) average cloudiness,is a surface temperature increase of 2AC.
Their computations also showed a considerable temperature decrease in the stratos-
phere as we would expect from our initial qualitative agrument.
There are some minor aspects of Manabe and Wetherald's paper which might not
secure universal acceptance, but it is reasonable to suggest that, with the exception
of a rather specialized area mentioned below, their computations are the best that need
be made for the effect of a change of CC>2 content on an atmosphere with specified
constant cloudiness. Refinement of their treatment of both radiation and convection
is possible, but the work would still be open to the (quoted) criticism which Moller
made of his own work. It seems reasonable to argue that an increase in atmospheric
CC>2 content would result in an initial tendency for warming at the surface and cooling
in the high stratosphere but there will be no justification for a forecast of the final
equilibrium temperatures until we have made an order-of-magnitude advance in the
complexity of atmospheric models, to include the distribution (and ideally also the albedo)
of cloud as a variable.
The aspect in which further computation might at present be rewarding concerns
the effect of CC>2 concentration change in the mesophere. At the upper limit of Manabe
and Wetherald's computation (around 40 km or 2.5 mb) doubling the CC>2 content from
its current value leads to a temperature decrease of about IOC. At about this height
computations begin to be complicated by the fact that CO2 is not in local thermodynamic
equilibrium. It is, however, much the most effective radiator in the atmosphere at
these heights and temperatures. The temperature and composition of the air in these
regions are linked through primary photochemical processes and secondary reactions,
some of which are sensitively temperature-dependent. The region of concern spreads
20
-------
through the temperature maximum and into the lower ionosphere where properties
important for radio propagation might be changed. An investigation of the consequences
of increased CC>2 content in this region of the atmosphere would probably be rewarding.
To the first order, it would be acceptable to hold tropospheric conditions constant.
5.2 Radiative Effects of Atmospheric Aerosol
Addition of a purely scattering material to the earth's atmosphere necessarily
increases the earth's albedo. The amount of solar radiation absorbed is reduced, the
re-radiated energy is reduced and the effective emissivity is not changed so the
effective radiative temperature must decrease. The detailed effect on surface tem-
perature is not of course obvious, but it is reasonable to expect a cooling on balance.
If, however, the added material absorbs as well as scatters solar radiation, it may
not increase the planetary albedo; if it absorbs terrestrial radiation it may tend to
increase surface temperatures by the "greenhouse effect," even though the overall
mean planetary temperature is reduced. M. Atwater has made some computations
indicating the magnitude of these effects which are summarized here: details are given
in Appendix 3.
It seemed useful to investigate an extreme but not inconceivable degree of aerosol
pollution, and because of availability of programs a near-surface rather than a stratos-
pheric aerosol layer was considered: a layer 300 m thick containing 106 particles cm"3
distributed as—
n(r) = 5.635 + 1018 r3 exp (- 45.1272r°'25) (r in Mm)
(which has a maximum concentration at radius 0.005 Mm) is consistent with the distri-
butions investigated by Petersen, Paulus and Foley [28]. (It reproduces closely the
distribution shown in Fig. 6 of their paper.) This mode radius is, of course, very
small for heavy near-surface pollution, but the properties of the layer as a whole are
not unrealistic and since the accent is on long-term effects, it is logical to postulate
a persistent aerosol. For a refractive index of v = 1.1 - O.li for solar radiation and
v = l.l - 0.25i for infra-red radiation*, we find for solar radiation-
Absorption coefficient a abs ~ 1.0 km
Scattering coefficient a scat ~ 0.5 km"1
and for terrestrial radiation—
a abs ~ 0.1 km"1
a scat ~ 0.002 km"1
o
These properties correspond to a horizontal visibility of 2.5 km and an Angstrom
turbidity coefficient for the layer of 0.2. The absorption coefficient is considerably
larger (about x 5) than that observed in an extensive haze layer about 600 m
thick over southern England (Robinson [29]) but less than some measured by Waldram
[30] in "heavy industrial haze." On a cloudless summer day in mid-latitudes the
*These refractive indexes were chosen to produce appropriate bulk optical properties
of the aerosol layer—see M. Atwater, Ph.D. Thesis, New York University.
21
-------
added absorption of solar radiation by this layer corresponds to a heating-rate for a
few hours around noon of about 50K per day. The heat concerned is, of course, lost
to the surface: the total effect of the aerosol on the lower atmosphere is to decrease
the availability of solar radiation and over a strongly absorbing surface to decrease
the actual heating of the atmosphere. The computation of extra cooling by thermal
radiation depends greatly on the air and surface temperatures assumed, but a reason-
able figure is 2 to 3 K per day over the layer as a whole, with warming near the
surface. With this particular aerosol over an absorbing surface, climatic effects
depending on long-wave radiation are considerably less than those depending on solar
radiation in most summer latitudes, but the long-wave radiation effects dominate in
winter.
It is obvious that if we are concerned with an aerosol which absorbs but does
not scatter, or with an absorbing and scattering aerosol over a high albedo surface,
the result will be a decreased global albedo. Atwater's computations (Appendix 3)
investigate this effect for a layer 275 m thick with various scattering and absorption
coefficients and various surface albedos, with some indication of the effect of zenith
angle and the ratio of forward to backward scatter. For example, the aerosol layer
considered above increases global albedo when the surface albedo is less than about
0.2, but decreases global albedo for larger surface albedos. This is a case of
extremely high aerosol absorption but the computations show, for example that most
industrial aerosols would reduce global albedo over a snow or ice surface with the
low solar elevations of sub-polar localities.
Measurement of the radiative effects of "pollutant" aerosol must take place
against the background of the "natural" aerosol which we have seen (Section 2.2 and
Appendix 1) to be more prevalent than "pollutant" aerosol over a large part of the
world. Lettau and Lettau [31] , for example, have shown that attenuation (though not
absorption) of direct solar radiation by aerosol at a desert location in Peru is com-
parable with that measured in the suburbs of London, England, before operation there
of the Clean Air Act; and during the recent "Bomex" observations in the western
tropical Atlantic layers of dust, presumably of Saharan origin, were observed to
absorb several percent of the incident solar radiation (A. J. Drummond, personal
communication).
Comparison of observed sky radiation in relatively unpolluted localities with
that computed for molecular scattering suggests that aerosol commonly scatters
downward 5 to 15 percent of the incident solar radiation. In apparently unpolluted
air over the English Channel, both Roach [32] and Robinson [29] observed an
absorption of solar radiation between 10,000 and 20,000 ft about 3 percent in excess of
that expected from gaseous constitutents, with an excess upward scatter of about
1 percent. More data of this type can be expected from the "Bomex" experiment. On
the other hand, some records of diffuse sky radiation taken on the Antarctic continent
can be explained in terms of molecular scattering only.
Light scattering observations—searchlight, laser and twilight—suggest that the
stratospheric particulates in the Junge layer scatter about the same amount of visible
radiation as does the clear air at the same height. Stratospheric scattering of the
total solar radiation appears to be significantly in excess of the Rayleigh value at all
22
-------
heights. Stratospheric scattering is too small an effect in absolute terms to be deduced
with confidence from the surface measurements of solar radiation normally made by
meteorological services, but it certainly appears that some Antarctic air masses have
a lower tropospheric concentration of aerosol than most stratospheric air.
Observations of a secular increase in atmospheric turbidity (a measure of
extinction of solar radiation in excess of that to be expected from a clean atmosphere)
at Washington, D.C. and Davos, Switzerland led McCormick and Ludwig [33] to raise
the whole question of long-term effects of pollutant aerosol. (Washington and Davos are
probably the only localities with really reliable records over a 50-year span: Davos
is an alpine station far removed from major pollution sources). Flowers, McCormick
and Kurfis [34] have examined the results of a network of stations measuring turbidity
and have established the pattern shown in Fig. 4 for non-urban stations within the
U.S. They appear to have established a simple and reliable indicator of total aerosol
content—natural and man-made—valuable as a long-term monitor.
23
-------
MONT. ' r
| N. OAK.
I
Fig. 4. Annual mean of turbidity over the United States excluding urban stations 1961—1966
(Flowers, McCormick and Kurfis).
-------
6.0 POLLUTANTS AND CLOUD CONDENSATION PROCESSES
An understanding of the physical processes involved in nucleation and an apprecia-
tion of the properties of natural nucleus populations is essential to a judgment of the
likelihood of modification of precipitation processes by pollutant aerosol. On the other
hand, there are broad thermodynamic arguments which suggest that conceivable levels
of pollution are not likely to affect climate on a global scale by way of nucleation effects
alone. To avoid an unbalanced accumulation of detail, only a brief overall descrip-
tion of the nucleation process is included in the body of this report: the necessarily
much fuller treatment by Dr. T. B. Smith and his colleagues is in Appendix 1.
There is much more likelihood of detectable, even serious, modification of
precipitation processes on a local and perhaps on a regional scale, and some of
Or. Smith's conclusions on these matters are mentioned here and repeated in Appendix 1.
6.1 General Nature of the Nucleation Process
Certain atmospheric particulates act as cloud condensation nuclei (CCN) or ice
nuclei (IN). CCN are those particles in the atmosphere on which water vapor con-
denses to form cloud droplets. IN are those particles which have the special property
of nucleating the ice phase in clouds either by nucleating supercooled droplets or by
serving as centers upon which ice is deposited directly from the vapor phase.
The concentration in the air of those CCN which are active at the maximum
supersaturation existing in a cloud determines the concentration of cloud droplets.
This is one factor controlling the efficiency with which cloud droplets can grow by
coalescence to form raindrops in a warm cloud. Thus, if the concentration of effec-
tive CCN (and therefore cloud droplets) is high the average size of the droplets will
be small and their growth to raindrop size will be difficult. This is thought to be the
situation for clouds forming in continental interiors. On the other hand, in maritime
air masses the concentration of CCN is quite small, the average size of the cloud
droplets is therefore larger than in continental clouds, and raindrops are produced
more readily. The addition of CCN to a cloud mass therefore increases the time
required for the development of precipitation. On the other hand, if relatively small
numbers of highly efficient CCN (so called giant nuclei) are introduced into a cloud
they may serve as preferential centers for condensation and these droplets may
increase rapidly in size to form raindrops. (This, of course, is the principle behind
the seeding of warm clouds with giant salt particles in order to enhance precipitation.)
In the case of clouds which extend above the 0°C level, the growth of ice
particles by the Bergeron- Findeisen process (distillation from supercooled droplets
to ice particles) provides another mechanism by which precipitation particles may be
formed. The concentration of ice particles is related to the concentration of active
ice nuclei in the cloud. Below a certain critical concentration of ice nuclei, the ice
particles can grow to precipitation size fairly readily. However, if the ice nuclei
exceed this"critical concentration the formation of precipitation may be hindered.
The nucleation of the ice phase in clouds has an additional and important effect,
namely, that it releases a significant quantity of latent heat which increases the
buoyancy of the cloud. Under certain environmental conditions, this can result in the
25
-------
"explosive" growth of the cloud. These ideas have received general confirmation in
experiments in which supercooled clouds are seeded with artificial ice nuclei.
6.2 Nature and Origins of Cloud Nuclei
Since the maximum supersaturations which exist in clouds rarely exceed
1 percent, only the larger (say > 0.05 fim) and generally hygroscopic particles in
the air act as CCN. Typical concentrations of CCN are 102 cm"3, whereas the total
concentration of particles may be of the order 105 cm"3. The surface of the earth
and the ocean are thought to be sources of CCN. Recently, certain industries (e.g.,
paper mills) and artificial and natural fires have been identified as sources of very
effective CCN.
Natural ice nuclei are very rare. Typically, only 1 particle in 10*1 jn ^e
atmosphere is effective as an ice nucleus at a temperature of -10°C. The exact
mode of action of an ice nucleus is still a matter of dispute. The relatively few
studies which have been made of the composition of natural ice nuclei indicate that
silicate minerals from the earth's surface are dominant. Certain industries (e.g.,
steel mills) emit large quantities of ice nuclei into the atmosphere. Measurements
made at three widely separated sites (Hawaii, Alaska, Washington) indicate that
under certain conditions ice nuclei may be advected over distances of thousands of
miles.
6.3 Pollution andNucleation— Condensation Nuclei
It is estimated that globally only a few percent of the CCN are man-made, but
in localized urban areas the number of artificial nuclei may exceed the natural
population. Pollution aerosol may also play a part in the activation or de-activation
of nuclei by coagulation with natural nuclei. Pollutants may produce nuclei by
secondary reactions in the atmosphere.
The effectiveness of additional CCN in modifying precipitation processes
depends in practice more on their size range and that of the natural population to
which they are added than on overall numbers. They change the rate at which the
coagulation process becomes effective in producing rain, not the nature of the process.
Some computations in realistic cases are set out in Appendix 1 where it is concluded
that, "the results can be viewed from the perspective of the overall mechanism
involved in the modification of the coalescence process. It is generally assumed, to
a first approximation, that the dynamics and lifetime of the warm cloud are not
affected by changes in the coalescence growth rate. This means that precipitation
changes can result only from changes in the storage of cloud water, either increased
or decreased. Since the water supply of most of the large clouds exceeds the cloud
storage capacity under natural conditions, it is only the smaller and marginal clouds
which are likely to be affected by the changes in particulate concentrations."
6.4 Pollution and Nucleation—Ice Nuclei
We have seen that natural ice nuclei are not abundant; it also appears that
very few pollutant particulates possess the ice-nucleating property. Details of
26
-------
some measurements are given in Appendix 1 which suggest that urban (specifically
Los Angeles) air may sometimes contain fewer ice nuclei than the surrounding rural
air—presumably because of deactivation by coagulation. On the other hand, large
numbers of IN have been detected in the effluent from steel mills, and some urban
atmospheres (specifically Seattle) have been found to contain an excess of IN. Two
further possible sources of pollutant IN are examined in Appendix 1, where it is
suggested that they are probably not significant. The first, suggested by Schaeffer,
results from reactions between lead compounds emitted in automobile exhaust gases
and iodine vapor, to produce Pbl, which is a highly efficient IN. The second is the
exhaust particulates of jet aircraft.
Direct measurement within the exhaust of a jet engine detected changes in ice
nucleus content which were not large enough to be considered significant, and the
importance of Schaeffer's mechanism has not been confirmed by independent observa-
tion in the atmosphere, perhaps because iodine vapor, even in the minimal concentra-
tions which would be significant, exists only as a rare local pollutant. To quote from
Appendix 1, "The foregoing studies indicate that the effect of air pollution on ice
nuclei concentrations may be quite variable, depending on the type of pollution, con-
centrations, etc. Values may range from the 1000/liter found in the French industrial
regions to changes of less than an order of magnitude to, finally, the Los Angeles
area where there is a tendency for lowest ice nuclei values to be associated with
heavy pollution. Schaefer's comments on iodine and lead reactions in automobile
exhausts can be viewed in the perspective of the measurements made in the two
metropolitan areas of Seattle and Los Angeles. In the one case, area concentrations
may have been increased by a factor of six as a result of automobile and/or industrial
sources. In the other case, with industrial sources more restricted than in most
areas, the ice nuclei counts in pollution tend to be lower than in unpolluted regions.
These comments support the general conclusion that local effects of the iodine and
lead reaction may occur but there appears to be no substantial evidence of a wide-
spread contribution from this source.
Long-term records of ice nuclei variations essentially do not exist. Partly
this results from the modifications and improvements in ice nuclei measurements which
have occurred in the past 15 years. Grant at Colorado State University, however has
measured ice nuclei routinely at Climax, Colorado, for the past 10 years. Fortunately,
these measurements have been consistent in terms of observational technique. Grant
(private communication) reports no definable change in ice nuclei counts over the
10-year period which might be considered as a background trend. It remains possible,
however, that areas such as the eastern sections of the United States or portions of
industrial Europe might be experiencing gradual increases in ice unclei content but
that an efficient natural removal process might serve to make the trend indiscernible
at Climax. The Climax data also fail to show any pronounced influence from polluted
areas to the west such as Los Angeles. This is in keeping with measurements made
in the Los Angeles area itself which fail to show widespread ice nuclei effects, even
in the source region itself."
27
-------
6.5 Observational Evidence of Precipitation Changes
6.5.1 Local Effects
Substantial evidence of precipitation modification is meager but significant.
The two outstanding and oft-quoted examples are an increase in reported rainfall
at La Porte, Indiana (Changnon [35]), and local decreases in Australia (Warner [36]).
Changes in precipitation of 25 to 30 percent have been found in these areas. Hobbs,
Radke, and Shumway [37] provide suggestive data relating precipitation increases
in Washington to the location of major industrial complexes. Increases of over
30 percent were found for the period 1947—1966 compared to 1929—1946. Miller [38]
found apparent increases in precipitation of the order of 15 percent over Long Island
and downwind of New York City. The remaining examples of precipitation effects
have been summarized by Changnon [39] and Peterson [40] and generally show
changes in precipitation of the order of 10 percent or less. It is useful to examine
these cases in the perspective of the preceding sections.
Changnon [ 39] has found that much of the increase in reported precipitation at
La Porte occurs during the warm season and that the number of thunderstorm days
is increased significantly and concludes that midwest urban areas produce significant
increases in convective activity. Principal precipitation effects come from an
increase in the number of days with 0.25 inch of rain or more. The conclusion was
reached that thermal and frictional effects were primarily responsible for the urban
effects in the midwest and that La Porte represented a unique situation due to an
unusual combination of urban, industrial, and lake contributions.
These conclusions are in agreement with the available results from advertent
modification experiments. Increases in precipitation of as much as 30 percent in
the annual rainfall in an area such as La Porte seems to require thermodynamic
seeding effects, i.e., through the release of additional convective activity. Micro-
physical effects such as have been observed at Climax should appear primarily in
terms of increased number of light precipitation days and should, as well, be more
related to stratiform cloud types. Ice nuclei, according to Langer [41] were
measured in concentrations of about 30/liter at -20°C. Results of seeding programs
at Flagstaff and elsewhere suggest that this concentration may not be sufficent to
produce the dynamic seeding effects found in the advertent cumulus seeding programs.
Condensation nuclei effects, although not discussed at length in the La Porte example,
have not been shown to result in such large precipitation increases. It is concluded that
the thermal and frictional effects, resulting in frequent updraft regions in a localized
area and the stimulation of convective motions in this area, are the most likely causes
of the pronounced increases indicated by the La Porte observations. A corollary of
this conclusion is that it should be possible to identify those days on which convective
activity is so poised that additional stimulation is sufficient to release the latent
convective instability in the preferred area. This identification problem has been
considered in the case of cumulus seeding with considerable success.
The Australian study by Warner has shown apparent decreases in precipitation
of the order of 25 percent. These effects have been attributed to the production of
large numbers of small cloud droplets and a consequent decrease in coalescence
28
-------
growth rate. Cloud droplet concentrations averaged about 900 per cm** in the inland
areas where cane smoke might have affected the cloud microstructure. Peak droplet
concentrations were over 2500 per cm3.
It has been indicated earlier that the principal effect of a slowdown in the
coalescence growth rate would be to prevent precipitation from developing in the
smaller clouds whose lifetimes are not particularly long. It has been estimated that
a decrease of 25 percent in total precipitation would result if precipitation were
eliminated from clouds with diameters less than 3 to 4 km. Larger clouds (with
longer lifetimes) might produce rainfall in much the same manner, regardless of
cloud droplet concentration. There is evidence in Warner's data that the number of
showery days has been decreased substantially but that the amount of rain per shower
day has not changed in a similar manner.
The concentration of cloud droplets reported by Warner is not extreme in
comparison to numbers found in other areas. In Flagstaff, concentrations of
700—1000/cm3 are frequent and occasional values of 2000/cm are observed. In
Flagstaff, however, substantial coalescence rain occurs in clouds of 3 to 4 km
diameter in spite of the large droplet concentrations. Number of droplets/cm3 should
not be the ultimate predictor of the occurrence of coalescence precipitation, and it
is entirely possible that a sufficient number of large particles may have been present
at Flagstaff to initiate the precipitation. It is concluded, therefore, that the Australian
data, if statistically sound, may show the result of added condensation nuclei but that
a droplet distribution devoid of large particles would be required to produce such a
marked change in precipitation as has been reported.
Hobbs, Radke, and Shumway [37] indicate that centers of increased precipita-
tion appear downwind of local industrial sources such as smelters or paper mills.
Precipitation increases are quoted at 30 percent for several areas when comparing
the industrially active, recent years (1947—1966) with the earlier years of 1929—1949.
The authors point out, however, that the entire northwest apparently experienced
heavier rain in the 1947—1966 period than in the earlier era. The value of 30 percent
includes both the effects of the wetter areal trend and the possible effects of industrial
sources. No attempt has been made to separate these two effects and a cursory glance
at the data suggests that no more than a 15 to 20 percent increase should be attributed
to the industrial effects. The authors attribute the apparent increase to the effects of
condensation nuclei released from the industrial plants.
Observations are described in the article by Hobbs et al, which indicate numerous
occasions of cumulus cloud development downwind of the particular plants in question.
On one occasion, a cloud street of 30 km in length was reported downwind of a paper
mill in a manner which can hardly be attributed to condensation nuclei. Evidence is
also given to indicate that the precipitation in the 1947—1966 period was more convec-
tively generated than in the earlier period. The size of the apparent increase (second
in magnitude only to that reported from La Porte), the cloud observations, and the
convective nature of the precipitation all suggests that dynamic effects on the precipi-
tation mechanisms should not be ruled out of consideration. The similarities to the
La Porte situation in terms of cloud developments and precipitation increases are
striking. In addition, calculations such as have been possible to date do not support
29
-------
such large precipitation increases as 15 to 20 percent (in annual rainfall) as a result
of stimulation of the condensation-coalescence process. It is concluded, therefore,
that the Washington data may afford another example of a thermodynamic or frictional
effect on precipitation in an environment somewhat different from that existing at
La Porte.
Data on Long Island (Miller [38]) suggest precipitation increases of around
15 percent downwind of New York City compared to surrounding areas. No detailed
examination of the concurrent environment conditions has been given. The pattern of
the increase, with isohyets symmetrical with respect to the island's longest axis,
are suggestive of dynamic effects rather than microphysical causes. It could be
hypothesized that convection over the island with reference to the cooler, surrounding
water surfaces might contribute to such an isohyetal pattern. As in the La Porte
case, the effect of New York City could be to initiate the convection in a consistent
location.
The remaining examples of precipitation changes are of the order of 10 percent
or less. Little physical documentation of the environment conditions is usually given
and the increases themselves are small enough to be more subject to doubt than the
preceding cases. Under these conditions, it is possible to comment on plausible
reasons for the precipitation effects if, in fact, they can be shown to be significant.
6.5.2 Regional Effects
In view of the possible influence of nuclei on warm cloud precipitation, as shown
by the Australian data, it was decided that long-term trends in summer rainfall in
several areas of the United States should be examined. In the light of probable
increases in pollution levels over a period of years in the eastern sections of the
country, widespread effects on precipitation might be expected to appear first in
summer precipitation amounts in these areas. June, July, August rainfall was
plotted for a period of years for several stations in the eastern part of the United
States. Data for Albany and St. Louis are shown in Fig. 5. Data for Nashville,
Cincinnati, and Philadelphia were also examined but show no particular trends over
the past 100 years or so.
The most pronounced summer rainfall trend found at any of the five stations is
shown in the curve for Albany in Fig. 5. Summer precipitation in this area appears to
have declined 20 to 25 percent over the period of the last 100 years. It can be argued
that this trend was particularly pronounced during the past 20 years, but such short-
term trends may occur as a result of a number of causes. At St. Louis, the general
trend was downward until the mid-1930's, after which evidence of rising summer
precipitation amounts is apparent.
Long-term trends in areal precipitation amounts could result from a variety of
causes. Along with the possibilities of nuclei and moisture effects, it is quite probable
that true climatic trends, identified with long-term circulation changes, may also play
a role in determining precipitation trends. As we have seen such long-term changes
may be associated with environmental changes, or may be a result of "almost-intran-
sitivity." It is important, therefore, that identification of such areas as Albany be
followed by more detailed studies to identify possible causes.
30
-------
CO
5,
in.
0)
S 12.5
c 10.0
o
$ 7.5
Albany
o
o>
5.0
2.5
1868
1888
1908
1928
1948
1968
I ln-
12-5
§ 10.0
5
o
o
EH
7.5
5.0
2.5
0
1868
I
1888
1908
1928
1948
1968
Fig. 5. June—August precipitation trends.
31
-------
7.0 THE BIOSPHERE AND LONG-TERM EFFECTS OF AIR POLLUTION
The long-term effects of air pollution on the biota of the earth are subtle, very
poorly understood, and presently receiving only very limited attention. These con-
clusions have been reached after a review of recent literature and research proposals
and personal contact with many of the leading investigators in the fields of air pol-
lution investigation.
In general, we need more meteorological and chemical information regarding
the fate of air contaminants. As suggested in Table I of Appendix 2 the terminal
nature of numerous gaseous pollutants is a chemical different from the source nature
of the pollutant. What are the terminal forms of oxides of sulfur, carbon and nitrogen?
What are the terminal forms of organic primary and secondary (photochemical)
pollutants ? Where are they deposited (over land, oceans) ? How are they deposited
(rain, dust)? In what quantities are they deposited?
Awareness of the fate (compositional, spatial, temporal) of air contaminants
will suggest the nature of potential problems.
a. Input of significant amounts of nitrates and nitrites (conversion
products of nitrogen oxides) into certain ecosystems might result in or
contribute to eutrophication of certain water resources.
b. Input of sulfates (conversion products of sulfur oxides), if large
enough (specific levels unknown and would vary depending on soil character-
istics), could influence soil pH and stimulate the microbial formation of
sulfides. Significant alterations in pH would profoundly effect nutrient
availability to plants and plant disease caused by certain soil pathogens.
Sulfide increases would accelerate corrosion potential of the soil.
c. Fate of oxides of carbon is extremely unclear. Many questions
exist regarding the global balances of CO2. Are plants growing better in
response to increased availability of CO2? Are temperatures being
altered? In the case of CO, the nature of the subtraction process remains
unknown. It may be microbial or photochemical. If microbial, what
microbes are important? In what ecosystem do they operate? Are they
under stress from any other environmental alterations?
Knowledge of the potential problems will permit the design of experiments,
employing plants and animals, capable of assessing the impact and monitoring the
influence.
More meteorological information is required with respect to the influence of air
pollutants on climatic and weather parameters.
a. What alterations are manifest in patterns of solar radiation ?
Because of wavelength specificity, the photosynthetic capacity of plants
may not be significantly reduced. The germicidal effect of light, however,
may be altered as meaningful subtractions may occur in the UV wave-
lengths. What is the significance of this subtraction on the ecology of air-
borne spores and other inocula of animal and plant disease agents?
32
-------
b. What is the significance of participate pollution in regard to
rain and fog patterns ? Changes in frequency and intensity of precipita-
tion patterns can have very significant effects on physiological and
pathological phenomena of plants growing in urban environments.
Plants currently being used to assess the presence of air pollutants are pri-
marily employed to monitor for relatively large amounts of air contaminants (source;
nature; form) and are evaluated by observing acute effects. Plants presently employed
in these programs are not used to measure chronic or long-term influences. Vegeta-
tive monitoring of long-term influences may require (if it is at all possible) different
plants which will react to the pollutant in its terminal form and which can respond to
relatively low levels. There is at least one very significant exception to the above
generalizations. In the case of accumulatory materials such as lead, fluroide and
chloride, plants may provide very convenient and accurate monitoring systems. They
do not appear to be used currently in this capacity.
In summary, the biological significance of the long-term influence of air pol-
lutants cannot be evaluated without greater information with regard to the physical
and chemical characteristics of the environment and environmental contaminants as
outlined above.
33
-------
8.0 MONITORING
8.1 Rationale of Monitoring Long-term and Large-scale Phenomena
Monitoring is described in Section 1.3 as the "purposeful, controlled, continuing
observation on a global or local scale of a pollutant or of an established or suspected
effect of a pollutant." Recent experience has suggested that everyone has his own
conception of monitoring and that some expansion of the idea of monitoring which is
adopted in this report is necessary. Here, monitoring implies selection as well as
purpose. We do not "monitor" a system by continuous watch on all its detail, but by
choosing some aspect as an indicator of the state of the system and observing this
aspect with the precision, frequency and geographical coverage which is necessary—
and just as important with no more effort or coverage than is necessary. The specu-
lative collection of data or the collection of data for occasional use—for example, for
the preparation of tomorrow's weather forecast—or for a limited research project is
not monitoring. For example (and it is not put forward as a sensible example), the
chemical determination of cyanide in river water, either because it might prove
interesting or specifically for the purpose of correlation with the respiration rate
of fish would not be considered "monitoring." On the other hand, systematic observa-
tions of the respiration rate of fish might be described as monitoring cyanide if the
correlation were specific and the procedure more sensitive, accurate, universal and/or
more economical than chemical methods. If, however, cyanide in the water had no
observable effect, or no known potential effect, other than that on the respiration
rate of fish, it would be an unnecessary complication to describe measuring the respira-
tion rate of fish as monitoring cyanide in the water. The essence of monitoring is
selection—selection of critical phenomena to be watched, selection of observables
critical to one or more of these phenomena, and selection and standardization of
methods of observation.
Our present task is to watch for long-term changes in climate induced by air
pollution and the accompanying changes in biota. Maintenance of a climatic record is
the task of weather services: a continually more detailed record is kept, but if it
should show changes we have seen that these cannot necessarily be attributed to air
pollution. We must, therefore, identify pollutants which could conceivably influence
climate in the long-term, study their concentration and distribution and study the
problem of monitoring them. We have identified CO2 and aerosol. COg has so long
a residence time that it is almost uniformly mixed. We can handle the minor varia-
tions by choice of site and monitor by direct measurement in a few localities. We
perhaps do not have sufficient information on the minor diurnal and annual variations
at chosen sites to monitor on any but a quasi-continuous basis at present, but we can
envisage the situation when a limited number of determinations each year at each
site will be sufficient to monitor an established temporal trend. The need then will be
for an exactly reproducible, preferably absolute, method, so that if we can establish
such a method now we should do so, even if it is more elaborate than less readily
standardized methods.
The problems with aerosol are very different. The particulate material is
highly inhomogeneous; it (or most of it) has a short life in the atmosphere and in any
one locality its incidence is as variable as the weather. Comprehensive global
34
-------
observation of the nature of atmospheric aerosol would be an enormous task. To reduce
the task, we inquire how aerosol reacts on climate, and find it does so through modifi-
cation of condensation processes and modification of the radiation field. We find that
two specialized types of particle are involved in condensation processes, and, perhaps
because of the short lifetime, we can identify a fairly constant "natural" background.
Study of long-term global effects requires the monitoring of this background, a problem
similar to the monitoring of CO^. Study of long-term localized climatic effects
requires the monitoring, on a local basis, of the contribution of local sources.
The radiative effects of aerosol can be studied by the bulk effect of the total
atmospheric content on solar and terrestrial radiation. Radiation measurements
might be considered a task of weather services, but are at present only partially
accepted as such. All types of particles are now involved and no "natural background"
fairly constant in time and space, has yet been established. The problem could, there-
fore, be approached by a broad quantitative survey of one particular radiative effect,
which need not in itself be more than an indicator of radiative interference by aerosol,
simultaneously with the start of a program involving a limited number of more com-
prehensive radiation measurements of higher precision selected with reference to
our current limited knowledge of the effects and distribution of aerosol. We would be
fortunate if the first attempt produced the optimum pattern of measurement for a
long-term monitoring system, and it is, therefore, essential that it should be so well
documented that its results could be accurately compared in some way with those of
future systems.
In the area of biological effects, we are in an even more rudimentary state of
knowledge—"The long-term effects of air pollution on the biota of the earth are subtle,
very poorly understood, and presently receiving only very limited attention" (Section 7).
Research is required rather than monitoring. But this research almost certainly
involves monitoring of concentration and rates of deposition of certain pollutants in
and on chosen ecosystems, just as research on the economic and medical effects of
pollution in cities calls for the monitoring of concentration and deposition rates in
which NAPCA is now engaged. As knowledge of long-term effects on biota is increased,
it may be possible to limit in quantity, and necessary to improve in quality, the measure-
ments involved. Simultaneously, the possibility will develop of using one or more
biological effects as an indicator or a warning of the likelihood of others.
8.2 Artificial Earth Satellites and the Monitoring of Long-term Trends
Artificial earth satellites offer the obvious advantage of global coverage of
observations, and the existence of operational meteorological satellites and of
experimental earth resources satellites makes it appropriate to consider their
possibilities for monitoring pollutants and their long-term effects. A comprehensive
study of the possibilities of detection of pollutants from satellite platforms has
recently been made (C. B. Ludwig, R. Bartle and M. Griggs [42]). This study
concerned itself in the main with currently available techniques and reading it leaves
the impression that detection of most pollutants is just feasible but that quantitative
assessment with any useful precision is a little beyond current techniques. Further
development of on-board techniques is clearly possible, and a more imaginative and
35
-------
purposeful deployment of satellite vehicles might ease the problems. However, once
one gets away from the concept of a space-sharing general-user satellite to consider
a system specifically designed for pollution detection, the cost becomes very high
indeed and cost-benefit becomes essential.
In the context of the study of long-term effects of pollution, we are concerned
with changes in concentration of fractions of one percent per year, and there is no
obvious immediate appeal in satellite methods of detection of pollutants, even allowing
for the global coverage. Identifying the two pollutants of major interest, CO2 and
aerosol, it does not at present seem that satellite techniques will in the next few
years offer any advantage over surface monitoring for CG>2, but that they may make
a valuable contribution—indeed may prove essential—to the study of the radiative
effects of aerosol. The albedo of the planet earth and its secular changes can be
measured most conveniently and directly, if not exclusively, by satellite techniques.
Current techniques are not sufficiently precise to allow long-term monitoring. They
have been developed in the context of operational meteorological satellites and make
a local measurement of albedo; a quantity varying rapidly in space and time, from
which the long-term global mean values we require must be deduced. We cannot
really assess the value to our problem of current work until we have some feel for
the year-to-year statistical variations of global albedo as deduced from these measure-
ments. No space experiment or projected experiment specifically designed to measure
the planetary albedo with high (fractions of 1 percent) precision has been publicised.
It would call for very high orbits, precision radiometry, and much patience; and it
would be very expensive.
The contribution of aerosol to the outgoing terrestrial radiation is required
for a full understanding of its long-term effects. It might be most readily detected
in the 10—12 /jm window, and study of existing IRIS interferograms might give some
feeling for possibilities. As has been suggested elsewhere, the requirements of
long-term monitoring of particulates and CO_ should be kept in mind when operational
equipment for remote temperature sensing is being designed.
36
-------
9.0 CONCLUSIONS— AREAS OF UNCERTAINTY
9.1 The Major Pollutants
9.1.1 Carbon Dioxide— Concentration and Life Cycle
The evidence of a global increase of CO2 concentration appears conclusive, but
the rate of increase is still in doubt. The atmosphere-ocean-biosphere partition of
CO2 is not fully understood (see Section 2.3).
Global monitoring is required, hopefully in an international program: A few
stations in carefully chosen situations should suffice. Combustion and vegetative
growth must be considered in choosing the stations— if their effects cannot be
eliminated they must be understood. Available methods of monitoring are sound
but not ideal (see Section 8) .
Theoretical and observational research on the partition of CCvj should be
encouraged. Isotope (C^) methods are indicated (see Sections 2.3, 8).
9.1.2 Aerosol— Concentration, Nature and Life Cycle
There is evidence of local and regional increases in gross aerosol content of
the atmosphere but not yet sound evidence of a global increase. To understand the
radiative effects of this aerosol the number concentration, size distribution, com-
position and refractive indexes of the particles must be known (see Sections 2.2, 5.2).
More observations of these details are required— these need not take the form of a
regular monitoring program (see below) but extensive tropospheric and stratospheric
sampling should continue (see Section 8) .
Residence time and removal processes are important areas of research.
Quantitative studies of large-scale removal— air mass cleansing processes— are
lacking (see Section 2.2).
9.1.3 Sulphur Dioxide— Transport and Life Cycle
The extent to which sulphate aerosol in the troposphere and stratosphere is a
product of pollutant SOo rather than part of the natural sulphur cycle should be
determined. One attack on the problem might be by detailed work on the chemistry of
sulphur-particulate pollution analogous to that which has thrown light on the photo-
chemical smog process (see Sections 2.2, 2.4).
The deposition of sulphate ion in rain is an associated problem with a long-term
biological interest. Careful and detailed long-term investigation of precipitation
chemistry as part of an ecological program is indicated. Forest and lake ecology in
Northeastern U.S.A. and Scandinavia are sensitive areas (see Section 7).
Regional and sub- continental air pollution models are of interest in this
connection (see Section 4).
9.1.4 Nitrogen Oxides— Transport and Life Cycle
There is possible interest in the long-term biological effects of nitrate rain-out.
The problems are similar to those for SC^. A quantitative budget of pollutant nitrogen
in the Los Angeles area would be interesting (see Sections 2.5, 7).
37
-------
9.1.5 HpO Concentration
Long-term effects of water emission from surface sources seem likely to be
very localized but two aspects of the emissions from aircraft cause concern—a
possible increase in cloud caused by present generation machines and a probable
significant increase in stratospheric f^O concentration if supersonic transport
becomes general. Careful organized observation would help with the first problem.
A minimal monitoring program for stratospheric water vapor could be initiated now.
The computational investigation of radiative flux divergence is in its least satisfactory
state near the tropopause (see Section 2.7).
9.1.6 Carbon Monoxide—Concentration. Transport and Life Cycle
Very recent theoretical work has opened a new aspect of this problem. No
serious geophysical or biological long-term effect of CO emission has been identified
(see Section 2.3).
9.2 Long-term Climatic Effects
Empirical correlation of observed climatic trends with observed pollution
trends does not establish causal connection (see Section 3.1). We must seek a physical
theory, by way of climatic models. The simpler models will probably always be
inadequate, but are suggestive. Current examples have weaknesses in the treatment
of meridional energy transport and the relation of albedo to temperature (see Section
3.2). The more complex current models are deficient in their treatment of cloud and
its reaction on the radiation field and albedo. Dynamically, they may be more
elaborate than is necessary for the long-term climatic application. Proper treatment
of the effects of aerosol will always be a major difficulty (see Sections 3.3, 3.4).
The more obvious aspects of climate are effectively monitored by weather
services, but the records should be examined for regional changes in areas subject
to changing pollution levels.
One aspect of climatic change which could be examined by detailed modeling is
the possibility of global effects following a regional change in some dynamically
sensitive area—e.g., a change in albedo or radiative flux divergence over the land
areas of eastern North America (see Section 3.3).
9.3 Other Long-term Geophysical Effects
Changes in the stratosphere and mesosphere are an aspect of climatic change,
but they may also involve changes in the lower ionosphere with consequent changes
in radio-propagation. The effect of changes of CO£ content on the temperature of the
mesosphere should be further investigated in the light of possible changes in chemical
reaction rates. It seems unlikely that surface-emitted NO could reach the mesosphere,
where it would be very reactive. Changes of f^O content would affect the rates of
some photochemical processes (see Sections 2.1, 2.3, 2.7).
9.4 Solar Radiation and Radiative Transfer (Other than as a Facet of Climatic Models)
The reduction of solar radiation by polluted atmospheres has not been shown to
have any major biological effects at its present level. It is severe at times and might
38
-------
have subtle psychological and socio-economic effects, so deserves examination in
its own right. It can also serve, suitably interpreted, as a monitor of aerosol content
and to provide indirect estimates of atmospheric albedo and absorption. As an instru-
ment for monitoring slow global changes remote, preferably high-altitude, stations
are called for. As a minimum they should record direct and diffuse solar radiation in
the U.V., visible, and I.R. They must pay careful attention to standardization. Diffuse
radiation in the I.R. is a sensitive indicator of aerosol loading (see Section 5.2).
The network of simple "turbidity" measurements established by NAPCA monitors
the geographical distribution of aerosol pollution, and should indicate statistically any
secular changes. It should be expanded geographically, and maintained. Calibration
standards must not be relaxed (see Section 5.2).
Turbidity measurements give a relative measure of total aerosol loading. They
should be supplemented in a few locations remote from pollution sources by observa-
tions of the height distribution of aerosol using stable or well calibrated methods.
Laser ('Lidar') methods appear to offer a good approach. The objective is stable
long-term monitoring of the aerosol loading in the upper troposphere and lower
stratosphere (see Sections 5.2, 8).
9.5 Nucleation and Precipitation
9.5.1 Nuclei Concentrations
There is insufficient information on the concentrations of ice and condensation
nuclei as related to various anthropogenic sources. There is a particular need for
more detailed data on the removal processes which operate in the atmosphere. It
is vital to determine the areal and quantitative extent of the nuclei changes now
taking place in the atmosphere and to isolate the possible long-term buildups in
concentration from the local variability. Two scales of nuclei observations should
be considered:
a. Global scale—A fixed network of up to 10 stations should be
established to measure condensation and ice nuclei on a routine basis.
A portion of these should be located in relatively clean air, i.e., Hawaii
and non-populated sections of the West Coast. A portion of the network
should also be located in the Northeastern United States where local
nuclei levels may be increasing regardless of the trend on a global
basis.
b. Local scale—In addition to the fixed network of nuclei stations,
it would be useful to have mobile units which could explore the effects
of individual source areas, determine their downwind extent, and obtain
useful information on the effectiveness of the removal processes.
(see Sections 2.2, 6.2, 6.3, 6.4).
9.5.2 Observational Studies
Primary evidence of inadvertent effects has come from statistical studies of
climatological data. Several areas of possible effect have been identified in this
39
-------
manner. Further progress in these cases will come from the development of a better
physical understanding of the causes of the anomalies. This can come about through
short, intensive field studies of nuclei, heat, and moisture fields in the area of
interest, together with modeling studies.
A detailed study of the environment in the areas of possible inadvertent effect,
e.g., La Porte or Seattle, should be made (see Section 6.5.1).
The variability in cloud droplet concentrations (particularly in summer cumulus)
should be explored for various areas of the country. It would be particularly important
to obtain more concentration data in the east and northeast United States where
numerous particulate sources may combine to influence large sections. Coalescence
processes in summer cumulus should be studied, requiring concurrent radar, cloud
nuclei, and cloud drop measurements (see Sections 6.5.1, 6.5.2).
Additional data are needed on free iodine concentrations in the atmosphere in
order to determine the importance of the automobile exhaust-iodine reaction (see
Section 6.4).
9.5.3 Model Studies
Many of the processes involved in the inadvertent nucleation effects can be
modeled with computer studies. This includes the coalescence process, plumes from
isolated sources, frictional and thermal "heat island" effects, and dynamic seeding
effects of ice nuclei. A very considerable effort is already being expended but there
is room for additional modeling work, particularly directed toward simple model
frameworks which can be used for parametric sensitivity studies (see Section 6.5).
9.5.4 Statistical Studies
Much has already been accomplished as a result of statistical studies in calling
attention to the problem of inadvertent modification. Additional studies in other
areas and in greater depth in already-identified locations are warranted (see Section 6.5).
9.5.5 Laboratory Studies
It seems inevitable that many of the inadvertent modification problems will
have to be resolved using in situ atmospheric measurements. Supporting laboratory
studies would be valuable, however. These should include additional work on auto-
mobile exhausts aimed at determining required iodine concentrations, studies of
nucleating properties of various materials and pollution processes, and the general
problem of the changes in nucleation properties due to agglomeration, radiation, etc.,
which occur in the atmosphere after release from the source (see Section 6).
9.6 Applications of Artificial Earth Satellites
Weather satellite systems introduce the possibility of monitoring albedo. It
seems likely that the precision attainable by present systems is insufficient; this
should be studied and improved when possible. Projected systems for global tempera-
ture observation may bring an opportunity for CO2 and aerosol monitoring (see
Sections 3, 8.2).
40
-------
In general, satellite methods do not seem obviously suited to investigation of
long-term air pollution problems—the enormous data-gathering capacity does not
appear to be particularly advantageous. Cost benefit analysis should be made of all
proposed satellite measurements, but the possibilities of combining with other
monitoring on weather and earth resources satellites should be kept in mind by
active feasibility studies (see Section 8.2).
10.0 PROJECTS FOR EARLY ATTENTION
Priorities must be assessed by a compromise between importance, likelihood
of early success, and magnitude and expense of the effort involved in relation to the
available total. I suggest that first attention should be given to the monitoring
projects which have begun, to ensure continuance, to allow a modest expansion, and
to improve methods. These projects are CC>2 measurement at a remote station,
solar radiation measurement at a remote station, and the widespread measurement of
atmospheric turbidity. The addition of measurements of all types of nuclei at one
or more remote stations should be considered at this stage, as should proposals for
the monitoring of stratospheric and high tropospheric aerosol, perhaps by "lidar"
methods.
NAPCA should also support projects for monitoring the planetary albedo,
particularly by measurement from satellites, though I doubt whether the necessary
precision will be attained at an early date.
Comprehensive monitoring of the chemical composition of precipitation, in
conjunction with ecological studies is suggested as a first step in the study of truly
long-term biological effects of pollution.
Turning from observation to the areas of understanding and prediction, there
is a requirement for further development of climatic modeling. NAPCA's particular
interest is in the long-term response to small but specific changes in atmospheric
constitution, and concentration on this aspect calls for some shift in the priorities of
groups presently engaged in this arduous and expensive work. Resources are stretched
even now, and an early review of facilities and potential by all interested agencies
might be useful.
Long-term effects on less than global scale are likely to cause increasing
public concern. The sub-continental scale transport of pollutants might be studied
by an extension of urban air pollution models or modification of weather forecasting
models. No great scientific advance is involved; the political interest suggests early
action.
More localized effects can properly be described as potentially long-term
modifications of climate. Urban effects on precipitation patterns and solar radiation
can properly be studied within a program concerned with "long-term geophysical and
biological effects of pollution." The relative ease of formulating specific projects,
and the increased probability of an early useful outcome should not be allowed to
divert too great a proportion of support from the more difficult larger-scale and
longer-term problems which will not be quickly solved.
41
-------
11.0 REFERENCES
1. E.Robinson and R.C. Bobbins, "Sources,abundance and fate of gaseous atmospheric
pollutants," 1968, American Petroleum Institute, New York, N. Y.
2. B. Bolin and C. D. Keeling, ''Large-scale atmospheric mixing as deduced from
the seasonal and meridional variations of CC>2," J. Geophys. Res., 68. .
pp. 3899-3920, 1963.
3. W. Seiler and C. Junge, "Decrease of CO mixing ratio above the polar tropopause,"
Tellus 21, pp. 447-449, 1969.
4. H. G. Horak, "Aerosol concentration and extinction in the earth's atmosphere,"
Los Almos Report LA-4023, Clearinghouse, 1969.
5. G. M. Shah, "Aerosols in the stratosphere," Canadian Aero.and Space J., 15.
pp. 321-326, 1969.
6. S. Twomey and T. A. Wojciechowski, "Observations of the geographical varia-
tion of cloud nuclei," J. Atmos Sci., 26, pp. 684—688, 1969.
7. P. A. Squires, "An estimate of the anthropogenic production of cloud nuclei,"
J. Rech. Atmos., 2, pp. 297—308, 1966.
8. S. Twomey, "On the nature and origin of natural cloud nuclei," Bull. Obs. Puy
de Dome, !_, pp. 1—19, 1960.
9. E. Hesstvedt, "Vertical distribution of CO near the tropopause," Nature, 225.
p. 50, 1970.
10. J. Pressman and P. Warneck, "The stratosphere as a chemical sink for CO,"
J. Atmos. Sci., 2J7, pp. 155—163, 1970.
11. G. R. Hilst, "The sensitivities of air quality prediction to input errors and
uncertainties," NAPCA Symposium on multiple source urban diffusion models,
Chapel Hill, North Carolina, 1969. (to be published).
12. G. D. Robinson, "Absorption of solar radiation by atmospheric aerosol as
revealed by measurement at the ground," Arch. Meteorol. Geophys. Biokl.
B12, pp. 19-40, 1962.
13. J. M. Craddock, "A contribution to the study of meteorological time series,"
M.R. P. 1051, Met. Office London, 1957.
14. H. Charnock and G. D. Robinson, "Spectral estimates from subdivided meteoro-
logical series," M.R.P. 1062, Met. Office London, 1957.
15. K. Myakoda, J. Smagorinsky, R. F. Strickler and G. D. Hembree, "Experimental
extended predictions with a nine-level atmospheric model," Mo. Wea. Rep., 97,
pp. 1-76, 1969.
16. E. N. Lorenz, "Climatic determinism," Met. Monog., j5, pp. 1—3, 1968.
17. J. S. Sawyer, "Notes on the response of the general circulation to changes in
the solar constant," Proc. Rome Symposium on Changes of Climate 1961,
UNESCO Paris.
42
-------
18. J. S. Sawyer, "Possible variations of the general circulation of the atmosphere,"
World Climate from 8000 to O BC, pp. 218—229, 1966, Roy. Met. Soc., London.
19. W. D. Sellers, "A global climatic model based on the energy balance of the earth
atmosphere system," J. Appl. Met. 8_, pp. 392—400, 1969.
20. L. R. Rakipova, "Changes in the zonal distribution of atmospheric temperature
as a result of active influence on the climate," Modern problems of climatology,
pp. 388—421, 1966, Leningrad, (Translation Clearinghouse AD 670-893).
21. G. D. Robinson, "Some current projects for global meteorological observation
and experiment." Quart. J. R. Meteor. Soc., 93_, pp. 409—418, 1967.
22. E. N. Lorenz, "The predictability of a flow which possesses many scales of
motion," Tellus 21, pp. 289-307, 1969.
23. F. H. Bushby and M. S. Timpson, "A 10-level atmospheric model and frontal
rain," Quart. J. R. Met. Soc., 93_, pp. 1—17, 1967.
24. NAPCA Symposium on multiple source urban diffusion models, Chapel Hill,
North Carolina, 1969. (to be published).
25. D. W. Fisher, This report, Appendix 2.
26. F. Moller, "The influence of changes in the CO2 concentration in air on the
radiation balance of the earth's surface and on the climate," J. Geophys. Res., 68.
pp. 3877-3886, 1963.
27. S. Manabe and R. T. Weatherald, "Thermal equilibrium of the atmosphere with a
given distribution of relative humidity," J. Atmos. Sci., pp. 241—259, 1967.
28. C. M. Peterson, H. J. Paulus, G. H. Foley, "The number-size distribution of
atmospheric particles during temperature inversions," J. Air Poll. Control
Assoc., 19_, pp. 795—801, 1969.
29. G. D. Robinson, "Some determinations of atmospheric absorption by measurement
of solar radiation from aircraft and at the surface," Quart. J. R. Meteor. Soc., 92.
pp. 263-269, 1966
30. J. M. Waldram, "Measurements of the photometric properties of the atmosphere,"
Quart. J. R. Meteor. Soc., 71^, pp. 319—336, 1945.
31. H. H. Lettau and K. Lettau, "Short-wave radiation climatonomy, " Tellus 21.
pp. 208-223, 1969.
32. W. T. Roach, "Some aircraft observations of fluxes of solar radiation in the
atmosphere," Quart. J. R. Meteor. Soc., 8J7, pp. 346—363, 1961.
33. R. A. McCormick and J. H. Ludwig, "Climate modification by atmospheric aerosols,"
Science, 156. pp. 1358—1359, 1967.
34. E. C. Flowers, R. A. McCormick and K. R. Kurfis, "Atmospheric turbidity over
the U.S., 1961-1966," J. Appl. Met., 8_, pp. 955-962, 1969.
35. S. H. Changnon, "The La Porte Weather Anomaly, fact or fiction?" Bull. Amer.
Meteor. Soc., 49, pp. 4—11, 1968.
43
-------
36. J. Warner, "A reduction in rainfall with smoke from sugar-cane fires—an
inadvertent weather modification," J. Appl. Meteor., 7_, pp. 247—251, 1968.
37. P. V. Hobbs, L. F. Radke and S. E. Shumway, "Cloud condensation nuclei from
industrial sources and their apparent influence on precipitation in Washington
State," J. Atmos. Sci., 2J7, pp. 81-89, 1970.
38. J. F. Miller, "Precipitation regimes over Long Island, N. Y.," ESSA Office of
Hydrology, 1967.
39. S. A. Chagnon, "Recent studies of urban effects on precipitation in the U.S.,"
Bull. Amer. Meteor. Soc., 50, pp. 411—421, 1969.
40. J. T. Peterson, "The climate of cities. A survey of recent literature," NAPCA
Pub. No. AP-59, 1969.
41. G. Langer, "Ice nuclei generated by steel-mill activity," Amer. Met. Soc. Proc.
1st Nat. Conf. on Weather Modification, pp. 220—227, 1968.
42. C. B. Ludwig, R. Bartle and M. Griggs, "Study of air pollutant detection by
remote sensors," NASA Report CR-1380, 1969.
44
------- |