|
Items in cart [1] |
Questions? Call 888-624-8373 |
| ||||||||||||
| Copyright © 2009. National Academy of Sciences. All rights reserved. Terms of Use and Privacy Statement |
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 34
Interpretation of Past
Climatic Variability
frond PaTeoenviron~ental
Inclicators
c)
INTRO DU CTI O N
CHARLES W. STOCKTON
University of Arizona
The primary objective of this panel is to evaluate the
national water supply in light of climatic variability. The
following questions then arise: How crucial is knowledge
of climatic variability to the design and operation of a
water supply system, its reservoirs, and its distribution
system? Should we incorporate newly gained knowledge
of climatic change into the design and operation of such
systems? If so, how much does climate vary? Can the
variability be established on the basis of historical rec-
ords? What is the probability of climatic change in the
next 25 to 50 years?
We must first define what we mean by weather and
climate. According to the Department of Transportation
Select Panel assessing the variability of the climate
(Mitchell et al., 1975), the term "weather" refers to the
total array of atmospheric conditions varying with time
and location on the earth's surface. "Climate" connotes
"average weather" but can be viewed from two perspec-
tives:
34
1. A purely statistical approach in which climate is the
sum of the weather as experienced at a point or over a
designated area of the earth for a given period of time.
2. A physical concept that recognizes climate as a basic
physical entity and weather as the momentary, transient
behavior of the atmosphere attempting to satisfy the re-
quirements dictated by the climate for horizontal and
vertical transfer of mass, momentum, and energy.
There seems to be disagreement even among experts as
to the need for understanding climatic variability in rela-
tion to water supply. There are those who believe that,
since most projects in water resources have an economic
life of from 40 to 100 years, and since there appears to
have been little or no obvious climatic change over the
past 200 years, the chance for natural climatic change in
the next 200 years is minimal, and, therefore, the ques-
tion is academic. For example, Chin and Yevjevich (1974)
purported to show that climatic variation could be re-
duced to a deterministic component based on the Milan-
kovich theory of astronomical cycles and a simple
OCR for page 35
Interpretation of Past Climatic Variabilityfrom Paleoenvironmental Indicators
Markovian stochastic component. From this position they
went on to state that "since most systems have been built
with the economic project life in the range 40 to 100
years, the chances are minimal that the expected natural
water supply would be significantly different during
these life spans than in the past 200 years." Furthermore,
"this question is, however, not crucial for the next several
generations of contemporary earth population, but rather
is more of an academic interest like many other human
concerns with the long-term future."
On the other hand, there are those who argue that
climatic variability is a part of life on the planet earth and
that it is to our advantage to recognize it, understand it,
and take it into consideration in our planning processes.
For example, Wallis and O'Connell (1973) studied the
power of various statistical tests to distinguish small sam-
ples taken from Markovian and more persistent generat-
ing mechanisms. They concluded that statistical tests
based on records of normal hydrologic length would usu-
ally lead one to believe that a Markov generating mecha-
nism adequately represents hydrologic reality; however,
because the tests have no power, this belief, while com-
forting, is likely to be erroneous. In a companion study,
O'Connell and Wallis (1973) showed that Markov and
more persistent generating mechanisms could lead to
very different estimates of reservoir firm yield for 50-year
design lives even when the generating mechanisms used
yielded samples with identical expected values for the
mean, variance, and lag-one correlation. They concluded
that it was essential that hydrologists and water-resource
planners understand the nature of climatic variability and
persistence.
A similar position was taken by Mitchell et al. (1975),
who stated:
The climate of the earth is now known beyond any doubt to have
been in a more or less continual state of flux. Changeability is an
evident characteristic of climate on all reasonable time scales of
variation, from that of aeons down to those of millennia and
centuries. The lesson of history seems to be that climatic var~a-
bility is to be recognized and dealt with as a fundamental
quantity of climate, and that it should be potentially perilous for
man to assume that the climate of future decades and centuries
will be free of similar variability.
The issue seems to revolve around the question, "How
variable has climate been in the past?" Presumably, if
atmospheric behavior is random in time, the definition of
climatic variability would be a straightforward exercise in
classical statistical sampling theory. One could estimate
climatic variability as precisely as desired merely by
choosing a long enough averaging interval. The problem
here is that, as we go back in time, our data base di-
minishes and knowledge of atmospheric variability be-
comes less detailed and reliable. However, we do know
enough about past climates to establish that long-term
atmospheric behavior does not proceed randomly in time.
Variations of climate from one geological epoch to
another, and from one millennium to another, are clearly
35
too large in amplitude to be explained as random devia-
tions from modern averages.
How, then, do we study long-term variability? Unfortu-
nately, climatic measurements do not extend back much
60°
4oo
20°
Do
20°
4oo
60°
6oc
4oo
20°
Go
20°
40C
6oc
60°
100 60 20 0° 20° 60°
t_:
,= ,,_ ,_
1 ' :.'_
_ _ of. ~
~ -1750-1759
_ 11
100° 1 40
_
180°
_ ~
180° 140° 100° 60° 20° 0° 20° 60° 100° 140° 180°
2;~:
180° 140° 100° 60° 20° 0° 2Q° 60° 100° 140 180
i'-
=- ~`
an,. ~
i .
:, ' to . !
· ~ He
FIGURE 2.1 Growth of the network of surface pressure obser-
vations and of the area that can be covered by reliable 10-year
average isobars (Lamb, 1969).
OCR for page 36
36
beyond the 1900's on any kind of an adequate spatial
coverage. Lamb (1969) shows this well in three maps
illustrating the growth of the network of surface baromet-
ric pressure observations and of the area covered by
reliable 10-year-average isobars (Figure 2.11. Obviously,
any intensive global or even hemispherical study of
climatic variability from instrumental records is limited
to the relatively short period 1900 to present. Even on a
more local basis, the longest continuous time series of
instrumental observations covers less than three cen-
turies.
Therefore, for longer time scales, climatic variations
must be inferred from historical evidence and from the
records of various natural phenomena linked in some way
to climate. Such paleoclimatic indicators (Table 2.1) dif-
fer greatly in the time spans over which they are ap-
plicable, in the degree of climatic detail they can provide,
in the aspects of climate to which they respond, and in
the fidelity of their response.
Reconstructions of the paleoclimatic record can lead
one to believe that changes of climate, such as those
associated with alternating glacial and interglacial stages
of the Pleistocene, are smoothly varying functions of time,
readily distinguishable from the much more rapid varia-
bility of year-to-year changes of atmospheric state. Hence,
to give a stable estimate of present-day climate, the aver-
aging interval would have to be long enough only to sup-
press year-to-year sampling variability but short in com-
parison with the duration of a glacial period. Unfortu-
nately, the apparent "smoothness" of atmospheric change
in the geological past is only an illusion, attributable to
the inadequate resolving power of paleoclimatic indi-
cators. Most such indicators act to some degree as low-
pass filters of the actual climatic chronology. Our more
recent experience, based on relatively higher-pass filters
such as tree rings, varves, ice-cap stratigraphy, and pollen
analysis applicable to postglacial time, suggests that the
state of the atmosphere has varied on most, if not all,
snorter scales ot time, as well as over the longer geologic
time scales.
, . a.
TRANSFER-FUNCTION ANALYSIS
Since about 1960, advances in mathematical and statisti-
cal techniques and the availability of high-speed com-
puters that enable researchers to handle large amounts of
data have for the first time made it possible to quantify
climatic parameters derived from secondary sources
(Fritts et al., 1971; Webb and Bryson, 19721. Furthermore,
these quantified climatic parameters are provided in a
form suitable for input into dynamic models of atmo-
spheric circulation (CHEAP Project Members, 19761. By
providing quantitative data for past conditions, it is be-
coming possible to model the dynamics of past circula-
tions and to test existing models of the present circulation
with regard to their power of explanation (Gates, 19761. At
the heart of these quantitative paleoclimatic records is the
CHARLES W. STOCKTON
concept of transfer-function analysis.
Let the matrix X be a defined set of response variables
that respond to climate measured over a specified realm
of time and space. Let C be a measured set of physical
indicators of climate, atmospheric or marine, measured
over the same time-space realm and assumed to be caus-
ally related to X. Let D be another set of physical parame-
ters of the system, independent of the response of X. (D
would typically include nonclimatic effects.) Then if
D = 0, the system consists of X, C, and a set of climatic
responsefunctions Re such that
X = Re(C).
(1)
If D +0, the total response function Rig must be con-
sidered, and
X = R~(CD).
(2)
The result is calibration of the climatic signal inherent in
the secondary series X, with measured values of the
climatic variable or variables of interest.
A fundamental problem of quantitative paleoclimatol-
ogy is to find a set of transferfunctions ~ such that C can
be estimated given X; i.e.,
C =~(X).
(3)
Generally, ~ is obtained by direct empirical methods and
not by inversion of Re (or R if. The X and C used to derive
the transfer function are the calibration data set. The X to
which the transfer functions are applied is the climatic
reconstruction data set.
In the use of any transfer function, the investigator must
make several basic decisions. There are fundamental
problems concerning the assumptions used in writing any
transfer function. Principal among these is the use of"the
present as a key to the past." For example, if elements of
the biota (in ocean-sediment samples) have evolved since
the fossil deposit was formed, the calibration and recon-
struction data sets are nonhomogeneous. Another prob-
lem is the no-analogue situation, for which fossil values of
certain taxa exceed the modern values used to derive the
transfer functions. Both problems exist in tree-ring
analysis. It is assumed that a tree responds to climatic
inputs in a similar fashion throughout its lifespan such
that one can make a homogeneous transition between the
calibration data set and the reconstruction data set via the
transfer function. One can conceive of a no-analogue
situation wherein climatic events that have occurred in
the past are not present in the calibration data set.
As a direct test of any transfer function, the recon-
structed data must withstand some sort of validation test
to determine the accuracy of estimates of past climate.'$In
general, five techniques are currently being used:
i. Direct check; In tree-ring work, meteorological rec-
ords are used to validate estimates of the reconstructed
climate. This usually requires "holding back" some por-
tion of the measured record from the calibration process
for use in checking reconstructed values.
OCR for page 37
Interpretation of Past Climatic Variability from Paleoenvironmental Indicators
TABLE 2.1 Characteristics of Paleoclimatic Data Sources (after Kutzbach, 1975)
Minimum Usual
Continuity Potential Period Open Sampling Dating
Variable of Geographical to Study Interval Accuracy
Data Source Measured Evidence Coverage (yr) (yr) (yr) Climate Inference
Ocean sediments
(cores, < 2 cm/
1000 yr)
Ancient soils
Marine shorelines
Ocean sediments
(common deep-
sea cores, 2-5
cm/1000 yr)
Ocean sediments
(common deep-
sea cores, 2-5
cm/1000 yr)
Ocean sediments
(common deep-
sea cores, 2-5
cm/1000 yr)
Layered ice cores
Isotopic compo-
sition of plank-
tonic fossils;
benthic fossils;
mineralogic com-
position
Soil type Episodic
Coastal features,
reef growth
Ash and sand
accumulation
Fossil plankton
composition
Continuous Global ocean
Episodic
Continuous
Continuous
Isotopic compo-
sition of plank-
tonic fossils;
benthic fossils;
mineralogic
composition
Oxygen-isotope Continuous
concentration
(long cores)
Closed-basin lakes Lake level Episodic
Mountain glaciers Terminal positions Episodic
Ice sheets
Bog or lake
sediments
Ocean sediments
(rare cores,
> 10 cm/1000 yr)
Layered ice cores
Layered lake
sediments
Tree rings
Terminal positions Episodic
Pollen-type concen- Continuous
"ration, mineral-
ogic composition
(normal core)
Isotopic composition Continuous
of planktonic
fossils; benthic
fossils; mineral-
ogic composition
Oxygen-isotope con-
centration, thick-
ness (short cores)
Pollen-type concen-
tration (annually
layered core)
Ring width anomaly,
density, isotopic
composition
Written records Phenology, weather Episodic
logs, sailing logs,
etc.
Archeological Varied Episodic
records
1,000,000 + 1000 +
Lower and mid- 1,000,000 200 +5%
latitudes
Stable coasts,
oceanic islands
Global ocean
(outside red
clay areas)
400,000 — +5%
200,000 500 + +5%
Global ocean
(outside red
clay areas)
Global ocean
(above CaCO3
compensation
level)
Antarctica;
Greenland
Lower and
midlatitudes
45° S to 70° N
Midlatitudes to
high latitudes
50° S to 70° N
Continuous Antarctica;
Greenland
Continuous Midlatitude conti-
nents
Continuous Midlatitudes and
high-latitudes
continents
Global
Global
37
+5~o Surface temperature,
global ice volume;
bottom temperature
and bottom-water
flux; bottom-water
chemistry
Temperature, precipita-
tion, drainage
Sea level, ice
volume
Wind direction
200,000 500 +
200,000 500 + +5%
100,000 + Variable
50,000 1-100
(variable)
50,000
25,000
(common)
1,000,000
(rare)
10,000+
(common)
200,000
(rare)
10,000+
200
20
Sea-surface tempera-
ture, surface salinity,
sea-ice extent
Surface temperature,
global ice volume;
bottom temperature
and bottom-water
flux; bottom-water
chemistry
Variable Temperature
Evaporation, runoff,
precipitation, temper-
ature
Extent of mountain
glaciers
Variable Area of ice sheets
+~%
+5%
10,000 + 1-10 +1-100
10,000 + 1-10 +1-10
1000 1 1
(common)
8000
(rare)
1000 + 1 1
10,000 + — Varied
Temperature,
precipitation,
soil moisture
Surface temperature,
global ice volume;
bottom temperature
and bottom-water
flux; bottom-water
chemistry
Temperature,
accumulation
Temperature, precipi-
tation, soil moisture
Temperature, runoff,
precipitation, soil
moisture
Varied
Varied
OCR for page 38
38
o
0.1
¢~ 0.2
o o.3
In
o
-
E 04
-
~ o.s
Q
O 0.6
~ 0.7
a:
0.8
0.9 .
_ 1 ~ . ~ ~ ~ ~ ~ ~
—1.0 - 1.5 - 2.0 - 2.5 100 80 60 40 36.2 36.0 35.8 35.6
OBSERVATION: 6180 (%0) CaCO3 (%) FAUNAL INDEX (S)
INTERPRETATION: Decreasing global Increasing CaCO3 Decreasing salinity
ice volume dissolution
2. Comparison of two or more independently derived
biologically based transfer functions: In ocean-sediment
analysis this could mean comparing results obtained by
using radiolaria to those obtained by using foraminifera.
In continental regions, the results obtained by pollen
analysis can be compared with those obtained by tree-
ring analysis, or results based on tree-ring data from one
site can be compared with those from another location.
3. Comparison of isotopic and biologically based esti-
mates: Results obtained by isotopic analysis of fossil
remains taken from ocean-sediment cores should corre-
spond to those derived from species associations.
4. Concordant estimate: Independently derived trans-
fer function applied to the same paleoclimatic indicator
should produce similar results. Discord can result from
the application of two different transfer functions based
on one paleoenvironmental group or perhaps from use of
one transfer function on more than one group of variables.
5. Synoptic consistency: On an intuitive basis, the spa-
tial pattern and absolute range of synoptic maps of recon-
structed climate must conform within reasonable varia-
tions. Using the high-speed computer and numerical
simulation techniques, intuitive evaluations can be made
more rigorous and inclusive. In the end, reconstructed
climatic data from all sources listed in Table 2.1 must fit
together temporally.
Another type of problem inherent in the use of transfer
functions to derive paleoclimatic information is as-
sociated with the mathematical manipulation of the data.
Obviously, the researcher wants to choose the transfer-
function technique that is most robust against various
CHARLES W. STOCKTON
FIGURE 2.2 Comparison of paleoen-
vironmental indicators of climatic records
for the past 1,000,000 years. Roman numer-
als indicate inferred climatic periods (mod-
if~ed from Figure A.14, U.S. Committee for
the Global Atmospheric Research Program,
1975).
types of distortion, most precise in terms of error, and
most accurate in terms of reconstructed climatic values.
Referring to Eq. (3), at least three such problems can be
singled out for appraisal.
The first problem is the selection and proper applica-
tion of appropriate statistical techniques. Generally, some
sort of multivariate technique is used. When this is the
case, eigenvectors are usually used as a mode of joint
behavior classification. A question then arises as to what
criteria should be applied for inclusion and whether to
use some sort of rotation. Most models currently being
used are linear. Is it valid to assume linearity, and, if not,
how does the use of a linear model affect the final results?
Is it the best policy to utilize transformations?
The second problem includes specifying the kinds of
variables to be included in X and the space and time to be
covered. Under what circumstances does ~ not exist?
How does one define the distribution of samples in time
and space to be used in the calibration data set?
The third problem is the selection of variables and
valid estimates of them for inclusion in matrix C. These
data must represent a homogeneous reconstruction. For
example, what climatic (or environmental) variables are
most likely to influence the response and to what degree?
Is the relationship linear, and, if not, is it reasonable to
assume that the response can be approximated by a linear
relationship? Is it wise to use secondary forms of vari-
ables such as barometric pressure when it is known that
the response is tied directly to such variables as precipi-
tation and temperature? Many biological and sedimen-
tary monitoring systems show significant lag in their
responses to climatic variation. It becomes essential to
OCR for page 39
Interpretation of Past Climatic Variability from Paleoenvironmental Indicators
assess this effect and to include it in the transfer-function
model.
RE S ULTS OF S TU DI E S
Although the recent attempts to quantify paleoclimate as
derived from secondary sources are plagued with prob-
lems, the degree of coherence in spatial and temporal
variation that is achieved between reconstructions de-
rived from different sources by different investigators has
been most encouraging.
Mitchell et al. (1975) have collected and assembled
climatic interpretations based on secondary sources.
These results have been used extensively in the rest of
this section.
LONG-TERM PALEOCLIMATIC INFERENCES
(GLOBAL AND HEMISPHERIC SCALE)
Major Ice Ages in the Past Billion Years
Geological evidence leaves little doubt that, during the
past billion years or so, the prevalent condition of macro-
scale climate was one of relative warmth as much as
10°C warmer than now—and almost total absence of
polar ice. This warm condition was, however, punctuated
by at least three major ice ages, each around 10 million
years long and separated by a few hundreds of millions of
years. Beginning roughly 50 million years ago, something
appears to have brought about a gradual cooling. This
cooling trend culminated, about 2 million years ago, in
the arrival of a new major ice age (the Quaternary), char-
acterized by a long sequence of perhaps as many as 20
major glacial-interglacial oscillations, which presumably
continue to grip the world today.
o
15
30
45
60
75
I ~ I ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 1 1
5 8 11 14 0 30 60 90 11 -38 -32 -28 0.5 ~.2 ~.9 -1.6
FAUNAL INDEX ARBOREAL ICE CORE 6180 IN
POLLEN 6180 PLANKTON SHELL
FIGURE 2.3 Comparison of paleoenvironmental indicators of
climatic records for the past 135,000 years (modified from Figure
A.13, U.S. Committee for the Global Atmospheric Research Pro-
gram, 1975~.
39
Glacial and Interglacial Stages of
the Quaternary
Detailed evidence of conditions within the Quaternary
shows that periods of glacial extension (glacials) and
retraction (interglacials) have alternated in a fairly regular
sequence (Kukla, 1961; Broecker and van Donk, 19701.
Figure 2.2 shows climatic records for the past million
years as deduced from the geological and biological rec-
ord. The first section of the figure shows the oxygen-
isotope curve for Pacific deep-sea core V28-238, inter-
preted as reflecting global ice volume (Shackleton and
Opdyke, 19731. The relatively rapid and high-amplitude
fluctuations are taken to indicate sudden deglaciations
and are designated as the terminations I to VII. The
second section shows the calcium carbonate percentage
in equatorial Pacific core RC11-209 (Hays et al., 1969~.
Low values are taken to indicate periods of rapid dissolu-
tion by bottom waters. The third section shows the faunal
index, reflecting changing composition of Caribbean
foraminiferal plankton, calibrated as an estimate of sea-
surface salinity in parts per thousand (Imbrie et al., 1973~.
Glacial periods are marked by the influx of plankton
preferring higher-salinity waters (Prell, 1974~. Note that
each of the three records reflects a climatic fluctuation or
"cycle" averaging about 100,000 years. This is particu-
larly true during the past 450,000 years. Each cycle
started with a short interglacial and ended with an
equally short extreme glacial peak. These extremes rep-
resent only 20 to 30 percent of the total duration of a
typical cycle (length about 100,000 years), and the glacial
itself can usually be subdivided into relatively warm
interstadials and cooler stadials.*
At the peak of the last glacial, 17,000 to 18,000 years
ago, an ice sheet 2 km thick covered the northern and
middle latitudes of North America as far south as New
York, and another sheet in Europe reached as far south as
Hamburg, Berlin, and Warsaw. Smaller ice caps and val-
ley glaciers covered large areas of the Rockies, Alps,
Andes, Hindukush, and many other mountain ranges.
Because the volume of ice on the continents was some
50X106 km3 greater than today (Flint, 1971), the oceans
stood about 100 m below their present level
(Bloom, 1971~. Atmospheric and oceanic circulation, as
reconstructed from available surface characteristics,
greatly differed from present means (Lamb, 1971~. The
mean annual temperatures were lower by about 3°C at the
equator and 10 to 12°C in the midlatitudes of the northern
hemisphere. The departure in the global mean was about
2°C (C~MAP Project Members, 19761.
During glacial maxima, vegetational and faunal zones
in temperate regions were displaced to lower altitudes
and latitudes as compared with their interglacial loca-
* Interstadials are regarded as moderately warm interglacial
periods not as extremely warm as present-day conditions. Simi-
larly, stadials are moderately cold.
OCR for page 40
40
lions. Highly continental climate, dry and with cold win-
ters, characterized Europe and Central Asia. As a result,
tundra and steppe replaced the pre-existing forests (Fren-
zel, 1967~. Greater continentality and desiccation are
similarly indicated for parts of Africa and South America
(Fairbridge, 1972~.
Figure 2.3 shows climatic records for the past 135,000
years: (a) A faunal index reflecting changes in foraminif-
eral plankton in a core west of Ireland. The index is an
estimate of August sea-surface temperature in degrees
Celsius (Sancetta et al., 19731. (b) The percentage of tree
pollen accumulated in a Macedonian lake (Van der
Hammen et al., 19711. High values indicate warmer and
somewhat drier conditions. (c) Oxygen-isotope ratio ex-
pressed as PRO in an ice core at Camp Century, Green-
land. This is interpreted as indicating changing air tem-
peratures over the ice cap (Dansgaard et al., 19711.
(d) Oxygen-isotope ratio in skeletons of planktonic
foraminifera in a Caribbean core, interpreted as changes
in global ice volume. High negative values reflect the
melting of ice containing isotonically light oxygen (Emi-
liani, 1964~.
The interglacial periods were characterized by climate
much like that of the recent past, with similar plant and
animal distributions. These were periods of retracted ice
sheets, high sea levels, and relative warmth.
The amplitude of the climatic variations associated
with the glacial cycles seems remarkably constant. It is
likely that global mean values of atmospheric or oceanic
variables differed little between successive interglacials
(Emiliani, 1973~.
Postglacial Climatic History
Although comprehensive evidence is scarce, the earth is
apparently now in an interglacial period. It began be-
tween 10,000 and 14,000 years ago with general warming
accompanied by the decay of the continental ice sheets.
The climate of this period is characterized by several
marked fluctuations, which appear to have occurred
mostly simultaneously in the northern and southern
hemispheres. This is especially true (Heusser, 1966) for
the drastic variations between 12,000 and 10,000 years
before the present (B.P). The limited data currently
available suggest that the cool and relatively wet climate
of the period was suddenly replaced by a worldwide
mild, even warm, period. Even more dramatic was the
later catastrophic readvance of the ice masses about
10,800 B.P., which killed entire forests in a period of
probably less than a century (Lamb, 1966~.
If we disregard some minor fluctuations during the
recession of the large continental ice sheets, which dis-
appeared completely about 6000 B.C. in Scandinavia and
about 4000 B.C. in northern Canada, the climax of
the postglacial warming was reached between 5000
and 4000 B.C. This was once more a worldwide
phenomenon, where the annual temperatures were
2 to 3°C warmer than today. Even in Alaska, Hey
CHARLES W. STOCKTON
were more than 1°C warmer. This period has been de-
fined as the "postglacial optimum" or "hypsithermal." Its
mild climate, together with the relative dryness in large
areas of North America and the Soviet Union, suggests a
poleward displacement of the subtropical anticyclonic
belt. The arctic sea ice had receded well north of its
present position, but there exists no evidence for a com-
plete disappearance of its central area north of about 80°
latitude. At the same time (and after), the Sahara and the
arid parts of the Near East were considerably more
humid; this means that in now completely barren, arid
areas there was a steppe vegetation produced by occa-
sional severe rainstorms. Some evidence also exists for a
northward extension of the tropical summer rain belt
(Lamb, 19661. During this time, many smaller mountain
glaciers completely disappeared, and the snow line was
about 300 m higher than today. The sea level gradually
rose to its present level but not above it (Shepard and
Curray, 19671; after this date it was mainly controlled by
the mass budgets of Antarctica and Greenland. A cool
episode occurred between 4000 and 3000 B.C., signaled
by He re-formation or expansion of mountain glaciers,
followed by renewed warming (Figure 2.4~. Cooling and
glacial advance took place again between 1400 and 500
B.C. The subsequent warming trend ended before A.D.
600, at least in western North America, where glaciers
again advanced, culminating about A.D. 900.
The climatic record becomes increasingly detailed and
reliable beginning about 1000 years ago, mainly because
of He availability of historical accounts in at least the
Norm Atlantic sector. The mild conditions of the postgla-
cial optimum were nearly reached once more during He
early Middle Ages, culminating about A.D. 1200 when ice
conditions around Iceland and Greenland were much less
severe than today (Figure 2.51. Annual mean tempera-
tures in southern Greenland must have been 2 to 4°C
above present averages (Table 2.2~. Oxygen-isotope ratios
in the Greenland icecap (Figure 2.6) confirm the warmer
climate there during tliis period. In England (Table 2.2
\ TREE GROWTH FLUCTUATIONS AT UPPER TREELINE A
| \` / ~ i Mean /~\ / ~,1:
....
3000
Expansion t
Contraction 1
~ I 1 1 1 1 · · I ~ I I · I I I 1 1 1 : I ~ I I ! 1 1 1 1 ~ ~ I I ~ ~ 1 1 ~ I I I I ~ ~ I 1' 1
2000 1000 B.C. 1 A.D. 1000 1971
IOLOCENE GLACIER
F LUCTUATIONS
FIGURE 2.4 Periods of low tree growth and glacial advance
indicating cool periods in late Holocene time. Tree-ring data are
from California; glacial data are mainly from northern hemi-
sphere (Demon and Karlen, 1973; LaMarche, 1974).
OCR for page 41
Interpretation of Past Climatic Variability from Paleoenvironmental Indicators
TABLE 2.2 Average Climatic Conditions over England and Wales (after Mitchell et al., 1975)a
Mean Temperatures (°C)
41
Annual Annual
Summer Winter rain- evapo-
Dates (July- (Dec.- fall ration
(approx.) Epoch Aug.) Feb.) Annual (mm) (mm)
A.D. 1901 - 1950 Recent 15.8 4.2 9.4 932 497
A.D. 155~1700 Little Ice Age 15.3 3.2 8.8 867 467
A.D. 115~1300 Little Optimum 16.3 4.2 10.2 960 517
90(~50 B.C. Subatlantic 15.1 4.7 9.3 96() 979 482
a After Lamb ( 1966).
and Figure 2.7), mean annual temperatures were about
1°C above recent normals. In contrast to the situation of
A.D. 500 to 900, when pronounced cooling of western
Norm America is apparently not reflected in the Euro-
pean record (Demon and Karlen, 1973), variations of the
climate of these parts of the nor~em hemisphere seem
remarkably similar during the past 1000 years.
Northern hemisphere data suggest that the period after
A.D. 1300 was one of widespread change to cooler condi-
tions; this period has been termed the historic "Little Ice
Age." Glaciers advanced in many parts of the world, but
this brief episode cannot be compared with the great
glacials lasting several tens of thousands of years.
Worldwide glacial recession accompanied global temper-
ature rises after 1895.
Kutzbach and Bryson (1975) present plots of frequency
versus percentage of variation in different climatic rec-
ords.-One of these is reproduced as Figure 2.8. From this
diagram it is apparent that, based on well-dated tempera-
ture records from Central England and Iceland and an
isotope record from Greenland, the data in the inter-
mediate zone (500 to 1000 years) show considerable per-
+0.8
+0.4
4"
~ r,
~ v
—0 4
-0.8
11
900 1100 1300 1500 1700 1900
Year (A.D.)
FIGURE 2.5 Departure of mean annual temperature in Iceland
inferred from extent of sea ice during the past 1000 years.
Departures from values for the period of meteorological record
(Bargthorsson, 1962, as presented by Bryson, 1974~.
sistence and the data at the higher frequencies (10 to 100
years) are nearly random. The authors point out the
shortcomings of the study, including We lack of replica-
tions from other records and the gap between periods of
500 and 1000 years, where spectral estimates are less reli-
able in a statistical sense. They also stress the need for
defining details of the climatic spectrum in the inter-
mediate range of 500 to 1000 years. Climatic fluctuations
at these time scales can have great impact on water
resources, yet this is the least known portion of We
spectrum.
REGIONAL EVALUATION OF PALEO-
HYDROLOGIC PHENOMENA
It is now quite apparent that, although adequate data may
be available for documentation of long-term climatic fluc-
tuations on a global or hemispherical scale, the amount
available for specific regions can be quite limited. How-
ever, recent studies have shown the type of regional
hydrologic information that can be obtained from paleo-
climatic indicators.
Within the framework of the national water demand
criteria, we decided to focus on two distinct regions—the
Southwest and the Northeast for concentrated studies.
-28 ~
20e_3°~{ :] Tar ; i;
,. . . . . .
700 500 300 100
Age, years before present
FIGURE 2.6 Oxygen-isotope ratios in ice from the Camp Cen-
tury Core, Greenland. Low values (top of graph) indicate low
temperatures. Vertical scale is relative departure of i80 constant
compared with a standard (Dansgaard et al., 1971).
OCR for page 42
42
~ 10.5
car .
U=J ID 10.0
4-= 9.5 .
c: ~ _ -
0.7
E
0.1 r
o.o 1 ~
_ ~
,` ,' N. hemisphere (40° - 70°N)
t `` (Mitchell, 1961 ) Oaf\
/ .
- 1 i+30 Q°-
Central England / c ~
`` ^` ~ ' \Ct' ] +5 ~ :t
\_—_'
MEAN ANNUAL TEMPERATURE
~1
TREE GROWTH AT UPPER TREELINE -I
_ i , . , , , . , . i
1000 1500 1960
Year (A.D.)
FIGURE 2.7 Estimated temperatures in Central England since
A.D. 900 compared with tree-ring variations in California indi-
eating general cooling in northern hemisphere between A.D.
1300 and 1900 (LaMarche, 1974).
These areas are, respectively, one of the fastest growing
areas of population and the area of present greatest popu-
lation. We have attempted to accumulate long-term in-
formation concerning climate for these two regions.
Southwestern United States
Stockton (1975) and Stockton and {acoby (1976) have
attempted to show how long-term total annual runoff
~ 40
a,
c,
o
~ 30
Q
. _
C1 Ad
N
o
10
wh ite
noise
~ CENTRAL ENGLAND. botanical record
11 .
',i
; ;
- i ;
1 1
.. .
1 1
· ~
/ CENTRAL ENGLAND,
/ historical record
/ I CE LAND,
/ / historical record
/ / CENTRAL ENGLAND
/ // instrumental record
·~ / G R E E N LA N D, ~ 1 80 record
· ~~\ ~7 ~._~,/f white noise continuum
`~ W<~_~
o
.01 .02 .03 .04 .05
f ~ (cycles per year)
30 20
P (years)
1000 100 50
FIGURE 2.8 Composite variance spectrum of temperature on
time scales of 10 to 103 years derived from instrumental, histori-
eal, botanical, and oxygen-isotope records (after Kutzbaeh and
Bryson, 1975~.
CHARLES W. STOCKTON
records can be reconstructed utilizing tree-ring data.
Their study provides detailed reconstructions of total
annual runoff for several subbasins within the Upper
Colorado River Basin. In addition, it suggests that, within
a larger basin, the tributary systems can show varying
degrees of persistence. For example, Figure 2.9 shows
the variance spectra computed for three tributary rivers
within the Upper Colorado River Basin. The drainage
areas of the three are not greatly different; however, the
degree of persistence in the reconstructed records differs
substantially. The Green River shows a greater tendency
for low-frequency variation than either the Colorado
above Cisco or the San Juan at Bluff. It appears that the
tendency for long-term climatic persistence may be
greater in the Green River Basin than for either the
Colorado above Cisco or the San Juan above Bluff. This
AU T O S P E C ~ R U M
Lags = 56 deg. f reedom = 14
10—
o
I!) 1.-
o
.:/
. .
. .
DECO LO R ~ DO at C I SCO
A....
1~, 1"~\ ·' ·: ~ (\\
I'll ~ '~j I ..
I ~
19.8 76
.:
i '_' , I ~ \ ~
; · . ·
. .
. .
. .
. .
a' ~ 11~\
! ! — ~
S A N J U A N ,! ~ `;
Y""1 .
A'
.
4.0 2.6 2.0
PERIOD (years)
FIGURE 2.9 Comparison of the sample autospeetral functions
for the long-term reconstructed runoff records for the Green
River at Green River, Utah; the Colorado River at Ciseo, Utah
(Colorado mainstem); and the San Juan River at Bluff, Utah
(Stockton, 1975~.
OCR for page 43
Interpretation of Past Climatic Variability from Paleoenvironmental Indicators
UNFILTERED
~ to—
TIC C air ~ ?
o
z
Ct
J
IS
Z
O O ~ T '
1600
—rid I ~ ~ ~ .
1800 1900
f ILTEREO
' I ' ' am'- 1 '' ' ' ' ~ ' ' ' ' 1 ' ' ' ' ~ ' ' ' ' 1 ' · ' ' ~ ~
1700 1800 1900
GREEtd RlvER AT GREEN Rl\/ER. UTAH
3'0
c
— :: ~~j ~~` \!~: ~ `~U'-t
. . ,
1600 1 700 1800
UNFI LTERE D
FILTERED
~ . ~ . .
1 900
~ lo
~ o ~
to
~5
~ ~ —~ ~ · ~ _ ~
1600
· ' 1 — -
1700 1800 1900
COLORADO RIVER At LEE FERRY. ARIZONA
FIGURE 2.10 Reconstructed hydrographs for total annual runoff for the Green River at Green River, Utah, and the
Colorado River at Lee Ferry, Arizona. In each case, the lower graph is the same data but with Me high-frequency
components (those with a frequency greater than 10 years) removed (Stockton, 1975).
suggests the need for evaluation of paleoclimatic varia-
tion in rather limited areas.
The long-term hydrograph for the Upper Colorado
River Basin as a whole does not exhibit the degree of
long-term persistence that is found in the Green River
reconstruction (Figure 2.10~. However, analysis of the
persistence does show that the series is significantly
different from random and is best modeled by a mixed
autoregressive-moving average scheme. In addition, the
reconstructed hydrograph shows that the early part of the
twentieth century was characterized by a period of
anomalously high sustained flow, the longest in the entire
450-year reconstruction. The gauged record alone would
not reveal this fact, so anyone depending solely on the
gauged record would obtain inflated estimates of mean
annual flow and variance. The reconstructed hydrograph
is consistent with the secular variation shown in tempera-
ture trends as illustrated by LaMarche (1974) (see Figure
2.7~. Also, the degree of persistence seems to be of the
same magnitude as that suggested by Kutzbach and Bry-
son (1975) for records of similar length (Figure 2.8~.
LaMarche (1973) studied tree-line changes in the
White Mountains of east-central California and found
43
that, between A.D. 1300 and 1600, abrupt climatic change
resulted in lowering of the timberline some 70 m. He
attributed this to an apparent climatic change to much
colder summers or to fairly cold summers and drier
springs, autumns, and winters. Judging from later work,
this condition apparently lasted at least up to the early
1900's. This example serves to illustrate two points. First,
that apparently climatic change can occur over a rela-
tively short time period (hundreds of years) and, second,
that there is additional evidence for the large flow
anomalies as reconstructed in the hydrograph for the
Upper Colorado River (Figure 2.10~.
From the foregoing evidence, it appears that within the
southwestern United States climatic change has occurred
during the past 500 or so years, that it has occurred over a
fairly short time span, and that it has been reflected in the
annual runoff, at least for the Upper Colorado River
Basin.
Northeastern United States
The only currently existing detailed regional analysis
using historical data for the northeastern United States is
OCR for page 44
44
FIGURE 2.11 Annual temperatures and
annual precipitation totals for the eastern
seaboard of the United States for the period
173~1967—a representative, reconstructed
synthetic series centered on Philadelphia
(after Landsberg et al., 1968~.
~ 45.0
', 44 n
CHARLES W. STOCKTON
58r
5'S
55
be 54
52
51
50
49
15
- 114
1740 1760 1780 1800 1820 1840 1860 1880 1900 1920 1940 1960
Year
47.0
46.C
13
1 1
..
10
120
_ 115
38.0 _
37.OI ~ 1 1 t 1_ ~ 1 1
1730 1750 1775 1790 1810
by Landsberg et al. (1968~. They used historical climatic
records along the eastern seaboard, centered on Philadel-
phia, and a least-squares technique to reconstruct provi-
sionally a 230-year record of temperature and precipita-
tion (Figure 2.11~. They noted a trend in the annual
temperature data principally "caused by the lack of cold
years since the turn of the 20th century." The trend is also
confirmed by a variance spectrum analysis. No mention is
made of the annual precipitation series, but Figure 2.11
illustrates the anomalous wet period extending from
about 1830 to 1880.
Although considerable paleoclimatic work has been
done in the northeastern section of the United States,
much of it is still qualitative. Many of the results are in the
form of pollen data and tree-ring data.
Using pollen data, Webb and Bryson (1972) presented
quantification of July mean temperature, summer precipi-
tation, and precipitation minus potential evaporation for
an area in north-central United States. This investigation
was later expanded by Bryson (1974) to cover a larger
area. These reconstructions cover the past 15,000 years
and show an extreme temperature drop (as much as 8°C)
at about 10,000 years B.P.
Blasing (1975) shows reconstruction of climatic types
on a national scale and indicates that certain climatic
anomaly patterns have been more prevalent during the
previous two centuries than in this one. However, his
conclusions are based on reconstructions of large-scale
~ ~ _-
1830 1850 1870 1890 1910 1930 1950 19 70
Year
atmospheric circulation patterns, derived from tree-ring
data in western North America.
CONCLUSIONS
The methods of quantitative paleoclimatology enable us
to increase our knowledge of many details of climatic
history. By increasing our knowledge of past climate, we
gain a valuable perspective to our view of climate of the
present and future.
There exist, at present, isolated time series that indeed
suggest important climatic changes at all time scales.
However, the job of transforming this information into
spatial maps so that we can study patterns of change with
adequate spatial detail is just beginning. Until these maps
exist, we cannot accurately characterize periods as warm
or cool, wet or dry except at specific locations. This will
not occur without concentrated research efforts and con-
siderable support in the future.
RE FE RE NC E S
Blasing, T. J. (1975). Methods for analyzing climatic variations in
the North Pacific sector and western North America for the last
few centuries. Ph.D. dissertation, University of Wisconsin,
Madison.
OCR for page 45
Interpretation of Past Climatic Variability from Paleoenvironmental Indicators
Bloom, A. L. (1971). Glacial-eustatic and isostatic controls of sea
level since the last glaciation, in The Late Cenozoic Glacial
Ages, K. K. Turekian, ea., Yale U. Press, New Haven, Conn.,
pp. 355-379.
Broecker, W. S., and J. van Donk (1970). Insolation changes, ice
volume, and the ~80 record in deep sea cores, Rev. Geophys.
Space Phys. 8, 169.
Bryson, R. A. (1974). A perspective on climatic change, Science
184, 753.
Chin, W. Q., and V. Yevjevich (1974). Almost-periodic, stochastic
process of long-term climatic changes, Colorado State U. Hy-
drology Papers, No. 65.
CHEAP Project Members (1976). The surface ofthe ice-age earth,
Science 191, 1131.
Dansgaard, W., S. T. Johnsen, H. B. Clausen, and C. C. Langway,
Jr. (1971). The Late Cenozoic Glacial Ages, K. K. Turekian,
ea., Yale U. Press, New Haven, Conn., pp. 267-306.
Denton, G. H., and W. Karlen (1973~. Holocene climatic changes,
their pattern and possible cause, Quaternary Res. 3, 155.
Emiliani, C. (1964~. Paleotemperature analysis of Caribbean
cores P6304-8 and P6304-9 and a generalized temperature
curve for the past 465,000 years, J. Geol 74, 109.
Emiliani, C. (1973~. Quaternary hypsithermals, Quaternary Res.
3, 270.
Fairbridge, R. W. (1972). Climatology of a glacial cycle, Quater-
nary Res. 2, 283.
Flint, R. F. (1971~. Glacial and Quaternary Geology, John Wiley
and Sons, New York.
Frenzel, B. (1967). Die Klimaschwanhungen des Eiszeitalters,
Brunswick.
Fritts, H. C., T. J. Biasing, B. P. Hayden, and J. E. Kutzbach
(1971~. Multivariate techniques for specifying tree-growth and
climate relationships and for reconstructing anomalies in
paleoclimates, J. Appl. Meteorol. 10, 845.
Gates, W. L. (1976). Modeling the ice-age climate, Science 191,
1138.
Hays, J. D., T. Saito, N. D. Opdyke, and L. H. Burckle (1969).
Pliocene-Pleistocene sediments of the equatorial Pacific, their
paleomagnetic, biostratigraphic and climatic record, Geol.
Soc. Am. Bull. 80, 1481.
Heusser, C. J. (1966~. Polar hemispheric correlation: palynologi-
cal evidence from Chile and the Pacific Northwest of America,
in World Climate,from 8000 to 0 B.C., J. S. Sawyer, ea., Royal
Meteorol. Soc., London, pp. 124-141.
Imbrie, J., J. van Donk, and N. G. Kipp (1973~. Paleoclimatic
investigation of a late Pleistocene Caribbean deep-sea core:
comparison of isotopic and faunal methods, Quaternary Res. 3,
10.
Kukla, J. (1961). Quaternary sedimentation cycle, Survey of
Czechoslovak Quaternary, Institut Geologiczny, Prace, 34,
pp. 145-154, Warszawa.
Kutzbach, J. E. (1975). Diagnostic studies of past climates, in The
Physical Basis of Climate and Climate Modeling, CARP Publi-
cation Series No.16, WMO, Geneva, Switzerland, pp. 119-126.
Kutzbach, J. E., and R. A. Bryson (1975~. Variance spectrum of
Holocene climatic fluctuations in the North Atlantic sector, in
Proceedings of the wMo/IAMAP symposium on long-term
climatic fluctuations, WMO, No. 421, pp. 97-104.
LaMarche, V. C. (19731. Holocene climatic variations inferred
45
from beeline fluctuations in the White Mountains, California,
Quaternary Res. 3, 623.
LaMarche, V. C., Jr. (1974~. Paleoclimatic inferences from long
tree-ring records, Science 183, 1043.
Lamb, H. H. (1966~. The Changing Climate, Methuen, London.
Lamb, H. H. (1969~. Climatic fluctuations, in World Survey of
Climatology 2~5), 17~249.
Lamb, H. H. (1971~. Climates and circulation regimes developed
over the northern hemisphere during and since the last ice
age, Paleogeog. Paleoclimatol. Paleoecol. 9, 125.
Landsberg, H. E., C. S. Yu, and L. Huang (1968~. Preliminary
reconstruction of a long time series of climatic data for the
eastern United States, Univ. Maryland Inst. Fluid Dynamics,
Appl. Math. Tech. Note BN-571, 30 pp.
Mitchell, S. M., V. C. LaMarche, E. S. Epstein, P. R. Julian, M. F.
Meir, and B. J. Kukla (1975~. Chap.2, Variability of the climate
of the natural troposphere, in Department of Transportation
Climatic Impact Assessment Program, Monograph 4.
O Connell, P. E., and J. R. Wallis (1973~. Choice of generating
mechanism in synthetic hydrology with inadequate data, Int.
Assoc. Hydrol. Sci., Madrid Symp., June.
Prell, W. L. (1974~. Late Pleistocene faunal and temperature
patterns of the Columbia Basin, Caribbean Sea, Geol. Soc. Am.
Special Paper.
Sancetta, C., J. Imbrie, and N. G. Kipp (1973). Climatic record of
the past 130,000 years in North Atlantic deep sea core V23-82:
correlation with the terrestrial record, Quarternary Res. 3, 110.
Shackleton, N. J., and N. D. Opdyke (1973). Oxygen isotope and
paleomagnetic stratigraphy of equatorial Pacific core V28-238:
oxygen isotope temperatures and ice volumes on a 105 and 106
year scale, Quaternary Res. 3, 39.
Shepard, F. P., and J. R. Curray (1967). Carbon-14 determina-
tions of sea level changes in stable areas, in Progress in
Oceanography, Pergamon Press, New York.
Stockton, C. W. (1975~. Long-term streamflow reconstruction in
the upper Colorado River basin using tree rings, in Colorado
River Basin Modeling Studies, proceedings of a seminar held
at Utah State University, Logan, Utah, July 1~18, 1975, C. G.
Clyde, D. H. Falkenborg, and J. P. Riley, eds., Utah Water
Research Laboratory, College of Engineering, Utah State U.,
Logan, Utah, pp. 401~41.
Stockton, C. W., and G. C. Jacoby (1976). Long-term surface
water supply and streamflow levels in the upper Colorado
River basin, Lake Powell Research Project Bull. (in press).
U.S. Committee for the Global Atmospheric Research Program
(1975). Understanding Climatic Change: A Program for Ac-
tion, National Academy of Sciences, Washington, D.C.
Van der Hammen, T., T. A. Wijmstra, and W. H. Zagwijn (1971~.
The floral record of the late Cenozoic of Europe, in The Late
Cenozoic Glacial Ages, K. K. Turekian, ea., Yale U. Press,
New Haven, Conn., pp. 391-424.
Wallis, J. R., and P. E. O Connell (1973). Firm reservoir yield;
how reliable are historic hydrological records, Hydrolog. Sci.
Bull. 18, 347.
Webb, T., and R. A. Bryson (1972~. Late- and post-glacial climatic
change in the northern midwest, USA: quantitative estimates
derived from fossil pollen spectra by multivariate statistical
analysis, Quaternary Res. 2, 70.
OCR for page 46
Representative terms from entire chapter:
colorado river
