Please note that this article was part of a talk given during a Royal Society Discussion Meeting on the Earth's Climate and Variability of the Sun Over Recent Millennia, held in London on the 15th and 16th of February 1989.
'Climate' is often used by historians to explain phenomena for which they cannot otherwise account. Accordingly, much of what has been written about climatic effects and climatic change must be read with extreme skepticism. Even though a disturbance may be obvious in the archaeological record, and it may be synchronous with a climatic event, a cause and effect relationship should be demonstrated before one can say with any degree of confidence that the evidence is secure. Only when a number of separate lines of investigation agree on the same thing are we safe in positing true climatic 'effect' or 'change.' This paper will focus on several instances in Mediterranean and Aegean archaeology where more or less satisfactory evidence for climatic change may be sought among a number of disciplines.
The title of this paper as originally assigned was "Archaeological
Evidence for Climatic Change." The qualifying phrase "and Non-Evidence" was
my addition, reflecting not just routine academic sophistry but a deep-rooted
suspicion that we often attribute to "climate" phenomena in the archaeological
record that we cannot otherwise explain. In layman's language climate becomes
the historian's cop-out. One example of the lengths to which some writers
will go should be sufficient: "Climate in Greco-Roman History," (Eddy 1980)
in which the author sets out to match two graphs, each one an inverted V, one
of the thickness of sequoia tree-rings in California, the second of
dedications and building starts in Africa between A.D. 138-244. The logic
seems to be, after the two V's are superimposed, that since the latitudes of
California and the Mediterranean are the same, so should be the climate. Six
data points in two thousand years allegedly prove this. So much for non-
evidence, but where might we look for something more satisfactory?
First, I have no doubt that climate does indeed change, both over the
long and short term, although, after recent reports announcing that there is
less than total agreement among the climatologists that the 'Little Ice Age'
ever existed and that there is doubt as to whether there is really a so-called
greenhouse effect after all, I wonder how the archaeologist is supposed to
provide information about events whose very existence has been called into
question.
Second, I have no doubt that we are in one way or another affected by
climatic changes, particularly from single catastrophic events. The severe
drought in the mid-western United States last summer and the billions of
dollars' worth of damage that resulted, all duly recorded by the press, is a
dramatic example of the degree to which we are at the mercy of the climate,
and the people who lost their livestock and had the banks foreclose on their
mortgages do not need to be informed by a Royal Society Discussion Meeting
that indeed 1988 was a bad year. But again I ask whether an archaeologist
digging 2000 years from now in the remains of a mid-western U.S. town and
coming down to the depressed 1988 level would be able a) to realize that there
had indeed been a horrendous drought, and b) to discriminate between the
effects [Page 646]
of the drought and some adverse aspect of, let us say, Reagan
economics. Other, less-dramatic climate changes have less obvious effects.
Here the problem is simply one of recognition. There is also the question of cause and effect. The same hypothetical
archaeologist continues to dig down to the 1930's level and finds the great
'dust bowl.' Does this have anything to do with the stock-market crash of
1929 or were the two events simply coincident? At the first International Conference on Climate and History at the
Climatic Research Unit, University of East Anglia, Norwich, in 1979, the point
was made repeatedly in both the printed review papers and the discussions
which went on all week that it is not so much climate that is important but
rather man's perceptions of it. Colleagues who work with the English Manor
House records used as an example a society where conventional father-to-son
wisdom is that you have to save one-third of your harvest in order to survive
until the next harvest and still have enough grain for the next year's
planting. Then a series of mild years comes along with abundant rain and sun,
and very quickly the old verities are forgotten. One-fourth of the harvest,
let us say, is saved, and the rest is sold for immediate profit. Then the
climate reverts to normal (except that people have forgotten what "normal"
is: the unofficial Norwich after-dinner estimate of how long our memories
work was about four years maximum), and people starve. Has the climate really
"changed?" Not at all. Human memory has once again played us false. How,
I wonder, is the archaeologist, measuring the human record as it bumbles along
in this fashion, going to be able to match it in any meaningful way with the
estimates of the palynologists (who, by the way, should have been represented
at this discussion meeting), the geomorphologists, and the meteorologists,
even presuming that nobody has made any faulty calculations? One final example: and here I speak as a member of our local volunteer
fire brigade. In October 1981 five inches (13cm) of rain in five hours sent the
Ithaca (New York) creeks into full flood, and for the next three days we
pumped out basements of houses along the banks of Fall Creek in which an
average of six feet (2m) of water had collected. As we pumped, we noticed that
there was not a single 19th century house in the lot. The old houses were up
on top of the bluff. Moral: obviously in the last century the creek had
flooded often enough so that any right-thinking house-builder built on high
and dry ground. But somewhere around the First World War enough time went by
without a flood to entice would-be money makers into subdividing the land
along the stream banks into house-lots and then building the houses which the
fire brigade had to pump. A glance at the flood-scoured walls of the gorge
through which the creek flows would have told any moderately competent
geologist that this was no place to build houses, either then or now, but
avarice won out over common sense. Antoine Meillet's great comment on his study of Indo-European
linguistics was "La linguistique est une système ou tout se tient," that is
to say: linguistics are a system in which everything hangs together, and so
it should be with archaeology and climate. Since much of my own work is with
tree-rings (Kuniholm and Striker 1983, 1987), I am continually reminded that
there are indeed periodicities of one kind or another in all the records with
which we work, but all too often our conclusions about climatic forcing
mechanisms or stimuli are drawn based on a single line of evidence without
independent corroboration and should therefore be treated with extreme
skepticism. Moreover, I find the vast majority of speculations on the
relationship between archaeology and climate to be precisely that: speculation
based often on inadequate or unverifiable data. Pollen diagrams down to the
end of the Upper Palaeolithic, for example, may be significant indicators for
climatic change, but after intensive agricultural
[Page 647]
exploitation begins in the Neolithic there is always the danger that a certain element
of anthropogenic 'noise' has been introduced into the palynological record. Moreover, how
should we expect various bodies of evidence to interact in the archaeological
record? There will be lags in some cases, parallel change in others,
divergent change in yet others. At best the possibilities are precarious.
A model might be the careful, cautious, exhaustive work of Emmanuel Le Roy
Ladurie whose Times of Feast, Times of Famine: A History of Climate Since the
Year 1000, first published in 1971, is now again in print and which makes even
better reading the second time around. Nothing is too humble for Le Roy's
attention: wine harvest dates, advances and retreats of glaciers, grain
prices, documentary evidence for crop yields, taxation rates, tree-ring
widths, etc., and his conclusions are equally humble: something indeed seems
to be going on with the climate, but he refrains from forcing conclusions upon
us. Twenty years ago the British palaeoclimatologist H. H. Lamb, while
commenting on the climatic changes that may have put an abrupt end to Bronze
Age civilization in Greece and in other areas of marginal agricultural
productivity (Carpenter 1966, and see below), issued a challenge to research-
ers in the Mediterranean and in the Near East (Lamb 1968):
A comprehensive and cautious survey of a wide variety of evidence is by
Butzer (1958). He finds it difficult to correlate evidence for short, minor
rainfall ameliorations with anything cultural. He laments the lack of
systematic study and systematic observations in both Anatolia and Iran. His
final conclusion is that although he believes small-scale variations of
climate occurred continually during historical times, he nevertheless strongly
negates any overall climatic change within the last 2500 years. More recent,
and indeed systematic, work (Van Zeist and Wright 1963, Van Zeist 1967, Van
Zeist et al., 1968) bears this out. Van Zeist concludes that after about 5500
B.P. the climate may have shown minor fluctuations but no major changes.
Despite millennia of written records, Mesopotamia is a frustrating area
with which to deal, partly because the evidence is very unevenly distributed,
partly because there has been hardly any research done on its ecological
microstructures (Nissen 1988). Plant cultivation must have been always
precarious at best with the bulk of the water coming through irrigation
channels [Page 648]
rather than from the wretched 200mm. per year rainfall of southern
Mesopotamia (Van Zeist 1969). The most entertaining single item for a
climatic event is Sir Leonard Woolley's finding of a water-borne layer, eight
to eleven feet (2-3m) thick, of clean mud and silt at Ur between 'Ubaid I and II
levels. Mrs. Woolley took one look at it and said, "Well, of course, it's the
Flood!" (Woolley 1954). Spoilsports (Lees & Falcon 1952) have cast doubt on
this, saying that the date is not contemporary with other early floods, and
that the major cause is tectonic subsidence rather than the simultaneous
rising of the Euphrates and the Tigris which can set southern Mesopotamia
awash (and see now Nützel 1976). Whatever the cause or combination of causes
(downward sea-level change, tectonic uplift, silting-in of the floodplain, or
climatic amelioration), H. J. Nissen's new (1988) book on Mesopotamian
prehistory concludes that southern Mesopotamia at last became inhabitable in
the 4th millennium B.C. only because the land was finally dry enough to farm
and raise animals. Another excellent but cautious recent paper (Neumann and
Parpola 1987) lays out climatological, meteorological, and Assyriological data
for a significant climatic change at the end of the Bronze Age (see also
Brinkman 1984).
The major review of Nile flooding and its causes is by Lyons (1905),
updated by Bell (1975) with an historical commentary particularly regarding
the low flood levels which might have helped bring on the so-called Little
Dark Age at the end of the Middle Kingdom sometime after ca. 1768 B.C. Bell
notes that this took place at a time when there was unusual uncertainty about
the proper order of succession to the Kingship, a general increase in
governmental instability, and numerous short reigns. She does not claim that
low water caused the concurrent political upheaval, but poor crops certainly
did not help matters. What bothers me about the evidence from the Nile is the
apparently deliberate falsification of the Nile records for revenue purposes
(Lyons 1905) which certainly should cast doubt on their utility as a reliable
source for climatic change unless supported by external evidence.
The most comprehensive recent study I have seen is by Horowitz (1952).
Yet his climatic evidence boils down to pollen analysis from a grand total of
two bore holes, one in Lake Hula and one in the Mediterranean, with one
uncalibrated radiocarbon date for each. More evidence is needed here. A
recent paper by Koucky (1987) on the Roman frontier in central Jordan goes at
long length into 570 year cycles of drought and plenty, and indeed one of his
graphs is quite impressive, but I am suspicious of his chronology, and his
references, albeit somewhat more recent than Eddy's, do not inspire
confidence.
Careful, competent work (they call their observations on climate
"speculations") is being done by Van Zeist and others in the Gröningen group,
but a major portion of their work (1975) concerns only prehistory for which
we have no corroborating evidence. In a limited study in southeastern Turkey
(1968) their pollen diagrams from three different lakes show that the climate
has not changed noticeably during the last 3000 to 4000 years.[Page 649] Clive Foss's (1979) explanations for the silting-in of the harbor at
Ephesus dodge between deforestation and possibly (my emphasis) climatic
change. In his own words: "the evidence is scattered and ambiguous, and
subject to much controversy and variety of interpretation...A theory of
climatic change, if it could be developed and maintained, could do much to
explain the decline of the ancient world and the conditions which prevailed
in the early Byzantine period." Last month David Whitehouse, whose skepticism about the lack of evidence
for climatic change at the end of the Roman period has been in print for some
time (Hodges and Whitehouse 1983), was kind enough to show me part of an
expanded manuscript on which he is now working concerning the apparent absence
of climatic change in Roman and early post-Roman times. After considerable
thought he concludes that the evidence is disappointing: the written records
of the River Tiber floods neither support nor contradict the hypothesis of a
Dark Age drought. A second class of evidence, Vita-Finzi's (1969) so-called
Younger Fill, can be shown, says Whitehouse upon reconsideration, to be
equally explicable without climatic change. A useful new book, Beyond the Acropolis: A Rural Greek Past (Van Andel
and Runnels 1987) has just appeared, with commentary on the Argolid, including
its climate, from the earliest times to the present. Speculations on droughts
in Attica or at least changes in the water levels in the 8th, 4th, and 2nd
centuries B.C. as attested by the filling in of Attic wells and the digging
of other, deeper wells have been published by John Camp (1979, 1982, 1984).
One obvious criticism of Camp's work is that wells go dry for a variety of
reasons besides drought, including heavy consumption of water by a large
population, but Camp is also able to bring to bear both epigraphical and
textual evidence for drought, shortage, and famine datable to specific years,
plus the need for Athens to import large amounts of grain from
abroad. We have also the Sanctuary of Zeus Ombrios or Rainy Zeus on the Top of
Mt. Hymettus (Langdon 1976), a rural sanctuary to which local farmers brought
offerings in inverse proportion and quality to the known prosperity of Athens.
When Athens was prosperous and importing goods from all over the Mediterranean
world, the Sanctuary of Rainy Zeus languished; but when Athens fell upon hard
times and was forced to patronize the local farmers, the sanctuary flourished.
In isolation this does not seem like very hard evidence, but taken together
with Camp's wells, and the inscriptions, and the literary texts, it seems to
me a promising beginning. Another useful new work, fulfilling a number of Lamb's desiderata, is
Peter Garnsey's Famine and Food Supply in the Graeco-Roman World: Responses
to Risk and Crisis (1988) which I received just last week and have not yet
been able to read with care. His commentary on the Eleusis First Fruits
inscription (Inscriptiones Graecae II2 1672), concerning the harvest of
329-328 B.C. which yielded enough wheat and barley to feed only some five-eighths of
the Attic population is both fascinating and frustrating because we simply do not have that kind
of detailed documentation for other years besides 329/8. Would that we had
an absolute Mediterranean tree-ring chronology in place for those centuries
in order to test these snippets of evidence, but for the moment we have only
disconnected dendrochronological sequences. Now that the Thesaurus Linguae Graecae is largely complete, containing
some 60,000,000 words from almost 3000 authors on Compact Disk, plus the
20,000 or so Attic inscriptions we have been
[Page 650]
adding at Cornell, we are at last in a position to begin part of the textual searching,
at least for the Greek world, of which Lamb spoke. Although some of the literary references
are no doubt metaphorical, others are probably every bit as literal as the 4th
century lead tablet from Dodona in which a supplicant asks the oracle: "Is the
present severe winter due to the impiety in the city?" (Parke, 1967) And of
course the cuneiform and hieroglyphic evidence needs to be searched as
well. After all these remarks about other people's work, here are a few
speculations of my own, arrived at during the last fifteen years' work of
building tree-ring chronologies in the northeast corner of the Mediterranean,
which I think might represent evidence rather than non-evidence for climatic
change.
The first is a typical short-term event. In 1874 in the Province of
Ankara, District of Keskin, a drought occurred of such devastating proportions
that 81% of the cattle and 97% of the sheep died. Of the population of
52,000, some 7,000 moved out of the district and 20,000 died (Christiansen-
Weniger and Tosun 1939). The traveller C. Naumann reported that in the
Provinces of Kastamonu, Ankara, and Kayseri 150,000 people and 100,000 head
of livestock (40% of all herds) died. Hunger and sickness through the 1873-1874
winter killed another 100,000 people (Naumann 1893). Plots of tree-ring growth for eleven sites in or near the area of the
reported catastrophe show sub-normal growth for 1873 and 1874, following four
other years of significantly sub-normal growth, as follows:
"Our knowledge of this three-thousand-year climatic sequence in the
middle and northern latitudes of Europe has been built up partly from
pollen analysis of numerous bog and lake site deposits and partly from
analysis of documentary records, especially those which reveal the
incidence of extreme warm or cold, wet or dry seasons. By comparison, the
Mediterranean is a strangely neglected region despite the many ancient
cultures there and the wealth of literature that has survived. Surely, we
need not remain for ever ignorant of which years in classical times had
unusually long, or short, dry seasons or of which winters were striking
for warmth or coldness. What seems most important, if we are to get at
the truth of these matters, is to encourage many more historians and
archaeologists to dig out the documentary and other sorts of evidence.
Every report which is specific as to year and place, giving or implying
the general character of a particular summer or winter, or the dates of
notable rains or floods at any season, is important. One would imagine
that Greek or Roman horticulturists must have recorded the actual dates of
the first and last rains. Any compilation of such reports from anywhere
and any period in the Mediterranean and Near East is badly needed."
Let us look around the Mediterranean now to see what, if anything, may
be said about archaeological evidence for climatic change:
Iran
Mesopotamia
Egypt
The Levant
Turkey
The Roman World
Greece
Example 1. Short Term Events: The Great Drought of 1873-1874 in
Turkey