Reid A. Bryson and Christine Padoch

On the Climates of History The development of objective, quantitative evidence of how climates or climata, and the associated biota, have changed significantly (even during postglacial and historical times) has expanded the possibility of the rational inclusion of the climatic factor in the study of history. Climatic variation has produced variation in both the quantitative and qualitative character of the economic base of cultures, nations, and societies. This new recognition is not a revival of environmental determinism; it implies neither that all environmental changes have a climatic cause, nor that all cultural changes have an environmental cause, and it does not rest on an assumption that the links between climatic and human history are simple or straightforward. This new appreciation of the role of climates and climatic change is rather an extension of well-known ecological principles.

One of these principles predicts that changes in community composition will result from shifts in relative competitive advantage when environmental factors change. In the human context, this principle suggests that the physical environment, and particularly the climate, gives a bias to the direction and success of the near-infinite series of decisions that make up the course of history. That is, with a shift in temperatures, or of amounts and timing of rainfall, the particular mix of resource use techniques characteristic of a population may well change, some occupations supplanting others as they become more profitable, less risky, and therefore more important. Clearly the size and rapidity of climatic change is crucial to the ability of societies to adapt or to cope.

Another of these principles is that of limiting factors. To use an example from limnology, nitrogen may be limiting in one lake, phosphorus in another. Lack of heat is limiting in the Arctic, lack of water in the desert. Climatic fluctuations may move temperature, precipitation, or other climatic thresholds and change the absolute limits of particular economic activities, altering previous patterns. New possibilities may also be opened or old patterns eliminated.

What then do we know about the climates of history?

DOCUMENTARY EVIDENCE There are scattered references to climatic change, and its impact on human activities, that go back to Greek and Roman times. Cyprian, about 250 A.D., commented on the diminution of winter rains and summer heat in Tunisia. Less clear is Avienus' description of the deserted and desert nature of the east coast of Spain in the fourth century A.D. It is not until the eighth or ninth century, however, that the documentary evidence becomes abundant enough to make possible a nearly continuous account for some regions.

As recently as 1962, reconstruction of past climates was largely qualitative, or at best semi-quantitative. The relative abundance of comments on the severity or mildness of winters, as gleaned from such sources as chronicles and diaries, gave a rough measure of the character of a decade. Similar relative measures of summer wetness were possible. More quantitative but less direct were the records of grain prices, numbers of prayers said for rain per year, duration and extent of sea ice, outbreaks of weather-related disease, famines, and the like. Interpretation and calibration were necessary, and one was never sure whether some of the variation from one year to the next might not be due to non-climatic economic or political factors.

At a conference on the climates of the eleventh and sixteenth centuries, the climatologists were surprised by, and enthusiastic about, the wealth of quantitative data on climate-related parameters that historians could provide for a reconstruction of climate, but none knew exactly how to interpret the data, even when gathered by the most skillful historians. The historians, on the other hand, wanted to know what the climates had been in order to interpret history.

A primary problem with the use of documentary evidence (other than documented instrumental data) in the reconstruction of past climates is the problem of compound parameters. Most of the documented data are derived from such items as grain prices, which depend on a variety of factors, both economic and climatic (and their interaction). The reconstruction of the variations of individual factors, such as summer temperature or rainfall, require as many different sets of data as there are interacting factors. This requirement is simply a restatement of the rule that the simultaneous solution of a set of linear equations requires as many equations as there are unknown quantities.

A second problem is that of critical lacunae. In a discussion of evidence of drought at the time of the disappearance of a particular culture, a palynologist once argued that a core from a nearby lake contained no evidence, in the pollen record, of drought at that time. When the original paper on the pollen analysis of that core was checked, it was found that the lake had totally dried up during that period and that there was thus no pollen evidence because there was no pollen preserved. It is likely that this problem may also be a function of historical records. The "times of troubles" have more fragmentary records than stable times.

Another problem arises with subjective records of climate such as "this has been the coldest winter in memory" — the data filter of human recollection. Such references are to the perceived normal and thus might record short-term, but probably not long-term changes. For example, Lamb's analysis of the chronicles from Russia shows a great variation in the number of severe winters prior to the time of Napoleon's invasion, but not after. Did the disappearance of severe winters really occur? That is hardly consistent with the rest of the European data, since severe winters in Western Europe are associated with the flow of very cold air out of Russia and central Asia. Or was there a psychological change, a severe winter being nothing to the Russians but intolerable to the invaders — a new pride that changed the perception of severe winters?

The resolution of these problems requires objective, quantitative data on climate which are independent of historical data. Although great progress has been made using documentary evidence in understanding the sequence of past climates, such parallel climatic data would eliminate interpretive problems, provide a calibration for proxy data, and open up new possibilities in the separation of environmental from cultural factors. Fortunately, there are some new developments in paleoclimatic science that should foster this deeper historic insight, especially for cultures without a written history.

OBJECTIVE, QUANTITATIVE CLIMATIC RECONSTRUCTIONS Climate varies on many scales of space and time. Interannual variations are usually highly correlated over short distances, but the correlation decreases with distance so that at some distance (often hundreds of kilometers) the correlation becomes negative. At still larger distances the correlation may again be positive and significant. These far-off positive correlations are often called teleconnections. The same is true of decadal and probably century-long variations. This situation is fundamentally a logical consequence of the atmosphere being a dynamic, interconnected whole, the large-scale dynamics requiring, through changes in the atmospheric wave structure, that some regions warm when others cool and some become wetter when others become drier. Precipitation often varies over shorter time and distance scales than does temperature.

There may also be variation in the rate of change. Some climatic changes appear to be rapid steps from one climatic state to another in some regions, especially those near climatic quasi-discontinuities (e.g., mean frontal positions), whereas others, far from such boundaries and tempered by maritime influences, appear to have more gradual climatic transitions.

There is no true substitute for an instrumental record of the climate at the place of interest to the historian. However, if the time of interest is more than a century into the past, the probability of such a record existing diminishes rapidly to zero. For the reasons mentioned above one cannot, even for the seventeenth through the nineteenth centuries, use whatever record happens to be available somewhere. The central England temperature record for the past few centuries is representative of central England, less so of Western Europe, still less so of Scandinavia, etc. It certainly must be regarded with some skepticism as being representative of the-hemisphere or globe unless there is corroborative evidence.

Fortunately, several methods have been developed for deriving objective, quantitative climatic data from proxy data. Most of them use biological data, although glaciological and chemical data may be useful in some areas. Unfortunately, year by year and month by month detail is not generally possible, even with the use of tree rings and annual laminae in sediments and glaciers. Trees blur several years together because of various storage mechanisms and respond to several parameters of the climate. For instance, if a drought has proceeded long enough for the weaker trees in an area to succumb, the roots of the surviving trees may have access to a larger area of soil, and thus to more water, and the tree may receive more light. The surviving trees may grow better, simulating more rainfall, despite a continuation of the drought.

However, even proxy records that are less than perfect and give decadal to century estimates have their use. Such records from the past few centuries, if they were taken closer to the historian's region of interest than the nearest instrumental record, may reflect the climate of concern as well or better. For earlier times they are essentially the only data.

A prime use of climatic data in historical analysis is to assess the degree of stress on the economic base of the people being studied; this is often an agricultural or fisheries base. For this purpose the state of the art in climatic reconstruction is sufficiently far advanced for considerable progress to be made, especially for cultures with no, or minimal, documentation. Even where the local conditions or chance result in no climatic reconstruction having been made as yet, the coherent structure of the atmosphere can provide a means of estimating the climate using data from far distant places with at least as much confidence as one should have in the use of an actual record from a single place a thousand kilometers away.

From the existing proxy records of environmental variation, primarily reflecting climatic variation, one may reach two conclusions:

1. It is not necessary to theorize about what the climate might have been on the basis of some putative cause of climatic change, at least for the past few centuries. There are adequate tools for assessing the proximal cause of environmental change without resorting to ultimate causes and theory, much of which is still untested against reality.

2. There have been times of globally synchronous, rapid, climatic (environmental) change. These should have been times of maximum stress for some cultures and improving times for others.

A TIME FRAMEWORK OF HISTORICAL ENVIRONMENTAL CHANGE In the context of the present discussion one can at least give a framework of times of apparent rapid change between climatic regimes. This framework does not provide the detailed local sequence of specific climatic data that one might wish to have for studies of local short-period history, but it does provide important background for the analysis of historical trends and developments.

Climatic data are often expressed in terms of means of temperature or totals of precipitation. That a decade was a degree or two colder than the preceding one does not sound impressive — especially to those of us who know that if we change the temperatures in our houses by a degree or two by changing the setting of the thermostat we really do not notice a significant difference. However, an often overlooked fact about climatic data is that small differences in the mean temperature may mean significant differences in the frequency of occurrence of extreme values. Even in the midwestern United States, a change of mean July temperature of 2–3 °F may be associated with a change in frequency of the extremes which are stressful to crops by a factor of five to ten. In Iceland a decrease of annual temperature of 1ºC reduces the growing degree days by 27 percent, illustrating that small changes may be critical in marginal areas.

If we restrict our attention to past cultures, as recognized by the scholars who study them and objectively date them, there are two significant dates — the endpoints. Intermediate dates may be regarded as a sampling from the cultural continuum. The earliest and latest dates that a culture would actually be recognizable are not generally known, however, since the probability of those points having been dated is small. However, we can treat the 150 or so radiocarbon dated cultures as a set of continua with only approximately known termini, and apply statistical tests to see whether there are significant times marked by global cultural change. The results of such a study show that there have been highly significant times of culture change, some being nearly ubiquitous. The significant dates from that study, and a later one appear in Table 1.

It should not be concluded from these dates that every society would have been experiencing climatic problems at those times, for nearly all climatic changes about which much is known have been such that some areas got wetter, some drier, some warmer, some colder, and some showed very little change. It cannot be assumed that, because some areas showed evidence of climatic change that conflicted with some other area, and some areas showed no change, the evidence is wrong. Furthermore, there might have been non-environmental factors operating on a global scale, although it is difficult to imagine what, other than geophysical processes, might have operated synchronously or at least within a century or two over most of the globe in prehistoric times.

An additional study was done to determine which of these times significant to culture history might have been times of environmental change. All dates reported in Radiocarbon as indicating times of environmental change were extracted, and statistically analyzed. There are hundreds of such dates, which tend to cluster at preferred times (Table 2). Work not yet completed suggests that the clustering is even sharper if the dates reported in Radiocarbon are adjusted to a uniform set of criteria as to data reduction and choice of beginning date for a change (rather than some being beginning, some middle, some previous peak, etc.).10

The dates in Tables 1 and 2 do not match very well but, given the uncertainty of radiocarbon dating and data selection, many are close enough to each other that the scholars of those periods should consider the hypothesis of climatic change as a factor in cultural change and historical events at those times. Indeed, it would be wise to consider the hypothesis for any time of global cultural change because geophysical changes have been more tightly linked globally than have human events.

There are more times of known climatic changes than are indicated in Table 2. Various proxy data series indicate, at least for Europe and North America, that important changes occurred around 1000 A.D., 1150–1200 A.D., C. 1450 A.D., and in the 1550–1600 A.D. period.

Summing up the studies mentioned above and many more, we may construct a tentative global sequence of climatic episodes, rather similar to the geological framework of epochs separated by significant changes or "revolutions" (Table 3). The names given come primarily from the Blytt-Sernander European sequence derived from palynological studies, since that time framework appears to be globally applicable. The post sub-Atlantic sub-episode names are taken from other work and are deemed preferable to such loosely defined terms as "little ice age." After all, glacial advances did not occur in all parts of the world.

It is not possible to summarize in a table all the detail that is known about each of the climatic episodes, and it is not possible, in general, to characterize even sub-episodes as "cold," for example. In almost every case data can be found to show that there was a varied direction of change, even though the times of change were globally common. Some examples will be outlined in the following section.