Featured Article » Nonfiction

Climate: The Little Ice Age After the Ring of Fire

Written by Iver P. Cooper

In winter 1634, James Byron "Jabe" McDougal "recalled that "the winter of 1631–32 had been quite a shock to himself and his fellow up-timers. Not only were they considerably farther north than they had been when Grantville was in West Virginia, but they were smack in the middle of what up-time historians had called the "Little Ice Age," which had begun some two centuries prior and would continue for another century, give or take. (Robinson, "Mightier than the Sword," Grantville Gazette 6).

****

So what was this Little Ice Age? The real Ice Ages were prolonged (as in millennia) periods of pronouncedly colder world or hemispheric temperatures in which the polar and continental ice sheets were of considerably greater extent than in historical times. There have been a dozen or so major glaciations over the last million years. A particularly big one occurred 650,000 years ago and lasted 50,000 years. However, the one that is usually considered the last Ice Age peaked about 20,000 years ago.

So that implies that a little ice age is one that is shorter and milder than that one, yet still noteworthy. Defining when a little ice age begins and ends is a bit tricky. Do you draw the line based on when a particular glacier advances or retreats, when a particular lake freezes or thaws, or when the grapes are harvested? If you rely on mean temperatures, then over how many years do you average them, and to what longer reference period do you compare that moving average? What if the temperature "breakpoints" are different in Iceland than they are in France?

Depending on who you ask, the Little Ice Age began in 1200, 1250, 1300, 1350, 1400, 1450, 1500, 1550, 1600 or even 1650. There is more agreement as to when it ended; 1850 is the year usually cited, but some would say 1870, 1900, or even 1920.

When people talk about the Little Ice Age (LIA) nowadays, they are mostly interested in the Big Picture: Was the LIA, viewed on some appropriate time scale, a global, a hemispheric or merely a European phenomenon? How much colder was the earth then than it is now? What caused it? Is the Earth warmer now than it was during the "Medieval Warm Period" that preceded the LIA? And to what extent is that warmth attributable to human activity (changes in albedo as a result of deforestation, or increases in greenhouse gases as a result of factory emissions)?

However, for those writing in the 1632 Universe, the Little Picture is what we need: what is the climate likely to be like in Germany, Italy, France, Scandinavia, England and in other areas of interest, each yearover the decade following the Ring of Fire (RoF)? (The RoF occurred up-time on April 2, 2000 and down-time on May 25, 1631 Gregorian calendar.) How does it compare to the climate in living memory, for both the down-timers and up-timers? What practical effect will it have on health, agriculture, transportation, communications, mining, industry and warfare?

In the first part of this article, I will provide some background as to the effects that climate can have on human society.

In the second part, I will try to fill in the Little Picture, based on the assumption that the Ring of Fire hasnot altered world climate; i.e., I can rely on modern reconstructions of historical temperature and precipitation averages in the areas and years of interest.

Finally, in the third part, I will consider the Ring of Fire as a meteorological phenomenon, and speculate about how much and for how long it could perturb weather and even climate.

PART I: EFFECTS OF CLIMATE ON HUMAN SOCIETY

Climate and Health

Excessive heat and cold can directly threaten human life. In studied regions of England and Wales (1993–2003 data), it was found that risk of mortality increased by 3% for every degree Celsius above the "heat threshold" (95th percentile of the mean daily temperature for the region), and by 6% for every degree below the "cold threshold" (the 5th percentile). In general, heat effects were seen once mean temperature reached 17–18^oC, and cold effect below 5^oC. (Hajat). At least in modern Europe and the United States, cold-related deaths are more common than heat-related ones, and that was even more likely to be true in LIA Europe.

The very old and very young, and those in poor health, are the most vulnerable to temperature extremes. However, the human body can adapt over time, which is why we can live in both cold and hot climates.

In addition, there are "cultural" as well as biological adaptations, and these can work in the short-term. In cold weather, one can wear heavier clothing, or go indoors and build a fire. In the 17th century, there was less that could be done about hot weather, of course. Especially since many Europeans thought that bathing was a bad idea.

Lives may be also be lost as a result of flooding caused by excessive rainfall, if the endangered population cannot flee to higher ground in time. Drought can also kill, if water has not been stored in advance. In hot, dry climates, dehydration is often associated with heat stress.

Even when climate extremes don't kill you outright, they can cause famine, which in turn reduces the body's resistance to infectious disease. "Malnutrition aggravated an influensa epidemic of 1557–8" (Mandia).

Normal seasonal variations may also have health consequences. Sometime around 400 B.C., Hippocrates declared, "The changes of the season mostly engender diseases." The basis for seasonality is not always clear. It may be related to increased pathogen (or disease vector) survival under particular temperature and humidity conditions, increased opportunity for transmission as a result of travel or overcrowding, or reduced host immunity or impairment of other host defenses (e.g., drying of the mucous membrane).

That said, some diseases definitely have seasonal propensities. In autumn and winter, we have influenza; in spring, measles; in summer, malaria (and in modern times, polio). (Dowell). In 1908 Manhattan, scarlet fever and measles were most common in March; there was a higher incidence of death from pneumonia and bronchitis from November through April; death from childhood diarrhea peaked in July–August, and cases of typhoid in August–September (North).

For malaria, the role of climate is well-understood. "Malaria transmission does not occur at temperatures below 16^oC or above 33^oC, and at altitudes > 2000m because development in the mosquito (sporogony) cannot take place. The optimum conditions for transmission are high humidity and ambient temperature between 20 and 30^oC. Although rainfall provides breeding sites for mosquitoes, excessive rainfall may wash away mosquito larvae and pupae." (Cook 1202). The northern limit for malaria in Europe has been the 15^oC July isotherm (Reiter).

While Europe was colder during the LIA, it wasn't cold enough to prevent malaria. However, a correlation has been reported between high (over 16^oC) summer temperatures in Kent and Essex parishes; Reiter speculates that the "hot weather . . . could certainly have increased the probability of transmission by shortening the extrinsic incubation period (the time required for the mosquito to become infective after feeding on an infected person)."

Yellow fever also is seasonal. In Trinidad, the density of one mosquito carrier was six times more common in the wet season (May–November) than in the dry season; bear in mind that in the tropics; the seasonal variation of temperature is small (Chadee).

Plague is rather more problematic. In Switzerland, 1628–30, the outbreaks were mostly between September and January, with November the month of highest frequency (Eckert). But other outbreaks favored summer, with peaks of mid-summer for Penrith 1597–8, Marseilles 1720, and London 1665, and late-summer for London 1625 and Debrecen 1739 (Welford).

****

The climatic deterioration was blamed on human misconduct. In Switzerland, Cysat wrote in 1600, "Unfortunately because of our sins, for already some time now the years have shown themselves to be more rigorous and severe than in the earlier past. . . ." (Pfister2007).

From blaming sins, it was a short step to looking for sinners. In Treves, Hans Linden's Gesta Treverorum blames the nigh-continuous crop failure of 1581–99 on "witches of devilish hate," and proclaims that "the whole country stood up for their eradication."

Accusations of causing "unnatural weather" or crop failure peaked when climate extremes disrupted agriculture. Moreover, it was generally considered unlikely that a single witch could control weather on a large-scale, which meant that the witch hunts were comparably large in scale (Pfister2007, Behringer).

In the 1620s, in Central Europe, there was a succession of extremely cold summers. For example, on May 24, 1626, there was a hailstorm in Stuttgart, "which brought hailstones the size of walnuts. . . ." Two nights later, ice formed, and crops failed. Witch-burnings in central Europe rose to a peak of over 500 a year, well above the "normal" (presumably, non-weather-related) level of the mid-16^th century of 100 a year. As late as 1630, "suspects still had to confess that they had been responsible for the extreme frost in May of 1626."

****

The wealthy, of course, get to choose where they live, and they live where conditions are healthiest. Lamb has pointed out that in Surrey, 20^th-century luxury housing is on the hilltops, whereas in the LIA, the favored sites were in the valley bottoms (LambCHMW 251).

Climate, Agriculture and Fishing

Jabe McDougal was not the only up-timer who has the Little Ice Age on his mind. One May, after the death of Mabel Jenkins in 1632 (Grid), Joe Jenkins grumbles that "there's snow on the ground" and "it's still here from February." He is worried that it won't be gone in time to plant corn and tomato, and adds, "If it weren't for the wheat, I could just up and starve with this here 'Little Ice Age.'" (Howard, "Golden Corn—A Tale of Old Joe on the Mountain Top," Grantville Gazette 9).

In the broader scheme of things, climate change can affect what crops can be raised in a particular part of the world. The ability of a plant to grow in a particular place is dependent on soil and climate.

Too much or too little heat, or too much or too little rain, can result in crop failure. Both droughts and floods can kill crops. Floods can be caused, not just by excessive rain, but by normal rain after a prolonged dry spell, as a result of which the soil has lost its normal ability to absorb water (Brooks 60).

If food cannot be rapidly and economically brought in from an unaffected area, crop failure leads to famine. Famine several years in a row can result in a major increase in illness, death or emigration, or in political unrest resulting in overthrow of the government or bloody suppression of a rebellion.

The down-timers in Thuringia are growing grain (primarily rye, barley, and spelt), vegetables, grasses for hay, and woad for dyeing. Of course, those are already adapted to the local climate. How will the plants that passed through the Ring of Fire, and are accustomed to the conditions of West Virginia in 2000, fare in LIA Germany?

There are complex plant-specific crop models available for predicting the combined, nonlinear effects of temperature and rainfall on plant development. These take into account changes in the sensitivity of the plant depending on its growth stage.

That's too complex for us, but we can look at what are called "cardinal temperatures"—minimum (base), optimum, and maximum (ceiling). Even those have their subtleties, as the cardinal temperatures may differ for germination, vegetative growth, and reproductive yield (which for grains is the crop yield).

Generally speaking, cool season crops (oats, rye, wheat, barley) have a base of 0–5^oC, an optimum of 25–31^oC, and a ceiling of 31–37^oC, and hot season crops (melons, sorghum) have a base of 15–18^oC, an optimum of 31–37^oC and a ceiling of 44–50^oC (Change, Climate and Agriculture 75).

Crop maturation is a cumulative process and crop scientists sometimes use the concept of growing degree days, awarding one GDD (^oF or ^oC) for each degree (^oF or ^oC) that the mean temperature on a particular day exceeds the base (some versions truncate if the temperature exceeds a ceiling). For example, wheat has a base of 40^oF; corn, 50^oF; and cotton, 60^oF. Insects also have GDDs; 50^oF for the European corn borer. (Fraise).

A decline in mean summer temperature has a double whammy. It reduces both the height and breadth (growing season length) of the GDD curve. In England, in the coldest years of the LIA (1695, 1725, 1740, 1816), summer temperatures were about 2^oC below the modern norm, and the growing season "was probably shortened by two months or even more." (LambCHMW 223).

The principal Indian crop in New England was maize, and there's reason to believe that the native strains required 2000 growing degree-days (GDDs), base 50^oF, to reach maturity. (The Indians also grew beans but these reached maturity more quickly.) In the 1960s, Connecticut, Rhode Island, Massachusetts, the Connecticut River Valley (NH-VT border), southeast New Hampshire and southwest Maine all were receiving at least 2000 GDDs (the area around Boston typically received over 2500 GDDs). A 2^oF reduction in mean July and mean annual temperatures would put all of New Hampshire, Vermont and Maine, as well as northwest Massachusetts, under the 2000 GDD mark (Demeritt).

Grantville is based on Mannington, located in Marion County, WV. According to the 1997 Census of Agriculture, Marion County had only one farm growing wheat and oats for grain. It had 251 farms producing hay (primarily from alfalfa). You can figure that alfalfa would be cut at 750 GDD, base 41^oF, to yield a fiber content 40% neutral detergent fiber. For 45% NDF, you would allow another 220 GDD (Pennington).

While there is no commercial production of corn in Mannington, canon says that there was a small quantity of seed corn available in Grantville as of the RoF (Weber, "In the Navy", Ring of Fire 1). There are also sunflower seeds, see Vance, "Second Chance Bird, Episode Two," Grantville Gazette 33. Sunflowers have a base of 44^oF and require a GDD of something like 2300. (Putnam).

We can compare these temperatures to those that are reconstructed for the places and times of interest.

Bear in mind that temperatures below the base temperature might not just stop growth, they might kill the plant altogether. Flowers and young fruits of fruit trees are often killed by mild frosts (0–5^oC) (Hatfield).

The USDA defines plant hardiness zones based on the extreme cold (expressed as the average minimum annual temperature) that a particular plant can tolerate. Zone 1 is -60^oF to -50^oF, zone 2 -50 to -40, and so on up to zone 11, 40 to 50. Each zone may be further subdivided into two subzones, "a" and "b," with "a" as the colder half. (Zones 0 and 12 are special cases; 0a is under -65^oF, 0b is -65 to -60, 12a is 50-55 and 12b is over 55.)

Plants vary in terms of what kind of climate they like. For example, the orange tree (Citrus sinensis) is considered hardy in zones 9a–11a, whereas the Scots pine (Pinus sylvestrus) grows in zones 1–4.

In 1990, the USDA prepared a map of North America depicting which areas are in which hardiness zones, based on their average annual lows (over the period 1974–86). This of course changes as the climate changes; in 2006, the Arbor Day Foundation updated the U.S. hardiness zones to reflect the most recent 15 years of data and perhaps half the U.S. (excepting California and Nevada) experienced a one zone (10^oF) increase.

The logic behind the hardiness zone definition is that even a brief exposure to a cold enough temperature will kill the plant. However, it ignores the fact that a plant may withstand a short exposure to say -5^oC yet be killed if there are too many days at 0^oC.

Also, it ignores the effects of day length, summer heat, wind, and the amount and distribution of rainfall, which in turn are influenced by latitude, elevation, continental position, and mountain barriers. The American Horticultural Society has a Plant Heat Zone map; the zones are based on the average number of days per year above 30^oC, thus accounting for summer heat. There are other zoning systems, that take additional factors into account, but we can't use them in LIA Europe because we lack some of the necessary data.

****

Besides the direct effects of climate on plant growth, there are also indirect effects. Plant pests are also affected by temperature; a warm winter may mean a bumper crop of insects in the spring. In late 17^th-century Switzerland, cool springs led to crop losses as a result of attacks of the parasite Fusarium nivale, which is active under snow cover (LambCHMW 206).

****

Domestic animals are also affected by climate. Animals can be killed by climate extremes, especially the combination of heat and drought. Even conditions that don't kill can reduce reproduction, growth rate, and milk production.

****

Considering domesticated plants and animals together, both temperature and precipitation can have significant adverse effects. The so-called LIA-type impacts are:

March, April: cold decreases forage for dairy animals and the volume of the grain harvest.

July, August: rain interferes with the harvesting of crops.

September, October: cold forces animals into the barn earlier and reduces the sugar content of vine-must; prolonged rain reduces area sown and nitrogen content of the soil (thus affecting the following year's productivity).

Pfister2006 has combined temperature and precipitation monthly data to arrive at a "biophysical climate impact factor."

****

Fish have preferred water temperatures. During the LIA, cod and herring moved south, hurting the fisheries of Norway, Scotland and the Faeroe Islands, but benefiting the English (Mandia).

Climate and Transportation

In 1630, the cheapest form of transportation was by water. However, except in far northern Europe, transport was dependent on liquid water; skating or skiing on ice or snow was fine for individuals but not practical for large-scale freight movement.

So that means that we need to ask when will geographically significant navigable rivers freeze and thaw, in which months will strategic harbors be closed by sea ice, and when will particular sea routes be endangered by icebergs.

Rainfall can also make a difference. In some parts of the world, rivers are navigable only for part of the year. Or in some years and not others.

Land transportation is also affected by climate. Snow can close a mountain pass, or simply make it slower to travel by road. Rainfall can turn dirt into mud, or make a ford impassable, or cause a flood that destroys a bridge.

On the other hand, the freezing of rivers (while not good for water travel) can make river crossings easier. In 1597–8, Matteo Ricci wrote that "once winter sets in, all the rivers in northern China are frozen over so hard that navigation on them is impossible and a wagon may pass over them." (Brooks 55).

Up-time transportation technology also has its vulnerabilities. Cold temperatures can reduce starter battery life, render fuel viscous, and cause engines to stutter. High temperatures make it easier for engines to overheat.

Climatic interference with transportation can make it more difficult to relieve a local famine by moving in food from elsewhere.

Climate and Communication

Prior to the RoF, messages traveled, at least over distances beyond line-of-sight, at pretty much the same speed as people and goods. On land, the fastest communications were those provided by a post horse system, and at sea, messages could be carried by a sailing ship built for speed and not burdened with a heavy cargo. The effect of climate on these channels of communication have already been discussed in the context of transportation.

The up-timers will be introducing radio and telegraph communications, and radio waves and electrical pulses travel at the speed of light. Of course, as a practical matter, it takes time for an operator to convert a message into transmissible form, and, at the receiving end, for another operator to convert it back again. If the message has to be relayed, then effective transmission times are increased. But it's still much faster than horse or ship.

Our climate is the result of the heating of the earth's land masses, oceans and atmosphere by solar radiation, coupled with the rotation of the earth about an axis tilted relative to its orbital plane.

The amount of solar variation emitted by the sun varies, and it turns out that there's a pretty good correlation between the number of sunspots and the solar output. All else being equal (and it rarely is), if solar output decreases, so will mean global temperature.

However, there is a more specific effect on radio communications. The solar radiation includes not only light photons, but also charged particles, and when those particles strike the earth's atmosphere, they ionize some of the air molecules. When solar output is high, the degree of atmospheric ionization is higher, and it is easier to bounce radio signals off the "ionosphere" so that they can travel longer distances. The principle is explained in much more detail in Boatright, "Radio in the 1632 Universe," Grantville Gazette 1.

Climate and Mining

Surface temperature doesn't have much of a direct effect on underground mining; the temperature underground is mostly a function of latitude. However, it can affect how easy it is to get miners and their goods to the mine, and to ship off the ore. A good case in point is that in the nineteenth century, cryolite could be mined in Greenland for only a small part of the year.

Rainfall is another matter. Drainage was a serious problem in both European and Japanese mines, and I imagine that in periods of heavy rainfall, the problem was exacerbated.

Climate and Industry

Industrial production presupposes the existence of healthy indoor temperatures. It is already a common practice to heat homes and shops during the winters in colder regions of the world. Factories in those climes will also need heating systems, and, if it gets colder, they will require more fuel (most likely wood or coal).

Summers in warmer regions are more of a problem, because the only form of cooling is ventilation. True air conditioning requires up-time technology. Fortunately, in those areas affected by the LIA, summer is not a major concern.

The effect of increased rainfall is a more subtle one; more rainfall will be associated with more humidity, which means more problems with decay (wood) and rust (iron). This may increase industrial demand, but it also means that the maintenance costs will be higher.

Climate and Warfare

The conduct of war is also affected by climate, both indirectly and directly. If harvests are poor, it will be difficult to feed the troops and their work animals. If roads are muddy or snow-covered, troop movements will be slow. If the soldiers are not conditioned to the local climate, and properly dressed for it, there will be weather-associated deaths.

Climate begets weather, and one of the more piquant examples of the effect of weather on warfare was the January 23, 1795 capture of the Dutch fleet by the cavalry of the French Republic. It was trapped to the lee of Texel Island by ice.

PART II: CLIMATE IN THE 1630s (OLD TIME LINE)

The Up-Timers' Perspective

The up-timers are coming from a West Virginia town. While Grantville is fictional, it is based on real-life Mannington, in north central West Virginia (Marion County). Climate data for Mannington goes back to 1948, but unfortunately it's spread over three different weather stations. For nearby Fairmont, there's continuous data from a single station.

Please note that interannual variability of even annual (let alone seasonal, monthly, or specific day of the year) temperatures is such that it is customary for weather services to calculate "climatological normals" over a thirty-year period.

Table 1 shows the sort of climate that the up-timers of Grantville are accustomed to. From this we can estimate seasonal average temperatures as follows: winter (DJF), 31.7^oF (-0.2^oC); spring (MAM), 51.0 (10.6); summer (JJA), 70.5 (21.4); autumn (SON), 54.2 (12.3). The average of the daily minimums for January was 20.4^oF.

(Climatography #81, #85)

Fairmont (ZIP code 26554) was in the 1990 USDA Plant Hardiness Zone 6A (average absolute annual minimum temperature in range -10 to -5^oF, -20.6^oC to -23.3^oC), and in zone 6-7 of the 2006 Arbor Day Foundation update.

In this part of West Virginia, the first freezing temperatures (end of the growing season) is typically in the first half of October, and the last freezing temperature (preceding spring planting) in the first half of May. http://www.accuracyproject.org/w-FreezeFrost.html

A Global Overview of the LIA

In 2002, Mann presented a figure comparing temperatures for the period 1000–2000 for eight different parts of the world. Mann considers the LIA to be 1400–1900, and my comments are based on the reconstructed annual means. I will call an LIA "low" if the temperature was less than the lowest value for that region during 1000–1400.

Northern Hemisphere: the lows are in the late-16^th, late-17^th, and late-19^th centuries, with highs in the early 17^th and mid 18^th centuries.

West North America: the deep lows are in the late-16^th and mid-19^th centuries, and a shallower but broader low appears in the 17^th. The highs are in the early-15^th, mid-16^th and late-18^th centuries.

Subtropical North Atlantic: there's a broad shallow low centered on 1700, and a high in the early- and mid-16^th century.

Western Greenland: The entire LIA was warmer than in the late-14^th century, but at its warmest in the early-15^th and coldest in the late-17^th and late-19^th centuries.

Central England: The LIA saw a long decline to the low of late-17^th century, then an improvement in the early-18^th century. Temperatures remained well below the broad peak of the 13^th century.

Fennoscandia: the deep low is just after 1600, and temperatures gradually recovered to a broad peak in the late-18^th and the whole 19^th centuries.

Eastern China: the biggest temperature drop of 1000–2000 was before the LIA, in the 12^th century. During the LIA, temperatures remained at or above their 14^th century levels, with a broad peak in the 19^th century.

Tropical Andes: the LIA really began around 1500 here, but there were no sharp lows. The lowest points are in the late-17^th and late-18^th centuries. The early-17^th century was cooler than the 15^th century but otherwise unremarkable.

Thus, the LIA was not simply a four-century cool period; it included warmer and cooler intervals, and these weren't synchronous between regions. However, it has been contended with some justice that it was a period of greater temperature variability.

1630s Europe: Historical Accounts

We can learn a lot about past climates from historical records. At the very least, they speak directly to the real-life consequences of weather conditions (droughts, floods, freezes, heat waves, and storms). And in some cases the historical records provide quantifiable information (e.g., the dates that particular lakes or rivers froze or thawed, the dates of harvesting grapes or other crops) that can be correlated with overlapping instrumental records so that the older temperatures may be inferred.

While our interest is particularly in the 1630s, we will from time to time look back at dates that would have been in living memory, and forward to the 1640s.

Pfister has constructed, by rating the severity of temperature and rainfall extremes in documentary accounts and weighting them together, an index of climate impact on European agriculture. There were major peaks in 1569–73, the late 1590s, 1614, and 1626–29. 1628 was a "year without a summer." (cp. Battaglia). "After 1630 the level of climatic stress drops substantially." The next peak, in the 1640s, was of about half the magnitude of the one in 1626–29.

Temperature increased from the 1620s to the OTL 1630s, and the number of witchcraft trials in eleven regions of Europe, standardized relative to the regional means, declined. In the OTL 1640s, they increased again, to higher than the 1620s level (Oster, Fig. 1).

Great Britain. According to Wikipedia/River Thames Frost Fairs, in the 17^th century, the Thames froze over at London in 1608, 1621, 1635, 1663, 1666, 1677, 1684 and 1695. With particular regard to the winter of 1635, the frost was severe from December 15 to February 11. It was followed by a warm and moist spring, and a very hot and dry summer and autumn. But the following winter (1635–36) was unseasonably warm. 1637 was also cold. The summers of 1636, 1637 and 1638 were all hot and dry (Marusek 116; LambCPFF 568).

During the LIA, the 25-year average of the English price of wheat increased from its low around 1500 to a high around 1650, then dropped to a shallower low in the late-18^th century, and then climbed to a greater high in the early-19^th century (Flohn 44; LambCPPF 462).

Note that during the coldest parts of the LIA (which for England was the late-17^th century), the growing season was shortened by 1–2 months compared to that of modern England (Mandia).

Scandinavia and the Baltic. Historical climatologists have found records of the date of ice break-up at the harbor of Riga (Latvia) going back to 1529. We know that in the 1620s, 1620–21 was a severe winter, 1622–23 was average, and 1625–6 was mild. And in the 1640s, 1642–3 was severe, 1648–9 was average, and 1649–50 was mild. But the data for the 1630s are missing. The average break-up date is March 24 in a mild winter, April 3 in an average one, and April 12 in a severe one, but the variability is fairly high. (Jevrejeva). For the 17^th century, the earliest date was Feb. 2, 1652 and the latest May 2, 1659 (LambCPFF 587).

There is also data, complete from 1600 on, for ice-breakup at Tallinn (Estonia), which is near the average western limit of the ice cover in the Gulf of Finland. (Tarand 192). The means for 1597–1629 were year-day 97.4 for Riga and 106.18 for Talinn, and for 1630–1662, they were 80.25 and 99.73 respectively. The estimated winter air temperatures for Tallinn were -5.84^oC and -4.72^oC for the two periods. And the "Ice Winter Severity Index" for the Western Baltic dropped from 0.73 to 0.44 (it was 0.02 in 1988–93). (Eriksson; Tarand 192).

Alas, the post-1622 Great Sea Toll records for Stockholm, recording the dates of first arrival and last departure for each shipping season, were requisitioned by the Swedish Army as—brace yourself—wadding for artillery. Nonetheless, useful records relative to the shipping industry have survived, and the climate observations from these records have been scaled and calibrated with overlapping instrumental data to reconstruct winter temperatures for Stockholm. These reveal that 1614–23 (-2.43^oC) and 1624–33 (-2.20^oC) were the second and fourth coldest decades since 1500. (The dangers of relying too much on the generic Little Ice Age label are shown by the fact that one of the five warmest decades, 1734–43, is within the conventional LIA.) The decade of 1634–43 was a bit warmer than 1624–33.

I have found reports of crop failures in Norway in 1632 and 1634 (GroveLIAAM, 67). These are probably attributable to the proximity of glaciers. The 1742 report of the court of inquiry on Elekrok stated "it was apparent to us that it was the nearness of the glacier which is the cause of crop failure on this farm . . . the ears on the side towards the glacier . . . were quite brown, and the other side green. . . ." (71).

The Netherlands. In East Friesland, on Sept. 1, 1637, there were great floods (Marusek 116). (The specific date won't be repeated in the new time line, but there may still be a propensity to flooding that autumn.)

For wheat prices, the first LIA climb came later than for England, possibly in the 1640s. Otherwise, the fluctuations were similar to those for England but smaller.

Germany. The walls of historic buildings at Tonning on the west coast of Schleswig-Holstein reveal that the flood of Oct. 11, 1634 reached a height of four feet above the ground surface. (LambCHMW 17). On Norstrand Island, 6,123 people drowned, and 50,000 livestock were lost (Rabeljee). This flood is mentioned in Grantville literature, but the up-timers didn't think to warn King Christian about it because they assumed it would be "butterflied away." Instead, it came ahead of schedule. See Boyes, "A Great Drowning of Men," Grantville Gazette 28.

The price of rye in Germany over four centuries has been analyzed. Peaks corresponded to a poor harvest—this could be because of climate, or because of warfare. Considering just 1590–1650, there were small peaks in 1590 and 1610, moderate ones in 1626, 1634 and 1649, and a large one in 1622. However, the worst one of all was that of 1816 (the "year without a summer") (Mandia, Fig. 17). Other than in 1634, the 1630s appear to have offered cheap rye—albeit not as cheap as in the "good years" of the 16^th century. The high prices in 1634 were probably attributable to plague (Pfister2007, Fig. 8).

France. On October 6, 1632, southern France was so cold that sixteen of Louis XIII's bodyguards died from exposure (Marusek 115). The winter of 1638 was also severe; in Marseilles, the "water froze around the ships."(116).

Wine grapes will reach maturity more quickly if the growing season (April–September) is warm, than if it is cool. Based on the extensive wine harvest data, the summers of 1634–39 were warmer than the 1599–1791 mean (Ladurier; Chuine).

French wheat prices followed a pattern similar to that of British, but the fluctuations were more moderate (Flohn 44).

Switzerland. Based on historical documentary evidence, Pfister constructed crude thermal (warm months-cold months) and wetness (wet months-dry months) indexes for Switzerland. The 1630s appear to be a little on the warm side, and markedly on the dry side. The 1640s were colder, although nowhere near as cold as the 1670s, 1690s, or 1810s, and not as dry (LambCHMW 204).

In the Alps, we have the very visible evidence of the advance and retreat of the glaciers. Unfortunately, in the 17^th century, we do not have good maps of their positions, and hence we have to rely on diaries and legal documents. In 1600–19, there are repeated descriptions of the destruction of houses, the failure of crops and the decline in tithes as a result of glacial advance (Ladurie 143ff).

It appears that the glaciers were more quiescent in the 1620s and 1630s, but they remained dangerous. A third of the cultivable land at Chaimonix was destroyed in 1628–30, and the Mattmarksee (influenced by the Allalin glacier) flooded in 1620, 1626, 1629, and 1633 (Aug. 21). In 1636, the people in the valley of Randa thought that the whole Zermatt glacier was coming down on top of them; forty people were killed by ejecta. In the 1640s, there were new glacial advances. In May, 1642, the Les Bois glacier was reportedly moving by "over a musket shot every day." Such ominous developments led to the famous June 1644 procession, led by the local bishop, to seek divine intervention to hold the glacier at bay (Ladurie, 165–173).

Wine harvest dates for fifteen locations in the Swiss Plateau and northwestern Switzerland have been used to reconstruct April–August temperatures. Harvest on year-day 285 indicated temperatures identical to those of the base period 1961–90; earlier harvests implying warmer temperatures. In 1632–33, temperatures were a little below base, whereas in 1634–39 they were higher, peaking at 1.78^oC higher in 1638. The temperature anomalies in 1640–43 were negative (Meier).

Northern Italy. The second quarter of the seventeenth century was not marked, as was the first one, by any "great" winters (enough for large bodies of water to have ice thick enough to support people) or even "severe" winters (causing the death of animals and plants) (Alfani Graph 1.3). The flooding of tributaries (Tanaro and Bormida) of the Po was perhaps half as common as in the preceding quarter-century, but nonetheless more common than in the next one.

In 1629, a landslide, triggered by heavy rain, caused loss of life and property in the hamlet of Onera. In 1629–30, a plague epidemic killed about 27% of the population of Northern Italy, but the extent to which climatic factors contributed to its occurrence remains in dispute (Alfani). In 1632, there were complaints about both heat and drought (Marusek 115). In the lower Po valley, cereal yields were "seriously reduced in the period 1590–1630, especially." That was, of course, attributable to the flooding (Grove 129).

The Italian price of wheat in the LIA reached a peak just after 1600, then descended to a broad low in the 18^th century, then climbed more moderately to twin peaks in the 19^th (Flohn 44).

Spain. The period 1575–1650 was "generally wet," at least in the southeast. 1617 and 1626 were "deluge" years, and "catastrophic floods were unusually frequent between 1571 and 1630, especially in Catalonia." (Grove 129). There was major flood activity in 1630–1650, too (Llasat Fig. 5).

The Black Art of Reconstructing Past Climates

Crude thermometers appeared in the 17^th century, and our oldest continuously monthly temperature records date back to 1659 (for central England). For that region, the coldest winter was in 1684, the coldest summer in 1725, and the coldest year overall was 1704 (Manley). Other early records are those for Berlin from 1697, for Hoofddorp and Zwanenberg/De Bilt in the Netherlands from 1706 and 1735, respectively, for Uppsala (Sweden) from 1739, and for St. Petersburg from 1726 (Flohn).

Clearly, this direct data doesn't tell us anything about what the temperatures were in 1631–39. However, it does help in calibrating "proxy" data.

A "proxy" is any observable variable of the fossil record (this term used in a broad sense) that can be reliably correlated with direct temperatures for part of the range of the record, so that the historical temperatures can be reconstructed for the rest of that range.

For our purposes, it isn't sufficient that the proxy be highly correlated with the actual temperature, it also must be "high resolution." For example, if we couldn't determine the age of a proxy value more accurately than the nearest decade, or if the proxy value reflected the temperatures over the preceding decade, then the resolution it offers is just decadal. We want resolution down to the annual level.

Here are some of the sources of high-resolution proxy data:

Ice Cores—the upper portion of an ice core exhibits a layered structure with annual variation; the light bands are formed by freshly fallen, clean summer snow and the dark bands are formed by old, dust-contaminated winter snow. The thickness of the light band is indicative of how much snowfall there was. Air bubbles in the ice preserve "fossil" air, in which the level of greenhouse gases can be measured. Also, oxygen isotope ratios are influenced by ocean temperatures. Obviously, ice cores are only available from a few parts of the world; notably Greenland, Antartica, and a few glaciers.

Tree Rings—the light colored layer grows in the spring and the dark colored one in late summer. Narrow rings are indicative of poor growth conditions, such as drought or severe winter. Tree ring data is available only where trees grow.

Corals—we can see annual variations in skeletal density and geochemical parameters. The light layers are from the summer and the dark layers from the winter. Oxygen isotope ratios are indicative of ocean temperatures. The most useful corals grow in shallow tropical waters.

Lake Sediments—these may exhibit seasonable variations (varving) in runoff sediment composition, which in turn are the result of summer temperature, rainfall, and winter snowfall.

Boreholes—the variation of temperature with depth has a detectable relationship to the history of temperature at the surface.

Speleotherms—these are stalactites, stalagmites and flowstones. Some provide annual resolution, as a result of visually detectable lamination, or a seasonal variation in trace elements. Layer thickness is related to surface rainfall and cave air temperature.

Historical accounts—these are most useful if they provide some kind of quantitative information.

There are technical problems with working with proxy data, but consideration of those problems is outside the scope of this article.

Climate Reconstructions: The North Atlantic Oscillation

In the mid-latitudes of the North Atlantic, the prevailing winds are from the west. These were convenient for mariners returning from the New World. However, those winds are also important because they bring moist air to Europe.

The direction and strength of the prevailing winds are controlled by the position and strength of a persistent low-pressure system over Iceland (the Icelandic Low), and a persistent high-pressure system over the Azores (the Azore High).

The atmosphere alternates between a state in which the pressure difference widens (positive phase, NAO+) and one in which it narrows (negative phase, NAO-). There are a number of ways the NAO may be quantified, but the simplest is as the normalized difference in pressure between a station in the Azores (or in Portugal) and one in Iceland. There is no significant periodicity in the switching between NAO+ and NAO-.

In NAO+, the westerlies are stronger, and more and stronger winter storms cross the north Atlantic, on a more northerly track. Temperatures are above average in the eastern United States and in northern Europe, and below average in northern Canada and Greenland and often in southern Europe, northern Africa and the Middle East. There is also above average precipitation in northern Europe, and below average in southern Europe. In NAO-, the effects are reversed.

The effects are strongest in winter. (NWS-CPC).

The North Atlantic Oscillation index has been reconstructed, on a seasonal basis, for 1500–1658 and monthly for 1659–2001 (LuterbacherNAO). Table 2-1shows its behavior for 1630–39. It can be seen that it was mostly in negative phase in that decade.

Climate Reconstructions: European Annual Average Temperatures

Looking first at reconstructed mean annual temperatures for Europe generally, Table 2-2A shows how the 1630s (with the years 1999-2000 for comparison) shape up.

^{LuterbacherTemp).}

If we consider just the temperature column, it's clear why the up-timers feel a chill in the air. However, the coldness ranks (30-362) provide some perspective. And it's worth comparing those temperature to the averages (Table 2-2B) for each century, for the Maunder Minimum (1645–1715), the whole LIA (1500–1850), and the modern period (1851–2004).

So, yes, we are in the LIA—but not in the worst decade, by any means.

Climate Reconstructions: European Seasonal Average Temperatures

There are several references to the severity of winter in 1632 universe canon. Watching TV in 1633, Joyce and Gary find that "the news was about broken armies, new business, and, of course, the weather and what the little ice age meant to their future." (Huff and Goodlett, "Wish Book" Grantville Gazette 12). In January 1634, Eric Krentz tells Thorsten Engler, "I always hated January even before an up-timer told me we're in the middle of what they call 'the Little Ice Age.' " (Flint, 1634: The Baltic War, Chapter 14). Later, in late March 1634, Admiral Simpson is "a bit surprised that the river [Elbe] hadn't frozen, although intellectually he'd understood that the past winter hadn't really been as cold as it had sometimes felt, Little Ice Age or not." (Chapter 31). In Flint and DeMarce, 1635: The Dreeson Incident, we are told that "Winter in Thuringia during the Little Ice Age encouraged the layered look." (Chap. 41).

So, how did LIA winters (and other seasons), compare to those the up-timers would have been acclimated to, and just how bad were the 1630s as compared to other parts of the LIA?

We define the seasons as DJF (winter, December-January-February, and I assume that winter for 1630 starts with Dec. 1629), MAM (spring), JJA (summer) and SON (fall). Table 2-3provides the reconstructions for each of 1630–1639, the actual values for 1999 and 2000, and averages of values for the decade 1630–39, the 30-year period 1620–1649, 1645–1715 (the Maunder sunspot minimum), 1500–1850 (LIA) and 1851–2004 (Post-LIA). The coldest (bold blue) and warmest (italic red) winter, spring, summer, fall and entire year are marked.

The springs (MAM) ranged in rank from 62^nd to 459^th coldest, the summers (JJA) from 92^nd to 424^th, and the falls (SON) from 60^th to 372^nd. For all three seasons, the mean for our decade was higher than the mean for the LIA.

Surprisingly, the mean for summer 1630-1639 was even higher than the mean for the post-LIA (1851–2004) period, although of course still cooler than the summers of 1999 and 2000.

We are lucky to have missed 1628, which was the thirteenth coldest summer (16.8^oC) of 1500–2004.

Climate Reconstructions: Central European Monthly Average Temperatures

Monthly temperatures have been reconstructed for central Europe in 1630s, based on documentary evidence (mostly from sites in the present Germany, Czech Republic, and Switzerland). Unfortunately, these are available just as anomalies, that is, the difference between the actual temperature and the average(s) for the base period 1961–1990 (Dobrovolny). I emailed Dr. Dobrovolny, asking him for the base temperatures, but he didn't reply. So I used the KNMI Climate Explorer to download the CPC GHCN/CAMS t2m analysis (land) data, gridded at 0.5° intervals, and calculated the 1961–1990 base for central Europe myself. I assumed that Dobrovolny defined central Europe the same as his coworkers did in LuterbacherSLP.

In Table 2-4, I show first my derived monthly temperatures for each year of 1630–39, and the average and standard deviation for the decade. Next, I provide my 1961–99 base numbers. There're no guarantees that Dobrovolny had exactly the same base averages; his region and his modern data could have differed from mine. But since I stated the bases I used, you can reconstruct Dobrovolny's anomalies by subtracting them out. Finally, I present the mean and standard deviation for the period 1766–1850, from Luterbacher's monthly gridded reconstructions.

Please note that with the exception of 1630 and the first four months of 1631, this monthly data is subject to perturbation by the RoF. Since it's time-averaged data, the impact won't be as severe as for daily weather, but there will be some effect.

Climate Reconstructions: Mapping Post-RoF European Seasonal Average Temperatures

Of course, pan-European averages are all well and good, but different parts of Europe would no doubt fare differently. Fortunately, Luterbacher's climate reconstruction provides reconstructed seasonal temperature for each point on a 0.5 degree by 0.5 degree grid, over the range 25W–40E longitude; and 35N–70N latitude.

It's said that a picture is worth a thousand words, and I have created some revealing images from Luterbacher's gridded temperature data. Using Climate Explorer, I have compared the decade 1630–39 with 1990–99. The figures compare the decades for (Fig. 1A) the entire year, (1B) winter, (1C) spring, (1D) summer and (1E) autumn.

And using the National Climatic Data Center's visualization tool, I have created comparisons of the coldest and warmest winters (2A), springs (2B), summers (2C) and autumns (2D) of the 1630s. Note that each season has a different temperature scale:

Winter (DJF): coldest 1635, warmest 1632, scale -20 to +10C;

Spring (MAM): coldest 1635, warmest 1636, scale -5 to +25C;

Summer (JJA): coldest 1630, warmest 1637, scale +5 to +35C;

Fall (SON): coldest 1635, warmest 1630, scale -5 to +25C.

Which parts of Europe are unusually hot and which are unusually cold is very strongly influenced by the position, areal extent and persistence of the high and low pressure areas (see North Atlantic Oscillation) and the position and strength of the jet stream. Stagnant (blocking) patterns lead to persistent weather conditions that influence monthly and even seasonal averages. On the west side of a stationary NH high, warm air is pushed north, and on the east side, cold air is dragged south. So you may be warmed or cooled depending on where you stand. Moreover, a slight shift in the location of the blocking pattern from one year to the next might mean that you face extreme cold in the first year and extreme heat in the second (LambWCHA 110).

Climate Reconstructions: Post-RoF Grantville Seasonal Average Temperatures

The place of greatest interest to the up-timers is, of course, the location in Thuringia where the RoF deposited Grantville. The center of the RoF was at approximately 11^o16' east longitude, 50^o40'12" north latitude. The closest Luterbacher grid point is 50.75N, 11.25E, and the reconstructed seasonal temperatures for this location are in Table 2-5A (with comparison to pre-RoF Grantville at the bottom of the table).