Guest Post by Willis Eschenbach
Since I was a kid I’ve been reading stories about “The Year Without A Summer”. This was the summer of 1816, one year after the great eruption of the Tambora volcano in Indonesia. The Tambora eruption, in April of 1815, was so huge it could be heard from 2,600 km away (1,600 miles). The stories were always about how the following summer was outrageously cold. Supposedly, the summer was so cold it was like having no summer at all.
Being a suspicious fellow, I got to thinking about that, and I realized I’d never seen any actual temperature data for the year of 1816. So I went off to find some early temperature data. I started with the ECA dataset, and downloaded the Daily Mean Temperature TG (162Mb). That revealed five stations with daily temperature records with starting dates before 1816—Stockholm, Bologna, Milan, Praha-Klementinum, and Hohenpeissenberg.
So once again, I found myself playing “Spot the Volcanoes”, as in my previous post on this subject. When I wrote that post, I hadn’t been able to spot the smaller eruptions of Pinatubo and other modern volcanoes, but Tambora was the big cheese, the grand gorgonzola of volcanoes. Surely I could find that one … so here’s the record from Stockholm.
Figure 1. Two ten-year periods from the early 1800’s in Stockholm
So the question is, which year is “The Year Without A Summer”? The year indicated by the blue arrow, or the year shown by the green arrow?
Actually, I fear that was a trick question. Here’s the same data, this time with the years indicated.
Figure 2. Two ten-year periods from the early 1800’s in Stockholm, including the dates.
As you can see, the 1816 “Year Without A Summer” actually was warmer than a number of other summers in Stockholm. It’s the third peak from the left in the top panel, and was above 20°C. Just in this tiny sample we see some six summers that were cooler than the summer of 1816 in Stockholm …
So, I looked at the other locations. Here are the other four European cities with records that cover the Tambora eruption—Bologna, Milan, Praha-Klementinum, and Hohenpeissenberg. In these, both the upper and lower panels are from the early 1800s. No more trick questions, in all cases, one or the other of the green and blue arrows actually indicates the “Year Without A Summer”.
Figure 3. Two ten-year periods from the early 1800’s in Bologna.
Figure 4. Two ten-year periods from the early 1800’s in Milan.
Figure 5. Two ten-year periods from the early 1800’s in Praha-Klementinum.
Figure 6. Two ten-year periods from the early 1800’s in Hohenpeissenberg.
That was all the daily temperature records I could find from that far back. There’s a monthly record from Armagh, in Ireland. Here’s that record.
Figure 6. Two ten-year periods from the early 1800’s in Armagh.
I’m sure that you can see the difficulty. If Tambora actually did something to the temperature, you sure couldn’t tell it from these records. Not one of them is readily distinguishable as missing a summer.
In “The Great Tambora Eruption in 1815 and Its Aftermath” (paywalled, Science Magazine, 1984), the author says (emphasis mine):
To Europeans and North Americans, 1816 became known as “the year without a summer” (41). Daily temperatures (especially the daily minimums) were in many cases abnormally low from late spring through early fall; frequent north-west winds brought snow and frost to northern New England and Canada, and heavy rains fell in western Europe. Many crops failed to ripen, and the poor harvests led to famine, disease, and so- cial distress, compounded by the aftermath of the Napoleonic wars.Tambora’s dust veil is often blamed by modern researchers for the cold summer of 1816. The argument given is that the stratospheric dust veil would have absorbed or reflected solar radiation that could otherwise have reached the ground (42). Not all regions,however, experienced abnormally low temperatures, and the preceding winter had generally been mild. Therefore, a few researchers deny that there was any (or at least a strong) connection with the volcano (39,43).
I’m leaning towards the “few researchers” that deny a strong connection with Tambora. What other records do we have? Well, over at KNMI I find the record for Manchester, England:
Figure 6. Two ten-year periods from the early 1800’s in Manchester.
Moving across the Atlantic, here’s the record from New Haven in Connecticut.
Figure 6. Two ten-year periods from the early 1800’s in New Haven, Connecticut.
I’m just not feeling the Tambora love here … where are the records of years without a summer? Or at least of a summer that’s significantly colder than its neighbors?
Don’t get me wrong here. I suspect that generally, the summer of 1816 was a bit colder than most summers. But as the graphs above show, in all of these datasets there are comparable summers within a few decades either side of 1816 that have summers that are as cool, or cooler, than the summer of 1816.
And I would guess that a careful search would reveal some records with cooler summers than the ones I’ve found here. But overall, let me suggest that over the years the Tambora story has gotten greatly exaggerated, just as we do today with our stories of “Cold? You haven’t seen real cold. Why, when I was a young man it was so cold that …”
Conclusions? Well, my main conclusion is what I’ve been saying for some time. The temperature of the earth is not particularly ruled by the changes in how much energy it receives. Tambora cut off a huge amount of sunlight, but the effect was small. Yes, some areas had a summer that was a bit cooler than most summers. And I’m sure there were certain locations where it hit harder than others. But overall? The thermostatic mechanisms of the planet kept Tambora from having a much of a cooling effect.
My best to all. I append all of the figures below, with the dates, so you can see the lack of effect. Note that in many of them, the temperature in 1815 was about the same as 1816 … and that despite the size of the volcano, if there was any effect, it was totally gone by 1817.
w.
If that’s what a really big volcano can do, I’m not impressed. Well, I am impressed, but what’s impressive is the strength of the thermostatic mechanisms that keep the earth’s temperature within a very narrow band. Even a huge volcano can’t put it out of sorts for much more than one summer, and even then not too much.
The place to look for the effect of volcanic eruptions on the climate is in food commodity prices. That is where climate change has its greatest impact on human society. In those records the Tambora eruption is unmissable.
Willis
The year you refer to was merely one of a number of increasingly cold years during a decade that turned out to be one of the coldest since the depths of the LIA 130 years previously. Decadal CET average 1810-1819 at 8.798C was the coldest decade since 1690-1699 .This decline happened BEFORE Tambora erupted;
The evolving climate can be seen here;
http://www.metoffice.gov.uk/hadobs/hadcet/
I compiled a complete chronology of weather from 1812 onwards to support an article I wrote about Charles Dickens who was greatly influenced by his cold childhood and 1816 fitted into the cold series of years he experienced.
http://noconsensus.wordpress.com/2010/01/06/bah-humbug/
tonyb
Willis Eschenbach says:
April 15, 2012 at 10:46 am
I’m not putting myself out as an expert about Tambora or Krakatoa. I’m not even putting myself out as an expert in 1816 weather in New Hampshire. At the time I merely wanted to collect some information that year in my adopted state and its granite monument to 1816.
Poking around on the web, I see this about Krakatoa:
http://www.independent.co.uk/news/science/how-krakatoa-made-the-biggest-bang-476616.html
Yeah, yeah, popular press and all that. You were right, I confess that 2,600 km sounded like it wasn’t very far, I should have checked. What I wanted to convery was that Krakatoa was louder despite being smaller. I hope I haven’t made a too very bad of a start at not being an expert.
Dr Dave says: “The UAH satellite data clearly shows a cooling following the Mt. Pinatubo eruption so in my mind that establishes the link.”
What reason do you have to suppose whatever data you are referring to would not have cooled if there had not been an eruption?
Don’t know what you got you doctorate in Doc
Willis: If you haven’t done so, you might want to look at the Wikipedia article on the Year Without a Summer and look at its map of temperature anomalies that came from the reconstruction made Luterbacher
http://en.wikipedia.org/wiki/File:1816_summer.png
http://www.giub.unibe.ch/klimet/docs/luterbacheretal_science.pdf
ftp://ftp.ncdc.noaa.gov/pub/data/paleo/historical/europe-seasonal-files/
Interestingly, nothing of this contradicts the information you have presented here. Cold in Europe in the summer of 1816 was centered on Paris, but wasn’t pronounced at the sites you reported in this post. If he was able to reconstruct lower temperature centered around Paris, Luterbacher presumably found records for more sites than you did, or than you chose to share with us. Were you cherry-picking? On the other hand, the map in Wikipedia doesn’t tell us what maps of other European summers looked like in the years decades around Tambora. The Wikipedia map reports temperature anomalies with respect to 1971-2000, so “normal” for this period (the end of the LIA and before increased CO2) would have been lower. It would be fun to see the same maps for a few other years besides 1816. Perhaps the -3 degC anomaly was common in that period.
The real cherry-picking with old records like this happens when alarmists associate ANY unusual LOCAL WEATHER with Tambora, a phenomena that theoretically effected global climate for at least a year. Unusually cold weather can be found somewhere in the world every summer (and even greater extremes will occur every few summers), whether or not a volcano erupts that year. (The same cherry-picking occurs when alarmists associate any local weather with carbon dioxide.)
According to reference 1 from Wikipedia, the “year without a summer” in New England consisted of unusual cold and snow in early June and an early frost in August, which combined with drought to decimate the corn crop. There was a summer – a summer with some very unusually cold weather (mostly in the late spring) – but the drop in average temperature wasn’t unusual. These cold events would have been unusual even if Tambora hadn’t globally lowered temperature by a modest amount.
Why can’t I find any evidence of the Dalton minimum’s cooling in those charts?
John West says:
April 15, 2012 at 9:15 am
Outstanding find, John. Here’s their information in graphic form.
Not only did England not notice in their food prices, they were at their lowest for the decade in 1816 …
I’ve added this to the head post.
w.
Excellent post, Willis! You have identified a prime example of the attribution problem: our desire to explain one noticeable thing by reference another. It is also an ‘exception reporting’ problem similar to that of the witch-lady giving you the evil eye and your pigs getting sick (an actual problem in Bermuda circa 1635).
Hmmm. Sounds like CAGW. But she got hung, we get taxes. Taxes are better.
Other variables NOT taken into account: The amount of grain in stores (literally: grains stored from the previous year) or grain available from other areas where crop failures (to the same degree, like maybe France) were not seen …
If one knew more about ‘markets’ and how grains etc. were handled at that time (storage, movement etc) one could better infer how scarcity/crop failures would affect price.
.
Willis you have to put prices of commodities in the historical perspective at the time:
During the Napoleon’s European experiment there was stockpiling of commodities by British government, increasing price to a large degree. Once Napoleon was defeated in Russia, there was huge surplus available throughout 1815 & 16, but if there was poor harvest in 1816 than prices would shoot up in 1817 again, but possibly not as high because of previous stockpiling.
http://www.vukcevic.talktalk.net/ECP1810-20.htm
Historic perspective : think of oil price during number of Israel-Arab conflicts. In Europe these things are in our DNA.
Willis
Further to my post at 11.54
This is the ref I made in my article (referrring to Britain)
“1816 was known as the year without a summer, snow fell very late and the summer never recovered. The winter proceeding it was severe. A volcanic eruption (Tambora: East Indies) disrupted wind patterns and temperatures greatly, affecting depressions, which tracked further south than usual, making the UK very cold and wet for the summer and beyond. In September the Thames had frozen and snow drifts remained on hills until late July.’
We perhaps need better overall context and cite the works of Turner amongst other artists, who painted the aftermath of volcanic ‘dust veils’ as they were known from 1807 to the 1830’s. It was reckoned that some 15 cubic kilometres of solid matter were thrown into the atmosphere by the Tambora eruption but there had also been major eruptions in 1812 on St vincent and Awu in Celebes and in 1814 in the Philippines.
Lamb wrote about these events on page 246/7 of ‘Climate History and the Modern World’ and cited high levels of dust veils between 1752 and the 1840’s. I have also seen the references to the weather of the era from contemporary accounts held in the Met Office archives. My surmise would be that volcanos did have periodic effects in some months in some places according to the wind patterns of the time.
tonyb
Scott says:
April 15, 2012 at 9:31 am
I just wrote a whole post on that very question, “Dronning Maud Meets the Little Ice Age“. I find the argument that the LIA was volcanically induced to be very weak.
w.
I usually hate to direct anyone to Wikipedia, but if you type in ‘Year without a summer’ quite a lot of history is returned concerning this time and the effects of several cold years. 1816 was the year when certain food exports were prohibited because of shortages, and when farms were abandoned as the farmers became hunters or went to work in the woods. There is enough history of this time to show how hard it was in at least the northern hemisphere..
Not a lesson in European history but a brief reminder
Willis this is map of Europe at 1812
http://images.wikia.com/genealogy/images/d/d4/Europe_map_1812.PNG
If you were running British government what would you do about stockpiling grain?
This is map of Europe just 3 years later in 1815
http://home.zonnet.nl/gerardvonhebel/euro1815.GIF
If you were running British government what would you do about the stockpiled grain?
Historical context is here HUGE factor.
During the Napoleon’s European experiment there was stockpiling of commodities by British government, increasing price to a large degree. Once Napoleon was defeated in Russia, there was huge surplus available throughout 1815 & 16, but if there was poor harvest in 1816 than prices would shoot up in 1817 again, but possibly not as high because of previous stockpiling.
http://www.vukcevic.talktalk.net/ECP1810-20.htm
Kasuha says:
April 15, 2012 at 9:31 am
Thanks, Kasuha. Neither humans, animals, nor plants relate or respond to average temperature. Averages are a mathematical construct with no reality. We relate and respond to actual temperatures. So I’ve used actual temperatures.
Were temperatures cooler in 1816? Sure. Was it significant? Not most places, near as I can tell. See the English food price chart above for how the plants and animals responded.
w.
Scott says:
April 15, 2012 at 9:34 am
Come back when you learn to spell my name, and your posts might get some traction. If you think frosts are the problem, see the English food prices above.
w.
Willis Eschenbach says:
April 15, 2012 at 10:49 am
I don’t believe I was able to find a daily temperature record in New Hampshire. I recall (but can’t find at the moment) that Henry Stommel’s Scientific American article showed a page from Thomas Jefferson’s journal, but that’s from Virginia. Apparently he had more crop damage than I recalled.
There may be some from Massachusetts, but I may be confusing that with weather observations from Cotton Mather from a long time before. I just checked my copy of “The New England Weather Book” by David Ludlum (a legend to New England weather freaks) and he didn’t cover 1816. (He must in other sources because Henry Stommel gives him credit, and, well, it’s inconceivable that Ludlum didn’t write about it. He does give July 9th, 1816 to “Waltham, Mass, frost in low places on July 8th and 9th; 44° at sunrise.” Hmm, July 8th, 1788: Canterbury Conn., hailstorm covered ground 34″ deep; flood followed.” We do have a “Sumptuous variety”!
If you really want, I could stop at the NH Historical Society and check. Their records are not easy to work with, I got the population figures I used from them, I think I had to use some of the towns’ census data.
One nice things about reports of snow, frost, breaking ice on water troughs, is that you don’t need to argue about the quality of the thermometer. There’s a lot to argue about still nights and radiational cooling, e.g., my father thought he saw some frost in a depression very early one July 3rd in central NH one year. Quite possible, we had a wonderful outbreak of dry Canadian air that we appreciated, though the beachgoers didn’t.
Vuk
Your post 12.32. You are right as regards historical context and about stockpiling which added to prices, but to that should be added profiteering (there were several crackdowns in the period) the impact of general economic depressions/booms and ones relating to the rural community, where the sudden withdrawal of men from farming due to the Napoleonic wars and their subsequent release afterwards caused major changes in levels of crop plantings, harvesting and storage, all of which directly impacted on prices in both directions.
http://eprints.worc.ac.uk/365/6/CHAPTER_4.pdf
tonyb
WIllis says: “Nylo, the graphs above show that there were a number of places that didn’t suffer the terrible summer”
I think you are missing the point Nylo was trying to make. When you point out other cold years, you seem to look through a few decade’s worth of data to find a good example of a cold summer. I get 1803, 1809, 1812, 1834, 1843, 1844 and 1844 for the alternate years. But when it was a cold summer in one spot in 1803, was it cold in ALL the other spots that year? We don;t know, because you purposely picked the alternate years to show that there were indeed other cold summers.
But 1816 seems to be below average in ALL the spots you looked at. That is, in itself is significant. 7 sites being below average for random data is definitely statistically significant (1/128). Now, these numbers are certainly correlated (cold weather in one city is correlated with cold weather in nearby cities), so this would lower the statical significance.
Willis, thanks again. Your followups to the comments, as usual, were worth reading. Your followup to the agricultural prices was good.
Aside from the crop lead from Izen, the best of the comments was the following by Frank: The real cherry-picking with old records like this happens when alarmists associate ANY unusual LOCAL WEATHER with Tambora, a phenomena that theoretically effected global climate for at least a year. Unusually cold weather can be found somewhere in the world every summer (and even greater extremes will occur every few summers), whether or not a volcano erupts that year. (The same cherry-picking occurs when alarmists associate any local weather with carbon dioxide.)
I want to reflect on Willis’s points from a couple of different points of view.
First, the “Year without Summer” is legend, but that is not to say it is untrue. Let us assume it as a valid, historical, if qualitative observation of a temporary climate change. If that assumption is valid than Willis has shown that his compressed display of Tmax,Tmin is a very poor proxy to detect that climate change.
Is the problem that Tmax,Tmin temperature records are poor measures of climate change? Heavens! That does attack the foundations of a multi-billion dollar research industry.
Maybe Willis’s display of the temperature records is ill-suited to detect this climate change. In the image that I see, we have 11 years compressed into <440 pixels. So each pixel can represent no less than seven days. Might that compression be hiding something significant? Or over emphasizing outlier data?
As some others here have pointed out, the Year Without Summer, for an agrarian community might be an indication of poor harvest, rather than an overall cooling. A cool wet spring that delayed planting or harmed the seedlings with a late frost could have made for a bad year. An early frost could have drastically cut the farm yield even if an Indian Summer followed. The Year without Summer might have amounted to no more than three cold stormy days in August that ended the growing season and rotted the crops in the field.
First, plot the Year Day Number of every day where Tmin is < Tloss (as Y), Time on X. Where Tloss is -1 deg C, -2 deg C, -3 deg C, etc. Keep only the highest point each Tloss in each year. Line by Year, Color by Tloss. This will be a scatter chart, connected within a year only, but the peak will show the last days of the killing winter or spring frosts. Can we pick out the volcano years on this plot? Cold spring years would show up as peaks.
Also, let’s Plot the temperature records as integrated Growth Degree Days. This is a little difficult because of the question of when is time zero? That seems to be a function of the Tloss in the previous paragraph. Also, there is the issue of Tbase for different crops. However, a first attempt could be to make a line chart with a new line for each (Tloss, and Tbase). The line for each year should end when Tmin < Tfreeze which ends the growth for the year. Poor yield years would show up as lows.
Perhaps he also demonstrates how incomplete knowledge of the ‘economy and markets’ as they apply to commodity pricing can hamper correct deductions and climate inferences too.
.
I have two anecdotal comments. I grew up in Indiana, which became a state by vote of Congress in 1816; the story was that the vote was close, and Indiana almost didn’t make it because of the weather reported from that summer: frost or snow in every month (at least in the northern half of the state), total failure of most grain crops, late frosts killing fruit crops (of which as yet there were few, as the state was still sparsely settled and orchards in most areas had had scant time to grow).
Since moving to Tidewater VA (Norfolk) in 1981, I have sort of absorbed various historical lore from that area and from Richmond, near which we live now. The tale is told that, in one of the winters of the Civil War (1862, I think), Hampton Roads harbor froze over hard enough for railroad tracks to be laid across the ice from Norfolk to Portsmouth, and locomotives with cargo cars attached crossed by those tracks several times before a thaw set in. I don’t know if this actually happened, though it is a good story. Needless to say, we have seen nothing of the sort since 1981.
Britain passed the Corn Laws in 1815, which banned import of wheat. They were specifically intended to raise the price of wheat. You can see in Willis’s graphic above that wheat prices rose after 1815. As for declining food prices I see price declines in more cold tolerant crops such as rye and oats. Indicating farmers switched to these crops after several years of cooler weather.
Vukcevic,
thank you for that very important piece of information. Except for the “Napoleon-driven” increase 25% increase seems typical after large volcanoes. http://virakkraft.com/volcanoes-foodprice.png
Thanks Tony, I think I made my point, Willis can take it or ignore it, the facts won’t change one way or the other whatever they happen to be.