Winter months in the Northern Hemisphere are responsible for most of the observed land warming in CRUTEM4. The summer months (June, July, August, September) essentially show no warming at all from 1850 to 1990. Overall, NH winters have warmed about twice as much as NH summers.
Winter(blue) is the average anomaly for Nov,Dec,Jan,Feb. Summer(red) for June,July,August,September
The same exercise for the southern hemisphere shows very little difference where of course the seasonal months are inverted. So why is the NH different and why?
Summer(blue) is (Nov,Dec,Jan,Feb) whereas Winter(red) is (June,July,August,September)
The SH is dominated by oceans and the coverage before 1900 was very poor. The SH is also sparsely populated having less human impact on the environment than the NH. Clearly the NH effect is real and shows that the seasonal variation in temperature between summer and winter has reduced for some reason since the 19th century. This reduction is mostly due to an increase in minimum winter temperatures rather than maximum summer temperatures.
This story is reminiscient of the observed reduction in the Diurnal Temperature Range (DTR). Several papers have shown that most ‘global warming’ has occured through an increase in minimum night-time temperatures rather than any increase in maximum daytime temperatures. Could there be an anthropogenic cause for both effects?
Perhaps an increase in winter clouds reduces winter heat loss, thereby reducing extreme cold temperatures. Could perhaps coal burning for indudstry and domestic heating prior to 1960 be partly responsible?
Is the slow increase in solar activity over the 20th century to blame? An increase in solar activity leads to a higher flux of UV producing more stratospheric ozone. This warms the stratosphere transfering heat to the dark polar winter as the lapse rate collapses. Could a new Maunder like solar minima bring back freezing winters to the NH?
Any other ideas?
A very new paper shows that a relationship exists between polar temperaures and solar activity in the tropics. Thanks to @@andrw100 Andrew Willias for this one which has just been accepted for publication in in Journal of Geophysical Research.
Stratospheric response to intraseasonal changes in incoming solar radiation
C. I. Garfinkel, V. Silverman,N. Harnik,C. Haspel,Y. Riz
Abstract
Two other things to consider,
1. UHI not being allowed for correctly.
2. improper data Adjustments.
Taking the case of the UK, it is much cooler now than during the 1997/8 El Nino period and also the mid 1970s, but not according to the “Quality Adjusted” data.
You’re right. The temperature in UK has hardly changed. Take this example – Portsmouth.
The red line in the top graph is the ‘corrections’ made by NOAA. They consistently subtract 0.5C from the past to generate a trend. CRUTEM4 shows no trend at all.
Anyway the UK would benefit from a 2C rise in temperatures. Summer weather might improve ! Growing season gets longer. English wines get better. What’s not to like about it ?
Your cloud cover hypothesis could be testable, at least to some extent, because ( if I remember correctly ?) cloud cover was/is one of the items recorded on an hourly basis by the meteo stations.
Bob,
The cloud data at least over the UK shows a correlation with surface temperatures.
https://clivebest.com/blog/?p=5353
cheers
Of course it does!
I know 🙂
Clive
Thanks for reminding me of that earlier post, which I thought was very good at the time.
So the question now arises, could you gain any more insights into the present post by constructing two time series for cloud cover ( – sunshine hours), one for the winter months and one for the summer months, and looking at the trends?
The earlier post shows quite a significant increase in sunshine hours from about 1360 to about 1460, but we don’t know if it is occurring in winter months, summer months , or spread evenly through the year.
Bob P.
Probably right but I would need cloud data for the whole world, NH and SH. ISCP data shows a decrease, but the ICCP still reject the results based on instrumentation issues even though the scientists involved now say these are corrected.
So the problem is finding the data.
I like the change in UV argument due to an increase in solar activity. This is now decreasing to a ‘Maunder” like minimum. If this is the cause then winters should get much colder again within 10 years.
Is it possible to use satellite data for this also? Seems to be a mandatory check nowadays, with so much fiddling of the surface temperature records going on.
I did the check with UAH and interestingly enough there is very small difference with winter showing slightly more warming. This means that it is purely a land based effect.
Clive, I seem to remember you doing some work in the past on Ice Ages.
As the Sun is going quieter and the Earth is cooling (real values, not Adjusted ones) we also appear to be getting a lot of Volcanic activity.
Could this Combination of Quiet Sun, ADO & AMO going negative and Volcanic activity be what starts new Ice Ages?
ie it is a timing issue.
We are actually at a minimum of summer insolation fot the northern hemisphere in the overall Milankovitch cycle, so the timing is right for another glacial period. However this current eccentricity cycle is at a minimum of the overall 400,000 year cycle, so some people thing this means that the current inter-glacial will last another 5-10,000 years. I am not convinced since the net effect of eccentricity is small. I am fairly sure other factors are involved – so yes anything is possible. No-one really understands everything about what triggers inter-glacials and what brings them to an end.
Clive-The winter warming versus summer warming thing is an interesting issue. Some details that any explanation should try to account for:
Winter warming has been strongest, generally, in dry anticyclonic (ie high sea level pressure) air masses:
Michaels, Patrick J., et al. “Observed warming in cold anticyclones.” Climate Research 14.1 (2000): 1-6.
Second, at least in the US in particular, warming has been greatest in the form of much warmer extremely cold days (that is, the coldest day of the year, whatever that day happens to be in a given year, has tended to be much warmer than it was in the late 1970s. Note that this is not “coldest date of the year, that is, we’re not comparing, say, Christmases to Christmases or Groundhog days to Groundhog days, we’re talking directly about the trend in the temperature extremes):
Knappenberger, Paul C., Patrick J. Michaels, and Robert E. Davis. “Nature of observed temperature changes across the United States during the 20th century.” Climate Research 17.1 (2001): 45-53.
I was able to basically verify both of these results myself. For the first, for example, I created these maps using GISS’s map tool:
Which more or less confirmed the spatial pattern found in the Michaels et al. paper (though I need to update to 2014)
And (using reanalysis data) I analyzed the warming rate in the US by temperature day rank within the year, and the relationship between average annual temperature and the temperature by day rank:
https://devoidofnulls.wordpress.com/2014/01/09/lessing-colding/
In general, in the US, when it gets warmer, it does so most strongly on the very coldest days. I think you’d probably find this is true in most Northern Hemisphere land areas where you could find data of adequate resolution.
Thise are interesting maps. This combination of a warming of the coldest days/nights of the year (winter) and concetrated in the coldest regions is striking.
I can’t help thinking that man has had an influence simply by keeping warm in extreme environments. To raise the temperature from -30C to -29C is far easier than raising +29C to +30C. Weather stations tend to be near habitation even remote one’s. People drive trucks up to them to collect data and service them. The growth of remote communities and transport may have contributed to the apparent warming. The UHI effect may be amplified in very cold areas.
I think it might just be physics. Because radiative heat loss is proportional to the fourth power of the temperature, it takes much less temperature change to balance out an additional increment of heat input when the body in question is already at a relatively high temperature. Of course, the Earth’s surface doesn’t mostly cool directly by radiation, but since water vapor would, as a greenhouse gas, tends to inhibit radiative cooling, that might explain why it’s not just the coldest, but especially the coldest and driest air mass which have had to warm the most. That should be true whether warming is induced by additional heat input-solar changes, waste heat from human activities as in UHI-or an inhibited rate of heat loss-such as CO2, but also anything that might inhibit other forms of heat loss, at least for surface changes or local changes.
Andrew,
I was thinking it was maybe the other way round. You need less temperature change when it’s cold to balance a given forcing DS
Climate sensitivity depends on temperature. For the same forcing DS (eg. doubling of CO2) the response (TCR) for two different initial temperatures T1 and T2 would be
DT2/DT1 = T2^3/T1^3
TCR(T2) = TCR(T1) * (T2^3/T1^3)
for T2 = 300K and T1 = 270K
TCR(Tropics) = 1.37*TCR(Arctic)
cheers
Here’s how I was thinking of it:
A black body at 270 K emits: 301.347 W/m^2, so to maintain that temperature it needs to receive the same amount of input energy. One additional W/m^2 of input is balanced by warming to about 270.22 K. On the other hand, a black body at 300 K emits about 459.300 W/m^2 so the temperature it needs to change to to balance an additional W/m^2 of energy input is about 300.16 K. So that’s the inverse of the factor you got.
Your right I got it the wrong way round!
For the same forcing DS
DS = 4.eps.sigma.T1^3.DT1 = 4.eps.sigma.T2^3.DT2
DT2 = (T1^3/T2^3) DT1
If T1 = 270C and T2 = 300C
DT2 = 0.73 * DT1
Clive Best,
Saw your comment about PUDAHUEL this morning over at Judith’s. Even though I created a WP account I still wasn’t able to reply to your comment over there. Hope you don’t mind my posting it here.
——————
clivebest | August 16, 2015 at 3:34 am
You might be interested in my comments in an earlier thread concerning duplicated data between stations. PUDAHUEL is one of them.
http://judithcurry.com/2015/02/09/berkeley-earth-raw-versus-adjusted-temperature-data/#comment-673206
I notified GHCN about the problem several months ago concerning the hundreds of years of duplicate data for various station pairs and they replicated my results. This was well prior to their new v3.3. They condescended to fix one of the station pairs having duplicates in v3.3, but have retained almost all the remaining duplicates.
————————————-
I can only infer retaining the duplicate data is necessary for some purpose. Perhaps it is related to your concerns.
Bob,
Thanks for the insight. That sounds very intriguing. First duplicate readings across nearby stations and then correct those which do not agree with regional trends.
…
Here is graphic of station comparisons I did a few months ago
-9999 comparisons are ignored. If a month was within 0.05C of the same month and year at another station I considered it a match. This is because USHCN reports temperature to 100th of a degree while most of the rest of the world reports only to a 10th of a degree. If I do exact matches the figures are marginally lower and USHCN doesn’t even have a match of 7 months. There are 3 times as many comparisons for USHCN compared to the rest of the world due to number of stations and length of the records.
It is literally trillions to one against two stations having 12 months of identical data during the same year. I consider it irresponsible of GHCN not to maintain as accurate a dataset as possible. Especially since they have known about it for months.
Maybe I should write up a short post and see if Anthony is interested.
Your result implies that there are about 200 stations with identical temperature measurements for 12 consecutive months in GHCN uncorrected. This would introduce a huge bias in the ‘Pairwise Homogenization’ algorithm that NCDC use to auto-correct the data to form V3C – the correcetd data. The reason is that the identical pairs will drag all close neighbours towards the same trend as the pair.
We should look into where geographically these identical pairs are located. If they indeed are in close proximity then I think it is potentially a very important bias in the whole global temperature analysis.
Yes I think it is worth writing this up.
In fall and winter, after harvest, cultivated land tends toward being darker and more absorptive (barren actually) than “wild” areas which have some kind of ground cover mostly of a lighter color (dry grass, fallen leaves etc.). During the period of the “pause”, farming has changed from fall plowing (very dark ground) to minimum tillage (much stalks/stubble lying on top). The acreage involved in the NH is enormous. Not much is made of this although, at first glance, it seems to match the warmer winter scenario and the “pause”. Does anyone know of any studies on this?
That is an interesting idea. Before about 1950 all farmers practiced crop rotation which would have left most fields green over winter. With the advent of mechanised farming and fertilizers this would have changed the surface albedo. Likewise any major change in farming practice such as that you describe would affect absorbed radiation. Furthermore the albedo would only need to change by a very small fraction to produce the pause. I am unaware if anyone has looked at this bu vegetation is monitored by satellite since the 1970s.
Clive, from a post by Willis at WUWT I have looked at this Site which was linked in a comment.
It is an awful lot of reading and quite heavy going.
But this guy does seem to know what he is talking about in great depth.
Perhaps you would like to take a look and see what you think.
http://climatorealist.neowordpress.fr/
That guy takes a Rudyard Kipling approach of intricate “just-so stories” to explain climate index behaviors. The proper way to do this is to eschew obfuscation and just apply first-order physics and see if the primary behavior emerges.
http://contextearth.com/2015/08/22/baby-d-model/
First-order physics works well for ENSO.