Digital Diatribes

A presentation of data on climate and other stuff

August 2008 Update on Global Temperature – HadCrut

Posted by The Diatribe Guy on August 18, 2008

So far, all the major temperature measures are in general agreement with regard to July 2008 versus June 2008 anomalies. HadCrut is in agreement with GISS in the relative anomaly comparison between July 2008 to July 2007. HadCrut and GISS derive their estimates from surface station (SST) readings (though HadCrut doesn’t publish their methodology and GISS does, as convoluted as it may be). The SST measures differ from the satellite data in the year over year comparison, with the satellites showing more dramatic cooling from last year.

The current HadCrut reading is 40.30, compared to 31.20 in June. This jump is not as large as GISS – it’s not quite a full tenth of a degree Celsius, whereas GISS was a quarter of a degree. However, both of the measures were nearly identical to previous year, with only the smallest of decreases (July 2007 was 40.6 for HadCrut).

As for where July ranks in the HadCrut data, the 40.3 reading is the coolest reading since 2004, and second coolest since 2000. So, for those looking at the short-term trends, anticipating that we are on the trip back down from the recent peak, the data can support that assertion. The data can also support the longer-term perspective of warming, since the current July reading is the eight warmest anomaly in the HadCrut record. I’m sure both sides will use the data as they desire.

In my opinion – as strictly a data guy – I have to subscribe to the shorter-term trend as the current dominant trend, riding a curve of an overall long-term warming trend. The difference in my perspective from traditional AGW warmers is that I believe the real trend is the overall trend that shows minor warming over the long-term that is in no way a concern. There are accelerated warming and cooling cycles around this long-term trend line, and we’ve experience accelerated warming over the last few decades. It makes sense to imagine that there will now be a cooling trend that pulls us back to – and perhaps even under for a time – the overall trend line. That overall trend is presented here:

The consecutive months trend of year-over-year cooling is only at three months, thanks to April’s reading being higher than April 2007. But 12 of the last 13 anomalies have had a cooler reading than the prior year.

The new reading maintained the cooling period from the prior month. It actually removed a month from the beginnning of the trend, while adding a month at the end. Here is that chart:

The last 12-month average anomaly is 29.9, which is the coolest 12-month reading since the period ending October 2001.

The predicted anomaly for the month of July was around 33, so the new anomaly came in a bit higher than predicted, though less than a tenth of a degree off. The projected anomaly for August is about 37 based on my predictive model. Depending on the individual models, the potential range of anomalies is from 19.3 to 48.3. Best estimate is in the 35.9 to 37.0 range.

The long-term predictive ability of the model is still questionable, and I will monitor that. It was built more with the specific error term minimized against the next month’s accuracy only. Still, logic dictates that the next month’s best prediction should present the subsequent month’s best prediction, and so on. Carrying this out a few years shows fairly steady temperature anomalies through the end of 2009, though slightly cooler. We will start to see cooler anomalies in 2010, with some becoming negative. I have only gone through 2012, but the model shows relentlessly negative anomalies from mid-2011 to the end of 2012. Here’s hoping my model’s a bust.

I present the next series of short-term charts for the interested reader. I’m not trying to ignore the longer term trendlines, which show positive slopes on the raw data, while declining slope trends. I am not presenting all of those because they are ultimately pretty similar to the GISS charts (though GISS has higher slope values, indicating more severe warming than HadCrut). In my previous post you can enjoy all the charts for various trend periods.


6 Responses to “August 2008 Update on Global Temperature – HadCrut”

  1. Jeff Id said

    I am a big fan of your site in reference to your data trend analysis. Good for several hundred views myself but I have never commented.

    I ran into an ice core dataset link below.

    This dataset “graph” shows variations much larger than the variations we’re seeing over a short term. It also shows that most of the time the earth is in a much cooler mode. The cyclic nature is obvious as well over 100k year increments.

    Seeing the graph at high resolution (zoomed in) in the last 5k years you can see clearly that the temperature has varied up and down by +/1 1.5degrees C quite often. You can also see the natural filtering of the dataset due to gas diffusion in the ice, so sharp spikes in temperature less than 100 years would be rounded off.

    I am interested in what you might be able to do as far as analysis of the variations in ice core data as compared to what we are seeing from actual measurements.

    I did a simplified review visually myself at

    Part #5 is relevant.

    I can work to find the tabular datasets if necessary. I think an FFT analysis to depict the max frequency and some kind of sigma analysis would be useful to show how our current temperature variation fits in the longer term scheme.

  2. Diatribical Idiot said

    Interesting. Thanks for the comments and the information. Life is pretty busy here and I haven’t had much time to devote to anything beyond the spreadsheets I have up and running, but I always welcome a look at other things from a data persepctive.

    I will bookmark these and get to them when things slow down a bit. Hopefully not too long.

    I have taken a look at the 100,000 cycles in the past “the pacemeker of the ice ages” so to speak, as it relates to CO2 levels, overall temperatures, and ice measures. But I have not done any particular analysis on the numbers myself.

  3. Jeff Id said

    I was up too late when I wrote my comment. I am looking at the data in the last 10k years from the Vostock Ice core. At this point it shows some large temperature variations quite regularly (not as good of a cycle as the 100k year trend.) To me the data looks like it shifts up or down more than 1 degree C every hundred years with very typical variations of 3 degrees peak to peak.

    I have downloaded the “Raw” temperature data and have begun researching the best way to analyze the variation. Visually it seems to me that our current temperature rise would fit within 1 to 1.5 standard deviations of the normal as far as the 100 year slope.

    Now I am researching why this seems so much more frequency and amplitude in the ice core reconstructions than the temperature reconstructions from things like tree rings and pollen. Now I’m really in deep. Why do I do this to myself?

    Can you suggest an appropriate statistical technique for the comparison of our current variation to ice core temperature variation so I could determine how unusual our thermal slope is? I’m sure you have more expertise in this than myself.

  4. Diatribical Idiot said

    I checked out your site, and I really like the write-up on your #5.

    From my understanding of the tree ring analysis, a paper I read and actually reviewed as part of my Landscheidt series by Eddy discusses the correlation of Carbon isotopes in tree rings with sunspots, due to the fact that fewer sunspots means more cosmic rays, which means more isotopes. Fewer sunspots also correlates with lower temperatures if there is low solar activity over a long period of time. However, even in this analysis, one of the real issues with good data analysis is the type of smoothing that occurs as you discussed, but also the lag in response. The carbon isotope studies, for example, really only provide insight into long periods of time of reduced solar activity because there is up to a 40 year lag depending on all sorts of other factors. But then there are situations where a number of influences align and the lag isn’t as long, so there can appear to be more of a spike, even though there may not have been one in reality.

    A couple months ago I took a look at a comparison between GISS and NCDC data and in the course of that looked at both correlation and standard deviation. There just doesn’t appear to be any significant deviation in recent times from the long-term trend of measured temps. The problem with trying to compare current measures to era-long ice core measures is complicated by a number of factors, and I’m not sure I can think of a great method of comparison.

    Another note that I thought of when you saw the spike around 8,000 years ago is a chart I saw in the Landscheidt paper I’ve reviewed. This, again, takes a look at carbon isotopes to assess changes in solar activity, but plots it against the sinusoidal curve of earth’s magnetism. It appears that swings occurred much more wildly both up and down about that time period. It may or may not help explain the spike, but it may be worth reviewing.

    I have been working on a much more complete model of projecting anomalies. It’s become one of those obsessions that probably serves no real purpose, other than I’m just really interested in this data. I’m getting there… I just haven’t posted much because that’s what I’ve been working on. Thanks for continuing to check the site.

    If only sleep weren’t required, I could stay on top of all this better. And now it’s almost football season to boot…

  5. terrain temp said

    Its perhaps a statement of the b******* obvious but when looking at temperature graphs which are linearised over a period of 150 years it should be remembered CO2 emissions have increase exponentially over this period to their current level of 26,400,000,000 (26.4Gt) per year.

    The current relatively cool period can be explained by looking at the pattern of the temperature record of the last thirty years, which is far from linear.

    There seems to be a pattern of rapid jumps after a period of slight cooling. This looks to be strongly correlated with the solar 12 year cycle.

    We are now in a deep solar minimum. If the previous pattern is followed we can expect higher world temperature anomalies in the next few years. The 1998 record will probably be broken.

  6. The Diatribe Guy said

    This post is a little dated, but I don’t disagree with you on the fact that temperature is not linear. The relationship to CO2 is quite suspect, however. If you read subsequent posts, you will see a number of data analyses that I believe help answer this question. First, I have applied sine waves to the data in a best format that greatly increases the obvious cyclicality of the temperature data. So much, in fact, that it becomes plainly obvious that the overall increase in temperature is, in fact, largely linear (probably not exactly, but a very flat curve nonetheless). Cycling around this approximate 0.4 degree linear increase, then are two substantial cycles, each approximately 40 years in length in total, that provide periods of more extreme warming, and then periods of cooling.

    I have also reviewed the sudden jumps in temperature, looking at the UAH data, and explaining the dangers of autocorrelation in observance of trends. And the jumps may well be related to solar cycle, but probably only inasmuch as the ocean cycles are related to the solar cycle. There’s a lot of conjecture there. The issue with sudden jumps simply has to do with the nature of the greenhouse effect of the earth. I am not talking about the increase in greenhouse gases by CO2, I’m just talking about the general nature of our atmosphere. If the oceans warm and heats things up suddenly, it takes a long freakin’ time to cool off, especially if the ocean is in a warm cycle. So, we had this super El Nino in 1998 that was really an aberration, but it occurred during peak PDO and rising AMO cycles, so after that peak we had more energy being thrown into temperature. This has little or nothing to do with CO2, other than a potential argument that cooling didn’t slow as much as it could have. The problem with this argument is that this actually means we have a more consistent climate over the years, while global warming alarmists argue that we see more severe swings.

    And I’ll take your wager. Based on the AMO and PDO, we will see cooling until about 2030. Bank on it. Ridicule me all you want, but everything right now is theory. But it will be realized soon enough to see who is wrong. (On a personal note, my conflict is that I’d like to be right from a pride/math analysis/avoid stupid public policy perspective, but I want to be completely wrong from a personal perspective. I hate cold weather, and I don’t exactly live in the tropics.)

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: