I’ve moved to WordPress: http://bobtisdale.wordpress.com/

Monday, March 23, 2009

The Large 1945 SST Discontinuity Also Appears in Cloud Cover and Marine Air Temperature Data

And A Curious Long-Term Dataset

UPDATE – March 24, 2009

I have acquired and read the Thompson et al (2008) paper “A large discontinuity in the mid-twentieth century in observed global-mean surface temperature”, and the follow-up articles “Hot Questions of Temperature Bias” by Forest and Reynolds and “Climate Anomaly is an Artifact” by Schiermeier. The similarities between the discontinuity in the Sea Surface Temperature data and those same shifts in Cloud Cover and Marine Air Temperature datasets were not addressed in any of the discussions.


Almost a year ago, the Thompson et al (2008) paper “A large discontinuity in the mid-twentieth century in observed global-mean surface temperature” caused a buzz in the climate change blogs. Thompson et al discussed the believed reasons for the sudden shift in global mean surface temperature in and around 1945.

Data sets used to monitor the Earth's climate indicate that the surface of the Earth warmed from ~1910 to 1940, cooled slightly from ~1940 to 1970, and then warmed markedly from ~1970 onward. The weak cooling apparent in the middle part of the century has been interpreted in the context of a variety of physical factors, such as atmosphere–ocean interactions and anthropogenic emissions of sulphate aerosols. Here we call attention to a previously overlooked discontinuity in the record at 1945, which is a prominent feature of the cooling trend in the mid-twentieth century. The discontinuity is evident in published versions of the global-mean temperature time series, but stands out more clearly after the data are filtered for the effects of internal climate variability. We argue that the abrupt temperature drop of ~0.3 deg C in 1945 is the apparent result of uncorrected instrumental biases in the sea surface temperature record. Corrections for the discontinuity are expected to alter the character of mid-twentieth century temperature variability but not estimates of the century-long trend in global-mean temperatures.

The “abrupt temperature drop” is visible in Figure 1, which illustrates the Global SST Anomalies (HADISST) from January 1930 to December 1959.

Figure 1

In the “The Independent”…
…Science Editor, Steve Connor, wrote, “The scientists point out that the British measurements were taken by throwing canvas buckets over the side and hauling water up to the deck for temperatures to be measured by immersing a thermometer for several minutes, which would result in a slightly cooler record because of evaporation from the bucket.

“The preferred American method was to take the temperature of the water sucked in by intake pipes to cool the ships' engines. Those records would be slightly warmer than the actual temperature of the sea because of the heat from the ship, the scientists said.

“Taking into account the difference in the way of measuring sea-surface temperatures, and the sudden increase in the proportion of British ships taking the measurements after the war, the result was an artificial lowering of the global average temperature by about 0.2C, said Professor Phil Jones of the University of East Anglia in Norwich.”

“The Independent” included a graph illustrating the effect the new corrections would have on the global temperature record. Refer to Figure 2.
Figure 2


And it’s doubtful that the “difference in the way of measuring sea-surface temperatures” (Phil Jones) impacted Cloud Cover and Marine Air Temperature measurements.

Figure 3 illustrates the COADS Global Ocean Cloud Cover data from Jan 1939 to December 1959. Its shift in 1945 is pronounced.
Figure 3

The MOHMAT4.3 Global Marine Air Temperature for the same period contains the same shift, as shown in Figure 4.
Figure 4

As does the COADS air temperature data illustrated in Figure 5.
Figure 5

Did the researchers consider the same shifts in other variables before deciding the SST data was in error?


I always download data for the entire term of the dataset, regardless of the period I’m investigating. It’s very easy then to smooth the additional data and plot the results. The long-term COADS Cloud Cover and the MOHMAT4.3 Global Marine Air Temperature data (smoothed with 37-month running-average filters) are shown in Figures 6 and 7. The magnitudes of the surges in the early-to-mid 1940s are clearly visible.
Figure 6
Figure 7

Now the smoothed COADS Air Temperature data from 1800 to 2007, Figure 8. I had to download the data a second time to assure that I hadn’t made an error. It clearly shows that the COADS Air Temperature data was higher in 1800 than it is today.
Figure 8

The raw COADS Air Temperature data, Figure 9, shows part of the reason for the anomalous curve. The data before 1850 appears unreliable, probably a function of measurement accuracy and data density, or lack thereof.
Figure 9

But if the data from 1800 to 1849 is removed, Figure 10, the COADS Air Temperature data continues to show early anomalies that are higher than present anomalies.
Figure 10


Not if we consider SST temperature reconstructions. The Subtropical South Pacific SST Reconstruction of Lindsey et al (2000) was discussed in my post on SST Reconstructions. It shows similar results with the early years warmer than present, Figure 11.
Figure 11

And the Mann et al NINO3 SST Reconstruction, when smoothed with a 30-year Gaussian-weighted filter as created by Jones et al (2001), shows a similar decrease in SST anomalies until about 1900. Refer to Figure 12. The Jones et al smoothing of the Mann data (and other ENSO Reconstructions and data) was discussed in my post Low Frequency ENSO Oscillations.

Figure 12

Unless the datasets were used to infill one another during the 1940s, or unless the "bucket adjustments" also somehow magically apply to Marine Air Temperature and Cloud Cover data,
the similarities in the shifts of the SST , the Cloud Cover and the Marine Air Temperature datasets would make one question the conclusions of the Thompson et al (2008) paper “A large discontinuity in the mid-twentieth century in observed global-mean surface temperature”.


The HADISST, COADS Ocean Cloud Cover, MOHMAT4.3 Marine Air Temperature, and COADS Air Temperature data are available through the KNMI Climate Explorer website:

Refer to the linked posts for the sources of the reconstruction datasets.


Anonymous said...

Wow. As usual, Bob, great work. It's hard to believe that climate variability can happen so quickly...and without anthropogenic forcings!

Keep up the great work!

Anonymous said...

I am a little confused by this statement:

Based on the similarities in the shifts of the Cloud Cover and Marine Air Temperature datasets, the Thompson et al (2008) paper “A large discontinuity in the mid-twentieth century in observed global-mean surface temperature” appears to be in error.

Should it say that it does not appear to be an error?

Bob Tisdale said...

Pearland Aggie: Thanks for noting that I needed to clarify my closing comment. I've revised it and included "Unless the datasets were used to infill one another during the 1940s, or unless the "bucket adjustments" somehow also apply to Marine Air Temperature and Cloud Cover data..."


Tips are now being accepted.

Comment Policy, SST Posts, and Notes

Comments that are political in nature or that have nothing to do with the post will be deleted.
The Smith and Reynolds SST Posts DOES NOT LIST ALL SST POSTS. I stopped using ERSST.v2 data for SST when NOAA deleted it from NOMADS early in 2009.

Please use the search feature in the upper left-hand corner of the page for posts on specific subjects.
NOTE: I’ve discovered that some of the links to older posts provide blank pages. While it’s possible to access that post by scrolling through the history, that’s time consuming. There’s a quick fix for the problem, so if you run into an absent post, please advise me. Thanks.
If you use the graphs, please cite or link to the address of the blog post or this website.