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

Tuesday, December 30, 2008

Nighttime Marine Air Temperature

I’ve moved to WordPress.  This post can now be found at Nighttime Marine Air Temperature

Historical Marine Air Temperature data are also available from the KNMI Climate Explorer website.

It’s most commonly referred to as Nighttime Marine Air Temperature (NMAT). The data presented in this post is the Met Office Historical Marine Air Temperature, version MOHMAT4.3. Refer to the Met Office Hadley Centre Observations Datasets webpage here:

Most presentations of the NMAT are for the Global and Hemispheric data sets. Figure 1 is Figure 3.4 from the IPCC AR4. Its representations of NMAT, what’s identified as HadMat, the green curves in cells A, C, and D, are significantly different from the data provided through the KNMI webpage, probably as result of the differences between the HadMat and the MOHMAT4.3 data sets.

Figure 1

This post illustrates the NMAT for ocean subsets, where possible, in addition to Global and Hemispheric datasets, all based on the MOHMAT4.3 data. The graphs are of the raw data and of the data smoothed with a 37-month running-average filter. All of the graphs end in May 2007, the last month of updates for the KNMI version of the MOHMAT4.3 data.

A minor problem: data is sparse in the late 1910s and the early 1940s for many of the datasets. For this reason, I’ve also included the graphs provided by KNMI in the lower right-hand corner of each illustration. They show the gaps in the data very well.


Figure 2 shows the Global NMAT from October 1904 to May 2007. The first thing that stands out is the peak in NMAT in the early 1940s, which was created by the sizable multiyear (1939/40/41/42) El Nino. (As noted in my post A Different Way to Look at NINO3.4 Data, that El Nino exceeds the intensity of the 1997/98 El Nino if the data is smoothed with a 25-month filter. Refer to Figure 2 in that post.) The second thing: using the smoothed curve as reference, current NMATs are only 0.06 Deg C higher than they were in the early 1940s.

Figure 2


The NMAT data from November 1905 to May 2007 for the Southern Hemisphere is shown in Figure 3. The gap in the data during the early-to-mid 1940s could not have come at a more inopportune time. It obliterated the impact of the multiyear El Nino. Note that there are no blank periods in the Southern Hemisphere NMAT data in the IPCC graphs for those periods. Refer back to Figure 1, cell d. There’s little to no data available, yet the IPCC shows it. Curious. Also note that, again using the smoothed data in Figure 3 as reference, there has been little, if any, rise in Southern Hemisphere NMAT since the mid-1980s.

Figure 3

Figure 4 illustrates the Northern Hemisphere NMAT data from January 1881 to May 2007. Prior to approximately 1905, there are significant breaks in the data, but since then, the data set is complete for the most part. There are a few months missing here and there, but no multiyear periods. Note how insignificant the difference is between the peak in the 1940s and the peak attributable to the 1997/98 El Nino. There’s approximately a 0.04 deg C difference.

Figure 4


The duration and number of breaks in the NMAT data for the South Atlantic, Figure 5, were so long and many it didn’t seem worthwhile to create my own graphs. I’m providing Figure 5, the corresponding graph from the KNMI website, simply as a reference.

Figure 5

The Indian Ocean NMAT data from February 1956 to May 2007 is shown in Figure 6. Highlighted is the dip and rebound from the late-19th Century to the mid-20th, which is consistent with many ERSST.v2 datasets. Also note how the NMAT for the Indian Ocean has not risen since the late 1980s.

Figure 6

Figure 7 shows the North Atlantic NMAT data from January 1867 to May 2007. The three spikes in the raw data stand out. The 1878 spike in NMAT coincides with a significant El Nino, similar the 1940s and the current stretch of elevated NMAT. The recent rise is also aided by Thermohaline Circulation/Meridional Overturning Circulation.

Figure 7

My favorite graph of the MOHMAT4.3 data is that of the Pacific Ocean NMAT from May 1921 to May 2007. Refer to Figure 8. Consider that the Pacific Ocean covers more of the global surface area than all land masses combined plus another Africa. And its NMAT in the early 1940s was higher than present values. It’s an indication of the magnitude of that multiyear El Nino. Somehow, that 1940s El Nino is not as significant in land surface temperature data.

Figure 8


There is little similarity between the NMAT data presented by the IPCC in AR4 and the Hadley Centre NMAT data available from through the KNMI website.

Monday, December 29, 2008

Ocean Cloud Cover Data

I’ve moved to WordPress.  This post can now be found at Ocean Cloud Cover Data
The KNMI Climate Explorer webpages contain a wealth of data. The start page is here:
The Monthly Observation data selection page is here:

Included are Cloud Cover data sets from CRU, ICOADS, and ISCCP. This post provides graphs of Ocean Cloud Cover for the globe and numerous subsets based on ICOADS 2-degree enhanced data. The CDC’s webpage is here:

Though the data does run back as far as 1800, it is sporadic until 1950, which is why the following graphs begin in 1950. The KNMI versions are updated as needed or as requested, so the cloud cover data ends mid-year in 2007. This is the reason the graphs of annual data end in 2006. (It would be nice to see what’s happened over the past two years, but I have not requested an update since this is only a blog, not serious data analysis.) The graphs that follow are also based on the raw okta data, not anomalies. An okta is a unit of cloudiness measurement, where a 0 oktas equals clear sky and 8 oktas equals complete cloud cover. OKTAS DO NOT DIFFERENTIATE BETWEEN CLOUD TYPE.

I will do a batch of anomaly comparisons in an upcoming post.


Figure 1 illustrates annual Global Ocean Cloud Cover from 1950 to 2006. The trend in cloud cover appears steepest from 1950 to 1960. The upward trend decreases from 1960 to the mid-1970s, then increases again until its peak in 1996, the year before the 1997/98 El Nino. After a dip and rebound, the Global Ocean Cloud Cover dropped steadily from 1998 to 2006.
Figure 1


The Ocean Cloud Cover for the Southern Hemisphere is shown in Figure 2. From 1960 to its peak in 1998, the underlying trend appears exponential. Afterwards, Southern Hemisphere Ocean Cloud Cover dropped through 2006.
Figure 2


In Figure 3, the Ocean Cloud Cover for the Northern Hemisphere is shown. Cloud cover rose quickly from 1950 to 1961. The trend flattened significantly from 1961 to 1983 until the year after the El Chichon volcanic eruption. The flat period coincidentally agrees the period of increased volcanic activity. Northern Hemisphere Ocean Cloud Cover rose again, with a minor fluctuation, until 1992. The severe drop in 1993 appears to be the result from the Mount Pinatubo eruption. Ocean Cloud Cover for the Northern Hemisphere then seems to react to ENSO until 2006.
Figure 3


Figure 4 shows the Tropical Ocean Cloud Cover from 1950 to 2006. The underlying trends are difficult to pick out, but Tropical Ocean Cloud Cover appears to accelerate upwards in the early to mid-1980s, peak in 1996 before the 97/98 El Nino, then peak secondarily in 1999 and 2001, before starting to decrease until 2006.
Figure 4


Figure 5 illustrates Ocean Cloud Cover for the Northern and Southern Hemispheres and Globally. Southern Hemisphere Cloud cover is significantly less than that of the Northern Hemisphere.
Figure 5

The Ocean Cloud Cover for the Atlantic, Pacific, and Indian Oceans are shown in Figure 6. The Atlantic Cloud Cover is slightly less than the Pacific, but the Indian is significantly less than the other two.

http://i44.tinypic.com/wvb9rs.jpg Figure 6

East and West Tropical Pacific Ocean Cloud Cover data are illustrated in Figure 7. Before 1986, there was a substantial difference between the two data sets, but in 1987, the Eastern Tropical Pacific Cloud Cover rose, reducing the difference.
Figure 7

What happened in 1987?


I found the apparent impact (assumed based on timing) of volcanic aerosols on Northern Hemisphere Ocean Cloud Cover trends, and only the Northern Hemisphere trends to that extent, to be the most interesting find in the ocean cloud cover data that I've downloaded so far.

Saturday, December 27, 2008

Tropical Cloud Variability

I’ve moved to WordPress.  This post can now be found at Tropical Cloud Variability
While investigating an entirely different topic, I ran across the June 25, 2005 presentation by Joel Norris of the Scripps Institute of Oceanography titled “Multidecadal Tropical Cloud Variability in Observations and GCMs”. It highlights the capabilities (or lack thereof) of GCMs at reproducing cloud cover variability. The presentation was made during the 5th International Conference on the Global Energy and Water Cycle in Costa Mesa, CA.

I was surprised to find that, when I Googled the title in quotes, there were only two responses.

The conclusions read:
•Between 1957-1976 and 1977-1996…
observed upper-level cloud cover increased over the central equatorial Pacific
decreased over the adjacent subtropics
decreased over the western tropical Pacific

•Physically consistent changes occur in precipitation and surface atmospheric circulation.
•These changes are much larger than the expected linear response to the observed increase in Niño3.4 SST.

•CCSM3 AMIP simulations largely reproduce the observed central Pacific upper-level cloud changes, albeit with relatively weaker magnitude.
•CCSM3, CM2.0, and CM2.1 IPCC 20thCentury runs do not reproduce the observed cloud changes.
•Interdecadal cloud variability in the CCSM3 and CM2.0 control runs is much weaker than the observed interdecadal cloud change.
•Interdecadal cloud variability in the CM2.1 control run has more realistic amplitude (but interannual SST variability is too high).

•The CCSM3, CM2.0, and CM2.1 20th Century simulations (if believable) suggest that the observed tropical cloud changes are not a climate response to anthropogenic forcing.
•Multidecadal tropical cloud variability is too weak in CCSM3 and CM2.0.


Correct simulation of cloud cover variability at various altitudes by GCMs should be a critical area of focus. Without it, GCMs can’t properly estimate the impacts of anthropogenic and natural forcings on global and regional climate.

Wednesday, December 24, 2008

Happy Holidays

I’ve moved to WordPress.  This post can now be found at Happy Holidays
I should resume posting again on Friday or Saturday.

Enjoy your Holidays.


Bob Tisdale

Sunday, December 21, 2008

The Indo-Pacific Warm Pool Index

I’ve moved to WordPress.  This post can now be found at The Indo-Pacific Warm Pool Index

I will be using the Indo-Pacific Warm Pool Index (IPWPI) in an upcoming post and I wanted to introduce it beforehand.

The IPWPI is defined as the average monthly SST for the Indo-Pacific Warm Pool greater than or equal to 28 Deg C. The coordinates listed for the Indo-Pacific Warm Pool are 20S-20N, 90E-180E. I ran across it in the PowerPoint presentation “Decadal Variability of the Indo-Pacific Warm Pool and Its Association with Atmospheric and Oceanic Variability in Winter and Summer” by Mehta and Wang, dated May 7, 2007.

In the presentation, Mehta and Wang also appear to have converted the difference between the IPWP monthly mean temperature and 28 Deg C into an anomaly. I have not taken this extra step in the following. Their illustration on Slide 5 also uses annual data from approximately 1945 to present, where I will use monthly data in the following from 1854 to present for the long-term data and from 1978 and 1980 to present for the short term. They also appear to use a different SST data set, possibly HADSST.


Figure 1 shows the IPWPI for the period of January 1854 to November 2008. The monthly data has been smoothed with a 12-month running-average filter. The major shift (~0.3 Deg C) in 1945 looks to be the result of the “bucket adjustment.” Hopefully, that correction will be made soon. Also note the dip and rebound that occurred in early 20th Century. Then there’s rise from the 1940s to present, which is obviously skewed by the bucket adjustment.

Note: The 0.3 Deg C shift is discussed in “A large discontinuity in the mid-twentieth century in observed global-mean surface temperature”, by Thompson et al (2008).

Figure 1

In Figure 2, the period has been shortened to the last 30 years. Note the significant step change in the IPWPI after the 1997/98 El Nino. Working back in time, there’s also a smaller upward step from 1994 to 1997, one from 1986 to 1988, and one from 1979 to 1982.
Figure 2

Adding scaled NINO3.4 data to the graph for timing purposes illustrates the cause of those step changes. Refer to Figure 3.
Figure 3

If the upward step changes in the IPWPI result from El Nino events, why then weren’t there step changes during the 1982/83 and the 1991/92 El Nino events? By overlaying scaled Sato Index data, Figure 4, the impact of the El Chichon and Mount Pinatubo eruptions on the IPWPI becomes apparent. They suppressed the 1982/83 and the 1991/92 El Nino events, and might have impacted the short upsurge in NINO3.4 SST anomaly that occurred during the Spring of 1993. In May of 1993, NINO3.4 SST anomalies reached 1.4 deg C, which is well above the threshold of an El Nino.
Figure 4


It is becoming more evident that the relatively high number of El Nino events since the late 1970s may be caused by the suppression of the El Nino events that occurred at the same time as the El Chichon and Mount Pinatubo volcanic eruptions. During those periods, the volcanic eruptions appear to reduce El Nino heat transfer within the tropics. It would therefore seem likely that the normal export of heat from the tropics poleward would be diminished as well.


The Smith and Reynolds Extended Reconstructed SST (ERSST.v2) data is available through the NOAA National Operational Model Archive & Distribution System (NOMADS).

Sato Mean Optical Thickness data is available from GISS:

Friday, December 19, 2008

The Lingering Effects of the 1997/98 El Nino

I’ve moved to WordPress.  This post can now be found at The Lingering Effects of the 1997/98 El Nino
A New Film

UPDATE - I removed the Blogger version of the video and imbedded the YouTube version.

This film, my second, presents observations about the lingering effects of the 1997/98 El Nino. Its backbone is the animation of Global Sea Surface Height available from the NASA Jet Propulsion Laboratory:
Specifically, It Was Edited From:

It uses that animation and graphs of OI.v2 SST data to illustrate the cause of much of the “Record” warming since 1996, which was a result of the long-lasting impacts of the 1997/98 El Nino on the East Indian and West Pacific Oceans. The reoccurring comparative graphs include:
-Scaled NINO3.4 SST Anomaly, used as reference for timing, and
-East Indian and West Pacific Ocean SST Anomalies.

It ends with comparative graphs of:
-NINO3.4 SST Anomalies VERSUS East Pacific, Atlantic, West Indian Ocean SST Anomalies,
-East Indian and West Pacific Ocean SST Anomalies VERSUS East Pacific, Atlantic, West Indian Ocean SST Anomalies,
-Those Two Data Sets VERSUS Global SST (60S to 65N), and
-Global SST Anomalies VERSUS Global Combined (Land+Sea) Surface Temperature Anomalies.

Copies of the graphs with TinyPic links are below.

The video is presented here in a smaller format than I’d prefer, so I’ve also provided a link to a YouTube copy. It’s larger and can be expanded to full screen. The video is just under 5 minutes in length. It’s longer than I originally planned. And because of its length, I had to shorten the time that some of the narrative remained on screen. Please stop and start, rewind, etc., as needed.


YouTube Link:

The reoccurring graph was the comparative graph of NINO3.4 SST anomalies and SST anomalies for the East Indian and West Pacific Oceans, Figure 1.
Figure 1

The second graph was a comparative graph of NINO3.4 SST anomalies and SST anomalies for the East Pacific, Atlantic, and West Indian Oceans, Figure 2.
Figure 2

The third graph was also a comparative graph. Figure 3 illustrated SST anomalies for the East Indian and West Pacific Oceans and SST anomalies for the East Pacific, Atlantic and West Indian Oceans.
Figure 3

Figure 4 illustrated Global SST anomalies, and SST anomalies for the East Indian and West Pacific Oceans, and SST anomalies for the East Pacific, Atlantic and West Indian Oceans.
Figure 4

And last, Figure 5 compared Global SST anomalies and Global Combined (Land PLUS Sea) Surface Temperature Anomalies, actually the Temperature of the Lower Tropospheric (TLT) from MSU AHU.
Figure 5


The NOAA Optimal Interpolated SST Analysis (OI.v2) Data is also available online through the NOAA National Operational Model Archive & Distribution System (NOMADS).
Instructions for Downloading:

Global Temperature (Temperature of the Lower Troposphere) anomaly data is available through the National Space Science & Technology Center, University of Alabama in Huntsville website, specifically:

Thursday, December 18, 2008

NINO3.4 Data Comparison--HADSST and ERSST.v3

I’ve moved to WordPress.  This post can now be found at NINO3.4 Data Comparison–HADSST and ERSST.v3
In a prior post NINO3.4 Data Comparison – HADSST and ERSST.v2, I illustrated the difference between the HADSST and ERSST.v2 versions of NINO3.4 data. As a follow up to that post, the paper “Improvements to NOAA’s Historical Merged Land–Ocean Surface Temperature Analysis (1880–2006)”, Smith and Reynolds (2008), JOURNAL OF CLIMATE, VOLUME 21, provides an explanation for the difference between the ERSST.v3 and HADSST versions. Refer to page 2293, or pdf document page 11 of 14.

“Changes in the Niño-3.4 SST anomalies between ERSST.v2 and ERSST.v3 are very small after 1950. Earlier in the record the two are also highly correlated, but there are times when the ERSST.v2 anomaly is greatly damped from a lack of sampling (Fig. 8). These times include years before 1880 and around 1918, shown more clearly in the difference (Fig. 8, bottom panel).”

Their Figure 8, both panels, is shown below.

The two versions (ERSST.v2 and ERSST.v3) of the NINO3.4 SST anomalies are similar to one another over the period where I illustrated in the earlier post that HADSST and ERSST.v2 diverge. Therefore, there should also be a significant difference between the HADSST and ERSST.v3 depictions of NINO3.4 SST anomalies. A comparative graph of the HADSST and ERSST.v3 versions of NINO3.4 are illustrated in Figure 1. The difference (ERSST.v3 Minus HADSST) is shown in Figure 2.
Figure 1

Figure 2


The explanation in the linked paper for the difference is:

“Both analyses are clearly producing consistent interannual variations. But there are important differences in this region in periods when sampling is sparse. In Niño-3.4 prior to 1950, HadISST is biased about 0.3°C warmer than ERSST.v3. Much of the bias is due to the use of different historical bias adjustments in the two analyses prior to 1942. Another important difference depends on the method used to compute low frequency variations. In HadISST they are computed by fitting data to a global mode, while here simpler averaging and filtering is used, as discussed above.”


The HADSST version of NINO3.4 SST anomaly data is available through the Royal Netherland Meteorological Institute (KNMI) webpage:

ERSST.v3b data sets are available from the NCDC here:
Note that the NINO data page has an error. The data changes from SST to SST anomaly in 1985.

I used an earlier version in the preparation of this post.

Monday, December 15, 2008

Sea Level Data: Global and Indian, Atlantic, and Pacific Oceans

I’ve moved to WordPress.  This post can now be found at Sea Level Data: Global and Indian, Atlantic, and Pacific Oceans

This post presents graphs of raw, smoothed, and annualized sea level data for Global, Indian Ocean, Atlantic Ocean, and Pacific Ocean data sets. The recently updated data is available through the University of Colorado at Boulder. Here’s a link to their Sea Level Change Overview (index) webpage:
and one to their Time Series Data Page:

In addition, graphs of the annual changes in sea level for those data sets are provided.

Note 1: You’ll note the graphs of the raw and smoothed data include the notation “Smoothed w/ 35-Period Filter”. The number of samplings varies per year, from 35 one year to 38 the next for example. Keep that in mind when viewing the data. To assure that the smoothing did not misrepresent the data, I’ve also presented the annual mean of the data sets in separate graphs.

Note 2: All of the data sets include the seasonal signals and they exclude the inverted-barometer adjustments.


Figure 1 shows the Global Sea Level for the period of December 1992 to September 2008. Global Sea Level appears to have risen at a reasonably constant rate until mid-2005, when the rate of rise decreased.
Figure 1

Looking at the annual data from 1993 to 2007, Figure 2, confirms the sharp deceleration in the rise of Global Sea Level in 2006 and 2007.
Figure 2

Figure 3 illustrates the annual change in Global Sea Level from 1994 to 2007. With the exception of 2006, all values are above zero, indicating that Global Sea Levels are rising. However, the annual changes indicate that the there has been a decrease in the rate of rise over the term of the data. The curves of Global Sea Level, Figure 2 and 3, would not have flattened out if the rate of rise was accelerating.
Figure 3


The Indian Ocean Sea Level data is shown in Figure 4. Based on the smoothed data, the Indian Ocean Sea Level has not increased since early 2007. Also note the multiple swings in sea level during 1996 and 1997, leading up to the El Nino of 1997/98.
Figure 4

Figure 5 illustrates the annual Indian Ocean Sea Level from 1993 to 2007. From 1999 to 2007, Indian Ocean Sea Level appears to have increased exponentially, but a look at the smoothed data in Figure 4 reveals that the rise has ended.
Figure 5

Viewing the annual change in Indian Ocean Sea Level, Figure 6, reveals the significant rise in 1998 associated with the 1997/98 El Nino.
Figure 6


The raw and smoothed sea level data for the Atlantic Ocean is illustrated in Figure 7. The significant drop in 2002 does correspond to a sudden drop in Atlantic Ocean SST anomaly. (Refer to Figure 10 in Atlantic, Indian, and Pacific Ocean SSTs Segmented By Longitude.) Also, the Atlantic Ocean Sea Level does not appear to have risen significantly since 2005 based on the smoothed data.
Figure 7

This is confirmed by the annual Atlantic Ocean Sea Level data. Refer to Figure 8. Atlantic Ocean Sea Levels remained relatively flat in 2006, then dropped in 2007.
Figure 8

The variability of the annual change in Atlantic Ocean Sea Level is illustrated in Figure 9.
Figure 9


The smoothed sea level data for the Pacific Ocean, Figure 10, also appears to have flattened and decreased in recent years.

Note that in 1997 the peak in raw data of sea level (approximately June/July) preceded the El Nino peak, which occurred in October 1997 (not illustrated). And note the second large perturbation in late 2002/early 2003 that corresponds with the El Nino at that time.

It is unfortunate that the number and timing of sea level readings do not correspond with the weekly or monthly values presented for SST so that comparative graphs of Pacific Sea Level versus Pacific and NINO3.4 SST anomalies could be presented. It would be interesting to compare the timing of the rises and falls in the different data sets.
Figure 10

Figure 11 illustrates the annual Pacific Ocean Sea Level data from 1993 to 2007. The impacts of the 1997/98 El Nino and subsequent multiyear La Nina are visible, as are the rise in 2002 and fall in 2003, which no longer appear disproportionate. Note that the Pacific Ocean Sea Level peaks in 2005 and has fallen in 2006 and 2007.
Figure 11

Figure 12 illustrates the annual changes in Pacific Ocean Sea Level and SST anomaly.
Figure 12


Thanks to Bill Illis for noting the flattening of the Pacific and Atlantic Ocean Sea Level curves in a comment on an earlier thread.


The source of the sea level data is discussed in the Introduction.

Saturday, December 13, 2008

Does Anyone Recall Any Other Data Sets with Curves of this Same Basic Shape?

I’ve moved to WordPress.  This post can now be found at Does Anyone Recall Any Other Data Sets with Curves of this Same Basic Shape?
This is quick post about an unusual similarity between the curves of two entirely different data sets.

The Armagh Observatory has numerous data sets unique to their location. One such data set is the hours of sunshine from 1880 to 2004. For the following graph, Figure 1, I converted their seasonal data to annual hours of sunshine. The data starts on page 86 of the paper titled “Meteorological Data recorded at Armagh Observatory from 1795 to 2004: Volume 10 - Daily, Monthly, Seasonal and Annual Hours of Bright Sunshine 1880-2004”, Butler et al.

Figure 1

The shape of the curve caught my eye, so I inverted it, Figure 2.
Figure 2

Look familiar?
Figure 3

Ponder that for a while.
Does anyone recall any other curves with the same basic shape?


The data bank of the Armagh Obsrvatory is found here:

The recently updated version of the Extended Reconstructed SST (ERSST.v3b), along with land surface temperature and combined (land + ocean) surface temperatures, are available in various latitudinal bands at:
The overview for the update is here:

Friday, December 12, 2008

The Step Change in HADSST Data After the 1997/98 El Nino

I’ve moved to WordPress.  This post can now be found at The Step Change in HADSST Data After the 1997/98 El Nino
When compared to the ERSST.v2, ERSST.v3b, and OI.v2 SST data sets (Figures 1, 2, and 3), there is a significant step in the HADSST global SST anomalies after the 1997/98 El Nino event. I’ve noted this in comments on numerous blogs over the past year, but have chosen not to offer an explanation for what appears to be the reason for the difference.

Figure 1
Figure 2
Figure 3


The Hadley Centre changed data sources for SST at 1998. [Note: That date was originally posted by me as 1978. I've corrected it.] This quote is from the Met Office’s Hadley Centre about the HADSST2 data set:

“Brief description of the data
“The SST data are taken from the International Comprehensive Ocean-Atmosphere Data Set, ICOADS, from 1850 to 1997 and from the NCEP-GTS from 1998 to the present.”

The following ICOADS link will bring you to the following explanation:

“The total period of record is currently 1784-May 2007 (Release 2.4), such that the observations and products are drawn from two separate archives (Project Status). ICOADS is supplemented by NCEP Real-time data (1991-date; limited products, NOT FULLY CONSISTENT WITH ICOADS).” [Emphasis added.]

The change of data set also helps explain why HADCRUT3 Global, Northern Hemisphere, and Southern Hemisphere data sets consistently run high since the 1997/98 El Nino when compared to other land and sea surface temperature data sets.


HADSST2GL data is available from the CRU website:

Extended Reconstructed SST Sea Surface Temperature Data (ERSST.v2) and the Optimally Interpolated Sea Surface Temperature Data (OI.v2 SST) are available through the NOAA National Operational Model Archive & Distribution System (NOMADS).

The recently updated version of the Extended Reconstructed SST (ERSST.v3b), along with land surface temperature and combined (land + ocean) surface temperatures, are available in various latitudinal bands at:
Don’t let the PDO in the address confuse you. There’s much more there. The overview for the update is here:http://www.ncdc.noaa.gov/oa/climate/research/sst/ersstv3.php

Thursday, December 11, 2008

Unheralded Changes in ERSST.v3 Data

I’ve moved to WordPress.  This post can now be found at Unheralded Changes in ERSST.v3 Data
This’ll be a quick post with only two illustrations.

Back in an August 8 post, My Curse on Data Set Updates, I noted the lack of updates to the ERSST.v3 data. At that time, the data ended in April 2008. The updates resumed last month, but I didn’t think much about it, other than I was pleased that the newer data set had been made current. (I’ll remove that August 8 post in a few days, since both data sets resumed their updates.) Then, in a comment here, Bill Illis noted that he’d read a recent remark at another website about a change in SST data. Yesterday, I noticed a very minor change in the webpage address for the monthly global SST data. There was a “v3b” after the data set name, as in ersstv3b.
The old address for the same data excluded the “v3b”. (The following link no longer functions.)

Did version “B” include the “bucket correction” corrections? Refer to “A large discontinuity in the mid-twentieth century in observed global-mean surface temperature”, by Thompson et al (2008):http://www.nature.com/nature/journal/v453/n7195/abs/nature06982.html

NOPE! It doesn’t include them.

Figure 1 is a comparative graph of the ERSST.v3 (old) and ERSST.v3b (new) versions of the Global SST anomaly data. The data has been smoothed with a 37-month filter. There are very minor changes when viewing the global data set.
Figure 1

Subtracting the ERSST.v3 (old) from the ERSST.v3b (new) data presents a better view of the differences. Globally, the update appears to have created a very minor increase in trend, ~0.022 deg C per Century.
Figure 2

Are the corrections small changes in all oceans or are they major changes in small data sets such as the Arctic and Southern Oceans? The updates aren’t mentioned in the main page of the data set.

We’ll find out in a future post. (I haven’t a clue, but my guess is that the changes are in the high-latitudes of the Southern Hemisphere.)

Thanks to Bill Illis for calling my attention to the update.


…are included in text of post.

Wednesday, December 10, 2008

Dip and Rebound Part 2

I’ve moved to WordPress.  This post can now be found at Dip and Rebound Part 2

I do occasionally receive comments to my posts. Some agree with the post, some point me in a different direction, and some disagree. A blogger going by the initials JC replied to my recent post Dip and Rebound and advised:
“There's been several studies examining why climate changed so dramatically in the early 20th century when forcing from greenhouse gases was low. Eg - Tett 2002 and Reid 1997. The dip and rebound was largely due to volcanic and solar forcing. Eg - strong volcanic activity in the late 19th century had a strong cooling effect coupled with a drop in solar activity. In the early 20th Century, volcanic activity dropped and solar activity rose.”
My reply:
“Thanks for the links, but a decline in solar forcing can’t account for the decrease in SST anomalies--not with the present opinions about climate sensitivities to changes in TSI. For SST’s to drop more than 0.3 deg C over 40 years, TSI would have had to have dropped more than 3 times the average of the trough-to-peak range of the last 3 solar cycles. That didn’t happen.“The effects of volcanic aerosols last only for a few years and would not be able to cause the decrease in SST over the 40 years illustrated in the ERSST.v2 version of Global SST anomalies. Note the timing of the decreases in global mean forcings illustrated in Figure 1 of the Hadley Centre paper you linked. They do not coincide with the dip and rebound. Therefore, volcanic aerosols do not work. Combining solar and volcanic aerosols won’t work either, because most of the change occurs in the Northern Hempisphere, confirming that it is not related to solar or volcanic aerosol forcings. “Again, the dip and rebound remains unaccounted for by the IPCC and CCSP. Now I can add the Hadley Centre to the list. Thanks.”
## ##
Much of the dip and rebound (Figure 1) can be explained with Thermohaline Circulation/Meridional Overturning Circulation, as I noted in a latter comment, but regardless of whether or not THC/MOC can or can’t be used, the fact remains that the IPCC and CCSP are using a natural variation (the dip) to amplify their claims of anthropogenic global warming.
Figure 1

The intent of this post is to further illustrate how volcanic aerosol and solar variations cannot account for the dip and rebound.


Figure 2 is a comparative graph of long-term global SST anomaly data (ERSST.v2) and the Sato Index of Mean Optical Thickness, which represent the effects of stratospheric aerosols from explosive volcanic eruptions. The Sato Index data has been scaled to produce a 0.35 deg C response to the Mount Pinatubo eruption, which is the midpoint of the range of 0.2 to 0.5 deg C of proposed values. As noted earlier in this thread and again in the illustration, the effects of volcanic aerosols last for only a few years after the eruption. This is reflected in the decay rates of the Sato Index. Note also that the start (~1870) of the decrease in SST anomaly is well before the Krakatau eruption in 1883 and that the decrease in SST changes direction (~1910) during the period of volcanic activity, which subsides ~1914.
Figure 2

The thought that volcanic eruptions are responsible for the decrease in SST anomalies from ~1870 to ~1910 doesn’t work.


Figure 3 illustrates monthly Sunspot Number from 1854 to 2008. The data is raw and smoothed with a 133-month filter. The period in question is marked. Since I have yet to find monthly TSI data over the term of the SST data, I’ll use sunspot number as a proxy for TSI.
Figure 3

In Figure 4, I’ve simply scaled the Sunspot data to the current consensus of a 0.1 deg C response in global SST to the trough to peak rise in solar irradiance for Solar Cycles 21 and 22. Even if the actual response is twice that number, it makes little difference as will be illustrated in the following. I’ll key on the smoothed data to make my point, since SST response to changes in solar irradiance is so long. The response time varies depending on the researcher and there’s considerable debate about it, so I won’t list a specific response time.
Figure 4

The raw sunspot cycle data has been removed in Figure 5, to eliminate the distraction and skewing of the scale. There is in fact a decrease in the smoothed solar impact data from ~1870 to ~1900, which might be the basis for the thoughts that the reduction in TSI is responsible for the decline in SST anomaly. BUT the scale is wrong. The reaction of SST, based on the current understanding of climate sensitivities to solar variations, is off by an order of magnitude, at least.
Figure 5

If the global SST anomaly data is compared to the scaled data of what is supposed to be the SST response to changes in Solar Irradiance, Figure 6, the failings of this line of though becomes obvious.
Figure 6

I wanted to determine the scaling factor required for the change in TSI to cause the drop in SST anomalies, so I took the comparison one step farther in Figure 7. To adjust the curve of the Sunspot Numbers, which have been scaled to SST impact, I needed to multiply the data by 22. Also, the period of 1950 to 1990 becomes difficult to explain.
Figure 7

The though that solar irradiance is the cause of the dip and rebound is fatally flawed.


Sea Surface Temperature Data is Smith and Reynolds Extended Reconstructed SST (ERSST.v2) available through the NOAA National Operational Model Archive & Distribution System (NOMADS).

The Sato Index data is available from GISS.

The monthly Sunspot data from January 1749 to October 2008 is available through NASA’s Marshall Space Flight Center.


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.