January 8, 2008

Musings on Satellite Temperatures

(January 9, 2008: An important update has been added to the bottom of this post)

A couple of interesting items have come to light recently regarding the temperature of the earth’s lower atmosphere as measured by satellites. Here we run briefly run through some of them, in no particular order.

1) Over at the blog site Climate Science, run by Dr. Roger Pielke, Sr., there is a piece alerting everyone to an “in press” article in the publication Journal of Geophysical Research, summarizing research by the Univeristy of Arizona’s Robb Randall and Benjamin Herman that further examines the existing differences between the lower tropospheric temperature histories developed by research teams at the University of Alabama (UAH) and at Remote Sensing Systems (RSS). For those new to the topic, producing temperatures from Microwave Sounding Unit (MSU) data gathered from orbiting satellites is not straightforward for a number of reasons, and at least two methodologies have been developed to do so (one by UAH and another by RSS) which do not produce identical results. The RSS data indicates that globally, temperatures in the lower atmosphere are warming up at a rate of about 0.17ºC/decade from January 1979 through December 2007, while the UAH data warm up at a slightly slower rate of 0.14ºC/decade. While the difference between them has been periodically examined, Randall and Herman add new information by examining shorter period trends within the overall record and comparing the trends with the best data available from a collection of weather balloon observations made during the same periods. Randall and Herman conclude that the RSS data incorporate an improper handling of “diurnal cycle effects” that permeates their methodology and imparts a small warm bias over global land areas. Thus, the slower warming rate in the UAH version of the satellite-derived temperature history of the earth’s lower atmosphere is probably a bit closer to the actual trends than is the RSS, faster-warming version.

2) Having just said that the UAH data is more accurate than the RSS data, comes the following disclaimer…in recent months/years, there is a warming bias from, you guessed it—diurnal cycle effects—that has crept into the UAH data, is progressively getting worse and has not yet been removed. It appears that the RSS data have a better handle on the current data, at least for the time being. In fact, in recent months, the UAH lower atmospheric temperatures may be as much as 0.2ºC too warm (see here for more details about the problem).

3) Figure 1 shows all of the above. We plot the monthly global temperature anomalies for the lower atmosphere as developed by UAH (top), RSS (middle) and the difference between the two (UAH minus RSS, bottom). Note how in the differences chart (Figure 1, bottom) there is a long-term downward trend (indicating that the RSS data is warming faster than the UAH data) but also a sharp uptick at the end, indicating the current problem in the UAH data.

Figure 1. Global temperature anomalies from the lower troposphere. January 1979 through December 2007. (top) Data compiled at the University of Alabama, (middle) data compiled by Remote Sensing Systems, (bottom) difference between the two datasets (UAH minus RSS). (data sources: UAH ; RSS )

4) The temperature anomalies in the RSS dataset for the two most recent months of November 2007 and December 2007 are below the long term (1979-1998) mean. This is the first time that this has occurred since January 2000. It’ll be interesting to see how temperatures evolve over the next couple of months. Will the global temperature anomaly stay close to, or even below the average for a while, effectively resetting global temperature back to the levels characteristic of the mid-1990s—and wiping out a good 3 or 4 tenths of a degrees C of warming? Or is this venture to cooler than normal conditions just a short-lived climate fluctuation? If the pattern of temperatures during the past three years is any indication (Figure 2) then perhaps the past two months are not an unusual fluctuation, but part of a longer period cooling trend. The next several months should shed more light on this question.

Figure 2. Global temperature anomalies from the RSS lower tropopsheric temperature dataset, January 2005 through December 2007. The zero line represents the 1979-1998 average (source: RSS).

5) The time has come to revisit a wager we proposed in these pages some 10 years ago. Back in 1998, a large El Niño warmed up the world and pushed global temperatures to unprecedented high levels. Realizing the unusualness of the situation, we proposed this friendly bet (see ‘Finally!’ at the bottom of this page):

“That 1998’s extremely warm temperatures were largely confined to one calendar year makes the annual record high temperature 1998 has established quite a difficult one to break.

If we were of a betting sort (and there are some nasty rumors going around that we are), we would be willing to wager that the 10-year period beginning in January 1998 and extending through December 2007 will show a statistically significant downward trend in the monthly satellite record of global temperatures.

Surely such a wager should sound interesting to those who think the planetary temperature will increase several tenths of a degree during that period.

No reasonable offers refused…”

We only received one offer, and it was unreasonable in that it was proffered more than half-way through the verification period—by then the odds were stacked fairly high against us (see here for the profferer’s reaction).

Now that the period of the bet has expired, we can see how we did. And, it turns out, that we would have lost, but not really by that much. Using the RSS data (recall that the UAH data currently is a bit messed up in recent years), the overall trend from January 1998 through December 2007 was indeed negative 0.06ºC/decade, but the decline was not statistically significant. But, we weren’t so far off. If the RSS global temperature anomaly remains at or near its current level for only the next 3-4 months, the temperature decline since 1998 will become statistically significant. Close, but no cigar!

In summary, interesting things seem to be ongoing aloft, so stay tuned over the next several months to see how they transpire.


Randall, R. M., and B. M. Herman, 2007. Using Limited Time Period Trends as a Means to Determine Attribution of Discrepancies in Microwave Sounding Unit Derived Tropospheric Temperature Time Series, Journal of Geophysical Research, doi:10.1029/2007JD008864, in press.

January 9th Update: Dr. John Christy, the main overseer of the UAH dataset wrote to us to tell us that we had misinterpreted the size of the recent warm bias in the UAH lower tropopsheric MSU temperatures. In fact, in an entry dated January 3, 2008 in the UAH above-mentioned readme file , Dr. Christy explains that, as it turned out, there never was much of a warm bias in the UAH data from recent months, and that upon further investigation, they found that they were mistaken in their estimates of its magnitude made in their December 19, 2007 entry in the same readme file. Further, Dr. Christy goes on to explain that it is his belief that the large difference between the UAH and the RSS temperatures over the course of the past several months (see the bottom panel of our Figure 1 above) rests with some undiagnosed problem (resulting in a spurious cooling) with the RSS data over the tropics. Dr. Christy, thus stands by the UAH data, in its current form (as depicted in the top panel of our Figure 1), as the record that is most representative of conditions in the earth’s lower troposphere.

What this means for us, is that we were not nearly as close to winning our bet as we had initially imagined. The trend in the UAH derived temperatures of the earth’s lower atmosphere for the most recent 10-year period (January 1998 though December 2007) is a positive 0.04ºC/decade (although it is not statistically significant). However, just like the RSS data, the UAH temperatures do show a decline over the past three years which may indicate that temperatures will continue their course towards the long-term average for the near future. But just as easily, they may reverse course and begin a warming trek once again.

As we suspected, things continue to be interesting, and the next several months will probably produce more noteworthy goings-on. So, be sure to check back often to see what’s new!

No Comments

No comments yet.

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.

Powered by WordPress