Relatively up-to-date radiosonde data is available from the Hadley Center, tropical (20N-20S) is here. Ratpac and Angell are not up to date. The tropical troposphere has been a source of disputes recently, but I haven’t seen any discussion of up-to-date radiosonde data, [Note: Luboš has a current discussion on radiosondes.)
You will recall the diagram illustrating a hot spot around 200 hPA in the tropical troposphere. Here’s a diagram from realclimate which is the first figure in their post entitled “Tropical Troposphere Trends”.
Here’s a simple plot of tropical 200 hPa radiosonde data to March 2008 from Hadley Center.
Only one month in the entire history of the radiosonde record since its commencement in January 1958 had 200 hPa and 150 hPa anomalies below -1.2 deg C. It was March 2008. The trend since January 1979, the start of satellite records, is -0.025 deg C/decade. Yeah, I know that it’s just one month, but it’s still a “record”. It would be interesting to calculate the odds of a negative record on the hypothesis of (say) a positive trend of 0.1 deg C/decade. (Note that these data sets are highly autocorrelated and that ARMA(1,1) coefficients are both significant with an AR1 coefficient of over 0.9 in an ARMA(!,1) model – something that reduces “significance” of any trend quite noticeably).
Note: As observed below, the GISS graphic shows the effect of doubled CO2, while the increase in CO2 levels since 1960 to date has been 20%, and about 15% since the start of satellite records in 1979. On the basis of a logarithmic impact, the first 20% increase accounts for about 26% of the impact; and the 15% increase since 1979 about 20%. So it should be noticeable in either data set. In other posts on radiosonde data, I’ve observed that there are many issues with inhomogeneity in radiosonde data.
More: CCSP 1-1 and HadAT Radiosonde Data
The U.S. Climate Change Science Assessment Report 1-1 , a report to which Douglass et al 2007 were in part responding, contained graphics illustrating both HadAT radiosonde data trends from 1979-99 and GISS projections over the same period, so it’s interesting to compare their results to the updated information here.
First here is a graphic from the CCSP report, which, inter alia, shows their calculations of HadAT radiosone trends for 1979-99, followed by my calculation of the same trends for 1979-March 2008. It’s interesting that the HadAT pattern hasn’t really changed that much even with the incorporation of 10 more years.
Top – Original caption: Figure 5.1: Vertical profiles of global-mean atmospheric temperature change over 1979 to 1999. Surface temperature changes are also shown. Results are from two different radiosonde data sets (HadAT2 and RATPAC; see Chapter 3) and from single forcing and combined forcing experiments performed with the Parallel Climate Model (PCM; Washington et al., 2000). PCM results for each forcing experiment are averages over four different realizations of that experiment. All trends were calculated with monthly mean anomaly data. Bottom — HadAt trends.
The CCSP report stated:
The pattern of temperature change estimated from HadAT2 radiosonde data is broadly similar, although the transition height between stratospheric cooling and tropospheric warming is noticeably lower than in the model simulations (Figure 5.7E). Another noticeable difference is that the HadAT2 data show a relative lack of warming in the tropical troposphere,66 where all four models simulate maximum warming. This particular aspect of the observed temperature-change pattern is very sensitive to data adjustments (Sherwood et al., 2005; Randel and Wu, 2006).
Below are their illustrations of GISS model projections 1979-99 compared to HadAt actuals. I’m not in a position to comment authoritatively on these graphics, but here are a couple of points that I find interesting. As I understand it, the top of the tropical tropopause is ∼18.7 km (70 mb). The GISS model shows warming right up to ~ 16 km (100 mb), both in the doubled and 20th century graphics, with cooling in the “blue” color code up to 25 km (25 mb). HadAt radiosonde data shows warming up to only about 12 km (250 mb), “blue” cooling from 12 km (250 mb) to 16 km (100 mb) and “purple” cooling from about 16 km (100 mb) to 25 km (25 mb) and higher.
The actual locus where additional CO2 has an immediate impact is at altitude, as more CO2 causes radiation to space to occur at a higher and colder altitude according to the Houghton heuristic cartoon. Getting the sign wrong in the 12-16 km is definitely a bit inconvenient and is not mere nit-picking and you can see why they are looking so hard at the observations to see if there’s some important inhomogeneity.
Karl, T. R., Susan J. Hassol, Christopher D. Miller, and Willieam L. Murray. 2006. Temperature Trends in the Lower Atmosphere: Steps for Understanding and Reconciling Differences. Synthesis and Assessment Product. Climate Change Science Program and the Subcommittee on Global Change Research. http://www.climatescience.gov/Library/sap/sap1-1/finalreport/sap1-1-final-all.pdf.