Review of Chapter 3
Chapter 3 asks the following: what do observations indicate about the changes of temperature in the atmosphere and at the surface since the advent of measuring temperatures vertically? This chapter describes observed temperature trends for three classes of data sets: surface observations (mainly in situ but including some satellite data), radiosonde measurements, and microwave sounding unit (MSU) observations for various levels in the atmosphere. Linear trends were computed for two periods: 1958-2004 for the surface and radiosonde data and 1979-2002 for all three data sets. The analysis includes three compilations of surface data, two compilations of radiosonde data, and three analyses of MSU data. A fourth data set, reanalyses products, are discussed but downplayed.
In general, the discussion is comprehensive and in most cases reflects an accurate interpretation of the results. The discussion of the three global surface temperature trend analyses is readable and yet discusses some of the important details. The general conclusion of a consistent warming signal at the surface seems well justified. Consistency between the two analyses of radiosonde data, for 1958-2004, is encouraging. The question to be addressed by the chapter is answered. However, the chapter needs to clarify a few points.
1. A major issue is the drop in temperature associated with the introduction of the Vaisala sonde. It is stated that this affects the stratosphere, but it is unclear how deeply this systematic bias might extend into the troposphere. This is an important research problem that should be addressed.
2. Mentioning the similarity of the basic data in the surface dataset while highlighting many of the potential problems (almost all of which have been adequately handled) sows doubts in the minds of readers. This gets picked up and emphasized in Chapter 6. The report should provide more explanation of how various problems in the data have been addressed and how this leads to some level of confidence in examining trends. For example, it can be easily shown by sub-sampling the surface data that the resulting hemispheric and global trends from the sub-samples would be almost exactly
the same. References should be made to the frozen grid analyses work done in the Climate Research Unit (CRU) in the mid-1980s to the mid-1990s.
3. In general, the chapter would benefit from a more careful dissection of the global mean and a recognition that radiosondes are not near global. It does not address global mapping or the need for evaluations at co-located sites of sondes (see Hurrell et al., 2000; Agudelo and Curry, 2004; and Free and Seidel, 2005). There are no latitude-time series presented. This chapter does identify differences over high latitude land as being the main reason for surface being larger than troposphere trends in the extratropics (page 30), but this is not carried forward to the Executive Summary or Chapter 5. Weakening or removal of inversions over cold land or ice is a very good reason why the surface should warm more and a good example of why the global mean should be dissected.
4. The chapter also places too much emphasis on linear trends. Only linear trends as a function of latitude are presented, however this presentation can hide many things. The claimed agreement between radiosondes is not shown except for the linear trends (e.g., Table 3.6.1) (see Free and Seidel, 2005). There is nothing on root mean square differences, which are very revealing (Hurrell et al., 2000), or on monthly differences (smoothing the time series can be misleading).
5. In a number of places, the assertion is made that the troposphere has warmed more than the surface. However, the differences in trends are often quite small, particularly for the 1958-2004 period. It is not clear that these differences are statistically significant. Although statistical significance is assessed for the trends themselves, no analysis of the significance of trend differences is presented. When comparing trend differences between two estimates of the same quantity (e.g., tropospheric temperatures from radiosondes and satellites), it is more appropriate to examine the trend of the difference time series, rather than trends for each time series individually (because the data contain similar overall variability). This is an omission that should be corrected, and the text should reflect the results of such an analysis. In particular, any statement about differences in warming should be weakened considerably if the differences are not statistically significant.
6. The trends calculated from reanalyses are downplayed because the input data sets are not homogenized. Although there is potential independent value of reanalysis products, it is not clear what trends from a reanalysis model mean in the context of temporally varying inputs. Therefore, the committee agrees with the decision of the authors to downplay this source of information.
7. The issue of regional land-use and land-cover changes is brought up in a number of places, but the implications are not clearly addressed. For example, in lines 94-96 it is suggested that regional land-use change must be considered in the development of land-based data sets. However, if regional changes are large enough to have a measurable influence on global temperature, then these changes will be sampled and detected by the existing land-based networks. As such, why is this an issue when analyzing the differences among the data sets? There is an issue related to land-use and land-cover changes that could be addressed here or in other chapters. In the modeling discussions in Chapters 1, 5, and 6, land-use and land-cover is considered to be a forcing (with uncertain magnitude in the past) that is incorporated in some models and not in others. The committee believes this is correct and that land-use and land-cover should be considered as a forcing. Any land-use and land-cover effects in observational datasets should
therefore be left in and not commented upon as a problem in Chapter 6. In other words, Chapter 6 cannot have it both ways the data are affected by land-use and land-cover change, so they are somehow wrong, yet this forcing is omitted from many models.
8. The Fu et al. results have the potential to be centrally important to the issue of tropospheric temperature trends and should be discussed more thoroughly in lines 863-868. Attempts to separate tropospheric and stratospheric contributions to trends are reasonable. They should not be rejected with the value statement that they are “controversial”. The only published criticism of the Fu et al. approach is by Tett and Thorne (2004), with other criticisms in the grey literature. The Fu et al. method has since been followed up by several studies which show that it is robust, including further research by Fu and colleagues and Gillett et al. (2004). The potential clarification that the Fu et al. method can contribute to the central issues is very significant.
9. The difference between the Remote Sensing Systems (RSS) and University of Alabama, Huntsville (UAH) trends is left as an open issue, with no relative value given. It is important to resolve this discrepancy, if possible. The trend difference in the mid-troposphere is the same size as the signal: zero for UAH and +0.1 K/decade for RSS. If no distinction can be made, then no conclusion can be drawn. Statements in lines 355-359 and elsewhere about discrepancies between RSS and UAH as being mostly due to the NOAA 9 satellite are misleading as can be seen by looking solely at the post 1987 period. In fact, examining differences between the two datasets, which are not shown in the report, reveals major issues remaining on adjustments for other satellites and diurnal cycle issues (especially as a function of latitude and in the tropics).
10. In lines 791-816, if the tropical tropospheric temperature profile behaves as a moist adiabat, which to an approximation it does, then the lapse rate is expected to decrease as temperature increases (i.e., as the surface warms, the troposphere is expected to warm more). This is the “global change theory” the authors refer to in Section 6.2.1. Therefore, it is no surprise that when the surface warms due to ENSO, the troposphere warms relative to the surface (line 798), or that when the atmosphere warmed in 1976-1977, the lapse rate dropped (line 802). These results are currently presented with no link to physical theory. The authors say that “the variation in tropical lapse rate can be characterized as highly complex, with rapid swings over a few years, superimposed on persistent periods of a decade or more”, but our guess is that much of this variation can be explained by changes in the mean temperature. Further, the authors say that the enhanced warming of the troposphere associated with surface warming gives “enhanced static stability” (lines 799 and 803). A reference should be provided for this statement. It should be noted that the troposphere did warm relative to the surface in the tropics during the 1997-98 El Niño event, which is a large signal. Also, the report should reference a study by Gettelman et al. (2002) on changes in stability. This study highlights the observed increases in Convective Available Potential Energy (CAPE) that are not replicated by models (which remove all CAPE), and so it is also relevant to Chapter 5 of the report.
1. The numerical system for numbering the figures is overly complicated and inconsistent. It would be simpler to number the figures 3.1, 3.2, 3.3 etc., rather than 2.4, 3.3, 4.4, 6.2, 6.2.2, 6.2.3, 7.1. In all the figures, the notations used to label the curves in the diagrams are different from the descriptions in the captions. For example, Figure 2.4 has the labels N, G, and U, and these are not defined in the caption. The same is true in different ways for 3.3, 4.4, etc. Also, without a very good color print, the different colored lines can be difficult to distinguish.
2. In lines 53-55, comparing results from more than one dataset also provides a better idea of the uncertainties or at least the range of results.
3. In lines 86-88, the statement that homogenization procedures are “quite successful” at addressing these issues should be more nuanced. While we are in agreement with the statement with regard to biases introduced by changes in time of observation, we are less confident that other issues (e.g., exposure changes) can so readily be addressed because there is often a lack of metadata.
4. In lines 107-111, the benefits of sea surface temperature (SST) over night marine air temperature (NMAT) are discussed without saying anything about what the relationship between SST and NMAT is likely to be (e.g., is SST a good proxy for NMAT?).
5. There should be a reference in line 163 to Jones et al. (1997, 2001). These papers give details of the procedure for allowing for changing numbers of observations through time.
6. This text in lines 180-183 is a bit wordy and does not follow on well from the previous sentence. The paper by Vose et al. (2005) should show that the differing techniques with the same data produce almost the same results.
7. In line 205, the text “since neither choice is optimal” suggests that there is a single optimal approach. This should be rephrased to “since each approach has advantages and disadvantages.”
8. In lines 229-231, the Radiosonde Atmospheric Temperature Product for Assessing Climate (RATPAC) data set incorporates different homogeneity adjustments before and after 1997. Has anyone evaluated the extent to which this might introduce an inhomogeneity into this data set?
9. Lines 294-296 state, “There is some ambiguity about whether the temperatures return to their earlier values or whether they experience step-like falls”. Surely this is just a matter of how best to describe the curves. A more important question is whether the observations agree with particular models (global circulation models or theoretical models). Has anyone suggested a plausible mechanism that would give a step-like cooling after a volcano (e.g., Douglass and Knox, 2005)?
10. In lines 297-299, is the interannual variability really mainly due to the Quasi-Biennial Oscillation (QBO)? If so, a reference should be provided.
11. The text in lines 299-301 makes it sound like the stratospheric cooling trend has been completely explained as a combination of the responses to stratospheric ozone depletion and cooling due to carbon dioxide. This is in disagreement with the Executive Summary, which indicates that the cooling cannot be fully explained by these forcings.
12. In lines 301-305, there are various descriptions of the curve including “the aforementioned step-like drops represent a viable alternative to a linear decrease”. What do the authors mean by “a viable alternative”? Presumably, they do not mean one based on a physically-plausible mechanism. Again, this seems to just be a discussion of how best to describe the curve, whereas the real issue is whether the observations agree with theoretical predictions.
13. In lines 320-323, the change to the Vaisala radiosonde in certain tropical areas is given as a possible reason for the differences in the two radiosonde data sets. What analysis has been done to suggest this possibility? Or is this statement made simply because the timing is coincidental?
14. The nomenclature of TMid-Trop-R and TMid-Trop-A are introduced without definition in line 358. At least a reference to Chapter 2, Figure 2.2 and related discussion should be included for those who may start reading here. Are these just the Microwave Sounding Unit (MSU) channels and their radiosonde integral equivalents, or something else? Also, the nomenclature in the figures and captions is inconsistent and not sufficiently defined.
15. TTrop-UW-A is introduced without definition in line 396.
16. In line 447, the National Centers for Environmental Prediction (NCEP)/National Center for Atmospheric Research (NCAR) reanalyses go back to 1948. It is probably best to ignore the period between 1948-57 as this study only goes back to 1958.
17. A reference to Simmons et al. (2004) might be needed in line 475, or a reference back to Chapter 2.
18. The Pielke and Chase (2004) reference in line 488 is missing from the reference list.
19. It is not completely clear what is meant in lines 502-505. Presumably this relates to the abrupt change in the late 1970s.
20. What do the authors mean by “it has been shown that such constructs are plausible” in lines 505-508? What criteria are used to judge their plausibility? Presumably it is just how well they fit the data. In this case, you could make a perfect fit to the data by regressing it on itself. Again, the real issue is whether the observations fit with theoretical predictions.
21. It is unclear what is meant in lines 515-518.
22. In lines 521-527, The reanalysis models tend to agree better with the climate model predictions than do the raw observations. Is this an alternate explanation of the differences between the reanalyses and the raw observations (i.e., if the reanalysis model has similar physics to the climate models, then its troposphere will warm more than its surface)?
23. It is not entirely clear what is meant in lines 542-544. Do the authors mean something like “trends in land air temperature in coastal regions are generally consistent with trends in SST over neighboring ocean areas”?
24. In lines 548-552, the authors do not mention the most obvious explanation for enhanced warming over land, namely the smaller effective heat capacity over land than ocean. Enhanced warming over land is seen in every climate change simulation and does not relate primarily to the phase of ENSO, though this could be a contributor. Better justification should be provided for a link between warmer temperature over Siberia and ENSO. Siberia encompasses a large area, so be more specific and provide a reference.
25. In lines 561-563, SSTs and NMAT have different trends for short periods owing to ENSO and changes in surface fluxes, as shown in other works.
26. In lines 568-570, these differences might be related to an increase in mean ship height above the sea surface.
27. While the explanation in lines 589-591 sounds plausible, has it ever actually been shown? Are the free tropospheric temperatures more highly correlated with maximum surface temperatures than with minimum temperatures? If there is not a published reference on this, then should be removed.
28. In lines 600-601, it is not clear whether this relative change in trend in the troposphere and surface is statistically significant in the recent era. Visually, it does not seem that impressive or obvious.
29. The comparison the authors make in lines 627-629 is equivalent to assuming that the Maryland stratospheric trend is the same as that in the other two datasets (since the Fu et al. approach is just to fit to a regression model).
30. In lines 655-656, “TLow-Strat-A” and “TLow-Strat-B” need definitions.
31. In lines 656-657, why is the cooling at the South Pole not more dramatic, especially given known problems over sea ice (Swanson, 2003) and the high ice sheet of Antarctica that greatly impacts channel 2? In fact, it looks like the cooling is larger in the northern hemisphere midlatitudes.
32. Replace “Soviet stations” in line 672 with “stations located in Russia and other countries of the former Soviet Union”.
33. The word “granularity” should be replaced in line 693.
34. In line 696, replace “noisy patterns that result” with “noise that results”.
35. The figure labeling (a, b, c and d) in line 704 is incorrect.
36. No mention is made of the Antarctic in line 713.
37. In lines 722-724, the sharp contrasts only seem to be around the western coasts of the Americas.
38. The unit for a lapse rate trend looks wrong in line 823. Surely it should be K km−1 decade−1 or something with the same dimensions.
39. Are there missing crosses for the surface in Figure 6.2b, or do they all overlap?