BLAME! the NoctambulantJoycean: Myth: Evidence Supports Curry's Claims Regarding Satellite-based Analyses and the Hot Spot

The outline for this post is as follows:

The Myth and Its Flaw
Context and Analysis (divided into 3 sections)
Posts Providing Further Information and Analysis
References

This is the "main version" version of this post, which means that this post lacks most of my references and citations. If you would like a more comprehensive version with all the references and citations, then please go to the "+References" version of this post.

References are cited as follows: "[#]", with "#" corresponding to the reference number given in the References section at the end of this post.

1. The Myth and Its Flaw

Climate models predict that in moist tropical areas, a region of the lower atmosphere known as the troposphere will warm more than Earth's surface. This region of greater warming is known as the "hot spot". Judith Curry, a critic of mainstream climate science, casts doubt on the hot spot's existence. She also casts doubt on a recent satellite-based analysis of tropospheric warming. Curry's claims on these topics constitute the myth this blogpost rebuts.

The myth's flaw: Curry's own sources undermine her claims, as do other published analyses of satellite and weather balloon data. This data supports the hot spot's existence and supports the satellite-based analysis Curry attempts to undermine. Curry obscures these points by citing unreliable sources and by not transparently explaining published research, consistent with her track-record when it comes to discussing climate science on non-peer-reviewed Internet forums. This allows Curry to distort the science surrounding carbon-dioxide-induced anthropogenic (man-made) global warming (AGW).

(A list of some of Curry's other distortions can be found in section 2 of "Myth: Judith Curry Fully and Accurately Represents Scientific Research")

2. Context and Analysis

Section 2.1: Curry's Sources Debunk Her on Position on the Hot Spot

Earth's atmosphere contains multiple layers. The layer closest to the Earth's surface air is known as the troposphere. Climate models predict that tropospheric warming in the tropics should increase with increasing height. See "Myth: The Tropospheric Hot Spot is a Fingerprint of CO2-induced Warming" for more on the mechanism behind this increased warming; however, knowing this mechanism is not required for understanding the points made in this blogpost.

The aforementioned tropical warming amplification is called the tropical tropospheric hot spot by many critics of mainstream climate science. As the critic Roy Spencer puts it:

"One of the most vivid predictions of global warming theory is a “hotspot” in the tropical upper troposphere, where increased tropical convection responding to warming sea surface temperatures (SSTs) is supposed to cause enhanced warming in the upper troposphere [1]."

Judith Curry, another critic of mainstream science on AGW, spreads misinformation regarding the hot spot. For example, she insinuates that the hot spot is a fingerprint of warming caused by greenhouse gases such as carbon dioxide (CO2). Yet solar-induced warming would also cause hot spot. In fact, any large surface warming in the tropics (especially warming of the tropical oceans) would cause a hot spot, as I explain in "Myth: The Tropospheric Hot Spot is a Fingerprint of CO2-induced Warming". Thus the hot spot is not a fingerprint of CO2-induced warming, contrary to what Curry suggests.

In addition to falsely portraying the hot spot as a fingerprint of CO2-induced warming, Curry also casts doubt on the hot spot's existence. Ironically, Curry's own source rebuts her on this point. To see why, first note that Curry lauds a re-analysis known as ERA-I, a.k.a. the European Centre for Medium-Range Weather Forecasts Interim re-analysis. Re-analyses such as ERA-I incorporate data from diverse sources, including satellites and weather balloons (radiosondes). ERA-I and other of re-analyses show the hot spot, with greater warming in the tropical upper troposphere (at an atmospheric pressure level of around 300hPa) than near Earth's surface; I discuss this further in "Myth: The Tropspheric Hot Spot does not Exist".

Figure 1 depicts the hot spot in ERA-I, in comparison to model-based predictions and projections of this hot spot:

Figure 1: ERA-I and model-based warming/cooling trends versus height, represented as the trend at a particular pressure level divided by the near-surface temperature trend at that latitude. The results at a particular latitude represent the average of the temperature response in both hemispheres at that latitudinal distance from the equator. The model-based results depict an average response from an ensemble of models included in the Coupled Model Intercomparison Project 5 (CMIP5). Results displayed are for 1979 - 2005 for (a) ERA-I, (b) the model-based temperature response to multiple factors including aerosols and increased greenhouse gases such as CO2, and (c) the model-based temperature response to increased greenhouse gases. Also displayed are (d) the model-based temperature response from 2006 - 2100 under Representative Concentration Pathway 8.5 (RCP8.5) [2, figure 1], in which humans continue to release large amounts of CO2. The tropics lie between latitudes 0° and 30°. Pressure decreases from the Earth's surface (near the bottom of the y-axis) to the troposphere to the stratosphere (near the top of the y-axis). The tropical troposphere lies below 150hPa.

In comparison to the model-based projections, ERA-I shows less warming amplification in the middle and lower troposphere. This is likely because ERA-I under-estimates middle and lower tropospheric warming, as admitted by the ERA-I team and other researchers. ERA-I also clearly shows a hot spot in figure 1, with upper tropospheric warming being greater than near-surface warming in the tropics. ERA-I therefore provides one line of evidence for the hot spot, as shown in a number of studies.

Scientists therefore made progress on the hot spot by incorporating diverse data sources into multiple analyses such as ERA-I, in accordance with a recommendation given by the United States Climate Change Science Program (CCSP); I discuss this further in "Myth: The CCSP Presented Evidence Against the Hot Spot's Existence". Curry should agree with this recommendation, in principle. After all, in reviewing the CCSP report Curry recommended that scientists sort out discrepancies between analyses, including by identifying errors in analyses. Though discrepancies remain between re-analyses (as I discuss in "Myth: The Tropspheric Hot Spot does not Exist"), all of the re-analyses, or all but one, show the hot spot.

So ERA-I and other re-analyses undermine Curry's attempts to cast doubt on the hot spot's existence, despite the fact that Curry cites ERA-I favorably. Thus Curry's own source rebuts here position. Consistent with this ERA-I result, radiosonde-based studies and satellite-based studies also confirm the hot spot's existence, with greater warming in the tropical upper troposphere than near Earth's surface, as I discuss in "Myth: The Tropspheric Hot Spot does not Exist".

To elaborate on the satellite-based results further, the following six groups generate satellite-based tropospheric temperature records:

a group at the University of Washington (UW)
a group at the National Oceanic and Atmospheric Administration Center for Satellite Applications and Research (NOAA/STAR or NOAA)
a group at Remote Sensing Systems (RSS)
Vinnikov et al. at the University of Maryland (UMD)
a group at the University of Alabama in Huntsville (UAH)
Weng and Zou at the University of Maryland

All of the analyses show the hot spot, with the exception of UAH. Figure 2 illustrates this point for four of these analyses:

Figure 2: HadCRUT4 tropical surface warming trends and tropical mid-to-upper tropospheric warming trends (in K per decade) from 20°N to 20°S above the land, oceans, and both land and oceans from 1979 - 2012. Tropospheric warming trends are from UW, NOAA, RSS, and UAH satellite data analyses. UW(GCM) and UW use different methods for processing the satellite data. The value in parentheses is the ratio of the tropospheric warming to the surface warming for a given tropospheric temperature trend [3, table 4]. The RSS tropospheric warming trend is spuriously low due to an error in homogenization. The RSS team later corrected this error. This resulted in a RSS tropical mid-to-upper tropospheric warming trend that is between the NOAA trend and the UW trend.

UAH's results likely differ from those of the other 5 research groups, because different groups use different data analysis methods and corrections for known artifacts/errors in the data. These corrections are known as homogenization and the artifacts are known as heterogeneities (I discuss homogenization in more detail in section 3.1 of "John Christy, Climate Models, and Long-term Tropospheric Warming", with examples of scientists validating homogenization techniques).

UAH has a long history of under-estimating tropospheric warming due to UAH's faulty homogenization. And UAH's analyses often diverged from analyses made by other research groups, in the mid-to-upper troposphere and other atmospheric layers. In light of this, other scientists critiqued UAH's homogenization methods. A startling example of this occurred with UAH's correction for diurnal drift.

Scientists need to apply a diurnal drift correction to account for the fact that satellite measurements occur at different times of day. Since temperature at noon will likely be warmer than temperature at midnight, correcting for this time-of-day effects remains crucial for discovering any underlying tropospheric warming trends. The RSS team revealed that UAH bungled the diurnal drift homogenization in a way that spuriously reduced UAH's tropospheric warming trend. The error occurred because Spencer's UAH team falsely assumed that the lower troposphere warmed at midnight and cooled at mid-day.

According to Spencer and his UAH colleague John Christy, correcting the UAH team's error increased their warming trend by ~40%. RSS' own warming trend at that time was even larger than this. When Spencer and Christy admitted this error, RSS members Carl Mears and Frank Wentz offered the following priceless reply [4; 15]:

Or as reportedly noted by Kevin Trenberth, one of Christy's supervisors in graduate school:

"[Trenberth] said he distanced himself from Christy around 2001, worried that every time a decision was called for in processing data, Christy was choosing values that gave little or no trend [16]."

(This quote is consistent with Trenberth's decades-long history of documenting Christy's distortions and correcting those who abused Christy's distortions in order to misleadingly minimize global warming)

Yes, one wonders why the UAH team adopted an obviously wrong adjustment that conveniently reduced their stated amount of lower tropospheric warming...

(In "John Christy and Atmospheric Temperature Trends", I summarize other examples of Christy distorting climate science in politically expedient ways)

Thus RSS identified a mistake in the UAH team's homogenization, as RSS has been doing for almost two decades. Conversely, UAH scientists pointed how an older RSS analysis under-estimated lower tropospheric warming, as I discuss later in this blogpost. UAH scientist Roy Spencer further suggested that one should opt for the flawed RSS analysis over Spencer's UAH analysis, if one was committed to showing as little global warming as possible. Despite Spencer suggesting that people engage in biased selection of data sources, this situation still illustrates that scientists acted in accordance with Curry's recommendation: scientists sorted out discrepancies between satellite-based analyses and identified errors in the UAH analyses. This process led to multiple analyses showing that hot spot, and cogent rebuttals to the lone satellite-based analysis (UAH) that did not show the hot spot.

Even some of these rebuttals were made and before RSS published corrections to their under-estimated trend, RSS team member Carl Mears accepted that satellite-based tropospheric records were more uncertain than surface-based records (based on his published uncertainty estimates and conference abstract). The U.S. Global Change Research Program makes much the same point, as does the climate scientist Andrew Dessler. But Curry contradicts these evidence-based points by claiming that satellite data is the best data we have, even as she goes on to dispute the corrections used for the satellite data, as I discuss in section 2.2.

So none of the aforementioned error-correction and evidence convinced Curry; she continues to cast doubt on the hot spot's existence, despite multiple satellite-based analyses, radiosonde-based analyses, and re-analyses showing the hot spot. She continues to cast doubt, despite the fact that she discusses results from RSS and ERA-I, two analyses that show the hot spot. Thus both RSS and ERA-I represent two cases in which Curry's own sources rebut her position on the hot spot.

Section 2.2: Curry's Sources Debunk Her on Position on RSS' Homogenization

Judith Curry's position became even more implausible when she shifts from discussing the hot spot, to critiquing RSS' recently improved homogenization methods. To understand the relevance of these improvements, we should first go into more detail about satellite-based tropospheric analyses. RSS generates at least three tropospheric analyses:

TLT : TLT represents trends in the lower troposphere, with some contribution from the mid-troposphere.

TMT : TMT focuses largely on trends in the mid-to-upper troposphere. Importantly, however, TMT also includes a significant contribution from the lower stratosphere, a layer of the atmosphere higher than the troposphere. TMT therefore contains some stratospheric cooling, as I discuss in section 2.3 of "Myth: Santer et al. Show that Climate Models are Very Flawed". Scientists have known about this stratospheric contamination of TMT since at least 1997; even UAH team member Roy Spencer is aware of it. This stratospheric cooling causes TMT to under-estimate mid-to-upper tropospheric warming.

TTT : TTT better represents trends in the mid-to-upper troposphere, by removing the stratospheric contribution to TMT. Figure 2 above presents TTT data from four satellite-based analyses, as does figure 5 below. Different research groups account for the TMT stratospheric contribution in different ways, with the exception of UAH and UAH team member John Christy. Christy and the UAH team fail to adequately correct for this cooling, since they object to the validated correction method used by RSS, UW, and NOAA/STAR.

In contrast to UAH and RSS, NOAA/STAR and UW do not produce a TLT analysis. Furthermore, UW generates a TMT and TTT analysis only for the tropics, while UAH, RSS, and NOAA/STAR produce TMT analyses for not only the tropics, but also for most of the globe.

UAH team member Roy Spencer identified errors in version 3 (v3) of RSS' TLT analysis; these errors caused RSSv3 to under-estimate lower tropospheric warming, especially after 1998. In 2017, the RSS team wrote a paper correcting their homogenization for their TLT analysis, after writing a 2016 paper correcting their homogenization for their TMT analysis. RSS incorporated these corrections into version 4 (v4) of their TLT, TMT, and TTT analyses, thereby increasing all three warming trends. This resulted in a RSSv4 tropical TTT trend between the tropical TTT trends for NOAA/STAR and UW, instead of the lower RSSv3.3 TTT trend shown in figure 2.

The improved homogenization for RSSv4 TMT disappointed Curry, even though these improvements accorded with her recommendation that scientists investigate discrepancies between satellite-based analyses and, if possible, identify errors in the analyses. She expressed concern that RSS' corrections could do away with a (supposed) post-1998 "pause [5]" in global warming (I discuss the "pause" more extensively in "Myth: No Global Warming for Two Decades, from 1997/1998 to 2016/2017", section 2.1 of "Myth: The IPCC's 2007 ~0.2°C/decade Model-based Projection Failed and Judith Curry's Forecast was More Reliable", and in section 3.4 of "John Christy, Climate Models, and Long-term Tropospheric Warming"). But Curry's concern lacks merit and shows that she did not pay sufficient attention to the ERA-I re-analysis she cited.

In the same 2016 blogpost in which Curry lauds ERA-I, Curry also discusses a 2016 paper that compares ERA-I's lower tropospheric temperature analysis to that of RSSv3 and UAHv6. That paper showed that the ERA-I, UAHv6, and RSSv3 lower tropospheric warming trends were very similar, as shown below:

Figure 3: Global, relative tropospheric temperature trends from 1979 - 2014 for ERA-I (ECMWF), UAH version 6.0, and RSS version 3.2. UAH, RSS, and ERA-I "TLT" trends represent the lower troposphere, while "ECMWF mid trop" represents the mid-troposphere [6, figure 3].

But as I previously discussed, in 2014 and again in 2016, the ERA-I team admitted that ERA-I under-estimates lower and middle tropospheric warming. Other scientists acknowledged this point as well. In contrast, ERA5, the update to ERA-I, yields a larger tropospheric warming trend (see figure 5 below). So if ERA-I under-estimates lower tropospheric warming and RSSv3 has lower tropospheric warming on par with ERA-I, as per figure 3, then RSSv3 likely under-estimates lower tropospheric warming as well. Thus one should expect RSS' corrections to increase lower tropospheric warming from RSSv3 to RSSv4, which is just what happened.

ERA-I therefore indirectly supports the RSS' correction to RSSv3. Curry should have been aware of this, since scientists pointed out ERA-I's under-estimation of tropospheric warming before Curry cited ERA-I and before she critiqued the corrections to RSSv3. Yet Curry still complains about RSS' corrections to RSSv3, even though the ERA-I analysis she cites implicitly supports the corrections she complains about.

Curry's defense of the (supposed) post-1998 "pause [5]" extends not only to critiquing RSS' troposphere analysis, but also to complaining about near-surface temperature trends. Curry unjustifiably attacks an NOAA analysis and an analysis from Cowtan+Way that both show substantial post-1998 near-surface warming (I rebut her critique in "Myth: Karl et al. of the NOAA Misleadingly Altered Ocean Temperature Records to Increase Global Warming"). Cowtan+Way and NOAA scientists argue that some temperature analyses suffer from poor global coverage, especially in the Arctic. This poor coverage causes these analyses to under-estimate post-1998 warming. Scientists have known about this issue since at least 2008. So Cowtan+Way, NOAA scientists, and other research groups correct for this under-estimation using various methods, resulting in an increase in the post-1998 warming trend.

Curry objects to the aforementioned corrections. She suggests she may not use Cowtan+Way's corrected analysis, but she proceeds to use it anyway. Moreover, in a blogpost Curry conveniently claims that re-analyses might not be of much utility in assessing Cowtan+Way's correction; this despite the fact that she later tries to compare ERA-I to Cowtan+Ways' correction, lauds the ERA-I re-analysis, and claims that re-analyses should be examined with respect to the corrections. Thus Curry's position appears to be internally inconsistent.

In any event, ERA-I, and another re-analysis known as JRA-55, confirm the results of Cowtan+Way's correction, as shown by the ERA-I team. The ERA-I team also comments on how Curry's discussion of Cowtan+Way's work under-estimates the relevance of re-analyses, and does little to effectively utilize re-analyses. Way also called out Curry on this, while citing other evidence in support of Cowtan+Way's correction method. Apparently Curry didn't bother to closely examine the re-analyses she lauds and claims should be examined. I... what?! So ERA-I supports the near-surface temperature correction Curry complains about, in addition to ERA-I validating the previously discussion results of the RSS correction Curry objected to.

Spencer also cites re-analyses such as ERA-I; he argues that the re-analyses better agree with the UAH analysis than with RSSv4, and that this supports UAH over RSSv4. But as we just saw, this actually argues in favor RSSv4 over UAH, since ERA-I under-estimates tropospheric warming. Spencer's criticisms do not carry as much punch when one remembers that his UAH colleague John Christy states that RSSv4 is a pretty good dataset, and advocates average RSSv4 with UAHv6 for presentations, instead of showing UAHv6 alone. In addition to citing re-analyses, Spencer also claims that radiosonde trends better agree with the UAH analysis than with RSSv4. Yet radiosonde-based estimates show more lower tropospheric warming than UAH over the post-1998 period of the supposed "pause".

And as with the re-analysis trends, radiosonde-based trends argue in favor of RSSv4 over UAH. This is because radiosonde trends under-estimate tropospheric warming during the post-1979 satellite era up to 2012, as I discuss in section 2.2 of "Myth: The CCSP Presented Evidence Against the Hot Spot's Existence". So an accurate satellite-based analysis should show significantly greater warming than the radiosonde-based trends. RSSv4 passes this test, while UAH and RSSv3 do not, as shown in the figure below:

Figure 4: Comparison of relative, lower tropospheric temperature (TLT) trends from 1979 - 2012 for satellite-based analyses and weather-balloon-based analyses, as presented by the RSS team. The figure covers specific regions where valid weather balloon data is available for each weather balloon analysis. The satellite-based analyses are RSS version 3.3, RSS version 4.0, UAH version 5.6, and UAH version 6.0. The weather balloon analyses are Hadley Center Radiosonde Temperature (HadAT), Radiosonde Observation Correction using Reanalysis (RAOBCORE), Radiosonde Innovation Composite Homogenization (RICH), and Iterative Universal Kriging (IUK) [7, figure 12]. RSS did not include Radiosonde Atmospheric Temperature Products for Assessing Climate (RATPAC), since RATPAC lacked the homogenization needed for a valid comparison [7, page 7712].

(At this point, a critic might claim that "it's suspiciously convenient that radiosonde analyses and re-analyses under-estimate lower tropopsheric warming." But this suspicion lacks merit, since climate scientists also admit when observational analyses over-estimate warming. So there is no conspiracy to always say that warming is under-estimated. For a more detailed response to paranoid conspiracy theories about climate scientists and homogenized analyses, see the response to objection 1 in section 3.1 of "John Christy, Climate Models, and Long-term Troposoheric Warming".)

Figure 5 below makes this point more apparent by comparing updated radiosonde-based, satellite-based, and re-analysis-based lower tropospheric warming trends. This figure was co-authored by UAH team member John Christy, RSS team member Carl Mears, and UW team member Stephen Po-Chedley:

Figure 5: Comparison of relative, lower and mid-to-upper tropospheric temperature trends (LTT and TTT, respectively; LTT is equivalent to the TLT discussed elsewhere in this blogpost) from 1958 - 2018 or 1979 - 2018 for weather-balloon-based analyses, satellite-based analyses, and four re-analyses. NASA/MERRA-2 begins in 1980. The data processing for UAH LTT is slightly different from the other analyses, causing its global trend value to be typically cooler by 0.01°C/decade in comparison to the other LTT trends [12, page S18].

A recent satellite-based global positioning system radio occultation (GPS RO; differs from the microwave-emissions-based satellite analyses in the figure above) paper showed that ERA-I, ERA5, and NASA/MERRA-2 under-estimate mid-to-upper tropospheric warming in comparison to 2002 - 2017 GPS RO results.

Despite co-authoring the above figure, less than four months before the figure was published, Christy still falsely claimed to non-experts that the weather-balloon-based analyses supported his UAHv6.0 analysis over RSSv4.0 and UWv1.0. Christy tries to justify this by shifting to his discussion of re-analyses from a March 2018 paper. But as the above figure shows, the re-analyses still undermine Christy's low UAH warming trend, and his shift to re-analyses tacitly retracts his previous claim that weather-balloon-based analyses support his position. I critique Christy's March 2018 paper in a parenthetical comment in section 2.2 below.

UAHv6 uses multiple satellite analyses (the mid-to-upper tropospheric TMT analysis, in combination with two other analyses) to generate its lower tropospheric temperature TLT trend. This represents a shift from how the warmer UAHv5 trend was generated. Yet Christy complains that using multiple analyses increases uncertainty, in contrast to "directly measured [13; 14]" analyses. Compounding Christy's self-admitted problem with UAHv6 TLT, UAHv6 TLT depends on a TMT analysis that under-estimates warming, as the UAH team admitted in a March 2018 paper; this 2018 paper unduly minimizes how much UAHv6 TMT under-estimates warming, for reasons I discuss later in this section. Thus the flawed UAHv6 TMT analysis helps explain why UAHv6 shows the least amount of warming in the tropical mid-to-upper troposphere and in the global lower troposphere, as shown in figure 5.

Moreover, all of the figure 5 analyses show more global lower tropospheric warming than UAHv6, except for the ERA-I analysis which is known to under-estimate lower and middle tropospheric warming. Christy even admits (in a deeply flawed March 2018 paper I discuss later in the section) that UAHv6 shows less mid-to-upper tropical tropospheric warming than does ERA-I. By slashing their lower tropospheric warming trend between UAHv5 and UAHv6, Spencer and Christy caused UAHv6's lower tropospheric warming trend to be lower than that of RSSv3 (see figure 4), despite the fact that Christy and Spencer admitted that RSSv3 contained a cooling bias. And UAHv6 even shows less warming than the HadAT radiosonde-based analysis (see figure 4), even though Christy admits HadAT that under-estimates tropospheric warming and Christy no longer cites HadAT. Moreover, an analysis of radiosonde-based trends in China showed that UAHv6 likely under-estimated lower tropospheric warming and mid-to-upper tropospheric warming, while RSSv4 under-estimated lower tropospheric warming. Thus the radiosonde-based analyses and re-analyses argue against the lower tropospheric warming trend from UAHv6 and RSSv3.

So Curry and Spencer cite radiosonde analyses that actually support RSS' corrections to RSSv3, contrary to what Spencer claims. Curry should be aware that radiosondes under-estimate tropospheric warming up to 2012, since the CCSP pointed this out in the report Curry reviewed. Yet in her blogpost comments on radiosonde trends, Curry did not address any of the published evidence on radiosondes under-estimating tropospheric warming. Instead she cast aspersions on scientists for "massag[ing]" and "interpret[ing]" the radiosonde data [8]. Given these aspersions, Curry might object to my above discussion of radiosonde analyses.

Curry's criticism fails for at least three reasons. First, her criticism does nothing to address published research that validates radiosonde homogenization methods, that estimates uncertainty in radiosonde trends, or that points out that radiosonde trends under-estimate warming. Second, her criticism represents a double standard where she discourages scientists from investigating and homogenizing radiosonde data, in the same blogpost in which she encourages scientists to investigate and homogenize satellite data. Third, Curry complains about "massag[ed]" and "interpret[ed]" radiosonde analyses [8], when these are some of the same radiosonde analyses Spencer uses to support his position. But Curry makes no comment on Spencer's use of analyses Curry considers to be illegitimately "massag[ed]" and "interpret[ed] [8]." Thus Curry may not believe her own criticism, or she may apply the criticism in an illegitimately selective way.

So Curry's critique of the radiosonde analyses lacks merit, and the radiosonde trends undermine her objections to RSSv4's homogenization. To bolster her suspicions regarding RSS' homogenization improvements, Curry copied a blogpost from Spencer. In this blogspost, Spencer critiqued RSS' v4 TMT corrections and he suggested that his critique would extend to RSSv4 TLT. Curry also adds her own comments to Spencer's analysis. Neither Curry nor Spencer make cogent points.

Both Curry and Spencer suggest that RSS increased the post-1998 warming trend largely by side-stepping issues with the calibration for two satellites: NOAA-14 and NOAA-15 (Zou also discusses issues with NOAA-15's calibration, though not in a way critical of RSS' work). Spencer argues that NOAA-14's calibration was worse than NOAA-15's calibration, such that one should exclude NOAA-14 data after 2001 and keep the NOAA-15 data for this time-period until 2007. He also suggests that NOAA-15 underwent less orbital drift than NOAA-14 during the time-period the two satellites overlapped; NOAA-14's greater drift would then (supposedly) warm its instruments overtime, skewing its results more than NOAA-15.

However, published research already showed calibration issues with some of the NOAA-15 instruments used by the UAH and RSS teams. And other researchers using a different set of instruments on NOAA-15, noted how NOAA drifted enough to skew the results from these instruments. Unfortunately, these researchers could not compare NOAA-15 to NOAA-14 since the particular instrument they needed on NOAA-14 had been corrupted since NOAA-14's launch.

Spencer and Christy's UAHv6 adjustments for NOAA-14 and NOAA-15, among other factors, decrease UAHv6 post-1999 warming in comparison to UAHv5. Figure 6 below depicts this. One can assess the accuracy of these adjustments by comparing UAH's trend over this time-period to trends from other data sources, such as radiosonde-based analyses. This comparison shows that the adjustments cause UAHv6 to under-estimate post-1999 warming, as per figure 7. RSS also under-estimates post-1999 warming, but not as much as UAH. This further undermines Spencer's case on NOAA-14 and NOAA-15 adjustments. Spencer and Christy sometimes conceal this point by excluding most of the post-1999 period when comparing their UAHv6 analysis to other sources, thereby creating the false impression that UAHv6's warming trend matches those sources. Christy knowingly does this in a two-faced manner; he claims the radiosonde-based trends support his UAH analysis when he speaks to non-experts, but admits that those trends conflict with his UAH analysis when he communicates with informed experts who are harder to fool.

Figure 6: Near-global TLT from 1979 - 2015 for UAHv5 and UAHv6 (the red line and black line, respectively), relative to a baseline of 1980 - 2010. The purple line represents the difference between the relative values of UAHv6 and UAHv5 at each time-point, depicted with an offset of 1°C downwards [17, figure 7].

Figure 7: Near-global TLT for weather-balloon-based analyses, satellite analyses, and re-analyses. Figure 5 above presents the corresponding temperature trends from these analyses, in the form of LTT. The data processing for UAH TLT is slightly different from the other analyses, causing its global trend value to be typically cooler by 0.01°C/decade in comparison to the other LTT trends [12, page S17 (also see 18; 19 - 22; 23 - 29, generated using 30, as per 31; 32)].

Despite co-authoring the above figure, less than four months before the figure was published, Christy still falsely claimed to non-experts that the weather-balloon-based analyses supported his UAHv6.0 analysis over RSSv4.0.

Spencer and Curry also insinuate that the RSS team avoided having UAH team members peer review RSS' paper, because the RSS team could not adequately respond to the UAH team's criticisms regarding NOAA-14 and NOAA-15. It is strange to see Spencer make this claim, since he (supposedly) predicted that no UAH team member would be asked to review RSS' TLT paper, while also admitting that UAH team member John Christy reviewed RSS' paper. And even if the UAH team did not peer review the paper, other competent experts could review the paper, including Weng, Zou, the UW research team and the NOAA/STAR research team. Spencer should be aware of these individuals, since he comments on their research. So Spencer should know better than to suggest that the RSS team avoided competent peer review in an attempt to avoid addressing the NOAA-14 calibration issue.

But the most decisive blow against Spencer and Curry's calibration insinuation is that RSS addressed their criticism in RSS' TLT paper. For instance, RSS notes that the calibration shift likely has a relatively small effect, contrary to Spencer's blogpost claims:

"As was the case for TMT, we suspect differences are caused by a spurious calibration drift in either NOAA-14 or NOAA-15 (or both). [...]. If we exclude MSU data after 1999 (implicitly assuming the error is due to NOAA-14), the long-term trend decreases by 0.008 K decade-1, and if we exclude AMSU data before 2003 (implicitly assuming the error is due to NOAA-15), the long-term trend increases by 0.007 K decade-1 [7, section 3h]."

The RSS team makes a similar point in the TMT paper Spencer commented on:

"If we exclude MSU data after 1999 (implicitly assuming the error is due to NOAA-14), the long-term trend decreases by 0.019 K decade−1, and if we exclude AMSU data before 2003 (implicitly assuming the error is due to NOAA-15), the long-term trend increases by 0.01 K decade−1) [9, section 3f]."

In a March 2018 peer-reviewed paper, Spencer and Christy again criticize RSS' treatment of NOAA-14 and NOAA-15, while citing radiosonde analyses and re-analyses. I critique this paper in the parenthetical comment below.

{Spencer's March 2018 paper still relies on ERA-I's artificially low warming trend and the artificially low post-1979 radiosonde-based trends that I discussed above. ERA5, the update to ERA-I, likely undermines Spencer's paper by showing greater tropospheric warming, as per figure 5. Spencer's paper then uses these analyses to claim RSS's tropical mid-to-upper tropospheric warming trend significantly differs from the average of other analyses. But Christy's figure 5 above rebuts this claim; RSS' value of 0.17°C/decade falls within the median range of 0.15 +/- 0.03°C/decade. RSS team member Carl Mears also co-authored a recent conference abstract that argues against the paper's conclusion, as applied TLT. Mears and UAH team member John Christy co-authored a 2017 report that supports the abstract's results, by showing that RSS' and UAH's updated TLT analyses showed less warming than radiosonde analyses and re-analyses for the past couple of decades. Moreover, Spencer's 2018 paper conflicts with the conclusions of published radiosonde analyses, along with the results of a published GPS radio occultation analysis on post-2001 upper tropospheric warming trends. The paper also continues Spencer and Christy's tradition of flawed comparisons of models vs. observational analyses; I discuss some of the relevant flaws in "Myth: Santer et al. Show that Climate Models are Very Flawed". And I cite some relevant literature on Spencer's paper in a separate multi-tweet Twitter thread [11].}

So this 2018 paper does not adequately rebut RSS' points regarding post-1999 trends. Thus this paper does not justify Curry's concern that the RSS analysis might illegitimately do away with the (supposed) "pause" in post-1998 warming. If anything, the radiosonde and re-analysis evidence suggests that RSS' treatment of NOAA-14 and NOAA-15 caused them to under-estimate lower tropospheric warming over the past couple of decades in their TLTv4 analysis. On this line of reasoning, UAH's TLTv6 treatment under-estimated warming even more egregiously. Spencer's NOAA-14- and NOAA-15-based critique of RSS, endorsed by Curry, therefore lacks merit.

Curry and Spencer's blogposts also criticize RSS' diurnal drift homogenization, since RSS' homogenization uses a general circulation climate model (GCM), instead of UAH's "empirical [5]" method of diurnal drift correction. One might remain justifiably skeptical of Spencer's comments on diurnal drift, since RSS pointed out the obviously incorrect diurnal drift correction that Spencer used during his time on the UAH team, as I discussed in section 2.1. However, Spencer also pointed out more subtle flaws in RSSv3, so one should not immediately dismiss Spencer's points out-of-hand. But further examination reveals that Spencer and Curry offer misleading comments on RSSv4's improved homogenization.

Spencer and Curry conveniently leave out how RSS canvassed multiple approaches for diurnal drift correction, using both model-based and empirical methods. Moreover, they avoid mentioning that empirical methods of diurnal drift suffer from their own problems, as discussed by United States Global Change Research Program. Ironically, these problems include calibration shifts, the very issue Spencer and Curry brought up as an objection to RSS' improved homogenization. The UW team also pointed out issues with UAH's diurnal drift correction. So taken together, these points cast doubt on UAH's "empirical" diurnal drift correction relative to RSS' model-based correction.

(Though Spencer and Curry do not make this argument, one might be tempted to argue that RSS's long-term tropospheric warming trends must match long-term warming trends from climate models, since RSS used a climate model in their homogenization. But this argument lacks merit, since using models for diurnal correction does not guarantee that the the resulting tropospheric trends will match multi-decadal trends from models. Case-in-point: the RSSv.3.3 TLT trend was lower than the average model-based TLT projection, even though RSSv3.3 TLT used climate models for diurnal drift homogenization. So one cannot both argue that RSS' analyses diverge from model-based projections, and argue that RSS' model-based diurnal drift correction guarantees that RSS' analyses will agree with model-based projections; such a position is self-refuting.)

In addition to presenting results from their model-based correction, the RSS team also presents results from UW's validated, "empirical" method of diurnal drift correction that uses satellites with more stable orbits (Zou employs an alternative, "empirical" method). The UW TTT trend remains significantly greater than UAH's TTT trend, as shown in figure 2 above. UW's TTT trend is also closer to TTT trends that use model-based diurnal drift correction, such as RSSv4 TTT (see figures 2 and 5 above). Moreover, UW's TMT trends is closer to RSSv4's trend than to UAH's trend, as shown in figure 8 below:

Figure 8: (A) Near-global and (B) tropical TMT trends from 1979 - 2015 for RSS, UAH, NOAA/STAR, and UW, relative to a baseline of 1979. Differences between satellite-based analyses for (C) near-global and (D) tropical TMT trends. For example, the green line "RSS V4.0 - STAR V3.0" displays the result of subtracting NOAA/STAR version 3 for a given year from RSS version 4 for that same year [9, figure 9].

So even if one preferred an "empirical" method of diurnal drift correction instead of model-based methods, that would not be enough to argue for UAH's analysis over RSS' analysis. Curry has no excuse for being unaware UW's analysis, since the RSS team compared their results to UW's results in the RSS TMT paper Curry discusses, as shown in figure 8 above. So the RSS paper Curry cited gave Curry some of the information she needed for debunking Spencer's point, even though Curry seems unaware that RSS' paper provides this information.

Section 2.3: Why Curry Ran Afoul of the Evidence

At this point, one might wonder why Judith Curry fell for Roy Spencer's blogpost critique of RSSv4 TMT, instead of investing more time in reading RSS' peer-reviewed TMT paper. There is no inherent problem with Curry (or anyone else) reading blogposts, as long as they pay sufficient attention to reputable, peer-reviewed scientific sources. Though peer review is not perfect, competent peer review helps prevent nonsense from being published, in contrast to the nonsense published on many denialist blogs. Many AGW denialists, AIDS denialists, young Earth creationists, and other science denialists evade this point by inventing paranoid conspiracy theories about peer review and by citing fake experts who's views contradict published evidence.

Curry should know better, since she once told a critic of her's to:

"Read the literature and the IPCC [the IPCC is the United Nations' Intergovernmental Panel on Climate Change] [10]."

Unfortunately, Curry fell short of her own standard. She made the mistake of relying on Spencer's non-peer-reviewed blogpost, instead of paying sufficient attention to the peer-reviewed scientific literature; she did not even pay sufficient attention to published studies on the ERA-I analysis and RSS analysis she cited. Curry was therefore left in a woefully uninformed position on the topics she discussed. Curry's hostility towards much mainstream climate science may explain why she attacked RSS' homogenization methods and cast doubt on the hot spot's existence, despite the fact that Curry cites published sources that debunk her position on these subjects. This is not the first time Curry has done something like this, as I discuss in "Myth: Judith Curry Fully and Accurately Represents Scientific Research".

Curry's example illustrate the dangers of solely relying on Internet critics (including me) for one's information on science; it always pays to read the peer-reviewed scientific literature, to ensure that Internet critics accurately represent what the scientific evidence shows and that they do not distort the evidence to support an ideologically-convenient position.