Reversing the Null Hypothesis is Bad Climate Science
The Kevin Trenberth Effect: Pulling Science Back to the Dark Ages - Part 1 Droughts and Heat waves
In my essay on the natural causes of Pacifica’s Coastal Erosion, I reported on how California’s coast has still not reached an equilibrium with sea levels that rose at the end of the last ice age. I also suggested the media and a few scientists give the public a false impression that all natural weather phenomenon and coastal erosion have been worsened by CO2-driven climate change. Pointing to a few leading perpetrators I wrote, “After centuries of scientific progress, Trenberth and his ilk have devolved climate science to the pre-Copernican days so that humans are once again at the center of the universe, and our carbon sins are responsible for every problem caused by an ever-changing natural world.” Such a strong statement deserves further elaboration. Although a highly intelligent scientist, to support his obsessive claims that CO2-caused climate change has worsened every extreme event, Trenberth has been tragically undermining the very foundations of scientific inquiry by 1) reversing the proper null hypothesis, 2) promoting methods that can not be falsified, 3) promoting fallacious arguments only by authority, and 4) stifling any debate that promotes alternative explanations.
As Carl Sagan wrote regards arguments from authority:
"One of the great commandments of science is, "Mistrust arguments from authority." ... Too many such arguments have proved too painfully wrong. Authorities must prove their contentions like everybody else.
Dr. Trenberth, via his well-groomed media conduits, preaches to the public that every extreme event - flood or drought, heat wave or snowstorm - is worsened by rising CO2. To fully appreciate the pitfalls of his “warmer and wetter” meme, you need to look no further than Trenberth’s pronouncements regards the devastating Moore, Oklahoma tornado. Although Trenberth admits, “climate change from human influences is difficult to perceive and detect because natural weather-related variability is large”, in a Scientific American interview, arguing only from authority, he cavalierly attributed CO2 climate change to a “5 to 10 percent effect in terms of the instability and subsequent rainfall, but it translates into up to a 33 percent effect in terms of damage.” But in contrast to Trenberth’s “warmer and wetter world” assertions, there was no warming contribution. Maximum temperatures in Oklahoma had been cooler since the 1940s.(The 14 USHCN stations surrounding Moore OK and used to compute average were Bartlesville, Buffalo, Cherokee, Claremore, Enid, Guthrie, Jefferson, Meeker, Miami, Mutual, Okeene, Paul’s Valley, Pawhuska, and Perry)
Clearly Trenberth’s simplistic “warmer and wetter” world assertion cannot be applied willy-nilly to every region. Climate change is not globally homogenous. It is regionally variable and the global average temperature is a chimera of that regional variability. Furthermore his claim of a “wetter world” is a hypothetical argument not supported by evidence. As seen in the graph below from the peer-reviewed paper Weather And Climate Analyses Using Improved Global Water Vapor Observations, there is little evidence of a steady increase in water vapor paralleling rising CO2. Even Trenberth’s own studies have concluded, “Total Precipitable Water vapor [TPW] variability for 1988–2001 was dominated by the evolution of ENSO [El Ninos].”
The El Nino effect is evidenced by peak water vapor coinciding with the 1998 El Nino. Since 1998, the atmosphere has been arguably drier, contradicting his CO2 driven "wetter world" hypothesis. Despite a multitude of contradictions, to garner support for his theories Trenberth insists on reframing the scientific method by reversing the null hypothesis. Instead of determining if CO2 had an effect on extreme weather beyond what natural variability predicts, Trenberth wants scientists and the public to blindly assume, “All weather events are affected by climate change because the environment in which they occur is warmer and moister than it used to be.”
In contrast to simply making the “Trenberth assumptions”, climate scientists use two main strategies to extract any possible CO2 effect. First based on physics, the consensus believes early changes in CO2 concentration exerted no significant climate impact, and extreme events happening before 1950 were due to natural variability. Thus historical analyses compare extreme events before and after 1950 to determine how they differ. But Trenberth has been maneuvering to make such CO2 attribution studies non-falsifiable by stripping recent extreme weather events from that historical framework. In the Washington Post, Chris Mooney pushes Trenberth’s “new normal” quoting,
“All storms, without exception, are different. Even if most of them look just like the ones we used to have, they are not the same.”
Trenberth’s “new normal” side steps historical scientific analyses. One would think a good investigative reporter would question Trenberth’s undermining of that scientific methodology, but Mooney is not a scientist. Ironically Mooney’s claim to fame was a book “The Republican War On Science”, about which Washington Post’s Keay Davidson wrote, "Mooney is like a judge who interprets a law one way to convict his enemies and another way to acquit his friends.” Evidently that is just the kind of journalist Trenberth and the Washington Post wanted. Mooney left Mother Jones and was hired by the Washington Post to write columns on climate change and serve as one of Trenberth’s media conduits. (Btw: the Real Science website is a great place to view headlines from the past illustrating great similarities between past and present extreme weather events.)
The second strategy relies on models that compare “the probability of an observed weather event in the real world with that of the ‘same’ event in a hypothetical world without global warming.” But this approach incorrectly assumes the natural variability is well modeled. Often the model’s “world without global warming” is assumed to be stationary but with a lot of “noise”. But that tactic generates false probabilities because our natural climate is not stationary but oscillating.
In 2012 climate experts met at Oxford University to discuss such attribution studies and the highlights were reported in Nature. Many experts suggested that due to “the current state of modeling any attribution would be unreliable, and perhaps impossible…One critic argued that, given the insufficient observational data and the coarse and mathematically far-from-perfect climate models used to generate attribution claims, they [attribution claims] are unjustifiably speculative, basically unverifiable and better not made at all. And even if event attribution were reliable, another speaker added, the notion that it is useful for any section of society is unproven.”
(Such concerns raise another question: if attributing a CO2 effect on any event like a heat wave or drought is nearly impossible, how reliable is any attribution of a global average temperature if those same extreme heat waves and droughts skew the global average?)
Color me an old-fashioned scientist, but our best practices demand we correctly establish the boundaries of natural climate change before we can ever assume rising CO2 has worsened weather events. But Trenberth and his ilk insist on reversing the null hypothesis. Instead of asking if a weather event exceeded natural variability, Trenberth insists we rashly assume CO2 has already worsened the weather. However most scientists share my concern about his maneuverings. As Professor Myles Allen from Oxford University said, “I doubt Trenberth’s suggestion [reversing the null hypothesis] will find much support in the scientific community.” Trenberth’s attempt to reverse the null hypothesis has been discussed previously by Dr. Judith Curry and by top rated skeptic blogs, and in a published paper by Dr. Allen “In Defense of the Traditional Null Hypothesis.
Nonetheless many papers are now being published that simply make Trenberth’s assumptions, and there is a growing rift between researchers who adopt Trenberth’s “new normal” tactics versus “old school” scientists. The different resulting scientific interpretations are well illustrated in peer-reviewed publications on droughts and heat waves.
A bank account serves as a good analogy to illustrate drought stress. Financial (hydrologic) stress results from changes in income (rain and snow) versus withdrawals (evaporation and runoff) and the buffering capacity of your reserves (lakes, wetlands and subsurface water). Old school science would demand researchers eliminate all confounding factors affecting hydrological stress before claiming any effect by a single variable like CO2. Here are a few confounding factors that are seldom addressed in papers that blame a greenhouse effect for higher temperatures and stronger heat waves and droughts.
The difference between “old school” science and Trenberth’s “new normal” is illustrated in contrasting interpretations of recent extreme droughts and heat waves. For example NOAA’s Randall Dole attributed the 2010 Russian heat wave to a lack of precipitation and a high-pressure blocking pattern that enhanced surface feedbacks. Dole had been studying the effects of blocking patterns for over 30 years since his research days at Harvard. Blocking high-pressure systems pump warm air northward on the systems western flank and trap that heat while clear skies increase insolation. In 1982 Dole had mapped out 3 regions most prone to blocking highs due to undulations of the jet stream. Those same blocking highs are also implicated in our more recent heat waves that “Trenberth’s school of climate change” trumpet as worsened by CO2. The 3 regions are 1.) Northeast Pacific where the “Ridiculous Resilient Ridge” typically produces California’s drought, 2) North Atlantic that affects Western Europe’s droughts and 3) over northern Russia generating heat waves every 20 years such as the 2010 heat wave.
Dole concluded in 2011 “the intense 2010 Russian heat wave was mainly due to natural internal atmospheric variability.” Dole’s historical analysis noted, “The July surface temperatures for the region impacted by the 2010 Russian heat wave show no significant warming trend over the prior 130-year period from 1880 to 2009” and he noted similar but slightly less extreme heat waves had occurred periodically over the past 130-year period. The more extreme temperatures could be attributed to “surface feedbacks” from the early season drought and landscape changes. Based on a proper null hypothesis Dole concluded, “For this region an anthropogenic climate change signal has yet to emerge above the natural background variability.”
Whether or not Dole is correct, Dole is a climate scientist we can trust. A trustworthy scientist, who cannot detect a difference between a recent extreme event and natural extreme events from the past, will simply report that they cannot detect an anthropogenic signal. Whether or not there was a CO2 global warming effect remains to be tested. In contrast less trustworthy scientists will push a non-falsifiable CO2 effect and argue natural variability “masked CO2 warming,” a warming Trenberth insists we must assume to be present.
Trenberth also appears to hate any scientific claim that weather was just weather. Accordingly he attacked Dole’s “heresy” via his internet attack dogs. Joe Romm blogged, “Monster Crop-Destroying Russian Heat Wave To Be Once-In-A-Decade Event By 2060s (Or Sooner)”, which provided Trenberth an opportunity to denigrate Dole’s analysis in a way not allowed in more staid scientific journals. Trenberth maligned Dole’s analysis as “superficial and does not come close to answering the question in an appropriate manner. Many statements are not justified and are actually irresponsible. The question itself is ill posed because we never expect to predict such a specific event under any circumstances, but with climate change, the odds of certain kinds of events do change.”
Seriously? Dole’s research was irresponsible because it found no CO2 effect?!? The great value of science to society is that it provides us with some measure of predictability that guides how we best adapt to future events. Dole simply asked, “Was There a Basis for Anticipating the 2010 Russian Heat Wave?” and concluded neither past weather patterns, current temperatures trends, historical precipitation trends or increasing CO2 could have prepared Russia for that event. The only predictability was that similar events had happened every 2 or 3 decades. But Trenberth has persistently argued the only “right question” to ask is “how much has CO2 worsened an extreme event, while Dole asked the more useful question. What triggers extreme Russian droughts and heat waves every 20 to 30 years?
Dole’s models, forced with sea ice or ocean temperatures, did not simulate the observed blocking patterns over Russia. Based on several modeling experiments Dole concluded results were “consistent with the interpretation that the Russian heat wave was primarily caused by internal atmospheric dynamical processes rather than observed ocean or sea ice states or greenhouse gas concentrations.” Yet despite Dole’s examination of a great breadth of contributing factors, Trenberth attacked Dole for being “too narrowly focused” because, of all things, Dole did not include July flooding in China and India, or record breaking floods in Pakistan in August. Trenberth was suggesting that that those floods were due to warmer oceans and thus global warming should have been blamed for worsening the Russian heat wave even though Dole’s modeling studies found no such connection.
But Trenberth had the tail wagging the dog - again! Due to the clockwise motion of a blocking High, warm air was pulled poleward and accumulated on the western side of the system driving the heat wave. In contrast the same system pushed colder air equatorward along the system’s leading eastern edge. As discussed in Hong 2011, when that cold air was pumped southward, it collided with warm moist air of the monsoons, and it was that cold air that increased the condensation that promoted extreme precipitation in some locales. Nonetheless, determined to connect CO2 warming to the Russian heat wave, it was Trenberth who was not asking the right questions. He should have been asking how much did a naturally occurring blocking pattern contribute to the southern Asian floods.
As was the case for the Russian heat wave, analyses of the historic heat wave for Texas and the Great Plains revealed no warming trend over the latter 20th and the 21st century. In Hoerling 2013, a team comprised of ten climate experts, mostly from NOAA, examined the Texas drought and heat wave. They reported “no systematic changes in the annual and warm season mean daily temperature have been detected over the Great Plains and Texas over the 62-yr period from 1948 to 2009 consistent with the notion of a regional ‘‘warming hole’’. Indeed, May–October maximum temperatures over the region have decreased by 0.9°C.” Thus those experts concluded the absence of observed warming since 1948 cautioned against attributing the heat wave and drought to any warming, natural or CO2 related. Likewise satellite data revealed a radiative signature of a reduced greenhouse effect and increased solar heating (Yin 2014). (In contrast CMIP5 modeled results suggested a 0.6°C warming effect since 1900).
In contrast Trenberth claimed on Romm’s blog, “Human climate change adds about a 1 percent to 2 percent effect every day in terms of more energy. So after a month or two this mounts up and helps dry things out. At that point all the heat goes into raising temperatures. So it mounts up to a point that once again records get broken. The extent of the extremes would not have occurred without human climate change.” But Trenberth’s 1% per day CO2 attribution seems absurd in a regions where maximum temperatures had decreased. His warmer and wetter world meme only obfuscated the issues and Trenberth was again asking the wrong question. The correct question was how much had the drought lowered surface moisture and reduced evaporative cooling that caused higher temperatures? In a region where there had been no increase in maximum temperatures, the amplified temperatures for this extreme weather event were likely the result of natural surface feedbacks caused by a lack of rain.
NOAA’s drought task force also reported on the following Great Plains drought and heat waves. They concluded this drought was likewise due to natural variability stating, “Climate simulations and empirical analysis suggest that neither the effects of ocean surface temperatures nor changes in greenhouse gas concentrations produced a substantial summertime dry signal.”
But no matter the level of expertise, Trenberth via his internet attack dog Joe Romm and his blog assailed the Drought Task Force with a less than an honest account. Trenberth assaulted their conclusions, “It fails completely to say anything about the observed soil moisture conditions, snow cover, and snow pack during the winter prior to the event in spite of the fact that snow pack was at record low levels in the winter and spring.” (But Trenberth’s denigration contrasted with a document-search for the term “soil moisture”, which found it was mentioned about 15 times including the sub-section title in big bold letters “Simulations of Precipitation and Soil Moisture”.) Trenberth’s mugging continued, “There is no discussion of evaporation, or potential evapotranspiration, which is greatly enhanced by increased heat-trapping greenhouse gases. In fact, given prevailing anticyclonic conditions, the expectation is for drought that is exacerbated by global warming, greatly increasing the heat waves and wild fire risk. The omission of any such considerations is a MAJOR failure of this publication.”
But that was a very odd comment for a top climate scientist! Anticyclonic conditions predict droughts will be exacerbated by natural feedbacks, not by global warming.
And again Trenberth failed to ask the right questions. If he believed a greenhouse effect exacerbated the drought by increasing evaporation, then he needed to ask why satellite data had been showing reduced downward long wave radiation and increased solar insolation that typically occur in dry clear skies? In contrast to Trenberth’s obfuscations, the Task Force had extensively discussed the meteorological conditions that inhibited the transport of moisture from the Gulf of Mexico, resulting in reduced soil moisture. Despite low snowpack, soil moisture had not been deficient in the spring. It was the lack of moisture transported from the Gulf that reduced summer soil moisture that raised temperatures and exacerbated the drought. Furthermore modeling experiments performed by the Task Force found precipitation was not affected by changes in sea surface temperatures or greenhouses gases. And historical analyses (as seen in Figure 7) revealed that despite global warming Central USA temperatures were lower than expected given the extreme dryness and expected surface feedbacks. So again Trenberth failed to ask the right questions. Why were temperatures higher during the droughts of the 30s when there was no increased greenhouse effect?
The 2011-2015 drought in California is the most flagrant example of the Trenberth Effect. California’s droughts are most often associated with natural La Nina conditions and a blocking ridge of high pressure that inhibits the flow of moisture from the Pacific to California. Another thorough analysis by NOAA’s Drought Task force again concluded, “the recent drought was dominated by natural variability.” In an interview with the NY Times co-author Dr. Hoerling stated, “It is quite clear that the scientific evidence does not support an argument that this current California drought is appreciably, if at all, linked to human-induced climate change.”
In support of the Drought Task Force’s conclusions, every study of the California drought has reported the major factor driving recent drought has been episodic rainfall deficits. Nonetheless despite the extreme rainfall shortfall there was no evidence of any trend in precipitation amounts or variability that could explain the recent lack of precipitation. Ridging patterns have always reduced rainfall, and the lack of a trend in precipitation contradicts recent claims that greenhouse gases are increasing the likelihood of a ridging pattern that was blocking precipitation (Swain 2014). Nonetheless media conduits for alarmism like Slandering Sou promoted Swain’s arguments. But Slandering Sou is not a scientist nor has she ever published any meaningful science. In contrast climate scientists like Dr. Cliff Mass readily pointed out Swain’s faulty analyses.
Furthermore there is no long-term precipitation trend as seen in the 700-year California Blue Oak study by Griffin 2015. The dashed blue line represents the extreme precipitation anomaly of 2014. For the past 700 years similar extreme precipitation shortfalls have equaled or exceeded 2014 several times every century. From a historical perspective, we can infer there is no evidence that rising CO2 has increased that ridging pattern that reduces rainfall and causes drought. More severe and enduring droughts happened during the Little Ice Age when temperatures were cooler. Clearly land managers and government agencies should prepare for severe periodic droughts whether or not CO2 has any effect or not, testifying to why the Oxford attendees saw little usefulness in CO2 attribution studies.
As expected Trenberth’s attack dogs assailed NOAA’s California report because it attributed drought to natural variability. Romm blogged that the drought would Soon Be More Dire. Over at the Washington Post, Mooney’s fellow yellow journalist Darryl Fears wrote “California’s terrifying climate forecast: It could face droughts nearly every year.” But Fears’ projection has already failed. Despite no precipitation trends, several authors blamed the California drought on extremely high temperatures. Michael Mann argued “Don’t Blame It on the Rain”. Blame it on global warming. To support warming assertions Trenberth blogged a fanciful analogy, “The extra heat from the increase in heat trapping gases in the atmosphere over six months is equivalent to running a small microwave oven at full power for about half an hour over every square foot of the land under the drought.” If that wasn’t fearful enough Trenberth added, “No wonder wild fires have increased!”
But historical analyses suggest the universe had unplugged Trenberth’s “microwaves” over most of California since 1940s, and wildfires were much worse during the Little Ice Age. As shown in the illustration below from Rapacciuolo 2014, observations show most of California, like Texas, had experienced a decline in the maximum temperatures since 1940. If maximum temperatures have not risen there has been no accumulation of heat and California appears to be insensitive to rising CO2. The question that Trenberth failed to ask is why did maximum temperatures decline in his “warmer and wetter” world?
Mao 2012 analyzed the drought in California’s Sierra Nevada and likewise found no trend in maximum temperatures. However assuming the minimum temperature trend was an expression of anthropogenic warming, he used the minimum trend to model CO2-warming effects on drought. But minimum temperatures have little effect on drought. Relative humidity is highest and approaches the dew point during the minimum. Due to daytime surface-heating, turbulent convection peaks around the maximum temperatures and increases evaporation and dries the soil dramatically. But turbulent convection is virtually non-existent when minimum temperatures are measured. Accordingly based on the minimum temperature trend, Mao 2012 found “warming may have slightly exacerbated some extreme events (including the 2013–2014 drought and the 1976–1977 drought of record), but the effect is modest; instead, these drought events are mainly the result of variability in precipitation.”
That brings us to the most recent example of how Trenberth’s “new normal” has undermined science. Williams 2015 claimed CO2 warming had worsened the California drought by 8 to 27%, a claim that was trumpeted by press releases and blogs. To his credit Williams did use a much better version of the Palmer Drought Severity Index (PDSI) that takes into account the physical processes causing a drought. He also pointed out that simpler versions of the PDSI, used by Diffenbaugh 2015, Griffin 2014 and others, had artificially amplified and overestimated the contribution of temperatures to drought (Sheffield 2012, Roderick 2009).
However Williams claimed to have separated anthropogenic warming from natural warming, but used the reversed null hypothesis to do so. Williams warned that he had assumed any warming trend was all anthropogenic. To determine natural temperature variability he simply subtracted his hypothetical anthropogenic-warming trend from California’s observed temperatures. Whatever remained was deemed natural variability. By assuming CO2 is responsible for any warming trend, alleviates climate scientists from the more arduous task of determining natural temperature variability. Furthermore instead of separating out the confounding factors that are known to contribute to higher temperatures, such as the PDO (Johnstone 2014) or landscape feedbacks (as discussed above), Williams simply acknowledged he did not account for those factors as a caveat, then went on to promote his human influence estimated in press releases suggesting he had scientifically linked CO2 warming to drought severity. Without accounting for all factors, Williams’s study was not a scientific evaluation, but simply an opinion piece. Still, as might be expected, Trenberth weighed in calling Williams analyses reasonable but conservative, and recommended that he drop the lower end (8%) of estimated human contribution.
But Williams and Trenberth never asked the right questions. How can scientists assume an anthropogenic warming trend if it hijacks the earlier warming trend before 1950, a trend that the consensus believes was all natural? How can scientists assume an anthropogenic warming trend when there has been no warming trend for maximum temperatures since 1950? How can scientists blame global warming for worsening droughts when other factors like the PDO, the drying of the California landscape and surface feed backs were never accounted for?
And more importantly, why should people ever trust Trenberth’s “new normal” science that undermines the very foundation of scientific inquiry. It is more than irksome that my taxes help pay Trenberth’s high salary and allow him to undermine the foundations of scientific inquiry.
In part two: Trenberth’s snowjob, I examine Trenberth’s fallacious argument that global warming causes more snow.