Blog Archive

Friday, April 12, 2013

Hoerling Howler -- uses model with no skill, publishes faulty report on NOAA website, lazy journalists fall for it (@ Seth Borenstein, Andrew Freedman, Suzanne Goldenberg -- way to go!) -- mega-drought continues

Yes, Climate Change Is Worsening U.S. Drought — NOAA Report Needlessly Confuses The Issue

NOAA has issued a report on a small part of the recent brutal droughts that have hit the United States over the past few years. The report — “An Interpretation of the Origins of the 2012 Central Great Plains Drought” — is needlessly confusing, scientifically problematic, and already leading to misleading headlines.
Dr. Kevin Trenberth, former head of the Climate Analysis Section at the National Center for Atmospheric Research, has sent to reporters a Commentary on the report, which I repost below. He concludes:
This report has some useful material in it describing aspects of the drought in 2012 in the central US. But it is quite incomplete in many respects, and it asks the wrong questions.  Then it does not provide very useful answers to the questions that are asked.
Indeed, it seems odd to do a 44-page report on the drought in the Central Great Plains (in the spring and summer of 2012) when so much of the Great Plains — and Southwest — have also been in a brutal extended drought that continues to this day.
As InsideClimate News reported two weeks ago:
Drought conditions in more than half of the United States have slipped into a pattern that climatologists say is uncomfortably similar to the most severe droughts in recent U.S. history, including the 1930s Dust Bowl and the widespread 1950s drought.
The 2013 drought season is already off to a worse start than in 2012 or 2011….
A variety of leading experts explained how human-caused warming worsened the 2011 drought (see Warming-Enhanced Texas Drought Is Once in “500 or 1,000 Years … Basically Off the Charts,” Says State Climatologist). And numerous scientific studies have projected that global warming will dry out the Southwest and at least parts of the Central Great Plains:
Amazingly, in the AP story, “Federal report says don’t blame global warming for freak of nature 2012 US drought,” the lead author, NOAA’s Martin Hoerling, dismisses any possible role of Arctic ice loss in the drought:
Researchers focused on six states — Wyoming, Kansas, Nebraska, Colorado, Missouri and Iowa — but the drought spread much farther and eventually included nearly two-thirds of the Lower 48 states. For the six states, the drought was the worst four-month period for lack of rainfall since records started being kept in 1895, Hoerling said.
He said the jet stream that draws moisture north from the Gulf was stuck unusually north in Canada.
Other scientists have linked recent changes in the jet stream to shrinking Arctic sea ice, but Hoerling and study co-author Richard Seager of Columbia University said those global warming connections are not valid.
Hoerling used computer simulations to see if he could replicate the drought using man-made global warming conditions. He couldn’t. So that means it was a random event, he said.
Yet, a 2012 Geophysical Research Letters study, “Evidence linking Arctic amplification to extreme weather in mid-latitudes,” found that the loss of Arctic ice favors “extreme weather events that result from prolonged conditions, such as drought, flooding, cold spells, and heat waves.” Indeed, as Climate Central explained in April 2012:
One does not have to look hard to find an example of an extreme event that resulted from a huge, slow-moving swing in the jet stream. It was a stuck or “blocking weather pattern” – with a massive dome of high pressure parked across the eastern U.S. for more than a week – that led to the remarkable March heat wave that sent temperatures in the Midwest and Northeast soaring into the 80s. In some locations, temperatures spiked to more than 40 degrees above average for that time of year.
The strong area of high pressure shunted the jet stream far north into Canada. At one point during the heat wave, a jetliner flying at 30,000 feet could’ve hitched a ride on the jet stream from Texas straight north to Hudson Bay, Canada. In the U.S., more than 14,000 warm-weather records (record-warm daytime highs and record-warm overnight lows) were set or tied during the month of March, compared to about 700 cold records.
According to the study, Arctic climate change may increase the odds that such high-impact, blocking weather patterns will occur. The study cites examples of other patterns that led to extreme events that also may bear Arctic fingerprints, including the 2011 Texas drought and heat wave….
In fact, in October 2012, NOAA scientists themselves led research that found Arctic ice loss was driving a shift in Arctic summer winds that in turn “could also bring about shifts in North American and European weather,” as NOAA’s own release pointed out.
Indeed, the NOAA release explained:
The effects of Arctic amplification will increase as more summer ice retreats over coming decades. Enhanced warming of the Arctic affects the jet stream by slowing its west-to-east winds and by promoting larger north-south meanders in the flow. Predicting those meanders and where the weather associated with them will be located in any given year, however, remains a challenge.
The researchers say that with more solar energy going into the Arctic Ocean because of lost ice, there is reason to expect more extreme weather events, such as heavy snowfall, heat waves, and flooding in North America and Europe but these will vary in location, intensity, and timescales.
The point is that just because Hoerling couldn’t replicate the drought with his computer simulations doesn’t mean climate change had nothing to do with the 2012 Central Great Plains Drought — let alone the entire 2012 drought and the current 2013 drought.
As I’ll discuss in a future post, other models have predicted that Arctic ice loss will drive drought in the U.S. Southwest and adjacent areas.
Finally, Dr. Kevin Trenberth argues in an extended analysis that in fact Hoerling’s analysis is seriously flawed:
This report has some useful material in it describing aspects of the drought in 2012 in the central US. But it is quite incomplete in many respects, and it asks the wrong questions. Then it does not provide very useful answers to the questions that are asked.
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. Soil moisture is touched upon in Fig. 8 from a model but not validated with any data or indices of vegetation health such as NDVIIn Colorado on 1 May 2012, snow pack was lowest on record since 1968 and just 19% of average. As a result there was subsequently no snow melt and soil moisture or runoff. All the heat went into desiccating vegetation and raising temperatures and there was no snow melt or evaporative cooling effects to be had. In their Fig. 1 they show that on 1 May 2012 the widespread drought was already present throughout the entire Southwestern parts of the country, and this of course translated into extreme fire risk by June when major wildfires caused havoc in Colorado and burned over 600 houses. The dryness and heat appeared to spread eastwards, even as Colorado had some rains in July. But none of this is mentioned.
In the experiments performed with climate models, no indication is given that the model used or the forecast results from several other models have any skill or utility at the task set them. The distinctive La Niña pattern in 2011 giving extremes of dryness in Texas and wetness further north was not simulated or predicted either! In the lower 48, it has been distinctly wetter after about the 1970s in all seasons other than winter, but none of the models simulate this. Not one! The model biases are not dealt with and their skill, or lack of it, is not given. They are not shown to be appropriate to the task at hand. There is a complete failure to provide any reasons to believe the results. Moreover the experiments are woefully incomplete. SSTs [sea surface temperatures] were specified but no attempt was made to include soil moisture, snow cover anomalies, or vegetation health, for instance.
It is well established that all events such as the one nominally analyzed have a large component from natural variability that create anticyclonic conditions (which are better seen at 300 hPa not 500 hPa). The question never addressed is what does global warming and human influences bring to that?  
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.
There are several other things that are quite misleading. They deal with the “morphology” of the drought. Certainly that should include a comprehensive description of the factors involved, including actual soil moisture, snow cover, vegetation, and evapotranspiration, for instance. This includes what they call the “proximate cause” but which is really just part of the description and not a “cause” at all. Yet that is quite incomplete.
Several aspects of the report comment almost as if the extreme dryness and heat did not really happen. It could not be simulated or predicted and no cause can be found, so did it really happen? The paper fails to offer a plausible way in which the whole drought evolved and the role of humans, and climate change, and even natural variability.
The fact is that such events have happened somewhere on the planet every year in recent times. Perhaps we could go back to the 2003 heat wave in Europe. Years of drought in Australia led to the exceptional drought and heat waves and wild fires in February 2009 in particular, and again in 2013 (with record flooding in 2010-11); Russia in 2010, and the US in 2011 and 2012.
Some things were mixed and turn out to distort the results: comparing the summer of 2012 with the whole year of 1934 and 1936. Both of the latter years were dryer than 2012 for the 48 contiguous states for the year as a whole. Why the focus was on the Great Plains (which were not defined) is not clear. In some Figures, Colorado is included but it is not a “Great Plains” State.
So wrt global warming and increased greenhouse gases, say we have 1 W m-2 of extra heat for, say, the first 6 months of 2012 and it has no where to go because there is no snow pack or moisture. That is 0.1 W m-2 per sq foot. I use a sq ft as this is about the size of a small microwave oven. So if we add this up for 180 days, it amounts to zapping everything with a microwave oven (every sq foot) at full power for 36 minutes. (That’s at 700 W). No wonder we had major wild fires here in Colorado. The report fails completely to deal with the cumulative effects of drought on heat and wild fire risk.

No comments: