Thursday, April 19, 2012

The Hadley Centre Jumps to Conclusions

Science is a process.  Scientific understanding is gained over time, through discussion, inquiry, agreement and disagreement. Indeed, while there can be agreement on some aspects (eg. hypothesis, methods, etc) they may disagree on the conclusions or the implications of the conclusions for science and society.  
Researchers with the Met Office Hadley Centre published a paper in Nature earlier this month. The article makes the following claim:
We have shown that volcanic and aerosol processes can drive pronounced multidecadal variability in historical NASST, which leads to improved (for the early twentieth century) or reproduces (for the later period) the observed historical trends. In these simulations, it is the inclusion of aerosol indirect effects that allows us to capture the magnitude and the temporal and spatial structure of SST variability. Our results show that volcanoes and, crucially (from a policy and climate impact perspective), anthropogenic emissions of aerosols can drive NASST variability resembling that which is observed. This work suggests that we need to reassess the current attribution to natural ocean variability of a number of prominent past climate impacts linked to NASSTs, such as Sahel drought.
The release of the paper was met with a number of questions and considerations from the scientific community.

Roger Pielke Sr. took issue with one of the authors' assumptions
Thus they neglect to  consider that the reason the models themselves do not reproduce these interactions [i.e. systematic climate shifts linked to multidecadal variability], is due, not to just the neglect of human aerosol forcing, but because the models inaccurately represent the multidecadal variability even in the absence of human climate forcings.
Nonetheless he suggests that while the article was an "informative study"  the conclusions were off the mark
It should have been written, however, as a sensitivity experiment not as an attribution study since the multi-decadal model simulations have not shown skill at predicting North Atlantic climate pattern variations prior to significant human climate forcings.
Judy Curry at her blog took issue with the how aerosols were modeled in the study
Specifying aerosol characteristics (which is mostly done here, esp sulfate) and then allowing interactive cloud microphysics and optics results in an overestimate of the aerosol indirect effect, since compensating dynamics and precip don’t influence the aerosols. 
She concludes that she is "unconvinced" by the paper and questions Nature's wisdom in publishing the article.

Likewise, others dismissed the paper all together and considered it an indication of a decline in credibility of Nature.  Pat Michaels of George Mason University and the CATO Institute suggested that
This paper marks, in my opinion, the death of credibility for Nature on global warming.  
May God rest the soul of Nature.
Presumably, accepting this perspective, the article was reprinted at Watts Up With That.

As commentary on the article in Nature, Amato Evan, pointed out that assuming that the article is spot on then one implication would be that
the AMO does not exist, in the sense that the temperature variations concerned are neither intrinsically oscillatory nor purely multidecadal.  
and therefore,
swings in hurricane frequency and intensity might therefore be the regional response to variations in the concentration of pollutant aerosols against a background of global warming, and thus completely man-made. 

And so, it would seem that there is much to discuss regarding the article, its findings, and its implications, if any.  Yet, the lead author has moved to quickly debunk over a decade's worth of research on Atlantic Ocean oscillations with one short Letter in Nature
Our research implies that far from being natural, these [SST] changes could have been largely driven by dirty pollution and volcanoes. If so, this means a number of natural disasters linked to these ocean fluctuations, such as persistent African drought during the 1970's and 80's, may not be so natural after all.
But the third author, Paul R. Halloran, realizes the preliminary nature of their conclusions
However, it's important to note that these findings are based on only one model, so further research using other next-generation climate models is required to shed further light on the mechanisms at play.
Why would some be interested in fast forwarding the process of discussion and further inquiry?

The publication of the research was quickly picked up by the financial industry.  The FT ran with an article titled "Study finds air clean-up linked to US hurricanes" (which is inaccurate).  The article further suggested that decreasing air pollution "helped to cause more disastrous hurricanes" (also, way unfounded).

The article continues, indicating the societal problem that the Center had in mind when producing the information and identifying its conclusions,

The study is likely to be closely read by the insurance industry, said Matt Huddleston, Met Office principal consultant for insurance. 
“The industry uses near­term estimates of how many hurricanes will hit land, and the Atlantic surface temperature is an important component of that,” he said. 
“The research means they can better understand why rates of hurricane landfalls may have varied, and also it raises the potential of prediction if there were short­term changes in pollution levels.” (emphasis mine)

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.