Controversy after IPCC Third Assessment Report Hockey stick graph

while ipcc third assessment report (tar) drew on 5 reconstructions support conclusion recent northern hemisphere temperatures warmest in past 1,000 years, gave particular prominence ipcc illustration based on mbh99 paper. hockey stick graph subsequently seen mass media , public central ipcc case global warming, had been based on other unrelated evidence. expert viewpoint graph was, newly published science, preliminary , uncertain, used publicise issue of global warming, , targeted opposing ratification of kyoto protocol on global warming.

a literature review willie , sallie baliunas, published in relatively obscure journal climate research on 31 january 2003, used data previous papers argue medieval warm period had been warmer 20th century, , recent warming not unusual. in march published extended paper in energy & environment, additional authors. bush administration s council on environmental quality chief of staff philip cooney inserted references papers in draft first environmental protection agency report on environment, , removed references reconstructions showing world temperatures rising on last 1,000 years. in , baliunas controversy, 2 scientists cited in papers said work misrepresented, , climate research paper criticised many other scientists, including several of journal s editors. on 8 july eos featured detailed rebuttal of both papers 13 scientists including mann , jones, presenting strong evidence , baliunas had used improper statistical methods. responding controversy, publisher of climate research upgraded hans von storch editor editor in chief, von storch decided , baliunas paper flawed , should not have been published was. proposed new editorial system, , though publisher of climate research agreed paper should not have been published uncorrected, rejected von storch s proposals improve editorial process, , von storch 3 other board members resigned. senator james m. inhofe stated belief manmade global warming greatest hoax ever perpetrated on american people , , hearing of united states senate committee on environment , public works convened on 29 july 2003 heard news of resignations.

toronto businessman stephen mcintyre downloaded datasets mbh99, , obtained mbh98 datasets request mann in april 2003. @ suggestion of sonja boehmer-christiansen, editor of social science journal energy & environment, mcintyre wrote article assistance of university of guelph economics professor ross mckitrick, energy & environment published on 27 october 2003. mcintyre & mckitrick 2003 paper (mm03) said mann, bradley & hughes 1998 (mbh98) hockey stick shape artefact of poor data handling , use of obsolete proxy records. criticism comprehensively refuted wahl & ammann 2007, showed errors in methods used mcintyre , mckitrick.

the statistical methods used in mbh reconstruction questioned in 2004 paper hans von storch team including eduardo zorita, said methodology used average data , wide uncertainties might have hidden abrupt climate changes, possibly large 20th century spike in measured temperatures. used pseudoproxy method mann , rutherford had developed in 2002, , them found regression methods of reconstruction tended underestimate amplitude of variation, problem covered wide error bars in mbh99. reasonable critique of reconstructions @ time, mbh singled out. other researchers subsequently found von storch paper had undisclosed additional step which, detrending data before estimating statistical relationships, had removed main pattern of variation. von storch et al. view graph defective overall refuted wahl, ritson , ammann (2006).

in 2004 mcintyre , mckitrick tried unsuccessfully extended analysis of hockey stick journal nature. @ stage nature contacted mann, bradley, , hughes, minor errors in online supplement mbh98. in corrigendum published on 1 july 2004 acknowledged mcintyre , mckitrick had pointed out errors in proxy data had been included supplementary information, , supplied full corrected listing of data. included documented archive of data used in mbh98, , expanded details of methods. stated none of these errors affect our published results.

the mcintyre , mckitrick comment accepted publication geophysical research letters. mcintyre & mckitrick 2005 (mm05) reported technical statistical error in mann, bradley & hughes 1998 (mbh98) method, said produce hockey stick shapes random data. claim given widespread publicity , political spin. scientists found issues raised mcintyre , mckitrick minor , did not affect main conclusions of mbh98 or mann, bradley & hughes 1999. mann himself had stopped using criticised statistical method in 2001, when changed on regem climate field reconstruction method. balance dense networks of tree-ring proxies against sparse proxy temperature records such lake sediments, ice cores or corals, mbh 1998 (and 1999) used principal component analysis (pca) find leading patterns of variation (pc1, pc2, pc3 etc.), objective method establish how many significant principal components should kept patterns put characterized original dataset. mcintyre , mckitrick highlighted effect of centering on 1902–1980 period rather whole 1400–1980 period have changed order of principal components warming pattern of high altitude tree ring data demoted pc1 pc4, instead of recalculating objective selection rule increased number of significant pcs 2 five, kept pc1 , pc2. removed significant 20th century warming pattern of pc4, discarding data produced hockey stick shape, subsequent investigation showed hockey stick shape remained correct selection rule.

the mm05 paper claimed 1902–1980 centering produce hockey stick shapes persistent red noise , methods exaggerated effect. tests of mbh98 methodology on pseudoproxies formed noise varying red noise white noise found effect caused small differences within uncertainty range , had no significance final reconstruction. mcintyre , mckitrick s code selected 100 simulations highest hockey stick index 10,000 simulations had carried out, , illustrations taken pre-selected 1%.

on 23 june 2005, rep. joe barton, chairman of house committee on energy , commerce wrote joint letters ed whitfield, chairman of subcommittee on oversight , investigations, referring publicity , demanding full records on climate research, personal information finances , careers, 3 scientists mann, bradley , hughes. sherwood boehlert, chairman of house science committee, told fellow republican joe barton misguided , illegitimate investigation apparently aimed @ intimidating scientists. u.s. national academy of sciences (nas) president ralph j. cicerone proposed nas should appoint independent panel investigate. barton dismissed offer, following boehlert s november 2005 request, national academy of science arranged national research council set special committee chaired gerald north, investigate , report.

the north report went through rigorous review process, , published on 22 june 2006. concluded high level of confidence global mean surface temperature higher during last few decades of 20th century during comparable period during preceding 4 centuries , justified consistent evidence wide variety of geographically diverse proxies, less confidence can placed in large-scale surface temperature reconstructions period 900 1600 . broadly agreed basic findings of original mbh studies had subsequently been supported other reconstructions , proxy records, while emphasising uncertainties on earlier periods. contested principal component analysis methodology had small tendency bias results not recommended, had little influence on final reconstructions, , other methods produced similar results.

barton s staffer contacted statistician edward wegman produced wegman report graduate student yasmin h. said, , statistician david w. scott, statisticians no expertise in climatology or other physical sciences. wegman report announced on 14 july 2006 in wall street journal, , discussed @ hearings of united states house energy subcommittee on oversight , investigations on 19 july 2006, , 27 july 2006. report not peer reviewed. reiterated mcintyre , mckitrick s claims on statistical failings in mbh studies, did not quantify whether correcting these points had significant effect. included social network analysis allege lack of independent peer review of mann s work: analysis has been discredited expert opinion , found have issues of plagiarism.