Thanks that was useful. I wonder why the BBC picked that 15 year timeframe to put to Jones.
I believe that the question was conjured up by Lubos Motl as the year at which statistical significance (@95% level) safely fails.There is a blog where Motl posted a comment using exactly these words but I can't remember the site. It might have been on Real Climate, but I'm not sure.
Rob,thanks for the video. I have been wondering about this for some time. I am a stats newbie and I have started to use R to plot trends. When I plot HadCRU data from 1995 I get a p value of 5.538e-07. This is statistically significant at the 99% level. Clearly I am doing something wrong!Could you explain, either here or at SkepticalScience where the 92% significance level comes from?
Could somebody please tell me what the statistical significance of global warming is calculated over the last 35 years? 35 years is important because that is when it first appears in the peer reviewed literature.My maths just isnt up to it. But it would I think be a useful rebuttal tool to have this figure.
PS thanks Rob for the Lubos Motl stuff. Have created a page for him on Source Watch.
"When I plot HadCRU data from 1995 I get a p value of 5.538e-07. This is statistically significant at the 99% level. Clearly I am doing something wrong!"Hi Steve,Are you plotting with monthly values? For temperature data, it is really most reasonable to use annual data. Really each datapoint should represent an independent replicate. Time-series data are not technically independent, but I think this is often overlooked. To be honest, I'm not sure how much that really matters for the purpose at hand anyhow. However, what represents a true replicate is indeed very important. Should you look at years, months, days, hours, seconds?! I think that the most appropriate replicate is indeed a year - you capture all seasons in both hemispheres and each replicate represents another trip around the sun. This has much more meaning than months, days, or seconds.Using months might not change your R2 very much (it might), but it will likely greatly influence your P value. With more "samples", you have more confidence. Thus, it's very important what constitutes a sample.Hope that helps.-Aldenp.s. a very accessible and excellent introduction to statistics is "A Primer of Ecological Statistics" by Gotelli and Ellison. It doesn't matter if you're into ecology or not - what's important is that it's readable! (many stats books are not)
"Could somebody please tell me what the statistical significance of global warming is calculated over the last 35 years?"Hi Muchacho,If we use the HadCRUT3v annual dataset from 1975 to 2009 (35 years) we have a statistical significance is less than 0.000000001. I could open up Matlab to get an exact number, but I'm too tired. :)
Pete-On the Skeptical Science blog you posted an article “On Statistical Significance and Confidence” article but it was removed, perhaps because it was considered to be off-topic, which may not be the case on this thread.It was not removed - I'm not sure what you're talking about.By far the greatest contribution to “Statistical Significance and Confidence” is the integrity of the raw data itself.Of course this is clearly important, but it's an entirely different topic to be addressed (which I likely will eventually). The topic that I was addressing here was that many people seem to be perfectly happy to conclude that warming has stopped based on the HadCRU data, even though it is easily shown to be nonsensical and unwarranted. Do I trust the scientific integrity of those trying to convince me that warming has stopped based on the last 15 years of HadCRU data? No.an article by John O’Sullivan “US Government in Massive New Global Warming Scandal – NOAA Disgraced” (Note 1) which reported on significant errors in data collected by the NOAA-16 satellite. On 11th John followed this up with “Official: Satellite Failure Means Decade of Global Warming Data Doubtful”The HadCRU data has nothing to do with satellites, nor do the GISS or the NCDC datasets. These are land-based. Also, it is important to note that the satellite sensors do not measure temperature - they are models of temperature based on microwave radiation from the atmosphere. They add an important contribution to our understanding of the climate system and they agree quite well with land-based datasets. There are two primary models used to produce the tropospheric temperature datasets, one from Remote Sensing Systems (RSS) and the other from the University of Alabama Huntsville (UAH). There is a lot of effort put into making these datasets reliable. Seeing as the articles from climatechangefraud.com don't actually mention anything about how these spurious high temperature values from one satellite affected either of the global temperature datasets (or if they were even incorporated into them), I find it extraordinarily presumptuous to conclude "Decade of Global Warming Data Doubtful." If one is going to make that claim, they have to demonstrate an important effect on actual global temperature values. The last time there was a glitch found in a global temperature dataset (for GISS U.S. values) there were stories of conspiracy(!), fraud(!), etc... What was the effect on global temperature? Nothing.If you're going to post about faulty satellite sensors, but please do so in a manner that informs our understanding of climate science. I'm not dismissing this as unimportant - it might indeed turn out to be - but the posts at climatechangefraud.com seem to be pure sensationalism and tell us nothing about how global temperature data may be affected. Please refrain from posting sensationalist stories and keep it to matters that have an ultimate and obvious bearing on our understanding of the climate system.-Alden
Alden, I’m not surprised that you don’t know what I was talking about. My apologies for the mix up. My misleading comment here arose out of making repeated changes to my original comment in an attempt to get it left on the Skeptical Science blog. I turned to your blog as a last resort and didn’t fully modify my comment for the new blog. It was MY comment that was repeatedly removed from the Skeptical Science thread, not yours. I should have started my fist comment here with:Alden, on the Skeptical Science blog I posted an article “On Statistical Significance and Confidence” article but it was removed, perhaps because it was considered to be off-topic, which may not be the case on this thread. Do you provide a facility for contributors here to revise their comments?Thanks for retaining my comment here because I appreciate your response to it rather than it being simply deleted.In common with others who are sceptical of the validity of The (significant human-made global climate change) Hyothesis (see * below), I have two major concerns about the claims regarding the mean global temperature trend over the past 150 years, first the reliability of the raw measurements themselves, then the effect of the statistical manipulations to which they are subjected. As we all know, raw data can be manipulated by statistical procedures to support different (often contradictory) arguments and politicians are skilled at doing this. The enormous uncertainties about global climate processes and drivers and the possible consequences of their effects make global climate change a perfect tool for politicians and other like-minded individuals and organisations. It is clear to everyone that the global climate change issue is highly politicised hence the arguments put must be treated with great suspicion.You give the impression of being satisfied that the raw temperature data are sufficiently reliable for you to give support to The Hypothesis. If this is the case then can you provide links to the evidence to support this so that I can pass it on to other sceptics?Current uncertainty about the NOAA-16 satellite data has fuelled our scepticism about the surface temperature data, as demonstrated by the speed with which the blogosphere has spread the news. Confirmation of the validity of the raw surface data would counter this.* There are those who ask why I refer to The (…) Hypothesis rather than AGW. The reason is simply that I have no argument against the theory that our use of fossil fuels may cause some global warming. I am sceptical about the significance of any such warming upon global climates. Best regards, Pete Ridley
Pete,I have in fact decided to delete your previous two comments. This is not because I won’t accept comments that challenge global warming, but because I won’t accept comments that point readers to false accusations of climate fraud."Current uncertainty about the NOAA-16 satellite data has fuelled our scepticism about the surface temperature data, as demonstrated by the speed with which the blogosphere has spread the news. Confirmation of the validity of the raw surface data would counter this."I have looked more into the matter and noticed that all of the hype about a few spurious values from the NOAA-16 satellite has to do with the AVHRR sensor, which is used to infer sea surface temperature. This has nothing to do with the MSU sensors that are used to measure global temperature. If you go to the RSS temperature analysis description (http://www.remss.com/msu/msu_data_description.html#rss_msu_data_analysis), you can see that NOAA-16 is not even used. The claim, “decade of global warming data doubtful”, was made without any investigation or research into the matter and is completely misleading. But why wait for “confirmation of the validity” before crying fraud? (see what happened last time: http://data.giss.nasa.gov/gistemp/updates/200708.html)There are many people who appear to be only interested in finding anything that might contradict global warming instead of thoroughly examining the evidence. To me, that’s not skepticism at all – that’s a preconceived conclusion. Like pretty much any other scientist I know, I’m a rather skeptical person (i.e. you have to really convince me). I am convinced that the temperature is rising and that humans are causing it, because the magnitude of the evidence is staggering, the stakes are too high, and the counter tactics are consistently shown to be empty and agenda-driven. There’s always going to be a sensor failing, there’s always going to be an incorrect citation. Science is not a house of cards, it is a jigsaw puzzle with pieces being put in, and sometimes incorrect ones coming out. By now, most of the pieces are in place, they clearly fit together, and the overall picture is unambiguous.-Alden
Alden, thanks for the response and for being prepared to discuss this. I’ll pass your comments on to John O’Sullivan. I don’t think you are correct when saying “There are many people who appear to be only interested in finding anything that might contradict global warming”. The disagreement is about the claims regarding the cause, the extent and the impact of any warming.I also disagree with your “By now, most of the pieces are in place, they clearly fit together, and the overall picture is unambiguous”. If that was the case there would be no need for all of the research that is being undertaken and proposed. Scientific understanding of the processes and drivers of global climates remains poor. Professor Barry Brook (Note 1) is Sir Hubert Wilkins professor of climate change at Adelaide University and scientific advisor to the Australian Government on climate change. Last year he commented that “There are a lot of uncertainties in science, and it is indeed likely that the current consensus on some points of climate science is wrong, or at least sufficiently uncertain that we don’t know anything much useful about processes or drivers” (Note 2 – please read the whole paragraph for the full context). Professor Brook is a staunch supporter of The (significant human-made global climate change) Hypothesis so then goes on to try to imply that 95% of the science is understood. I tried repeatedly to get Professor Brook to clarify his comment about those uncertainties but he never did explain to me what he meant.It is worthwhile listening to what Professor John Beddington had to say in January (Note 3), which included the gross understatement “..that scientists had perhaps not been as good at communicating the value of uncertainty to the general public .. ”. As Boddington said in January (Note 5) “There is a fundamental uncertainty about climate change prediction that can’t be changed”.It is that poor scientific understanding of global climate processes and drivers that forces supporters of The Hypothesis to speculate about future trends and other agenda unrelated to climate change that causes them to speculate about a catastrophic trend. Much of this speculation is agenda driven, especially by politicians, environmentalists, power-hungry individuals and yes, even scientists, who are only human after all.You may be “.. convinced that the temperature is rising and that humans are causing it, because the magnitude of the evidence is staggering .. ” but there are plenty sceptics who disagree with you. On top of that, if recent opinion polls are to be believed, the general public are becoming increasingly sceptical.NOTES:1) see http://www.adelaide.edu.au/directory/barry.brook 2) see http://bravenewclimate.com/2009/04/23/ian-plimer-heaven-and-earth/ 3) see http://www.actoncopenhagen.decc.gov.uk/en/ambition/achievements/february/john-beddington-audio 4) see http://www.timesonline.co.uk/tol/news/environment/article7003622.ece Best regards, Pete Ridley.
Pete,I said that "I am convinced that the temperature is rising and that humans are causing it." Did I say that we know everything about climate and that we are absolutely certain about future climate projections? No. You are putting words in my mouth. You are also using the term "uncertainty" in a very particular way that ignores its true meaning. We are no longer certain that the Greenland ice sheet will remain. We are no longer certain that sea ice will remain. We are no longer certain that the climate we depend on will remain the same. We are no longer certain that the oceans will overturn the way they always do. These are very unsettling uncertainties. A large part of the uncertainties that you are referring to are very troubling indeed - not comforting.I'm quite certain that there's nothing I can say that will convince you though. For instance, it can be readily shown that all the big noise made about the NOAA-16 satellite has absolutely zero impact on our understanding of global temperatures, and yet that doesn't seem to bother you. You don't feel mislead? You were straight-up lied to: "decade of global warming data doubtful" based on completely irrelevant information.False accusations of fraud by people who don't actually bother to do any research into the matter certainly bother me.Please do not leave anymore comments unless it has to do with the topic of the original post.
Currently your Google Reader subscribers are seeing "(title unknown)" to each individual blog post.I highly recommend you provide a blog post title to each blog post. Currently the title appears incorporated in the presentation which is fine once we navigate to a blog post page, but we start by reviewing rows of many blog posts from many bloggers in our Google Reader aggregation window. I also highly recommending getting the print on black background as well. Great presentations by the way, I hope your work gets spread all over.
Alden,You state near the end of your presentation that using short time sets is one of the, "most used tricks in the skeptics' playbook". That comes close to using the term 'skeptic' in the pejorative sense where I believe I've seen you rightly claim skepticism is a feature of good science.If you are not willing or ready to get down in the dirt and use 'denialist' for the frauds who perpetuate such falsehoods I understand. But I suggest either way not insinuating that skeptics are the people who currently dominate misconstruing the science since they are few and far between. Skepticism insinuates a level of informedness and comprehension I have not seen revealed in those challenging the science in the non-peer-reviewed media.
Hi Michael,"Currently your Google Reader subscribers are seeing "(title unknown)" to each individual blog post."I'm aware of this and it has to do with the fact that I've tweaked Blogger quite a bit to make the site look how it does. Seeing as it seems that you have to go to the site anyhow to actually view the presentation, I'm not too concerned about how it looks in a reader - it's just more like a notification. I'll look into it more though.
Hi Michael,Yes I agree that the term 'skeptic' is often not used very well, but I really don't think it muddles the point I'm trying to make. The term 'climate skeptic' has taken on its own meaning beyond the more traditional idea of a skeptical scientist, i.e. not easily convinced.
ah - I believe I fixed the title issue
Hi Alden,Can you put your great work on youtube please?Many forums have embedded players for youtube videos. That would get the matter closer to the audience.Thanks a million.Keep up the good work.Charlie.
Umm very good your graphs just show how installing temperature stations that don't meet the benchmarking criteria set over 100 years ago would show a 0.5c rise.If over-time you decide to put electronics into your weather stations and paint them in white enamel paint its no wonder the temperature would be 0.5c warmer.
"installing temperature stations that don't meet the benchmarking criteria set over 100 years ago would show a 0.5c rise."This post is about the interpretation of the HadCRU temperature data. The conclusion that global warming has stopped is utterly not supported by these data as they are.But obviously the quality of the data is extremely important to the greater question. Those who seem to be convinced that surface temperatures are biased and incorrect never seem to also point out that surface temps and satellite temps give us the same conclusion: substantial recent warming. These are two completely different methods for analyzing the Earth's temperature and they produce statistically indistinguishable temperature trends.And never mind that the vast majority of warming goes to the oceans - which have shown robust warming.I'm hoping to have a post on this topic within the next month.
Very Nice, but you have to be convinced that the data in themselves are correct.....What about selective temperaturedata (closing measurepoints in antarctica, urbanisation around other poits etc?)
"but you have to be convinced that the data in themselves are correct....."Please see my comment above. One could go on endlessly about various potential issues with the surface temperature record without asking whether or not the data are corroborated by other sources. Every other measurement of the Earth's temperature and heat content matches up very well with what the surface station data tells us. We could throw out all of the station temperature data and there would still be unequivocal global warming.-Alden
Thanks for the video. I have been wondering about this for some time. So glad to have gone on this post.
The DATA is significant, but how good is the measurement (i.e., what does the data mean)? This is a more important question.
JonQPublic - see comments above.
Post a Comment