Talk:Extreme weather

From Wikipedia, the free encyclopedia

	This article related to meteorology and/or specific weather events is part of WikiProject Meteorology and Weather Events, an attempt to standardize and improve all articles related to weather or meteorology. You can help! Visit the project page or discuss an article at its talk page.
Stub	This article has been rated as Stub-Class on the assessment scale.
Mid	This article has been rated as Mid-importance within WikiProject Meteorology.

[edit] 2004 Nova Scotia blizzard deaths?

Where did the figures of 9 and 100 dead in the two 2004 Nova Scotia blizzards come from? I don't remember hearing about any deaths, let alone 100.

[edit] Extrapolation graph

The cost of extreme weather is rising very rapidly. source data: IPCC, 2001. Some of the cost increase is due to added exposure such as building on the coast, and some of it is due to global warming.

I am replacing this graph because it is directly pertinent to this page, because it was adjusted from an earlier version based on comments at Talk:Global warming, and because it very precisely predicted the preliminary 2005 results from the 1950-1998 IPCC data. Click on the graphic for more information. —James S. 19:34, 7 January 2006 (UTC)

"very precisely" is false - something is either precise, or it is not, and your data falls within a very, very wide range you've allowed, which is designed specifically for imprecision. furthermore, the 2005 information is informal, and cannot be considered valid for a graph that purports to be based upon data. it is irrelevant to an article about extreme weather. extreme weather is weather that has already occurred, and for which we have data, and examples. your 'predictions' are not informational. the 'window' you've given yourself allows for you to be "right" given just about any future costs. it's little different from guesses. this is an encyclopedia, not a place where we speculate about a future that has not arrived. as well, the costs associated with extreme weather are far more a factor of population and economic growth, rather than on the weather itself. if a category three hurricane hits a deserted island, no costs are incurred, even though the hurricane's power is 'extreme'. if the same island later becomes populated with a million people and a thousand five star hotels, and is again hit with a category three hurricane, then the costs in human lives and property will be astronomical - but not owing to any greater power in the storm itself. furthermore, according to wikipedia guidelines regarding original research (WP:NOR, and more specifically Wikipedia:No_original_research#What_counts_as_a_reputable_publication.3F), this graph is inappropriate content to be added. on this, and many other bases, i am removing the graph. Anastrophe 19:57, 7 January 2006 (UTC)

I mean by "very precisely" that the predicted $200 billion is very close to the prelimiary 2005 EDS estimate, especially given the width of the confidence interval to which you are apparently referring. Extrapolation is simply a way of displaying the underlying data, which in this case was published in 2001 by the IPCC. Your scare-quoted complaints about the duration of extrapolation are addressed by the presence of the confidence interval. —James S. 20:11, 7 January 2006 (UTC)

why must we continue rehashing this debate? your graph is original research. your confidence interval means you'll be "right" given almost any scenario. "very precisely" and "very close" constitutes an oxymoron - hardly scientific. the graph is 'junk science', pure and simple. the costs are primarily sociological in origin; your graph is designed to suggest that the costs are due to weather becoming more extreme. why must we do this over and over? your graph is original research, and definitely falls within those bounds based on the 'reputable publication' criterion. there is only one person on wikipedia who thinks this graph has any probative value at all. take your graph to a peer reviewed journal, or to the IPCC, or whomever, submit your 'findings' to them, once you manage to get it _genuinely_ published, not vanity published, you might find receptivity to it increase from the current nil to slight. Anastrophe 20:23, 7 January 2006 (UTC)

I am not referring to the precision of the confidence interval, I am referring to the precision of the mid-interval prediction (the solid black line) in its accuracy of predicting 2005 preliminary estimates from the 1950-1998 data. Your assertion that "costs are primarily sociological in origin" is not supported by any sources; it is contradicted by the Association of British Insurers, which claims 20% [1] and by the analysis of Choi, O. and A. Fisher, "The Impacts of Socioeconomic Development and Climate Change on Severe Weather Catastrophe Losses: Mid-Atlantic Region (MAR) and the U.S." Clim. Change 58, 149 (2003), and by the Wuebbles and Hayhoe, "Climate Change Projections for the United States Midwest," in Mitigation and Adaptation Strategies for Global Change, 9, 335 (2004.)

Do you have any sources at all supporting your repetitive assertions? —James S. 20:41, 7 January 2006 (UTC)

your sources above do not appear to present your graph. the onus is upon you to show that your graph is not original research; you have provided no citations showing your research has been presented in a peer-reviewed journal of any kind. the originator of a graph that has been charged to be original research is *not* the person who makes the decision as to whether it is or is not original research, since there is a conflict of interest at work. that's why a peer-review is required for such research - it eliminates that conflict of interest (your peers determine whether it is valid, not you yourself). provide a pointer to a reputable peer-reviewed journal that has published your graph, and your analyses therein. until then, your graph remains original research. to repeat from the WP:NOR policy: "Reputable publications include peer-reviewed journals, books published by a known academic publishing house or university press, and divisions of a general publisher which have a good reputation for scholarly publications." your graph does not appear in any such publication, unless i'm mistaken. please stop posting your original research to wikipedia until it has been peer-reviewed per the qualifications listed above. i'm removing the graph. this will constitute my third reversion, so please do not post the graph again. Anastrophe 21:06, 7 January 2006 (UTC)

Display of an extrapolation trend, as revised after review, and reflecting the best known fit to the underlying data is not considered original research by the scientific community. However, it is a valid way of displaying the underlying data published by reputable sources. Of all the versions of the graph created in 2005, the graph included and shown above was by far the best at predicting the preliminary 2005 data which only because available on 2 January 2006. I believe that the graph is valuable, accurate, correctly reflects the confidence of the fit, and carries an important message. —James S. 02:25, 8 January 2006 (UTC)

(back left)"as revised after review" - please provide a link or reference to the organization(s) that has/have reviewed your work. "carries an important message" - this suggests that you believe the prediction tells us something significant. what is that message, specifically? you are suggesting, in the legend to the graph, that some of the predicted increase in cost is due to global warming. the problem is that that's an extremely vague assertion. if you are charting trends in costs, then you should be able to quantify the sources of those increased costs specifically, rather than vaguely. this points out the flaw in your argument - you believe that it is global warming that is causing the increase in costs, but you don't have evidence of that. if the increase in costs of extreme weather were the result of global warming, then the future trends would tend to ramp up at approximately the same rates, but unless i'm mistaken, they do not. please overlay your graph with this: .

you'll see that your predictions are completely out of bounds with predictions for increases in global average temperature. if you review the changes in global climate over the last 100 years, and compare them to the changes in costs of extreme weather over the last 100 years, you'll see there's no correlation either. between about 1910 and 1945, the global average temperature rose a bit more than .4 degrees celsius. between about 1975 and 2005, the global average temperature rose about .4 degrees celsius. if global warming were the cause of the increased costs, then there should have been similar increases in cost between 1910 and 1945, and 1975 and 2005. global warming *may* have contributed in some small - and non-quantifiable - measure, but the majority of the cost increase was cultural, and sociological. in short, you are making a prediction - and claiming contributing factors - which are unsupported by the data. your predictions and conclusions aren't supported by the existing data; your research is speculative, rather than informative; it is original research, as your opinion of the probative value is not supported by any peer-review. in short, your graphs do not belong in an encyclopedia. or in wikipedia, for that matter. Anastrophe 06:06, 8 January 2006 (UTC)

The original graph was reviewed extensively in Talk:Global warming, and the suggestions there were incorporated in December 2005, using the same 1950-1998 data to produce the revision which predicted the 2005 preliminary figure published in January 2006. The important message is that the financial costs are increasing much more steeply than most people imagine. There is no question that global warming increases evaporation and transpiration, leading to increased precipitation. Global warming also directly increases average wind speed through more rapid daytime thermal expansion of air and the resulting increased barometric pressure gradients. Those factors both contribute to stronger storms, as explained in the peer-reviewed citations above. The idea that "most" of the increased costs are sociological (such as more building on the coast) is contradicted by the Association of British Insurers, who say that 80% of the costs can be eliminated by 2080 by reducing greenhouse gas emissions. The Choi & Fisher and Wuebbles & Hayhoe papers both also suggest that the increase is due more to increased storm strength than societal factors. The reason that storm strength increases aren't expected to rise in parallel with temperature is due to at least two factors: first, there are positive feedback effects (water vapor in cloud and transparent form is a greenhouse gas) and because increases in sea level from polar ice cap melting make an otherwise-equivalent coastal storm more costly. I appreciate the opportunity to respond to your concerns but when you assert that the conclusions as shown in the extrapolation are unfounded, I believe you are mistaken. I can not claim that the .jpg costs graph above is "original" because it is a product of others' review, and it is not "research" because it merely reflects the best possible fit (in the correct, nonnegative domain) of the underlying published data. The graph is important and should be included. —James S. 06:55, 8 January 2006 (UTC)

(back left)I'm afraid i'm going to have to take this point by point, which is not well handled by the wiki format. your text will be in italics, mine, not.

• The original graph was reviewed extensively in Talk:Global warming, and the suggestions there were incorporated in December 2005, using the same 1950-1998 data to produce the revision which predicted the 2005 preliminary figure published in January 2006.

you seem to be suggesting that comments and discussion on wikipedia's discussion pages constitute "peer review". that is not the case. please see WP:NOR.

• The important message is that the financial costs are increasing much more steeply than most people imagine.

how have you determined what most people imagine? do you have some poll results you can cite? the message of increasing costs is more than adequately demonstrated by the IPCC graph, [[2]], which shows the actual costs and their increasing nature.

• There is no question that global warming increases evaporation and transpiration, leading to increased precipitation. Global warming also directly increases average wind speed through more rapid daytime thermal expansion of air and the resulting increased barometric pressure gradients. Those factors both contribute to stronger storms, as explained in the peer-reviewed citations above.

none of this is in dispute. unfortunately, the chart of cost increases, compared with the increase in global temperatures over the last 100 years, does not support your conjecture. again - .4 degree C rise from about 1910 to 1945. another .4 degree C rise from about 1975 to present. if global warming were the cause of the increased costs, then if one graphs the costs over the last 100 years, one should see rises of nearly identical proportion during those intervals, with an approximately flat interval from 1945 to 1975, when global temperatures did not rise (in fact, showing a small average drop). it should be trivially easy to obtain this cost data for the interval prior to the IPCC graph.

• The idea that "most" of the increased costs are sociological (such as more building on the coast) is contradicted by the Association of British Insurers, who say that 80% of the costs can be eliminated by 2080 by reducing greenhouse gas emissions.

this is a nonsensical comment. the last time i checked, it was 2006, not 2080. with all due deference to the association of british insurers, i don't believe they are experts in climatology. i also strongly dispute predictions so far into the future, on the simple basis that they are fiction - pleasant fictions to entertain perhaps, but fictions nonetheless. there are no time machines. too little is understood about global warming to suggest what the climate will be like in 2080. what i am talking about, which should be patently obvious considering my rejection of your predictions and most others, is the current data that we already have. the interval covered by the IPCC graph is also an interval of unprecedented economic development in this country. The US real GDP in 1950 was $1.7 trillion (year 2000 basis). in 2004, it was $10.7 trillion. the population was 151 million in 1950, 293 million in 2004 [3]. The implications of that should be about as subtle as a tree falling on you. the increase in cost from extreme weather is due to massively expanding wealth of our country over that interval. many other countries have experienced similar expansions.

• I can not claim that the .jpg costs graph above is "original" because it is a product of others' review, and it is not "research" because it merely reflects the best possible fit (in the correct, nonnegative domain) of the underlying published data.

it is original because it has not been peer reviewed. review by other wikipedia editors does not constitute peer review. this should not need be repeated again and again. your graph extrapolates beyond the published data, it does not restrict itself only to the published data. as such, it is original research. another issue worth noting - since the preliminary cost estimates came out recently, you've been claiming that your graph has been validated because it 'correctly' predicted the increased costs suggested by that preliminary data. here's the problem: please read the IPCC graph again. please review the data from which you generated your graph. then please review bullet point number two of this very article. the error should be obvious. the fact that you've generated further graphs from a wholey incorrect datapoint - if not embarrassing - should at least get you to stop and review your methods, and to ask yourself how careful you are being in generating these graphs. the IPCC graph is entitled Global Costs of Extreme Weather Events. the preliminary cost estimates you have cited, and incorporated into your graph, are US only! of course, i fully expect you to rationalize the mistake, since it 'falls within the confidence intervals', even though it's the wrong data. i've devoted too much time to this already. the conclusions you draw from your graphs are deeply flawed, and it is your conclusions that are primarily in dispute. but, i give up. others can try to talk sense on this matter. Anastrophe 09:03, 8 January 2006 (UTC)

one final note. as i pointed out above, the increase in GDP since 1950 has been nearly ten-fold. then take note of the comments on [this] graph, particularly "Yearly economic losses from large events increased 10.3-fold from US$4 billion in the 1950s to US$40 billion per year in the 1990s (all in 1999 US$).". how much clearer can it be? it's not the weather (or at least, it's not much). it's the increase in wealth. Anastrophe 09:09, 8 January 2006 (UTC)

[edit] What is the best way to display the trend information?

If not a graph, then what? Numbers? A paragraph word problem? --James S. 00:55, 1 March 2006 (UTC)

The problem is you are making strong predictions (unmotivated by outside sources) and obscuring the underlying data. If you stuck closer to the data, presented a graph like this, and described it fairly (e.g. noted that some of the change is related to non-environmental factors like income and population growth) then it should be fine. Dragons flight 01:08, 1 March 2006 (UTC)

Well, you found that after I'd been looking for months. Thank you. And you know your graphs are much more attractive than mine, because I'm an audio stats guy, and my graphic design training was mostly along the lines of developing film from a phototypesetter. So, please, boldly fix the graph you think should be, with as much or as little information about the trend and its confidence interval as you feel is appropriate. --James S. 18:22, 4 March 2006 (UTC)

Also, I forgot: I am sorry I accused you of bad faith regarding the orientation of the x-axis from left to right. At the time, I was being accused of POV-pushing and I couldn't believe that people actually held the other point of view after the amount of government effort being paid to push it had been disclosed. Please accept my apology. --James S. 18:26, 4 March 2006 (UTC)

[edit] WMC removed reference formatting!

What the fuq!?!? Can you explain the revert WMC? J. D. Redding

You're not the only pereson who got in trouble with the arbcomm: Wikipedia:Requests_for_arbitration/Climate_change_dispute_2 William M. Connolley 11:56, 6 March 2006 (UTC)

Not a very subtle personal swipe ... does nothing for your actions though ... the "Choosing citation format" back up the full citation (eg., comprehensive reference information) more than the inline versionj you revereted to. Please explain more fully ... otherwise the fuller source info shuold be put back in. J. D. Redding 12:14, 6 March 2006 (UTC) (PS., listing the article and other info help when the lik becomes defunct .. which happens quite often over time ...)

If you want to repeat the entirety of the arguments in the arbcomm case, I suggest you do it in some corner of your user pages where it won't waste everyone elses time William M. Connolley 14:07, 6 March 2006 (UTC)

A good reference section w/alphabetized listing by author or web handle, if no author is given, would be a good addition. Then perhaps a Harvard style notation (standard in most science pubs) could take the place of the number in the direct inline link. This would be the best in my view and avoid the absurd multiple clicking required to check out a linked reference. I am a strong supporter of direct inline links. This is the web, not a book. Vsmith 16:33, 6 March 2006 (UTC)

[edit] Global daily highest temperature

Hi, does anyone know of a source somewhere on the internet that lists extremes for the previous day, I wish to plot the daily highest temperature with google maps but cannot seem to find any source for this information despite hours of googleing. Htaccess 12:27, 18 March 2006 (UTC)

[edit] Strange tornado force-scale recalibration article

Does this article strike anyone else as strange? --James S. 22:26, 31 March 2006 (UTC)

[edit] Tornado force scale adjusted

Missourian News

By KURT AUSTIN, March 31, 2006

Forty-four tornadoes touched down across Missouri on March 11-12, resulting in 10 fatalities and 107 injuries, according to updated information from ground and aerial surveys done by the National Weather Service’s Missouri offices.

Several of the tornadoes in mid-Missouri, including the one that killed four in Renick, were given an F3 rating on the Fujita scale used to estimate wind speeds based on damage. A tornado in Monroe County was rated F4, a classification for tornadoes with winds between an estimated 210 and 261 miles per hour.

At this time next year, tornadoes causing the same amount of damage will be classified as having lower wind speeds based on research showing the existing scale overestimates wind speed.

The old scale, named after famed storm researcher Tetsuya Theodore Fujita and adopted in 1971, was thought to be limited by meteorologists in its ability to draw a correlation between damage and wind speed.

“The new scale refines the wind speed,” said Pat Slattery, spokesman for the National Weather Service’s regional office in Kansas City. “The corrections will make the enhanced scale ratings more accurate and consistent.”

The new scale will be known as the Enhanced Fujita Scale and will go into effect in February.

Anthony Lupo, associate professor of atmospheric science at MU, said the change will not affect the public, but rather will improve the accuracy and consistency of tornado intensities.

“It is a transparent change as far as the public is concerned,” Lupo said. “But they are going to get a much better picture in the end of how severe a tornado may have hit.”

He said the change will make a bigger difference in Missouri and the Midwest than elsewhere because of its proximity to tornado alley.

The new scale, developed by a forum of meteorologists and wind engineers nationwide in cooperation with Texas Tech University’s Wind Science and Engineering Research Center, will continue to rate tornado categories from F0, the weakest, to F5, the strongest. However, the wind speeds are lowered in every category except F0 to be more accurate by taking into account damage indicators — that is, the type of structures damaged — and the degree of damage to those indicators.

For example, the enhanced scale will differentiate between hardwood and softwood trees as well as buildings such as schools, churches or homes. In the assessment of those structures, eight degrees of damage will be used ranging from visible damage to complete destruction. Field tests and training of the process are under way nationwide.

The National Weather Service is the only agency with the authority to provide an official Fujita scale rating. The rating represents the highest estimated wind speed that occurred during the life cycle of the tornado.

This is true. It is actually a quite good new system, which takes into account the different degrees of damage for different structures, leading to a better maximum wind speed estimate. Also, the new scale is based on 3-second wind gust speed. There is more at Fujita scale. Runningonbrains 12:51, 7 June 2006 (UTC)

[edit] Anyone still think I'm cooking numbers?

Just out of curiosity, does anyone still think I spend time fiddling with trend projections in order to come out with particularly freightening results?

I swear, I've never done that. Every time I've ever run the best models available to me on ordinary personal computers, people have accused me of cooking the fit somehow to come up with a worst-case scenario. It happened in 2000, when I did this carbon dioxide trend extrapolation (R²=0.98, logistic sigmoid consistent with resource consumption by growing population, predicting leveling off around 813 ppm about year 2800) and of course it happened 2005-6 here on Wikipedia.

Nobody ever accuses me of trying to come up with a best-case scenario. Maybe in 2010 someone will accuse me of that. --James S. 06:45, 9 April 2006 (UTC)

[edit] EPA etc

I removed the EPA link [4] since when you look closely it doesn't actually support the text. It says that there are more disasters, but it doesn't explicitly link them to GW.

I also toned down the WMO bit - commondreams had got carried away. WMO didn't say quite that.

And the TAR is always good.

William M. Connolley 21:32, 11 April 2006 (UTC)

[edit] ABI

Can't find it at the moment, but this post by RP Jr is worth reading re the ABI report William M. Connolley 21:05, 24 April 2006 (UTC)

[edit] Lead paragraph

The lead section to this article seems very sloppy and may not be NPOV. Would someone care to weigh in, or possibly fix it? Climatology is not my area of expertise. Runningonbrains 12:46, 7 June 2006 (UTC)

[edit] Probative Value of List?

looking at the listing of extreme weather events that has grown in this entry - which now makes up the majority of the entry - it strikes me as exceedingly misleading. aside from a handful of extreme weather events i (and others) took the time to research a year ago and add information for, the the listings for 2005 and 2006 overwhelm the page in the minutia they record. this is a function - i believe - of the ease of recording events as they happen contemporaneously, not of the actual frequency of extreme weather events. for that matter, there's no real evidence to suggest that the majority of these events are in any way associated with extreme weather - they are simply weather events that caused loss of life in one degree or another. by that measure, the flooding of the lower mississippi in 1927, which cost 300,000 lives, was clearly the greatest "extreme weather" event of last century...frankly i think the list should be tossed, as i see no scientific rigor being applied in the determination that X event actually constitutes "extreme weather", according to the definition that opens this entry. Anastrophe 17:39, 22 July 2006 (UTC)

I agree with you to an extent. I was simply adding all of these events here because there's nothing in the way of anything that documents weather events (and I'm pulling most of these from the NOAA Extreme Weather site). If we can come up with what should be included here, then I'd be willing to trim down this list and help with adding what should be included. bob rulz 19:34, 22 July 2006 (UTC)

So is this becoming List of extreme weather events? If so, we can keep severe weather as an article on types of severe weather, the way it is now. -Runningonbrains 16:24, 20 October 2006 (UTC)

I've been thinking of several ways we could change this article, but to move all of the stuff and to write a good article about extreme weather is a monumental task that I'm not up to at the moment, and haven't been for a while. I agree that the article should be split, I just don't know the best way to do it. If someone else wants to Be Bold and do it, then that would help tremendously. I just don't have the time. bob rulz 22:26, 20 October 2006 (UTC)