|
|
Post by marchesarosa on Jul 7, 2011 15:10:27 GMT 1
See, Louise, extreme weather is normal!
|
|
|
|
Post by speakertoanimals on Jul 7, 2011 16:06:54 GMT 1
That's a daft statement! By definition, extreme weather is out of the ordinary, out in the tail of the distribution.
The relevant question then isn't whether (weather?) or not it happens at all, but whether or not you can say anything meaningful about the frequency of such events.
Are extreme weather events becoming MORE frequenct, or becoming more extreme, that's the question. And a simple list of past and present 'extreme' events adds nothing to answering that, just becomes a list of menaingless anecdotes. Keeps M exercising her cut&paste keys though, as of they didn't already get enough exercise...................
|
|
|
|
Post by marchesarosa on Jul 7, 2011 16:23:55 GMT 1
On the contrary, extreme weather is part of the definition of the normal distribution.
There is no way of knowing whether extreme weather events are becoming more frequent because we do not have the necessary length of records to establish such a claim.
|
|
|
|
Post by speakertoanimals on Jul 7, 2011 16:49:33 GMT 1
Seems you don't understand the normal distribution either -- I was referring to the 'tail' of such a distribution, as the place where extreme weather lives.
Ah, another vain hope based on ignorance of statistics! That is what statistics can tell us, WHETHER or not what we are seeing is:
1) Just an unlikely accident, sometimes even rare events happen more frequently that we naively might expect!
2) An indication that actually the whole distribution has shifted.
3) We don't have enough data to say either way
Why do you ASSUME that the answer is the one you would like, then continue as if it had been decided?
|
|
|
|
Post by marchesarosa on Jul 7, 2011 18:08:58 GMT 1
Weather is by definition prone to extremes and the "normal distribution" applies to weather as much as to any other variable.
Where have you been living?
The validity of statistics depends upon the existence of valid and comprehensive data to measure.
If you think the range of weather variables that it has been possible to estimate in the last few decades amounts to enough data to make sweeping "global" historical statements, poor old you!
|
|
|
|
Post by marchesarosa on Jul 7, 2011 18:20:45 GMT 1
Xie et al. collected hail size information from stations throughout China and at the end of their analyses, they state “Here, we found no significant long-term trend in hail size based on the proportion of severe hail indirectly in the four regions in China, suggesting that hail size, as an important aspect of hail climatology, may not be sensitive to the intrinsic natural variability or climate change in the last 2–3 decades.” Oops, another failure to link extreme precipitation to global warming.www.worldclimatereport.com/index.php/2011/07/01/extreme-precipitation-update/#more-496------- Is hail size an index of extreme weather, STA? If so, no measurable change in China, you'll be relieved to hear!
|
|
|
|
Post by marchesarosa on Jul 7, 2011 18:26:12 GMT 1
A team of German and French scientists decided to have a look at the situation in Europe, and they begin their article curiously stating
“Assumptions about an increase in extreme flood events due to an intensified hydrological cycle caused by global warming are still under discussion and must be better verified. Statistical flood analyses indicate that simple mechanistic explanations are not straightforward and that complex seasonal and regional effects have to be considered. Some historical flood records even indicate that flood frequencies were higher during colder periods, challenging the hypothesis of a correlation between the frequency of extreme floods and a warmer climate.” We knew we were going to like this one!www.worldclimatereport.com/index.php/2011/07/01/extreme-precipitation-update/#more-496
|
|
|
|
Post by speakertoanimals on Jul 7, 2011 19:01:30 GMT 1
I see that you really don't understand the normal distribution.
The normal distribution is one PARTICULAR continuous distribution, that may or may not apply to ACTUAL results in some area or other. It's useful when it comes to reading errors etc because of some sexy piece of maths (the central limit theorem), but it doesn't necessarily apply to weather per se.
The 'normal' in 'normal distribution' shouldn't be confused with 'normal' when you mean 'usual'.
So, whatever the distirbution IS for whatever weather-related thing you happen to be measuring (which may be continuous, like average weekly rainfall, or average temperature, or maximum temperature -- or may be discrete, like the number of hurricanes in one year), then there will usually be extremes. If the distribution shifts, then we might expect the frequency of various 'extreme' events to change. If you're really unlucky, what was formerly 'extreme' may become more normal (in the everyday sense). Or you may see even more extreme events than you saw initially.
Nope, because any idiot can see that if you're looking at hail (which is a good choice, because hail is a fairly extreme and un likely event, unlike rain), and asking, is hail more extreme, there are two obvious things to look at that MIGHT vary. So, size of hailstones is one, frequency of hail storms is another. They just found that in the case they considered, the size didn't seem to be sensitive, that's all. And it need not be, as far as I can see -- climate changing so that the conditions that breed hail storms become MORE common doesn't nexcessarily mean that the conditions within each hailstorm change in such a way that the hail itself becomes bigger. In short, the maximum size of hail may be determined by other factors which aren't sensitive to climate change.
Well, you're in no position whatsoever to make ANY judgement on that, given that you don't even know enough stats to know what the normal distribution IS.
Plus you've made a silly error. The point about stats is that you can do statistics on whatever data you do have, and part of that statistical analysis is telling you how likely, or unlikely whatever predictions you make are, or how sure you can be about hypotheses you test, based on how much data you had in the first place!
In basic terms, if I'm looking at possible deviations from some mean, the AMOUNT of data I start with determines the ERROR on my estimate of the mean, and also determines the ERROR on my estimate of the natural variability. So, the correction, if you like, for a limited amount of initial data, is automatically included in the statistical analysis.
Indeed, without that, there'd be no bloody point DOING stats.
So, suppose I had ONE sample height (2m, say). Okay, my sample mean is 2m, but my error on that mean is formally infinite -- I can say NOTHING about whether my next measurement indicates a deviation fom the former mean. but if I had TWO measurements of height, then I get at least a finite (but still very large) error on my mean.
If I have thousands of measurements, then the error on the mean and my estimate of the natural variability becomes better and better. In shortm, what stats says is the MORE data, the better your ability to detect deviations, and stats tells you EXACTLY how sure you can be that you have detected a deviation.
So, take home message -- rely on the stats, because intuition is usually TOTALLY wrong. Hence your opinion is worthless twice over (at least) -- and I have a decent amount of data on the past validity of your opinions to back me up on that!
|
|
|
|
Post by speakertoanimals on Jul 7, 2011 19:10:59 GMT 1
This is actually fairly trivial to get your head round.
1) Extreme weather events have always occured
2) climate change may mean more extreme extreme events, or just more events of the same extremity, or something else. Either way, we are looking for a CHANGE in the pattern.
3)Changes in patterns -- we need a description of the past 'pattern', and stats to tell us whether or not what we think we have seen as a change is actually statistically significant or not.
4) Anecdotal evidence along the lines of -- oooh, its bloody wet/cold/warm/snowy/ice-free is totally useless
5) Non-statistical estimates of the likelihood or meaning of supposedly extreme events is totally useless. We are totally CRAP when it comes to that sort of risk-assessment or assessment of significance. We are a species that can see meaning in the fact that our car registration ends with a seven, or believes that using our lucky numbers in the lottery will gives us a better chance of winning. As such, we are totally useless at assessing the significance of rare events that either effect us, or luckily miss us.
6) As a consequence of 5), non-specialists opinion of matters that are actually statistical in nature can (probably :-) ) be dismissed as totally worthless.
|
|
|
|
Post by marchesarosa on Jul 7, 2011 19:23:33 GMT 1
...we’ll leave you with this incredible one sentence summary of an article by Laurens Bouwer in the Bulletin of the American Meteorological Society – a statement that still has us shaking our heads: “Lacking significant impact from anthropogenic warming so far, the best way to assess the potential influence of climate change on disaster losses may be to analyze future projections rather than historical data.” So we should disregard the lessons from past data on hail in China, extreme precipitation events in Hawaii, floods in southern Germany, or the historical occurences of other extreme weather, and instead turn to climate model projections for guidance? We can only hope that he is kidding. www.worldclimatereport.com/index.php/2011/07/01/extreme-precipitation-update/#more-496---------- Why does that statement remind me of STA prating that estimating ocean volume is more useful than knowing changes in ACTUAL relative sea level! |
|
|
|
Post by speakertoanimals on Jul 7, 2011 20:08:28 GMT 1
Don't be daft. It's quite obvious that what he meant was:
1) Our climate predictions have to be based on past data, and the best science. Else they'd be useless. hence the chinese hail, etc etc
2) We DON'T have sufficient historical data on the sort of disaster losses that we MAY encounter due to climate change. We are talking about the HUMAN aspect here, which is different now to in the past. No more odd than the fact that native americans living in tents made of buffalo skin would have a different 'disaster loss' in a major earthquake than we would see if the same size major earthquake hit California today. We'd have to project what might happen if that big an earthquake hit, because we don't have enough real data on the effect of such an earthquake on a modern city. Just as (unfortunately) no one managed to predict in detail what a big tsunami hitting Japan would do to nuclear power stations, based on historical records of japanese tsunamis.
Is that clear? The records tell us how big a tsunami might be, the climate records can help us understand how the climate might change, but we can't use the historical records to predict what might be the detailed effects of such an event (although I will admit that the historical records ought to have told the engineers that the seawall should have been higher!).
Do you have a problem with basic english comprehension, as well as basic maths and stats.........................
|
|
|
|
Post by principled on Jul 7, 2011 20:20:49 GMT 1
STA This is actually fairly trivial to get your head round. 1) Extreme weather events have always occurred, our problem is that beyond couple hundred years we have no exact evidence about the extent of these events (see 4) 2) Climate change may mean more extreme extreme events, or just more events of the same extremity, or something else. Either way, we are looking for a CHANGE in the pattern, and to do this we would like to have information going back more than a couple hundred years. (see 4) 3)Changes in patterns -- we need a description of the past 'pattern', and stats to tell us whether or not what we think we have seen as a change is actually statistically significant or not. Our problem comes when we try to get widespread data more than a couple hundred years old. (See 4) 4) We have much anecdotal evidence along the lines of -- oooh, its bloody wet/cold/warm/snowy/ice-free , but we can't use this statistically. But it can be a useful adjunct to our estimation of past events where data are sparse. 5) Non-statistical estimates of the likelihood or meaning of supposedly extreme events is totally useless. We are totally CRAP when it comes to that sort of risk-assessment or assessment of significance. So, once we get the climate models to mirror real climate, we'll be able to ditch Fred and his bottle of shark oil. 6)Once 5) is accomplished, the non-specialists opinion of matters that are actually statistical in nature will be able to be dismissed as totally worthless. In the meantime,... See STA. It's easy. P  |
|
|
|
Post by marchesarosa on Jul 17, 2011 9:11:25 GMT 1
Posted on July 16, 2011 by News Staff From NTD Television english.ntdtv.com/ntdtv_en/news_china/2011-07-15/snow-covers-sichuan-state-highway-318-in-summer.htmlSnow Covers Sichuan State Highway 318 in SummerWhile people in other regions of China, are enduring the hot summer, a bizarre scene of drifting snow covered the Zheduo Mountain section of the Sichuan State Highway 318 on Wednesday. The snowfall measured more than 12 inches deep. The temperature was less than 32 degrees Fahrenheit. Heavy snow and fog almost paralyzed traffic on the state highway—as vehicles moved bumper to bumper on the snow-covered road. While the snowfall was creating a traffic jam, some people enjoyed the unusual sight, and got out of their cars to take pictures. [Tourist]: "I came from Nanjing City of Jiangsu Province. I've come to Sichuan for a tour. I never saw snow in July. It is indeed rare. Even in winter, I rarely saw snow in Nanjing." Meteorologists offer two reasons for the sudden snowfall in summer. The first is the altitude of Zheduo Mountain, which is more than 13,000 feet high. The second reason is the continuous rain with low temperatures in the region. See also wattsupwiththat.com/2011/07/16/curious-summer-weather-in-china/ |
|
|
|
Post by marchesarosa on Jul 17, 2011 9:22:02 GMT 1
TrueNorthist says
In BC — the southwest coast of Canada that is — we have had the coldest, wettest, spring/summer ever(?). You wouldn’t know it from the press though. Ask any “ordinary” person and they will tell you it has been very, very cold. Environmental Canada on the other hand — the official source of all things weather, says it is reallllllllllly waaaaaaaaaaay hotter than ever before. My sources in EC tell me they know that it is colder than anything seen for 100 years... There is a heavy snow-pack above 5000 feet, and it is dreary and cold all around. I understand that LA is suffering much the same fate. Cool and wet.
|
|
|
|
Post by marchesarosa on Jul 17, 2011 9:29:10 GMT 1
Anticlimactic says
On the BBC weather forecast for this weekend [From the UK Met Office] the presenter said “Think ‘Autumn’ “!
-------
I'm beginning to think I was right in enjoying the mini heatwave in April. Looks like that was the best of the British "summer".
Isn't it strange how the Met Office meme of "barbecue summer" now seems so passé?
|
|
