Dr. Jeff Masters' WunderBlog

Quoting LostHobbit:
As a student of mathematics and of meteorology I love reading your blog from time to time. I am not sure as what to make of the Global Warming issue. I have read “An Inconvenient Truth” and use it to point out good and bad statistics to my classes. Personally I think the world is warming, but is it due to man or part of the evolutionary process of our planet, I don’t know. Some of the processes we have documented, like the 88-year sunspot cycles, are long-term and climate change might be even longer. As you begin to talk about the misinformation that is out in the media about global warming, I would love to see you address some basic concerns I see with data collection, data analysis, and to quote a popular movie, “where’s the money?” After all we know very well that liars figure and figures lie all the time. As soon as any issue becomes politicized, the lies increase dramatically on both sides of an issue as both sides cherry pick their data to prove their point (remember the intelligence leading up to the Iraq war on WMD).
First, your graph (from NASA) on 2008 global temperature has at the bottom a graph from 1880 to present (baselined to the 30 year 1950 – 1980 average) that shows a “global” surface temperature with 3 bars of uncertainty (about 1890, 1945, and 2005). These bars are decreasing over time. On the surface this makes perfect sense. A confidence interval’s margin of error (assumed to be at least part of the uncertainty listed in the graph) is made up of three components, a confidence level (unknown as it is not listed with the graph), a variance term (taken directly from the data), and a sample size term. Since the confidence level is not listed, not much can be said about it in any detail; only general comments can be made. IE, the higher the confidence level the broader the margin of error; conversely the lower the confidence level the smaller the margin of error. The lack of transparency on the confidence level is a red-flag to mathematicians that someone might be skewing results – since a 10% confidence level would have a smaller margin of error and look more “scientific” than a 90% confidence level that might be off the graph.
The variance term brings with it a whole set of issues on data collection. First, what is the precision of the measurements? According to mathematics any change that is below the precision of the measuring device is in the noise of the data collection. For example a thermometer that measures temperature to 1° C would have a precision of 0.5° C. Any changes less than that precision in average temperature data is in the noise level of the collection device and therefore unreliable. Surely the precision of our measuring devices have changed over time making our measurements more accurate and this leads to the lessening of uncertainty expressed in the data. But it also makes drawing conclusions on small changes in earlier collected data more problematical. The second area of concern with the variance term is in the location of the measuring devices themselves. If we look at a continuum of data collection points, most of the ones that have a long history associated with them are in the middle of population centers. Over time these population centers have become temperature “hot points” (as any infrared satellite imagery can document). Increases in temperature over time in these “hot points” could make measurements from these “hot points” to be possible outliers in the overall data. Since humans tend not to populate the distant cold lands, the readings, especially the further back we go, from these areas would be underrepresented in the data and could possibly skew our model’s output (global temperature). Since most of the surface of our planet is oceans, we probably have these areas underrepresented in our data, again especially the further back we go as well. All this to say, that from a statistical standpoint, we have a lot of things that contribute to the uncertainty in the data.
Obviously as the number of collection points increase (our sample size increases), the uncertainty in our data should decrease. We need to be careful to avoid convenience sampling techniques as any statistician would tell you that they are unreliable. Without better insight to the distribution of the collection sites, I couldn’t begin to make any comment on whether that would apply here or not.
For those of us without access to the data, we have to trust others to accurately analyze and interpret the data. Try and research the material on the internet and you will be buried in opinion pro and con on any issue – with lots of purported documentation. With so many examples of peoples’ and companies’ lack of integrity in these studies, many times it comes down to who paid for the study to be done and what vested interests did the investigator have in the study. Obviously if several studies concluded that our climate change is due to a natural evolutionary process of our planet and there’s very little we can do about it, then I don’t think these people will get any more money to study the issue (not from a political viewpoint, but a scientific – no need to spend more money on a settled issue). Insight into these area would be helpful.

Quoting presslord:
...but Dolly still did a lot of damage to the "bosom" of the Rio Grands Valley....

Quoting all4hurricanes:
Does the post analysis say anything about any other storms like Gustav or Fay?

Quoting all4hurricanes:
Ok lets have a poll on when the first storm of the season will form who thinks Ana will form
1 in January, February, March, or April.
2 May 1-15
3 May 16-31
4 June 1-15
5 June 16-30
6 July 1-15
7 July 16-31
8 After July
I,m going for #5 for now

Quoting TampaSpin:


Yep ....there should be a front coming thru about next Friday.....

Quoting all4hurricanes:
Ok lets have a poll on when the first storm of the season will form who thinks Ana will form
1 in January, February, March, or April.
2 May 1-15
3 May 16-31
4 June 1-15
5 June 16-30
6 July 1-15
7 July 16-31
8 After July
I,m going for #5 for now

Quoting theshepherd:
664
What I read was a paraphrase of the same old subject that has been beaten black and blue here. Nothing new.
Have I ever read a medical journal article?
Yes,dear. Have you ever given a distemper shot to a half crazed Leopard???

Quoting theshepherd:
mornin' conchy
"Men dominated"?
Of course.
Nothing less could be expected of the Dominant Paradigm.

Quoting NEwxguy:
Ok another artic front coming through this weekend,enough already!!!!
Surfmom your quietly waiting for spring,me on the other hand am not being so quiet.
WHERE THE HECK IS SPRING!!!!!!!

Quoting conchygirl:
Morning All: A beautiful 32 here this morning and more frost - lost a few plants the last few days.

Surfmom: You were missed on Rob's blog last night. Men dominated!

Quoting charlottefl:



Spring.. I'll take summer at this point :)

Quoting Skyepony:
Big chunk of Antarctica's pennesula is about to break off. POES flies over it a few times a day, for another few weeks it can't get a good reading through clouds but that's no reason to toss the data on clear days. Ecosystems are creaping toward the poles with record melts occuring in many places. The argument that the oceans are undermeasured, with satalites, countless buoys & even the cruise ships getting in on collecting data doesn't make much sense. Your not gonna find the data out there in a media article, you don't need to turn to research either. Start with the links Dr Master's left above & begin to learn about the satellites, find their data pages. It is out there & available to look at. Right around the 15th it gets grouped into a nice monthly package that is pretty easy to understand. Looking beyond the data it's obvious our world has gotten a little hotter. It would be irresponsible to throw the bushel out over a few bad apples.

The ocean covers 71 percent of the Earth's surface and contains 97 percent of the planet's water, yet more than 95 percent of the underwater world remains unexplored.