HelloZeph.
Nice to see a semi reasonable post in the thread. I brought up the NCDC data because of two vastly different 'papers' using the data, in regards to winter trends. One used a very short time period to show winters getting colder, which I pointed out was too short a period, and the other was a completly dishonest attempt to show drastic warming, by cherry picking the time period. Neither showed the bigger picture, but the "getting colder" one at least featured the NCDC site which provided the data for both 'papers'. The topic hasn't been the same since.
Short form for rj: Even random noise with no real trend or underlying physical process, will allow one to calculate a trendline by regression, and for inappropriately short periods that trend will rarely be zero.
There has been a huge amount of noise in this topic about temperature data, about trends, time periods and more. It was even claimed that you can make the data say anything by the time period you pick for a trend line. I pointed out that if all the data point are going in one direction, that would be impossible.
Just calculating a regression slope does not automatically indicate a real "trend" of the sort that's worth discussing
I think part of the problem is confusing is language, and the mystery of statistics, in which anything is possible, or possible to dismiss. The following data could be illustrative, and educational.
Trend: 0.145 ±0.157 °C/decade (2σ)
β=0.014526 σw=0.0019048 ν=17.050 σc=σw√ν=0.0078651
Does that show global temperatures are warmer?
Trend: 0.134 ±0.089 °C/decade (2σ)
β=0.013430 σw=0.0010786 ν=16.855 σc=σw√ν=0.0044281
Does that show warming?
http://www.skepticalscience.com/trend.php
The first one is 20 years, the second 30 years.
The key is in distinguishing statistical and physical significance. Just using the website to calculate a slope doesn't do that.
Well there is the difference between the imaginary world and the real world. In the real world, temperatures actually mean something, and an increase (or decrease) in temperatures, over time, are evidence. Not just something that you can dismiss as mathematically meaningless.
People are not disputing the numbers (now that you know how to use the website correctly), but the meaning (real world significance) of the calculated numbers.
Well there was the reason I said "semi reasonable" in regard to your opinions. Of course the numbers were disputed. I mentioned winters were getting colder, showed short term and long term trends that showed exactly that, and suddenly it didn't mean anything, or it was cherry picking, or "a state doesn't mean anything", or that nobody could understand what was being said. I posted multiple links to hypothesis trying to explain the colder winters, and it was described as confusing.
That right now the US, the UK, Europe, Mongolia and China are all experiencing record cold, and snow, due to both the extended cold periods, as well as how late it is, makes the issue of winter trends quite relevant.
People tend to be influenced by what happens to them. All the experts in the world telling them it's one of the warmest years ever in history, well, the climate expert is going to lose every time. It's why I joked about the possible reason nobody was bothering to counter the fatalistic discussion in this topic. Because they might be sick of the cold and snow.
I pointed out that March had been unusually cold, record cold in Florida (little did I know at the time that would continue into April), and February data was used to show how wrong my observations were. It looks very much like March will end up being one of the coldest on record.
It's all quite complicated, but a concise explanation of what is considered valid, and what is not, would go a long way towards educating us normal folk about what the experts are trying to say. Or dismiss as meaningless.
Because it seems 30 years of NCDC for Illinois is valid if it show warming, but anything that shows cooling doesn't mean anything.
