3 ms·
Well I guess we would want to know the standard deviation of these numbers (for each month), and therefore work out how often we would expect to see such a 1.35
by mathgenius 11y ago
Well I guess we would want to know the standard deviation of these numbers (for each month), and therefore work out how often we would expect to see such a 1.35C difference from the mean in febuary. Is it 1 in 100 years? or 1 in 100000? or what?
- ekianjo 11y agoYeah, and the problem is that our database for reliable temperature measurements is counted in decades and not even hundred of years. So hardly telling.
- mathgenius 11y agoA few tens of febuaries is enough to get an estimate of the standard deviation, no? And also, don't we have ice core samples going back thousands of years? This should indicate more about the occurrence of large scale variations in temperature. I'm not going to argue minutea because I don't have a spare life that I can devote to climate science. I just assume they are getting it right.
- ekianjo 11y ago> A few tens of febuaries is enough to get an estimate of the standard deviation, no? Yes, but even if this point is a significant outlier, there is no telling that it's not random variation (p-values only talk about probabilities, they never prove anything not imply causal relationships) and on top of that you don't have a model to predict the future. Forecasts out of existing data values/limits are usually extremely unreliable.