Except it's not quite statistically significant. The SkS trend calculator gives a 15-year trend of 0.135 ± 0.217°C/decade (2σ), which is a statistical signfiicance of 79%, less than the canonical (and arbitrary) standard of 95%. The large uncertainty just shows that 15 years is too small of an interval to make statistically meaningful conclusions about trends. If the 15-yr trend uncertainty is 0.217°C/decade (2σ), the trend would need to be 0.213°C/decade for a statistical significance of 95% -- and the UAH record has had a 15-year trend of 0.213°C/decade or greater only 32% of the time. (Yes, I know that the uncertainty varies somewhat with the interval, but this estimate is good enough for blog work.)
The last 5 years are still the warmest 5 years in the UAH LT record:
If, as appears to be the case, that's a trailing 5 year average of monthly results, then the slope over the last 15 years is actually based on 20 years of data, i.e. you haven't actually refuted what deniers of anthropogenic climate change claim about the pause and they would no doubt accuse you of being deliberately deceptive in apparently implying that you have.
ReplyDeleteThe slope I'm quoting is the slope of the monthly data, not the slope of the moving average.
ReplyDeleteI guess this feature in UAH data is causing a migration to use RSS instead. In certain circles.. ;)
ReplyDeleteYes, you're right.... And to be truthful, there is a large discrepency between UAH and RSS data lately, that needs to get resolved before either of them can be taken are good estimates of atmospheric temperatures.
ReplyDelete