The weather stations are key. This is what always bothered me even back when I sort of believed some of the warming scare hype back in the 90's. If you keep adding new weather stations, you are adding new data points into the data set! For example, if you have 100 stations in the US that were all created in 1800, you have 100 data points that span the last 220 years. Now you add 200 more over time, some in the 1900s, 1920s, 1950s, 2010 etc. Those data points only go back for a handful of years but the data is mashed into the overall average. This is bogus! And of course decommissioning the ones in colder areas as you mentioned has another impact.
They think that they can just extrapolate some maths to fill in the gaps and that makes the data perfectly valid and unquestionable. This is what makes the data so easy to manipulate to tell the story you want, there's so much room for tweaking the variables and weights in the extrapolation models. And that's exactly what they got caught doing in the ClimateGate emails with the "Hide the Decline" shenanigans. They start with an answer and fit the models to match, which is literally the complete opposite of how science is supposed to be done.
The weather stations are key. This is what always bothered me even back when I sort of believed some of the warming scare hype back in the 90's. If you keep adding new weather stations, you are adding new data points into the data set! For example, if you have 100 stations in the US that were all created in 1800, you have 100 data points that span the last 220 years. Now you add 200 more over time, some in the 1900s, 1920s, 1950s, 2010 etc. Those data points only go back for a handful of years but the data is mashed into the overall average. This is bogus! And of course decommissioning the ones in colder areas as you mentioned has another impact.
They think that they can just extrapolate some maths to fill in the gaps and that makes the data perfectly valid and unquestionable. This is what makes the data so easy to manipulate to tell the story you want, there's so much room for tweaking the variables and weights in the extrapolation models. And that's exactly what they got caught doing in the ClimateGate emails with the "Hide the Decline" shenanigans. They start with an answer and fit the models to match, which is literally the complete opposite of how science is supposed to be done.