All based on believing the chain-of-custody of the data from the sensors to the presented documents have in no possible way been altered. Starting at the reader of the sensors all the way through the charts and graphs producers.
That’s if those same sensors have a valid calibration seals and corresponding calibration certifications that also have a chain-of-custody that proves none of them are falsified or altered. All sensors are within the calibration times prior to their expiration and due date for the next calibration.
REAL science is tough. It requires verifiable proof, not just “because I said so”. Including independent organizations launching surprise inspections of the sensors and documentation.
The sensors placement should not be within 200 meters of any manmade structure or any natural cover.
Really good science requires double or triple blind methods.
Those devising the study don't collect the data, and those collecting the data don't analyze it. This last group if possible doesn't have a clue what the hypothesis is.
One of the most repeatedly verified results of science is that scientists tend to find what they expect to find if such methods aren't used. Even with the most honest intent in the world, they will subconciously manipulate the data. And, of course, when money, careers and prestige are on the line, the most honest intent doesn't always apply.
This is why you don’t put an official NOAA temperature sensor over concrete
Posted on April 21, 2008 by Anthony Watts