If you say:
I hypothesize that temperature is changing and I say. "Sources of Error: I cannot measure the arctic reliably."
And then say "It's getting warmer on the long trend, but on the short term, the noise results in no warming trend." You've fulfilled your study.
If you then have a follow up study and you say "I filled the gaps in my data and my hypothesis was further confirmed and even the short term change was warming." You're still following empiricism.
It would be like dropping a ball and a feather and saying "Hmm, my theory that objects would fall at the same rate appears to be wrong. Source of error: Different amounts of air resistance."
Rerun the test: "Eliminating a potential source of error changed my result."
TWiTfan: HE CHANGED THE CONDITIONS TO PROVE HIS THEORY!
Of course you change the conditions. You identify sources of error and then hopefully in the future can eliminate them with closer study for more accurate results. that *IS* empiricism.