Satellites do not measure sea surface or tropospheric air temperatures. They measure some voltage variations from sensors that are affected by various frequencies of electromagnetic radiation and, when integrated with location and orientation data, generate a data product that needs interpretation by experts in a way that may or may not correlate with actual sea surface or tropospheric temperature.

The article is pointing at some inconsistencies in some, but not all, satellite data products. It's not quite saying "if the observations do not fit the theory, the observations are wrong" so much as "if the interpretation of the observations do not fit the theory, the interpretation of the observations may be wrong".

https://skepticalscience.com/Roy_Spencer_arg.htm

Denying global warming at least since 1997, using anything he finds most convenient at the moment.

https://cornwallalliance.org/2009/05/evangelical-declaration...

"But our analysis reveals that several observational datasets — particularly those with the smallest values of ocean surface warming and tropospheric warming — appear to be at odds with other, independently measured complementary variables (e.g. tropical temperature and moisture)"

Which suggests that unless all of the physics is wrong there could be anomalies in either the satellites or the ways its data is processed.