Occam's razor. The "two circle" idea requires some gymnastics, while attributing the discrepancy to an approximation doesn't.
Occam's razor is great if you have no information or are unable to infer more information but in this case we can do both and we should leave occam's razor to apply after we gather all the data and inferences.
There were either one of two style vessels. We know the author refers to a rim and before we are given any data as to size it would be logical to assume it was a rim that was larger than the body. A vessel with a rim that is flared out is likely especially in that it was to be used to bath in. So before even considering the author's neasurements I see a vessel with a rim that is larger than the main body in my minds eye.
The author provides us with the diameter of that rim. With this information I can easily calculate the diameter if I wanted to using some approx of PI.
But I still don't know the height or the circumferance of the vessel. The author then provides us with a measurement that would not be inconsistant with a flared rim.
Lets apply Occam's razor knowing that the logical shape of the vessel is likely a flared rim and the measurements are constiant wiht a flared rim then the simplest explanation is the measurements are the actual (or approximate) diameter of the rim and the actual (or approximate) circumfurance of the body of the vessel. If I apply the measurement as being the diameter of the rim for the one that specifically refers to the rim