The Kemet equivalent circuit is unusual, I didn't yet see it in other literature. But it's still a simplified circuit and you shouldn't read too much into it. Better measure the real complex impedance curve. And if it's important for you to have an exact model, see which model can be fitted best to the empirical impedance curve.
The fact that ESL varies with frequency doesn't make it impossible to measure. It just means that a value you measure at some particular time, under one set of conditions, will be different from the value you measure at another time under different conditions.