Isn't induced EMF in an antenna dependent on the power of the electromagnetic wave?
Indeed it is not - for given antenna geometry, the induced EMF due to external wave depends only on electric field of the wave where the antenna is. The total induced EMF includes also self-induced EMF and the latter depends on behaviour of current in the antenna and its geometry (it's complicated). But neither of these are directly related to total power of the EM wave. For all practical situations, the available power or possibly absorbable power is much smaller than total power of EM wave.
The induced EMF in an antenna is often given as E⋅h, where h is the height of the antenna and E is the electric field strength in V/m, but this does not seem to capture the power of the electromagnetic wave in the far field (i.e. the multiplication with the magnetic component related by the impedance of free space) and hence the correct transfer of power.
Yes, induced EMF due to EM wave is line integral of EM wave's electric field; in case of straight line dipole antenna directed along the line of polarization of the wave, it is electric field of the wave at the antenna times length of the antenna. Also yes, this quantity does not capture available power, and neither it should be expected to.
Electric field or EMF of external EM wave for a given antenna geometry do not imply anything about available power that the antenna can draw from the EM wave. It tells us the maximum value of voltage that can be induced on the antenna terminals, and in practice this voltage is slightly lower due to the fact current has to be allowed to flow (the antenna is connected to a circuit). But it does not tell us induced current in the antenna.
The power drawn from the EM wave depends not only on external EMF, but also on details of electric current as function of position and time, that oscillates in the antenna. It is like with force and velocity in mechanics: if the current ("velocity") is high and in phase with the external electric field of the wave ("force"), then absorbed power can be high. This requires tuning antenna parameters and it results in condition of impedance matching/adaptation - in order to extract as much power as possible, impedance of the connected circuit has to be tuned to impedance of the antenna.
I read that the antenna automatically transforms the impedance of free space to the impedance of the antenna. Does this mean that the induced voltage is actually $$Eh\sqrt{\frac{Z_0}{Z_a}}$$ where $Z_0$ is the impedance of free space and $Z_a$ is the internal impedance of the antenna? If the antenna acts as a transformer, surely it would be.
Even if the formula for voltage would be correct, this would not be correct description in terms of physics. There is only single voltage on the antenna terminals, there is no transformation between two voltages going on like in HV-AC transformers. Yes, induced EMF has different value than induced voltage, but transformer does not transform EMF into voltage - it transforms one voltage to another voltage, using induced EMF. So the description in terms of transformer is not very physical.
There are other situations in physics where EMF and voltage magnitudes are not equal to each other - for example, real inductor with ohmic resistance. We do not say real inductor transforms EMF into voltage like a voltage tranformer with windings and core - we just accept that EMF and voltage are two different concepts, with possibly different values. It is only in special idealized circumstances they have the same magnitude (perfect conductor solenoid).