For vertical incidence and sound moving from medium 1 to medium 2 we have for the transmission coefficient
T=2 * (rho2 * c2)/(rho2 * c2 + rho1 * c1)
rho is density {take sin to 90 deg in formula given by @neilrichards)}
then for air to water the transmission coefficient becomes close to 2 and for water to air the transmission coefficient becomes about 4.5 e-4
sound generated in water reflects easily on the surface, and sound generated in air transits easily into water. In fact detecting acoustically the overflight of an helicopter or an aircraft can been done.
Note however, that this does not mean that the sound intensity is the same. The formula hold for sound pressure values and for the same pressure value, the sound intensity in water is about 3710 times lower than in air.
Edit: One consequence of the acoustic impedance mismatch between air and water is that sound detection may be impacted. If your define your detector as intensity ratio, then you must consider the acoustic impedance mismatch (and not only pressure squared) between sound in air and in water.
comment: A technical consequence is that sound sensors must be generated such that no impedance mismatch occurs. Microphones do not work well underwater, hydrophones do not work well in air. They both work in the other domain, but not so good.