Ah, driving high frequency sonar transducers, been there done that.
The usual thing is to use a tuned transformer (Resonate the secondary winding inductance with the transducer fixed capacitance) as part of a fifth order or so matching network, then drive the whole thing with a square wave, the transducer will do a good job of turning it into a sine in the working fluid. Note that your driver will probably have to deal with a load that is not in fact resistive, active snubbers can be your friends.
These transformers can be small as the duty cycle is usually negligible, but the voltages can be huge so kapton tape is a good thing, as is litz wire to keep the Q where you want it.
Usually the input match is some form of L network, often with deliberate series resistance to make the bandwidth a little wider, do make sure you work from a G/B plot done in water, they measure **very** differently in air.
Are you sure that plot is impedance rather then admittance? I would expect the resonance to be a low impedance as the devices model as series resonant devices with an annoyingly large shunt cap and the acoustic radiation resistance in the series arm.
73 Dan.