I am surprised at their assertion
to prevent thermal deterioration of electronic parts, the oven temperature is generally kept at 70C, about ten degrees lower than the turnover temperature of SC-cut resonators.
It sets the tone for the whole research.
Would it not be better to get better components and more accurate temperature controller? I can imagine that if you stick a modern design - like IHR - in a double oven then you can achieve better stability than typical classic OCXO with less stress on the components.
Conceptually, this dual mode idea might work, but repeatability will probably be awful - I reckon frequency difference between the modes will vary wildly in a batch - with spurious modes in the way and all over the place. This will probably make every oscillator into a unique hand-built part.
And in the end all it solves is temperature stability [of sub-optimal operation point] and does not address ageing, contamination, frequency jumps, pressure dependence, etc.
Leo