you are true - geophysical methods may be helpful in solutionn of your task. But not IP. Turn your attention to magnetotelluric, magnetometry, gravity and certainly seismic. If you have available reflection seismic results with sufficient long time record (at least 8 sec), you can perhaps reach Moho discontinuity. Magnetic is useful for revealing of hidden volcanics (subvolcanic - hypoabysal bodies) that can usualy have positive influences on the increasing of thermal gradient. It depends, which results are available.
I suggest you use magnetometry and magnetotelluric. With depth analysis and appropriate software obtained from magnetometry; the corresponding curie temperatures could be obtained from magnetotelluric. I wish you good luck.
depending on the target depth and available temperature well logs (to be used for calibration) it is possible to estimate the temperatures at large depths from EM sounding data (magnetotellurics, magnetovariational data, CSEM). Please, have a look at the monograph (Spichak V., Zakharova, O., 2015. Electromagnetic geothermometry. Elsevier, Oxford), where you may find the fundamentals of the new technique and examples of its application in different geological environments.
Please, do not hesitate to contact me ([email protected]) for more details,