Subsea Power Distribution Unit

We have a subsea power coupling module that provides isolated 550 VAC to our subsea control pod on a well in the Gulf of Mexico. The power line is actually termed as an umbilical, with the two cores or wires running from our power unit on the platform to the subsea field. The length is in the order of 50 miles or so. In our coupling unit we have an Integra power meter which supplies our voltage, current, THD, power factor, apparent power, reactive power, etc, data. We also have a Bender A-ISOmeter IRDH375 that monitors our insulation resistance using Adaptive Pulse Measurement AMP. The ISOmeter has an alarm output contact that opens our contactor in the event of an low (600kOhm) insulation reading. These faults are common in older fields such as this, but this seems to be an exception. The power meter documentation for the Integra 1350 says the power factor screen will display any number between 1 and -1, i.e. -.76 which would indicate a power factor of 76 that is lagging (inductive reactance I assume). Our situation is that the power factor is shown to be a .46, which is indicative of a LEADING power factor, and very low at that. The norm for our systems is around -.67 to -.78. I was assuming at first that the power factor was leading due to the large amount of capacitance due to the length of the umbilical. I know that our lines are usually in the negative range in respect to power factor due to the step up transformer in the coupling unit and the step down transformer in the subsea control unit. There are no other components in the line other than the two transformers and 1 large varistor in the coupling unit to suppress transients. The problem is a low insulation alarm, and resulting trip in our supply. We cannot seem to correlate this event with any other. The power coupling unit itself is run off a UPS with a stable and conditioned supply. We do see a 20 volt increase in the subsea control unit, followed by a drop of the power factor from .46 to .35, then we will see the ISOmeter go into alarm and trip the circuit. Any ideas what could cause such a low, leading power factor? What would be the effects of a low, leading power factor? Would there be any connection between the low power factor and a low insulation resistance, or could the varistor be failing and cause such a low insulation resistance?