Doesn't the prop RPM have to do with the issue? During the climb sequence, the reduction of Np will give an increase in the torque values, as:
TQ=Power/Np
So if the Np decrease more than the power, the TQ values must rise (for ATR 42-500 TQ typically rising from 90% to ~109% after selecting CLB, while the actual power decreases) Just guessing...
On the oter hand, on ATR 42-500, when changing from CLB to CRZ, the torqes (and the fuel flow and ITT) increase, with Np being constant at 82%. Seems that CRZ pwr is higher than CLB. Nobody (at least in my company) is able to explain this phenomenon...
Greetings
Smola