As the flaps were transitting to 10 degrees, the combination of airframe drag, loss of lift and airspeed, along with the increased AOA necessary to maintain 2300ft, may have triggered the stick shaker and pusher, which disconnected the autopilot. At that point, the stabilizer trim may have been adjusted significantly nose up, compensating for the drag and lift loss to maintain that altitude. This might explain the immediate 31 degree nose pitch up, after the autopilot disconnected.
As an aircraft with positive static longitudinal stability (like the DHC-8-400) decelerates, if no control inputs were made, the nose would fall. The autopilot applies increasing aft stick commands to hold level flight. The autopilot will command the pitch trim to run to reduce the sustained control force that it is holding. The result is that the autopilot is always slightly slow to move the trim, and the control force is never completely trimmed out. If the autopilot disconnects, and nothing else has changed, the nose will fall slightly.
There has to be another reason for the increase in pitch attitude to 31 degrees nose up. It was probably due to the sum of pilot aft control input, and the pitching moment due to the large change in power that he made. I am told that the FDR on the -400 records the force the pilot puts on the controls, so it should be possible to eventually learn what the actual pilot control inputs were.