A soft-computing cyclone intensity prediction scheme for the Western North Pacific Ocean

A soft-computing cyclone intensity prediction scheme (SCIPS) is introduced using an artificial neural network (ANN) approach and adding ocean heat content, as an additional predictor to the normally used atmospheric parameters, to predict tropical cyclone intensity change in the western north Pacific Ocean. We used 1997-2004 data to develop and validate this scheme. The ANN-based estimations have been compared with observations and estimations using the multiple linear regression (MLR). SCIPS performance improves upon MLR as the lead hour increases from 12 to 120 h and also for high intensifying cyclones.