Numerical prediction of tropical cyclone motion


Student thesis: Master's Thesis

View graph of relations


  • Cheuk Kwan LAW


Awarding Institution
Award date22 Jun 1996


The motion of the tropical cyclones during August and September 1990 is studied with the non-divergent barotropic vorticity equation in spherical coordinates. The model was initialized using the Final Analysis wind field from the Tropical Cyclone motion (TCM-90) Experiment. The sensitivities of the model to the following parameters in predicting the movement of tropical cyclone (TC) were investigated: (a) the initial position of the vortex centre; (b) the structure of the outer wind profile of the vortex, and (c) the environmental steering. The effect of the initial position error on the forecast error was studied by dividing the sample into two groups based on the initial position error. Significance tests showed that the 12-h mean forecast error (MFE) depends on the initial error but the MFE after 24-h has less dependence. The sensitivity of the model to the vortex structure was then tested. The results suggested that although the Final Analysis data provided a better definition of the flow related to TC movement, the representation of the magnitude and pattern of the asymmetry in the outer circulation was still insufficient. In addition, the coarse resolution of the model and boundary effects biased the forecasts significantly. The sensitivity of the model to environmental steering was investigated by inserting a bogus vortex with different steering flow (the persistence vector defined as the operational estimate of motion over the previous six hours). The movement of the bogus vortex could be forced to become closer to the best track. The magnitude and orientation of the advection of the asymmetric vorticity by symmetric flow did not change much even though the asymmetric flow increased. However, the advection of the symmetric vorticity by asymmetric flow increased in magnitude and the vorticity advection pattern rotated in the direction of the persistence vector. An appropriate persistence vector together with a proper bogus vortex therefore play an important role in improving the prediction of TC movement.

    Research areas

  • Cyclone forecasting, Tropics, Cyclones