Considerable advancements have recently been made in understanding tro
pical cyclone motion. Based on these new understandings, the requireme
nts for accurate tropical cyclone motion prediction with dynamical mod
els can be specified. Four issues related to dynamical track predictio
n are the initial specifications of the environmental wind field, the
symmetric vortex and the asymmetric vortex structure, as well as the a
dequacy of the models to predict the time evolution of each of these t
hree components of the total wind field. Recently developed barotropic
and limited-region or global baroclinic models are examined in terms
of these issues. The capability of the Hurricane Research Division bar
otropic model to provide skillful track forecasts to 48 h retrospectiv
ely substantiates that tropical cyclones motion is governed by barotro
pic dynamics to first order. Limited-region baroclinic models are demo
nstrated to have many of the numerical characteristics, physical proce
ss representations, and initial condition specifications that will be
required to properly predict tropical cyclone tracks. In particular, t
he semioperational Geophysical Fluid Dynamics Laboratory model most cl
osely addresses all of the above issues, and has demonstrated potentia
l for markedly improved tracks for a small sample of cases. Finally, t
he inclusion of some aspects of tropical cyclone structure in the init
ial conditions of global baroclinic models has improved their track pr
edictions. Thus, the outlook is for a significant improvement in dynam
ical track predictions.