Tropical cyclone forecasting with model-constrained 3D-Var. II: Improved cyclone track forecasting using AMSU-A, QuikSCAT and cloud-drift wind data

Generally, the three-dimensional variational (3D-Var) data assimilation technique does not impose constraints on the numerical model such as the dynamics and physics, which are used in a 4D-Var technique. On the other hand, adopting the 4D-Var technique requires a large amount of computer resources, which limits its practical application. In Part I, a 3D-Var method was proposed by minimizing the distance between observations and model variables, and the time tendency of model variables which makes the optimized initial conditions satisfy the constraints of full dynamics and physics in the numerical model. The forward and adjoint models used in this method are the same as those in a 4D-Var method but are only integrated one time step to calculate the time tendency and gradient. Because a numerical model is adopted as weak constraint, this technique is labelled as MC-3DVar. Here the National Center for Atmospheric Research/Penn State Mesoscale Model 5 (MM5) is used. In this paper (Part II), AMSU-A retrieved air temperatures are assimilated into 32 tropical cyclone (TC) cases using this framework. The results show a significant decrease in track forecast effors. Meanwhile, one case-study of assimilating AMSU-A temperature, QuikSCAT sea-level winds, and cloud-drift winds gives dramatic track error decreases. The study shows that the assimilation of these data with MC-3DVar improves TC forecasts, and more satellite data give better performances. Copyright © 2007 Royal Meteorological Society.