A typhoon bogus data assimilation (BDA) scheme is built in the MM5 3D-Var system with mainly three components: a) An algorithm to specify the typhoon bogus sea-level pressure (SLP) and wind profiles is constructed in the MM5 3D-Var system based on the typhoon report issued by the regional typhoon center and the 3D-Var background; b) The errors of the bogus observations are empirically assigned according to our knowledge of the typhoon SLP and wind distributions; c) The MM5 3D-Var system includes typhoon bogus observation operators. Since typhoon bogus observations can have large differences from the 3D-Var background fields, special treatment is made to allow the typhoon bogus data retained in the minimization procedure. Numerical experiments are conducted using Typhoon Rusa (2002) case in the Northwestern Pacific Ocean near landfall to the Korean Peninsula. It is indicated that the 3D-Var bogus algorithm works well and can improve the prediction of typhoon track and intensity. With MM5 3D-Var BDA, the initial typhoon vortex is balanced with the dynamical and statistical balance embedded in the 3D-Var system. It can generate a well-defined warm core structure in the typhoon initialization. Compared with bogussing in the 3D-Var background fields, MM5 3D-Var BDA results in less spin-down/up problems in the subsequent typhoon forecast. Sensitivity experiments show that assimilation of the only bogus SLP data produces too strong a typhoon intensity, while assimilation of the only bogus wind data produces much weaker intensity prediction than the observations. Assimilation of both bogus SLP and wind data obtains the best initialization and prediction for this typhoon case. Its landfall location, time and intensity are very close to the observation in the experiment.
