Extended Abstract (136K)JP4.4
Using 4d-VAR to move a simulated hurricane in a mesoscale model
Ross N. Hoffman, AER, Lexington, MA; and J. M. Henderson and S. M. Leidner
Four-dimensional variational analysis (4d-VAR) is a data assimilation technique that has promise for calculating optimal perturbations for weather modification. By modifying the initial conditions of a mesoscale model, we attempt to influence the future path of a tropical cyclone and thus minimize potential damage over land areas in the simulation. We present results from two preliminary simulations. The first uses unmodified initial conditions ("unperturbed"), while the second uses initial conditions modified by 4d-VAR ("controlled"). For the controlled case, we provide the Penn State/NCAR Mesoscale Model 5 4d-VAR system with a target model state, valid 6 h into the simulation, in which the unperturbed model vortex has been repositioned 112 km to the west. 4d-VAR attempts to determine the optimal controlled simulation which simultaneously minimizes the size of the initial perturbation and the difference (using a quadratic norm) between the controlled model state and the target.
The controlled model vortex is displaced over 100 km to the west during the first 6 h of the simulation so that the controlled forecast at 6 h closely matches the target model state. Perturbations to the initial conditions are small and are concentrated in the lowest layer temperature field. Examination of the convective precipitation during the first 6 h of the simulation reveals that convection is redistributed to the west of the vortex center. During the next 36 h, the controlled vortex track remains approximately 100 km to the west of the unperturbed track.
Joint Poster Session 4, The Use of Mesoscale Models in Severe Storms Forecasting (Joint with 21SLS and 19WAF/15NWP)
Tuesday, 13 August 2002, 3:00 PM-4:30 PM
Previous paper Next paper