Structures of adjoint sensitivities to potential vorticity for specific initial and final norm are investigated for a short-range cyclone forecast in a three-dimensional quasigeostrophic (QG) model. Moreover, adjoint sensitivities to potential vorticity are compared with nonlinear sensitivities calculated for the same cyclogenesis case in the QG model. The adjoint sensitivities using different initial and final norms (e. g., total QG disturbance energy and potential enstrophy) show approximately similar characteristics for the horizontal and vertical structures and evolutions. Consistent with previous studies, the horizontal structure of the adjoint sensitivity is smaller for the energy norm than for the potential enstrophy norm. The dynamical mechanism of cyclone development by adjoint sensitivity coincides with that of nonlinear sensitivity, with slight differences in the region of sensitivity maxima over the upstream (nascent) low for the adjoint (nonlinear) sensitivity. The adjoint sensitivities show different vertical distributions from the nonlinear sensitivities. Consistent with the sensitive regions denoted by singular vectors and error evolution in the QG model, maxima of the adjoint sensitivities are located at both the upper and lower boundaries, with prominent secondary peaks in the lower to mid-troposphere of the domain. The level of the secondary maxima changes depending on the initial and final norm used. The secondary peak is located in the lower to mid- (mid-) troposphere for the total QG disturbance energy (potential enstrophy) as the initial and final norm. Based on the correspondence in the level of the sensitivity maxima in the interior of the domain between the adjoint and nonlinear sensitivities, adjoint sensitivities may serve as an alternative to nonlinear sensitivities given the enormous computing expenses in nonlinear sensitivity calculation.
Bibliographical noteFunding Information:
The authors wish to thank two anonymous reviewers for their valuable comments. This study was supported by the Korea Meteorological Administration Research and Development Program under Grant CATER 2006–2102. The first author is grateful to Prof. Michael Morgan for discussions at an earlier stage of the work.
All Science Journal Classification (ASJC) codes
- Atmospheric Science