Understanding the Sensitivity of Cyclogenesis Using Adjoint Analysis and the Link with Targeted Observations

While dynamical processes associated with extratropical cyclogenesis are well known, it remains unclear what effect certain perturbations of these dynamical forcings have on cyclone development. From a modeling perspective, this is useful to understand, as slight changes in the pertinent dynamics have the ability to change drastically a NWP model forecast. In order for a forecast to remain as accurate as possible, it is useful to assimilate additional observations located in regions sensitive to the development of a cyclone. Through the use of adjoint modeling, specific regions where the initial cyclone state is most sensitive can be calculated. These sensitive regions can then be targeted for additional observations which can be inserted into a NWP model forecast to produce a more accurate forecast. To achieve this, it is first important to understand what changes to cyclone development are garnered from a perturbation inserted into a region of sensitivity. This is done by calculating a set of optimal perturbations using an adjoint model based on some change that is wished to be achieved in the cyclone at the final state, for example, a 5hPa deeper cyclone. These optimal perturbations can then be used to perturb a cyclone simulation’s initial conditions in the specific regions where the initial cyclone state is most sensitive. Once the perturbation has been inserted in the model, synoptic diagnosis of the deviation of the perturbed forecast trajectory from a control forecast trajectory can be used to examine the dynamical impacts of the perturbation on cyclone development. A case study utilizing the WRF model of an explosively deepening October 2010 North American cyclone that caused $18.5 million in damages will be conducted using adjoint sensitivities in conjunction with other synoptic diagnostics. This study will render a better understanding of how different dynamical perturbations impact the life-cycle of a rapidly developing cyclone, and can also be applied to the development of more modest cyclones. The knowledge gathered can then be applied to a study of not only what type, but also when and where it can be important to try and assimilate additional observations into a NWP model forecast. The goal is that the result will produce a more robust forecast for events that are characterized by varying degrees of forecast uncertainty.