Adjoint sensitivity of an idealized extratropical cyclone with moist physical processes

An adjoint model (MAMS1) that includes parametrizations for convective (subgrid-scale), and non-convective (grid-scale) precipitation, and surface latent-heat flux is used to investigate an idealized extratropical cyclogenesis. The adjoint sensitivity information demonstrates the effects that perturbations of model variables and parameters at various times during the cyclone life cycle have on forecast cyclone intensity. For a nonlinear trajectory that includes precipitation processes and surface latent-heat flux, the accuracy of the tangent-linear and adjoint model is much higher when moist physical processes are included. Inclusion of moist processes in the adjoint model increases sensitivity magnitude compared with sensitivity obtained with a dry adjoint model, but does not alter the primary spatial pattern of sensitivity. The larger cyclone deepening rates that occur with the inclusion of moisture are related to latent-heat release from condensation of water vapour in areas of the middle and lower troposphere (the warm-front region) that are strongly sensitive to temperature perturbations in both dry and moist cyclone simulations. The effects of diabatic heating on cyclone development are interpreted as a reinforcement of dry baroclinic instability and not a separate development mechanism (which would have a unique non-baroclinic sensitivity signature). The sensitivity patterns explain why favourably positioned latent-heat release is an ingredient that can lead to explosive baroclinic development. Cyclone intensity is very sensitive to the vertical distribution of temperature perturbations, so this feature of diabatic heating is critical to the cyclone forecast. An increase in the transfer coefficient C-E, for the surface latent-heat flux can intensify the cyclone by adding moisture to the lower troposphere in the cyclone warm sector before the release of latent heat by precipitation processes. Perturbations of C-E have more effect on cyclone intensity than perturbations of the transfer coefficients involved in surface sensible-heat flux and surface stress during most of the cyclone life cycle.