Perturbation forecast modellingThe Met Office Perturbation Forecast (PF) model consists of a set of linear equations that are linearized about a trajectory of the Met Office Unified Model (UM). The PF model and its adjoint are essential to the variational data assimilation (DA) procedure. In the usual DA formulation the PF model would be known as the "tangent linear model" which is actually the Jacobian of the nonlinear model that can be used to evolve a perturbation of a nonlinear model variable forward in time. However, at the Met Office, the PF model is not precisely a tangent linear approximation to the full nonlinear UM. Instead, the PF model contains a version of the linearized equations of the UM plus a set of parametrizations of its own which serve to model moisture, cloud microphysics and other important nonlinear physical processes. The perturbation variables are a smaller set of prognostic variables than are used in the UM; they are 3D wind field, density, pressure, potential temperature and total mass of water per unit mass of air. At the Met Office, the PF model is typically run at a lower spatial resolution than the UM and the output is re-gridded onto the nonlinear model's grid.
|Improving cloud microphysics in the PF model||Cristina Charlton-Perez|
You are visitor number 5171 to this page.