METO614: Numerical Weather Forecasting and Predictability
Instructor:
Dr. Eugenia Kalnay (3431 CSS; (301) 405-5370; ekalnay@atmos.umd.edu)Office hours: Mon-Wed-Fri: 9-11
Lectures: MWF 1:00-1:50 Room: CSS 2114
This course should provide you with a solid understanding of modern numerical weather forecasting and predictability, concentrating on basic concepts and essential developments.
Text: We will finish writing a new book to be published by Cambridge U. Press! It is available on the web in html and Word97. Outdated but still useful books: Haltiner and Martin (1980), Thompson (1961). New book on computational methods by Dale D Durran (Springer, 1999).
Homework and laboratory:
There will be 5 homework sets; we will build simple models: shallow-water equations model, a vorticity model, and test basic concepts on these models.
1. Historical overview (about 5 lectures)
Introduction: V. Bjerknes vision. Richardson's experiment. Charney et al. Filtered models.
Data assimilation. Beginning of operational NWP. Primitive equation models: Global and Regional models. Ensemble forecasting. Nonhydrostatic mesoscale models. Forecast skill: overview
2. The continuous equations (about 6 lectures)
Primitive equations. Basic wave solutions. Filtered models. Shallow water equations
Horizontal and vertical coordinates. Spherical coordinates. Conformal mapping.
Nonhydrostatic equations
3. Discretization of the equations (about 8 lectures)
The wave equation and the heat equation: space and time finite differences. CFL
Explicit and implicit numerical schemes. Spectral and finite element discretizations.
Vertical coordinates: Pressure, sigma, theta, eta. Boundary conditions for regional models.
4. Parameterizations of subgrid-scale physical processes (about 4 lectures)
Radiation. Surface processes, boundary layer
Clouds: stratified, convective, prognostic equations for clouds
Coupling with land and ocean models
5. Examples of regional and global models (about 2 lectures)
NCEP operational models: global model, Eta model, Regional spectral model
Nonhydrostatic models: ARPS. Postprocessing, MOS, Perfect Prog, Kalman Filter
6. Data assimilation (about 8 lectures)
Successive corrections, Barnes, nudging. Optimal Interpolation and 3D VAR. Brathseth
4D VAR and approximations; Kalman filtering, ensemble Kalman Filtering
Initial balance, NLNMI, Lynch filter
7. Atmospheric predictability and ensemble forecasting (about 8 lectures)
Historic overview: Chaos. Lorenz (1963), Lorenz (1965) Singular vectors and Lyapunov vectors
Early ensemble forecasting: Epstein, Leith, LAF, MCF, SLAF
Operational ensemble forecasting: Breeding, Singular vectors and ensemble data assimilation. Ensembles of operational models. Predictability of the atmosphere-ocean-land system
Grading: Homework and lab: 40%; mid-term exam: 25%; Final exam: 35%