FALL 2008
SYLLABUS
METR 5303 - Objective Analysis
(or, more appropriately,
Objective Analysis, Initialization and Data Assimilation)
Instructors: Dr.
Fred Carr (fcarr@ou.edu) and Dr. Ming Xue (mxue@ou.edu)
Office
Hours: Carr - MWF, 9:00 - 11:00 am (or by appointment)
Xue - TF, 11:00am - 12:30pm (or by appointment)
When and Where: TR 1:00-2:15 pm; Rm. 5930 NWC
Prerequisites: MATH
3113 (ODE) and 3113 (linear algebra);
ENGR 3723 (numerical methods) or equivalent. Or permission of instructor.
Texts: R. Daley, 1995, Atmospheric Data Analysis, Cambridge
University Press, 472pp
E. Kalnay, 2002, Atmospheric Modeling, Data Assimilation and Predictability, Cambridge University Press, 341 pp.
Also, Selected material from journals and review articles. A web site has been set up for the
course at http://twister.ou.edu/OBAN2008/.
Tests: Exam 1: Tuesday, October 14
Exam 2: Tuesday, November 18
Final Exam: Wednesday, Dec. 17, from 1:30-3:30 pm.
Grading Policy: In-class exams (2): 15%
each
Computer
assignments (5) 40%
Final
Exam 30%
Objectives: This course is designed to improve our
understanding of what is (or should be) done to "raw" observations
before they are used in diagnostic studies or in numerical weather prediction. With the avalanche of data from new
observing systems (Profilers, radars, numerous satellites, ASOS, ACARS, etc.),
it is important to understand these data and the techniques used to optimize
their information content. Of particular importance are the procedures used to analyze data onto
regularly-spaced grids for the purpose of diagnostic computations or for
initial conditions for numerical models.
Numerous objective analysis techniques will be presented. The concepts of balancing or
initializing the data will be explained.
New developments in the use of the model equations to achieve the balance and to assimilate indirect observations (four-dimensional data
assimilation, adjoint techniques, ensemble Kalman filtering, etc.) will be
presented. The data assimilation
methods will also be discussed in the context of optimal estimation theory. Although
much of the literature on these subjects concerns large-scale NWP, techniques
with promise for use in mesoscale models will be emphasized.
Tentative List of
Topics:
1. General
comments on observing systems
2. Objective
Analysis
(a) General concepts; function fitting
(b)
Cressman, Barnes and Bratseth techniques
(c)
Filtering concepts
(d) Statistical analysis
(i) optimum
interpolation
(ii)
multivariate O.I.
3. Three-Dimensional Variational Analysis and Data Assimilation (3DVAR)
4. Four-dimensional
Data Assimilation – Conventional Approaches
(a) Historical approaches
(b) Newtonian relaxation or nudging
5.
Four-dimensional Data Assimilation – Variational Approaches
(a) 4DVAR concept,
(b) Adjoint techniques for
minimization – 4D
6. Kalman
Filters
(a)
Classic Kalman filter and extended Kalman Filter
(b)
Ensemble Kalman Filter
7. Special
Topics (if time allows)
(a) Methods used in current
operational forecast systems
(b) Methods for mesoscale and
storm-scale prediction
(c) Observing System Simulation
Experiments (OSSE)
Computer
programs will be written as part of the homework assignments. Thus working knowledge of a programming
language is required. By the end of
the course each student will have at least three working objective analysis
codes and have worked one or two simple variational analysis problems. Note that homework determines a
significant part of the final grade, so that your efforts there will be
rewarded.
All students are expected to be familiar with and abide by the OU
Academic Misconduct Code.
Information on this code is at http://www.ou.edu/studentcode