Instructor: Prof. Ryan Torn (442.4560)

 
Course Objective: This course will describe methods of evaluating the ability of a model to predict the evolution of the atmosphere, which is an example of a chaotic system. In addition, the course will cover techniques used to address the challenges associated with these systems, including ensemble prediction systems, improving a model's initial conditions through data assimilation and advanced methods of improving model forecasts.

 
Prerequisites: Atm 562, familiarity with linear algebra, probability, partial differential equations and computer programming.

 
Class meets: TR 9:00-10:20 in BB 141

 
Syllabus

 
Assignments

Data Assimilation with Lorenz 63, Python Code

Final Project Description

 
Reading

Probability Primer (courtesy of Greg Hakim, U. Washington)

Linear Algebra Primer (courtesy of Greg Hakim, U. Washington)

Lorenz (1963)

Lorenz (1969)

Date	Paper	Presenters
3 September	Buizza and Palmer (1995)	Tyler and Alex
3 September	Hodyss and Majumdar (2008)	Cam and Lauriana
17 September	Durran and Weyn (2016)	Scott and Brennan
17 September	Baumgart et al. (2019)	Cam and Lauriana
8 October	Kleist and Ide (2014)	Kevin and Jannetta
8 October	Y. F. Zhang et al. (2018)	Adam and Peyton
22 October	Otkin and Potthast (2019)	Jeremiah and Nathalie
22 October	Lang et al. (2012)	Kevin and Jannetta
29 October	Palmer (2001)	Scott and Brennan
29 October	Eckel and Mass (2005)	Mansour and Matt
12 November	Romine et al. (2014)	Jeremiah and Nathalie
12 November	Ancell and Hakim (2007)	Mansour and Matt
19 November	Szunyogh et al. (2002), Update from Hamill et al. (2013)	Adam and Peyton
19 November	Demirdjian et al. (2020)	Tyler and Alex

 
Links

Parameter Estimation: Cloud Estimation Sea-Breeze GWD Parameter Estimation

TC Targeting: Flight Track (Fig. 1), Track Improvement (Fig. 5), Other Storm Improvement (Fig. 4), JHT Project Information, JHT Target Website

Observation Targeting: FASTEX Flight, Mean Forecast Improvement, Breakdown by Case, Breakdown by Obs. Location

Observation Impact: Impact Distribution, Impact By Type, Real-Time Monitoring

Ensemble Sensitivity: Climatological Sensitivity (Fig. 3), Case Study (Fig. 4)

ETKF: Midlatitude Example (Fig. 2)

Sensitivity Analysis: Adjoint Sensitivity for Cyclone (Fig. 7), Singular Vector Sensitivity for Cyclone, Tropical Cyclone Track (Fig. 10)

Ensemble Forecasting

Ensemble Verification: Rank Histograms

Stochastic Backscatter: Model Energy Spectra, Pattern

Stochastic Perturbed Parameters: ECMWF Perturbed Parameters, SPPT vs. SPP

Stochastic Perturbed Physics Tendency: Pattern 1, Pattern 2, Pattern Loop, Forecast Results

Multi-Physics Ensemble Graphics: Convection, Bayesian Model Averaging, Multi-Physics MJO

Multi-Physics Ensemble: SREF Configuration, SSEF Configuration, Canadian Global Ensemble

Bred Vectors: Bred Vector Schematic, ETKF Perturbation Method

Ensemble Diagram: Ensemble Diagram

Data Assimilation

Bias Correction: Mean Bias, Variational Bias Correction (Dee 2005, QJRMS)

Satellite Radiances: AMSU Weighing Functions, AIRS Channels, Satellite Count, New Satellite Count

EnKF: Covariance Comparison

4D-VAR General Schematic, ECMWF Implementation

Single Variable Variational DA: same observation variance, larger variance at second point

Bayesian Data Assimilation (From DART system)

Dynamical Systems

Singular Vector Differences (Fig. 1) (From Palmer et al. 1998)

Error Energy Spectra

Lorenz Model: Control, dual attractor (r=28), single attractor (r=13)

Henon Map

Introduction Material

Henon Map

Gaussians: Uniform Variance, Non-Uniform Variance, Covariance

Observation Impact Comparison

Observation Targeting

ECMWF Observations: Types of Observations, Distribution

Lorenz Model: Model Trajectory, Control Forecast, Perturbed Forecast

Ensemble Examples: GFS Ensemble, Regional Ensemble

 
WRF Links

WRF Users Guide

WRF Technical Description

WRF Download

NOMADS GFS Download

NCEP FNL Analysis Download

Real-Time GFS