Skip Navigation
JCSDA Logo and banner

2014 JCSDA Seminars


Title

Correcting for Position Errors in Variational Data Assimilation

Summary Slides, (PDF, 19.96 MB)

Presentation audio stream, (MP3, 60.06 MB)

Speaker Thomas Nehrkorn
AER
Date Monday, November 10, 2014
2:00 p.m. - 3:00 p.m.
Conference Center, NOAA Center for Weather and Climate Prediction,5830 University Research Court, College Park, MD
Abstract

Show Abstract

Forecasts used in variational data assimilation schemes for the atmosphere and ocean often exhibit significant position errors. In standard data assimilation approaches position errors often result in background errors that have complex error correlations, are not normally distributed, and are particularly difficult to correct. A number of approaches have been proposed to measure and correct position or phase errors, ranging from techniques originating from image processing such as "image warping" or "morphing" and "optical flow," object-oriented verification measures, to alignment approaches aimed at data assimilation improvements.

We present an implementation of a displacement scheme to correct phase errors based on the feature calibration and alignment (FCA). In its original formulation, a set of two-dimensional displacement vectors is applied to forecast fields to improve the alignment of features in the forecast and observations. These displacement vectors are obtained by a nonlinear minimization of a cost function that measures the misfit to observations, along with a number of additional constraints (e.g., smoothness and non-divergence of the displacement vectors) to prevent unphysical solutions. Results from this implementation will be compared with a more recent implementation within the WRF-Var (WRFDA) algorithm, in which the nonlinear minimization is replaced by the (linear) conjugate gradient inner-loop minimization combined with outer loop nonlinear adjustments, and the ad-hoc penalty function constraints are replaced by an error-covariance representation of the displacement vectors.

Remote Access Video: 1. Go to JCSDA Seminar and click on the seminar title
2. Enter your name and email address.
3. Click "Join Now".
4. Follow the instructions that appear on your screen.
Audio: USA participants: 1-866-715-2479, Passcode: 9457557
International: 1-517-345-5260
Contact

Erin Jones


Title

Satellite Data Assimilation at ECMWF

Presentation file posted here when available.

Powerpoint version, (PPTX, 15.35 MB)

Presentation audio stream, (MP3, 118.29 MB)

Speaker Stephen English
ECMWF
Date Monday, October 20, 2014
11:00 a.m. - 12:00 p.m.
Conference Center, NOAA Center for Weather and Climate Prediction,5830 University Research Court, College Park, MD
Abstract

Show Abstract

ECMWF assimilates observations from over 50 satellite instruments, and monitors many more in the operational system. Recently the main focus has been in three broad areas:

1) Full (or at least more complete) utilisation of hyperspectral sounder data; 2) Moving more moisture sensitive observations into the so called "all-sky" assimilation framework, where we explicitly model the impact of cloud and precipitation on the observations, and allow these observations to influence the analysis; 3) An increasingly sophisticated treatment of observation errors, notably for Atmospheric Motion Vectors (AMVs), and observation error correlations for radiances.

In the first area, ECMWF has been able to demonstrate that for cloud-free scenes assimilation of IASI observations as radiances or Principal Components (PCs) can give the same impact, and a better impact when more of the spectrum is assimilated affordably through the PCs. Indeed, this advantage is sufficient that the assimilation of cloud-free PCs is, in general, equal to and often better than the assimilation of radiances in clear and overcast scenes, and for channels peaking above cloud. In the second area, ECMWF has been assimilating microwave imager data for many years through the all-sky route, but now humidity sounders are also being moved to the all-sky framework, with noticeable positive impact, especially on vector wind forecast scores. In the third area, the impact of AMVs has notably increased through the use of scene dependent errors, and experiments are showing not only can we cope with observation error correlation, we may even be able to exploit it in a positive way in some situations. In addition, the seminar will briefly discuss the status of Radio Occultation assimilation, progress with data from the Feng Yun satellites, new observation impact diagnostics, and a perspective on current and future directions for data assimilation in general at ECMWF.

Remote Access Video: 1. Go to JCSDA Seminar and click on the seminar title
2. Enter your name and email address.
3. Click "Join Now".
4. Follow the instructions that appear on your screen.
Audio: USA participants: 1-866-715-2479, Passcode: 9457557
International: 1-517-345-5260
Contact

Kevin Garrett


Title

Analysis and Prediction of Clouds using Satellite All-Sky Radiances: Lessons Learned and Perspectives

Presentation file posted here when available.

Powerpoint version, (PPTX, 517.19 MB)

Presentation audio stream, (MP3, 94.66 MB)

Speaker Thomas Auligne
NCAR
Date Wednesday, September 17, 2014
2:00 p.m. - 3:00 p.m.
Conference Center, NOAA Center for Weather and Climate Prediction,5830 University Research Court, College Park, MD
Abstract

Show Abstract

The initialization of cloud parameters is one of the main frontiers for improving short-term prediction skills in numerical weather prediction (NWP) models. We will present the latest developments at the National Center for Atmospheric Research (NCAR) in building a capability to accurately initialize cloud parameters based on all-sky satellite observations assimilated in the Weather Research and Forecasting (WRF) numerical model. Two approaches have been considered:

1) A nowcasting system with a simplified cloud fraction. The analysis is designed to closely fit radiance observations and the NWP model dynamically transports clouds without involving the physics. This method is computationally efficient and useful for short-term forecast since it does not suffer from the typical imbalances often associated with model re-initializations.

2) A multivariate analysis including the microphysics parameters. It relies on a hybrid ensemble/variational data assimilation system with an augmented control variable for clouds. Specific developments to address non-linearities in the observation operator and non-Gaussian error distributions will be presented. We will also introduce original algorithms to correct for position errors and estimate the analysis error within the variational data assimilation system.

The current analysis shows a sustained impact on cloud forecast, yet there are still many challenges data assimilation will need to address to optimize the use of satellite all-sky observations.

Remote Access Video: 1. Go to JCSDA Seminar and click on the seminar title
2. Enter your name and email address.
3. Click "Join Now".
4. Follow the instructions that appear on your screen.
Audio: USA participants: 1-866-715-2479, Passcode: 9457557
International: 1-517-345-5260
Contact

Kevin Garrett


Title

Uncertainties in Forward Passive Microwave Brightness Temperatures and What to Do about Them

Summary Slides, (PDF, 9.56 MB)

Presentation audio stream, (MP3, 67.89 MB)

Speaker Jeff Steward
UCLA, Joint Institute for Regional Earth System Science and Engineering
Date Tuesday, May 6, 2014
1:30 p.m. - 2:30 p.m.
Conference Center, NOAA Center for Weather and Climate Prediction,5830 University Research Court, College Park, MD
Abstract

Show Abstract

Forward simulations of cloudy satellite radiances are fraught with uncertainty especially in the microwave regime. The scattering parameters that determine satellite brightness temperatures critically depend on complex, unknown size and shape distributions of liquid and solid hydrometeors. These uncertainties can lead to differences of 50 K or more in forward calculations alone, and therefore this issue must be addressed before attempting to assimilate this data into operational models. In this talk, I will quantify some of the various sources of uncertainty in passive microwave radiances (e.g. TRMM/TMI). I will also present an elegant method we have developed at UCLA and JPL to assimilate these uncertain observations. This method is based upon statistical inference using the directions of the most certain correlations. The uncertain directions can be neglected, thus reducing the order of the problem as well as regularizing the mapping. The observation error covariance formulation also becomes straightforward and elegant. We present the encouraging results of our microwave brightness temperatures simulation in both 1D-Var and Ensemble Kalman Filter frameworks.

Remote Access Video: 1. Go to JCSDA Seminar and click on the seminar title
2. Enter your name and email address.
3. Click "Join Now".
4. Follow the instructions that appear on your screen.
Audio: USA participants: 1-866-715-2479, Passcode: 9457557
International: 1-517-345-5260
Contact

Kevin Garrett

Modified December 4, 2014 11:12 PM
NOAA | NWS | NASA | Navy Weather
Air Force Weather | Website owner: JCSDA
Level A conformance icon, W3C-WAI Web Content Accessibility Guidelines 1.0 and Valid HTML 4.01 Icon
icon: valid HTML 4.01 transitional. Level A conformance icon, W3C-WAI Web Content Accessibility Guidelines 1.0