Use of Satellite Data in Local Analysis and Modeling Dan Birkenheuer Forecast Systems Laboratory Bou

1
Use of Satellite Data in Local Analysis and
ModelingDan BirkenheuerForecast Systems
LaboratoryBoulder, CObirk_at_fsl.noaa.gov303
497 5584http//laps.fsl.noaa.gov/birk/comap/comet
/index.htm
2
Data Available to the Forecaster
Imager Sounder
GOES GVAR
Derived Products
CIRA
RAMSDIS
Imager Sounder
Imager Sounder
Imager
FSL
NESDIS
Derived Products
Imager Sounder Derived Products
FX Alpha System
Imager Derived Products
Imager Sounder
SBN feed Image data
AWIPS
LAPS
Imager Derived
Imager
3
Image data from Satellite Broadcast Networks (SBN)

• Channel Enhancements 2, 3, 4, 5
• Error as a function of temperature
• Error as a function of 8-bit count value

4
Various Approaches for Assimilation

• Sounding Retrievals
• Variational Approach

5
Sounding Retrievals

• The basic approach
• Obtain a set of radiances for specific field of
  view
• Using first guess atmosphere (chosen by some
  method) compute radiances with radiative transfer
  model
• Modify first guess through various methods (some
  iterative) such that a new atmospheric profile
  generates radiances similar to observed
• Call the revised profile a retrieved sounding

6
Sounding Retrievals (cont.)

• Once the retrieval step is complete, it is used
  in analysis like a RAOB sounding
• On the plus side
• We produce soundings independent of our analysis
  scheme
• We use soundings like RAOBs in the analysis
• On the minus side
• Retrievals are NOT RAOBs, soundings have a FOV
• Processing retrievals off line ignores other data
  that might be available.

7
Variational Assimilation

• The basic approach
• Obtain a set of radiances for a specific field of
  view
• Use a forward radiative transfer model with other
  data sources and their respective forward models
  and error characteristics to simultaneously
  arrive at a solution that best fits the error and
  measurements of all the data.

8
Variational Assimilation (cont.)

• On the plus side
• Direct use of radiances in analysis (no point
  approximations made).
• Solution utilizes all data in one computation
• The solution treats radiances for what they are
  radiances, not soundings
• The solution is sensitive to the error
  characteristics of all data, balancing radiance
  error appropriately.

9
Variational Assimilation (cont.)

• On the minus side
• Hard problems to solve (the math is more
  sophisticated, but computers are better now than
  they were 10 years ago).
• The analysis system has to be rewritten
  completely to take advantage of this technique
  over using one that has been running in place
  using traditional methods.

10
Approaches to radiance assimilation in LAPS (a
bit of history)

• Horizontal Shape Matching (HSM)
• Functional minimization (1DVar)
• Illustration

11
The zero-dimensional variational problem
(Illustration)

• Data and functional
• Forward model (Radiance equation)
• Graphical solution
• Analytical solution

12
Theory of GOES LAPS Moisture Adjustment

• Graphical adjustment
• Summary
• Impact on Analysis

13
Impact on Local Forecasting

• RAMS showed little temperature change
• Thermal bias only influenced by GOES impacting
  cloud presence and surface gradients
• RAMS shows positive GOES impact in bias
  throughout the moisture run

14
Better Data Mean Better Analysis

• An experiment contrasts the utility of 8-bit
  image data and 13-bit sounder data
• More impressive results occur with more channels,
  adding 7 sounder channels to the moisture
  analysis allows analysis of more levels

15
GOES Derived Precipitable Water (GVAP)

• Real-time experiment in progress
• Paired runs with and without GVAP archived since
  May 99
• Statistics show positive improvement at most
  levels
• Need to make the data available in real-time.

16
(No Transcript)
17
Plans

• Extend moisture analysis to 2 layer moisture
  using GOES sounder data and current variational
  algorithm
• Done -- in-house
• Planned upgrade to AWIPS if/when AWIPS gets
  sounder radiances (a possibility)
• Switching to OPTRAN forward model
• Interfacing LAPS to polar satellites
• Bringing thermal and moisture analyses together
  variationally using sounder data (LAPS II)
• Continue to upgrade LAPS in AWIPS

18
Summary

• Variational analysis methods are preferred
• Use all the data sources in defining atmospheric
  state
• Use error characteristics of the data
• SBN 8-bit image data can effectively be used in
  LAPS-AWIPS for analysis
• Sounder 13-bit data enables better vertical
  resolution