Introduction to the Gridpoint Statistical Interpolation GSI Analysis

1
Introduction to the Gridpoint Statistical
Interpolation (GSI) Analysis

• Global Modeling Branch Briefing
• Thursday, 07 July 2005

2
Overview of GSI

• 3d variational assimilation based on the SSI . .
  .
• Formulated in physical space
• Analysis (Control) Variables
• ?, ?u, Tu, q, Psu, O3 (not ?, ?)
• Multivariate relation differs (non-trivial to use
  things like full balance equation on a grid)
• Grid point definition of background error
• Spectral definitions replaced with recursive
  filters
• Vertical EOFs replaced with recursive filters
• All statistics are a function of height and
  latitude
• Improved efficiency, coding, and documentation

3
Multivariate Relation (balance)

• Tb G? ?b c? Psb w1? w2 ?u

Projection of ? at vertical level 25 onto
vertical profile of balanced temperature (G25)
Percentage of full temperature variance explained
by the balance projection
4
Background Error Covariance (?)
a
b
c

• Standard deviation of variance (interval 0.5 x
  106 m2 s-1)
• Horizontal length scale estimate (interval 100
  km)
• Vertical length scale estimate (interval 1 grid
  unit)

5
Single Observation Analysis
Single zonal wind observation (1.0 ms-1 O-F and
error)
u increment at ?0.3 (interval 0.1 ms-1 ) from SSI
u increment at ?0.3 (interval 0.1 ms-1 ) from GSI
6
Single Observation Analysis
Cross Section at 180o
u increment (black,interval 0.1 ms-1 ) and T
increment (color, interval 0.02 K) from SSI
u increment at (black, interval 0.1 ms-1 ) and T
increment (color, interval 0.02K) from GSI
7
Sigma T62L64 Cycling/Tuning
GFS 500 hPa Day 5 Anomaly Correlation
8
T254L64 Cycling (Initial Hybrid Tests)
GFS 500 hPa Anomaly Correlation (25 cases from
August 2004)
9
T254L64 Cycling (Initial Hybrid Tests)
GFS Tropical Wind Vector RMS Error
10
T382L64 (768x384) GSI/SSI Comparison
Analysis Valid at 00Z 03 July 2005Both
experiments started 10 May 2005
11
Single Analysis Increment Comparison
12
GSI/SSI Guess and Analysis Difference Fields
13
Operations vs GSI Hybrid Parallel
GFS 500 hPa Day 5 Anomaly Correlation
14
Operations vs GSI Hybrid Parallel
GFS Tropical Wind Vector RMS Error
15
Thoughts/Work in Progress . . .

• We have yet to really begin to take advantage of
  being in physical space
• Situation dependent/anisotropic background error
• With considerable tuning/experimentation, the GSI
  seems to perform as well as SSI. . . at lower
  resolution
• GSI performs well relative to SSI in terms of
  tropical RMS at all resolutions (impact of
  latitude-dependent balance)
• GSI generally yields worse scores in the
  extratropics (dropouts seem to be larger)
• Balance Issue?
• Attempting to add a dynamic constraint term to
  complement the statistical balance (work in
  progress)
• Further evaluation of GSI parallel needs to be
  done (with help from EMCers and new real-time
  parallel)