Precipitation Error Characterization over the Global Oceans

1
Precipitation Error Characterizationover the
Global Oceans

• 16 June 2004

2
Oceanic Precipitation League

• Yuter, Wood, Horn - University of Washington
• Wilheit - Texas A M University
• Sobel - Columbia University
• Kingsmill University of Colorado
• Foufoula-Georgiou University of Minnesota
• Chandrasekar Colorado State University
• Braun NASA Goddard Space Flight Center

3
3-Pronged Approach

• Bridge analysis of observations and refinements
  of algorithms
• GV site routine products
• Focused measurements for physical validation
• Global error characterization product

4
Potential Oceanic GV Supersites
5
Error CharacterizationDevelopment Challenges

• Has not been done before
• User expectations vague
• Difficult to define requirements

6
Prototype Advantages

• End-to-end skeleton working early
• Test drive and refine concepts
• Frequent feedback from users
• Functionality added incrementally

7
Prototype Global Oceanic Error Characterization

• Input TRMM TMI and PR (Vers. 5)
• 1B11, 1C21, 2A12, 2A23, 2A25
• Oceanic rain certain pixels
• Grids of 2000 x 2000 km2 oceanic area
• Daily product based on accumulated statistics for
  previous 30 days
• Compare
• Probability distributions
• Statistical variables

8
(No Transcript)
9
Global Error Product Will Address

• How different?
• Relative error statistics
• Why Different?
• Diagnostic information
• Target User Groups
• Forecasters
• Algorithm Refinement
• Data Assimiliation

10
GSFC PPS
End User
Satellite Error Characterization
GPM Orbit Products
Daily 1 day accum 30 day products
OPL Viewer
GV Precip Characterization
GV Routine Products
Info for Data Assim.
Models
Distilled Info
Focused Measurements
GV Data Processing
11
Prototype Will Compare

• Spatial scales
• PR and TMI native resolutions
• PR rescaled to TMI 10 GHz scale (30 x 60 km2)
• Data types
• Surface Rainrate (Rsfc)
• Vertically-Integrated Precipitable
  Liquid Water Content
• Emission Tb
• Rainy Area

12
Conditional Rsfc PR TMI
13
Conditional Rsfc PR TMI
48
13
14
Tropical West Pacific near Kwajalein
15
Kwajalein GV radar 37,388,868 4 km2 pixels
Number of Pixels
Log10(R)
16
South Central Pacific
17
PR pixels rescaled to TMI 10 GHz scale Rainy
Area of 30 km x 60 km pixels
18
Rainy area of 10 GHz scale pixels
19
Need Feedback
20
Conclusions

• Global Error Characterization Prototype Skeleton
  Working
• PR vs TMI Rsfc Bias Varies Regionally and
  Temporally
• Differences between PR and TMI
• Rain/No Rain Screening
• Min Rsfc Threshold
• Can Impact Level 3 Products
• Rainy Area of TMI 10 GHz scale pixels
• Regional Variations
• Strongly Skewed to Small

21
The End
22
EFOVs
Satellite subpt 15 km between swaths
EFOV
IFOV start
IFOV end
Along-track axis
PR Swath (245 km)
TMI Swath (872 km)
23
(No Transcript)
24
PR Rainy Pixel Rsfc Mean
25
TMI Rainy Pixel Rsfc Mean
26
PR-rescaled Rsfc Average