An%20Examination%20of%20the%20Diurnal%20Cycle%20in%20the%20NCEP

1
An Examination of the Diurnal Cycle in the NCEP
GFS (and Eta) Model Precipitation Forecasts
(during NAME)
John Janowiak, Valery Dagostaro, Vern Kousky,
Bob Joyce Climate Prediction Center/NCEP/NWS,
Camp Springs, MD Meteorological Development
Laboratory/NWS, Silver Spring, MD
30th Climate Diagnostics and Prediction Workshop,
Penn State University, October 26, 2005
2
MODEL FORECASTS
Study Period July 12 August 15, 2004
(during NAME field campaign)
Every 3 hours after 1st 12 hours from 00Z runs
(12h, 15h, , 36h) 1o 1o x lat/lon grid
3
VALIDATION DATA

• A new member of the MORPH family RMORPH
  (Research-quality)
• Land CPC daily gauge analysis (Higgins Shi)
  disaggregated by CMORPH
• satellite estimates
• Ocean CMORPH only
• Interpolated to match space-time resolution of
  model forecasts

4
VALIDATION DATA

• A new member of the MORPH family RMORPH
  (Research-quality)
• Land CPC daily gauge analysis (Higgins Shi)
  dissaggregated by CMORPH
• satellite estimates
• Ocean CMORPH only
• Interpolated to match space-time resolution of
  model forecasts

5
RMORPH Gauge Comparison
6
Precipitation Difference from Validation Data
7
Difference in Frequency () by Intensity
Eta - RMORPH
GFS - RMORPH
Percentage Difference
8
Distribution of Heavy Rainfall Events
9
(Land only)
10
Mean 3-hour Precipitation
11
HARMONIC ANALYSIS Applied to period mean
diurnal cycle
12
HARMONIC ANALYSIS Applied to period mean
diurnal cycle
13
(No Transcript)
14
of Daily Mean Precipitation _at_ 35oN
15
Precipitation _at_ 35oN
16
CONCLUSIONS

• Peak daily rainfall 3-6 hrs too early in Eta (SE
  US)
• and GFS (eastern U.S.)
• Better phase agreement over NAME Tier-1
• Models propagate rainfall well (central US -gt
  east )
• Light rain rate OVERforecast in the models
• heavy events UNDERforecast