1. Challenges in Using GOES Data Within Operational Numerical Models Dr. Louis W. Uccellini Director National Centers for Environmental Prediction NASA Science Community Workshop on Polar Orbiting IR and MW Sounders Greenbelt, MD November 1, 2010

2. 2 Background: Historical Perspective in the use of LEO/GEO Data for Forecast Application Summary Overview

3. 3 Background: Historical Perspective in the use of LEO/GEO Data for Forecast Application

4. 4 GEO Principally used by forecast community: became the “eyes” for the forecaster –Timely access a key Sheets, R.C., 1990. The National Hurricane Center – Past, Present, and Future. Wea. Forecasting, 5: 185-232. “The greatest single advancement in observing tools … was unquestionably the advent of the geosynchronous meteorological satellite. If there was a choice of only one observing tool for use in meeting the responsibilities of the NHC, the author would clearly choose the geosynchronous satellite with its present day associated accessing, processing, and displaying systems …”

5. 5 LEO Higher spatial resolution The backbone of the Global Observing System for numerical models Was not readily available to the forecaster in “real time”

6. 6 Today Distinction is becoming blurred –LEO data is more readily available in “real time” for forecaster applications –GEO is used quantitatively in numerical models (winds, land surface, SST) NOAA-18 Image of CA fires GOES Sounder-Derived Precipitable Water

7. 7 QuikSCAT Hurricane Force Extratropical Cyclone kts Geostationary IR Intense, non-tropical cyclones with hurricane force winds Feb 09, 2007, North Atlantic

8. Hurricane Force Extra-tropical Cyclones Detection and Warning Trend using QuikSCAT 2000-2009 8 Hurricane Force Warning Initiated Dec 2000 Detection increased with: -Forecaster familiarity -Data availability -Improved resolution -Improved algorithm QuikSCAT Launch Jun 99 Hurricane Force Wind Warning Initiated Dec 00 25 km QuikSCAT Available in N-AWIPS Oct 01 12.5 km QuikSCAT available May 04 Improved wind algorithm and rain flag Oct 06 Totals A-289 P-269 558

9. 9 The SST (top) and OHC (bottom) in the pre-storm environment for Hurricane Katrina. The storm intensity and positions from the NHC best track are indicated by the circles. Mainelli, et. al., 2008, Wea. Forecasting, 23, 3-16 OHC: Ocean Heat Content derived from oceanic T and salinity climatology, SST analysis, and radar altimetry sea height anomaly (SHA) fields created from a blend of Jason-1 data and the Geosat Follow-On (GFO)

10. Record Values 10

11. Impact with COSMIC AC scores (the higher the better) as a function of the forecast day for the 500 mb gph in Southern Hemisphere 40-day experiments: –expx (NO COSMIC) –cnt (old RO assimilation code - with COSMIC) –exp (operational code - with COSMIC) COSMIC provides 8 hours of gain in model forecast skill starting at day 4 !!! Cucurull et al., 2010, WAF

12. Skill score dropouts impact NCEP’s global model performance in Northern and Southern Hemispheres Dropouts are defined by 5-day anomaly correlation (AC) scores < 0.70 For example, the 00Z Feb. 03 2008 case, using GPSRO data alleviated a dropout in the Southern Hemisphere. Looking into lack of bias in COSMIC as important influence on data analysis (Courtesy of DaNa Carlis, NCEP) SH 5-day AC scores: GFS=0.65 (NCEP’s model) GDAS=0.69 ECMWF=0.83 First guess+nodata=0.70 First guess+conven=0.68 First guess+conven+amsua=0.70 First guess+conven+airs=0.75 First guess+conven+amsub=0.77 First guess+conven+mhs=0.78 First guess+conven+gpsro=0.79 First guess+conven+mhs+amsub=0.78 First guess+conven+gpsro+mhs+amsub=0.87 COSMIC also Produces Positive Impact on “Dropout” Case COSMIC capable of alleviating ‘dropouts’ in the Southern Hemisphere

13. 13 Mean(RSS ‐ RO_AMSU) = ‐ 0.99 Std (RSS ‐ RO_AMSU) =1.67 R=0.99 Mean(UAH ‐ RO_AMSU) =0.02 Std (UAH ‐ RO_AMSU) =2.06 R=0.99 Mean(NOAA ‐ RO_AMSU) =-0.49 Std (NOAA ‐ RO_AMSU) =0.5 R=0.99 UAH gives smallest mean TLS difference (0.02 K) compared to COSMIC GPSRO adjustment NESDIS’ SNO produces smallest std dev (0.5 K) compared to COSMIC Adjustment Two years (2006-2008) of COSMIC TLS: -- too short to detect temperature trend -- provides insight on constructing climate data records from longer time series from other data records (Figure Courtesy B. Ho and C.Z. Zou) RSS UAH SNO

14. COSMIC-II Data Distribution Comparison of sounding distribution over three hour periods between COSMIC and COSMIC-II is shown. 8000-12000 profiles per day Average profile within 45 minutes Full vertical profile All weather Day and night No instrument drift or calibration Global coverage 0.001 TEC Unit relative Electron Density Profile 10% S4 index uncertainty – 0.1 Could evolve to 15-20 min max latency with satellite to satellite links

15. Summary 15