1. Michiko Masutani NOAA/NWS/NCEP/EMC RSIS/Wyle Information Systems Introduction to OSSE and Summary of NCEP OSSEs http://www.emc.ncep.noaa.gov/research/OSSE

2. Nature Run: Serves as a true atmosphere for OSSEs Observational data will be simulated from the Nature Run Nature Run

3. DATA PRESENTATION OSSE Quality Control (Simulated conventional data) OSSE DATA PRESENTATION Quality Control (Real conventional data) OSSE DA Real TOVS AIRS etc. DA Simulated TOVS AIRS etc. NWP forecast Existing data + Proposed data DWL, CrIS, ATMS, UAS, etc Nature Run Current observing system NWP forecast GFS OSSE

4. ● OSSEs help in understanding and formulating observational errors ● DA (Data Assimilation) system will be prepared for the new data ● Enable data formatting and handling in advance of “live” instrument ● OSSE results also showed that theoretical explanations will not be satisfactory when designing future observing systems. Need for OSSEs ♦Quantitatively–based decisions on the design and implementation of future observing systems ♦ Evaluate possible future instruments without the costs of developing, maintaining & using observing systems. Benefit of OSSEs

5. Full OSSE There are many types of simulation experiments. We have to call our OSSE a Full OSSE to avoid confusion. ● Nature run (proxy true atmosphere) is produced from a free forecast run using the highest resolution operational model. ● Calibration to compare data impact between real and simulated data will be performed. ● Data impact on forecast will be evaluated. ● A Full OSSE can provide detailed evaluations of the configuration of observing systems.

6. Summary of results from NCEP OSSE at NCEP Wintertime Nature run (1 month, Feb5-Mar.7,1993) NR by ECMWF model T213 (~0.5 deg) 1993 data distribution for calibration NCEP DA (SSI) withT62 ~ 2.5 deg, 300km and T170 ~1 deg, 110km Simulate and assimilate level1B radiance –Different method than using interpolated temperature as retrieval Use line-of- sight (LOS) wind for DWL –not u and v components Calibration performed Effects of observational error tested NR clouds are evaluated and adjusted

7. OSSE Calibration Compare real data sensitivity to sensitivity with simulated data Impact of withdrawing RAOB winds

8. Results from OSSEs for Doppler Wind Lidar (DWL) All levels (Best-DWL): Ultimate DWL that provides full tropospheric LOS soundings, clouds permitting. DWL-Upper : An instrument that provides mid- and upper- tropospheric winds down only to the levels of significant cloud coverage. DWL-PBL: An instrument that provides wind observations only from clouds and the PBL. Non-Scan DWL : A non-scanning instrument that provides full tropospheric LOS soundings, clouds permitting, along a single line that parallels the ground track. (ADM-Aeolus like DWL) Zonally and time averaged number of DWL measurements in a 2.5 degree grid box with 50km thickness for 6 hours. Numbers are divided by 1000. Note that 2.5 degree boxes are smaller in size at higher latitudes. Estimate impact of real DWL from combination.

9. ADM-Aeolus with single payload Atmospheric LAser Doppler INstrument ALADIN  Observations of Line-of-Sight LOS wind profiles in troposphere to lower stratosphere, up to 30 km with vertical resolution from 250 m - 2 km, horizontally averaged over 50 km every 200 km  Vertical sampling with 25 range gates can be varied up to 8 times during one orbit  High requirement for random error of HLOS LOS accuracy = 0.6*HLOS  Operating @ 355 nm with spectrometers for molecular Rayleigh and aerosol/cloud Mie backscatter  First wind lidar and first High Spectral Resolution Lidar HSRL in space to obtain aerosol/cloud optical properties (backscatter and extinction coefficients) Atmospheric Dynamics Mission ADM-Aeolus

10. Doppler Wind Lidar (DWL) Impact without TOVS (using 1999 DAS) Time averaged anomaly correlations between forecast and NR for meridional wind (V) fields at 200 hPa and 850 hPa. Experiments are done with 1999 DAS. CTL assimilates conventional data only. 3 3 8 8 No Lidar (Conventional + NOAA11 and NOAA12 TOVS) Scan_DWL_Lower Scan_DWL_upper Non_scan_DWL Scan_DWL Synoptic Scale Total scale 200 mb V 850 mb V

11. Forecast hour Doppler Wind Lidar (DWL) Impact With TOVS (using 2004 DAS) 3 3 1.2 Dashed green line is for scan DWL with 20 times less data to make observation counts similar to non-scan DWL. This experiment is done with 2004 DAS. No Lidar (Conventional + NOAA11 and NOAA12 TOVS) Scan DWL_Lower Scan DWL Uniformly thinned to 5% Scan DWL_upper Non-scan_DWL Scan_DWL Synoptic Scale Total scale 200 mb V 850 mb V

12. 2004 DAS CTL (Conventional + TOVS) Data Impact of scan DWL 1999 DAS vs. 2004 DAS Scan DWL with 2004 DAS 1999 DAS CTL - CTL(99) Differences in anomaly correlation Synoptic Scale Total scale 200 mb V Scan DWL with 1999 DAS - 1999 DAS CTL

13. 2004 DAS CTL (Conventional + TOVS) Data Impact of scan DWL 1999 DAS vs. 2004 DAS Scan DWL with 2004 DAS 1999 DAS CTL - CTL(99) Differences in anomaly correlation Synoptic Scale Total scale 200 mb V Impact of 2004 DAS Impact of DWL Impact of DWL + better DAS Scan DWL with 1999 DAS - 1999 DAS CTL

14. T62 CTL (reference) (Conventional data only) Data Impact of scan DWL vs. T170 T62 CTL with Scan DWL T170 CTL T170 CTL with Scan DWL ◊ CTL X CTL+BestDWL - CTL(T62) Differences in anomaly correlation Synoptic ScaleTotal scale 200 mb V Synoptic Scale

15. T62 CTL (reference) (Conventional data only) Data Impact of scan DWL vs. T170 T62 CTL with Scan DWL T170 CTL T170 CTL with Scan DWL ◊ CTL X CTL+BestDWL - CTL(T62) Differences in anomaly correlation Synoptic ScaleTotal scale 200 mb V Impact of T170 model Impact of DWL Impact of DWL + T170 At smaller scales, adding DWL with scan is more important than increasing the model resolution At planetary scale increasing the resolution is more important than adding DWL with scan

16. Effect of Observational Error on DWL Impact Percent improvement over Control Forecast (without DWL) Open circles: RAOBs simulated with systematic representation error Closed circles: RAOBs simulated with random errors Green: Best DWL Blue: Non-Scan DWL 3 8 3 8 Time averaged anomaly correlations between forecast and NR for meridional wind (V) fields at 200 hPa and 850 hPa. Experiments are done with 1999 DAS. CTL assimilates conventional data only.

17. DWL-Upper: An instrument that provides mid- and upper-tropospheric winds only to the levels of significant cloud coverage. Operates only 10% (possibly up to 20%) of the time. Switched on for 10 min at a time. DWL-Lower: An instrument that provides wind observations from only clouds and the PBL. Operates 100% of the time and keeps the instruments warm as well as measuring low level wind DWL-NonScan: DWL covers all levels without scanning (ADM mission type DWL) Targeted DWL experiments (Technologically possible scenarios) Combination of two lidars

18. 200mb 100% Upper Level 10% Uniform DWL Upper 10% Upper Level Adaptive sampling (based on the difference between first guess and NR, three minutes of segments are chosen – the other 81 min are discarded) Doubled contour (Feb13 - Mar 6 average ) Non-Scan DWL

19. CTL No Lidar (Conventional + NOAA11 and NOAA12 TOVS) 100% Lower + 10% Uniform Upper 100% Lower + 100% Upper 100% Lower + 10% Targeted Upper 100% Lower + Non-Scan Non-Scan only 100% Lower Anomaly correlation difference from control Synoptic scale Meridional wind (V) 200hpa NH Feb13-Feb28 DWL-Lower is better than DWL-Non-Scan only With 100% DWL-Lower DWL-Non-Scan is better than uniform 10% DWL-Upper Targeted 10% DWL-Upper performs somewhat better than DWL-Non-Scan in the analysis DWL-Non-Scan performs somewhat better than Targeted 10% DWL-Upper in 36-48 hour forecast

20. OSSEs with Uniform Data More data or a better model? Fibonacci Grid used in the uniform data coverage OSSE Data and model resolution - Yucheng Song Time averaged from Feb13-Feb28 12-hour sampling 200mb U and 200mb T are presented Skill is presented as Anomaly Correlation % The differences from selected CTL are presented 40 levels of equally-spaced data 100km, 500km, 200km are tested

21. 1000km RaobT170L42 analT62L28 fcst CTL 1000km RaobT62L28 anal & fcst U 200 hPa 500km RaobT62L28 anal & fcst 500km Raob T62L64 anal T62L64 fcst 500km Raob T62L64 anal T62L28 fcst T 200 hPa Benefit from increasing the number of levels Increasing the vertical resolution was important for high density observations. High density observation give better analysis but it could cause poor forecast High density observations give a better analysis but could cause a poor forecast. 5

22. 1000km RaobT170L42 analT62L28 fcst CTL 1000km RaobT62L28 anal & fcst U 200 hPa 500km RaobT62L28 anal & fcst 500km Raob T62L64 anal T62L64 fcst 500km Raob T62L64 anal T62L28 fcst T 200 hPa Benefit from increasing the number of levels Increasing the vertical resolution was important for high density observations. High density observation give better analysis but it could cause poor forecast T170 L42 model L64 anl & fcst High density observations give a better analysis but could cause a poor forecast. L64 anl & fcst L64 anl L28 fcst 5 500km obs

23. Results from OSSEs for Doppler Wind Lidar (DWL) All levels (Best-DWL): Ultimate DWL that provides full tropospheric LOS soundings, clouds permitting. DWL-Upper : An instrument that provides mid- and upper- tropospheric winds down only to the levels of significant cloud coverage. DWL-PBL: An instrument that provides wind observations only from clouds and the PBL. Non-Scan DWL : A non-scanning instrument that provides full tropospheric LOS soundings, clouds permitting, along a single line that parallels the ground track. (ADM-Aeolus like DWL) Zonally and time averaged number of DWL measurements in a 2.5 degree grid box with 50km thickness for 6 hours. Numbers are divided by 1000. Note that 2.5 degree boxes are smaller in size at higher latitudes. Estimate impact of real DWL from combination.

24. Forecast hour Doppler Wind Lidar (DWL) Impact 3 3 1.2 Dashed green line is for scan DWL with 20 times less data to make observation counts similar to non-scan DWL. The results show definite advantage of scanning. This experiment is done with T62 model and SSI. No Lidar (Conventional + NOAA11 and NOAA12 TOVS) Scan DWL_Lower Scan DWL Uniformly thinned to 5% Scan DWL_upper Non-scan_DWL Scan_DWL Synoptic Scale Total scale 200 mb V 850 mb V

25. D2+D3: Red: Scan upper+scan Lower D1: Light Blue closed circle: Best DWL (D1) with scan Rep Error 1m/s R45: Cyen dotted line triangle: D1 with rep error 4.5m/s (4.5x4.5≈20) U20: Orange: D1 uniformly thinned for factor 20 (Note this is technologically difficult) N4: Violet:D1 Thinned for factor 20 but in for direction 45,135,225,315 (mimicking GWOS) S10: Green dashed: Scan DWL 10min on 90 min off. No other DWL D4 : Dark Blue dashed: non scan DWL One or Four non-scan-DWL NH V500 Zonal wave number 10-20

26. GWOS is better than ADM-Aeolus in the SH, Zonal wave 1-3 T W GWOS Like lidar ADM U V D2+D3: Red: Scan upper+scan Lower D1: Light Blue closed circle:scal-DWL repE1m/s R45: Cyen dotted line triangle: repE 4.5m/s U20: Orange: uniformly thinned to 5% N4: Violet:Four direction S10: Green dashed:Scan DWL 10min on 90 min off. D4 : Dark Blue dashed: non scan DWL

27. Scan-DWL has much more impact compared to non- scan-DWL with same amount of data. If the data is thinned uniformly 20 times thinned data (U20) produce 50%-90% of impact. 20 times less weighted 100% data (R45) is generally slightly better than U20 (5% of data) In fact U20 and R45 perform better than D1 time to time. This is more clear in 200hPa. Simple GWOS type experiment (N4) showed significant improvement to D4 only in large scale over SH but not much better over NH and synoptic scale. Without additional scan-DWL,10min on 90 off (S10)sampling is much worse than U20(5% uniform thinning) with twice as much as data.

28. Using T62 model forecast skill drop after 48 hr but this will improved with T170 particularly in wave number 10-20 The results are expected to change with GSI particularly with flow dependent back ground error covariance. OSSE with one month long T213 nature run is limited. Need better nature run. All results have to be confirmed by Joint OSSEs with Joint OSSE NR

29. END