1. Status of the Regional OSSE for Space-Based LIDAR Winds – Feb01
John McGinley
Adrian Marroquin
John Smart
Linda Wharton
NOAA/Forecast Systems Lab
10/25/2019

2. Goal of OSSE
Assess impact of space-borne LIDAR winds on regional weather forecasts
Compare to the Global LIDAR OSSE
10/25/2019

3. Collaborators NOAA/ETL NOAA/NWS/NCEP LIDAR Simulation/ Data
Background grids
Advice
10/25/2019

4. Sub-Tasks
Nature run with independent model – MM5 – for 9-day period Feb 5-13, 1993
Archival and storage of grids
Data extraction
Creation of “operational” data base (progress to here)
Ingest of data by RUC assimilation system
Continuous assimilation and forecasts for test period
Validation of LIDAR, no-LIDAR runs with control run
Write Report
10/25/2019

5. Flow Diagram RUC Standard Obs Extractor LIDAR OBS Simulator ACARS
Obsv Metadata
Satellite
Track Data
ACARS Track
Data
Flow Diagram
Standard Obs
Extractor
LIDAR OBS
Simulator
ACARS
Simulator
MM5
Nature Run
ECMWF Bkgrnd
Control Archive
Std Data Files
LIDAR
Data Files
ACARS
Data Files
RUC
Validation
Scheme
Results
RUC Forecast Fields
NCEP w/
Ideal LIDAR
NCEP Bkgnd
wo/ LIDAR
NCEP Bkgnd
w/ LIDAR
10/25/2019

6. Nature Run Case 5-15 Feb 1993 MM5 Model version 3.0
Non-Hydrostatic, complete cloud microphysics, radiation, surface processes
740 x 520 x 31 – 10km grid – 13 variables
Output every 15 minutes – 874MB: 115 GB/day (flat files)
Computer: JET - HPTi MPP (256 processor – 42 are used)
Integration time: ~ 0.8 forecast time
5-9 Feb 1993 – Spin-up
10-15 Feb – Output and archive every 15-min
10/25/2019

7. Domain
10/25/2019

8. Nature run on 9 Feb – at 5km
Satellite vis at 9/18Z
10/25/2019

9. Nature run on 9 Feb – at 1.6km
Satellite vis at 9/18Z
10/25/2019

10. Observation Extraction
Fixed Site: RAOB, Surface, Wind Profiler, Radar VAD
utilize obs generators from 1996 OSSE for KSC and VDB
Mobile Platforms: ACARS, LIDAR
Specify aircraft tracks time and space
Specify 8-position sampling beam in time and space
Provide wind components, cloud species, droplet spectra
Modeled Error
Random instrument error
Correlated meteorological noise
Mean and Spatial Variance
Data Files
netCDF – identical to FSL database
10/25/2019

11. Observation Processing
10/25/2019

12. Simulated Random Errors
Surface
RAOB
VAD WSR88D
PROFILERS
ACARS
Windspeed
0.5m/s
3.1 m/s
2 – 4 m/s
u,v comp
2.5 m/s
1.5 m/s
Direcetion
2.5 deg
2 –14 deg
Fcn spd
5 deg
Temp
0.5 C
Dewpoint
0.75 C
3.0 C
INSTs
DATA
10/25/2019

13. Data Assimilation Use RUC 20-km model run at FSL
Ingest all simulated fixed-site observations
Ingest ACARS and LIDAR radial winds
Three data options:
Standard set + LIDAR
Standard set only
Standard + Idealized LIDAR
Three background/ boundary options:
NCEP background derived from LIDAR assimilation
NCEP background, no LIDAR
NCEP background with Idealized LIDAR
Nature Run initialization and boundary conditions
10/25/2019

14. Data Assimilation Table
NCEP Global w/o LIDAR
NCEP
With LIDAR
With Ideal
LIDAR
Yes
Standard – no LIDAR
Maybe
Standard
+ LIDAR
Standard + idealized
BC’s
Nature
Run
DATA
Background
Only
10/25/2019

15. Assimilation Timeline
Hourly Assimilation and
24 Hour Forecasts
Spin Up
Feb
1993
10/25/2019

16. Validation Procedures
Compare RUC Forecasts to Nature Run
Compute errors in state variables
Assess precipitation skill
10/25/2019

17. RUC Data Assimilation and Verification
Schedule
Nature Run
Conventional Data Extraction /Preparation
LIDAR Data Extraction/ Preparation
RUC Data Assimilation and Verification
Report
Preparation
Jan Feb Mar Apr May Jun Jul Aug Sep Oct 01
10/25/2019

18. Current Status and Accomplishments
Nature Runs and Database
Prepared ECMWF Background – Gaussian to lat/lon
Converted to MM5 from RAMS
Set up MM5 v 3.0 on Jet
Completed 48-hour run for 5-6 Feb 1993 for data extraction
Provided to FSL, ETL, and Simpson Weather Associates
Coordinated format for LIDAR beam position with ETL
Coordinated model parameters needed for simulating LIDAR obs
Modified MM5 to compute mean and variance of all LIDAR fields – tested with convective case 28 Jun 1999
Archived representative ACARS tracks
Completed Nature Run for 5-15 Feb 1993
10/25/2019

19. Immediate goals (next few weeks)
Complete Standard Observation Processing
Obtain Satellite Tracks from ETL
Pass Data to ETL for LIDAR simulation
Run RUC forced by Nature Run only
Run RUC with all “no-LIDAR” Scenarios
10/25/2019

20. Problems Control Runs and Database
Had to convert to MM5 from RAMS – RAMS not runnable on Jet (DEC-Alpha environment)
Had to develop code for estimating mean and variance of LIDAR-measured quantities and convince ourselves that grid point variance could be related to spatial variability
Early Jan – Found ECMWF Background had incorrect u,v components – 2 day and 11-day simulations had to be rerun
10/25/2019

21. Ongoing Issues and Concerns
Global Nature Run low resolution
Nature Run clouds and climatology
10/25/2019

22. Nature run cloud water – OK?
10/25/2019

23. Features seem to compare well
Satellite vis at 9/18Z
10/25/2019

24. Summary Nature Run completed Conventional data files being prepared
(50% finished)
Data assimilation experiments to begin in Feb.
10/25/2019