1. OSE/OSSE activities at NOAA ESRL Global Systems Division
Lidia Cucurull (*)(**)
Chief, Global Observing Systems Analysis (GOSA) Group
Ruifang Li(*), Hongli Wang(*), Jason English(*),
Tanya Peevey(*), and Kirk Holub(*)
(*) NOAA ESRL Global Systems Division, Boulder, CO
(**) NOAA AOML, Miami, FL
13th JCSDA Technical Review Meeting & Science Workshop on Satellite Data Assimilation
College Park, MD, May 2015

2. Outline
Introduction to the Global Observing Systems Analysis (GOSA) Group
GOSA activities
Observing System Simulation Experiments (OSSEs)
Observing System Experiments (OSEs)
Support to the COSMIC-2 Program (GNSS Radio Occultation mission)
NOAA ground-based GPS observations (GPS-Met)
Summary
COSMIC-2

3. GOSA Director’s Office NOAA OAR ESRL Global Systems Division
Kevin Kelleher, Dir.
Jennifer Mahoney, Deputy
NOAA OAR ESRL
Global Systems Division
Research Support Services (A&R)
Information Tech. Services (ITS)
Global Observing Systems Analysis (GOSA) Group
Chief: Lidia Cucurull
GSD Senior Scientist DTC Science Advisor
GOSA
Earth Modeling Branch (EMB)
Acting Chief: Stan Benjamin
Deputy: Tim Schneider
Advanced Technology and Outreach (ATO) Branch
Chief: John Schneider
Evaluation and Decision Support (EDS) Branch:
Chief: Mike Kraus
Assimilation
Model Physics
EV&O
HPC
FIQAS
Forecast System Evolution
Global Modeling
Earth System/Chem
I&V
IWDS
Decision Support
Systems Support
Model Assessment

4. GOSA Group Recently formed – 1 October 2014
Maximize and optimize the uses of current and future global observations to improve numerical weather prediction forecast skill in NOAA’s models. This includes algorithm development, management, and science support
Quantitatively evaluate the complementarity of different observing systems through OSSEs and OSEs to help NOAA management prioritize mission designs in a cost-effective way
Why is GOSA so important? OSSEs save Taxpayer $$ as they allow to analyze tradeoffs in the design and configuration of proposed observing systems (e.g. coverage, resolution, accuracy and data redundancy)

5. (1) OSSEs: with RO
Collaborative effort with NOAA AOML, JCSDA, NASA GMAO, and other partners
Investigate the benefits of increasing the number of assimilated GNSS RO profiles on global weather forecast skill
NCEP’s Global Data Assimilation System (GSI/GFS)
Focus on current and future RO constellations, including commercial (e.g., GeoOptics, PlanetIQ) and non-commercial options (COSMIC-2)
Conduct preliminary OSSEs with a sub-optimal configuration set up (completed)
Follow-up with more rigorous OSSEs (ongoing effort to be completed by Dec 2015)
Configuration setup for preliminary OSSEs:
T511 ECWMF Nature Run (~ 40 km)
NCEP’s global data assimilation system is T382 with non-hybrid GSI analysis – operational configuration is T1534 with hybrid GSI
RO observation is refractivity – operational configuration assimilates bending angle, a less derived product
Assimilation of RO profiles stops at 30 km – operational configuration assimilates up to 50 km
All observations are assumed error-free except for satellite radiances, which have biases added on

6. Preliminary OSSEs with COSMIC-2NH SH
-Little impact from full C2
-C2-EQ makes things slightly worse
-Results are not statistically significant
-Significant positive impact from C2-EQ and full C2
-Results are statistically significant
Forecast hour
OSSENOGPS: control without RO observations (0 RO satellites)
OSSECTRL: control, all observations (6 RO satellites)
C2EQ: control + COSMIC-2 equatorial (12 RO satellites)
C2PO: control + COSMIC-2 equatorial + COSMIC-2 polar (18 RO satellites)

7. Preliminary RO OSSE results
Overall, increasing the number of assimilated RO satellites from 6 to 18 results in better weather forecast skill: 18 satellites is better than 12 satellites; 12 satellites is better than 6 satellites
There is a lot of room to optimize the assimilation system
Results are currently being investigated with more detail
These preliminary studies do not use the state-of-the-art OSSE system
Repeat experiments with a state-of-the-art OSSE configuration (ongoing effort to be completed by Dec 2015)
Higher resolution Nature Run (7 km GMAO NR)
Random errors added to the observations
Hybrid GSI in DA system
Higher horizontal resolution in DA system
Bending angle forward operator instead of refractivity operator
Extending the top of the RO profiles from 30 to 50 km

8. Example of Simulated bending angle profile
NCEP Advanced
Bending Angle Model
(NABAM)
Ruifang Li

9. (1) OSSEs: with UAS
Sensing Hazards with Operational Unmanned Technology (SHOUT) to Mitigate the Risk of Satellite Observing Gaps
Perform OSSEs to:
Evaluate the impact of current and future instruments on the Global Hawk in terms of global forecast skill
Help selecting instruments to fly in future flight campaigns, including the use of targeting observation techniques
Focus on mid-latitude (winter) storms
T511 Nature Run from ECMWF – OSSE capability with 7 km GMAO NR is an ongoing effort and synthetic observations are currently only being generated for a summer period
Make use of the NCEP’s Global Data Assimilation System (GDAS) at a lower resolution (T382)
Conduct Observing System Experiments (OSEs) with real FY15 and FY16 flights:
Operational version of NCEP’s GDAS
Real observations

10. Hongli Wang, Tanya Peevey, Francisco Alvarez, Jason English
CON NR SA SB
When examining the storm system south of Alaska it is clear that both SA and SB are closer to the NR compared to CON. This location and over Alaska and western Canada is where improvements seem to be most significant. Specifically, the trough in the CON is due west of that found in the NR, resulting in the CON forecast both underestimating and overestimating the wind speeds at 250 mb.
Hongli Wang, Tanya Peevey, Francisco Alvarez, Jason English

11. (2) OSEs: Complementarity of satellite systems
Investigate the impacts on the skill of NCEP global forecasts due to a loss of the NOAA and AQUA MW and all RO soundings
Two extreme scenarios are considered
MW instruments from NOAA-15/18/19 and AQUA have not reached the end of their life before JPSS-1 is launched
MW instruments from NOAA-15/18/19 and AQUA are lost before JPSS-1 is launched
Evaluate the impact of losing all RO observations in both extreme scenarios
Results of the study might be considered pessimistic (worse case) as all systems are not likely to fail before there are some replacements
Extreme cases will indicate relative importance of these losses – will provide significant signal
Cucurull L., and R. A. Anthes, 2015: Impact of Loss of Microwave and Radio Occultation Observations in Operational Numerical Weather Prediction in Support of the US Data Gap Mitigation Activities, Wea. Forecasting, 30, 2,

12. AC 250-mb Temperature Largest benefit from MW, but small differences
Largest benefit from RO, and larger
differences
500-mb, AC gph, NH
Removing only RO makes the least difference and removing both AMSU-A and ATMS makes the largest difference
The relative impact of ATMS and AMSU-A varies with forecast time and the impact of each system is positive compared to ROOnly
ATMS may mitigate to some extent the loss of other MW sounders
Differences are not statistically significant
500-mb, AC gph, SH
Worst two scenarios include the loss of RO out to 7 days, and results are stats. significant out to 3 days, but not beyond
ATMSOnly shows the lowest score and removing ATMS improves the forecasts slightly – but not stat. significant

13. Fit to Radiosonde (Temperature)
MW tend to increase biases, while RO
tends to reduce it
The impact of ATMS is mostly neutral

14. (3) Support to COSMIC-2
GOSA has a formal agreement with NWS and NESDIS to support GNSS RO research and associated R2O
The agreement includes Chief Scientist, RO monitoring, Cal/Val, and algorithm development activities
Algorithm development work currently focuses on improving the assimilation of RO observations in the lower troposphere
NBAM (NCEP’s bending angle method) is used to assimilate bending angle observations in the operational configuration
An upgraded version of NBAM (NABAM, “A” for “Advanced”) is being developed to better address rays that cross the lower troposphere, in particular in the moist tropical troposphere
We must address this issue in the GSI in preparation for the large amount of observations that COSMIC-2 and other GNSS RO missions will bring

15. NBAM vs NABAM

16. Summary
GOSA Group has been established (in coordination with NOAA’s QOSAP Program) to coordinate OSSE efforts to help NOAA management prioritize mission designs in a cost-effective way
Both OSEs and OSSEs are necessary to quantitatively evaluate the benefits of observations in weather forecasting
OSEs allow the evaluation of the impact of current observations in weather forecasting
OSSEs provide a rigorous, cost-effective approach to evaluate the potential impact of new observing systems, alternate configurations and deployments of existing systems, and to optimize observing strategies.
GOSA is engaged in ongoing R&D to quantify impacts of additional GNSS RO measurements on operational forecast models (e.g., what is the saturation point?)
GOSA has a formal agreement with NWS and NESDIS to support GNSS RO research and associated R2O (in support of the COSMIC-2 Program)
GOSA is transitioning its ground based GPS application to the private sector