1. Presented by: Vijay Tallapragada Chief, Global Climate and Weather Modeling Branch Based on Work Done by EMC DA Team, Land Surface Team and GCWMB 2015 NCEP Production Suite Review GLOBAL MODELING 1

2. Table of Contents 1)The GDAS/GFS is being upgraded to 4D- Hybrid En-VAR System 2)Land surface improvements to address summer time warm/dry biases in surface fields 3)Hourly output fields through 120-hr forecasts 4)New model upgrade evaluation strategy early in the process: do not wait until last 30-day NCO parallels! 5)Global plans for 2016+ 2

3. Next GFS/GDAS in 2016 The 4D Hybrid En-Var 4-D hybrid All-sky AMSU-A Radiances SATWND ob changes CRTM v2.2.1 Aircraft ob changes Modified thinning/weight in time Bug fixes and optimization for GSI New ob monitoring Upgrade data assimilation monitoring package t-3 t=0 t+3 4D increment constructed by figuring out best combination of 4D ensemble perturbations Weights constant throughout window Use temporal correlations within each member to extract time information in obs 4D Schematic 3

4. 4 New All-sky AMSU-A Radiances and Other Model Changes All-sky AMSU-A Radiances Situation-dependent observation error inflation; AMSU-A obs error re-tuned; All-sky radiance bias correction strategy (Zhu et al. 2014) Additional QC: cloud effect (Geer et al. 2013) and emissivity sensitivity screening; Normalized cloud water control variable; New static background error variance and correlation lengths for cloud water; Only initial implementation of this capability More data coverage: Thick clouds that are excluded from clear-sky assimilation are now assimilated under all-sky configurations Modified relocation/tracking to allow hourly relocation AVHRR satellite winds and aircraft moisture data are also assimilated. Forecast Model and Product Changes Semi-implicit upgrade, Convective gravity wave upgrade, Tracer adjustment upgrade Corrections to land surface to reduce summertime warm, dry bias over Great Plains Improved icing probability products and new icing severity product Hourly output through 120-hr forecast

5. Addressing Warm, dry bias over Great Plains late July and early August many complaints from field offices GFS too warm and too dry particularly over the Great Plains. noticed last summer in the parallel GFS some parameters were refined to address the issue. Change rsmin for grassland and cropland from 45 to 20 Extensive sensitivity tests --refinement significantly reduced the warm/dry biases too late to put into the last January GFS implementation. further tests conducted with proposed parameter changes. huge, significant improvement in 2m T, 2m Td, 10 m winds over Great Plains, midwest, Southeast, both in bias and rms error against observations, some improvement Northeast --little impact elsewhere in west, Alaska T 2m Td 2m 5

6. New Model Upgrade Evaluation Strategy 6 GCWMB real time (pr4devb) period: 2015070100 - real time2015070100 GCWMB 2015 summer retrospective (pr4devbs15) period: 2015041500 - 2015120100 (230 days)20150415002015120100 GCWMB 2013 summer retrospective (pr4devbs13) period: 2013041500 - 2013120100 (230 days)20130415002013120100 NCO 2013-2014 winter retrospective period: 2013110100 - 2014060100 (212 days)20131101002014060100 NCO 2014 summer retrospective period: 2014050100 - 2014120100 (214 days)20140501002014120100 GCWMB 2014-2015 winter retrospective (pr4devbw14) period: 2014110100 - 2015070100 (242 days)20141101002015070100 GCWMB Special retrospective for H. Sandy period: 2012101700 - 201213100 (15 days)2012101700201213100 Involve field in real-time and retrospective evaluation of science upgrades Identify case studies and provide data for extended evaluation period beyond last 30-day parallel NCO 30-day parallel is only for IT evaluation Completion of retrospective runs-Feb. 1, 2016 EMC testing, external evaluation complete—Feb. 19 EMC CCB approval—Feb. 22 30-day IT evaluation—April Followed by operational implementation

7. Hurricane tracks through day 7 with statistical significance & tropical cyclogenesis verification EMC producing gempak files from real time parallel; Western Region and WPC will use for side by side maps using MAG Synoptic and daily precip verification for real time parallel, GFS soundings, near surface variables at representative stations for GFS, GEFS and GFSX on EMC web pages Continuity objective score-needs long term development, first attempt available G2O (near surface verification) for all 4 cycles (currently plotted for 0z only) Jet stream MODE verification—Tracey Dorian Dec. 17 at MEG CAPE—working on MODE verification Retrospectives—Standard verification page—against own analyses, fit to obs, g2o, hurricanes, precip scores (ETS, MODE?) Evaluation plans for Q3FY16 GFS/4D-Hybrid EnVar Case studies Hurricane Sandy Case studies from Western, Central and Southern regions WPC—unspecified—high impact cases Alaska—Atsani, other tropical extratropical transitions Examine time-means, systematic errors Subjective assessment of PBL structure and other fields by MEG, Revisit MEG cases Where to find the evaluation results: http://www.emc.ncep.noaa.gov/gmb/wd20rt/vsdb/pr4devbs13/ http://www.emc.ncep.noaa.gov/gmb/wd20rt/vsdb/pr4devbs14/ http://www.emc.ncep.noaa.gov/gmb/wd20rt/vsdb/pr4devbs15/ http://www.emc.ncep.noaa.gov/gmb/wd20rt/vsdb/pr4devbw13/ http://www.emc.ncep.noaa.gov/gmb/wd20rt/vsdb/pr4devbw14/ http://www.emc.ncep.noaa.gov/gmb/STATS_vsdb/ More links will be made available as we add them…… stay tuned… 7

8. --EMC responds very quickly to near surface problems over CONUS --land surface changes—reduces summertime warm/dry bias BUT doesn’t correct diurnal cycle --new GFS significant improvements over operational GFS in short range forecasts --hurricane tracks a concern— including at days 6 and 7 add estimates of statistical significance --forecasts of tropical cyclogenesis being assessed --lot to assess—need to engage community -- MDL early evaluation indicated “no harm done”, no need for immediate recalibration of MOS Highlights of Q3FY16 GFS/GDAS Evaluation Results 8

9. 2015/16 Global Implementations GFS T1534 (~13 km) GCWMB – January 2015 GEFS (higher resolution/new perturbations) Yuejian Zhu – December 2, 2015 CFS V2.2.0 (data assimilation upgrade) Suru Saha – January 2016? WAM (developmental parallel up to 600 km) George Milward – On Track NGAC (major aerosol upgrade) GCWMB – January 2016 GFS (4D hybrid) – April/May 2016 9

10. GEFS Changes (Dec. 2nd 2015) V10.0.0V11.0.0 GFS ModelEuler, 2012Semi-Lagrangian, 2015 Resolution 0-192 hT254 (52km) L42 (hybrid)T L 574 (34km) L64 (hybrid) Resolution 192-384 h T190 (70km) L42 (hybrid)T L 382 (52km) L64 (hybrid) Output resolution1 O x 1 O 0.5 O x 0.5 O and 1 O x 1 O Output frequency6h3h the first 8 days; 6h the rest 10 Moving from BV-ETR approach to EnKF A major scientific shift Unification of DA and Ensemble Generation Direct link to the hybrid 3D-Var EnKF DA system Perturbations are 6h EnKF forecasts with adjustments: Tropical Storm Relocation Centering of perts on ensemble ctl analysis Stochastic perturbation (STTP) upgrade Fine-tune amplitude for changes in model and perturbation method Turn off surface pressure perturbations for tropics to reduce the spread growing of geopotential height

11. Significant upgrade that includes: Install GSI REL-5.0.4 Install Obsproc_Prep.v3.5.0 Code and script changes to conform to NCO standardization Modernization of some diagnostic codes Expected Benefits: Improvements to GSI analysis Improved satellite bias correction scheme Satellites introduced during the last 5 years can now be assimilated in CFSv2 Potential to include new or replacement data sources Improved software serviceability Analysis upgrade supports maintenance and servicability for the next 5 years 2015 CFSv2.2.0 Upgrade Parallel run for Dec2013-Dec2015 11

12. 12 Plans for 2017: Transition to NEMS GSM NOAA Environmental Modeling System (NEMS) version of GFS using ESMF Global atmosphere divided into dynamics and physics components (easier to switch to NGGPS dynamics in the future) Using common write component outside of science components May use NUOPC physics driver if ready Dynamics GSM with a two time-level Semi-Lagrangian Semi-implicit (SLSI) dynamics with a reduced linear transform grid. T2046 spectral resolution: 4096x2048 (~ 9.8 km near the equator) Vertical: Either 64 or 128 hybrid σ – p layers. Time step – 450s Reference pressure – 101350 Pa for stable two-time level semi-implicit scheme – reduced divergence damping Possible Physics Options Simplified Higher Order Closure (SHOC) sub-grid scale turbulence with prognostic TKE Unified deep convection parameterization (Chikira-Sugiyama with Arakawa-Wu extension for scale-aware extension Improving the interactions of cloud macrophysics, cloud microphysics, and radiation Morrison Double-moment cloud microphysics scheme

13. Simulation of Hurricane Sandy: Impact of resolution & physics 13 NEMS/GSM run at T2046: opr physicsNEMS/GSM run at T2046: SHOC+CS

14. Future of Global Modeling at NCEP: NGGPS and NEMS / ESMF Modular modeling, using ESMF to modularize elements in fully coupled unified global model ( + ionosphere, ecosystems, …… ) 14

15. NGGPS Dycore: Selection in 2016; Target Implementation in 2019 Selecting a new dynamic core for global model to serve the NWS for the coming decades. –Architecture suitable for future compute environments. –Non-hydrostatic to allow for future convection-resolving global models. 18 month process to down-select candidate cores. 5 year plan to replace operations. Core  NEMS  applications. –GSM-NH (EMC) –MPAS (NCAR) –FV3 (GFDL) –NIM (ESRL) –NEPTUNE (NRL) –NMMB-UJ (EMC) Potential GFS Upgrade Schedule Dycore Selection Schedule 15

16. Atmosphere Model including Dynamics  t, u, v, w, T, , p, z, q x, c x, a x destaggered Tendencies and Updates Init Mode Dynamical equations, advection, horizontal mixing, diffusion. Radiation Deep and Shallow Cumulus Surface Layer PBL and Vertical Mixing Micro- physics Modified Kalnay Rules Layer NUOPC Physics Driver Schematic Output Diagnostics fields rates budgets others Atmospheric Physics Driver (init, run, finalize modes) Initialize Physics Tables and Databases Finalize Mode. standard interface for model physics NGGPS physics Version 1.0 delivered June 2015 16

17. GEFS Plans Short term (2-3 years) – Configurations: –Will use NEMS/GFS –Increasing horizontal resolution, –Increasing memberships, –Extend forecast to 35 days Science: –Improving initial perturbations for TC relocation –Improving model uncertainty – introduce new stochastic physics –Reduce model bias –Improving probabilistic forecast skills for several weather, and HIW Reforecast: –Run retrospective reforecast for downstream applications Long term (3-10 years) Configurations: –Fully coupled model system – NGGPS –Comparable resolution to cover week 3&4, monthly forecast (45 days) Science: –Optimum initial perturbations –Stochastic parameterizations –Improving probabilistic forecast skills for day 1-30, include HIWs Reforecast: –Run reforecast (with new reanalysis) for downstream applications 17

18. NEMS GFS Aerosol Component (NGAC) updates NGAC Phase 2 (Q2FY16): Multispecies aerosol forecasts o Extend the dust-only system to include sulfate, sea salt, and carbonaceous aerosols o near-real-time biomass burning emissions o Newer GOCART version o Atmosphere physics is upgraded to the latest operational GFS physics package : o New products to support downstream applications o Total Aerosol Optical Depth (AOD) and AOD from each species o Single scatter albedo and asymmetric factor for total aerosols at 340nm o 3D mixing ratios for each aerosol species o Running twice per day at 00Z and 12Z Future Plans o Phase 3:Aerosol analysis using VIIRS AOD (FY18 implementation, tentatively)near-real-time biomass burning emissions o Enable aerosol impacts on medium range high resolution weather forecasts (GFS/GDAS) o Build aerosol-chemistry-climate interaction in the next generation of Climate Forecast System (CFS) o Evaluate the Impact of Cloud-Aerosol- Precipitation Interaction (CAPI) schemes on Rainfall Forecast in the NGGPS o Enable global atmospheric constituents forecasting capability with aerosol impacts on various time scales o Provide quality atmospheric constituents forecast products to serve a wide-range stakeholders 18

19. Thanks for your attention. Questions? 19