Post on 29-Jun-2020
Department of Commerce // National Oceanic and Atmospheric Administration // 1
Use of Operational Ocean
Satellite Data in Operational
Ocean Forecast Systems:
Opportunities and Challenges
Avichal Mehra1 and Jim Cummings2
NWS/NCEP/EMC/Modeling and Data Assimilation Branch
1NOAA/NWS/EMC; 2IMSG at NOAA/NWS/EMC
First International Operational Satellite Oceanography Symposium 2019
June 18-20, 2019, College Park, MD
Department of Commerce // National Oceanic and Atmospheric Administration // 2
Outline
○ Current Operational Ocean Forecast Systems
Global
Regional/Hurricane
○ Operational Ocean/Marine Data Assimilation Systems
GODAS (Current)
RTOFS-DA (Near-Future)
Marine JEDI (Future)
o Operational Monitoring/Metrics
o Challenges
Department of Commerce // National Oceanic and Atmospheric Administration // 3
GODAS
3DVAR
Ocean Model
MOMv4
fully global
1/2ox1/2o (1/4o in tropics)
40 levels
Atmospheric Model
GFS (2007)
T382 64 levels
Land Model Ice Mdl SISLDAS
GDAS
GSI
6hr
24hr
6hr
Ice Ext6hr
Climate Forecast System
http://cfs.ncep.noaa.gov/cfsr
Climate Forecast System (CFS v2) at NCEP
Department of Commerce // National Oceanic and Atmospheric Administration // 4
Real Time Ocean Forecast System (RTOFS v1.2) at NWS/NCEP
Strong collaboration with US Navy, leveraging core HYCOM
and ocean data assimilation developments at NRL.
Eddy Resolving Ocean Modeling and Initialization
Coupled Modeling for Hurricanes (Air- Sea-Wave flux interactions, mixing)
Data Assimilation and Analysis for Global Weather scales
Coupled Ecosystem Forecasting (Biogeochemical, NPZD, tracers)
RTOFS presently based on HYCOM
Department of Commerce // National Oceanic and Atmospheric Administration // 5
Operational HWRF (Hurricane Weather Research
and Forecast) System
5
• HWRF continuously improved in the past
six years through support from HFIP
• Successful community modeling approach
for accelerated transition of research to
operations
o Three nested domains at a horizontal
resolution of 13.5/4.5/1.5 kms
o Formal vortex initialization and GSI-
based data assimilation
o Actively coupled to Ocean models in all
global basins
o Actively coupled to Wave model
(WW3) in NHC basins
o Inner-core DA including real-time data
from recon missions (TDR,
dropsondes, SFMR etc.)
Expanded
international
partnerships &
collaborations
2017 HWRF
Results from annual
retrospectives
Department of Commerce // National Oceanic and Atmospheric Administration // 6
HMON: Hurricanes in a Multi-scale Ocean coupled Non-hydrostatic model
HMON: Implements a long-term strategy at NCEP/EMC for multiple static and moving
nests globally, with one- and two-way interaction and coupled to other (ocean, wave,
land, surge, inundation, etc.) models using NEMS-NUOPC infrastructure. Precursor
to FV3-based systems.
Operational HMON v2.0at NWS/NCEP
HM
ON
Department of Commerce // National Oceanic and Atmospheric Administration // 7
Operational Hurricane Forecasting Models
HWRF/HYCOM HMON/HYCOM HWRF/MPIPOM
North West Pacific, North Indian Ocean, Southern Hemisphere Basins
North AtlanticBasin
North Atlantic, North East Pacific, Central Pacific Basins
IC come from climatology and SSH-based feature model
IC from Global RTOFS nowcast
IC from Global RTOFS nowcast
North East Pacific, Central Pacific Basins
Department of Commerce // National Oceanic and Atmospheric Administration // 8
Global Ocean Data Assimilation System(GODAS)
To initialize ocean state for seasonal to intra-seasonal forecasting system (CFSv2)
Model : GFDL Modular Ocean Model (MOM4p0)
Horizontal resolution : 1.0 degree by 1.0 degree with meridional 1/3 degree resolution in Tropical
region.
Vertical resolution : 40 levels with 10 m near surface to 506 m near bottom
3dVAR developed at NCEP
Observational Data sets used
Hydrostatic data : XBT, CTD
Mooring data : TAO, TRITON, PIRATE
Float Data : Argo
Satellite date : Jason 1 (2)
Surface Fluxes : NCEP/DOC (R2) Reanalysis
Sea Surface Temperature : NCEP/DOC (R2)
Sea Surface Salinity : weakly restoring to Climatology (WOA09)
Synthetic Salinity profile : based on observed T and climatological relations in T and S
Department of Commerce // National Oceanic and Atmospheric Administration // 9
Ocean
Obs.
Ocean Data
QC
3DVAR
HYCOM
SST: GOES, MSG, AMSR-E,
AATSR, NOAA (GAC,LAC),
METOP (GAC,LAC), Ship/Buoy
Profile Temp/Salt: XBT,
CTD, Argo Float, Fixed/Drifting
Buoy Altimeter SSH: Jason2,
Jason3, Altika, CryoSat2
Sea Ice: SSMI, SSMI-S
Glider: Slocum, Spray-CTD,
Sea-Glider
Ocean Color: VIIRS
Stage 1: Preliminary data
sensibility error check
Stage 2: External
data error check
Stage 3: Internal
data error checkInnovations
Increments
First
Guess
Forecast
Fields +
Prediction
Errors
Adaptive
Sampling
Data Impacts
Forecast
Component
Analysis Component
(QC+3DVAR)
Stage 4: Adjoint
sensitivities
RTOFS - Data Assimilation: 3DVAR (Based on Navy’s Coupled Ocean Data Assimilation (NCODA)
Oceanographic implementation of the 3-D
variational (3DVAR) technique
Unified and flexible oceanographic
analysis
Global temperature, salinity, geopotential,
flow fields from all sources of operational in
situ and remotely sensed ocean
observations
NCODA 3DVAR is tightly coupled to an
ocean data quality control (QC) system
Extension to biogeochemical (BGC)
components is in progress
9
Department of Commerce // National Oceanic and Atmospheric Administration // 10
Unified DA Effort - JEDI
o The Joint Effort for Data assimilation Integration (JEDI) is a collaborative development spearheaded by the JCSDA
Next generation unified data assimilation system
For research and operations (including R2O/O2R)
For various components of the earth system, including coupled
Mutualize as much as possible without imposing single approach• Open-development software –model: in addition to supported releases, community, developers can
obtain and collaborate on latest development branches
Collaborative teams – NOAA, NASA, US NAVY
o The Marine JEDI DA system for NOAA/NCEP is through SOCA (Sea-ice Ocean Coupled Assimilation)
10
Department of Commerce // National Oceanic and Atmospheric Administration // 11
Operational Ocean Forecasts: Monitoring and Verification
o Methodology:
Focus on global and basin-wide domains and specific water mass
observations
Use real time data (remote, in situ)
Design diagnostics based on established mature science
Evaluate system performance and product quality
Adopt standard definitions and community-wide formats:
A good starting point – GODAE validations; (Oceanography, Vol. 22, No. 3: 128-143)
11
Department of Commerce // National Oceanic and Atmospheric Administration // 12
Operational Ocean Forecasts: Monitoring and Verification
o Classes of Metrics:
CLASS 1: 2D horizontal Fields ; T, S, U, V, and W at fixed depths;
SSH, MLD.
CLASS 2: Vertical sections/ moorings (T,S,U,V)
CLASS 3: Integrated quantites (transports etc.)
CLASS 4: Test and compare performance of analyses and forecasts.
12
Department of Commerce // National Oceanic and Atmospheric Administration // 13
Operational Ocean Forecasts: Monitoring and Verification
13
Class 1: RTOFS vs MADT SSHA
Most of the differences are in regions of large
variability
Better representation of warm pool/cold tongue in
equatorial Pacific in parallel.
RTOFS v1.1 RTOFS v1.0Observations
Class 2: Obs vs RTOFS v1.1 (parallel) vs RTOFS
v1.0 (operational); Temperature Cross Sections
Department of Commerce // National Oceanic and Atmospheric Administration // 14
Operational Ocean Forecasts: Monitoring and Verification
14
Class 3: Florida Cable Transports RTOFS vs Cable Class 4: RTOFS vs SST
Integrated transports at 27 N Red: 192 hr forecasts; Green: all forecasts;
Blue: nowcast; Black: #points
Department of Commerce // National Oceanic and Atmospheric Administration // 15
Operational Ocean Forecasts at NCEP:
Data related Challenges
Coupled Modeling and Data Assimilation:
o Modern architectures for coupling, object oriented methods, community-based
unified systems (UFS, JEDI)
o Enable and explore coupled modeling and DA
o Modern cross-component databases
Advanced Ocean Observation Systems:
o Adaptive in-situ systems for rapid response deployment (e.g. for Hurricanes)
o Use of forecast uncertainties for enhanced observation systems
o Use of advanced/improved sensors on future satellites
o Accurate better representation of observation errors and bias
Department of Commerce // National Oceanic and Atmospheric Administration // 16
Operational Ocean Forecasts at NCEP:
Data related Challenges
Enhanced Probabilistic Guidance:
o Use of ensembles and super-ensembles; perturbations based on
observations
o Advanced monitoring/verification for improved uncertainty measures
o Co-ordination amongst national and international efforts
Modern Computing Architectures/ Cloud Servers:
o Software Engineering practices
o Hybrid CPU/GPU platforms
o Cloud-based resources
Department of Commerce // National Oceanic and Atmospheric Administration // 17
Thank You!
Department of Commerce // National Oceanic and Atmospheric Administration // 18
Operational Ocean Forecasts at NCEP: Users
NOAA:
o NWS (EMC, CPC, NHC, OWP);
o NOS (Coastal & Estuarine OFS; IOOS, WCOFS);
o OAR (PMEL, AOML, CPO, ESRL);
o NESDIS (Ocean Color, Sea Ice, SSS , ADT)
Non-NOAA Federal:
o US Coast Guard
o JCSDA,
o FDA, USGS, other federal agencies
Others:
o Universities
o International Projects
o Private Industry
Department of Commerce // National Oceanic and Atmospheric Administration // 19
Operational Ocean Forecasts at NCEP: Status
Global Models:
o Current:
Real Time Ocean Forecast System (RTOFS) based on HYCOM for weather scales at 1/12°
Climate Forecast System (CFS) based on MOM4 for sub-seasonal to seasonal scales at ½ °
o Future:
Unified Forecast System (UFS) with MOM6 coupled to FV3 and CICE6, initially for sub-seasonal
to seasonal applications
Regional (Hurricane) Models:
o Current Status:
HWRF coupled to MPIPOM and HYCOM at 1/12°
HMON coupled to HYCOM at 1/12°
o Future:
Hurricane Analysis and Forecast System (HAFS) with FV3 based Hurricane model coupled to
HYCOM (initially) and MOM6 when available.