HFIP Ocean Model Impact Tiger Team (OMITT)

1
HFIPOcean Model Impact Tiger Team (OMITT)
Chair and co-chair H.-S. Kim, G.
Halliwell Team L. Bernardet, P. Black, N. Bond,
S. Chen, J. Cion, M. Cronin, J. Dong, I. Ginis,
B. Jaimes, B. Liu, E. Sanabia, N. Shay, V.
Tallapragada, B. Thomas, E. Uhlhorn, and L. Zhu
Institutions EMC, DTC, HRD/AOML, PhoD/AOML,
PMEL, USNA, Navy, URI, UMiami, and UWashington
Wednesday March 4, 2015 The 69th
Interdepartmental Hurricane Conference March 2-5,
2015 Jacksonville, FL
2
Overview

1. Objectives
2. Approaches
3. Groups and Primary Tasks
4. Progress
5. Near Future Plans
6. Related Presentations at IHC

3

• 1. Objectives
• Determine the benefit of coupling ocean models of
  various complexity
• to the hurricane atmospheric model, by assessing
  the impacts of the
• ocean components on the HWRF forecasts and the
  sensitivity of these
• impacts on air-sea interface, surface flux, and
  atmospheric parameters.
• Background
• Several studies have demonstrated already in the
  past, e.g. Lee and Chen (2014), Yablansky and
  Ginis (2014), and Wada and Usui (2007 and 2010).
• Q1 What would this effort be different from the
  past research/effort?
• Q2 Observation limits the effort. What are
  options?
• Q3 What do we expect at the end of one year?
• Year-End Goal
• Submit Recommended strategies to improve ocean
  model performance in TC forecasts, prepare a
  report and present results in 2015 HFIP annual
  meeting.

4

• 2. Approaches
• Models
• 1. 3D Ocean couple
• POM (w/ HWRF), HYCOM (w/ HWRF), and NCOM (w/
  COAMPS-TC)
• Performed in 3 basins NAtl, EPac, and WNP.
• 2. 1D Ocean couple Mixing physics
• M-Y and KPP w/ momentum balance in a column
• No simulations done, but can be conducted on
  demand.
• Reserve this for Ideal Runs.
• 3. Specified SST
• a) Persistent SST from GFS/GDAS (RTG), NCODA and
  RTOFS nowcasts
• - objective what is the impact of
  mesoscale variability of SST.
• (Preliminary Results in
  slides 11-12)
• b) Updated SST (from global RTOFS products)
• Can be performed in all 3 basins.

5
3. Groups and Primary Tasks
Real cases evaluate and compare with obs.
Black, Chen, Cion, Halliwell, Jaimes, Kim, Liu,
Sanabia, Shay, Uhlhorn, Zhu
Data tools collect, archive at DTC, and
distribute Bernardet Black Cion Halt Kim
Sanabia Shay Uhlhorn
Design simulations e.g., run real cases on
demand or sensitivity study Chen, Dong,
Halliwell Ginis Thomas Zhu
Communicate at biweekly telecons with goals
of 1. Build synergy bwn modelers and field
scientists, 2. Exchange diagnostic tools and make
them available, 3. Update progress, and 4.
Assist publications.
Kim et al. 5
6
4-1
4. Progress

• Defined (as of February 2015)
• Models, Groups, Regular telecon times,
  Definitions done
• Storms, Metrics, Data tools Repository almost
  complete
• Primary Storms of Interest
• Cases selected from the 2014 operational, and
  upgraded TE runs, based on
• a) the strength of ocean influence, and
• b) the availability of ocean and atmosphere
  observations
• Runs by non-coupled, and coupled but with
  different ocean models as operational or
  experimental exercise in 3 basins N. Atlantic
  (NAtl), E. Pacific (EPac), and W. N. Pacific
  (WNP).
• NAtl
• Edouard (2014), Isaac (2012), and Gonzalo (2014)
• EPac
• Julio (2014) and Iselle (2014)
• WNP
• Fengshen (2014), Pabuk (2013), Kammuri (2014) and
  Soulik (2013)

7
4-2
4. Progress

• Data
• DTCs Mesoscale Model Evaluation Testbed (MMET)
  repository
• of simulations and observations.
• Observations
• The SS funded ocean obs. - leverage
• Target obs. air-borne, unman vehicle, ?
• Mooring obs. KEO, TAO, NDBC, ?
• Lagrangian obs. gliders, drifts, floats, ?
• Remote sensing obs. MW, IR, Wind, ?
• Simulations
• Operational HWRF (HWRF-POM),
• Experimental HWRF (HWRF-HYCOM) and COAMPS-TC, and
  
• Ideal cases non-coupled, 1D and 3D ocean
  coupled.

8
Mesoscale Model Evaluation Testbed (how it can
help HFIP OMITT)
4-3
Model and observation data sharing by L.
Bernardet, 13th February 2015
4. Progress

• What Mechanism to assist research community with
  initial stage of diagnostics and testing, with
  the goal of leading to model improvements
• Can provide
• Observational datasets
• Forecasts and verification from selected model
  configurations
• Code for conducting model runs
• Where Hosted by the DTC served through
  Repository for Archiving, Managing and Accessing
  Diverse DAta (RAMADDA)

http//www.dtcenter.org/eval/meso_mod/mmet
9
4-4
DTCs hwrf-contrib for tools Repository by C.
Halt Feb. 13, 2015
4. Progress

• Metrics
• Use consistent definitions
• Oceanic Mixed Layer depth (MLD)
• storm footprint
• surface layer
• Diagnostic depths

• Diagnostic Tools
• DTCs hwrf-contrib SVN repository for
  diagnostic tools.

http//www.dtcenter.org/HurrWRF/developers/contrib
/overview.php
10
4-5
4. Progress

• Sensitivity to SST (J. Dong and H-S Kim)
• Experiment design
• Non-coupled HWRF
• SST
• Constant SST (25, 27 and 28oC)
• GFS/GDAS (RTG) available daily
• NCODA available daily and/or 6 hourly
• global RTOFS available hourly
• Approach
• Hurricane Edouard (2014) atmospheric conditions
  but using persistent SST.

• Experiment 1 (next 2 slides)
• Spatially uniform vs. varying SST, but persistent
  (Non-Coupled HWRF)
• constant SST 27 and 28oC
• spatial varying SST 00Z SST from GFS, NCODA and
  RTOFS

11
4-6
4. Progress

• Sensitivity to SST (J. Dong and H-S Kim)

Preliminary Results for Experiment 1 Non-coupled
HWRF using persistent SST (spatially uniform vs.
varying)

• Intensity two groups
• Large difference between constant SSTs.
• Less differences among spatially varying SST.
• Constant SST results in under-prediction for 27oC
  and over-prediction for 28oC.
• All GFS, NCODA and RTOFS SST show a similar
  forecasts.

Track a similar bias pattern northward for
0-30 hrs, southwestward for 36-66 hrs, and
eastward for 72-120 hrs. Except 28oC SSTs track
is excellent for 36-60 hrs.
12
4-6
4. Progress

• Sensitivity to SST (J. Dong and H-S Kim)

1. Constant SST large error at later lead
times and the same for bias to negative in Pmin
for 28oC and positive for 27oC, and opposite for
Vmax. 2. RTOFS (pink) less spin-up/down
overall small error and bias at later lead times,
cf others. 3. NCODA (yellow) relatively large
bias and errors, starting from 24-hr and
persistent. 4. GFS (green) comparatively
better performed than NCODA, except abnormal peak
in bias around 24 hr.
Preliminary Results ? Best w/ RTOFS SST
? Worse w/ constant SST
Error
Bias
Note different color codes from the previous
slide.
13

1. Complete a write-up of the OMITT Work Plan
2. Build tools to extract and estimate a minimum set
   of TC parameters reduction of the simulation
   outputs for a easy reference to the OMI
   investigation
3. Build a database for storms of interest
4. Complete building of 3D HYCOM coupling Ideal case
5. Transition Tools to Operation
6. Assist Field Observation
7. Reinforce collaborations e.g., invite
   atmospheric scientists to the telecon

14
6. Related Presentations at IHC

• P01 Upper Ocean Observations in Hurricane
  Edouard Uhlhorn et al.
• P04 Support for Users and Developers of the
  Hurricane WRF Model Bernardet et al.
• P11 Kuroshio Extension Observatory (KEO)
  Measurements of the Upper-Ocean Response to
  Tropical Cyclones in the Western North Pacific
  Bond et al.
• S1-05 2014 AXBT Demonstration Project
  Operations Summary and Research Update
  
  
  Sanabia and Black
• S3b-02 Improving the Ocean Component of the
  Operational HWRF and GFDN/GFDN Hurricane Models
  
  
  Ginis et al.
• S5a-01 Advanced Operational Global Tropical
  Cyclones Forecasts from NOAAs High-Resolution
  HWRF Modeling System Tallapragada
• S5a-06 Proposed 2015 NCEP HWRF Hurricane
  Forecasting Model Trahan et al.
• S5b-03 Sensitivity of Ocean Sampling for Coupled
  COAMPS-TC Prediction Chen et al.
• S5b-05 NOAAs Use of the Coyote UAS in Hurricane
  Edouard to Enhance Basic Understanding and
  Improve Model Physics Cion et al.

15
Thank you Questions?