The Transition of Hurricane Earl (2010) from a Category Four

Hurricane to a Large Extratropical Cyclone

Sim D Aberson1

Altuğ Aksoy1,2

Kathryn Sellwood1,2

1NOAA/AOML/Hurricane Research Division2UM/Cooperative Institute for Marine and Atmospheric Studies

Fourth International Workshop on Extratropical Transition21 May, 2012

Mont Gabriel Lodge, QuebecMonday, May 21, 12

NOAA Hurricane Field Program/Intensity Forecasting Experiment (IFEX)

Multi-year experiment collaboration between NOAA Research, Operations, Satellite Information Service and Aircraft Operations

The goal is to increase understanding of intensity and structure changes to ultimately improve forecasts

One of the experiments that has been in the program is the Extratropical Transition Experiment

The Hurricane Ophelia (2005) flights were reported at IWET3. Since then, we have done a second set of flights into Hurricane Earl.

Monday, May 21, 12

Six P3 missions and three G-IV missions (centered at 0000 UTC each day) from before maximum intensity to just before transition was completed, including just before intensification during transition

Monday, May 21, 12

Model and DA developmentConcurrent with IFEX, HRD has been developing a new high-resolution operational regional hurricane forecasting system (Hurricane Weather Research and Forecast Model - Gopalakrishnan et al. 2011) and a data assimilation system (the Hurricane Ensemble Data Assimilation System - Aksoy et al. 2012) for incorporation of high-resolution aircraft data into the model initial conditions.Some studies have looked at the high-resolution core structure of transitioning TCs using models (eg. Davis et al. 2008, Evans and Hart 2008), but we have not yet had observations to comfirm these results.

Monday, May 21, 12

HEDAS Cycling Workflow• Run for cases (2008-2011) when NOAA

Airborne Doppler Radar data were available (84 cases)

• Uses 1452 processors on NOAA’s t-jet T T + 126 hT – 6 h

Ensemblespin-up

DA cyclingwith EnKF

Deterministic HWRFx forecast from ensemble

mean

Real-timeobservation

pre-processing

Ensembleinitialization from GFS-

EnKF Mean of final analysis

1-h cycles

Monday, May 21, 12

HRD’s Hurricane Ensemble Data Assimilation System

(HEDAS)

• Forecast model:– HRD’s Experimental HWRF (HWRFx)– 2 nested domains (9/3-km horizontal resolution, 42 vertical levels)– Static inner nest to accommodate covariance computations

Inner nest size: ~10x10 degrees– Ferrier microphysics, explicit convection on inner nest– Model integration begins at the time of the final analysis

• Ensemble system:– Initialized from GFS-EnKF (NOAA/ESRL) ensemble– Initial ensemble is spun up for 3-4 h before assimilation begins– 30 ensemble members

• Data assimilation:– Square-root ensemble Kalman filter, EnKF (Whitaker and Hamill

2002)– Assimilates all realtime aircraft data on the inner nest

NOAA P-3, NOAA G-IV, USAF C-130– Covariance localization (Gaspari and Cohn 1999)

Monday, May 21, 12

C-Band weather radar

GPS navigation

Stepped-frequency Microwave Radiometer (new)

Aircraft Data of Interest

HumidityGust probe

C-Band weather radar

Chemistry inlet

RadiometerLower-fuselage weather radar

Launch tubes and chutes:

Pylons

Tail Doppler radar

Radar altimeterGPS

(Stepped-frequency Microwave Radiometer, cloud physics, aerosol)

P-3 Aircraft(convective core)

G-IV Aircraft(environmental)

HEDAS assimilates dropwindsonde, Doppler radar, flight-level, and SFMR observations from

all aircraft (mostly NOAA and AF)

Tail Doppler radar (new)

Dropwindsonde

Dropwindsonde, AXBT, AXCP, AXCTD

Observation Error

Doppler wind speed 2 ms-1

flight-level/dropwindsonde temperature 0.5 K

flight-level/dropwindsonde u/v 2 ms-1

Stepped-Frequency Microwave Radiometer Variable, mean ~5 ms-1

Monday, May 21, 12

The assimilation of aircraft data improves track forecasts by up to 10% versus the standard initialization during the critical watch and warning periods.  [The poor result at 108 is due to a very few Ike forecasts.  The majority of the forecasts at 108 h are improved.]

Homogeneous Homogeneous

Pairwise Homogeneous

** **

*

* - means of two samples are different at 90% ** - means of two samples are different at 98%

Monday, May 21, 12

Homogeneous Homogeneous

Pairwise Homogeneous

* - means of two samples are different at 90% ** - means of two samples are different at 95%

Standard Verification

The assimilation of all aircraft data improves intensity forecasts by nearly 25% versus the standard initialization during much of the critical watch and warning periods.

** **

**

****

Monday, May 21, 12

GOES IR imagery

during six flights (one

missing from NRL archive)[next: realtime Doppler

radar analyses]

Monday, May 21, 12

Monday, May 21, 12

Monday, May 21, 12

HEDAS analyses at upper levels where there are fewer scatterers for Doppler analyses alone

Monday, May 21, 12

200 hPa circulation disappears in last 24 h. Circulation at 500 hPa weakens slower than that below and above.

Monday, May 21, 12

Monday, May 21, 12

Radar and HEDAS analyses show that, in time, the radius of maximum wind speed increases and the profile flattens. This can also be seen in the raw flight-level data from the P3s. (Evans and Hart 2008)

Like the wind field, the warm core is maintained best in middle levels and weakens most rapidly near the tropopause, but also weakens more rapidly at low levels. Evans and Hart (2008) did not specifically show the latter.

Evans and Hart (2008) also showed that wind field expansion begins in the upper troposphere and progresses downward, but these analyses do not necessarily show that in this case.

Monday, May 21, 12

6-h rain rate (to allow for spin-up) shows change of symmetry to most rainfall on left side during transition

Monday, May 21, 12

Very little tilt when Earl was at its strongest

Monday, May 21, 12

Quite a lot of tilt (to SE at low levels, then to NW (>40 km) to upper levels as transition begins

Monday, May 21, 12

Tilt actually deceases as transition progresses and Earl strengthens

Monday, May 21, 12

Davis et al. 2008 looked at tilt of the circulation with height and its relation to shear.

Synoptic-scale shear remains < 40 ms-1 throughout

Since these analyses are only snapshots every 12 h, they cannot be used to confirm their results on adjustments to shear (precession, formation of new center, baroclinic development, etc.) and relationship between precipitation and tilt except in this particular modeling framework.

Analyses of individual flight legs might provide some insight, but lack of scatters near vortex center precludes a close look at tilt.

Doppler Wind Lidar may help in this regardMonday, May 21, 12

Future plansThese HEDAS analyses were created with older, less accurate version of the model. In addition, not all variables were updated in analyses, causing large spindown (bias) issues.

Upgrade to current operation version of HWRF is now complete.

Addition of high-density CIMSS winds, GIV Doppler data to analysis.

ET experiment is still in the HFP book. If anyone has any suggestions for modifications, please let me know. Planned retirement of the P3s in the next few years makes conducting another experiment soon imperative.

Monday, May 21, 12

Monday, May 21, 12

Monday, May 21, 12