Probabilistic Hurricane Inundation Surge Height (PHISH) Guidance

1
Probabilistic Hurricane Inundation Surge Height
(PHISH) Guidance

• 2011 NOAA Hurricane Conference
• Anne Myckow - MDL/NWS
• Arthur Taylor(MDL), John Kuhn (OS21)

2
SLOSH Products

• Deterministic / Historic Runs
• P-Surge
• Probabilistic Storm Surge
• Response (lt48 hr of landfall)
• MEOW
• Maximum Envelope Of Water
• Readiness (48hr 120 hr of landfall)
• MOM
• Maximum Of the MEOWs
• Planning / Mitigation (gt120 hr of landfall)

3
Case Study Hurricane Ivan

• Top Left Real-time deterministic SLOSH run for
  Ivan at advisory 54, about 10 hours before
  landfall
• Note large (14ft) surge in Mobile, small (3ft)
  surge for Pensacola

• Bottom Right Hindcast best track SLOSH results
  for Ivan
• Track forecast off by approx. 25 mi
• Note significant surge in Pensacola (7 10ft),
  missed by the deterministic

4
Katrina, Adv 23
Single track forecast for Katrina, with cone of
uncertainty represented in grey
5
Katrina, Adv 23

• Create more forecasts, first varying across the
  track
• Distribution based on known historical errors
  with NHCs forecasts
• Chose to have forecasts spaced Rmax apart at the
  48hr forecast point

6
Katrina, Adv 23
More forecasts created by varying size (pictured
left) as well as intensity and forward speed for
all possible tracks, 3 variations for
each. Results in 270 total possible forecasts,
run within multiple basins p-surge completes its
run in approx. 30min
7
Probability of Surge gt 5 feet
Katrina adv 23
8
Surge Height Exceeded by 10 of Ensemble Members
Katrina adv 23
9
Rationale for PHISH

• P-surge gives results above datum, which can be
  confusing for users, but results above ground
  level will eliminate datum issues.
• How to move p-surge to above ground level?
• Subtract land from p-surge products
• Could work for exceedance product
• Unable to subtract land from a probability
• Subtract land before combing into probabilities
• Expert users may still need above datum product,
  so cannot replace p-surge
• Create a new product (PHISH)

10
PHISH as of 2011
11
PHISH Example (Probability)
P-surge (above datum) Probabilistic product
PHISH (above ground level) Probabilistic product
12
PHISH Example (Probability)
P-surge (above datum) Probabilistic product
PHISH (above ground level) Probabilistic product
13
PHISH Example (Exceedance)
P-surge (above datum) Exceedance product
PHISH (above ground level) Exceedance product
14
PHISH Example (Exceedance)
P-surge (above datum) Exceedance product
PHISH (above ground level) Exceedance product
15
PHISH in 2012

• PHISH implementations at NCEP
• PHISH data transmissions via AWIPS
• WMO headers
• Cumulative probabilities for 0-6, 0-12, 0-18,
  0-78
• Incremental probabilities for 0-6, 6-12, 12-18,
  72-78

16
Future Work

• SLOSH Tides gtgt PHISH Tides
• Use more recent basins
• Shift all basins to NAVD88
• Inundation maps with 30m DEMs
• Possible routes
• User subtracts DEM from p-surge exceedance
  product
• Wont work for probability products
• Provide PHISH products at high resolution
• Large amount of data transmission (May need
  paradigm shift)

17
Working Towards Tropical Cyclone Inundation
Products
Hurricane season starts
Current Phase
Research and Development
Experimental
Operational
Operational Enhancement
Depends on Slosh Tide

• Interim product discontinued
• 2 Psurge without tide is currently
  operational

18

• Questions?

19
SLOSH (may not need slide)

• Sea, Lake and Overland Surges from Hurricanes
• Finite differencing model developed by the
  Meteorological Development Laboratory to predict
  storm surge
• Overland flooding
• Parametric wind model for forcing
• Structured grid with finer resolution near shore,
  and coarser offshore
• Models sub-grid features such as levies, barrier
  islands, and river channels
• Does not include
• Tides, waves, river flow
• Tides can be conservatively estimated by
  initializing the grid at high tide