Docstoc
EXCLUSIVE OFFER FOR DOCSTOC USERS
Try the all-new QuickBooks Online for FREE.  No credit card required.

Tropical Cyclone Analysis with QuikSCAT

Document Sample
Tropical Cyclone Analysis with QuikSCAT Powered By Docstoc
					Hydrometeorological Prediction Center
Tropical Cyclone Responsibilities and
        Tropical Cyclone QPF




                    Michael J. Brennan
               Science and Operations Officer
                Hydrometeorological Prediction Center

             NOAA in the Carolinas Hurricane Conference
                            6 May 2008
6 May 2008                                                1
               Outline
• History of HPC’s Tropical Cyclone (TC)
  Responsibilities
• Synoptic Reasoning and Track
  Forecasting
• Tropical Cyclone QPF
• Inland Advisories
• NHC COOP

  6 May 2008                          2
      History of HPC’s Tropical Cyclone
           Program Responsibility
• HPC (and its predecessor, NMC) have provided
  track forecast guidance to NHC for at least the
  last 30 years
  • Initially done because NMC had access to much more
    global model guidance than NHC
• NHC COOP program
  • HPC/NMC has been the official NHC backup since at
    least 1978
  • Overhauled and modernized in 2002-2003
• HPC began providing the rainfall statement for
  NHC’s public advisories in 2005

 6 May 2008                                        3
                Synoptic Reasoning and
                  Track Forecasting
• HPC provides track
  forecast guidance
  (points) to NHC for TCs
  west of 60°W
• HPC’s synoptic
  reasoning provided on
  Hurricane Hotline
                            Day 6 Forecast of Ernesto
  conference call
• Day 6 & 7 TC positions
  for HPC medium range
  charts coordinated with
  NHC in daily Noon
  conference call
   6 May 2008                HPC Medium Range Prog Valid 12Z 31 August 2006
                                                                          4
                  Synoptic Reasoning and
                    Track Forecasting
• HPC takes advantage of
  expertise in large-scale
  pattern diagnosis to
  compliment NHC track
  forecast
• HPC track forecast not
  constrained by continuity
  and can react more to
  run-to-run changes in
  model guidance
• Ensures HPC has working
  forecast in place for all
  TCs west of 60°W in case
  backup for NHC needed

     6 May 2008                            5
        Tropical Cyclone Rainfall




6 May 2008                          6
             U.S. Weather Fatalities




6 May 2008                             7
       TC Rainfall – Major Contributor to
            Annual Precipitation in
              Southeastern U.S.
2005 Annual Rainfall (mm)                  2005 TC Rainfall (mm)
                      10-20 % of annual rainfall
                             from TCs




               Images Courtesy Frank Marks (AOML/HRD)
  6 May 2008                                                   8
               Factors Impacting Rainfall
             Distribution in Landfalling TCs
•     Storm size (positive) – the larger the TC, the more it
      rains at any given spot
•     Storm track (location)
•     Diurnal cycle – core rainfall overnight, outer band
      rainfall during the day
•     Topography – enhances rainfall in upslope areas, but
      decreases past the spine of the mountains
•     Wind shear (negative) – leads to a quicker drop-off in
      rainfall for inland TCs
•     Interaction with nearby features (troughs, fronts/jets),
      extratropical transition can greatly modify rainfall
      distribution


    6 May 2008                                                   9
         Depth of Upper Trough Causing
                Recurvature Key
• Storms which drop most of their rain
  left of track recurve due to significant
  upper troughs in the westerlies
    • Rainfall streaks out well north of the
      system due to jet streaks moving
      around the upper trough and
      frontogenesis at the trough’s leading
      edge
    • Trough transitions from positive to
      negative tilt during interaction with
      TC
• Storms where most rain occurs to
  the right of track are steered
  predominantly by shear lines or
  through a break in the subtropical
  ridge
    • Rainfall tends to be concentrated
      near and right of track

      6 May 2008                                                           10
                                               Atallah et al. (2007) Mon. Wea. Rev.
                                 Bertha (1996)
                      250-mb height and wind




                                                 Bertha moving
 00Z 13 July
                                                 around break in
                                                subtropical ridge
                                               Upper trough and jet
                                                well northwest of
                                                 Bertha’s track
 15Z 13 July



Norman W. “Wes”
Junker


         6 May 2008                                             11
                              Floyd (1999)
                      250-mb height and wind




                                               Floyd interacts with
 00Z 16 Sep
                                                 positively-tilted
                                               upper trough more
                                                     directly
                                               Precipitation forms
                                                 in pronounced
                                               jet entrance region
 12Z 16 Sep

Norman W. “Wes”
Junker




         6 May 2008                                            12
 Bertha (1996) vs. Floyd (1999)




6 May 2008                        13
         “Predecessor” Rainfall Events
 • Coherent area of rain displaced poleward of TC
 • Maximum rainfall rates exceeded 100 mm (~ 4 in.)
 in 24 h
 • Moisture transport from TC toward PRE

                            PRE
 47 PREs associated with 21 TCs were identified
during the period from 1998-2006 (~2 PREs per TC)
~1/3 of all U.S. landfalling TCs produced at least one
Slides courtesy            PRE
Matt Cote &
Lance Bosart –
          Five   cases where TC did not make U.S. landfall
                              Gaston
SUNY Albany

    6 May 2008                                               14
                 1800 UTC 040830 WSI NOWRAD Radar Mosaic
           Southeast Recurvatures
• 7/11 (64%) produced
at least one PRE
• 16 PREs from 7 TCs
• Influential
geographical features:
   - Gulf of Mexico
   - Atlantic Ocean
   - Appalachians

 • Approximate point of
 PRE formation




6 May 2008
        Slides   courtesy Matt Cote & Lance Bosart – SUNY Albany   15
                        PRE Statistics
                                              Separation Distance
                                                1086 ± 482 km
                                                Median: 935 km
                                                 Event Duration
                                                   14 ± 7 h
                                                  Median: 12 h
                                                     Time Lag
                                                   45 ± 29 h
Bosart and Carr (1978) conceptual
                                                  Median: 36 h
model of antecedent rainfall from
          Agnes (1972)
  6 May 2008
          Slides   courtesy Matt Cote & Lance Bosart – SUNY Albany   16
  Track-Relative Location of PREs

                         PRE Locations Relative TC Track (1998-2006)
                         PRE Locations Relative to to TC Track 1998–2006
                    30


                    25                            Potential for flooding in
                                                    Potential for excessive
                                26                   areas not directly
                                                  flooding beginning before
   Number of PREs




                    20                            impacted by TC rainfall
                                                      arrival of TC rainfall
                    15


                    10
                                                       12
                    5                                                          9
                    0
                         PRE Left of TC Track   PRE Along TC Track   PRE Right of TC Track
                                                Relative Locations




6 May 2008
        Slides             courtesy Matt Cote & Lance Bosart – SUNY Albany                   17
      Southeast Recurvatures Composite
                             Time of PRE Initiation
                     Matt Cote & Lance Bosart – SUNY Albany




        700 hPa Ht (dam) and UVM (μb s-1)   925 hPa Ht (dam), θe (K), and 200 hPa wind speeds (m s-1)

• Significant midlevel trough with            • PRE forms:
weak UVM well poleward of TC                     - in right-entrance region of
                                                   intensifying upper-level jet
• Deep meridional flow transports
tropical2008
    6 May moisture up East Coast                   - on western edge of θe ridge
                                                                             18
             Tropical Cyclone QPF




6 May 2008                          19
                    Tropical Cyclone QPF
• HPC provides rainfall guidance for all TCs with land
  impacts
    • Rainfall statements included in NHC’s public advisories
      and HPC’s advisories on inland systems
    • Detailed QPF guidance to WFOs and RFCs for river and
      flood forecasting

TROPICAL STORM ERNESTO ADVISORY NUMBER 30
NWS TPC/NATIONAL HURRICANE CENTER MIAMI FL
1100 PM EDT THU AUG 31 2006

RAINFALL TOTALS OF 5 TO 10 INCHES ARE
POSSIBLE FROM NORTHEASTERN SOUTH CAROLINA
INTO THE MID-ATLANTIC STATES...AND THE
SOUTHERN AND CENTRAL APPALACHIANS THROUGH
SATURDAY...WITH ISOLATED MAXIMUM AMOUNTS
OF 15 INCHES.



       6 May 2008                                          20
                        How Much QPF?
• Depends on multiple factors
     •     Availability of moisture
           •     Moisture at T=0
           •     Moisture flux – more important than the initial moisture!
     •     Intensity of Precipitation
     •     Speed and direction of movement
     •     Training cells
     •     Structured banding
     •     Enhancement by boundaries and/or topography
     •     Precipitation efficiency
•     No “proven” Quantitative method for QPF Yet!
    6 May 2008                                                               21
                       HPC TC QPF
• HPC TC QPF process
  incorporates:
  • Global models
      • GFS, ECMWF
  • Mesoscale models
      • NAM
  • Ensembles
      • SREF
  • Hurricane models
      • GFDL, HWRF
  • Climatological models
      • R-CLIPER
  • Satellite techniques
      • TRaP
  • Extensive in-house TC rainfall
    climatology and research
  • Official NHC track forecast

    6 May 2008                       22
        HPC TC QPF Methodology
•   Forecasts made in 6-h increments from 0-84 h and in
    one 48-h period for days 4-5 by 3 forecast desks:
    •        Day 1 QPF
    •        Day 2/3 QPF
    •        Medium Range (Days 4-5)
•   Starting point is model closest to NHC track forecast
    •        Usually GFS
•   Locate relevant synoptic scale and mesoscale
    features (troughs, fronts, etc.)
•   Use conceptual models, current structure, and pattern
    recognition to modify/shift QPF
•   Use TRaP and recent satellite and radar imagery to
    analyze current structure and rainfall rates

6 May 2008                                              23
            HPC TC QPF Methodology
•   Look at forecast storm-relative shear and 200-mb winds to further shift/limit
    QPF
•   Use climatology (PRISM, R-CLIPER, TC Rainfall Climatology) to:
    •    Adjust amounts and numerical guidance biases
    •    Depict areas of terrain that could be significantly impacted
    •   Provide a reality check
•   Create TC rainfall statements for the Public Advisories
•   Day 1-3 forecasts issued by 06z/10z/18z/22z
•   Day 4-5 forecasts and 5-day total graphic issued at 00z/12z




    6 May 2008                                                                24
                 Day 1-3 QPF                                    Day 1-5 QPF
                TC Rainfall Climatology
             United States TC Rain Maxima By State




6 May 2008                                           25
                     Ensembles
•   Can help quantify
    degree of confidence
    in a forecast and
    identifying range of
    possible solutions
•   Examine spaghetti plot
    of precipitation
    isohyets from
    ensemble members to
    see areas of overlap
    and spread
•   Don’t use ensemble
    mean QPF – dampens
    extreme events
                             1” isohyets from SREF members
    6 May 2008                                           26
                      HPC QPF Verification
• 24-hour Day 1 QPFs have
  been verified since 1961
• Verified on common 32-km
  grid based on manual
  observational precipitation
  analysis performed at HPC
• TC QPF verification has been
  performed on a “regional”
  scale for the past 3 seasons
                  from a TC may
    • Precipitation
    straddle more than one region
    • Other systems may influence
    precipitation in a region
                                    HPC QPF Verification Regions
• New for 2007 – ECMWF QPF
  verification



     6 May 2008                                             27
                      Verification Metrics
Threat Score
            Correct                        .




 (Forecast + Observed - Correct)

 Measures fraction of observed and/or
 forecast events that were correctly
 predicted; Sensitive to hits, penalizes
 both misses and false alarms

 For a perfect forecast:
   Correct = Forecast = Observed
     to yield a TS of 1
 The worst possible forecast:
     Correct = 0, yields a TS of
 zero



     6 May 2008                                28
                   Verification Metrics
Bias
           Forecast    _
           Observed

Gauges accuracy of areal/station
  coverage of specified
  threshold amount, regardless
  of accuracy in location

Ideal forecast:
       Forecast = Observed
            Bias = 1




      6 May 2008                          29
               Seasonal Verification
• Verification of all Day 1 QPF for TCs and TC remnants
  with precipitation impacts in the CONUS
   • 2005: Arlene, Cindy, Dennis, Katrina, Ophelia, Rita, Tammy,
     Wilma (8)
   • 2006: Alberto, Ernesto (2)
   • 2007: Andrea, Barry, Erin, Gabrielle, Humberto, Henriette (East
     Pacific), TD 10, Noel (8)
• Seasonal threat score and bias from HPC, GFS, NAM,
  and ECMWF computed
• Major changes occurred to all three models during this
  period
   • Difficult to identify consistent biases from season to season



  6 May 2008                                                         30
      2005 Seasonal Verification
                 Summary Statistics for Major Landfalling Hurricanes - 2005
               Arlene, Cindy, Dennis, Katrina, Ophelia, Rita, Tammy & Wilma
                                Day 1 Threat Score and Bias


       0.6                                                                        2
                                                                                  1.5
       0.4
                                                                                  1
       0.2
                                                                                  0.5
        0                                                                         0
             0.5         1        2          3          4        5            6
                                      Threshold (.in)

              HPC TS            GFS TS             NAM TS            ECMWF TS
              HPC BIAS          GFS BIAS           NAM BIAS          ECMWF BIAS


•NAM tends to under-forecast TC precipitation, especially at higher thresholds
•GFS has best model bias from 1-6 in.; NAM best at 0.5 in.
•ECMWF TS competitive with GFS across all thresholds; ECMWF bias slightly lower
than GFS
•HPC Bias generally too high, especially at higher thresholds
•HPC TS better than the NAM and GFS at all thresholds for tropical systems in 2005
6 May 2008                                                                              31
                                                        2005 Seasonal Verification
                                                        HPC D1 TS Improvement over NAM/GFS
                                                      Summary Statistics for Tropical Cyclones - 2005
                                            Arlene, Cindy, Dennis, Katrina, Ophelia, Rita, Tammy & Wilma
% Improvement over NAM -




                                                                                                        1313.79 %                   3393.15 %
                                500%
                                                                                408.80%
                                400%
                           Day 1 TS




                                300%
                                                                     173.15%
                                200%

                                100%                      50.45%
                                             24.38%
                                      0%
                                              0.5            1          2           3                                    4            5
                                                                     Threshold (inches)                                              Summary Statistics for Tropical Cyclones- 2005
        NAM                                                        HPC improvement                                           Arlene, Cindy, Dennis, Katrina, Ophelia, Rita, Tammy & Wilma

                                                                                                                   25%
                                                                                        % Improvement over GFS -




                                                                                                                                                                                 19.53%
                                                                                                                   20%
                                                                                                                                                             14.23%    14.53%
                                                                                                Day 1 TS




                                                                                                                   15%
                                                                                                                                                 10.60%
                                                                                                                   10%                  8.26%
                                                                                                                             6.37%
                                                                                                                   5%                                                                       3.13%

                                                                                                                   0%
                                                                                                                              0.5         1         2           3            4     5           6
                                                                                                                                                        Threshold (inches)                32
                                           6 May 2008
                                                                                                   GFS                                                   HPC improvement
       2006 Seasonal Verification
                   QPF Summary Statistics for Landfalling Tropical Cyclones-2006
                                       Alberto and Ernesto
                                 Day 1 Threat Score and Bias


         0.6                                                                           4
         0.5                                                                           3.5
                                                                                       3
         0.4                                                                           2.5
         0.3                                                                           2
         0.2                                                                           1.5
                                                                                       1
         0.1                                                                           0.5
              0                                                                        0
                  0.5        1        2          3          4        5             6
                                          Threshold (.in)
                  HPC TS            GFS TS             NAM TS            ECMWF TS
                  HPC BIAS          GFS BIAS           NAM BIAS          ECMWF BIAS


•ECMWF had highest model threat score across all thresholds through 4 in.
•GFS and ECMWF had biases near 1 through 3 in. threshold
•GFS had pronounced high bias at 4 in. and above
•HPC showed most improvement in TS at higher amounts at the expense of high bias
•NAM showed lowest TS and worst bias for amounts of 3 in. or greater
 6 May 2008                                                                                  33
       2007 Seasonal Verification
                    Summary Statistics for Landfalling Tropical Cyclones-2007
                                 Day 1 Threat Score and Bias


        0.6                                                                           2.5

        0.5                                                                           2
        0.4
                                                                                      1.5
        0.3
                                                                                      1
        0.2

        0.1                                                                           0.5

         0                                                                            0
              0.5         1          2          3          4         5          6
                                         Threshold (.in)
               HPC TS              GFS TS            NAM TS              ECMWF TS
               HPC BIAS            GFS BIAS          NAM BIAS            ECMWF BIAS


•GFS provided best overall model QPF guidance
•HPC showed low bias at highest thresholds, reversing trend of 2005 and 2006
•ECMWF performance worse than previous years, especially above 2 in. threshold
•NAM continued to have lowest TS and severe low bias at high amounts
•HPC had a better or the same TS as the best model for all 2007 TC total stats Days
1-3 (except amounts > 3” on Day 2)
 6 May 2008                                                                                 34
                                   Gabrielle
•Isolated area of heavy rainfall
along central NC coast
•Not well-predicted by models
which struggled with track and
structure of the system
•HPC showed significant
improvement on heavier
amounts, particularly on day 2
                                                                 35   7
                                                             1




                                    Gabrielle Rainfall Analysis by David Roth - HPC


      6 May 2008                                                               35
                                    Gabrielle
    Day 3 QPF valid for 24-h ending 12Z 10 September 2007
                              Gabrielle - September 10, 2007
    HPC                         Day 3 Threat Scores and Bias                GFS

             0.4                                                      2
         0.35                                                         1.8
                                                                      1.6
             0.3
                                                                      1.4
         0.25                                                         1.2
             0.2                                                      1
         0.15                                                         0.8
                                                                      0.6
             0.1
                                                                      0.4
         0.05                                                         0.2
   NAM        0                                                       0ECMWF
                      0.5           1          2          3    4
                                        Threshold (.in)
                   HPC TS           GFS TS          NAM TS     ECMWF TS
                   HPC BIAS         GFS BIAS        NAM BIAS   ECMWF BIAS

         • HPC showed some skill at 1” threshold
         • All models had zero bias for amounts greater than 0.5”

6 May 2008                                                                        36
                                  Gabrielle
   Day 2 QPF valid for 24-h ending 12Z 10 September 2007
                            Gabrielle - September 10, 2007
   HPC                        Day 2 Threat Scores and Bias
                                                                          GFS

         0.35                                                       4.5
           0.3                                                      4
                                                                    3.5
         0.25
                                                                    3
           0.2                                                      2.5
         0.15                                                       2
                                                                    1.5
           0.1
                                                                    1
         0.05                                                       0.5
   NAM       0                                                      0   ECMWF
                    0.5           1          2          3    4
                                      Threshold (.in)
                 HPC TS           GFS TS          NAM TS     ECMWF TS
                 HPC BIAS         GFS BIAS        NAM BIAS   ECMWF BIAS

   • HPC successfully forecasted location of maximum rainfall and
     showed major improvement over the models for amounts > 0.50”
   • Higher HPC TS at 3” came at expense of high bias
     • NAM
6 May 2008   had best model forecast – only model to depict any amounts         37
    > 1”
                                      Gabrielle
    Day 1 QPF valid for 24-h ending 12Z 10 September 2007

    HPC                        Gabrielle - September 10, 2007                  GFS
                                 Day 1 Threat Scores and Bias

              0.3                                                         5
             0.25                                                         4
              0.2
                                                                          3
             0.15
                                                                          2
              0.1
             0.05                                                         1

               0                                                          0
   NAM                  0.5          1          2
                                          Threshold (.in)     3    4     ECMWF
                    HPC TS           GFS TS          NAM TS       ECMWF TS
                    HPC BIAS         GFS BIAS        NAM BIAS     ECMWF BIAS

    • HPC and model QPF displaced too far northeast
    • None of the models hit the amounts > 1”
    • The NAM had the best TS for 0.50” & 1”
    • All of the models had a high bias through 2” except ECMWF
6 May 2008                                                                           38
       HPC Inland TC Advisories




6 May 2008                        39
                 HPC Inland Advisories
• HPC issues advisories on TCs that have moved inland and weakened
  below tropical storm strength but still pose a threat of heavy rains and
  flash floods in the CONUS or adjacent areas within Mexico
• Advisories continue even if the low-level TC remnant center dissipates
  as long as flood threat remains
• HPC advisories cease when threat of flash flooding has ended
• Advisories issued at 03, 09, 15, and 21 UTC under the following
  headers:
    • Atlantic: TCPAT(1-5); WTNT3(1-5) KWNH
    • East Pacific: TCPEP(1-5); WTPZ3(1-5) KWNH
    • Links to HPC advisories can be found on HPC webpage, NHC webpage
      and graphical TWO




    6 May 2008                                                         40
               HPC Inland Advisories
• Primarily focus on potential for heavy rain and flash
  flooding
• Contents of HPC advisories include:
   • Headline describing the current event
   • General summary of current watches, warnings, and advisories
   • Description of the current location, maximum sustained winds,
     movement, and minimum central pressure of the system
   • Table listing precipitation induced by the system
   • Description of the evolution and forecast for the system
   • Statement including time of next HPC advisory or, if the final
     HPC advisory, where to find subsequent information
   • Forecast for the position of the surface center (if it still exists) as
     long as the system is a flash flooding threat for the CONUS


  6 May 2008                                                             41
HPC Inland Advisories Issued Annually
                                   2002-2007
                           Annual Count of HPC Inland TC Advisories


  50
                                       45
  45

  40                                                    38

  35         34

                                                                             30
  30
                    26
  25

  20

  15

  10
                                                                       7

   5

   0
             2002   2003              2004             2005           2006   2007
                                               Year




6 May 2008                                                                          42
             NHC Continuity of
              Operations Plan




6 May 2008                       43
                        NHC COOP
• HPC is the designated
  backup for NHC (since at
  least late 1970s)
• HPC will assume forecast
  responsibility for Atlantic
  and East Pacific TCs in the
  event of communications
  outage or other problems
  at NHC
• Approximately 20 HPC
  forecasters trained in:
   • TC analysis and forecasting
   • Use of ATCF software
   • Hurricane Program Policy
    6 May 2008                     44
                            NHC COOP
• Major technical effort to
  maintain software, scripts,
  code and to update with
  annual changes to
  products, graphics, etc.
• Annual in-house exercise,
  planned and unannounced
  backup drills performed to
  maintain proficiency
• Implementation of COOP
  designed to be seamless
  and transparent to users
    • Operational NHC
      products reproduced
    • NHC website maintained




      6 May 2008                       45
                       NHC COOP
• COOP program overhaul began in 2002-2003
  • 2002
        • Simple communications tests to ensure NHC products could
          be transmitted from HPC
  • 2003
        • All GS-14 and many GS-13 forecasters trained for NHC
          backup
        • Began back-up drills on active TCs (issuing full forecast
          package)
  • 2005
        • Began issuing East Pacific TWO on the 1st Tuesday of every
          month to test communications and proficiency




 6 May 2008                                                           46
               NHC COOP
• HPC has never officially taken over for
  NHC, but there have been at least two
  close calls:
  • 2003 or 2004: NHC requested backup due to
    communications problems, but recovered
    before HPC had to issue products
  • Hurricane Andrew (1992): NHC sent
    forecasters to NMC for backup prior to the
    storm hitting Miami, but NHC remained up
    during and after the storm

 6 May 2008                                 47
              Acknowledgements
• HPC Staff, especially
  •   Mark Klein
  •   Jessica Clark
  •   David Roth
  •   Wes Junker
• Chris Lauer of TPC




 6 May 2008                      48
                          Questions??




                      NESDIS
                               Tropical/      QPF             Media
              Basic                                           Room
                               Winter Wx
Surface       Wx                      Model
Analysis                              Diag.
                                                    Medium Range




                  http://www.hpc.ncep.noaa.gov/
                 Michael.J.Brennan@noaa.gov
 6 May 2008                                                           49

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:4
posted:3/12/2012
language:
pages:49