1

Conference CallFebruary 12, 2012

• Recent project news• Climatology paper outline• Modeling results• Next steps

2

•Accurate prediction of the tropical cyclone wind field after landfall is one of the greatest challenges for operational forecasters.

•Many past studies have examined the evolution of the tropical cyclone wind field after landfall (Wong et al. 2008; Bhowmik et al. 2005; and Kaplan and DeMaria 2001).

•The results of these studies have not been routinely incorporated into the techniques used by NWS forecasters in operational prediction.

This work is funded by NOAA Grant NA10NWS4680007

Introduction

A New Approach to Improved Inland Wind Forecasts for Landfalling Tropical Cyclones

Bryce Tyner and Anantha AiyyerDepartment of Marine, Earth, and Atmospheric Sciences

North Carolina State University

Acknowledgements

Bhowmik, S.K.R., S.D. Kotal, and S.R. Kalsi, 2005: An Empirical Model for Predicting the Decay of Tropical Cyclone Wind Speed after Landfall over the Indian Region. J. Appl. Meteor., 44, 179-185.

Kaplan, J., and M. DeMaria, 2001: A note on the decay of tropical cyclone winds after landfall in the New England area. J. Appl. Meteor., 40, 280-286.

Wong, M.L.M., J.C.L. Chan, and W. Zhou, 2008: A Simple Empirical Model for Estimating the Intensity Change of Tropical Cyclones after Landfall along the South China Coast. J. Appl. Meteor. Climatol.,47, 326-338.

References

Ongoing/Future WorkWind Speed Climatology

Data and Methods

•The overall goal of this study is to improve the wind speed and wind gust forecasts associated with tropical cyclones.

•The primary focuses for improvement are in the land reduction factors and gust factors used in the TCMWindTool.

Purpose

•Hourly wind speed observations for Maryland, Virginia, North Carolina, and South Carolina were obtained from the North Carolina State Climate Office CRONOS database for the months June-November 2005-2010. These wind speed observations were used to develop a climatology of sustained wind speeds in the region.

•Hourly wind speed observations were compared to surface wind analyses from the Hurricane Research Division’s H*Wind project.

•Sustained speed forecasts of recent tropical cyclones affecting the region were compared to both H*Wind forecasts and surface wind speed observations. Ernesto (2006), Gabrielle (2007), Hanna (2008), and Irene (2011) were landfalling tropical cyclones in the region with forecast data available. Cristobal (2008) and Earl (2010) were also examined, although they did not make landfall in the region.

•Land decay factors were analyzed for all tropical cyclones affecting the region 1950-2008 (not shown). Only the ten closest stations to the center of the tropical cyclone were used in this analysis.

Numerous past studies have shown surface winds can be best described by a Weibull distribution. The two-parameterWeibull distribution is defined as:

where:η = scale parameter,β = shape parameter (or slope),

Irene (2011)

•Areas of highest elevation and locations near the coastline have the largest shape and scale parameters, indicative of highest mean wind speed values as well as largest spread in wind speed distributions.

•There is a distinct minimum in mean sustained wind speeds as well as the spread in the wind speed distributions for much of Virginia and west-central North Carolina.

•Topographic influences lead to significantly different mean wind speeds and distributions for areas in the same local region.

•The highest gust factors (>1.4) were observed in areas where the sustained wind speeds were much weaker. This occurred in areas further inland as well as after the storm passage in many areas. Weaker wind speeds were also consistent with higher variability in the gust factors.

•In areas with strongest sustained wind speeds, gust factor values were lower, near 1.2, and variability in the gust factors was also lower.

•Many coastal sites and areas near the mountains did not observe a significant increase in average values and variability in gust factors after the storm passed.

•Boundaries of WFOs appear in analysis•Raleigh WFO appears to have smallest difference between forecast and HWind analysis (where forecasters used strongest land reduction factors of 33%)•Strongest over prediction of wind speeds present in coastal regions

NDFD – HWIND Analysis at (a) 0730 27 August, (b) 1330 27 August, (c) 1930 27 August, and (d) 0130 28 August (UTC)

(a) (b)

(c) (d)

Output Grid: Land Reduction Factors Ouput Grid: Gust Factors

Grid File: Land Use/Terrain Data

Basic Storm InformationStrength of StormSize of StormAngle of storm approachStorm propagation speed

NHC Best Track Data

Grid File: Thermodynamic/Environmental ConditionsStatic StabilityBoundary layer conditionsET Transition?Cold Air Damming?

Proposed Final Product to be Used by NWS Forecasters:

Simulated base radar reflectivity (left) and KMHX WSR-88Dradar reflectivity (right) near the time of landfall for Irene (2011)

Radar Reflectivity (dBZ)

•18/3 km nested run initialized 48 hours prior to landfallaccurately forecasted the location of landfall as well asstorm strength (model central SLP=950 hPa, best trackSLP=952 hPa)•The time of landfall was delayed in the model by approx.6 hours. Sensitivity studies are being conducted to produce a simulation with a more accurate landfall time.•Once a realistic simulation is produced, WRF-LES simulations will be conducted in order to examine the turbulence profiles for various locations in the model domain.

3

•The highest gust factors (>1.4) were observed in areas where the sustained wind speeds were much weaker. This occurred in areas further inland as well as after the storm passage in many areas. Weaker wind speeds were also consistent with higher variability in the gust factors.

•In areas with strongest sustained wind speeds, gust factor values were lower, near 1.2, and variability in the gust factors was also lower.

•Many coastal sites and areas near the mountains did not observe a significant increase in average values and variability in gust factors after the storm passed.

4

Climatology Paper

• Proposed Title: “A Climatological Look at Sustained Wind Speeds and Gusts of Landfalling Tropical Cyclones in the Mid-Atlantic, 2005-2011”

• Proposed Journal: Weather and Forecasting• Submission Date Goal: May 2012

5

Climatology Paper Outline

• Introduction– Define Weibull distribution– Define gust factor– Literature survey on land decay papers and gust factors for the

region

• Methods/Data– NHC Best Track Data– HWind Surface Wind Analyses– Surface observations—1 min data for all ASOS stations– NC CRONOS data base (for land decay)– NDFD forecasts

6

Climatology Paper Outline

• Results– Weibull distributions of wind speeds and gusts during

times of landfalling TCs vs. climatology– Land decay for 10 (?) closest points to storm from

CRONOS database with high order polynomial fit– NDFD verification for storms (using both CRONOS

and HWind Analyses)– Spatial gust factor analysis for storms using ASOS

data– Frequency of missing observations from ASOS data

7

Model Runs for Irene (2011)

As of 2/21/12

8

Runs Conducted Thus Far

All runs:• Initialized 8/25/1200 UTC • Terminated 8/29/12 UTC• Approximately 48 hours prior to best track

landfall

9

Runs Conducted Thus Far

Run 1: • 18 / 6 km nested run• Kain Fritsch convective scheme• ERA_Interim data• Model SST used

10

Runs Conducted Thus Far

Run 2: • 9 / 3 km nested run• K.F. convective scheme (outer domain)• ERA_Interim data• Model SST used

11

Runs Conducted Thus Far

Run 3: • 18 / 6 km nested run• K.F. convective scheme (outer domain)• GFS analysis data• Model SST used

12

Runs Conducted Thus Far

Run 4: • 18 / 6 km nested run• Kain Fritsch convective scheme• ERA_Interim data• RTG SST data from 8/26/00 UTC used

13

Plots to Compare

• Simulated vs. radar mosaic• Best track SLP at time of landfall• 10 m winds vs. HWIND at time of landfall• Landfall location• Track after landfall

14

Landfall Verification: Run 1

SIMULATED LANDFALL TIME: 08/27/1800 UTCLANDFALL CENTRAL SLP: 951 hPa

BEST TRACK LANDFALL TIME: 08/27/1200 UTCLANDFALL CENTRAL SLP: 952 hPa

15

Landfall Verification: Run 2

SIMULATED LANDFALL TIME: 08/27/1900 UTCLANDFALL CENTRAL SLP: 950 hPa

BEST TRACK LANDFALL TIME: 08/27/1200 UTCLANDFALL CENTRAL SLP: 952 hPa

LANDFALL LOCATION ABOUT SAME AS IN RUN #1, BUT SLIGHTLYSLOWER AND STRONGER

16

Landfall Verification: Run 3

SIMULATED LANDFALL TIME: 08/27/1600 UTCLANDFALL CENTRAL SLP: 952 hPa

BEST TRACK LANDFALL TIME: 08/27/1200 UTCLANDFALL CENTRAL SLP: 952 hPa

LANDFALL LOCATION ABOUT SAME AS IN RUN #1, BUT SLIGHTLYQUICKER AND WEAKER

17

Landfall Verification: Run 4

SIMULATED LANDFALL TIME: 08/27/1500 UTCLANDFALL CENTRAL SLP: 961 hPa

BEST TRACK LANDFALL TIME: 08/27/1200 UTCLANDFALL CENTRAL SLP: 952 hPa

LANDFALL LOCATION ABOUT SAME AS OTHER RUNS, BUT MUCH WEAKER AND SLOWER

18

HWIND ANALYSIS, 8/27/1330 UTC Model Run 4, 8/27/1300 UTC

19

Landfall Verification: Run 4

GENERAL STORM STRUCTURE IS CORRECTLY SIMULATED (BANDS AHEAD OF STORM, WESTWARD EXTENT OF PRECIPITATION, EASTERN SIDE WINDS)

20

NEXT RUN

• RTG SST data updated every 6 hours• Once we have a run we are “happy with”,

we can begin WRF-LES simulations

21

Current/Future Work

• Work on writing climatology paper• Get WRF run for Irene that is “reasonable”• Reading on hurricane WRF LES

simulations (Zhu et al. 2008, Chen et al. 2010, etc)

• Develop first WRF LES run for Irene, consult with Dr. Sukanta Basu

22

Output Grid: Land Reduction Factors Ouput Grid: Gust Factors

Grid File: Land Use/Terrain Data

Basic Storm InformationStrength of StormSize of StormAngle of storm approachStorm propagation speed

NHC Best Track Data

Grid File: Thermodynamic/Environmental ConditionsStatic StabilityBoundary layer conditionsET Transition?Cold Air Damming?