motivation

21
The Quantitative Precipitation Forecasting Component of the 2010 NOAA Hazardous Weather Testbed Spring Experiment Faye E. Barthold 1,2 , Michael J. Bodner 1 , David R. Novak 1 , Richard Bann 1 , Robert Oravec 1 , Bruce Sullivan 1 , Andrew R. Dean 3 , Israel L. Jirak 3 , Christopher J. Melick 3 , Ryan A. Sobash 4 , Adam J. Clark 5 , Fanyou Kong 4,6 , Steven J. Weiss 3 , and Ming Xue 4,6 1 NOAA/NWS/Hydrometeorological Prediction Center 2 I.M. Systems Group, Inc. 3 NOAA/NWS/Storm Prediction Center 4 University of Oklahoma 5 NOAA/National Severe Storms Laboratory 6 Center for Analysis and Prediction of Storms

Upload: lala

Post on 19-Mar-2016

27 views

Category:

Documents


2 download

DESCRIPTION

The Quantitative Precipitation Forecasting Component of the 2010 NOAA Hazardous Weather Testbed Spring Experiment. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Motivation

The Quantitative Precipitation Forecasting Component of the

2010 NOAA Hazardous Weather Testbed Spring Experiment

Faye E. Barthold1,2, Michael J. Bodner1, David R. Novak1, Richard Bann1, Robert Oravec1, Bruce Sullivan1, Andrew R. Dean3, Israel L.

Jirak3, Christopher J. Melick3, Ryan A. Sobash4, Adam J. Clark5, Fanyou Kong4,6, Steven J. Weiss3, and Ming Xue4,6

1NOAA/NWS/Hydrometeorological Prediction Center2I.M. Systems Group, Inc.

3NOAA/NWS/Storm Prediction Center4University of Oklahoma

5NOAA/National Severe Storms Laboratory6Center for Analysis and Prediction of Storms

Page 2: Motivation

Motivation• Flash flooding is a leading cause of weather

related deaths in the U.S.– Typically occurs in the warm season

• Current operationalmodels don’t handleconvection well

• HPC threat scorelowest in the warmseason

HPC Monthly 1.00" Threat Score (February 2006 - November 2010)

0

0.1

0.2

0.3

0.4

0.5

0.6

Page 3: Motivation

QPF Component• GOAL: Determine whether high resolution

convection allowing guidance provides added value over the current operational models

• Forecasting Activities– Probabilistic 6 hr QPF forecasts

• valid at 00Z and 06Z• 0.50” and 1.0” thresholds• Indicate highest possible

amount within any 1.0” area– Forecast discussion– Evaluation of experimental

forecasts and modelperformance

Page 4: Motivation

High Resolution Model Data

Provider Model Delta X Notes Label

CAPSWRF/ARPS 26 member ensemble

4 km

Multi-model, multi-physics, multi-IC ensemble system with radar assimilation

SSEF

CAPS WRF-ARW 1 km ARPS 3DVAR ICs and NAM LBCs CAPS-ARW1

NCAR WRF-ARW 3 km RUC ICs and GFS LBCs WRF-NCAR3

NSSL WRF-ARW 4 km NAM ICs and LBCs WRF-NSSL4

NCEP/EMC WRF-NMM 4 km NAM ICs and LBCs WRF-NMM4

NCEP/EMC WRF-ARW 5.1 km NAM ICs and LBCs HRWE-ARW4

Page 5: Motivation

Experimental Ensemble Products

– Probability matched mean—combines the spatial pattern of the ensemble mean QPF with the frequency distribution of the rainfall rates to provide a more realistic ensemble rainfall intensity forecast (Ebert 2001)

– Ensemble maximum precipitation—maximum QPF from any ensemble member

– Neighborhood probabilities—probability of an event within ~80 km of a point

Page 6: Motivation

Forecast Valid 00Z 21 May 2010

6hr NSSLQ2 QPE valid 00Z 21 May 2010 NAM12 6 hr QPF (24 hr forecast)

Page 7: Motivation

Forecast Valid 00Z 21 May 2010

6hr NSSLQ2 QPE valid 00Z 21 May 2010 WRF-NSSL4 6 hr QPF (24 hr forecast)

Page 8: Motivation

Forecast Valid 00Z 21 May 2010

6hr NSSLQ2 QPE valid 00Z 21 May 2010 WRF-NMM4 6 hr QPF (24 hr forecast)

Page 9: Motivation

Forecast Valid 00Z 21 May 2010

6hr NSSLQ2 QPE valid 00Z 21 May 2010 HRWE-ARW4 6 hr QPF (24 hr forecast)

Page 10: Motivation

Forecast Valid 00Z 21 May 2010

6hr NSSLQ2 QPE valid 00Z 21 May 2010 SREF mean 6 hr QPF (27 hr forecast)

Page 11: Motivation

Forecast Valid 00Z 21 May 2010

6hr NSSLQ2 QPE valid 00Z 21 May 2010 SSEF mean 6 hr QPF (24 hr forecast)

Page 12: Motivation

Forecast Valid 00Z 2 June 20106hr NSSLQ2 QPE valid 00Z 2 June 2010 NAM12 6 hr QPF (24 hr forecast)

Page 13: Motivation

Forecast Valid 00Z 2 June 20106hr NSSLQ2 QPE valid 00Z 2 June 2010 WRF-NSSL4 6 hr QPF (24 hr forecast)

Page 14: Motivation

Forecast Valid 00Z 2 June 20106hr NSSLQ2 QPE valid 00Z 2 June 2010 WRF-NMM4 6 hr QPF (24 hr forecast)

Page 15: Motivation

Forecast Valid 00Z 2 June 20106hr NSSLQ2 QPE valid 00Z 2 June 2010 HRWE-ARW4 6 hr QPF (24 hr forecast)

Page 16: Motivation

ResultsSubjective Verification

2010 HWT Spring ExperimentHigh Resolution Model Performance Compared to the

NAM/SREF

9/3013/4211/28

14/28

24/42

26/40

12/3018/42

7/287/28

7/42

3/40

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

SSEF WRF-NSSL4

CAPS-ARW1

HRWE-ARW4

WRF-NMM4

WRF-NCAR3

Model

Perc

enta

ge

improved guidance worse guidance

Page 17: Motivation

ResultsObjective Verification

Courtesy Adam Clark, NSSL

Page 18: Motivation

ResultsEnsemble Products

– Ensemble mean—useful, provided a realistic depiction of precip amounts and coverage

– Probability matched mean—question about validity of using this technique on a national scale

•Future work: recalculate using a regional scheme

– Neighborhood probabilities—probabilities often too high and coverage too broad

•Future work: recalculate using different smoothing parameters

– Ensemble maximum precipitation—not useful, values unrealistically high

Page 19: Motivation

Convection-Allowing Guidance at HPC

Up to five runs available per cycle• NCEP High Res Window ARW (east & west)• NCEP High Res Window NMM (east & west)• NCEP NMMB (new NAM)• NCEP NMM• NSSL WRF (ARW core)

547 AM EDT TUE JUL 20 2010VERY FRUSTRATING QPF PATTERN…PIECES OF SHRTWV ENERGY FIRING CNVCTN WHICH THEN…BEGINS TO TAKE ON A LIFE OF ITS OWN…THE BULK OF MODEL GUIDANCE HAS WOUND UP BEING TOO FAR NORTH WITH THE AXIS OF HEAVIEST PCPN. THE HI RES ARW HAS DONE A MUCH BETTER JOB THAN NCEP AND NON-NCEP MODEL SUITES IN SHOWING THIS SRN DISPLACEMENT...

Page 20: Motivation

Conclusions• While not perfect, high resolution convection

allowing models are useful and can improve warm season QPF forecasts

• SSEF represents a transformational improvement in warm season QPF

• Experiment impact at HPC– Forecasters interested and using high-res guidance– New data available (WRF-NSSL4 and NMMB nest)– Renewed emphasis on verification of excessive rainfall

events

Page 21: Motivation

Future Plans• Add the 06-12Z forecast period to capture the nocturnal

precip max

• Add afternoon forecasting activities– Update the 00-06Z forecast period– Forecast for the 06-12Z forecast period

• Expand the available post-processed guidance– e.g., development and evaluation of bias-corrected QPF

• Expand hydromet participation

• 2010 data available for post-analysis