2016 17 Weather Prediction Center WPC Winter Weather

2016 -17 Weather Prediction Center (WPC) Winter Weather Desk Operations • Dan Petersen, Bruce Veenhuis, Mike Bodner, Mark Klein, Chris Bailey

Winter Weather Forecast Desk Updates • • Days 1 -3 Probabilistic Snow/Ice Forecasts Revised freezing rain accumulation calculation Days 4 -7 Winter Weather Forecasts Collaboration updates • Forecast Verification

• 2016 -17 Winter Weather Desk Operations and Changes Workflow: Winter Weather Desk remains operational 24 hours, 7 days a week Day Shift 12 -21 z Night shift 00 -09 z. Coverage and forecast updates between shifts 21 z-00 z and 09 -12 z event-driven Probability/percentile/deterministic forecasts of snowfall and freezing rain over the contiguous US in 12, 24, 48 and 72 -hour increments Revised winter weather web page and graphics

Probabilistic Winter Precipitation 2015 -16 Forecast Ensemble Composition (day shift) 63 members total (SREF 41 percent of all members) • 26 SREF members • 25 ECMWF ensemble members, randomly selected (00 Z) • 1 NAM operational run (12 Z) • 1 GFS operational run (12 Z) • 1 ECMWF operational run (00 Z) • 1 Canadian Global Model (CMC) operational run (00 Z) • 1 ECMWF ensemble mean (00 Z) • 1 GFS ensemble mean (06 Z) • 5 GFS ensemble members, randomly selected (06 Z) • WPC deterministic forecast

2016 -17 Probabilistic Winter Precipitation Forecast Ensemble Composition (day shift) 70 members total (63 members last season) New Members: • 10 GFS ensemble members (06 Z, up from 5 201516) • 1 High res WRF ARW (00 z) • 1 High res WRF NMMB (00 z) • 2 prior cycle runs of GFS after ARW/NMMB Days 2 -3 ftp: //ftp. wpc. ncep. noaa. gov/pwpf/conus_2. 5 km

2016 -17 Forecast Ensemble Composition (day shift) Day 1 Day 2 Day 3 1 WRF ARW 00 Z GFS 00 Z 1 WRF NMMB 00 Z GFS 06 Z 1 NAM Nest 12 Z NAM Nest 12 z Parent NAM 12 z 1 GFS 12 Z 1 Canadian Model (CMC) 00 Z 1 ECMWF Operational 00 Z 1 ECMWF Ensemble Mean 00 Z 1 GEFS Ensemble Mean 06 z 10 GEFS Ensemble Members 06 z 1 WPC Deterministic 26 SREF Ensemble Members SREF Ensemble 09 Z Members 09 Z SREF Ensemble Members 09 Z 25 ECMWF Ensemble Members 00 Z WPC Deterministic ECMWF Ensemble Members 00 Z

Web Site Display Changes for 2016 -17 New Winter weather page http: //www. wpc. ncep. noaa. gov/#page=wwx

Web Site Display Changes for 2016 -17 Transition to ESRI Maps http: //www. wpc. ncep. noaa. gov/pwpf/wwd_accum_probs. php

2016 -17 Freezing Rain Accumulation (FRAM) Calculation Freezing rain accumulation rates decrease as precipitation rate increases(top), increase with wind speed (middle), and decrease as wet bulb temperatures increase (bottom) Sanders and Barjenbruch, 2016 http: //journals. ametsoc. org/doi/pdf/10. 1175/WAF-D-15 -0118. 1

2016 -17 Freezing Rain Accumulation (FRAM) Calculation • e. The median Ice to Liquid Ratio for elevated horizontal ice accumulation was 0. 72: 1 with a 25 th percentile of 0. 50: 1 and a 75 th percentile of 1. 0: 1 WPC will adopt the equation covering the precip rate, which accounts for 80 percent of the weight of the factors. Ice to Liquid Ratio = 0. 1395 x (P)exp-0. 541 where P=Precipitation Rate http: //journals. ametsoc. org/doi/pdf/10. 1175/WAF-D-15 -0118. 1

Days 4 -7 Probability Forecasts Probability of exceeding 0. 25” of combined snow/sleet Forecasts for 2016 -17 will be provided from 12 z to the following 12 z (24 h intervals) for Days 4, 5, 6, and 7 both on day and night shift http: //www. wpc. ncep. noaa. gov/wwd/internal/pwpf_d 47/pwpf_medr. php http: //www. nws. noaa. gov/os/notification/tin 16 -34 day 4 -7 winter_ndfd. htm

Days 4 -7 Probability Forecasts for Day 4, Day 5, Day 6, and Day 7 Probability of WPC QPF >=. 25 “ Ensemble Probability of Snow/Ice Pellets > 0. 25”

Days 4 -7 Probability Forecasts Methodology Probability of Frozen Precipitation • Determination of precipitation type via GEFS, CMCE members • Average of GEFS + CMCE ensemble mean QPFs used to disaggregate Days 4 -5 and Day 6 -7 WPC QPF • http: //www. wpc. ncep. noaa. gov/wwd/pwpf_d 47/WPC_Win. Wx_PDD. pdf

Winter Weather Product Days. Day 4 -74 -7 Probability Forecasts We edit draft probabilities, including those from the ECMWF ensembles (upper right) 24 -hour probability of at least 0. 10” of winter precipitation(ending 2/17/2015, 12 Z)

Winter Weather Watch Collaborator Summary option Probability option Forecasts of probability of exceeding WFO watch/warning criteria, with summary maps where probabilities exceeded 30 or 50 percent. Displays show summary graphics and probabilities for each time step www. wpc. ncep. noaa. gov/wwd/internal/watchcollab/watch_collaborator. php

Winter Weather Watch Collaborator When you select the probability forecasts, The forecasts are available in 6 hour time steps controlled by the forward and backward buttons, play, rock, etc. www. wpc. ncep. noaa. gov/wwd/internal/watchcollab/watch_collaborator. php

Winter Weather Watch Collaborator • Graphics have both a 30 percent contour and 50 percent probability of exceeding 12/24 hour winter storm watch criteria • 24 hour winter storm watch criteria for snow(in) http: //www. wpc. ncep. noaa. gov/wwd/internal/watchcollab/watch_collaborator. php

New WFO/WPC Forecast Snow Comparison Graphic WPC-NDFD snow difference graphic issued once the draft WPC snow amounts are sent for internal collaboration with WFOs http: //www. wpc. ncep. noaa. gov/wwd/internal/comp_snow/compsnow. php? day=2

2015 -16 WPC Winter Weather Forecast Verification

2015 -16 Low Tracks Verification ECMWF lowest distance error for low positions amongst models 48 -72 hours, GFS+ ECMWF blend for combination of models

Improving Collaboration and Verification: Common Analyses NOHRSC Snowfall Analysis (V 1) -poor quality in data sparse areas and terrain Relies on snowfall reports and not other observed precipitation. Improvements to gridded NOHRSC snowfall analysis will integrate QPE and use of HRRR model snow forecast (starting in Jan? ) 2014 -15 Season January 27 -28, 2015 Snowstorm NOHRSC Operational Development

WPC Deterministic Snow Verification 2014 -15 Day One Snow Threat Score Over the CONUS (Coop data) Automated Ensemble 0. 4 Final Better 0. 3 0. 2 0. 1 0 4" 8" 12" Final forecasts did not improve upon the Automated Ensemble for 4, 8, but did for 12” Automated Ensemble: NAM + GFS + ECMWF + SREF members + GEFS members + ECMWF members

WPC Deterministic Snow Verification 2015 -16 Day One Snow Threat Score Over the CONUS (Coop data) Automated Ensemble 0. 4 Final Better 0. 3 0. 2 0. 1 0 4" 8" 12" Final forecasts improved upon the Automated Ensemble for 4, 8, and 12” Automated Ensemble: NAM + GFS + ECMWF + SREF members + GEFS members + ECMWF members

WPC Deterministic Snow Verification 2015 -16 Day One Snow Frequency Bias Over the CONUS (Coop Data) 4. 5 4 3. 5 3 2. 5 2 1. 5 1 0. 5 0 Automated Ensemble Final High bias above this line 4" 8" 12" Bias performance show major improvement for both automated and human forecasts, and even have a low bias for 12”

Summary • Winter Weather Desk (WWD) operational 24 hours/day , with day shift 12 -21 z and night shift 00 -09 z • New WPC NDFD comparison snow accumulation forecast graphic to aid collaboration process • WPC has adopted the Sanders and Barjenbruch model for freezing rain accumulations (FRAM) • Days 1 -3 Probabilistic Snow/Ice Forecasts now at 70 members to reduce SREF weighting • Experimental Days 4 -7 Forecasts are issued to public and available in NDFD • Verification using Coop Observer reports shows slight high bias for 4, 8 inches but slightly low bias for 12 inches

Questions or Comments? Dan Petersen (Dan. Petersen@noaa. gov) WPC Forecast Operations Branch (301) 683 -0784 (Winter Weather) (301) 683 -1530 (QPF, Lead Forecaster)

Winter Weather Desk Resources • Winter Weather Desk Methodology: http: //www. wpc. ncep. noaa. gov/wwd/internal/ WPCQPFto. Winterweatherdeskmethodology. pdf • Probabilistic Winter Forecasts: http: //www. wpc. ncep. noaa. gov/wwd/internal/pwpf_nws. mp 4, ftp: //ftp. wpc. ncep. noaa. gov/pwpf/conus_2. 5 km/ https: //www. youtube. com/watch? v=Ryf. VMFdo. WDI&feature=youtu. be • Probability Forecasts available on Satellite Broadcast Network: http: //www. nws. noaa. gov/os/notification/tin 15 -45 pwpf. htm • Snow to liquid ratio calculations: http: //www. wpc. ncep. noaa. gov/wwd/internal/S LR_reference. pdf • Virtual Lab presentation on precipitation type (Chenard, Schumacher): https: //vlab. ncep. noaa. gov/web/guest/vlab-forum

WFO/WPC Collaboration by Year 450 400 350 300 2010 -11 250 2011 -12 2012 -13 200 2013 -14 150 2014 -15 100 2015 -16 50 0 WFO questions Comments (WWD) Phone Calls 46 Multi-office collaboration calls, a new record volume

Collaboration Guidelines Snow and freezing rain forecasts: WFO-WFO collaboration guidelines listed at http: //www. nws. noaa. gov/directives/sym/pd 01002001 curr. pdf Use these for WPC-WFO Collaboration 24 hour Snow accumulation: • >2” WPC/WFO difference snow <6” • >4” WPC/WFO difference snow <8” • >8” WPC/WFO difference snow >12” 24 hour Ice Accumulation: • >0. 1” difference for WPC/WFO ice <0. 5” • >0. 2” difference for WPC/WFO ice <1. 0” • >0. 4” difference for WPC/WFO ice >=1. 0

Conference Call Collaboration Guidelines • Have a call if Winter Watch Collaborator shows >30% chance of snow/freezing rain exceeding local watch/warning criteria over 2 or more County Warning Areas • Calls facilitated by Regional Operations Center (ROC) on day shift and WPC at night/weekends/holidays when no ROC support • Calls held 0615/1815 Z, with WPC issuing a NIMNAT message in AWIPS to provide call-in #/time • WPC focuses on model differences/preferences, forecast confidence • WFO focus on items impacting their watch/warning/advisory decisions and forecast items impacting decision support

Central Region (CR) QPF/Winter Weather Test • 0530 z/1730 Z CR offices run a multi-model superblend (minus the ECMWF) to determine draft snow/ice accumulations, and then collaborate (WPC can use ECMWF in blend) • Last year’s CR test had 40% continuity, this year’s will have 0% • Please share your draft snowfall accumulations, precipitation type and snow ratios, and ice ratios with WPC in AWIPS Intersite Coordination Grids! • CR WFOs have about 2. 5 hours for collaboration, which overlaps with WPC collaboration window @0615 z/1815 z

WPC Deterministic Snow Verification 2014 -15 Day One Snow Threat Score Over the CONUS (NOHRSC analysis http: //www. nohrsc. noaa. gov/) 0. 25 Automated Ensemble Final 0. 2 Better 0. 15 0. 1 0. 05 0 4" 8" 12" Final forecasts did not improve upon the Automated Ensemble for 4, 8, and 12” Automated Ensemble: NAM + GFS + ECMWF + SREF members + GEFS members + ECMWF members

WPC Deterministic Snow Verification 2015 -16 Day One Snow Threat Score Over the CONUS (NOHRSC analysis http: //www. nohrsc. noaa. gov/) 0. 25 Automated Ensemble Final 0. 2 Better 0. 15 0. 1 0. 05 0 4" 8" 12" Final forecasts did not improve upon the Automated Ensemble for 4, 8, and 12” Automated Ensemble: NAM + GFS + ECMWF + SREF members + GEFS members + ECMWF members

WPC Deterministic Snow Verification Better 2015 -16 Day One Snow Frequency Bias Over the CONUS (NOHRSC analysis http: //www. nohrsc. noaa. gov/ ) 4. 5 4 3. 5 3 2. 5 2 1. 5 1 0. 5 0 Automated Ensemble Final High bias above this line 4" 8" 12" Both automated consensus and human final forecasts have a high bias, especially for 12”

WPC Deterministic Snowfall Forecast 2015 -16 Day Two Snow Threat Score Over the CONUS (NOHRSC Analysis) 0. 35 Automated Ensemble 0. 3 Final Better 0. 25 0. 2 0. 15 0. 1 0. 05 0 4" 8" 12" Final forecasts outperform the Automated Ensemble for 12 inches, but barely for 4 and 8”

WPC Deterministic Snowfall Forecast 2015 -16 Day Two Snow Threat Score Over the CONUS (Coop Observer Analysis) 0. 35 Automated Ensemble 0. 3 Final Better 0. 25 0. 2 0. 15 0. 1 0. 05 0 4" 8" 12" Final forecasts outperform the Automated Ensemble for 8 and 12 inches, but barely for 4 inches

WPC Deterministic Snowfall Forecast 2015 -16 Day Two Snow Frequency Bias Over the CONUS (NOHRSC) 5 Automated Final Ensemble 4 3 2 Perfect 1 0 4" 8" 12" Final forecasts have a high bias, especially for 12”

WPC Deterministic Snowfall Forecast 2015 -16 Day Two Snow Frequency Bias Over the CONUS (Coop Observer) 5 Automated Final Ensemble 4 3 2 High bias above this line 1 0 4" 8" 12" Automated and Final forecasts have a much lower bias, especially for 12”

WPC Deterministic Snowfall Forecast 2015 -16 Day Three Snow Threat Score Over the CONUS (NOHRSC) 0. 35 0. 3 Automated Ensemble Final Better 0. 25 0. 2 0. 15 0. 1 0. 05 0 4" 8" 12" Automated Ensemble forecasts outperform the final for 8 and 12” forecasts

WPC Deterministic Snowfall Forecast 2015 -16 Day Three Snow Threat Score Over the CONUS (Coop Observer) 0. 35 0. 3 Automated Ensemble Final Better 0. 25 0. 2 0. 15 0. 1 0. 05 0 4" 8" 12" Final forecasts outperform the Automated Ensemble for 4, 8, and 12” forecasts

WPC Deterministic Snowfall Forecast 2015 -16 Day Three Snow Frequency Bias Over the CONUS (NOHRSC) 5 Automated Ensemble Final 4 3 2 High bias above this line 1 0 4" 8" 12" • Final forecasts have a high bias, especially for 12” • Bias may be associated with snow analysis problems (bias higher than verification from WPC snow analysis)

WPC Deterministic Snowfall Forecast 2015 -16 Day Three Snow Frequency Bias Over the CONUS (Coop Observer) 5 Automated Ensemble Final 4 3 2 High bias above this line 1 0 4" 8" 12" • High bias now reversed to low bias, especially for 12” • NOHRSC high bias may be associated with snow analysis problems (no data=no accumulating snow)

WPC Probabilistic Snowfall Verification Day 1 PWPF Verifying frequency over CONUS (Coop Data) 80 Verifying frequency 70 60 Lo w Hi bi gh as 50 bi a s 40 Perfect 4"PWPF 30 20 8"PWPF 10 0 10% 40% 70% Forecast frequency Events forecast more frequently than observed

WPC Probabilistic Snowfall Day 1 PWPF Verifying frequency over CONUS (NOHRSC Data) 80 Verifying frequency 70 60 Lo w Hi bi gh as 50 bi a s 40 Perfect 4"PWPF 30 20 8"PWPF 10 0 10% 40% 70% Forecast frequency Events forecast more frequently than observed

SREF Forecast Plume of Snow Amounts for Washington DC (DCA) from 21 z 20 Jan verifying 12 z 24 Jan 2016 Forecast amounts range from 0 to 40 inches, with a mean of 15 inches DCA event record 28: Over dispersive! Observed 18. 8” http: //jimladueview. blogspot. com/ http: //www. weather. gov/lwx/winter_storm-pr#Top%20 Day%20 Snowfall

SREF Forecast 24 Hour Precip (left) vs. Observed (right) verifying 12 z 23 Jan 2016 SREF Precip extended too far northwest into southern IL/IN/OH http: //www. emc. ncep. noaa. gov/mmb/ylin/pcpverif/daily/20160123/

WPC snow to liquid ratio (SLR) calculation 2016 -17 Snow to Liquid Ratio Method: NAM Rime Factor modified Roebber Snow Ratio + NAM Rime Factor modified Baxter Climatology Snow ratio + GFS Roebber Snow Ratio+ Baxter Climatology Snow ratio + 11: 1