Ensembles and the Short Range Ensemble Guidance Website at the SPC David Bright NOAA/NWS/Storm Prediction Center Norman, OK Southern Region Teletraining March 6, 2008 Where Americas Climate and Weather Services Begin

Outline Introduction Applications in High Impact Forecasting –Fire Weather –Severe Convective Weather –Winter Weather (via real-time webpage) Summary

Outline Introduction Applications in High Impact Forecasting –Fire Weather –Severe Convective Weather –Winter Weather (via real-time webpage) Summary

Hail, Wind, Tornadoes Excessive rainfall Fire weather Winter weather STORM PREDICTION CENTER HAZARDOUS PHENOMENA

Hazardous Weather Forecasting The Challenge: High impact events often occur on temporal and spatial scales below the resolvable resolutions of most observing and forecasting systems Key premise: We must use knowledge of the environment and non-resolved processes to determine the spectrum of possible hazardous weather, where and when it may occur, and how it may evolve over time

Ensemble Guidance at the SPC Develop specialized guidance for the specific application (severe weather, fire weather, winter weather) Design guidance that… –Helps blend deterministic and ensemble approaches –Facilitates transition toward probabilistic forecasts –Incorporates larger-scale environmental information to yield downscaled probabilistic guidance –Aids in decision support of high impact weather Gauge confidence Alert for potentially significant events National Weather Center, Norman OK

Commonly Used Ensemble Guidance at the SPC Mean, Median, Max, Min, Spread, Exceedance Probabilities, and Combined/Joint Probabilities –Basic Weather Parameters Temperature, Height, MSLP, Wind, Moisture, etc. –Derived Severe Weather Parameters CAPE, Shear, Supercell and Sig. Tornado Parameters, etc. –Calibrated Probability of Thunderstorms and Severe Thunderstorms –Goal of the SPC SREF webpage: Share some useful information that we have and others may not

NCEP/EMC Short Range Ensemble Forecast (SREF) EMC SREF system (21 members) –87 hr forecasts four times daily (03, 09, 15, 21 UTC) –North American domain –Model grid lengths km –Multi-model: Eta, RSM, WRF-NMM, WRF-ARW –Multi-analysis: NAM, GFS initial and boundary conds. –IC perturbations and physics diversity –Recently added bias-correction to some fields (not covered and not on SPC webpage)

Note: All products shown are available in real-time at the SPC website

Outline Introduction Applications in High Impact Forecasting –Fire Weather –Severe Convective Weather –Winter Weather (via real-time webpage) Summary

SREF 500 mb Mean Height, Wind, Temp

SREF 700 mb Mean Height and SD (dash) SD

SREF Mean PCPN, UVV, Thickness

SREF Pr[P12I >.01”] and Mean P12I =.01” (dash)

SREF Pr[RH < 15%] and Mean RH = 15% (dash)

SREF Pr[WSPD > 20 mph] and Mean WSPD = 20 mph (dash)

SREF Combined or Joint Probability Pr [P12I < 0.01”] X Pr [RH < 15%] X Pr [WSPD > 20 mph] X Pr [TMPF > 60F]

Pr [P12I < 0.01”] X Pr [RH < 10%] X Pr [WSPD > 20 mph] X Pr [TMPF > 60F] SREF Combined or Joint Probability

Diagnostics and Analysis Example: Fosberg Fire Weather Index (FFWI) –A nonlinear empirical relationship between meteorological conditions and fire behavior. (Fuels are not considered!) –Derived to highlight the fire weather threat over “small” space and time scales FFWI = F (Wind speed, RH, Temperature)  0 < FFWI < 100  FFWI > ~50-60  significant fire weather conditions  FFWI > ~75  extreme fire weather conditions

SREF Median Fosberg Index + Union (red) Union of all members (Fosberg = 50) Median Fosberg Index Intersection of all members (Fosberg = 50)

SREF Pr[Fosberg Index > 60] and Mean FFWI = 60

SREF Maximum Fosberg Index (any member) Extreme values

SPC Operational Outlook (Uncertainty communicated in accompanying text) Flames are whipped by high winds as Odessa firefighters try to gain control of a grass fire that burned about 7,500 acres Monday afternoon, Feb. 25, Firefighters from Midland and volunteer firefighters from West Odessa, Gardendale and South Ector County assisted in fighting the fire.

Outline Introduction Applications in High Impact Forecasting –Fire Weather –Severe Convective Weather –Winter Weather (via real-time webpage) Summary

SREF 500 mb Mean Height, Wind, Temp

SREF 500 mb Mean Height and SD (dash)

SREF 850 mb Mean Height, Wind, Temp

SREF Precipitable Water (Spaghetti = 1”)

SREF Pr[MUCAPE > 2000 J/kg] & Mean MUCAPE=2000 (dash)

SREF Pr[ESHR > 40 kts] & Mean ESHR=40 kts (dash) “Effective Shear” (ESHR; Thompson et al. 2007, WAF) is the bulk shear in the lower half of the convective cloud

SREF Pr[C03I >.01”] and Mean C03I =.01” (dash) C03I = 3hr Convective Precipitation

SREF Combined or Joint Probability Pr [MUCAPE > 2000 J/kg] X Pr [ESHR > 40 kts] X Pr [C03I > 0.01”] Probability of convection in high CAPE, high shear environment (favorable for supercells)

Diagnostics and Analysis Example: Significant Tornado Parameter (STP) –A parameter designed to help forecasters identify supercell environments capable of producing significant (> F2) tornadoes (Thompson et al. 2003) STP = F (MLCAPE, MLLCL, Helicity, Deep shear)*  STP > ~1 indicative of environments that may support strong or violent tornadoes (given that convection occurs) * An updated version (not shown) includes CIN and effective depth

SREF Median STP, Union (red), Intersection (blue) Union of all members (STP = 1) Median STP Intersection of all members (STP = 1)

SREF Pr[STP > 5] and Mean STP = 5 (dash)

SREF Pr[0-1KM HLCY > 150 m^2/s^2] & Mean 0-1KM HLCY=150 m^2/s^2(dash)

SREF Pr[MLLCL < 1000m] & Mean MLLCL = 1000 m (dash)

Pr [MLCAPE > 1000 J/kg] X Pr [MLLCL < 1000 m] X Pr [0-1KM HLCY > 100 m^2/s^2] X Pr [0-6 KM Shear > 40 kts] X Pr [C03I > 0.01”] SREF Combined or Joint Probability: STP Ingredients Probability of Significant Tornado Environment

SREF 12h Calibrated Probability of Severe Thunderstorms Severe Thunderstorm = > 1 CG Lightning Strike in 40 km grid box and Wind > 50 kts or Hail > 0.75” or Tornado

SREF Probability of STP Ingredients: Time Trends 48 hr SREF Forecast Valid 21 UTC 7 April 2006 Prob (MLCAPE > 1000 Jkg -1 ) X Prob (6 km Shear > 40 kt) X Prob (0-1 km SRH > 100 m 2 s -2 ) X Prob (MLLCL < 1000 m) X Prob (3h conv. Pcpn > 0.01 in) Shaded Area Prob > 5% Max 40%

SREF Probability of STP Ingredients: Time Trends 36 hr SREF Forecast Valid 21 UTC 7 April 2006 Prob (MLCAPE > 1000 Jkg -1 ) X Prob (6 km Shear > 40 kt) X Prob (0-1 km SRH > 100 m 2 s -2 ) X Prob (MLLCL < 1000 m) X Prob (3h conv. Pcpn > 0.01 in) Shaded Area Prob > 5% Max 50%

SREF Probability of STP Ingredients: Time Trends 24 hr SREF Forecast Valid 21 UTC 7 April 2006 Prob (MLCAPE > 1000 Jkg -1 ) X Prob (6 km Shear > 40 kt) X Prob (0-1 km SRH > 100 m 2 s -2 ) X Prob (MLLCL < 1000 m) X Prob (3h conv. Pcpn > 0.01 in) Shaded Area Prob > 5% Max 50%

SREF Probability of STP Ingredients: Time Trends 12 hr SREF Forecast Valid 21 UTC 7 April 2006 Prob (MLCAPE > 1000 Jkg -1 ) X Prob (6 km Shear > 40 kt) X Prob (0-1 km SRH > 100 m 2 s -2 ) X Prob (MLLCL < 1000 m) X Prob (3h conv. Pcpn > 0.01 in) Shaded Area Prob > 5% Max 50%

Severe Event of April 7, 2006 First ever Day 2 outlook High Risk issued by SPC More than 800 total severe reports –3 killer tornadoes and 10 deaths SREF severe weather fields aided forecaster confidence

Outline Introduction Applications in High Impact Forecasting –Fire Weather –Severe Convective Weather –Winter Weather (via real-time webpage) [ ] Summary

SREF 500 mb Mean Height, Wind, Temp

Probability Matched Averaging [PM] 1)Calculate the ensemble mean on the grid (use PDF for location) 2)Create a ranked distribution of ensemble mean grid point data (use PDF for values) 3)Create a ranked distribution of the raw data from ALL MEMBERS (n members) 4)On the grid, replace the ranked value of the ensemble mean with every n th value of the ranked raw data 5)Location of the ensemble mean is retained, but value replaced with actual data Rank (Min to Max) Data Values (i,j) Ebert, E.E., 2001: Ability of a Poor Man's Ensemble to Predict the Probability and Distribution of Precipitation. Mon. Wea. Rev., 129, 2461–2480. Ensemble mean Raw data (all members)

Probability Matched Averaging [PM] 3h Precip - Arithmetic Mean 3h Precip - Probability Matched Mean

Estimating the snowfall from the model QPF Forecast Time 3 hr QPF: PCP Rate: Snow ratio is adapted from Boone and Etchevers (2001; AMS J. Hydrometeorlogy): Snow ratio = 1000 / ( T), where T is the maximum saturated temperature surface to 300 mb AGL [Allowable range: Min 7:1 and Max 40:1; 10:1 at 0 degC] Snowfall = Snow ratio at output time x precipitation over previous 3 hours Snow rate (inches per hour) = Snowfall / 3 hrs -or- Instantaneous precipitation rate is converted to an instantaneous snow rate per hour at the start and end times. If hourly snow rate > 3hr total snowfall, hourly rate = 3hr value. Available data:

Example: Dendritic Growth Zone (DGZ) – Very efficient growth (assuming water vapor is replenished) – Peak growth rate -14 to -15C in low-to-mid troposphere – Accumulate rapidly Search ensemble members for: UVV > 3 cm/second -11 C < Temp < -17 C Layer depth > 50 mb RH in layer > 85% Diagnostics and Analysis

AMS (Emanuel 1985) Enhancement in the upward branch of the frontal circulation in the presence of low MPV. Typical values of MPV Low values of MPV MPV Moist Potential Vorticity

Outline Introduction Applications in High Impact Forecasting –Fire Weather –Severe Convective Weather –Winter Weather (via real-time webpage) Summary –Interpreting Ensemble Means, Verification, and Storm-scale

Be careful interpreting ensemble means Consider this location…mean wind ~10 kts (15 mph) F075 SREF - Valid 00 UTC 11 Nov 2007 Mean PMSL; Mean Wind; Mean Thickness

F075 SREF - Valid 00 UTC 11 Nov 2007 Probability wind speed > 30 mph Probability the wind speed > 30 mph ~70% Probability doesn’t appear to fit mean

F075 SREF - Valid 00 UTC 11 Nov 2007 Maximum speed (mph) at each grid point L L L L L Maybe, but doesn’t explain the apparent discrepancy. Is the distribution positively skewed?

F075 SREF - Valid 00 UTC 11 Nov 2007 Spaghetti of low center positions Low center spaghetti Considerable uncertainty in the location of the low.

F075 SREF - Valid 00 UTC 11 Nov 2007 Probability wind speed > 30 mph (given u < 0) Conditiional probability: Given U < 0 Probability the wind speed > 30 mph ~50% May need to consider “feature relative” averaging

Perfect Forecast No Skill Perfect Forecast No Skill Calibrated Reliability (5 Aug to 5 Nov 2004) Calibrated Thunder Probability Climatology Frequency [0%, 5%, …, 100%]

Hail >.75”Wind > 50 ktsTornado All 24h forecasts F24 – F63 from 15 April to 15 October 2005 SevereVerification 09 UTC and 21 UTC SREF ROC Area=.84 Ave Hit = 15% Ave Miss= 2%

Verification Reliability Diagram: All 3 h forecasts (F00 – F63); 35 days (Oct 1 – Apr 30) Economic Potential ValueReliability

SREF 3h Calibrated Probability of Thunderstorms Thunderstorm = > 1 CG Lightning Strike in 40 km grid box What about dry thunderstorms? Photo from John Saltenberger

SREF 3h Calibrated Probability of “Dry” Thunderstorms Dry thunderstorms = CG Lightning with < 0.10” precipitation

F027: Valid 00 UTC 15 May hr UH > mi Supercell Detection Probability Updraft Helicity > 50 m 2 /s 2

F027: Valid 00 UTC 15 May hr Linear mode + 25 miles Squall Line Detection Probability Linear Mode Exceeding 200 miles

Radar 00 UTC 15 May 2007 Radar Reflectivity Valid 02 UTC 15 May 2007

Radar 00 UTC 15 May 2007 > 35 dbZ Radar Reflectivity Valid 02 UTC 15 May 2007

SPC SREF Products on WEB Questions/Comments…