Impact of the Loss of QuikSCAT on NOAA NWS Marine Warning and Forecast Operations Joseph Sienkiewicz 1 Michael J. Brennan 2, Richard Knabb 3, Paul S. Chang 4, Hugh Cobb 2, Zorana J. Jelenak 5, Khalil A. Ahmad 4, Seubson Soisuvarn 4, Affiliations 1 NOAA Ocean Prediction Center David Kosier 1, 2 Tropical lprediction i Center George Bancroft 1 3 The Weather Channel 4 NESDIS StAR OVWST, Barcelona Spain 5 UCAR 1
High resolution Ocean Surface Vector Winds (OSVW) fully integrated into operations Twice daily near complete wind field Increased situational awareness Improved ddetection and warning capability Enhanced understanding of winds over oceans Allowed focus on extreme cyclones (Hurricane Force) QuikSCAT May 19, 2010 OVWST, Barcelona Spain 2
Without QuikSCAT Forecasters now soley rely on: Satellite Interpretation Skills NWP Analyses and short-term forecasts ASCAT Ship and buoy observations Situational awareness is reduced Consistency and confidence reduced May 19, 2010 OVWST, Barcelona Spain 3
Impact of QuikSCAT Loss Tropical Analysis and Forecast Branch 10 year climatology of Gulf of Tehuantepec gap wind events based on QuikSCAT showed an average of 6 to 8 storm force events per cold season TAFB issued a total of 15 Gale and 5 Storm warnings for the area during the 2009 2010 cold season based on improving NWP model output, but could not get confirmation of the Storm warnings from ASCAT Brennan et al. (2010) Wea. Forecasting Typically 1 ASCAT pass per day hits this region, compared to 2 QuikSCAT passes on most days From Jan Mar 2010, TAFB issued 20 Gale warnings in their Atlantic waters, 9 in the Gulf of Mexico, and 4 in the Caribbean May 19, 2010 OVWST, Barcelona Spain 4
Case Example 8 14 January 2010 Gulf of Mexico Gale/Tehuantepec Storm Surface analysis, ship/buoy obs and ASCAT every 6 h from 8 14 January May 19, 2010 OVWST, Barcelona Spain 5
Case Example 8 14 January 2010 Gulf Gale/Tehuantepec Storm Only 5 ASCAT hits on the Gulf of Tehuantepec in a 7 day period Strongest ASCAT winds 40 45 kt on 10 Jan No ASCAT data in region from 03 UTC 9 Jan through 15 UTC 10 Jan and from 15 UTC 10 Jan to 04 UTC 12 Jan 3 storm force ship observations Only a couple ASCAT gale force retrievals in far SW Gulf of Mexico Summary of lost capability bl due to loss of QuikSCAT Ability to capture storm force winds Full outline of area impacted by gale force winds in strongest events Ability to compare the model wind field with observations over a large spatial area Decreased forecaster confidence for severity, coverage, and timing of most extreme events May 19, 2010 OVWST, Barcelona Spain 6
Specific events Dec 9, 2009 Hawaii big wave event Tropical Storm Nida other than satellite data and without QuikSCAT data the ability to measure winds for these islands has been severely hampered making it difficult to verify any tropical system. Jan, 2020 C/V Horizon Hunter damaged en route to Guam diverted to Hawaii May 19, 2010 OVWST, Barcelona Spain 7
Hurricane Force Observations (Oct 07 ~ May 08 ) 250 200 150 212 188 100 50 0 2 39 38 19 QS -12.5 km QS - 25 km AS - opr AS - new GFS OBS
QuikSCAT vs. NCEP Global Forecast System OVWST, Barcelona Spain May 19, 2010 9
April 23, 2010 Forecaster tools the Quandary Satellite imagery interpretation etat Result 986 STORM Surface observations Reality do not know intensity ASCAT NWP Analyses/forecasts May 19, 2010 OVWST, Barcelona Spain 10
% Wind Error QSCAT 25 km NRT vs. GDAS Aug 08 Jun 09 GALE STORM HURRICANE FORCE May 19, 2010 OVWST, Barcelona Spain 11
Hurricane Force Extra tropical Cyclones Detection and Warning Trend using QuikSCAT 2000 2009 70 60 WARNING CATEGORIES Pre- QSCAT50 1. GALE 34-4747 kt 2. STORM >48 C y 40 c kt l o n e 30 s QSCAT ERA 1. GALE 34-4747 kt 2. STORM 48-63 kt 3. HURCN FORCE > 64 kt 20 10 Hurricane Force Warning Initiated Dec 2000 Detection increased with: -Forecaster familiarity -Data availability -Improved p resolution o -Improved algorithm Hurricane Force Wind Warning Initiated Dec 00 25 km QuikSCAT Available in N AWIPS Oct 01 23 24 23 12.5 km QuikSCAT available May 04 15 22 37 Improved wind algorithm and rain flag Oct 06 33 31 34 64 51 39 49 45 34 Atlantic Pacific Totals A 289 QuikSCAT Launch 14 Jun 99 11 9 P 269 558 0 1998-99 1999-00 2000-01 2001-02 2002-03 2003-04 2004-05 2005-06 2006-07 2007-08 2008-09 May 19, 2010 OVWST, Barcelona Spain 12
Forecaster estimated Hurricane Force cyclones Sep 09 Apr 2010 40 35 31 30 25 20 15 12 10 8 With 5 QSCAT 0 Atlantic 2009 10 Atl below previous 5 40 yr average (43) DJFMA (32.6) 22 Pacific Pac at avg. (40.2) DJFMA (29.8) 28 Atlantic *highly subjective post QuikSCAT Pacific May 19, 2010 OVWST, Barcelona Spain 13
10 Forecaster estimated Hurricane Force cyclones Monthly Distribution 2009 2010 vs. 2004 2009 average 9 8 QuikSCAT Failure Nov09 7 6 5 Atlantic Average 4 Atlantic 2009 10 3 10 Pacific 6% reduction vs. 5 yr avg. 2 1 9 8 QuikSCAT Failure Nov 09 0 Sept Oct Nov Dec Jan Feb Mar Apr 7 6 5 Pacific Average Atlantic 33% reduction vs. 5 yr avg. 4 Pacific 2009 10 3 2 1 0 May 19, 2010 OVWST, Barcelona SeptSpain Oct Nov Dec Jan Feb Mar Apr 14
0.8 0.7 0.6 0.5 0.3 0.2 0.1 Forecaster verification 48,96 Pacific, Atlantic Probability of Detection (POD) 0.4 POD 48_PAC_HF POD 48_PAC_Storm POD 48_PAC_Gale Pac 48 Atl 48 0 0 03-09 ONDJFM * 09-10 (DJFM) Only 0.8 0.7 0.6 0.5 0.4 POD 48_ATL_HF 0.3 0.2 0.1 POD 48_ATL_Storm POD 48_ATL_Gale 08 0.8 0.8 0.7 0.7 0.6 0.6 0.5 0.5 04 0.4 POD 96_PAC_HF 04 0.4 POD 96_ATL_HF 0.3 POD 96_PAC_Storm 0.3 POD 96_ATL_Storm 0.2 POD 96_PAC_Gale 0.2 POD 96_ATL_Gale 0.1 0.1 Pac 96 0 0 Atl 96 May 19, 2010 OVWST, Barcelona Spain 15
0.8 0.7 0.6 0.5 Forecaster verification 48,96 Pacific, Atlantic False Alarm Rate(FAR) 0.8 0.7 0.6 0.5 0.4 48_PAC_HF 0.3 0.2 0.1 48_PAC_Storm 48_PAC_Gale Pac 48 Atl 48 0 0 0.4 48_ATL_HF 0.3 0.2 0.1 03-09 ONDJFM * 09-10 (DJFMA) Only 48_ATL_Storm 48_ATL_Gale 0.8 0.7 0.6 0.5 0.8 0.7 0.6 0.5 0.4 96_PAC_HF 0.3 0.2 0.1 0 96_PAC_Storm 0.4 96_ATL_HF 0.3 96_ATL_Storm 96_PAC_Gale 0.2 96_ATL_Gale Pac 96 0.1 Atl 96 0 May 19, 2010 OVWST, Barcelona Spain 16
Summary Have lost OVW consistency for dfii defining warnings relying heavily on imagery, NWP and ASCAT less observed HF cyclones (Atlantic) nearly all not verified by observations cyclone verification suggests forecasters growing greliance on NWP to estimate warning categories (HF cyclones) Need to compare NWP analyses to forecaster estimated warnings NWP (GDAS) low bias for warning categories (compared to QSCAT) less storm force in Gulf of Tehuantepec less able to define warning areas need to examine other orographic jets (Greenland Tip Jet) W Pacific cyclones reduced detection capability Atlantic hurricanes impact still to be determined May 19, 2010 OVWST, Barcelona Spain 17
Impact of QuikSCAT Loss Tropical Cyclones NHC has not had any TCs develop in its AOR since the failure of QuikSCAT Likely impacts going forward into 2010 season: Loss of useful data for the analyzing the location, intensity, and structure of TCs, particularly those that are not sampled by aircraft reconnaissance Even ASCAT hits often fail to sample the entire TC circulation Reduced capability to identify winds above tropical storm force due to resolution and retrieval algorithm issues Degraded analysis of 34 kt wind radii in major hurricanes not sampled by aircraft reconnaissance Fewer passes over formative systems QuikSCAT was occasionally helpful in making the decision to initiate advisories May 19, 2010 OVWST, Barcelona Spain 18
Great Circle Ray Tracing North coast New Guinea Dec. 2008 GFS Dec 04 0000 f045 QS Dec 05 2100 KTS GFS Dec 04 1200 f033 GFS Dec 05 0000 f021 May 19, 2010 OVWST, Barcelona Spain 19