Upload
randolf-horton
View
218
Download
0
Tags:
Embed Size (px)
Citation preview
Updated Radar-Based Techniques for Tornado Warning Guidance in the Northeastern United
States
Brian J. Frugis & Thomas A. WasulaNOAA/NWS Albany, New York
NROW XVI – 4-5 November 2015
Motivation
• CSTAR study has been completed to update the V-R Shear Technique for tornado warning guidance using 8 bit, high resolution radar data– This was originally done for the area around the
Albany CWA in 2011-2013, but has been expanded this year for tornadoes over the entire Northeastern US
– Will the updated version of this study help detect tornadoes that were previously difficult to predict?
Motivation(continued)
• Impact-based warnings require knowledge regarding the estimated strength of the tornado– Ongoing studies have been looking at estimating
tornadic strength from the maximum height of tornadic debris from events from the Southeast & Plains, but Northeastern United States tornados have not been examined in this matter.
– Although limited due to the small amount of tornadic events, are the results for these Northeastern US Events similar to the findings for other parts of the US?
Original V-R Shear Study(LaPenta et al. 2000)
Maximum observed gate-to-gate shear below 3 km was found to be useful in identifying tornadic storms (LaPenta et al. 2000)
S=Vr/(D*1800)
Shear (S) is measured in units of s-1, rotational velocity (Vr) in knots and D is the diameter of which S is calculated in n mi.
Vm is the maximum strength of the mesocyclone, which can be found using a vertical cross-section over a 3.5 n mi width
Radar data was 4 bit and had to be range normalized due to radar resolution and beam spreading
Expanded V-R Shear Study8-bit radar data from 76 tornadoes from 2003-15 were examined using Gibson Ridge Software’s GR2Analyst. This expanded the dataset that was previously just 41 tornadic events near the ALY CWA that was originally completed in 2013.
21 null cases (storms with strong rotation that prompted tornado warnings but didn’t produce a tornado) were also looked at.
Shear was calculated for each tornado using a diameter of 0.5 n mi for all tornadoes from 2008-15 within 60 n mi of radar (since super-res data was available). The old methodology for determining D was used for tornadoes from 2003-08.
The maximum velocity differential of the mesocyclone (Vm) was calculated for each storm
Results of S and Vm were compared to the nomogram graphic created by the LaPenta study
Using the V-R Shear Method
Rotational Velocity of tornadic couplet from 0.5° SRM measured over 0.5 nm width (about 10 nm from KENX radar located just south-southeast of tornadic couplet)Vr=42.5 kts, S=0.0472 s-1
Cross-Section of SRM over 3.5 nm width of Mesocyclone. Max Strength: 89 kts (Vm=Max inbound + Max Outbound of Meso)
Updated V-R Shear Study Nomogramfor Entire Northeastern United States
(n=76)
Group I: No Clear Signal Between Tornadic & Non-Tornadic Mesocyclones; no F2/EF2 or strongerGroup II: 71% Tornadic & Mainly Weak Tornadoes Group III: 100% tornadic; mainly F1+/EF1+
Tornadic Debris Signature Detection (TDS) Methodology(Adapted from Entremont and Lamb 2015)
1) Find a tornadic couplet in SRM with gate-to-gate rotation.
2) Have Correlation Coefficient (CC) less than 0.90 co-located with the couplet.
3) Have Reflectivity (Z) values 35 dBZ or greater co-located with the couplet.
4) Have Differential Reflectivity (ZDR) values around zero co-located with the couplet.
5) If all these criteria are met, check next radar elevation slice above and continue the process until they are not met anymore. The top slice where all criteria is met is the Max TDS Height.
1951 UTC 22 May 2014 Duanesburg-Delanson, NY EF3 KENX Radar 0.5° Elevation AngleTop-left: SRM (kts), Top-right: CCBottom-left: Z (dBZ), Bottom-right: ZDR (dB)
Tornadic Debris Signature StatisticsNortheastern US Events 2012-2015
69 tornadic events from 2012-2015 from the Northeastern United States were examined for the presence of a Tornadic Debris Signature (TDS).
Out of those 69 tornadoes, a TDS was present in 24 of them (35%).
Of these 24, 12 tornadic events occurred within supercell thunderstorms and 12 were within Quasi-Linear Convective System (QLCS) thunderstorms.
All EF2+ strength tornadoes occurring in the Northeastern US between 2012-2015 displayed a TDS.
EF-0 EF-1 EF-2 EF-3 EF-4 EF-5
5 14 4 1 0 0
Maximum Height of TDSAverage & Median Values
EF0 (5) EF1 (14) EF2+ (5)0
1000
2000
3000
4000
5000
6000
7000
8000
9000
Maximum Height of Tornadic Debris (in feet)
Average Median
For the 24 events examined between 2012-2015, the average & median values increase with increasing tornadic strength, especially once reaching the EF2 level.
This is similar to results found by Entremont and Lamb (2015).
More storms will be examined in future years as the dataset expands to see if these trends continue.
Maximum Height of TDS vs. EF Scale Strength vs. Range
Weak tornadic events (EF0) were generally only see close to the radar, as debris generally didn’t get very high and beam height rising over distance would prevent these further away storms from being seen.
Hard to seen much difference between EF0/EF1 storms with this plot.
0 10 20 30 40 50 60 70 800
2000400060008000
1000012000140001600018000
Max TDS Height vs. EF Scale Strength vs. Range (n=24)
EF0 EF1 EF2+
Range (nm)
Max
TD
S H
eigh
t (fe
et)
Maximum Height of TDSNortheastern US Tornados 2012-2015
Box and Whisker Plot
EF0 (5) EF1 (14) EF2+ (5)0
1000
2000
3000
4000
5000
6000TDS
Height (ft)
Max TDS Height vs. Tornado Strength (n=24)
EF0/EF1 are similar, but jump up seen with EF2+ events. Entremont and Lamb (2015) also saw a similar signal in their data as well (which contained 181 events).
Normalized Rotation(NROT)
• NROT is a derived product found in GR2Analyst that attempts to assess rotation for each range bin, regardless of distance from radar
• Values range from -5 to 5 and are unitless– Software documentation
notes that over 1.00 is significant and over 2.50 is extreme
NROT display from GR2Analyst from 2247Z 29 May 2013. NROT values reached a maximum of +1.36 just prior to tornado touchdown.
Normalized Rotation (NROT)Data & Analysis
• 137 tornadic events from 2003-2015 from the Northeastern United States were examined using Gibson Ridge’s software GR2Analyst.
• The value of Normalized Rotation (NROT) was recorded at the onset for each tornadic event.
• In addition, 42 null events were examined as well. Null events showed enough rotation to warrant a tornado warning, but a tornado was not reported.
Normalized Rotation (NROT)As Compared to Storm Strength
EF0 (49) EF1 (75) EF2+ (13)0.00
0.10
0.20
0.30
0.40
0.50
0.60
NROT Values at Tornado Formation (n=137)Northeastern US Tornadic Events 2003-2015
NROT
Normalized Rotation (NROT)Null vs. Tornadic Events
Null Events (42) Tornadic Events (137)
0.00
0.10
0.20
0.30
0.40
0.50
0.60
NROT Values at Time of Actual Tornado or Tornado Warning for Null Events (n=179)
Northeastern United States 2003-2015
NROT
Conclusions:Expanded V-R Shear Study
• The updated V-R Shear technique shows similar results as the previous study done by LaPenta et al. in 2000, but now accounts for the improvements in radar resolution. – Expanded dataset over entire Northeastern US helps increase confidence in
study results
• Weak Tornadoes (EF0/EF1) continue to be difficult to detect via radar– Expanded V-R Shear Study still shows little skill differentiating between null
mesocyclones & weak tornadoes
• Strong tornadoes for the Northeast (EF2+) show a strong correlation between increasing amounts of low-level shear and the increased strength within the mid-level mesocyclone
Conclusions:Northeastern US TDS Study
• As also shown by Entremont and Lamb 2015, the higher Max TDS Heights are correlated with an increase in tornado strength.
• Most pronounced jump in Max TDS height occurs with transition between weak (EF0/EF1) to strong tornados (EF2+)
• Small amounts of tornados since launch of Dual-Polarization are a limiting factor, but results are similar to larger study by Entremont and Lamb
Conclusions:Normalized Rotation (NROT)
• NROT is not always a clear discriminator on its own for tornadic vs. non-tornadic storms.
• There also isn’t a strong signal seen in increasing tornadic strength with increasing NROT values either.
• NROT could be useful if combined with other mesoscale information– WFO BGM has done work regarding this & has shown this at
previous NROW meetings