Lightning and Radar Applications Using WDSSII

1
Lightning and Radar Applications Using WDSS-II

• Scott D. Rudlosky
• Ph.D. Candidate, Florida State University
• Henry E. Fuelberg
• Professor of Meteorology, Florida State University

2
Outline

• Warning Decision Support System Integrated
  Information (WDSS-II)
• Introduction and Description
• Data Sources
• Lightning, Model-derived, and WSR-88D
• Early WDSS-II research
• Positive lightning research
• Current Techniques
• Storm Type Identification (NSSL)
• Incorporation of Lightning Data
• Future Work
• Total Lightning and GOES-R GLM
• Severe vs. Non-Severe Storms

3
WDSS-II

• WDSS-II has three components
• Suite of multi-sensor weather algorithms for
• combining and interrogating radar data
• diagnosing hail, lightning and precipitation
• A 3-D display for viewing
  multi-sensor data and algorithm
  outputs.
• Application programming interface
  (API) library in C.

WDSS-II related information obtained from
http//www.wdssii.org/
4
WDSS-II

• Real-time Algorithms
• Research system at NSSL/OU
• Several agencies and private companies license
  WDSS-II to run in real-time.
• Algorithm Communication
• Real-time or archived cases
• Algorithm Creator
• Allows users to modify existing, or to create new
  algorithms.

Schematic obtained from http//www.wdssii.org/
5
WDSS-II

• Very complex software
• Steep learning curve
• Provides an extremely advanced starting point
• NSSL discussion forum
• Archived solutions to problems
• Facilitates user feedback
• Becoming more user friendly
• Training modules now available
• Merging of multiple data sources
• Radar, lightning, model-derived, and satellite
  data
• Allows for differing formats to be combined

6
Data

• Radar Data
• WSR-88D Level II data
• Melbourne, Florida
• Cloud-to-ground lightning data
• National Lightning Detection Network (NLDN)

• Rapid Update Cycle (RUC) Model Data
• 20 km resolution
• Lighting Detection and Ranging (LDAR)
• Total Lightning Data
• Kennedy Space Center

7
LDAR Data

• Lightning flashes are derived with a spark
  consolidation algorithm
• Murphy et al. (2000), Nelson (2002), and McNamara
  (2002)
• Examined in many different forms
• Volume or density of sources
• Initiation points
• Counts and/or densities
• Additional Forms
• Flash extent densities
• GOES-R GLM proxy data

Algorithm Derived Flash
15 min LDAR Source Density
8
Previous Approach

• Linking Individual Cells with Lightning
• Storm Cell Identification and Tracking (SCIT)
• Common linking methods
• Predefined radii
• Visual inspection
• Some combination
• No approach is perfect
• Accuracy is time consuming
• Necessitates case study mode
• Our Early Approach
• Define grid or designate box
• Count flashes

SCIT Cells
Default grid Grid boxes are used individually or
in combination to track cells and lightning.
9
Previous Approach
Individual Cell
Individual Cell
Group of Cells
Group of Cells
10
WDSS-II Techniques

• SCIT Algorithm
• Intermittent tracking
• Temporally dependent
• Limited storm characteristics
• WDSS-II Algorithms
• Merging of Radar and RUC data
• Near-storm environment information
• Additional radar parameters
• e.g., Reflectivity at various isotherm levels

3 May 2007 Left Vertically Integrated Liquid
(VIL) Below Precipitation Rate
Reflectivity 0 C
Reflectivity -10 C
Reflectivity -20 C
11
WDSS-II Techniques

• w2segmotion Algorithm
• Hierarchical K-Means clustering method
  (Lakshmanan et al. 2007)
• Identification of storm cells
• Motion estimates
• Create forecasts
• K-Means Clusters
• Features on the scale of 10 km2
• Any gridded field
• e.g., Reflectivity, VIL, LMA, etc.
• Advection of fields
• Improved temporal resolution
• Provides forecasts

Reflectivity 0.5 Scan
KMeans Cluster of Composite Reflectivity over
Reflectivity 0.5
12
WDSS-II Techniques

• NSSL Storm Type Identification
• Clusters are defined by 30 dBZ mean composite
  reflectivity
• Clusters track additional fields
• i.e., VIL, MESH, POSH, SHI, echo top, and low
  level shear
• Algorithm computes statistics
• i.e., maximum, minimum, average, count, standard
  deviation, and change with time
• Decision tree determines type

13
WDSS-II Techniques

• Storm Type Example (NSSL)
• Determined using the
• Size, speed, aspect ratio, low-level shear, max
  VIL, MESH, mean reflectivity, and orientation
• Storm types include
• Isolated supercell, line, pulse, and
  non-severe

14
WDSS-II Techniques

• FSU Algorithm Modification
• Storm type algorithm
• Storm type is dependent on season and geography
• Algorithm can be trained
• User created algorithms
• Clusters can be based on other fields
• Incorporation of lightning data

10 June 2007 Cross-section of Composite
Reflectivity, LDAR sparks during corresponding
1 min period.
15
WDSS-II Techniques

• Temporal Resolution
• Segmotion attempts to fill the gaps
• Lightning data are more frequent
• We are creating new algorithms
• FSU Algorithm Development
• With clusters based on LDAR
• 1-min resolution
• Modification of storm type algorithm
• Consider lightning parameters
• Short-term forecasts
• Total and cloud-to-ground
• Combination of above
• Severe vs. Non-severe
  determination

10 June 2007
Cross-section of Composite Reflectivity, 1 min of
LDAR sparks between 3 and 4 min after the
volume scan.
16
Continuing Research

• Optimizing the Use of Lightning Data in Severe
  Storm Warning Assessment
• Research Objectives
• Total lightning data
• LDAR/LMAs are regional networks
• Few NWS WFOs have access to data
• Determine relationship between total lightning
  and radar
• GOES-R Global lightning mapper (GLM)
• Goodman et al. (2008)
• Total lightning for all NWS WFOs
• Apply LMA relationships to GLM
• Sterling, VA Lightning mapping array (LMA)
• Examination and/or development of proxy data
• Can data be used similarly?
• Develop algorithms for risk-reduction

17
Continuing Research

• Optimizing the Use of Lightning Data in Severe
  Storm Warning Assessment
• Approach
• Automate dataset preparation
• Minimize manual inspection
• Maximize accuracy
• Examine many cases
• Transition out of case study mode
• Statistical Software
• Determine relationships
• Probabilistic scheme to determine severity
• WDSS-II Algorithm
• Consider parameters and trends
• Utilize motion estimates
• Probabilistic forecast of severity

18
Summary

• WDSS-II
• Capabilities
• FSU Algorithm Modification
• Transition away from case studies
• Total Lightning Approach
• Relate total lightning to radar parameters
• CG (NLDN) and IC (LMA/LDAR)
• GOES-R GLM
• Development/Evaluation of proxy data
• Can data be used in a similar fashion to LMA
• Severe vs. Non-Severe
• Develop algorithms
• Statistical software packages for relationships
• WDSS-II for real-time determination