LIGHTNING METEOROLOGY I: ELECTRIFICATION AND LIGHTNING ACTIVITY BY STORM SCALE
TALKING POINTS
Slide 1 – Title
Welcome to this VISIT teletraining session on Lightning Meteorology. This session was developed by Bard Zajac and John Weaver at the NOAA/NESDIS branch located at CSU/CIRA (Colorado State University / Cooperative Institute for Research in the Atmosphere) with contributions from those individuals listed.
During the next 90 minutes, we will discuss electrification and lightning activity in isolated storms and mesoscale convective systems (MCSs) using a mix of theory and AWIPS case studies. At the end of the session, you should have a better understanding of thunderstorms and know how to use lightning data in most convective weather scenarios.
Slide 2 – Teletraining Tips
No comment.
Slide 3 – Introduction
The theoretical concepts presented in this session apply to roughly 80–90% of the isolated storms and MCSs occurring during the warm season. Lightning activity in other storms is examined in Lightning Meteorology II. Storms examined in Part II include severe-positive strike dominated storms and winter storms.
Slide 4 – Objectives
In this teletraining session, the ice-ice collisional charging mechanism is considered the primary charging mechanism in thunderstorms. The requirement of ice for electrification means that thunderstorms have common characteristics that can identified from satellite and radar. We will examine the charge distributions in thunderstorms and see how they control the occurrence of lightning. With an understanding of electrification, charge distributions and lightning, positive and negative cloud-to-ground (CG) lightning can be used to infer precipitation location and intensity and storm lifecycle. Throughout this session, lightning data is integrated with other data sets.
Slide 5 – Overview
No comment.
Slide 6 – Sec 1: introduction to review section
The lifecycle of a typical isolated thunderstorm is separated into four stages: shallow cumulus, towering cumulus, mature cumulonimbus and dissipating cumulonimbus.
Slide 7 – Sec 1: shallow cumulus (16 frames)
Frame 1: Figure shows a shallow cumulus in the process of growing into a towering cumulus. The cloud is mostly updraft with weak vertical motions. Downward motions occur only along the edges of the cloud and are associated with turbulence and entrainment. Most of the cloud lies below the freezing level but a part of the cloud penetrates above the freezing level and even above the –10C level. The –10C level is noted since it is an important temperature threshold for ice nucleation, to be discussed later in this slide.
Frames 2-3: Question (given to an office) highlights moist convective processes: lifting of moist air from the boundary layer, adiabatic expansion and cooling, water supersaturation and condensation.
Frames 4-5: Question (taken by the instructor) highlights how CCN provide an active surface for condensation and droplet formation.
Frames 6-7: Question (given to an office) highlights collision-coalescence as the main droplet growth process in warm clouds. This question leads to a more thorough discussion of collision-coalescence in the next frame.
Frame 8: Figure shows the growth of cloud droplets into rain drops. Initially droplets grow by condensation on CCN, but this process is not efficient for droplets larger than 10 m in radius due to curvature effects. Once droplets reach this size, the main growth process is collision-coalescence. Over the course of many collisions, droplets grow into drizzle drops and eventually rain drops.
Frames 9-10: Question (given to an office) highlights the fact that water does not necessarily freeze when cooled below freezing. The term, supercooled droplets, is important to remember.
Frames 11-12: Question (taken by the instructor) highlights how IN provide an active surface for ice nucleation. The –10C temperature threshold represents an average temperature at which various IN become active (IN include soil mineral, volcanic ash, etc.).
Frame 13: Figure depicts ice nucleation involving IN. Ice nucleation can occur without IN. Riming and spontaneous nucleation are two important examples discussed later in this section.
Frames 14-15: Question (given to an office) highlights the growth of ice crystals through deposition (i.e., the movement of water molecules directly from vapor to solid under ice supersaturation).
Frame 16: Figure summarizes the dynamics and microphysics of a shallow cumulus that is just starting to develop ice. The figure depicts varying amounts of cloud liquid water below the freezing level (see legend). Variations in cloud liquid water content (CLWC) are caused by variations in updraft strength and associated water supersaturation. The figure also depicts the presence of supercooled liquid water above the freezing level and the presence of small ice particles above the –10C level.
Slide 8 – Sec. 1: towering cumulus (7 frames)
Frame 1: Figure shows a towering cumulus cloud with a large volume of updraft and a downdraft forming at mid-levels. The tower is tall, growing rapidly and is starting to produce precipitation-sized ice particles at mid-levels. These particles are starting to descend, but have yet to fall out of the cloud. The cloud top is approaching the –40C level and the tropopause is assumed to be around –60C.
Frames 2-3: Question (given to an office) highlights the fact that small ice particles of various origins are lofted to upper-levels.
Frames 4-6: Question (taken by the instructor) highlights the formation of graupel through riming (i.e., the immediate freezing of supercooled droplets on contact with larger ice particles).
Frame 5: The three photos depict riming, starting on the left with a stellar plate that has grown large enough to descend with respect to supercooled droplets and accrete a dense coat of rime. The middle photo shows the same stellar plate after accreting even more rime. The photo on the right shows conical graupel. Graupel is the stage of riming when the embryo (a stellar plate in this case) can no longer be discerned. Note that riming is a positive feedback mechanism: as an ice particle accretes rime, its radius and fall speed increases and it collides with a larger number of supercooled droplets per unit time. Graupel becomes hail when supercooled droplets no longer freeze immediately upon collision (due to latent heat of freezing, the graupel surface warms). Hail normally accretes supercooled droplets in both wet and dry growth modes.
Frame 6: The concept of a particle balance level is worth mentioning here. The balance level is the height in the storm where a particle's fall speed equals the updraft speed. For small supercooled droplets and small ice crystals, the balance level is found at upper-levels where the updraft is weak. For large ice particles accreting rime, the balance level found at mid-levels where the updraft is stronger. In this region of the storm, ice particles are near-stationary as they collect upward-moving supercooled droplets. Riming ice particles eventually gain enough mass to fall through the updraft.
This discussion indicates that large ice particles collide frequently with both small supercooled droplets and small ice crystals. Ice-supercooled droplet collisions cause riming. Ice-ice collisions cause electrification. How? Laboratory cloud chamber studies show that electrical charge is transferred between ice particles as they collide in the presence of supercooled droplets. This ice-ice collisional charging mechanism is examined in the next section.
Frame 7: The main distinction between the towering cumulus stage and shallow cumulus stage is the formation of graupel at mid-levels due to riming. As graupel grows and eventually descends, a downdraft is initiated by precipitation drag.
Frame 8: The appearance of radar echo > 30 dBZ at mid-levels indicates that graupel has formed and that this graupel may reach the surface as convective precipitation within a short time period (as soon as 10 minutes in come cases).
Slide 9 – Sec. 1: mature cumulonimbus (3 frames)
Frame 1: The sounding is assumed to have sufficient instability and shear to separate the updraft and downdraft. The storm has the potential to maintain itself for at least a short period of time.
Frames 2-3: Question (given to an office) highlights the fact water cannot exist in the liquid phase at temperatures colder than –40C. Even in the absence of IN, supercooled droplets freeze around –40C due to spontaneous or homogeneous nucleation.
Frame 4: The main distinctions between the mature cumulonimbus stage and the towering cumulus stage is the extension of precipitation/downdraft to the surface and the formation of an anvil. The storm continues to produce heavy precipitation so long as the updraft provides supercooled droplets for riming.
Frame 5: Radar echo > 30 dBZ extends from mid-levels to the surface. This echo is associated with graupel/hail above the freezing level and with raindrops and graupel/hail in various stages of melting below the freezing level. Some frozen precipitation may reach surface. Range of reflectivities listed is approximate.
Slide 10 – Sec. 1: dissipating cumulonimbus (2 frames)
Frame 1: Figure shows the storm after most heavy precipitation has fallen out.
Frame 2: Above the freezing level, the cloud is mostly glaciated with little to no supercooled droplets present due to weakening updraft. Light to moderate precipitation may still fall out of the storm.
Slide 11 – Sec. 1: review questions (2 frames)
Frames 1-2: Questions (given to an office) highlight the radar echo associated with graupel as well as the formation of graupel by riming.
Slide 12 – Sec. 2: introduction to electrification
Laboratory cloud chambers can create the microphysical environment of a thunderstorm at mid-levels. This environment comprises supercooled droplets, ice crystals, and ice undergoing riming. Riming ice is reproduced using a rotating metal rod that accretes rime as it collides with supercooled droplets. The charging experiment uses a rotating metal rod attached to sensitive electrical equipment.
Slide 13 – Sec. 2: ice-ice collisional charging mechanism
It is beyond the scope of this session to examine the process (or processes) responsible for charge transfer during ice-ice collisions, especially since charge transfer is not well understood. For this reason, we ask forecasters to take the this information at face value.
Slide 14 – Sec. 2: dipole charge structure and lifecycle (8 frames)
In this slide, graupel-ice crystal charging mechanism is applied to the four-stage thunderstorm lifecycle.
Frames 1-2: No comment.
Frames 3-4: Figures focus on charge generation at mid-levels.
Frames 5-6: Figures focus on charge generation at mid-levels as well as the advection of charge. Positively charged ice crystals are lofted to upper-levels and negatively charged graupel is either suspended at mid-levels by the updraft or descends towards the surface. Normal dipole is an important term to remember.
Frame 7-8: Figures focus on the lifecycle stage when riming and charge generation have ended but charge advection continues. Positively charged ice crystals are carried downshear in the anvil, forming the tilted dipole charge structure. Negatively charged graupel falls out of the storm.
Slide 15 – Sec. 2: review questions (2 frames)
Frame 1: Questions (given to an office) highlight the key points of electrification including the requirement of graupel for charging.
Slide 16 – Sec. 2: introduction to Fort Collins (FCL) case
No comment.
Slide 17 – Sec. 2: FCL sounding (2 frames)
Frame 1: The Denver evening sounding is warm and moist throughout the troposphere. The Denver sounding is similar to an average sounding from the tropical western Pacific (TOGA COARE in green).
Frame 2: The potential for deep convection is low based on the parcel trajectory plotted. Cloud tops will be much lower than the tropopause. The potential for CG lightning is low based on the small positive area (CAPE) above –10C. The weak CAPE above –10C suggests that graupel may not form. Without graupel, charging and lightning do not occur.
Note that vertical wind shear may be sufficient to support long-lived storms.
Slide 18 – Sec. 2: FCL IR4 & CGs (20 frame loop)
Loop shows 15-minute GOES-8 IR4 imagery and 15-minute CG lightning data from 23:15 to 05:00 on 28–29 July 1997.
Cloud tops as cold as –70C and copious lightning around the Denver area indicate that the Denver sounding is not representative of these storms. Post-event analysis indicates that mid-level drying and cooling occurred. However, the Denver sounding appears to be representative of the Fort Collins area with cloud tops between –30C and –40C and infrequent lightning.
Slide 19 – Sec. 2: FCL radar time-height analysis
Plot shows a time-height cross-section of maximum reflectivity over the location in southwest Fort Collins where it rained the most. Radar data were collected by the WSR-88D in Cheyenne, WY over the full storm period from 17:25–22:25 MDT on 28 July 1997. Maximum reflectivities were calculated over a cylindrical volume with dimensions of 14 km in height and 10 km in diameter. The O's on the x-axis indicate cloud-to-ground (CG) lightning strikes associated with the storms over Fort Collins. The 0C and –10C isotherms are plotted and are estimated from the Denver sounding launched at 00:00 UTC on 29 July (Slide 17). Two black lines are also plotted. One traces the 40 dBZ echo and the other traces rain mass flux. The red arrows indicate four periods of heavy precipitation.
The following points are discussed: 1) the CSU-CHILL dual-polarization radar, located 40 miles east-southeast of Fort Collins, was scanning during the time period plotted. Dual-polarization data was used to diagnose the fraction of ice versus water above the freezing level; 2) during the first two periods of heavy precipitation, the storms were shallow and produced no CG lightning. The CHILL radar indicated that the fraction of ice was less than 25%; 3) during the second two periods of heavy precipitation, the storms were slightly deeper and produced CG lightning. The CHILL radar indicated that the fraction of ice was greater than 75%; 4) a radar reflectivity threshold of 45 dBZ at –10C is chosen to distinguish the first two periods of heavy precipitation from the second two periods. This threshold appears to identify ice particle sizes and concentrations necessary to produce CG lightning (i.e., high concentrations of millimeter-sized graupel). Note that the 45 dBZ at –10C threshold does not guarantee CG lightning: no lightning is produced between 200 and 240 minutes after 17:25 MDT, even though the threshold is met.
Slide 20 – Sec. 2: FCL radar four-panel display (5 frames)
The radar time-height cross-section in the previous slide took research meteorologists several days to produce. How can the operational meteorologist assess the potential for CG lightning in real-time? The instructors encourage forecasters to use the AWIPS four-panel radar display in conjunction with a proximity sounding.
Frames 1-4: Radar data from the WSR-88D in Denver and temperatures derived from the Denver sounding shown in Slide 17. Five-minute CG lightning data is also plotted. These four frames correspond to the four periods of heavy precipitation discussed with the previous slide. The radar/temperature data highlight the shallow nature of the first two periods of heavy precipitation and the greater depth of the second two periods of heavy precipitation. As in the previous slide, the 45 dBZ at –10C threshold distinguishes those storms that produce CG lightning and those that do not. However, Frame 4 shows that this threshold does not work at all times. For this reason, we consider the radar threshold a necessary but not sufficient condition for CG lightning.
Frame 5: This four-panel plot shows storms southeast of Denver. Three of the storms meet the 45 dBZ at –10C threshold, yet not all produce CG lightning. Again, this reflectivity threshold is a necessary but not sufficient condition for CG lightning.
Slide 21 – Sec. 2: FCL summary
Radar and satellite thresholds are approximate, but capture physical characteristics of thunderstorms. The radar threshold represents the minimum ice content necessary for charging and CG lightning. The satellite threshold represents the minimum cloud depth necessary for ice crystals to separate from graupel, thus forming the normal dipole.
We ask the offices to test these thresholds locally. Published research shows that the radar reflectivity threshold may vary between 35-45 dBZ. Informal research at CIRA suggests that satellite threshold works throughout the year and in most locations.
The –30C cloud top temperature threshold appears to be most useful in cases with low instability such as the Fort Collins case. In cases with greater instability, the threshold may not be as meaningful since clouds are growing rapidly and may penetrate the –30C level before they start to produce CG lightning. The –30C cloud top temperature threshold does not work well with small, isolated storms. GOES IR data cannot resolve the cloud tops of these storms and, thus, the IR temperatures can be much warmer than -30C. This topic will be discussed in Lightning Meteorology II.
To summarize this section, we have identified the minimum ice content and cloud depth associated with thunderstorms. In the next section, we assume that these thresholds are met and examined how the normal and titled dipole charge distributions control the occurrence of CG lightning.