CSPP Geosphere imagery, above, centered on LaCrosse, WI (the bluish thermal signal of the Mississippi River is also apparent) shows the development of brighter purple pixels starting at 0636 UTC. That change is perhaps easier to view in the slow stepped animation below. The changes in the RGB shows the initiation of a fire at a facility that recycles railroad ties (news link) in La Crosse. The initial small fire was in a shed before spreading to an adjacent mass of railroad ties.

---

How did the Next Generation Fire System do with this event? That is shown below. The first detection — at 0646 UTC — appeared shortly before 0650 UTC. Once that alert has occurred, you can click on the triangle to have access to the created imagery, showing the first NGFS detection, and from there you can open a RealEarth instance (here, for this example) that includes all imagery.

NGFS microphysics imagery, below, shows very subtle color changes between 0631 and 0636 and 0641 UTC before the NGFS identification of a fire pixel at 0646 UTC.

The Night Microphysics animation at the top of this blog post includes the signal of a cloud band, and the signal of the fire is lost in that animation as the cloud band moves over La Crosse. A particular strength of NGFS fire detections is their persistence in the presence of clouds, as shown in the animation below from 0741 to 0806 UTC. The cloud signal is apparent in the NGFS microphysics, and the fire detection persists through the cloud band’s passage.

View only this post Read Less

Hurricane John was an intense cyclone that affected the Pacific coast of Mexico for a about a week in late September 2024. (Click here to see an approximate path). The storm was noteworthy in that it made landfall, dissipated, re-developed, and then made landfall again. The animation above shows the lifecycle of the system, starting as a tropical wave south of Mexico and ending with a slow approach to the coast on the 26th and 27th. John weakened as it moved along the coast, before dissipating late on the 27th.

Hurricane John made landfall just after 0300 UTC UTC on 24 September. At 0200 UTC, shown below, a well-developed eye is apparent in satellite imagery, and strong convection is wrapped around that eye. The structure of the hurricane was quickly disrupted once inland, and by 1800 UTC on the 24th, the National Hurricane Center ceased advisories on the system, with the caveat that the system could still generate heavy rain, and might redevelop.

At 1200 UTC on 26 September 2024, shown below, John has redeveloped into a hurricane, with curved bands in the infrared imagery, to the east of a large mesoscale convective system. John weakened as it moved parallel to the coast for the next 30 hours, dissipating at 2100 UTC on 27 September.

View only this post Read Less

GOES-16 Visible imagery (Band 2, 0.64 µm) from Mesoscale Sector 1, above, show the evolution of Helene for about 90 minutes shortly after sunrise on 26 September. Strong convection is wrapping around a center that occasionally has an eye-like characteristic. At the time of the animation, Helene was a very large (in area) Category 2 hurricane with maximum sustained winds of 100 mph. The visible imagery shows, along the western edge, anticyclonic cirrus outflow (and cyclonic low-level inflow) in addition to the compelling eye convection.

Helene is destined to interact with a mid-tropospheric system over the mid-Mississippi river valley that is shown below in the GOES-16 airmass RGB animation (the area of red/orange over western KY/TN).

---

Helene is likely to produce prolific rains over the southeastern United States. MIMIC Total Precipitable Water fields for 1300 UTC on 26 September, above (from this site), suggest a stream of moisture moving north-northwestward from the Bahamas/tropical Atlantic into the southeastern United States. This in addition to the moisture moving northward with Helene itself will likely mean very heavy rains.

---

As sunset approached, Helene was in the northeast Gulf, nearing landfall, as a much more organized Category 4 storm. Visible imagery, below, shows convection surrounding a more prominent eye.

The GLM instrument observed lightning throughout the eye during this time of intensification, as shown below. Lightning within the eyewall is a trait of some hurricanes that are strengthening rapidly.

---

Helene poses significant risks to the southeastern US. Landfall is expected after sunset on 26 September, with a large storm surge and heavy rains. For the latest on Helene, visit the National Hurricane Center website, and also the National Weather Service Forecast Offices in Atlanta/Peachtree City, Tallahassee and Jacksonville.

View only this post Read Less

A positively-tilted upper-level trough generated severe thunderstorms across the eastern U.S. recently. The associated surface low-pressure system with attendant cold front spawned numerous severe storms from Mississippi to Ohio to Virigina.

The LightningCast model picked up on this developing convection, providing 10-30 minutes of lead time to lightning initiation in many convective cells (measured from the 30% probability threshold). LightningCast uses an AI model and GOES-R Advanced Baseline Imager inputs to predict the next-hour probability of lightning.

Near Bowling Green, Tennessee, convective cores ramped up quickly into thunderstorms and affected the regional airport with flash rates of 70 flashes per 5-minutes within 5 miles of the airport. The meteogram below shows the probability of lightning near the airport, (the red line is derived from the GOES-16 5-minute CONUS scan and the yellow is derived from the GOES-16 1-minute mesoscale scan) which increased well ahead of observed lightning (blue dots) from the Geostationary Lightning Mapper (GLM).

Further west In Oklahoma, on the backside of the upper-trough, a supercell thunderstorm produced hail up to 2″ in diameter. The storm popped up from below a cloud deck, as observed from GOES-18 (GOES-West) and quickly became electrified. LightningCast provided 17 minutes of lead time to lightning initiation (measured from the 30% contour).

The storm was severe-warned 20 minutes after electrification, and 15 minutes thereafter began producing hail with diameters in excess 1″. Severe hail was reported across the Oklahoma City metro area. From the animation below, we can see the storm’s rightward lurch to the south. The storm contours are from the ProbSevere version 3 system.

ProbSevere version 3 improved upon version 2 in this supercell, with a 55% probability of severe just prior to the initial NWS warning, whereas version 2 was at 16% (see image below). An offline analysis revealed that the top-5 most-important predictors for the storm at this time were:

• Total flash rate (12 flashes/minute)
• Lapse rate 0-3 km (8.7 C/km)
• Effective bulk shear (55 kt)
• Max MESH (0.56″)
• Satellite growth rate (2.1%/minute [moderate])

Mid-level rotation parameters and mid-level storm-relative wind (39 kt) were the next highest contributing predictors. ProbSevere v3 regularly provided improved predictions of severe weather, often extending potential lead time to initial hazards and situational awareness to forecasters. Both LightningCast and ProbSevere v3 will hopefully be operational in 2025.

View only this post Read Less