ProbSevere is a collection of nowcasting products for convective hazards. Two such products are LightningCast and IntenseStormNet, which are AI models that use solely satellite data to make predictions. LightningCast predicts the probability of lightning in the next 60 minutes, whereas IntenseStormNet detects the strongest portions of storms, which are likely to produce tornadoes, severe hail, and severe wind gusts. Both use data images from satellite instruments.

Last week, storms erupted on the South Dakota / North Dakota border within the warm sector of a surface low pressure system. The animation below shows the evolution of LightningCast (left) and IntenseStormNet (right) over time.

One important goal of LightningCast is provide actionable lead time to lightning initiation. Using GOES-16 GLM flash-extent density, we see that LightningCast provided about 30 minutes of lead time to the first flash in the southern cells, and then 25 minutes of lead time to the first flash in the later northern cell, when measured from the cyan 25% probability contour. As storms matured, we noticed how the probability contours were more packed to the west, the upstream side of the convection, whereas they were more diffuse in the downstream direction of convection (eastward). The diffuse nature of the eastward contours indicates less certainty on where lightning will be within the next hour, where the storm cores and associated anvil clouds were heading. We believe this is a signal that the LightningCast model learned the approximate motion of storms from a single timestamp of GOES-R ABI images. More research is underway using time sequences of images to see if that can improve implicit motion estimates and lightning predictions.

The northern cell in central North Dakota impacted Bismarck Airport. The lightning dashboard below shows how LightningCast probabilities from both GOES-East and GOES-West increased initially over the airport, then were fairly steady, and then rapidly increased over about 10-15 minutes before the first lightning strikes were detected within 5 miles of the airport by the both GOES-East and GOES-West GLM instruments. These dashboards, available to forecasters at airports, many stadiums, and over wildland fires, help to sync the LightningCast probability information and GLM-observed flashes with potentially vulnerable locations of interest, facilitating the National Weather Service’s decision support for key entities.

Looking again at the animation above, the IntenseStormNet uses images of 0.64-µm reflectance, 10.35-µm brightness temperature, and GLM flash-extent density to predict a probability of “intense” convection. The product really highlights the intense most regions, where strong overshooting tops and “cold-U” signatures reside, often colocated with robust “bubbly” texture from the visible reflectance. Notice the fact that many reported severe weather events werre associated with these features, created by the strong storm updrafts. Nearby pronounced anvil-edge gradients in the ABI channels and a core of GLM flash-extent density are also important features for contributing to higher probabilities.

IntenseStormNet output is a predictor in ProbSevere v3, machine-learning models that combine radar, satellite, lightning, and environmental data to predict the next-hour probability of severe hail, severe wind gusts, and tornadoes. IntenseStormNet can also be used as a stand-alone satellite-only severe-weather nowcasting tool in the absence of radar, such as in parts of the western U.S., in Canada, in oceanic regions, and in parts of Latin America.

LightningCast should be operational at NOAA/NESDIS in early 2025.

View only this post Read Less

VIIRS data from NOAA-20 at the CIMSS Direct Broadcast download site, above, at 1734 UTC on 2 September were used to create True Color imagery, and SST fields over the largely-clear Great Lakes. Light winds associated with persistent high pressure over the Great Lakes (below) meant suppressed vertical mixing, leading to warm water temperatures on the lake surface. Most of Lake Superior’s surface is warmer than 60^oF, with portions of Lake Erie, Saginaw Bay, and Green Bay near 77^oF. This is also shown in the AWIPS Screen-capture with the same data, below.

Early on 1 September, skies were also mostly clear, and the VIIRS ACSPO SSTs are shown below in concert with Day Night Band visible imagery. The warmest temperatures in the color enhancement (white) are 77^oF (that is, 25^oC).

Clear skies continued overnight on 2-3 September, and another beautiful scene was produced. Lake Surface Temperatures and the Day Night Band at 0740 UTC on 3 September are shown below.

True-Color imagery from VIIRS and Lake Surface Temperatures are shown below from 1834 UTC on 3 September. Note the complete coverage over western Lake Superior. The anomaly in the analysis above has been rectified. Smoke is also apparent in this image over northern WI and central MN.

The smoke shows up well in the VIIRS AOD field, below (and in the GOES-16 fields as well here)

View only this post Read Less

Nighttime Microphysics RGB Imagery (Quick Guide) from the CSPP Geosphere site, above (direct link to image), from early on 1 September, shows a distinct color change from the southern US to the northern US, stretching east-northeast from Oklahoma to Pennsylvania. This abrupt change from deep purple to light purple heralds a change in moisture — as one might expect in early September as cooler, dryer air starts its inexorable move south as Summer wanes. The cyan color of low clouds is also present along this boundary (and in central Iowa). Part of the color change will be controlled by temperature as well: Blue in the RGB is a function of the Band 13 brightness temperature.

MIMIC Total Precipitable Water fields, above, from 0700 UTC, also show the sharp gradient in total precipitable water along this boundary as values drop from near 2″ over Arkansas to closer to 1/2″ over central Missouri. Blended TPW fields, below, taken from the CIRA SLIDER, show a similar sharp boundary.

How did Precipitable Water (PWAT) change across this boundary in the 0000 UTC Soundings on 1 September? Skew-Ts, below (from the Wyoming Sounding site), from Lincoln IL (left, in the dry air), Springfield MO (center) and Little Rock AR (right), show PWAT values of 15, 49 and 51 mm, respectively!

Hat tip to Jochen Kerkmann, EUMETSAT, for noting this event today and sending out an email.

View only this post Read Less

True-color imagery, above, shows the evolution of the Remington Fire in extreme northern Wyoming on 22 August 2024. (Geography note: The Wyoming/Montana border in the image is at 45^oN Latitude). The initial report for this fire occurred at 8:07 AM (MDT)/1407 UTC on 22 August 2024 (note that the time at the Watch Duty website (2:33 PM, link), below, refers to the report time at that site). This was after the NGFS detection. Fire initiation occurred in a sparsely populated region of northeast Wyoming, so it should not be a surprise that the initial report was after the fire was underway.

In contrast, the weather satellite-based Next Generation Fire System detected fire at this location nearly a day earlier, starting at 2143 UTC (3:43 PM Mountain Time) on 21 August (!) as shown below. This is an excellent example showing NGFS’s importance for wildfire detection in remote regions where human eyes (or webcams) aren’t available to monitor things. The animations below show the GOES-based Fire Temperature RGB and the NGFS Microphysics RGB. The complex nature of the RGBs is in part due to cloudiness occurring over the region. (Here’s the visible imagery at 21:47:54) An NGFS detection is shown by the orange polygons at 21:47:54 UTC on 21 August. Northern Wyoming was in a region of elevated fire risk as shown in this graphic from SPC.

An important feature of the NGFS RealEarth instance is that imagery can be probed. By clicking on the NGFS pixels (that is, the orange polygons), users can view information about the surface properties in the region of the developing fire. The Probe for this NGFS polygon is shown below, including the latitude/longitude (just south of the Wyoming/Montana border), and the physical description of the land — mostly vegetation — grass and chapparal — and very little urban areas. This information can help a forecaster determine the likelihood of fire and risk to populations. It might also influence a decision to commit fire-fighting resources.

NGFS output is available at this website. Previous blog posts on NGFS capabilities are here.

View only this post Read Less