Overview: Week 1, Day 3

 10AM

Day three kicked off with a loaded discussion, with OCTANE, Synthetic GXI Imagery, and the Lightning Stoplight Tool taking center stage. One interesting topic came from the presentation of Lightning Stoplight, and how to interpret regions with no data next to pixels in the 0-10 minute range (red). Would a site manager interpret that area as 'safe'? This also led into a discussion on how forecasters would present these data to partners, with most in the group stating that they wouldn't feel comfortable giving the Lightning Stoplight or LightningCast tools to EMs without considerable on-site guidance or training.

The forecast discussion was split four ways between Texas, Mississippi, West Virginia, and Colorado. We ended up going to the San Antonio, TX (EWX) office for a supercell and large hail threat, and the Charleston, WV (RLX) office for widely scattered convection with a marginal wind and hail threat. Forecast groups were also now tasked with two mock-DSS events, forcing them to monitor multiple events on the LightningCast superdashboard product. The Lightning Stoplight tool was not in AWIPS until 21 Z, but forecasters were advised to use the web display from NASA SPoRT.

1PM

I started the day by going west to Hawaii to collect SZA imagery at sunrise from GOES-West and sharing these with the forecasters. How much more 'usable' information can SZA provide compared the traditional imagery, and how much sooner if possible?

4 PM

After several discussion regarding DSS, LightningCast, and how to interpret the Synthetic GeoXO Imagery, we were able to get the Lightning Stoplight product into our AWIPS-II! Forecasters stated in the morning discussions they want to immediately union these data with LightningCast, and some procedures were made to help them out at the beginning of ops tomorrow. I created an animation of the Lightning Stoplight that includes GLM, ENTLN flashes, and ABI Clean-IR imagery. I set the opacity of the Lightning Stoplight to 80% so the colors stand out, but you can still see some detail underneath. We'll see how well the forecasters like this information and if they make any new displays. This animation shows the dissipation of a thunderstorm on the western side of the scene, and you can see the loss of GLM FED data, while the Lightning Stoplight product changes colors and shrinks in size to indicate a decreasing lightning hazard.

The day ended with a quick look at sea-breeze thunderstorms in eastern Cuba for the SZA imagery. Its interesting to see the low clouds and how long they stay present in the SZA's modified DCPD RGB. We'll see how the forecasters like this in the morning.

Kevin

Simulated GeoXO/GXI Imagery Shows Boundary Associated with CI near Wichita Falls

Figure 1: Clockwise from top-left: (a) 5.15 µm simulated imagery, (b) Simulated water vapor transmittance (0.91 µm-based), (c) 0.64 µm ABI, (d) 10.35 µm ABI

Two simulated GeoXO/GXI products demonstrated skill in identifying the location of convective initiation near Wichita Falls, TX.

Although the HRRR model on which the synthetic imagery is based did not initiate convection soon enough, the imagery did identify the approximate location of initial storm formation near Wichita Falls.  The simulated 5.15 µm imagery (upper-left panel of Figure 1), which focuses on low-level water vapor, showed a southwest-northeast oriented tongue of water vapor extending into the Wichita Falls area.  Even in the model, convection did eventually form on this boundary, and visible and infrared imagery from ABI (bottom panels) confirm the actual location of convection initiation close to that location.  The 0.91 µm-based WVT product (upper-right panel) also showed the boundary, though not with the same level of contrast in this case.  This may suggest that the moisture pooling was focused in the boundary layer, eliciting a larger response at 5.15 µm, or perhaps a different color table could be used on the WVT product to better depict differences in total precipitable water.  The topic of color tables for the WVT product has come up in the discussions with the forecasters.

 -VortexTilting

Monitoring Downburst Potential: A Multi-Parameter Analysis over Mississippi and Alabama

To identify the area of greatest concern for severe convective windstorms during the afternoon of 28 April 2026, the AWIPS composite analysis in Figure 1 seamlessly integrates advanced satellite and radar products. This display overlays LightningCast lightning probabilities and OCTANE cloud phase distinction imagery with Jackson, MS (KDGX) NEXRAD radar reflectivity. This composite highlights a rapidly developing, intense cluster of thunderstorms moving across east-central Mississippi and advancing into western Alabama. Notably, the OCTANE imagery reveals a dense field of altocumulus clouds—indicated by the prominent blue shading—positioned just south of the main thunderstorm area. This cloud field is indicative of a highly dynamic and potentially volatile mid-level environment feeding into the
convective complex, raising situational awareness for severe downburst activity in the immediate downstream environment.

Figure 1: AWIPS composite image of Lightningcast lightning probability, OCTANE cloud phase distinction and Jackson, MS NEXRAD (DGX) reflectivity at 1939 UTC 28 April 2026.

Shifting from a horizontal to a vertical perspective, the pre-convective environment near Hattiesburg, MS, is examined in Figure 2 below using both GOES satellite-derived and Rapid Refresh (RAP) model sounding profiles. There is excellent agreement between the observational satellite data and the mesoscale model output. Both profiles exhibit a classic, convectively unstable "hourglass" thermodynamic signature. Furthermore, the lower troposphere displays a pronounced surface-based "inverted-V" profile. This specific thermodynamic structure is highly favorable for the generation of severe wet microbursts and downbursts. Based on these profiles, the Microburst Windspeed Potential Index (MWPI) calculates a Wind Gust Potential (WGP) ranging between 46 and 53 knots (53–61 mph), signaling a legitimate threat for damaging straight-line winds at the surface.

Figure 2. Comparison of GOES-19 and RAP-model derived soundings near Hattiesburg, MS at 1900 UTC.

To confirm the mesoscale trends, we consult the high-resolution HRRR model sounding over Hattiesburg, MS, via the AWIPS Volume Browser. The HRRR confirms the presence of a deep, surface-based inverted-V signature as shown in Figure 3.

The Thermodynamic Significance of the "Hourglass" Profile:

The "hourglass" sounding is a notorious precursor to severe downbursts. It is characterized by a moist boundary layer near the surface, a pronounced layer of very dry air in the mid-troposphere, and moist air aloft (creating an hourglass shape between the temperature and dewpoint lines on a Skew-T diagram).

Figure 3. AWIPS display of an HRRR-model derived sounding near Hattiesburg, MS at 2000 UTC.

 -StormRangerWX

Big Ol’ Storms, Amazing Satellite Products!

 Today’s focus for my group was over WFO FWD where discrete supercells developed, eventually congealing into a more linear storm mode. Being on the DSS desk today, I primarily used lightning products today, though I did peep at a few others. I wanted to take a moment to highlight the GeoXO 5.15/0.91 products for very clearly defining a moisture gradient over northeast TX that would eventually go on to be our focus point for CI.

Figure 1: WV/WVT/Itob/VISob Compare highlighting distinct moist/dry boundary

LightingCast:

Provided DSS using LightningCast V2 overlaid with OCTANE CONUS CTC. I really like using the products together as the increasing LightningCast probs nudge me to either contact or not contact an EM. Overall, I thought the product did very well. I did notice something odd in the dashboard that I spoke with the developer about - instances of GLM flashes were observed at a time there was a downward trend with LightningCast. Figure attached below.

Figure 2: LightningCast V2 vs V1

For documentation, here was the primary procedure I used today to make decisions:

Figure 3: OCTANE CONUS CTC w/ LightningCast (yellow circle is 8-mile radius DSS event)

Using all of the above mentioned products, here was a CWA graphic I put together:

Figure 4: DSS graphic

Feedback for Day 2

LightningCast: For the Dashboard, I feel like V2 did a better job than V1. The amount of variables being measured is a little confusing, and over time, I noticed I was more just looking at a trend rather than all of the different variables.

Stoplight: Though I never got to give the all clear for my DSS event, Stoplight matched up very well with LightningCast and radar trends.

OCTANE: Made some progress in terms of understanding some of the OCTANE products today, thank you for answering all of my questions, Jason! I really liked that we had 2 live weather events to compare the difference in the wind profiles with OCTANE. Yesterday, we observed a lot of greenish hues, indicative of stronger low-level flow. That element was a bit more lacking in FWD today, and the imagery corroborated that with more blue hues. Jason also showed us examples of cloud top divergence, with more blue/green colors indicating deceleration. After getting to know these products a little better today, I feel less overwhelmed by it.

GeoXO: I continue to be very impressed with this product's moisture and boundary detection. The example I shared above was very impressive and mostly lined up with the 00Z HRRR.

SZA: Did not look at SZA today

-simoom

Tuesdays HWT

Main issue today was to do IDSS near Delhi MS. We were using lightning cast and lightning stoplight. We were also issued a few warnings. Fortunately, we had a storm move over the region so we were quite busy for at least an hour or so.

Our team created some loops of the data. A long loop of GLM and ltg prob is below …

The blinking circle was our DSS site. Lena was our “emergency manager” for the day.

We had thorough conversations with Lena and Kevin regarding using the tow lightning products in operations. We agreed that the products are very useful, but would not work all that well to send these products directly to our partners.

Note that we had a thorough conversation on the slack channel and you can follow the “blow by blow” discussions there.

We also had a great conversation with Lena regarding how we interact with our emergency managers.

Regarding Lightncast, I believe that the planview products are good. I prefer looking at the imagery like this though >>>>

Note that the data probabilities on the awips image above go from 0% to 100 percent.  Why is it important to show the data this way instead of contours? The reason is you can see the VERY beginning of when the ltg probs start to show up. Using contours, you have to wait until the 10% prob shows up. Note that I do not have sat pix imagery overlaid, The reason why I do this with AWIPS is I can have the sat pix data opened up in another panel and use the cursor to read the data. Also the contours (or the image for that matter) would get “lost” (or buried) within the colors of the sat pix imagery.

We used the Lightning Stoplight product today. This was helpful in letting us know when the last flash occurred in the 10, 20 and 30 increment time frame. However I am hesitant with sharing this directly with the EMs.

I envision that the Stoplight product and the LightningCast product will be married in the future, and this will make these two products more useful.

I can not say enough how valuable the lightningcast product is. Prior to this we did not have a good idea when lightning would be a threat, and would have to carefully watch radar data and do cross sections or monitor to -10C level. With lightningcast, you have the probabilities calculated for you in real time every minute (mesosector) or every 5 minutes for CONUS.(However waiting 5 minutes is wayyyyyyyy toooooo lonnnnnnnnnnnnng.

We talked about the SZA product. This is a no brainer and this satellite product should be implemented operationally as soon as possible. Being able to see the detail in the sat pix data towards sunset is priceless.

-Mesovortex

Cloud Top Cooling Signals

The Octane Cloud Top Cooling product provided a strong indication of the continued southwestward development of storms in advance. While in the towering cumulus stage of development, the CTC indicated the stronger updrafts. Later on, those towering cumulus continued to develop into mature thunderstorms. While highlighting the areas that ended up growing, I didn’t notice it flagging any of the bubbling cumulus in areas where storms did not develop. This raises my confidence in the product as being able to use it to identify areas of greater concern. It may be worth noting that high level cirrus hid any potential updraft development in the southern portions of the area. However, in this specific case that proved to be non-consequential 

Figure 1: OCTANE CONUS cloud top cooling at 18:51 zulu. Cloud top cooling can be seen on the line extending SW from the then dominant supercell.

Figure 2: Radar screenshot from roughly the same time as Figure 1.

Figure 3: The OCTANE CONUS cloud top cooling at 1911 zulu. Showering more development and continued signals on the SW flank.

Figure 4: Radar corresponding with Figure 3. Storms continued to develop and intensify in the same region highlighted 20 minutes prior in Figure 1.

-wxboi

April 28th Hazmat DSS

 By Blizzard

Warm Sector Growth: Jaxson MS

The day started this morning with thunderstorm development from the overnight outflow boundary that remained in the southeast. This was quickly replaced by 1500 to 2500 j/kg of MUCAPE. There was 40 kts of shear, which would help to sustain any pre-frontal convection. With dewpoints in the warm sector approaching 75° in Northeast Louisiana. A very shallow cap was present, but quickly eroded from solar heating. Once the cap was eroded we got rapid growth of a thunderstorm complex. This complex moved slowly across Northeast Louisiana and Central Mississippi throughout the afternoon.

DSS operations for a Hazmat: Northeast Louisiana

We coordinated with a hazmat cleanup site that was near the Louisiana/Mississippi border. The emergency manager had required the heads up on lightning within a 10 mile radius of the location. In the figure below we see a developed thunderstorm complex that is moving west to east across the segment of the 10 mile radius. The OCTANE product allowed us to build confidence on a strengthening thunderstorm near the hazmat clean up area. We can see a very moist inflow on the southwestern edge of the complex. It was good that we were able to see a very distinct inflow because we received a report of 1.75” hail. This inflow allowed the thunderstorm to continue to strengthen.

Cessation of Lightning:

The biggest question the public or our partners will ask, is when will the lightning end. We used a variety of tools to determine when it is good to give the all clear. They were the lightning stoplight, lightning cast, and radar. With this event we heavily relied on the stoplight and radar, due to a lag in the lightning cast probabilities. While lightning cast had identified lower chances it would still put over 50% probabilities when the threat had been over for 30 minutes.  

The first graphic shows a quick downward response to the cessation of lightning. However, on the second graphic we see lightning probabilities remain elevated for a long period of time. Using other tools like lightning stoplight and radar, it makes no sense that we should be holding the all clear until the probabilities go near 0. While lightning cast may not be perfect, it is another tool in the toolbox.

Overview: Week 1, Day 2

10AM

Day 2 of the experiment brought another Moderate Risk region from the SPC, driven by a 30% CIG 2 Hail outlook. Thunderstorms were expected to form along an advancing dry line, with a diffuse temperature and moisture gradient in place from southern Oklahoma to northern Alabama as of 15Z. After our morning debrief, the Fort Worth-Dallas office (FWD) was a pretty easy pick for us to target today. After more discussion, we decided our second WFO should be further east ahead of some early morning convection in the Jackson, MS (JAN) office. Unfortunately the Lightning Stoplight product continued to face ingest issues and was unavailable today, and we hope to have it in AWIPS tomorrow. In the meantime forecasters viewed that data on the SPoRT Lightning Viewer and in GR2Analyst.

1 PM

Forecasters today were tasked with issuing simulated Severe Thunderstorm and Tornado warnings in WarnGen on AWIPS-II, along with providing decision support for mock-events that we created for each office. Today FWD was tasked with mock-DSS for a Red River BBQ Cookoff near Gainesville, TX, and JAN was tasked with mock-DSS for a hazmat spill near Waverly, MS. Shortly after operations began, CI occurred just west of the FWD CWA, and thunderstorms were maturing just west of the JAN CWA. This gave everyone plenty of storms to observe, and the forecasters with the chance to issue warnings and communicate hazards through decision support.

I spent part of the day in ops viewing the SZA imagery product on the west coast shortly after sunrise. When comparing SZA data in a modified Day Cloud Phase Distinction RGB against the traditional RGB, the differences are stark right at sunrise, and features stand out sooner. How much sooner is something we need to determine. There was a nice discussion in the morning about how SZA imagery may benefit WFOs in higher latitudes (main example: Alaska). I also compared the modified SZA RGB against CIRA's GeoColor imagery, and found that more interesting than anything.

When zooming into the SZA imagery in north Texas, it appears the 0.69µm (channel 2) imagery has 'gaps' in highly reflective features. My first hypothesis is that the solar zenith adjustment is pushing reflectance factor values above 130% and creating plotting issues in AWIPS, but I feel like traditional ABI visible imagery saw similar artifacts in AWIPS after launch that had to be corrected.

Kevin

New LightningCast Product Tested

 Forecasters were in PAH and STL today getting acquainted with the new satellite products for this year. Although ProbSevere Lightningcast has been at the testbed before, one of the new products for this year was a “super” dashboard. These super dashboards display multiple LightningCast time series, which could be used for decision support services (DSS).

On the top bar, users have the option to change the number of time series displayed, either in a grid or in a custom layout. Users can then select the office they are interested in, change the time range, and add the lightning time series graphs from the Stadiums/DSS or Airport time series. Forecasters can also easily share with their DSS partners and toggle the legend. If a forecaster wants to get a more detailed view of one of the individual time series, clicking on the time series title will open it in a new tab. Forecasters look to use this product throughout the week at the DSS events that pop up in their CWAs.

Figure 1: Example of the ProbSevere LightningCast Super dashboards displaying multiple time series for the Paducah, KY CWA.

Figure 2: Zoomed in close-up of different options forecasters can change.

-Aurora

PDS Tornado Watch: Impressive Sounding at PAH

 Very impressive thermodynamic structure evolving over southwestern Kentucky at mid-afternoon (2000 UTC) 27 April 2026 as apparent in the RAP model analysis sounding over Paducah (Figure 1) with a PDS tornado watch issued for western Kentucky and Tennessee at 1945 UTC. The MetPy-generated skew-T profile and vertical theta-e sounding indicate the increasing convective instability over the mid-Mississippi Valley in a region of strong surface convergence and enhanced cumulus cloud development. To support the monitoring of developing convective storms over southeastern Missouri with a convergence zone in place over western Kentucky, using OCTANE cloud top cooling and LightningCast lightning probability products clearly shows atmospheric destabilization and a ramp-up in lightning activity in Figure 3. The RAP model sounding profile indicates downburst wind gust potential > 50 knots in the PAH area.

Figure 1. TOWR-S VuSkew screenshot of a RAP model sounding near Paducah, KY (PAH) at 2000 UTC 27 April 2026.

Figure 2: MetPy-generated RAP model skew-T plot (top) and theta-e sounding profile (bottom) at PAH at 2000 UTC showing strong convective instability over southwestern Kentucky.

Figure 3. AWIPS screenshot showing LightningCast lightning probability overlying OCTANE cloud top cooling and cloud phase detection products at 2026 UTC 27 April 2026.

-StormRangerWX