Landspout tornados and severe thunderstorms impacted southeast Colorado and southwest Kansas during the afternoon of 21 May 2020 (spc reports). GOES-East satellite imagery was used by NWS PUB forecasters to analyze the cumulus cloud field leading up to convective initiation, and gauge convective evolution as storms matured.

Analyzing 1-min visible imagery during the two hours between 1859 and 2059 UTC, the cu field across southeast Colorado from just north of Holly to Campo became increasingly agitated, with the greatest clumping and vertical growth occurring between Lamar and Holly (Fig 1). These cumulus cloud trends were signs of imminent convective initiation. Additionally, forecasters diagnosed significant low-level cyclonic vorticty given the evolution and character of the broad cu field in the imagery across southeast Colorado into southwest Kansas. Finally, the movement of cu clouds to the west with towering cu tops and orphan anvils drifting to the east highlighted the presence of substantial low-mid-level shear.

The insight gleaned from satellite imagery along with additional knowledge about the environment gave forecasters confidence that a landspout tornado threat existed with any storm that might develop in the area, in addition to the large hail threat. The trends in satellite imagery also helped forecasters to message this risk to core partners and the public. Two tweets were sent during this period, highlighting the threat for strong to severe thunderstorms (Fig 2), and landspout tornados (Fig 3).

Visible imagery during the following two hours (2059 – 2258 UTC) showed continued upscale growth of cumulus clouds across the region, including convective initiation and eventual development of strong-severe thunderstorms (Fig 4). Multiple instances of landspout tornados were confirmed during this period near the CO/KS border in association with the growing cumulus clouds.

In addition to the landspout threat, forecasters were motioning the development of severe thunderstorms. As convection continued to grow upscale and mature from Springfield to north of Holly/Lamar, forecasters paid close attention to IR-Window imagery (Fig 5). The IR imagery showed where the most rapid convective initiation was occurring, and was used to help forecasters determine which cell(s) would become most dominant. The most impressive and persistent cold cloud top temperatures and overshooting top developed within the broader region of convection with a storm between Lamar and Holly. This storm also quickly produced an above anvil cirrus plume, indicating a particularly strong updraft and severe potential. Severe thunderstorm and tornado warnings were issued with this region of development, which produced confirmed tornados and baseball sized hail.

Combining the VIS and IR into a sandwich image combo reveals the development of a dominant cell quite well with this case through the blending of texture and brightness temperature information (Fig 6). Details of the overshooting top and above anvil cirrus plume become obvious.

Viewing an animation covering the full duration of this event, the evolution of the cu field in the context of RAP analyzed sfc vorticty and 0-3 km MLCAPE (good fields for assessing landspout potential) is observed (Fig 7). The large values of sfc vorticty from the model generally agree with what is implied from the satellite imagery. Values of 0-3 km MLCAPE increased over and east of Campo during the afternoon. The region of overlap between the two fields near the CO/KS border is where landspout tornados were observed.

Bill Line (NESDIS and CIRA) and Klint Skelly (NWS PUB)

Very dry and breezy conditions developed across southeast Colorado during the afternoon of 20 May 2020 ahead of a broad western US upper trough and behind a surface dryline. With deep vertical mixing occurring behind the dryline, RH values fell to around 10 %, and winds gusted 30-40 knots. A wildland grass fire initiated north of Kim, CO during the early afternoon.

A GOES-East 1-min meso sector was available over a region including the wildfire during the day in support of convective warning operations (Fig 1). A watchful eye would notice the flickering of a faint hot spot below thin cirrus clouds by around 1945 UTC. With 5x more images , a forecaster can be more confident in wildfire detection earlier when using 1-min imagery over 5-min imagery. A simple gray-scale color table applied to ABI single-band 3.9 um imagery has proven to be a very reliable method for detecting wildfire starts, especially when 1-min imagery is available.

Later, as the fire continued to grow, an impressive smoke plume developed. The 1-min visible imagery shows the rapid growth of the plume, and even development of pyrocumulus adjacent to a colder cloud shield. IRW brightness temperatures associated with the pyrocu reached -20C shortly after 0100 UTC.

A long animation of the wildfire combines visible, SWIR and IRW imagery along with RAP surface analysis RH and wind gust fields for the duration of the event. The animation shows the fire initiating as wind gust speeds increase over 30 knots and RH values drop below 15%. The fire continues to burn hot well after sunset, before decreasing in intensity as wind speeds died down and RH values rose.

A 4-panel image of the wildfire and smoke plume during the early evening provides a variety of unique RGB displays for tracking the hot spot and associated smoke plume in unison. At this time, the hot spot was quite expansive, and pyrocumulus developing within the smoke plume.

S-NPP and NOAA-20 VIIRS imagery also captured the wildfire during it’s early stages, and during the evening. The same gray scale color table is used for each VIIRS/ABI image pair.

The 375 m I-4 band (3.74 um) imagery captured the wildfire earlier than was possible in the 2-km ABI imagery, and pinpointed it to a smaller region (a single dark pixel at 1915 UTC; Fig 5). At 2000 UTC, the fire had spread to a few pixels, but was still quite faint in the ABI imagery. However, as was shown, the animations of 1-min ABI imagery (not possible with VIIRS) allowed for confident detection of a hot spot by this point. By 2051 UTC, the wildfire hot spot was obvious in both VIIRS and ABI, but with VIIRS narrowing down the location to a much smaller area.

During the evening, both instruments continue to detect heat associated with the wildfire/burned area, with VIIRS pinpointing the burn region more precisely (Fig 6).

Bill Line (NESDIS and CIRA) and Mike Umscheid (NWS DDC)

Two rounds of severe convection developed over the Texas Panhandle during the afternoon and then evening of 07 May 2020. Storms developed amid increasing forcing associated The Amarillo NWS forecast office (AMA) utilized GOES-East imagery to track development.

GOES-East water vapor imagery highlighted a well-defined upper low digging southeast into the central US plains, with periodic and much more subtle perturbations in the W/NW flow to it’s south (Fig 1). Large scale forcing associated with these features aided convective development along surface/low-level boundaries.

The first round of severe storms developed over the southeastern portion of the Texas Panhandle along a dryline during the afternoon. Leading up to convective initiation, the AMA forecasters monitored the 1-min Day Cloud Phase Distinction RGB imagery for signs of imminent convective initiation (Fig 2). By 2055 UTC, it was apparent from the imagery that convective initiation was most most likely in the near-term over the southeast portion of the panhandle along the AMA/LUB county warning area (CWA) border. The cumulus cloud field was well established and maturing, cloud color was transitioning from cyan to green (glaciation), and orphan anvils were generated (failed initiation attempts). Shortly thereafter, deep convection initiated successfully just south of the CWA border in LUB’s area.

Forecasters from AMA noted the value of 1-min VIS/IR sandwich combo imagery as the convective scenario evolved (Fig 3). Of note early on was a new, stronger updraft developing ahead of the original updraft, quickly producing an above anvil cirrus plume. By 2216 UTC, a new updraft developed ahead of the main storm per cooling of cloud tops and increased cloud texture. This area of convection developed within the AMA CWA and forced the issuance of a severe thunderstorm warning. A storm split was apparent in the imagery a little later, by 2230 UTC, with continued cooling of cloud tops on the northeast cell (left split) and development of an OT. The right split accelerated to the southeast and maintained a large OT and prolific AACP. Forecasters noted the early signs of new updraft development and storm split were apparent in the 1-min satellite imagery just prior to radar.

An uninterrupted loop overlays MRMS MESH on the sandwich imagery, revealing the long hail swath associated with the main cell and right split, and apparent hail development from the forward convection in the southeast corner of AMA CWA, and eventually with the left split (Fig 4).

Later on that evening, additional convection developed across SW KS and the OK/TX Panhandles along a cold front within increasing large scale forcing for ascent ahead of the aforementioned shortwave. AMA forecasters noted the value of the 10.3 um IR window channel at night for analyzing this second round of convection (Fig 5). The imagery was utilized to identify the main updrafts of the stronger cores aloft given the appearance of overshooting tops, and to infer which updrafts were more likely to sustain themselves given cloud top IR temperature trends.

An animation of grayscale IRW and GOES-derived CAPE for the full-duration of the event shows a corridor of increasing instability ahead of the dryline during the afternoon, along/within which the initial round of convection developed and evolved. Instability remained high to the north ahead of the pressing cold front, with the second round of convection developing along the northwest CAPE gradient (cold front) as analyzed in the satellite product.

This blog post from CIMSS highlights the value of NUCAPS products from this event.

Bill Line (NESDIS and CIRA) and Kaitlin Rutt (NWS AMA)

Strong winds on the western portion of a deep surface low in association with an exiting shortwave trough resulted in lofted/blowing dust from the IA/NE/MO/KS Missouri River Valley (MRV) southeast across central Missouri. This is a favorable region for winds to go unimpeded, and the lofted dust is likely due in part to sedimentation/silt left behind from the big spring flooding of 2019. Further, it is still early in the planting season, so fields are susceptible to having dirt erosion. Local media shared photos of blowing dust across I-29, which runs along the eastern side of the MRV (Fig 1).

The Omaha NWS office noted the blowing dust on social media in the nighttime microphysics RGB (Fig 2). This RGB is effective in detecting lofted dust due primarily to the inclusion of the 12.3-10.3 um split window difference (SWD), which has been shown to be an effective ABI channel combination for dust detection (see previous blog post).

Given the conditions, the NWS in Omaha issued a Special Weather Statement for “reduced visibility due to blowing dust” along I-29, and the NWS in Pleasant Hill mentioned, “reductions in visibility as a result of the blowing dust” in their Wind Advisory.

The SWD-IR combo procedure perhaps provides the best depiction from ABI of the blowing dust across the region for the duration of the event (Fig 3). The blowing dust appears as very dark gray to black (very low positive or negative difference values), while most clouds will appear as colors with the inclusion of cold brightness temperatures from the IR window channel. Small cumulus clouds will be lighter gray) The lofted dust is diagnosed to initiate near northwest Missouri in the presence of a tight surface pressure gradient and 40+ knot wind gusts. The lofted dust is carried southeast over I-29 and into central Missouri as it wraps around the southwest and southern portion of the surface low. The animation, lasting 10 hours, depicts the long duration of this event.

The blowing dust can also be diagnosed in the 500 m 0.64 um VIS channel. The color table used is modified to highlight lofted material (Fig 4). The blowing dust becomes most apparent near sunset, when forward scattering of dust particles toward the satellite is heightened.

Finally, a modified Dust RGB provides a great depiction of the blowing dust, while capturing cloud details as well (Fig 5). In this example, the dust appears as deep magenta, low/cumulus clouds dark blue, stratus or mid-level clouds as red, and cirrus clouds as black. The RGB recipe is shown in Fig 6.

SNPP and NOAA-20 consecutive overpasses provided higher resolution (750 m vs 2 km for GOES) VIIRS M-band SWD imagery near the beginning of the event, allowing for slightly more details (spatially) to be gleaned from the lofted dust. (Fig 7).

Bill Line (NESDIS and CIRA) and Andrew Ansorge (NWS DMX)

Severe thunderstorms developed across the southern plains on 4/22/2020 in association with a shortwave trough traveling east across the TX PH, OK, and N TX. GOES-East water vapor imagery shows the evolution of the large scale feature and associated thunderstorm development, with RAP analyses quantifying the wave and it’s influence on the surface pressure and surface/mid-level wind fields (Fig 1). The vorticity max becomes increasingly well-defined as it accelerates east across OK, with associated dry descending air through west Texas and moist ascending air over east TX/OK into the southeast US. Additional imagery from some of the notable thunderstorms will be highlighted in this blog post.

Thunderstorms were ongoing across the northern half of OK and initiating across the southern half of Oklahoma during sunrise on the 22nd. Analyzing GOES-East 1-min visible imagery, shadows associated with the low sun angle provided excellent detail about storm initiation and storm top features such as overshooting tops (OTs), above anvil cirrus plumes (AACPs) and anvil gravity waves (Fig 2).

During the early-mid afternoon hours on the 22nd, additional convection developed across southern Oklahoma near the interface of a dryline/cold front/warm front surface triple point. A GOES-East 1-min VIS-IR sandwich combo allows for a more detailed analysis of the thunderstorms than a either single channel alone (Fig 3). This sandwich combo was created using a VIS linear color scale with a semi-transparent IR overlay, and can be found on the STOR VLAB page. The procedure maintains details from the VIS while also including quantitative information from the IR that can be visualized and sampled. Features such as OTs and AACPs, as well as cooling/warning trends, are easily diagnosed in the imagery. Numerous NWS severe thunderstorm and tornado warnings were issued with these storms, and are shown in the animation. Note, while the base of the storms fall within the polygons, the storm tops are oriented north of the polygons due to parallax. These storms did produce tornados, severe hail, and severe wind gusts.

Further south in east Texas, a long-lived severe thunderstorm produced a track of tornado and severe hail/wind reports. This storm had an impressive appearance in GOES-East imagery, with persistent and large OTs, long-lived AACPs, and very cold tops (<-80C at times). A simple procedure in AWIPS combines the aforementioned VIS-IR sandwich combo with IR imagery to allow for a smooth transition between the two products (make the low end of the VIS transparent). A 2-hour animation of GOES-East 1-min imagery centered over the the storm using this procedure is shown in Figure 4.

A longer, 5-min animation shows the full 10-hr duration of the storm, and instead uses the “Daylight Transition” image combination feature available in AWIPS (Fig 5). Feature-following zoom is also used to provide a storm relative view of convective evolution.

Finally, a similar animation but highlighting GLM Flash Extent Density is available in Figure 6. The storm consistently produces an abundance of total lightning, with periodic lightning jumps/dips throughout the evolution.

Bill Line, NESDIS/CIRA

Severe thunderstorms developed across the southeast US ahead of a potent shortwave trough during the day/evening of 19 April 2020. Water vapor imagery captured the strengthening shortwave as it accelerated east across the southern plains and into the southeast (Fig 1). RAP analysis overlay helps to quantify the shortwave, revealing the sharpening 500 mb trough in the height field, increasing 500 mb wind field ahead of the wave (including wind speeds over 70 knots), and the vorticity max.

By the late evening hours, the primary region of thunderstorms had shifted east into Alabama, Georgia, and the Florida Panhandle. Unfortunately, NWS TAE required emergency backup during a period covering at least 0545 – 0745 UTC, resulting in NWS Houston taking over warning issuance. To make matters worse, data from area radars was unavailable/intermittent to forecasters, primarily during the ~0600 UTC to 0700 UTC timeframe (Fig 2). Therefore, the warning forecaster was required to rely on satellite and lightning data for warning decisions.

The primary tools used following the loss of radar data was GOES-East IR imagery (Fig 3) and total lightning data from Earth Networks (not shown in examples), and later, GLM (Fig 4) as well. Cloud top trends in the IR imagery were monitored closely, including brightness temperature trends and overshooting tops. Persistence of overshooting tops and cold temperatures for storms that had previously appeared strong/severe in radar (and were warned on) aided confidence in continued warning issuance, as did rapid increases in lightning activity. Similarly, warming of storm tops, loss of OTs, and decreases in lightning density provided confidence in letting warnings expire.

Earth Networks lightning data was also used to account for parallax in the satellite imagery, helping a forecaster determine more accurately where the updraft core was located geographically. In this case, storm tops would have been displaced roughly 10-14 km to the NNW in the GOES-East imagery and GLM data.

***AWIPS display/procedures shown in this blog post were created after the event, and are similar but not exactly that used by the warning forecaster***

Bill Line (NESDIS and CIRA) and Sean Luchs (NWS HGX)

A hail storm moved south-southeast across portions of southwest Kansas during the late afternoon hours of 19 April 2020. This storm produced copious amounts of hail along a path of about 15 miles.

The evolution of the cu field across central Kansas can be analyzed in detail using the the GOES-East Day Cloud Phase Distinction RGB. The cu field evolves from flat, liquid clouds (cyan), to a more agitated cu field with clumping clouds and enhanced vertical growth and glaciation (green; Fig 1). With convective initiation, vertical growth is rapid, and cloud tops transition to yellows and reds with cooling and continued glaciation of cloud tops. A glimpse of the hail swath (green) is seen in the wake of thunderstorms, east of Kinsley.

The sky cleared off just enough prior to sunset such that an enhanced version of the GOES-East Day Cloud Phase Distinction RGB product could remotely sense the hail swath (Fig 2).

Numerous pictures of significant hail accumulation were received by NWS Dodge City on social media, some of which showing 4 to 5 inches of accumulation (Fig 3).

The hail swath stuck around through the cool evening hours, and was still apparent in GOES-East imagery the following morning (Fig 4, Fig 5).

GOES-East provides a fairly high resolution confirmation as to where exactly accumulating hail fell following passage of a thunderstorm and clearing of clouds. While MRMS provides a good estimate to where hail fell and how large it was, it doesn’t lend much detail about hail accumulation. The satellite imagery provides an observation of the hail swath, or where the more significant accumulating hail occurred. In this case, the swath diagnosed in GOES imagery matched up well with the MRMS MESH track, but the actual length of the accumulation of hail did not extend as far south into Kiowa County as what may have been suggested by MRMS MESH.

Mike Umscheid (NWS Dodge City, KS) and Bill Line (NESDIS and CIRA)

GOES-East upper level (6.2 um) water vapor imagery with GLM and RAP analysis overlay depicted the Friday night – Sunday morning evolution of a storm system that brought severe thunderstorms to the southern US (Figure 1). A closed upper low initially centered over southern California on Friday accelerated east across northern Mexico Saturday as it evolved into an open wave, and eventually progressed ENE into the southern plains and southeast Sunday into Monday morning. Water vapor imagery shows the drying/warming descending air (warm colors) wrapping around the southern portion of the strengthening shortwave in conjunction with the intense upper level jet. Strong ascent is apparent east/northeast of the upper energy below the exit region of the upper jet represented as regions of cooling (cool colors and white) and developing thunderstorms. The RAP overlays help one to conceptualize what is being observed in the more detailed (temporally and spatially) water vapor imagery.

Additionally in Figure 1, a shortwave trough is diagnosed digging southeast across the northwest US Saturday and then east across Colorado into the central high plains on Sunday as it swings around the base of a broad upper low slowly sinking south into the far north-central US. This system brought a swath of snowfall to the Rockies, central high plains, and into the Midwest, which is apparent in the GOES-East Day Cloud Phase Distnction RGB as green (Fig 2).

Visible/GLM FED Sandwich combo imagery with NWS warning polygons during the daytime shows the development and evolution of thunderstorms throughout the day, and the relationship between lightning activity and strong (warned) storms (Fig 3). GLM is effective in highlighting the location of updraft cores and updraft trends within a messy cloud field and long flashes extending outward from the main updrafts and into the anvils. The image combination includes quantitative information from GLM without sacrificing the vital texture information from the VIS.

Bill Line, NESDIS and CIRA

A low pressure system deepened rapidly over the Atlantic just off the Carolina coast during the day on 01 April 2020. GOES-East water vapor imagery and RAP surface analysis provide a great visualization of the strengthening low as it progressed east away from the coast (Fig 1). Dry/warming descending air (warm colors) is evident wrapping around the low from the south, while ascending/cooling air (including deep moist convection) is obvious further east, north, and west of the low.

As the low strengthened, very gusty winds developed at the surface, prompting the issuance of a Hurricane Force Wind Warning by the NWS. HRRR model analyses indicate widespread wind gusts in excess of 50 knots wrapping around the southern and eastern portion of the low (Fig 2).

GOES-East low-level DMWs, while not abundant nearest the center of circulation where cloud streets developed and winds were likely strongest, did produce several wind vectors over 50 knots near the low center, indicating strong flow just above the surface.

METOP-B ASCAT observed surface winds in excess of 50 knots wrapping around the western portion of the low between 1500 and 1600 UTC (Fig 4a). Later between 1800 and 1900 UTC, AMSR2 measured surface wind speeds between 50-60 knots on the western portion of the low, and 40-50 knots wrapping around the southern quadrant, between 18-19 UTC (Fig 4b).

An area of clearing was present adjacent to deep moist convection near the low center during the morning hours after sunrise. GOES-East visible (0.64 um) imagery captured rough seas (white caps) and associated/implied sea spray under the clear skies, confirming gusty winds reaching the surface during that period (Fig 5). The white caps/sea spray is diagnosed by regions of higher reflectance compared to nearby calmer seas. A modified gray-scale colortable is utilized in order to best highlight the white cap/lofted sea spray signature.

A zoomed in feature following animation provides an alternative means for viewing the evolution of the white caps through the morning (Fig 6). The phenomenon is most prolific extending south and east from the region of thunderstorms in the center of the scene.

The longer wavelength 0.865 um imagery provides better contrast for detecting the white caps (compared to clear sky calmer seas) given less influence of atmospheric aerosols compared to at the shorter wavelengths (Fig 7). However, spatial resolution is degraded by 4x compared to the 0.64 um channel with ABI, reducing clarity of the feature.

The white caps can be diagnosed in the snow-cloud RGB as regions of bright blue (Fig 8).

Finally, the milky appearance of the white caps was easily apparent in the Geocolor (true color) imagery (Fig 9).

A longer feature following animation shows white caps and sea spray continued along the southern portion of the low center into the afternoon and early evening hours (Fig 10).

Bill Line, NESDIS and CIRA

A remarkable shortwave trough brought widespread strong to severe thunderstorms to the Mississippi River Valley region during the day/evening of 28 March 2020.

GOES-East 6.2 um water vapor imagery with RAP field overlays provides an excellent view of the synoptic setup and evolution during this event (Fig 1). Features such as the positions of the low and upper level jet core as depicted in model analyses (such as the RAP here) are confirmed/corrected through analysis of the water vapor imagery.

Focusing in on the Midwest, convection developed along a dryline and warm front, with thunderstorms eventually producing large hail and tornadoes. The GOES-East Split Window Difference (SWD) and Infrared Window Combo imagery (available on the STOR) captures the evolution of the dry air north and east around the southeast portion of the deepening cyclone, and convective initiation along it’s leading edge (Fig 2). The GOES imagery provides better horizontal spatial and temporal resolution when compared to surface observations and surface analyses (hourly RAP surface equivalent potential temperature shown here). The simple gray scale color table of the SWD shows relatively dry low-level air as darker gray, with relatively moist regions lighter gray. An overlay of IR window is provided for cold brightness temperatures (clouds) as non-gray colors). The dry air is diagnosed surging northeast through eastern Kansas into northern Missouri and southern Iowa. Convection develops along the dryline, as diagnosed in the SWD imagery, across Iowa.

While the GOES-East TPW derived product also captured the punch of drier air northward and dryline boundary evolution (at lower spatial res), the CAPE product only ever shows values less than 500 j/kg (Figs 3 and 4). This is lower than what was observed by radiosondes and what was computed by model analyses and the SPC mesoanalysis.

One-minute imagery from GOES-East was available across the region during this event. Day Cloud Phase Distinction 1-min imagery showed deepening cumulus clouds along the dryline, with a transition from cyan to green indicating glaciation and imminent convective initiation (Fig 5). Continued vertical growth and transition to yellow and red colors indicates further glaciation and cooling of cloud tops, and that convective initiation has occurred. Shortly thereafter, the first lightning flashes occurred with these storms per GLM FED data. FED values then increase quickly leading up to the first reported tornado. All of this is apparent in real-time given the very low latency (<1-min) of the 1-min imagery.

Polar passes from SNPP (x2) and NOAA-20 meant three VIIRS images within about a 1.5 hour timeframe. This imagery provided higher spatial detail in the cloud field as convection began to initiate along the dryline. Day Cloud Phase Distinction RGB imagery created from the VIIRS 375 m I bands allows for the diagnosis of highly detailed glaciation trends in the cloud field (Fig 6).

Further south, strong thunderstorms developed within the broader cloud shield over northeast Arkansas, with one storm producing a tornado that caused damage and injuries across Jonesboro, AR. GOES-East 1-min visible imagery with GLM semi-transparent overlay shows increasing visible texture through the broad cirrus shield as the storm approaches Jonesboro from the southwest (Fig 7). FED values increase quickly as well just prior to/during the development of a tornado, with MRMS low-level rotation tracks confirming significant rotation as the storm advanced through the town.

Bill Line, NESDIS and CIRA