Satellite Liaison Blog

Strong northwesterly winds across North Dakota snowpack resulted in widespread blowing snow on 18 Jan 2022. The blowing snow reduced visibility considerably (less than 1 mile) in many areas, as was highlighted by NWS Bismarck, ND here (below) and here. NWS offices leveraged satellite imagery in their forecasts/analyses and communication of blowing snow during this event.

NWS Bismarck leveraged GOES satellite imagery to help shape their understanding of the blowing snow event and influence their forecast products. Specifically during the late morning, satellite imagery of blowing snow resulted in the addition of a county to the Winter Weather Advisory: “Quick update to add Divide County into the Winter Weather Advisory given satellite imagery suggesting significant blowing snow continuing from eastern parts of the county upstream into southern Canada.”

In the early afternoon, NWS/BIS provided a detailed update to the forecast/analysis of blowing snow, shaped by several observational data sources, including trends in satellite imagery: “Web camera trends suggest the lowest visibilities are more variable than ASOS/AWOS sensors alone would suggest, while satellite imagery suggests the most significant blowing snow is related to well-defined plumes that are occurring in Horizontal Convective Roles (HCRs). For the most part, those HCRs are relatively widely-spaced, leading to the variability in visibilities spatially, and temporally as the HCRs shift slightly with the background flow. Satellite imagery does suggest the most widespread blowing snow plumes are centered over Burke County and vicinity, where impacts are likely most significant. In the end, we continue to monitor trends for the need for any Blizzard Warning upgrades, but are holding off for now. Changes with this update cycle were mainly focused on observational trends through the afternoon hours, with no significant adjustments.”

GOES-East satellite imagery referenced in the discussion, specifically the default Day Snow-Fog RGB available in AWIPS, is shown in Fig 1. The plumes of blowing snow organizing into HCRs are easily diagnosed in the imagery streaming south across North Dakota. These are highlighted in the imagery as subtle difference in color (brighter red) compared with the background snow-covered surface, and the shadowing along the northern edges of the tall plumes.

As was introduced in this recent blog post, an attempt to further highlight areas of blowing snow, and associated HCRs, is made with an experimental Blowing Snow RGB (Fig 2). Regions of blowing snow, especially the HCRs, further stand out against non (or weaker)-blowing snow, snow-covered background.

Afternoon VIIRS passes provide a detailed view of the blowing snow plumes and HCRs via a similar experimental Blowing Snow RGB (Fig 3).

The afternoon NWS/BIS forecast discussion included further analysis blowing snow in satellite imagery. “Satellite shows well-defined blowing snow plumes embedded in Horizontal Convective Rolls (HCRs), which have been becoming somewhat more widespread in that corridor as temperatures fall to 0 F or below, increasing the ability for snow to be lofted, especially in areas where the heavy snow fell late last week and strong winds today eroded any crust….We will however continue to monitor the situation, especially since satellite trends do suggest some increase in HCRs and related blowing snow plumes recently, perhaps as the boundary layer depth shrinks a bit. Interestingly, those satellite images also suggest blowing snow is being transported as far south as Burleigh County, where the pre-existing snowpack on the ground is indeed sufficiently crusted to not be broken even with the winds today.”

NWS Grand Forks, ND also leveraged satellite imagery during this blowing snow event, discussing: “… along with horizontal convective rolls are producing snow showers and based on satellite trends, have increased in coverage, pushing into the central valley.” Downstream, NWS Aberdeen was monitoring satellite imagery for blowing snow: “Based off of satellite, we are seeing blowing snow in North Dakota continuing to push southeast. With the snow in our area crusted over, that should limit any lofting of snow from the strong winds…”

NWS/FGF further took advantage of satellite imagery by including it in their public messaging of the hazard (also below).

Compare the above RGBs with single-band VIS and IR imagery in Figs 4 and 5, respectively. While the presence of blowing snow/HCRs can be realized given the shadowing in the VIS and slight temperature difference in the IR, the exact location and extent of blowing snow, especially non-HCR areas, is much more difficult compared to in the multi-spectral, RGB images.

Bill Line, NESDIS and CIRA

Low, thin clouds and fog developed across northern Illinois snowpack overnight into the morning hours of 13 Jan 2022. Comparison of single band imagery with multispectral Imagery and level-2 products remind us of the advanced methods we have for analyzing low clouds and fog in complex scenes, in this case, during the day. One-minute imagery was available over the region during the period, which is quite valuable when monitoring the real-time evolution of low clouds near TAF sites.

Starting off with visible imagery, the scene is encompassed with a lot of light gray to white, representing both snow and cloud cover, diagnosed by stationary and non-stationary movement, and shadows (Fig 1). Near the IA/IL border, some movement is noted representing likely low clouds, but the delineation between low clouds and snow cover is difficult.

While Geocolor establishes a more familiar scene from space, including adding color to non-snow surface features, it doesn’t do much to help to separate clouds from snow cover (Fig 2).

IR Window imagery helps capture the higher (colder clouds), but again doesn’t help much with differentiating potential low clouds from the similar-temperature snow surface (Fig 3).

By combining the VIS, IR and Snow/Ice bands, we are left with a picture that separates snow cover (green) from low (liquid) clouds (cyan, light blue) from high (ice) clouds (pink/red) in the Day Cloud Phase Distinction RGB (Fig 4). High detail is maintained in the features with the inclusion of the 0.5 km VIS. The reflectance components can easily/quickly be adjusted in AWIPS (lower max) to draw out features during low light situations. The nearly stationary and semi-transparent low clouds across northwest Illinois, which could not be diagnosed in previous imagery, is easily observed in the RGB against the snowy background.

Similarly, the Day Snow-Fog RGB combines the Veggie Band, Snow/ice Band, and Fog Difference to separate snow cover from clouds (Fig 5).

The FLS Probability level-2 product, now available in AWIPS, combines satellite and NWP information to give a probability of MVFR, IFR, and LIFR conditions (and Cloud Thickness). During the day, the areal extent of coverage matches well with appearance in imagery, though with reduced spatial detail and some difficulty in multi-layer cloud scenes (Fig 6).

One method for incorporating the quantitative fog products into AWIPS workflow is to underlay the three fog probability products in your typical Cloud imagery display. In the example shown, sampling the area will not only show the RGB readout information, but also the probabilities of the various restrictions, for that location (Fig 7). This type of display allows one to view the imagery but still get the benefit of the quantitative information without taking away valuable screen real estate, while also helping to connect the quantitative product to the imagery.

Bill Line, NESDIS and CIRA

Previous blog posts and presentations have introduced the ability to diagnose plumes of blowing snow in GOES ABI imagery. In particular, the feature becomes apparent in the Day Snow-Fog RGB, primarily due to contrast of the plumes with surrounding environment in the Snow/Ice band and Fog Difference, and shadowing. Recent snowfall and strong winds resulted in a widespread blowing snow event across southern Saskatchewan/Manitoba into North Dakota on Jan 8. Surface observations and webcams reported considerable visibility reductions within the region of blowing snow as detected by ABI and VIIRS, including some areas to less than 1 mile. The blowing snow was pointed out by Carl Jones (NWS/FGF) on Twitter:

Blowing snow plumes 💨 💨 💨 finally showing up in North Dakota. Thick batch coming out of Saskatchewan. #ndwx #skwx pic.twitter.com/ItSY57Audv

— Carl Jones (@Wx_Jones) January 8, 2022

An experimental Blowing Snow RGB captured the blowing snow well across the region, as a shade of gold to dark yellow compared to the red snow-covered background and blue of clouds. Some plumes of blowing snow appear to develop into HCRs at times (as was documented in this paper), and have shadows associated with them in the imagery. Recall, this experimental RGB is similar to the Day Snow-Fog RGB available in AWIPS, but with the higher resolution 0.64 um band replacing the 0.87 um band, and with ranges tweaked to better highlight the blowing snow feature.

Compare the Blowing Snow RGB in Fig 1 with the Day-Snow-Fog RGB in Fig 2.

A similar experimental Blowing Snow RGB can be applied to VIIRS 375 m I band imagery, providing excellent spatial detail (Fig 3 and Fig 4). The individual plumes/HCRs of blowing snow can better be observed, as well as more subtle areas of blowing snow. Combining NOAA-20 and S-NPP, an animation of four VIIRS swaths can be created within a period of ~2.5 hours. Of course, ABI has the advantage in showing the longer evolution of blowing snow with time. Utilized together, one gains an ideal understanding of the feature and its evolution.

Bill Line, NESDIS and CIRA

A strengthening shortwave trough ejecting into the central US plains resulted in widespread damaging winds, wildfires, and visibility-reducing blowing dust on 15 Dec 2021.

NWS forecast offices leveraged satellite imagery throughout the event. In particular, water vapor imagery was used to track the progression of the mid-upper low, including areas of most pronounced subsidence (drying), where strong synoptic winds were most likely (Figure 1). Water vapor imagery also showed abundant gravity waves emanating from the center of circulation, helping to characterize the turbulent system. During the early afternoon from NWS CYS:

“Mesmerizing GOES WV loop for Windsday as a strong, progressive low continues to deepen and move off to the northeast into central KS/NE. Back edge of this system continues to move through the CWA with drying starting to become evident along the South Laramie Range aided by downsloping winds.”

and from NWS BOU slightly later:

“Winds have clamed down in most areas. However, satellite data shows enhanced subsidence in the nrn foothills. This is allowing for wind gusts in the 50 to 60 mph range in the normal windy areas.”

NWS Pueblo forecasters mentioned they used “all 3 water vapor channel’s to get a good feel of the intensity of the system.”

The system brought widespread wind gusts over 60 mph across the plains, including several reports over 100 mph. NWS Pueblo provides a great summary of the event. Some of the strongest wind gusts for southern Colorado occurred during the morning hours as the surface low deepened and strong subsidence developed across the area. GOES-East 1-min Geocolor imagery with NWS LSRs overlayed captures this period in Figure 2, just prior to the lofting of widespread blowing dust.

Over the next couple of hours, GOES-East 1-min Geocolor imagery captured widespread lofting of dust, particularly in the Pueblo area, and over the far eastern CO plains (Fig 3). Forecasters leveraged GOES-East satellite imagery to detect and track the blowing dust: “Satellite imagery shows more widespread blowing dust over the eastern plains with some roads closed due to winds and visibility. Issued dust storm warning for the plains east of the Interstate 25 corridor. The I25 corridor remains in a blowing dust advisory with some areas reporting visibilities below 1 mile.”

A day long Geocolor animation shows continuing evacuation of dust from southeast Colorado and surrounding areas, and transportation across the central plains (Fig 4).

In the infrared, the 10.3 minus 12.3 um split window difference (SWD), discussed numerous times on this blog, depicts the dense blowing dust plume clearly as areas of relatively dark gray to black (negative values; Fig 5). The SWD is available to forecasters in AWIPS with a different default color range and table. However, a grayscale color table such as what is shown can easily be employed.

The Dust RGB, which includes the SWD and is available in AWIPS, reveals the blowing dust as pinks/reds, along with information about clouds (Fig 6).

The Dust-Fire RGB, available in AWIPS on request, similarly includes both the SWD and IRW, but also adds the 3.9 um SWIR in order to reveal wildfire hot spots as well (Fig 7). The RGB is also able to differentiate dust from clouds. Given the strong gusty winds, numerous hot spots developed and spread across the plains, appearing as red pixels amongst the blowing dust (green). Note that the infrared dust products continue to capture dust after sunset, which is not possible with the vis/nir methods, such as Geocolor.

The DEBRA-Dust product, available online and in AWIPS upon request, was leveraged by both NWS PUB and BOU for dust detection during the event (Fig 8). This product also leverages the SWD, along with more advanced techniques, to highlight blowing dust. From NWS/PUB: “the dust -DEBRA (CIRA) product on Colorada State’s RAMMB/CIRA Slider page was very helpful to monitor how bad the dust was getting across the far eastern plains.” And NWS/BOU wrote in an AFD: “The DEBRA dust satellite product shows the extent of the dust across SE CO and all of W. KS. There is a fair amount of dust across the northeast corner of Colorado, but with mid level clouds we can’t see it on satellite but ASOS and trained spotters have been all over it. Will keep dust in the forecast through this evening across mainly the northeast corner.” DEBRA-Dust also works at night.

Another useful “windy day” display that can be made in AWIPS combines the Geocolor product with below zero SWD (yellow), SWIR hot spots (red), and GLM flash points (blue), resulting in a multipurpose display that captures blowing dust, wildfires, smoke, and thunderstorm details (Fig 9).

Finally, VIIRS providing even more detailed views of the blowing dust and wildfire hot spots in the early afternoon via the Natural Fire Color RGB (Fig 10). The exact locations of dust initiation, and the thickest plumes, can be diagnosed in the imagery, along with a detail of the wildfire hot spot location and size.

Bill Line, NESDIS and CIRA

Input from NWS/PUB forecasters, and NWS text products

A broad mid-upper level trough and associated shortwaves moving across the country resulted in numerous weather impacts, including but not limited to heavy snow, blowing dust, and severe storms. A detailed analysis of GOES water vapor imagery is provided by NWS/LUB, with associated imagery shown in Fig 1 and Fig 2.

“08Z analysis indicates a gradually amplifying, positively-tilted trough pivoting over the Great Basin with a shortwave trough embedded within the broader flow and rounding the base over southern Nevada, with the infant stages of a developing baroclinic leaf ahead of it as mid-level frontogenesis intensifies over the Intermountain West. A secondary, more-compact shortwave trough was analyzed over northern Sonora with a pair of vorticity lobes (induced via horizontal shearing instability due to the largely barotropic state of the atmosphere over the eastern two-thirds of the CONUS) rotating northeastward over the Texas Panhandle and the east-central Great Plains; all of which is seen on water vapor imagery.”

VIIRS Day Cloud Phase Distinction RGB imagery provided a high resolution (375 m) look at the snow cover (green) before (Dec 3) and after (Dec 11) the snowstorm over Colorado (Fig 3). The Colorado mountain snowpack had been well below normal prior to the storm.

Strong gusty winds (widespread over 35 knots, localized over 50 knots) developed across the southern Plains south of the surface low as the mid-upper level jet advanced over the region and deep vertical mixing set in during the afternoon. The relative position of the this jet core can be analyzed in water vapor imagery as strong cold to warm poleward temperature gradient extending southwest from the main shortwave. The 250 mb Jet as analyzed by the RAP NWP model aligns with that per the water vapor analysis (Fig 4).

The strong winds across dry antecedent lands resutled in widespread blowing dust across the southern plains. As has been discussed numerous times on this blog, the GOES 10.3 – 12.3 um Split Window Difference (SWD) captured lofted dust as near zero to slightly negative values compared to otherwise mostly positive values. The SWD in this case captures numerous large and long lasting dust plumes during the day, lasting into the early evening (Fig 5). Dust is not only diagnosed by the values, but also by the unique motion and streaky appearance of said values. Visibility was periodically reduced significantly within the blowing dust plumes.

Given the dry and windy conditions, rapid growth and spread of wildfires, should any develop, was also a concern. Several wildfires did indeed develop, including within the NWS/OUN forecast area. Forecasters at NWS/OUN were motioning GOES Imagery for wildfire hot spots, and blowing dust, writing:

“Several hot spots associated with wildfires have been observed on satellite so far this afternoon, as strong winds and low RH continue to lead to near-critical fire weather conditions. Clouds have been a bit more extensive than forecast across western portions of the area, with fires most prevalent so far within
clearing across central into southwest Oklahoma and western north Texas, on the western edge of the low-level thermal ridge. Expect fire risk to continue through the afternoon ahead of a cold front currently advancing into far northwestern Oklahoma….For this forecast, will mention MVFR restrictions in blowing dust at KLAW and KSPS ahead of the front closest to dust plume evident on satellite.”

Since wildfires and blowing dust often occur together, a Dust-Fire RGB was developed to highlight both in a single image, in addition to masking out clouds. The Dust-Fire RGB is shown over parts of OK and TX during the afternoon of the 10th (Fig 6). Blowing dust appears as relatively bright green, clouds generally as blue, and wildfire hot spots as red. The cold front is also diagnosed pushing south through the Panhandles and N OK by the end of the animation.

One-minute Imagery from GOES-East was available to forecasters to monitor the wildfire threat. Geocolor imagery also captured the blowing dust (and smoke plumes) well during the day, and can be combined with SWIR imagery to also highlight wildfire hot spots (Fig 7).

Natural Color Fire RGB imagery from VIIRS provides a detailed view of a wildfire, its smoke plume, and a large dust plume in North Texas near the S OK border (Fig 8).

Further east across the southeast, severe thunderstorms developed as the broader system brought anomalously moist low level air into the region along with favorable wind shear. GOES-East 1-min imagery was also available over this region to aid forecasters in tracking thunderstorm development and evolution. The thunderstorm associated with the Quad State tornado (moving from southwest to northeast across domain) exhibited a consistent and strong updraft per the cold OT, periodic AACP, and persistently high GLM flash rate (Fig 9).

Bill Line, NESDIS and CIRA

Gusty winds, low relative humidity, and dry fuels resulted in high fire danger across parts of Colorado in mid-November, 2021. These conditions allowed for the growth of the Kruger Rock Fire near Estes Park, CO on 16 Nov 2021.

GOES-East shortwave IR imagery, however, did not detect the wildfire hot spot during the morning and afternoon, despite the fire quickly growing to over 75 acres. However, when one viewed the GOES-West shortwave IR imagery, the hot spot was easily diagnosed from initiation through the afternoon. An animation comparing SWIR imagery from GOES-East and GOES-West is shown in Figure 1.

The explanation for this appears to be a couple of factors. The most significant factor was a persistent and optically thick mountain wave cloud, oriented above a point just east of the fire, blocking the heat signal to GOES-East, but not to GOES-West. Note that at 105W, this location is in the middle (longitude) of the two GOES satellite (137.2W and 75.2W), meaning pixel sizes are similar. Another 2-panel GOES-East/West comparison is shown in Figure 2, but this time comparing Natural Fire Color RGB Imagery, which shows the hot spot and cloud cover in the same image. GOES-West has an unobstructed view of the fire, while a cloud appears to mask the view from GOES-East. A similar case of a mountain wave cloud blocking a wildfire hot spot signature occured last eyar with the Cameron Peak Fire.

Another factor that may have weakened the SWIR hot spot signal from GOES-East compared to GOES-West was that the wildfire was burning on a west facing slope. Therefore, while the GOES-West satellite had a direct and unobstructed view of the heat signature, the GOES-East satellite view of the heat signature may have been partially degraded due to terrain blocking. Figure 3 shows the satellite comparison again, but a little later in the afternoon as the primary wave cloud shifted to the east. The hot spot signature becomes apparent in the GOES-East imagery, but not to the degree that it was earlier from GOES-West. Note, during the same time period, additional clouds developing from the west begin to obstruct the view of the hot spot from GOES-West. Not a clear comparison, unfortunately, so the influence of terrain is not straightforward with this case.

VIIRS passes during the early afternoon provided yet another view that captured the wildfire hot spot signal during the early afternoon. The VIIRS SWIR imagery is available at 375 m resolution from the I bands (compared to 2 km at nadir from ABI), and is a component of the Natural Fire Color RGB (Fig 4). Parallax causing shifting of clouds also plays a role here, from scan to scan, depending on where the viewing target lies within the VIIRS swath. Terrain correction allows the hot spot to remain at the same location from scan to scan. The ~1850 UTC scan captured the fire heat signature beneath the cloud, albeit under what appeared to be a relatively transparent one. Subsequent scans detected the hot spot from the clear sky.

The example underscores the importance of forecasters viewing imagery from both GOES-East and GOES-West when possible in such situations. Timely VIIRS imagery can also be useful in determining the exact location of the hot spot consistently from scan to scan and in more detail, and when viewing among clouds and on terrain.

Bill Line, NESDIS and CIRA

GOES Imagery has long been a valuable tool for fog detection and tracking, both during the day and night. With GOES ABI, forecasters are able analyse low clouds and fog in even better detail, spatial and temporally. Further, the additional spectral bands allow for more insight into the features through the generation of multispectral products such as RGBs.

NWS Bismark, ND forecasters leveraged GOES-East ABI Imagery in tandem with area webcams and surface obs to analyze fog evolution prior to and after sunrise on 09 Nov 2021, informing their forecast decision-making. While GOES imagery provides a great look at the spatial extent of low clouds and fog and their evolution with time, the webcams and surface obs allow the user to determine actual surface conditions associated with the satellite cloud feature. At 1213 UTC, the forecaster issued an update to the Area Forecast Discussion that read: “GOES Nighttime Microphysics shows extent of fog in southwest North Dakota, possibly nudging into the south central this morning. Based on cameras and coverage depicted on satellite imagery, coverage is somewhat patchy, but has lowered visibility at Hettinger down to 1/4 mile. An SPS was issued to address fog this morning for this area. Fog will gradually dissipate mid-morning as boundary layer mixing increases after sunrise.”

The Nighttime Microphysics RGB imagery that goes with the above discussion captures the low clouds and fog as a light orange/light brown/yellow color, compared to the pink clear-sky land background, blue/magenta bodies of water, and red (thick) to black (thin) high clouds (Figure 1). The imagery shows the development and northward surge of low clouds and fog from beneath exiting upper level cloud cover over a 5-hour period ending just prior to sunrise. The webcams (and METARs) confirmed conditions beneath the suspected low clouds, per the discussion.

Forecasters continued to analyze satellite imagery and webcams after sunrise to inform their decision-making, as the low clouds and fog hung around. At 1505 UTC, Forecasters wrote: “Latest satellite and webcams indicate fog over the southwest and south central is beginning to dissipate. We did extend the areas of fog farther east and extend the SPS for fog through 16 UTC. Otherwise no changes to the current forecast.” During the daytime, multispectral products such as the Day Cloud Phase Distinction (Fig 2) and Day Snow Fog (Fig 3) RGBs may be utilized, in addition to visible imagery (Fig 4). In the three hour period following sunrise, these animations show the fog and low clouds dissipating through the morning hours, with surface conditions confirmed by webcams and METARs.

Bill Line, NESDIS and CIRA

A series of potent mid-latitude cyclones brought an active period of weather to much of the US during the end of October. GOES Water Vapor imagery with GLM overlay captures the large scale features influencing the active period from the morning of Oct 24 through the morning of Oct 28 (Fig 1). The animation begins with the much discussed atmospheric river pushing across the west coast, characterized by relatively cool BTs in the water vapor imagery due to the presence of abundant atmospheric moisture and cloudiness.. The associated shortwave makes landfall on the US coast thereafter and advances across the southwest US, as drying/descending air coupled with cooling ascending air is diagnosed in the water vapor imagery with the amplifying trough. During the second half of the animation, this system continues east across the southern plains into the southeast, resulting in several rounds of severe weather. At the same time, another shortwave trough exits the midwest into the northeast, forcing the development of severe weather along its path. This system merges with another shortwave moving northeast along the east coast, developing into a strong nor’easter. The pulses of severe weather are apparent as increase in cloud cover (very cold BTs) and development of lightning (yellow-red overlay).

A GOES water vapor imagery analysis, typical of NWS forecasters, was presented by NWS AMA in their AM 26 Oct AFD ahead of expected severe weather forced by the approaching western US system: “The latest water vapor imagery reveals very dry subsiding air moving along the base of an upper level trough near the AZ/CA border. Lift associated with this system is now evident in east central AZ up across the Rocky Mountains. As this system advances east, strong jet level winds will move into the Panhandles, with some stacked jet alignment across the western zones this afternoon.”

A great water vapor analysis was presented by NWS ALY for the East Coast events during the morning of 26 Oct: “GOES 16 water vapor imagery shows the 3 upper-level features that will come together over the next several hours to produce a rapidly deepening coastal low. There is a southern stream shortwave off the East Coast, another shortwave currently over Virginia and North Carolina, and a remnant upper low over Michigan. What was initially the primary surface low weakened earlier tonight over Pennsylvania, and the new coastal low has developed in response to the southern stream disturbance off the Coast of the Carolinas and has already intensified to 995 mb. With this classic Miller Type B storm track, this coastal low is expected to continue to intensify as it moves northeastward during the day today.”

Focusing on 26 Oct, the western shortwave trough brought blowing dust and severe weather to the southern US plains. The split window difference (with IRW overlay for cold BTs), captured several plumes of blowing dust (low to neg SWD, or dark gray to black in this scale), including in southeast AZ, Chihuahuan Desert, and west Texas all within the strong/dry southwest flow (Fig 2). The SWD also reveals the presence of the dryline extending north across west TX through west Kansas (sudden transition from dark grey (dry) to light gray (moist), west to east). Convection initiates along this boundary toward the end of the animation.

Related to the blowing Dust, NWS EPZ wrote in their PM 26 Oct AFD: “Blowing dust still a concern, not many plumes visible from satellite though a few obs sites are reporting visby reductions. Willcox Playa [in southeast AZ] looks like it is starting to shed blowing dust off to the east across the Lordsburg area.”

The GOES TPW product also captures the deeper moisture gradient present along the N-S oriented dryline, along which convection subsequently develops (Fig 3).

Focusing on convective initiation along the dryline over the W OK/TX panhandle border, a 1-min GOES-East VIS/DCPD RGB side-by-side animation shows the value of the RGB (Fig 4). By 2222 UTC, the DCPD RGB begins to show glaciation within the cu field (transition from cyan to green to yellow), signification initiation is imminent. Meanwhile, in the VIS, one can diagnose vertical cloud growth and clumping of cu, but cannot classify glaciation. Note, the recipe for the DCPD RGB was modified slightly in this case, lowering the upper end of the VIS and Snow/Ice band components to account for the lower sun angle and resulting lower reflectance values.

Convection would go on to initiate, with the first lightning detected by GLM at 2248 UTC, and first convective warning at 2321 UTC (Fig 5).

An IR animation shows the evolution of convection, including severe storms, across Oklahoma through the evening (Fig 6). The GLM Flash points overlay (as opposed to the gridded overlay) allows one to view the IR image and GLM data within a single image without blocking the important cloud top information of the IR. The rapid dropoff of lightning activity toward the end of the animation is coincident with cloud top warming and weakening of the convective line.

The VIIRS DNB/NCC product captured active convection further south slightly after the animation above (Fig 7a and 7b). The two images provide a more detailed view of convection (visible imagery at night), and even captures a lightning flash, in the second image, associated with convection developing ahead of the main line in southern Texas.

Another line of severe storms developed across the southeast US on the 27th. GOES-East 1-min VIS and GLM flash points captured individual convective updrafts and potentially stronger storms embedded within the broader line of convection (Fig 8).

Bill Line, NESDIS and CIRA

The period of 09/27/21 to 10/02/21 was very stormy stretching from the western Bering Sea to the British Columbia and Alaska panhandle regions. Three storm-force lows moved generally west to east with a final storm achieving hurricane force before making landfall just north of Sitka, Alaska. The AirMass RGB imagery from Himawari and GOES-17 along with scatterometer (ASCAT) data available from the Metop-A/B/C satellites provides forecasters at the NWS Ocean Prediction Center with some powerful tools to monitor and predict this series of high impact storms. This blog post will show some examples of imagery, surface maps, and ASCAT data used during this time.

A picture or in this case, an animation is worth a million words. There is so much happening in this animation above that it’s hard to explain any one feature without your eyes being drawn to another feature. There are two storm force lows, a couple gales, and a recurving typhoon. Can you tell which is which? I will break down what is happening here in the following images.

Now with the NWS/OPC surface analyses overlaid, you can tell where the surface low pressure systems, high pressure systems, and recurving typhoon (lower left) are located throughout the animation. For our first storm-force low of interest, I’ll show some AirMass RGB and ASCAT imagery along with surface maps to show the portion of the storm that is the strongest. For the AirMass RGB imagery, the red colors indicate descending dry air that is rich in ozone and potential vorticity. This means the brighter the red coloring, the stronger the warming (drying) in the 350 mb to 700 mb level and the more ozone and potential vorticity available to the storm system. The green coloring is higher moisture and typically, a warmer mid-troposphere, while the bluish, purple coloring is higher moisture, but cooler mid-troposphere.

As you see in the above images, the OPC surface map at 0000 UTC on 09/30/21 was very active with two storm-force lows, a developing storm force in the eastern Pacific, a couple gales, and a recurving typhoon near Japan. The strongest storm force is the one in the northern Bering Sea, west of Alaska with winds of 55-60 kt (65-70 mph) near the Siberian coast (white circle in ASCAT image) as a barrier jet forms as the air quickly converges into a meso- to micro-scale jet due to the strong storm circulation and terrain near the coast. This storm would fill and drop southeast rather fast (>30 kt/35 mph) and begin to merge with the developing gale (near the forecast 998 mb label) south of the Aleutians.

The two animations above show the next phase of this multi-day evolution ending with the strong hurricane-force low that moves towards the Alaska panhandle. Note how fast features are moving south of the Aleutian Islands into the Gulf of Alaska thanks to a strong upper-level jet. I’ll break down this series of storms below.

In this next set of images, another storm-force low is developing as it approaches Queen Charlotte Island, British Columbia, Canada at 0000 UTC on 09/30/21. At this time the storm is producing winds to 45 kt (50 mph), but strengthening rapidly. By 0600 UTC on 09/30/21, the storm achieves storm-force verified by excellent ASCAT coverage with winds 50-55 kt (60-65 mph) nearing the southern tip of Queen Charlotte Island and the Hecate Strait. The storm makes landfall the following hour and later dissipates over the mountains of British Columbia.

Although I won’t cover this next storm-force low beyond the ASCAT image above, winds did approach 45-50 kt (50-55 mph) near southwest Kodiak Island (white circle) which just meets the storm-force warning criteria. This low would eventually provide additional vorticity advection for the hurricane-force low farther south.

The above sequence of GOES-17 AirMass RGB imagery shows the evolution of the hurricane-force low that develops along an active baroclinic zone. These types of storm evolutions are called Shapiro-Keyser cyclogenesis and can lead to some of the strongest (wind perspective) extratropical cyclones.

Thanks to some very timely ASCAT-C overpasses, the hurricane-force low was verified with winds peaking around 75 kt (85 mph), which is very intense for a non-tropical cyclone.

The wind/wave analysis above shows significant wave heights associated with the hurricane-force approaching 41 ft! It is possible individual waves were significantly higher.

Although not available in real-time for forecasters, the above SAR image shows some incredible detail in the winds field associated with the hurricane-force low as it approached Sitka, Alaska during the early morning hours of 10/02/21. Note the strong wind gradient to the northwest of Queen Charlotte Island which depicts the powerful cold front. This was associated with a squall line that produced some rare lightning strikes and wind gusts of around 95 kt (109 mph) at the Ketchikan Airport in the southern Alaska panhandle. Please see the NWS Juneau twitter feed for more information. Meanwhile, multiple wind gusts of 70-90 mph were reported along the coast, while the hurricane-force sustained winds slowly diminished as the storm filled and approached land.

This opening act to the fall/winter baroclinic season in the North Pacific was very impressive and with a Sea Surface Temperature Anomaly map that looks this warm and a strong temperature anomaly gradient across the North Pacific, I expect we will be seeing a plethora of storm and hurricane-force lows in the coming weeks. In the meantime, it has quieted down, though I assume it’s very temporary.

Thank you for reading!

The active wildfire season over the western US this summer has resulted in periodic days of smokey skies downstream. Forecasters are concerned with the detection of smoke since it has impacts on local air quality and visibility, and is the source of public inquiries. During the daytime, smoke is easily diagnosed by forecasters using GOES and VIIRS visible, near-IR and multispectral products, such as Geocolor. At night and in the absence of these channels/products, smoke is much more difficult to discern using IR-only-based products save for the most dense of smoke plumes. For example, ABI IRW imagery from the evening of 19 Sep 2021 is shown in Fig 1. Signs of smoke are not readily apparent.

A nighttime smoke detection option available to forecasters in AWIPS is the VIIRS Day Night Band (DNB) Near Constant Contrast (NCC) product. Available from both SNPP and NOAA-20, this product is available 2-4 times per night over a given location across the CONUS due to overlapping swaths. During the early morning, pre-dawn hours of 19 Sep 2021, forecasters at NWS Bismark, ND were analyzing the NCC imagery. They noted in surface obs slightly reduced visibility at locations within the CWA (8 SM in southwest ND), and highlighted in an AFD update: “VIIRS Near Constant Contrast satellite product at 08Z shows areas of smoke across eastern Wyoming and Montana, pushing into western North Dakota.” Figure 2 captures the 08Z NCC product at hand.

An animation of the three overnight VIIRS passes captures the evolution of smoke across the region (Fig 3). The smoke becomes less apparent later in the evening with shifting viewing angle, but can still be tracked with careful analysis. During the 1.5 hour period, the smoke is observed shifting north and east across the BIS CWA.

BIS went on to mention: “HRRR Smoke guidance indicates smoke moving eastward alongside the thermal ridge.” The VIIRS NCC provides a check on HRRR smoke model analyses and forecasts overnight (Fig 4). Combined, the forecaster could make a assessment on where smoke was at present, and where its influence would be during the day.

This case provides a great example of a forecaster leveraging the VIIRS NCC product to diagnose smoke coverage across the region overnight, confirm the likely source of reduced visibility, and provide a check on model smoke output.

Bill Line, NESDIS and CIRA