Hurricane Ian made landfall along the Florida Gulf Coast near Cayo Costa at 1905 UTC UTC on 28 August 2022 as a strong category 4 Hurricane, bringing with it intense winds, storm surge, and rainfall amounts. NOAA GOES and JPSS satellites provided imagery and products that weren’t only visually pleasing, but were critical to operations in tracking the storm.

In the days leading up to, and including, landfall, the National Hurricane Center continuously requested 30-second mesoscale sector imagery from GOES-East (GOES-16). The imagery captured the storm in tremendous detail, helping forecasters to pinpoint the center of circulation in real-time, track thunderstorm activity both in the eyewall and in the tornado-producing outer rain bands, and help to determine when landfall occurs. The high temporal resolution imagery was also leveraged in communicating the storm via social media and other media outlets.

Starting with the last sequence of images prior to landfall, GOES-East 30-second-updating 500-meter-spatial-resolution Channel 02 Visible imagery provided the best detail of Hurricane Ian. Focusing on the center of circulation, or eye, one can diagnose a fairly large eye, the presence of meso-vorticies within the eye, and thunderstorm activity within the eyewall (Fig 1).

Now stepping back a few days, GOES-East 1-min VIS/IR sandwich imagery from the morning of the 26th captured active convection in great detail, as was noted in the NHC 11 AM forecast discussion (Fig 2): “The satellite presentation of Ian has improved this morning. Deep convection has increased within the inner core during the past several hours, with an expanding central dense overcast noted in recent satellite imagery.” The sandwich imagery combines the benefits of great spatial detail from the VIS with quantitative brightness temperature information from the IR.

GOES-East began collecting 30-second imagery later in the day on the 26th. The imagery toward sunset captures the center of circulation and associated convection, but no clear eye yet (Fig 3). The 5 PM update from NHC provides a basic view into how the satellite imagery was being leveraged: “The 18 UTC satellite classifications from SAB and TAFB were a consensus T4.5/77 kt, but the continued improvement in satellite structure warrants raising the initial intensity to 85 kt for this advisory.”

The NWS Weather Prediction Center routinely leveraged GOES Satellite imagery while monitoring Ian, including during the preparation of their Mesoscale Precipitation Discussions (MPDs). For example, from early on the 27th discussing thunderstorms over Florida: “These storms were tied to distant outer rainbands to the north of Hurricane Ian (Ian’s center located just south of western Cuba at 06Z) within a broad region of divergence aloft / northern outflow channel noted on infrared satellite imagery. (Fig 3.2)”

During the evening of the 26th into the 27th, Ian advanced across western Cuba. By the morning of the 27th, Ian emerged off the coast of Cuba, with NHC noting at 11 AM: “The well-defined eye of Ian emerged off the coast of western Cuba about an hour ago.” GOES-East 30-second visible imagery reveals this evolution (Fig. 4).

By 5 PM on the 27th, the storm continued to strengthen, and NHC states: “The eye of Ian remains well-defined on visible imagery, although radar data from Key West suggest that an eyewall replacement could be in the initial stages.” GOES-East 30-second Geocolor imagery zoomed out demonstrates the massive size of Ian (Fig. 5). NHC forecast track, cone, and intensity can be loaded into AWIPS and displayed with imagery in order to provide additional context.

A zoomed out full-day GOES-East 2-min VIS/IR sandwich animations shows the evolution of the storm on the 27th, including the large and open eye, and outer rain bands and convection (Fig 6). The Storm Prediction Center had a Slight Risk for severe thunderstorms (driven by the tornado threat) across all of south Florida. Numerous Tornado and Marine Warnings were issued by local NWS offices for supercells associated with Ian (plotted in the below animation), and 11 tornadoes were reported.

By sunrise on the 28th, Hurricane Ian was making it’s final approach toward the Florida Gulf Coast. Early GOES-East 30-second visible imagery captured a very large eye, initially obscured by cirrus but clearing (Figure 7).

Refer back to Figure 1 for 30-second imagery capturing the landfall of Hurricane Ian during the early afternoon of the 28th.

After landfall, NHC noted in the 5 PM discussion: “While there hasn’t been much in situ data recently, satellite images show that the eye has become more cloud filled, and Tampa Doppler radar data is indicating a gradual reduction in winds.” A full day, 1-min-updating visible animation of Hurricane Ian’s evolution on the 28th is shown in Fig 7.5. Note the filling of the eye after landfall as was noted in the NHC discussion.

A zoomed out, full-day 2-min VIS/IR Sandwich imagery animation from the 28th from GOES-East is shown in Fig 7.3.

A longer-term animation of GOES-East Ch-13 IR animation shows the organization of Hurricane Ian within the GOES-East CONUS sector from the morning of the 25th in the Caribbean when Ian was a Tropical Storm, to around landfall during the afternoon of the 28th when Ian was a Major Hurricane (Figure 8).

An animation covering the time period from early on the 27th through landfall on the 28th shows the total lightning activity, captured from the GOES-East GLM (Figure 9). Throughout the progression through the Gulf, there were numerous bursts of lightning associated with eyewall convection and with thunderstorms in the outer bands.

GOES-18, currently in post-launch testing, also collected 30-second imagery of Hurricane Ian as it made it’s approach through the Gulf of Mexico into Florida (Figure 10). From the GOES-West position, the imagery was captured at a steep viewing angle (far from the satellite subpoint), providing a unique perspective of the storm. GOES-18 ABI Imagery during this period was preliminary and non-operational.

A zoomed out animation of GOES-18 VIS/IR Sandwich Imagery captures the storm in the 5 hours leading up to landfall (Fig 10.2).

VIIRS imagery from the NOAA-20 and S-NPP satellites provided very high resolution (375-m) imagery of Hurricane Ian as the storm was making landfall on the 28th. The IR imagery measured cloud top brightness temperatures colder than -73C, and the high spatial resolution IR and VIS and overhead (near nadir) look provided details about the shape and size of the eye, and thunderstorms within the eyewall and outer bands (Fig 11). The VIIRS Natural Color RGB from NOAA-20 provided additional details about the clouds (ice vs water) and surrounding clear sky areas by combining multiple channels into a single multispectral imagery product (Fig 12).

By the late morning/early afternoon of the 29th, Ian had traversed across the Florida Peninsula and progressed back over water in the Atlantic. With land and increasing shear taking its toll on the storm, the low-level circulation was now exposed. Day Cloud Phase Distinction RGB imagery combines the Ch13 IR, Ch05 NIR, and Ch02 VIS into a single image in order to provide information about cloud top phase. With tropical cyclones such as Ian at this point, the imagery helps to differentiate ice/high clouds (bright red/green/yellow) vs water/low clouds (bright blue/cyan), which itself makes it easier to identify the low-level circulation, thunderstorms, and low stratus clouds (Fig 13).

Bill Line, NESDIS/STAR

Mid-September brought a period of active Fire Weather, resulting in several large wildfires and widespread smoke, to Pacific Northwest US. GOES and VIIRS Imagery provided stunning and operationally useful information about the environment and fires and smoke.

A GOES-West animation captures the ignition (or re-ignition) of wildfire hot spots coinciding with the development of dry air across the northwest US (Fig 1). The animation includes GOES-West Water Vapor Imagery (gray) with the derived TPW as an overlay (colors) and derived product wildfire hot spots (yellow pixels). As the dry air expands from the 7th to the 10th, numerous hot spots, including several large ones, appear. From the 10th to the 12th, moisture increases from the south and the west in association with a tropical cyclone (Kay) and a mid-altitude cyclone, respectively, helping to calm the fire activity.

Zooming out and focusing just on only GOES-West water vapor imagery, in the early part of the loop, we diagnose fairly strong northwest flow in the upper levels along with periodic disturbances across the Pacific Northwest (Fig 2). During the second half of the loop, increasing moisture and cloudiness are observed to develop across the region in associated with Kay from the south and the mid-latitude storm from the west.

Three of the significant wildfires during the period will be discussed below, using a GOES-West “Geocolor-Fire” procedure in AWIPS. The imagery display used to depict each wildfire includes a base of Geocolor in order to highlight the wildfire smoke plume and surrounding clouds and vegetation, with an overlay of ch07, ch06, and ch05, in order to show details of the wildfire hot spot. The swir/nir overlay are included as varying degrees of transparency and colors with final colors ranging from red to yellow to white for “lesser” to “greater” wildfire intensities (similar to with the Fire Temperature RGB). While this procedure is great for communicating/depicting a given wildfire and tracking smoke coverage, it should not be used for new fire hotspot detection since relatively “cool” hot spots will be missed. It is recommended to stick to basic ch07 single band imagery for that task.

On the 8th, the Mosquito Fire located between Sacramento and Lake Tahoe experienced a period of rapid growth and smoke output. The fire, which began on Sep 6, grew from under 7000 acres on the morning of Sep 8 to 29,000 acres in the afternoon of the 9th. The fire had grown to 46,000 acres by Sep 12. The Geocolor-Fire procedure captures the large and intense heat signature, along with the dense smoke plume as it races east over lake Tahoe (Fig 3). Additionally, periods of pyrocumulus clouds are observed over the fire. One-minute temporal resolution allows one to observe the hot spot and smoke plume evolve in ~real-time and with smooth animations, as opposed to waiting for relatively choppy animations with longer timesteps.

Forecasters and satellite analysts are aware that smoke is particularly difficult to detect at night in satellite imagery given the absence of reflectance channels, and weak to no signal in infrared channels. Therefore, although currently only available over any given location of the CONUS 2-4 times per evening within a timespan of 55 min – 3.5 hours, VIIRS Day Night Band Near Constant Contrast Imagery is a valuable operational tool for detecting and tracking wildfire smoke at night.

For the Mosquito Fire, overnight, smoke is observed maintaining a presence from Modesto north to Sacramento, and east over Lake Tahoe and Reno. This knowledge can be useful for aviation forecasts, temperature forecasts, and general communication to the public. The light emitted from the wildfire itself can also be observed in the imagery, providing more information on which parts of the fire are more or less active. In this case, the wildfire is located just left of center in the image, just south of I-80, and appears as a “v-shape”.

On Sep 9, the Cedar Creek Fire in Oregon, which began on Aug 1, grew from 33,100 acres in the morning to 51,800 acres the morning of the 10th. The fire had grown to 86,000 acres by Sep 11 PM. Geocolor-Fire imagery shows an intense hot spot and smoke plume already by sunrise, with smoke streaming over Eugene, OR all day. (Fig 5)

Overnight VIIRS NCC product shows the densest portion of the smoke plume shifting north north of Eugene and toward Portland. The wildfire is the bright area in the lower right part of the image, and is burning strongest on the western front, with continued slightly weaker burning to the east.

Finally, The Bolt Creek fire east of Seattle initiated early in the morning on Sep 10, and grew to 7,600 acres by the early afternoon of the 11th. Geocolor-Fire Imagery captures some of this rapid growth during the day of the 10th, along with a growing smoke plume as it drifted west over northern parts of Seattle (Fig 7). Occasionally “bursts” in updraft activity over the fire sent smoke slightly higher in the atmosphere and subsequently to the east as it tapped into the westerly flow aloft.

The VIIRS NCC during the following evening captures widespread smoke across the area along with increasing cloud cover. Focusing in on the wildfire located in the center of the image, a relatively dense plume of smoke is diagnosed streaming north over the northern Seattle suburbs. A relatively bright ring of light marks the outer perimeter of much of the fire.

Bill Line, NESDIS/STAR

GOES-West Operational Interleave ended today, Sep 8, at 1601 UTC (began on 8/1). This means that GOES-West Imagery in AWIPS switched from GOES-18 back to GOES-17. All other GOES-West products remain from GOES-17 as they have been. NWS AWIPS users will likely not notice any difference, as we have come out of the GOES-17 ABI Warm Period, which resulted in degraded GOES-17 LWIR imagery due to the Loop Heat Pipe (LHP) cooling system issue. GOES-18 post launch checkout will continue, and the next Interleave period will take place during the next warm period, in the 10/13 – 11/16 period. GOES-18 is expected to become the operational GOES-West Satellite in early January 2023.

See ch08 water vapor imagery in Fig 1 for an AWIPS example of the transition from G18 to G17 imagery. One may notice that ,in this band, the imagery becomes slightly less “crisp” after the transition back to G17 due to the now minimal impacts of the G17 ABI LHP issue as we come out of the warm period.

Coming out of Interleave, the GOES-West satellites captured significant growth of the Mosquito Fire located in California between Sacramento and Reno. A 6-hour-long blended GOES-17 1-min Geocolor+SWIR+NIR Imagery product revealed the development of a large and hot heat signature, in addition to an impressive smoke plume with persistent pyrocumulus clouds (Figure 2).

Bill Line, NESDIS/STAR

Through the first week of September 2022, a strong upper Ridge developed across the western US and large wildfires burned across the Pacific Northwest. Increased flow and embedded disturbances across the northern part of the ridge helped to accelerate significant wildfire smoke east across the northern Rockies into the Great Plains and midwest (Fig 1). NWS offices were monitoring the evolution of the smoke, as it was dense enough to reduce visibility, impact temperatures, and reduce Air Quality and threaten the health of certain groups. Satellite imagery is the primary tool forecasters utilize to observe and track the evolution of wildfire smoke. The imagery is supplemented by other observational tools such as ASOS, webcams, air quality sensors, etc in order to better quantify impacts of the smoke. Finally, the HRRR-Smoke model, in which satellite data are vital components, is leveraged to make predictions of where the smoke will evolve over the next few hours to couple of days.

While wildfire smoke can be analyzed in single-band visible satellite imagery, animations of RGB imagery, and specifically True Color Imagery such as in daytime Geocolor, provide the best method for diagnosing and tracking wildfire smoke during the day. Geocolor may not yet be available in all NWS offices, but can be accessed online at the CIRA Slider webpage and GOES Image Viewer. As has been mentioned in past blog posts and is illustrated below, forward scattering plays an important role in the appearance of aerosols, such as smoke, in satellite imagery. In Geocolor imagery from the 6th, Smoke is much more apparent in GOES-West (-East ) imagery during the morning (evening; Fig 2 and 3). Hence, it is important forecasters have access to, and leverage, imagery from both satellites.

A perusal of central US NWS Area Forecast Discussions from the past couple of days indicate widespread discussion of the smoke given its impacts. A few of the discussions specifically mentioning the use of satellite imagery or HRRR Smoke are included below.

First on the 4th, NWS Bismarck, ND analyzed the smoke in Natural Color RGB imagery, available in AWIPS, and its potential impacts: “Natural Color RGB satellite showing nicely wildfire smoke being pulled into the upper low across Montana. Look for this to enter into western North Dakota by this evening.” Note in Figure 4, the smoke becomes most apparent later in the afternoon as it advances east across Montana into North Dakota, from the GOES-East position as the sun sets to the west.

NWS BIS also leveraged satellite imagery on social media to help message smoke impacts.

Later in the evening of the 4th, NWS Aberdeen, SD discussed the earlier appearance of smoke in satellite imagery, as well as the HRRR smoke predicted evolution of smoke for the next two days: “Satellite imagery verifies that there is smoke building into the northern plains region from western CONUS forest fires. The latest available RAP/HRRR smoke model output suggests there is some potential for lofted smoke to be thick/dense enough to affect high temperatures today mainly across the Missouri River valley and west river counties in this CWA. Overspreading more of the CWA overnight and on Tuesday, a larger area of the CWA could see some of that temperature-impacting smoke overhead.”

A HRRR Smoke “Vertically Integrated Smoke” 48 hr forecast animation is shown in Figure 5, predicting smoke to round the northern and eastern periphery of the ridge and fill in across the central US through the evening of the 6th. The predicted location of smoke in this forecast matches well with that in the animations of Fig’s 2 and 3.

On the morning of the 6th, NWS Boulder discussed the state of the smoke as depicted in satellite imagery and the HRRR smoke model, and expected influences on weather for the day: “The huge fire in Idaho is also generating a lot of smoke that is coming south along the periphery of the anticyclone, per the latest HRRR Smoke and GOES-16 imagery. The smoke will not be particularly thick, but should be enough to keep temperatures below 100 across all but the northeast corner of Colorado.” Refer to Figures 2 and 3.

NWS Boulder created a neat social media graphic on the 6th highlighting the satellite depiction of the wildfire hot spots and smoke plumes.

During the afternoon of the 6th, NWS OAX discussed the smoke per Geocolor Imagery and it’s anticipated evolution per HRRR Smoke model: “There is some smoke aloft across the mid CONUS as noted on Geocolor satellite imagery, and the HRRR smoke model indicates this smoke aloft could become more prevalent over our area through Wednesday night. Really not expecting any surface smoke in our coverage area, just most of a hazy sky appearance.” Refer to Figures 2 and 3.

NWS Omaha shared a Geocolor Graphic on Twitter pointing out the location of smoke across the region.

By the morning of the 7th, widespread smoke was present across the middle of the United States, including practically thick smoke over the northern US Rockies and adjacent high plains (Fig 6). NWS offices were already noticing an impact on temperature recovery during the morning, with BIS mentioning: “Temperatures have been slow to warm due to smoke moving in from fires out west. Therefore, high temperatures have been decreased a few degrees and may have to decrease them more with the next
package” and Rapid City writing: “Visible satellite imagery this aftn shows thick mid- to upper-level
smoke across wrn SD drifting ewd with time. This smoke limited incoming solar radiation and associated heating thru the morning hours.”

Other excellent social media posts that messaged the wildfire smoke via satellite imagery and HRRR smoke graphics, along with annotation, include:

• Dodge City leveraging VIS, Fire Temperature RGB, and HRRR Smoke.
• Albuquerque sharing Geocolor from the STAR GOES Image Viewer.
• Goodland using Natural Color imagery.
• Pueblo with HRER Smoke output.
• Des Moines shared a video annotating Geocolor Imagery in CIRA Slider.

Bill Line, NESDIS/STAR