The GOES-17 (now GOES-West) ABI has a cooling system issue that results in degraded image quality during certain times of the day. This anomaly was discovered last summer during post-launch testing, and steps have been taken since then (and continue) to optimize ABI performance.
Additionally, CIMSS is working on data fusion techniques to further mitigate the effects (read here and here).

Detailed information about the cooling system issue and how/when it affects GOES-17 imagery can be found here. To summarize:

• Bands impacted: IR bands 8-12 (3 water vapor, cloud top phase, ozone) and bands 15-16 (Dirty and CO2 IR). This issue also affects the RGB’s, derived products, and channel differences that use the impacted channels.
• Bands not impacted: VIS, Near-IR, shortwave IR, and IR window bands
• Time of year of impact: before and after the vernal and autumnal equinoxes.
• Time of day of impact: morning hours, ~0900 UTC – 1700 UTC. Greatest impact around 1300 UTC.

Presently (mid Feb 2019), the degradation of the impacted GOES-17 bands has been apparent, and will continue to worsen through the end of February. Quality will improve in March, before worsening again in April.

GOES-17 low-level (band 10) water vapor imagery from today, 15 Feb (0815 UTC – 1900 UTC) shows the imagery quality degrade, become unusable, and then improve during the overnight/morning hours (Fig 1). In this animation, there is about 4.5 hours (1115 UTC – 1545 UTC) when the imagery is unusable, and about 2 hours each before and after complete loss when the quality is degraded, but still usable. Band 10 (along with Ozone Band 12 and CO2 band 16) experiences the most degradation in image quality as a result of the cooling issue.

Corresponding imagery from IR Window band 13 appears normal, as expected (Fig 2).

Corresponding imagery from the Airmass RGB, which uses three of the channels impacted including band 10, of course shows degradation and loss similar to that of band 10 (Fig 3).

Next, a 4-panel animation comparing the three water vapor channels (all impacted) and the 10.3 um IR channel (not impacted, Fig 4).

Finally, a 24-hr full disk loop (starting 0230 UTC) showing all 16 GOES-17 ABI channels on 15 Feb 2019 (Fig 5). Not bands 1-7 (VIS, near-IR, shortwave IR) and 13-14 (clean IR window) and have no degradation.

Bill Line, NWS

It was announced today (11 Feb) that GOES-17 will become the operational GOES-West satellite tomorrow (12 Feb) at 1800 UTC. GOES-15 (previous GOES-West) will continue to operate from the 128W position through early July 2019. The GOES-17 GOES-West position is 137.2W. Figure 1 shows GOES-17 Full Disk (15-min) water vapor imagery from today.

Bill Line, NWS

Freezing dense fog developed across the far eastern Colorado plains and surrounding states during the early morning hours of 6 Feb 2018. By mid morning, the dense fog began to lift and retreat east, revealing hoar frost had been deposited on surfaces. The hoar frost was first apparent to NWS PUB forecasters in the GOES-16 Day Cloud Phase Distinction RGB, where dark green colors appeared along the outside edge of the retreating low clouds (Fig 1).

This dark green color comes from a relatively high reflectance (compared to bare surface, but not as high as with fresh snow) in the VIS, combined with warm temperatures in the IR, and low reflectance in the snow ice band. All of these details indicate a blanket of ice crystals on the surface. Analyzing the snow ice single-band imagery, it further appears that there was indeed a layer of ice left behind by the retreating fog, as ice crystals have a very low reflectance (Fig 2). In the channel 2 VIS, the frost layer is subtly apparent as slightly higher reflectance compared to the bare surface, but slightly lower reflectance compared to the low clouds (Fig 3). In all examples, the frost is seen quickly retreating/melting as the low cloud shield also retreats, exposing the frost to the sun.

Finally, webcams in the area confirm the thick layer of hoar frost (Fig 4 and 5).

Bill Line, NWS

A bitterly cold polar airmass dipped into the north central United States during the last week of January 2019. During the early morning hours of the 30th, temperatures dropped into the -20s to -30s (F) over a wide swath of the upper midwest, while daytime highs on the 30th were not expected to get out of the negative teens. With breezy northwest winds also present, wind chill values dropped below -50F in many locations (Fig 1). GOES-16 provided some intriguing views of the cold airmass as it impacted the region during the week.

GOES-16 IR window imagery 15-hour animation showed the broad scope of the cold air as it expanded across the midwest during the evening of the 29th into the morning of the 30th (Fig 2). Clear-sky brightness temperatures below -30C were widespread, with values below -40C found in the Dakotas and Minnesota. These values corresponded to brightness temperatures typical of mid to upper level clouds.

The GOES-16 Land Surface Temperature (LST) derived product uses GOES-16 imagery to estimate the skin temperature of the surface (Fig 2). The LST product uses the 11.2 um and 12.3 um IR ABI channels in its calculation. Temperatures below -40C were sampled across the region, with readings as low as -50C sampled just north of the border in Manitoba.

Of course, the airmass was very dry as well. The extent of the dryness is exemplified when viewing the three water vapor channels in conjunction with the 10.3 um IR channel (Fig 3). The 7.3 um (low-level) and then 6.95 um (mid-level) water vapor channels show brightness temperatures just slightly cooler than the 10.3 um IR window channel (~-40C) within the cold region, indicating they are all sensing in a similar layer of the atmosphere (near surface) and that there is a lack of moisture in the atmosphere since these moisture-sensitive channels are sensing so close to the surface. Further, surface features such as lakes and rivers are easily apparent in the 7.34 um channel, and to a lesser extent, 6.95 um. Usually, due to absorption by water vapor, these channels are sensing progressively higher in the atmosphere and are not able to detect surface features. The 6.19 um (upper level) water vapor channel is only running ~5C colder than than the other two, further demonstrating the degree of dryness in the atmosphere.

Fig 5 is similar to Fig 4, but zoomed in and sampling a point near the IA/MN border. Notice the land features apparent in the IRW, LLWV, and MLWV, particularly the rivers in Illinois.

Another way to view the location in the atmosphere within which these channels are sensing is to analyze weighting function plots, available from CIMSS. The weighting functions for the GOES-16 water vapor channels over GRB are shown in Fig 6. These soundings compare well with those from a standard midlatitude winter atmosphere with 10% column moisture (Fig 7), it is obvious that greatest contribution to these channels is coming from much lower in the atmosphere than is typical. It is also apparent that, since the weighting functions are peaking lower than normal, that the atmosphere is quite dry. In fact, the 7.34 um channel is peaking at the surface, which is not surprising considering we are seeing surface features in the imagery.

The GOES-16 Total Precipitable Water (TPW) Derived product reinforces the degree of dryness, showing a wide region of values less than 0.1″ (Fig 8).

00Z 1/30 Soundings in the region support the GOES-16 TPW estimates, including apparent daily record values less than 0.05″ at ABR, INL, GRB, and MPX. Example in Fig 9 is from MPX

Fig 9: 00Z 30 Jan 2019 MPX sounding (left) and MPX TPW sounding climatology (right).

Bill Line, NWS

An upper level storm system brought winter weather and severe weather to a large portion of the US on 18-19 Jan 2019. GOES-16 water vapor and IR imagery show the extent and evolution of the mid latitude cyclone on the 19th (Fig 1).

Snow left by the system the previous day and evening was obvious in Day Cloud Phase Distinction RGB imagery (green) as low clouds (aqua/light blue) dissipated (Fig 2). Rapid melting of the snow was apparent during the day, especially in Texas.

Further east, thunderstorms along and near the cold front produced damaging winds and a few tornados. GOES-16 1-min imagery was available over the region to support forecast operations. The GOES-16 GLM revealed an overall lack of lightning activity along the cold front. However, it did highlight the location of (particularly strong) updraft cores and strengthening/weakening updraft trends (Fig 3).

Finally, an undular bore was captured in GOES-16 imagery traveling south along the east coast of Mexico (Fig 4).

Bill Line, NWS

A recent snowfall and relatively moist airmass in place made for a complicated scene over southern Colorado on 13 Jan 2019. As has been shown in previous blog posts, the Day Cloud Phase Distinction RGB is a great tool for differentiating low clouds (light blue) from snow (green) from high clouds (reds) during the day. In this example, the RGB imagery was utilized to monitor low cloud and fog evolution across much of the CWA in order to keep the grids/forecast accurate, but especially near KPUB, KCOS, and KALS in order to provide accurate TAFs. 1-min imagery was available from one of the GOES-17 default mesoscale sectors to provide even more detailed cloud analysis. Compare the RGB (Fig 1) with VIS (Fig 2). Cloud analysis is much easier in the RGB, particularly in areas where low clouds are present over/next to snow cover.

Note the sharp gradient in snow cover south of the Arkansas River (KPUB to KLHX to KLAA), where northerly flow becomes upslope, aiding the development of snowfall. Near and north of the river to the Palmer Divide, northerly flow is downslope, which limits snow potential in such a case. Snow is apparent again further north along the Palmer Divide from KCOS through KDEN, where northerly flow is upslope again.

Notice the stationary area of relatively high reflectance just south of KPUB? It turns out that is steam emitted from a local Power Plant (Figure 3)!

Bill Line, NWS

GOES-17 data are preliminary and non-operational.

GOES-16 has periodically been scanning 1-min mesoscale sectors in South America for the RELAMPAGO Research Mission since this past October. Strong thunderstorms in Argentina were captured in the mesoscale sector on 11 December 2018. 

Focusing in on one of the storms for a 1.5 hour period (FIg 1-3), a persistent and large overshooting top is clear, along with a downstream above anvil cirrus plume, both indicating a particularly strong updraft and possible severe weather at the surface.  Shown are VIS, IR, and a VIS/IR sandwich combo.

Also shown are 15-min loops and wider views of the storm cluster moving across Argentina.

A GOES-16 1-min sector captured strong thunderstorms developing within a corridor of relatively high atmospheric moisture on Dec 14. The GOES-16 Total Precipitable Water product showed values up to 2.5″ within a large corridor of 2″+. 

The field study continued into 2019. Widespread thunderstorm activity was captured by the GOES-16 1-min sector over Argentina during the morning of 6 Jan. Figures 6 and 7 show ongoing convection after sunrise, including several storms exhibiting overshooting tops and above anvil cirrus plumes, which are indicative of particularly strong and potentially severe updrafts. Figure 8 shows the rapid development of a thunderstorm over a 95 minute period.

On 15 Jan 2019, 1-min visible imagery over Argentina provided an intriguing visualization of atmospheric shear. Loading Derived Motion Winds (DMWs) as well (updating every 5-min in mesoscale sectors), the shear can be better quantified. Sampling some winds in the area, we see 10-15 knot winds from 160 degrees at around 950 mb and 30 knot winds from 310 degrees at 550 mb yields a deep layer shear magnitude of around 40 knots.

Bill Line, NWS