Satellite Liaison Blog

GOES-16 imagery in AWIPS-II transitioned to Fixed Grid today at 1527. See below for information and examples.

•  What
  • GOES imagery in NWS AWIPS is now be on native geostationary grid (fixed grid)
  • Up until now, fixed grid data has been remapped to a different projection
    • Result is oversampling of native fixed grid data
    • Smaller pixels, appearance of more detail, smoothed data, different reflectance/BT
• Result of using Fixed Grid Data
  • Increase accuracy and clarity of imagery
  • Reduce latency, bandwidth, disk storage
  • Decrease AWIPS load time
  • Daytime Convection RGB Composite

The example below (Figure 1) over eastern Kansas uses 1-min VIS and IR imagery at 1526 UTC (old) and 1527 UTC (new fixed grid) to exemplify the transition. When using the CONUS projection in AWIPS, the pixels now appear diagonal (away from position of the satellite sub point) and larger (with distance away from satellite sub point). Considering the data are no longer smoothed across pixels, the imagery and features within (cu, anvil edges, overshooting tops) appear crisper.

Figure 1: 19 June 2018 GOES-16 VIS and IR transition to Fixed Grid over eastern Kansas. Full res

A longer 1-min animation centered over the time of transition (Figure 2).

Figure 2: 19 June 2018 GOES-16 1-min VIS and IR imagery over eastern Kansas. Transition to fixed grid occurs at 1527 UTC, roughly midway through the animation. Full res

Here is a video from the OPG on the transition: https://www.youtube.com/watch?v=Dg8VacIUl38&feature=youtu.be

– Bill Line, NWS

GOES-16 visible imagery observations of the Ute Park Fire smoke plume on June 2 yielded some interesting phenomenon (Fig 1). Early on in the loop and before the fire really gets going for the day, the plume is trapped under a stable layer. As the mountain slopes heats up to the west and north of the wildfire, flow under the inversion drives the smoke west/north toward the heated/lower pressure slopes. After the wildfire becomes more active and heats up substantially during the early afternoon, the smoke plume penetrates through the inversion and taps into the westerly flow aloft, accelerating east. Winds at the surface increase from the south, and smoke in the lower levels accelerates north. Toward the end of the loop, gravity wave “ripple” features are diagnosed emanating away from the fire within the smoke layer still trapped under the inversion.

Figure 1: 2 June 2018 GOES-16 5-min 0.64 um visible imagery centered over Ute Park Fire in northern New Mexico. Full res

Bill Line, NWS

GOES-16 1-min satellite imagery was available over the southern High Plains for monitoring of severe convection on 30 May 2018. One storm in particular displayed rapid intensification, as displayed in Figure 1 below. Using a 1-min updating sandwich image combo, rapid cooling and overshooting top development is easily diagnosed with the southernmost storm. The overlaid tracking meteogram trend graph is another way to show the rapid cooling/strengthening that occurred. The storm top cooled roughly 13C in 20 minutes! Severe wind damage and 2″ diameter hail were reported with this storm about 10 minutes after it was at its coolest.

Figure 1: 30 May 2018 GOES-16 1-min VIS/IR Sandwich combo. Overlay: Tracking Meteogram graph of minimum 10.3 um brightness temperature during period of animation. Severe reports are also overlaid. The northernmost wind report is associated with a different storm, while the southern wind and hail reports are associated with the storm of interest. Full res

Bill Line, NWS

Severe thunderstorms developed across the high plains during the afternoon on 28 May 2018, Memorial Day. In this post we focus on convection over the southern High Plains. In addition to storms developing off of the high terrain of Colorado and advancing east to the adjacent plains, convection also fired along a dryline over KS/OK/TX/NM. Using the 10.3 – 12.3 split window difference from GOES-16, forecasters could locate and track (in 5-min intervals) the precise position of the moisture gradient as early as the morning hours, prior to even the first cu development. This difference has been discussed in previous posts. With this straightforward linear grayscale color map, darker grays represent relatively drier air at the low levels, while lighter grays indicate increased low-level moisture. The key with the split window difference is to identify the gradients in the field. As an overlay on the difference is the 10.3 um IR window channel with warm brightness temperatures transparent, so only the cold cloud tops are highlighted. In this event, the aforementioned dryline is clearly evident in the difference imagery, along which convection develops. A relatively dry slot is diagnosed pushing north into southeast Colorado, west and north of which the airmass is moist again. Severe thunderstorms developed off the high terrain and into this moist region as well.

Figure 1: 28 May 2018 GOES-16 5-min Split Window Difference (Gray scale) with 10.3 um IR overlay (color scale; warm temperatures made transparent). Full res

Bill Line, NWS

An abundance of strong thunderstorms developed across portions of the US High Plains on 18 May 2018, producing severe criteria hail, winds, and a few tornados. 1-min satellite imagery from GOES-16 was available over the region for this event per a request by the Storm Prediction Center. Notable during this event was the abundance of satellite cloud top signatures apparent with the storms during the early evening. The VIS+IR Sandwich image combo highlights these features quite well. This image combo is created using the 0.64 um VIS as an underlay, and 10.3 um IR as an overlay. The 0.5 km VIS captures the detail in the imagery, while the IR captures the quantitative brightness temperature information. The IR overlay is fully transparent below a certain threshold to allow for detailed analysis of cumulus clouds pre-CI using the VIS alone. The coldest brightness temperatures, which are associated with the cloud tops, are made semi-transparent (40%). Cloud top features apparent in this animation are overshooting tops (cooler temperature, increased texture, shadow), above anvil cirrus plumes (warm anomoly downstream of overshooting top, relatively smooth), and gravity waves (ripples) emanating from the updrafts. The rapid anvil expansion with these storms is also indicative of strong updrafts. The persistence of these features indicates long-lived strong updrafts.

Figure 1: 18 May 2018 GOES-16 1-min VIS+IR Sandwich image combo. Full res

Bill Line, NWS

A Supercell thunderstorm brought accumulating hail to east-central Colorado on 14 May, just east of I-25 and north of Castle Rock. The hail on the ground was observed by GOES-16 products. The 0.5 km, 0.64 um VIS revealed the hail swath quite well as the storm advanced east (Fig 1). The Day Cloud Phase Distinction RGB, which utilizes the 0.64 um channel (green) in addition to the 1.6 um NIR channel (blue) and 10.3 um IRW channel (red), made the hail swath even more apparent (Fig 2). The ice is highly reflective in the VIS and lowly reflective in the 1.6 um channel, like the thunderstorm clouds. However, unlike the clouds, the hail swath is warm because it is being sensed at the surface. The result is a hail swath (dark green) that contrasts nicely with the surrounding clouds (yellows, reds).

Figure 1: 14 May 2018 GOES-16 5-min 0.64 um VIS. Hail swath is apparent north of Castle Rock. Full res

Figure 2: 14 May 2018 GOES-16 5-min Day Cloud Phase Distinction RGB. Hail swath is apparent north of Castle Rock (green). Full res

Bill Line, NWS

Winds gusting out of the north at over 40 mph spread across the southeast Colorado plains during the morning and early afternoon of 26 April 2017. Given the strong winds, dust was picked up from two separate areas as detected by the GOES-16 10.3 – 12.3 split window difference (Figure 1). Lofted dust detection using GOES-16 has been described in previous blog posts.  Using the GOES-16 imagery and burn scar maps, it was quickly determined that the blowing dust was emanating from two very recent burn scars: the 117 fire in southern El Paso County and the Badger Hole fire in eastern Baca County. In this example, the dust signal in the imagery was subtle, but the moving pixels of relatively darker gray, combined with the location relative to the burn scars and confirmation via a webcam, gave us confidence that we were seeing blowing dust. The linear gray color scale was adjusted to make the dust areas more obvious. The dust was not easily apparent in visible imagery or radar imagery (KPUX radar was not available).

Figure 2: 26 April 2018 GOES-16 5-min 10.3 – 12.3 um split window difference imagery (gray scale) and IR window channel imagery for cold clouds (color). Yellow circles surround the regions of blowing dust, captured in the imagery as moving pixels of dark gray. The two recent burn scars are outlined in red. Full res

Given what was observed in GOES-16 imagery and confirmed in webcams, areas of blowing dust was added to the forecast grids (Figure 2) through the morning into the early afternoon, when wind speeds would begin to subside. The GOES-16 imagery was used to outline the narrow regions where blowing dust was occurring and would likely continue to occur. The AFD was also updated to read: “GOES-16 split-window difference imagery and webcams indicate blowing dust emanating off of the recent 117 Fire burn scar in southern El Paso County and Badger Hole Fire burn scar in eastern Baca county. Winds out of the north are gusting over 40 mph in these areas, carrying dust well south, reducing visibility and air quality in areas. Have added blowing dust to the grids through the early afternoon, after which wind speeds will begin to decrease.”

Figure 2: 1600 UTC 26 April 2018 PotBlowingDust grid (left) and Weather grid (right). BD = Blowing Dust. Red outlines are burn scar areas. Blowing dust regions were outlined with reference to GOES-16 split window difference imagery. Full res

If someone in eastern Baca County were to check their forecast during this period, it would look like that in Figure 3.

Figure 3: NWS website screen capture for 3 NE Walsh, CO in eastern Baca County. Full res

Bill Line, NWS