Satellite Liaison Blog

Given active convective and tropical weather in recent weeks, there have been many days with multiple GOES-16 meso sector requests from the field, re-positioning the sectors from their default locations. On 15 August, NHC requested a sector to aid in monitoring Hurricane Gert out in the Atlantic. WFO Albuquerque requested a sector due to their radar being down in a SPC marginal risk for severe.

GOES-16 VIS/IR sandwich combo of Gert provided a great view of thunderstorm activity near the center of the storm as well as gravity wave features east of the center (Fig 1). The differential motion at different vertical levels is also quiet mesmerizing to see in 1-min imagery.

Figure 1: 15 August 2017 GOES-16 1-min sandwich VIS/IR combo of Hurricane Gert in the Atlantic Ocean. Full res: https://satelliteliaisonblog.files.wordpress.com/2017/08/20170815_sw_gert_ani2_anno.gif

Meanwhile back over the US, widespread convection developed across the high plains in the other 1-min sector (Fig 2). Features indicative of very strong updrafts were present with these storms, including overshooting tops and above anvil cirrus plumes. MRMS indicated hail over 1″ in diameter with many of these storms. The 1-min imagery allowed for real-time monitoring of these features as an assessment of updraft health (strengthening, maintaining strength, weakening), as well as quick identification of new convective initiation which was quite rapid given the large amount of instability. Significant hail of 2″ diameter (Lime size) was reported with a storm near Hugo in east-central Colorado.

Figure 2: 15 August 2017 GOES-16 1-min VIS over Colorado. Full res: https://satelliteliaisonblog.files.wordpress.com/2017/08/20170812_1min_hp_ani_anno.gif

Severe storm activity persisted in east-central Colorado by sunset. One particular storm that developed rapidly near Matheson, CO in Elbert county around 23z produced a reported 1.75″ size hail (golf ball). Of note with this storm observed in the GOES-16 1-min VIS was the long-lived above anvil cirrus plume, signifying a strong updraft consistently penetrating the troposphere (Fig 3).

Figure 3: 15 August 2017 GOES_16 1-min VIS over east-central Colorado. Full res: https://satelliteliaisonblog.files.wordpress.com/2017/08/20170812_1min_co_ani3_anno.gif

-Bill Line, NWS

“The GOES-16 data posted on this page are preliminary, non-operational data and are undergoing testing. Users bear all responsibility for inspecting the data prior to use and for the manner in which the data are utilized.”

Imagery from GOES-16 ABI greatly enhances our ability to detect and differentiate cloud types. During the day, visible imagery makes this task fairly simple. At night, however, it is difficult to detect low clouds and differentiate cloud types using the raw IR imagery alone.  In this post we investigate the varying cloud types/heights over Colorado/New Mexico and the southern High Plains on 12 August 2017 using GOES-16 image combinations.

The 10.3 – 3.9 um difference from GOES satellites has long been used to detect low clouds and fog at night. The higher spatial and temporal resolution from GOES-16 allows us to now detect and track low cloud and fog evolution with more precision. The first animation below is the 10.3 – 3.9 um fog difference over the region during the pre-dawn hours (Figure 1). The blacks and darkest grays represent a positive difference, or water clouds with small droplets. The negative values, or light grays to white, represent ice/high clouds. The fog difference is great for differentiating low-mid/water clouds from upper/ice clouds, but does not allow one to further discern cloud type/height.

Figure 1: 12 August 2017 GOES-16 fog difference. Full resolution

With the increase in spectral bands available with GOES-16 (16 vs 5), we now have the ability to create more advanced image combinations, including RGBs. An RGB is created by combining three channels or channel combinations, each of which contribute to the color red, green, or blue. One such RGB known as the advanced nighttime microphysics RGB, is particularly helpful for further differentiating cloud types/heights at night. The ingredients for this RGB are as follows:

• Red: 12.3 – 10.35 um for cloud thickness. Large contributions come from low positive to negative difference values (thick clouds or clear skies and dry atm). Small contributions will come from high positive difference values (thin clouds or clear skies and very moist atmosphere).
• 10.35 – 3.9 um (green) for ice vs water cloud. Large contributions come from high positive difference values (water clouds). Small contributions come from values near 0 (land) and negative values (ice clouds).
• 10.35 um temperature. Large contributions come from warmer temperatures (low clouds or clear skies). Small contributions come from cold temperatures (high clouds).

The nighttime microphysics RGB below is for the same time period and domain.

Figure 2: 12 August 2017 GOES-16 nighttime microphysics RGB. Full resolution

Lets analyze the imagery at 1142 in Figure 3. First let’s start with the water clouds, those that appeared dark (positive difference). The nighttime microphysics , in this case, draws out three fairly distinct colors from these dark areas. The widespread aqua colors across the eastern plains are low clouds, and the most likely areas of fog in the scene. The area has fairly strong green and blue contributions, telling us these are low water clouds. Surface METAR observations within this region did in fact range from fog to ceilings <500 ft. The patch of clouds in south central Colorado has more of a green tint to them due to higher contributions from red and green (a thicker cloud) and lower contributions from blue (slightly cooler cloud). These are still likely low clouds, but less likely to be fog. METAR’s confirm ceilings around 1300 ft. In central Colorado, the light green to tan colors tell us these are actually mid level clouds (water and possibly ice mixed in), as they have an even lower contribution from blue (even cooler clouds) but still have contributions from red and green. METAR’s indicate ceilings of 10,000 ft with these clouds. The nighttime microphysics RGB allowed for these three areas of water clouds to be differentiated, something the fog difference alone could not do.

Figure 3: 12 August 2017 1142 UTC GOES-16 nighttime microphysics RGB.

Analyzing the same figure, we also see that areas of upper clouds, which appear light gray to white in the fog difference (negative values), can be broken into multiple groups. High thick high clouds appear the most red since they have little to no contribution from both blue (cold) and green (ice), but moderate contribution from red (thick). High thin clouds appear a dark blue to almost black given little to no contribution from red (thin) and green (ice), and little contribution from blue (cold cloud but semi-transparent).

The land surface color will vary, and other features such as fire hotspots and snow cover can be diagnosed in the nighttime microphysics RGB.

Quick guides and other training resources for the GOES-16 RGB’s will be available in the coming months.

-Bill Line, NWS

“The GOES-16 data posted on this page are preliminary, non-operational data and are undergoing testing. Users bear all responsibility for inspecting the data prior to use and for the manner in which the data are utilized.”

Widespread severe convection developed across eastern Colorado ahead of a cold front in an environment characterized by favorable deep layer shear and instability on 10 August 2017. Given this, SPC had a slight risk for severe storms out for a broad area, and would later upgrade a portion of the region to Enhanced. Given the threat, WFO-Pueblo requested 1-min imagery from GOES-16. Convection would develop quickly off the high terrain and move east into the plains. These storms produced widespread severe hail, including a few reports up to baseball size. The GOES-16 1-min animation below shows rapid development of a storm on the Colorado/Kansas border which would quickly produce at least golf ball sized hail. The overshooting tops and above anvil cirrus plume indicate a particularly strong updraft and likely severe.

Figure 1: 10 August 2017 1-min VIS along Colorado/Kansas border. Full res: https://satelliteliaisonblog.files.wordpress.com/2017/08/20170810_1min_vis4.gif

-Bill Line, NWS

“The GOES-16 data posted on this page are preliminary, non-operational data and are undergoing testing. Users bear all responsibility for inspecting the data prior to use and for the manner in which the data are utilized.”

Hurricane Franklin, the first Atlantic Basin Hurricane in the GOES-R era, approached the east coast of Mexico during the evening of 09 August 2017. One strategy to observe a smooth day/night transition, discussed in a previous blog post, involves combining visible and infrared imagery. For this case, we take it a step further by including the sandwich VIS/IR combo during the day, and transition to just IR at night. The sandwich combo, also discussed in previous posts, has a VIS underlay with a semi-transparent IR overlay of only the coldest cloud tops. GOES-16 was undergoing testing on the 9th, so no 1-min meso sectors were available, leaving “only” 5-min imagery to observe Franklin.

Figure 1: 08 August 2017 VIS/IR Sandwich combo during day transition to IR at night for Hurricane Franklin. Full resolution: https://satelliteliaisonblog.files.wordpress.com/2017/08/20170809_hurr_transition_sw2_anno.gif

-Bill Line, NWS

“The GOES-16 data posted on this page are preliminary, non-operational data and are undergoing testing. Users bear all responsibility for inspecting the data prior to use and for the manner in which the data are utilized.”

Over the weekend Tropical Storm (TS) Franklin formed off of the coast of Honduras, a prime location for viewing with GOES-16.  NHC forecasters have found the higher spatial and temporal visible imagery very useful for analyzing genesis in the past.  However, TS Franklin formed overnight, a time when visible satellite imagery is not available.

NHC 3-day Forecast Track,Initial Wind Field and Watch/Warning Graphic for Tropical Storm Franklin on 06 August 2017, 11 PM EDT – the first advisory after being classified a Tropical Storm. (http://www.nhc.noaa.gov/archive/2017/FRANKLIN.shtml?)

Fortunately, there are several new capabilities available from GOES-16 for nighttime viewing – including the Nighttime Microphysics (Advanced) RGB product (originally created by EUMETSAT).  The Nighttime Microphysics RGB product uses infrared channels to provide information about the fundamental properties of clouds including drop size and phase (water, mixed phase, ice) at nighttime. It can help distinguish between ice phase clouds at high elevations and water phase clouds at lower elevations, providing a pseudo three-dimensional view of the atmosphere. NHC forecasters found this product particularly useful in determining whether or not TS Franklin had a closed surface circulation, a necessary condition for formation.  

From the 5am Tropical Cyclone Discussion, it’s clear that the GOES-16 Nighttime Microphysics imagery provided forecasters with information that other data sources could not:

In the ASCAT image below, there are no westerly winds along the south side of the pre-Franklin disturbance (which was situated just off the coast of Honduras at this time), which would suggest the surface circulation was not yet closed.

ASCAT-A pass mentioned in the above TCD.  Source:  https://manati.star.nesdis.noaa.gov/datasets/ASCATData.php/ASCATData.php

However, in the GOES-16 Nighttime Microphysics imagery, the motion of small, low-level clouds along the south side of the pre-Franklin disturbance could be seen to have a westerly component of motion, indicating that the circulation may in fact be closed after all.

Full-disk GOES-16 Nighttime Microphysics RGB imagery over the pre-Franklin disturbance from 0-6 UTC on 8/7/17.  **THIS DATA IS PRELIMINARY AND NOT OPERATIONAL**

These low-level cloud motions were much easier to see in the 5-minute resolution imagery, which was available because the pre-Franklin disturbance was within the CONUS domain at this time.

CONUS GOES-16 Nighttime Microphysics RGB imagery over the pre-Franklin disturbance from 0-6 UTC on 8/7/17.  **THIS DATA IS PRELIMINARY AND NOT OPERATIONAL**

-Andrea Schumacher, CIRA/CSU and Michael Folmer, CICS/UMD

“The GOES-16 data posted on this page are preliminary, non-operational data and are undergoing testing. Users bear all responsibility for inspecting the data prior to use and for the manner in which the data are utilized.”

The 2017 Summer Experiment at the Aviation Weather Testbed has begun, it’s focus centered on mock ups of the operational Traffic Flow Management (TFM) Convective Forecast desk (TCF), and on the Digital Aviation Services ceiling and visibility effort. Yesterday was a spin up day, allowing forecasters to peruse various new datasets as they explored the mock desks.

While the tropics were not in the overall experiment plan, it was difficult not to turn the attention southward to Franklin, particularly when new Geostationary Lightning Mapper (GLM) data was available.  The GLM data is optical, meaning that the concept of detecting lightning is much different from typical terrestrial networks. Therefore, developers at the AWC have been exploring various methods of displaying this data as a grid. Below is an example of the GLM groups, a ten minute average estimated every two minutes.

GLM group lightning density associated with Franklin on August 7th – a ten minute average ever two minutes. THIS DATA IS PRELIMINARY AND NON-OPERATIONAL!

Participants also were able to look at groups as an energy density, also averaged over ten minutes every two.

GLM group energy associated with Franklin on August 7th – ten minute average every two minutes. THIS DATA IS PRELIMINARY AND NON-OPERATIONAL!

These grids were designed for the AWC’s N-AWIPS systems, both created on an eighth of a degree grid over the entirety of the GOES-16 domain. When zoomed in, the grid points are noted to be pretty coarse. However, these particular displays were designed with AWC tropical and global forecasters in mind, folks whose perspective is broader. In their eyes, this coarse resolution isn’t a big issue. For others, such as TCF and Convective SIGMET, it may be that a similar display with higher resolution will need to be created. Additionally, gridded plots of the group energy density are also being added to the selection for later in the experiment.

Why gridded plots? As mentioned, GLM is an optical sensor, collecting lightning data in a very different manner than ground based networks… i.e. point data from that optical sensor does not equate to point data taken from a ground-based sensor calibrated to pick up a specific frequency. Therefore, it is important for forecasters to examine data with a different perspective than they have with other lightning data in the past in order to glean it’s value. For that reason, AWC has started the exploration with grids.

Over the next two weeks, we plan to take a cautious approach, keeping the known data quality issues in mind while collecting feedback and exposing forecasters to this new exciting data.

With dry and breezy conditions across large portions of the northwest US, wildfire activity has been ramping up in recent weeks. Yesterday (8/1), multiple ongoing wildfires exploded during the daytime hours, producing large smoke plumes. A GOES-16 RGB (hotspot, smoke, burn scar) discussed in previous posts captured the hotspots and smoke plumes well (Figure 1).

Figure 1: 01 August 2017 GOES-16 hotspot/smoke/scar RGB. Full res: https://satelliteliaisonblog.files.wordpress.com/2017/08/20170801_wildfire_rgb1_final_anno.gif

– Bill Line, NWS

“The GOES-16 data posted on this page are preliminary, non-operational data and are undergoing testing. Users bear all responsibility for inspecting the data prior to use and for the manner in which the data are utilized.”