Tropical Storm Beta in the Gulf of Mexico

September 20th, 2020 |

GOES-16 Imagery and Derived Motion winds at 1346 UTC on 20 September 2020 (Click to enlarge). ABI Imagery includes Visible (Band 2, 0.64 µm), the near-infrared ‘Cirrus’ (Band 4, 1.38 µm) and the upper water vapor (Band 8, 6.19 µm) and lower water vapor (Band 10, 7.34 µm) infrared imagery. Derived motion winds for 1346 UTC near 1000 mb (green), 850 mb (yellow), 700 mb (orange), 500 mb (cyan) and 300 mb (purple) are also shown

Tropical Storm Beta was in the northwest Gulf of Mexico on 20 September 2020. Visible imagery (with GLM overlain) shows two principal regions of convection, one near the center, and one in a long feeder band to the east of the storm. Derived motion winds (this image includes a legend that links vector color to level) show cyclonic low-level motion in the northwest Gulf of Mexico, divergent motion at 500 mb, and strong westerly outflow at 300 mb.

‘Cirrus Channel’ (Band 4, 1.38 µm) near-infrared imagery shows the considerable upper-level cloudiness associated with the central convection and the convective band east of the center. There is also abundant storm outflow to the east and north of the storm.

Visible imagery and low-level winds show cyclonic motion at low levels.  The convection is displaced to the east because of southwesterly shear over the storm (shown below, in an image take from this site).

200-850 mb wind shear over the Gulf of Mexico, 1400 UTC on 20 September 2020 (click to enlarge)

Both upper-level and low-level water vapor imagery show very dry mid-tropospheric air over Texas. MIMIC Total Precipitable Water, below, (source), also shows the significant dry air over the continent. (Hurricane Teddy is also apparent in the western Atlantic).

Hourly MIMIC Total Precipitable Water estimates for the 24 hours ending 1400 UTC on 20 September 2020 (Click to enlarge)

Is the dry air influencing the development of this storm? Low-level flow (850-700), below, from this site, shows weak easterly flow. Low-level flow is from regions of deep moisture. Upper-level flow (200-700 mb) shows motion from the (dryer) west. These two different airflows are influencing the development of Beta.

Mean 850-700 mb flow at 1200 UTC, 20 September 2020 (Click to enlarge)

Mean 700-200 mb flow at 1200 UTC, 20 September 2020 (Click to enlarge)

For the latest on Beta, refer to the pages of the National Hurricane Center (direct link for Beta).

Hurricane Gilbert: 1988 as seen by GOES-7

September 14th, 2020 |

Hurricane Gilbert (1988) is one of the most intense Atlantic-basin hurricane on record. NOAA’s GOES-7 offer both visible and infrared views of the storm. These images are from the VISSR mode. What is unique about the view from the geostationary orbit, is that it allows both large / synoptic scale views as well as finer (mesoscale) views. 

Visible band


GOES-7 Visible images from September 10-17, 1988. [click to play animation | MP4]

A week-long visible loop of the Hurricane Gilbert as it moves across the Caribbean and through the Gulf of Mexico. Tropical Storm Florence can also be seen near Louisiana, early in the animation. 

Gilbert. GOES-7 Visible

GOES-7 Visible images from September 12-15, 1988. [click to play animation | MP4]

A GOES-7 visible loop over the time period of maximum intensity. 


GOES-7 Visible images from September 13, 1988. [click to play animation | MP4]

The highest spatial resolution visible GOES-17 imagery of Hurricane Gilbert. Note the horizontal striping due to the photo-multipler tube technology that was then used. 

Infrared window band


GOES-7 IR images from September 10-18, 1988. [click to play animation | MP4]

Above is a “large-scale” view of the GOES-7 infrared longwave window band covering September 10-18, 1988. Tropical Storm Florence can also be seen near Louisiana, early in the animation. 

A more “zoomed in” view:


GOES-7 IR images from September 12-14, 1988. [click to play animation | MP4]

All the IR images have been color-enhanced to highlight the coldest temperatures. 

Visible and Infrared window bands

GOES-7 Full Disk

GOES-7 combined visible and infrared full disk image from September 13, 1988. [Click to enlarge.]

A much larger file (18 MB) of the same day/time as above. This is a combined image, with the visible band, along with the cold pixels from the infrared band (color). 

Swipe between GOES-7 Visible and Infrared bands.

Fade between GOES-7 Visible and Infrared bands. (Using this software.)

NOAA GOES-7 data are via the University of Wisconsin-Madison SSEC Satellite Data Services.



Using GOES ABI and deep learning to nowcast lightning

September 2nd, 2020 |

NOAA and CIMSS are developing a product that uses a deep-learning model to recognize complex patterns in weather satellite imagery to predict the probability of lightning in the short term. Deep learning is a branch of machine learning based on artificial neural networks, which have the ability to automatically learn targeted features in the data by approximating how humans learn.

A convolutional neural network (CNN) was trained on over 23,000 images of GOES-16 Advanced Baseline Imager (ABI) data to predict the probability of lightning, within any given ABI pixel, in the next 60 minutes. The CNN was trained using 118 days of data collected between May and August of 2018. Images of GOES Geostationary Lightning Mapper (GLM) flash-extent density (created with glmtools) were used as the source of lightning observations. Note that GLM is an optical sensor that observes both in-cloud and cloud-to-ground lightning.

The CNN currently uses four ABI channels: band 2 (0.64-µm), band 5 (1.6-µm), band 13 (10.3-µm), and band 15 (12.3-µm). Bands 2 and 5 are only utilized under sunlit conditions. Utilization of additional channels and time sequences of images is under investigation. The model uses a semantic image segmentation architecture to assign the probability of lighting in the next 60 minutes to each pixel in the image. The model is very computationally efficient, only needing 30 seconds to process the ABI CONUS domain and 3 seconds to process an ABI mesoscale domain using multithreading on a 40-CPU linux server.

Currently, the model only utilizes satellite radiances. Thus, it can be applied to nearly any spatial domain covered by the ABI or an ABI-like sensor (e.g. AHI). Based on near-real-time testing, the model routinely nowcasts lightning initiation with 10-30 minutes of lead-time. We expect the skill and lead-time will increase as new predictors (e.g. more ABI fields, NWP, radar where available) are added to the model.

Below are a sampling of recent examples. The base images are 0.64-µm reflectance, with GLM-derived flash-extent density overlaid as filled semi-transparent polygons. The flash-extent density is the accumulated number of flashes within the previous 5 minutes. The CNN-derived probabilities are displayed as contours at select levels (near-real-time output is available through RealEarth).

The overall objective is to improve lightning nowcasts in support of aviation, mariners, and outdoor events/activities. Beyond improving the CNN, our work will focus on packaging the output into actionable information for forecasters and other decision makers.

A cold front in Iowa


Thunderstorm development on sea-breeze boundaries in Florida and the Bahamas


Diurnally and orographically forced storms in the Southwest U.S. and Rocky Mountains


Storms in central Oklahoma, on the edge of Hurricane Laura’s cloud shield


A couple of examples over the Northeast U.S.


A boundary of convection on the southern bank of Lake Ontario


Storms in a warm sector in IL/IN/OH, perhaps along an outflow boundary


Southeast U.S. offshore region

The background image in this example transitions from 10.3-µm brightness temperature to 0.64-µm reflectance, while the flash-extent density enhancement also changes, in an attempt to enhance contrast.

First Rapid Scan Satellite Imagery of Volcanic Ash Plumes: July 1980 (Mount St. Helens)

July 22nd, 2020 |



SMS-2 Visible and infrared (IR) from July 23, 1980. The red square represents the approximate location of Mount St. Helens.  [click to play animation | MP4]

The main modern Mount St. Helens eruption was May 18, 1980 — yet there were also later paroxysmal eruptions, such as on June 12/13, 1980. Geostationary satellite imagery from NASA’s SMS-2 (Synchronous Meteorological Satellite) monitored two more Mount St. Helens eruptions on July 22th (local time), 1980, as shown above. Note that in “UTC-time”, the eruption took place on July 23rd. A similar side-by-side SMS-2 visible and infrared animation.  This may be the first* “rapid scan” imaging of a volcanic ash plume (with a 3-minute cadence for almost an hour), where “rapid scan” is defined as satellite imagery less than 5 min apart.

There is a long history of rapid scan imaging from geostationary imagers, including from SMS-1/2, ATS-1, ATS-3, GOES-1, GOES-7 series, GOES-8 series, GOES-14 , Meteosat, etc and of course, AHI and the GOES-R series ABI where 1-min imagery is routine. Here’s a page where users can search historical meso-scale sector locations from the University of Wisconsin-Madison SSEC Satellite Data Services.  

The monitoring of volcanic ash plumes and their attributes have greatly increased from 1980 to today. Moving from qualitative (somewhat after the fact imagery) to quantitative applications (that are much more timely)! Due to the large number of volcanoes, coupled with the increase in satellite observations, satellite observations are key in monitoring the world’s volcanoes for aviation safety and other uses. More on volcanic ash monitoring.


A similar loop as above (SMS-2 Visible and IR from July 23, 1980), but the in mp4 format. Both the day before and after, SMS-02 was in a routine scan mode of imagery every 30 minutes. The rapid scan imagery was just on July 23, 1980 for approximately one hour, starting at 00:14 UTC. 

This webpage allows to customize the loop speed of the SMS visible and infrared side-by-side animation. This uses the hanis software. 

SMS-2 Visible from July 23, 1980

SMS-2 Visible from July 23, 1980 covering approximately one hour. The red square represents the approximate location of Mount St. Helens.  [click to play animation | MP4]

The shadows from the plume are evident. 

A longer duration (4-hr) SMS-02 IR animation (mp4) or (animated gif). The red square represents the approximate location of Mount St. Helens.  Note the less than ideal image navigation. 


NOAA’s GOES-3 was also operating, although not in a rapid scan mode, so imagery was every 30 minutes. 

GOES-3 IR July 1980.

GOES-3 IR July 23, 1980 over 4 hours. The red square represents the approximate location of Mount St. Helens.  [click to play animation | MP4]

The two pulses are clearly evident. 


Thanks to Jean Phillips, the SSEC Data Services, and the Scott’s (Bachmeier and Lindstrom). NASA SMS-2 and NOAA GOES-3 data are via the University of Wisconsin-Madison SSEC Satellite Data Services. More GOES-R series information

* There may have been rapid scan satellite observations of volcanic ash plumes prior to this case in 1980, and if you know of any, please contact T. Schmit.