ProbSevere products over the Southern Plains

May 3rd, 2021 |

The NOAA/CIMSS ProbSevere portfolio contains AI models for nowcasting convective weather. I’ll use Monday’s severe weather over the Southern Plains to highlight several of them.

A strong cold front spawned numerous severe-hail, wind, and tornado producing storms over Texas and Oklahoma, aided by very large values of convective available potential energy (CAPE; > 4000 J/kg).  You can see numerous storm reports in Figure 1.

210503_rpts Reports Graphic

Storm Prediction Center’s preliminary severe storm reports for May 3rd, 2021.

Probsevere version 2 (PSv2) is an operational set of models at NOAA, which predict the probability of severe hail, severe wind, and tornadoes, in the next 60 minutes. The models are storm-centric, and the models’ domain is the entire contiguous United States (CONUS).  These models use MRMS (radar), GOES (satellite), short-term NWP, and terrestrial-based lightning observations to generate probabilistic guidance of severe hazards. Figure 2 shows output from an experimental version (PSv3), which includes additional MRMS, GOES, and NWP fields as predictors in a machine learning model.

Figure 2: ProbSevere v3 contours (colored, around storms), MRMS MergedReflectivity, and NWS severe weather warnings (yellow and red boxes) for storms over the Southern Plains. The second outer contour around some storms is colored by the probability of tornado.


Another ProbSevere product is a convolutional neural network that uses GOES-R ABI and GLM images to detect regions of intense convection, and is often correlated with strong overshooting tops, “bubbly-like” texture in visible imagery, strong lightning cores, and the cold-U/above-anvil cirrus plume signature. The intense convection probability (ICP) can be run on the 1-minute mesoscale scans as well as 5-minute CONUS sector scans aboard the GOES satellites. The ICP does not require radar data, and may also be able to operate on data from satellites with similar intruments (e.g., Meteosat Third Generation). ICP output is being used as a predictor in the experimental ProbSevere v3.


Predicting when and where lightning will occur is also important for many users, such as mariners, aviators, and outdoor event managers. The probability of lightning model (PLTG) is also a convolutional neural network, using images of visible, near-infrared, and longwave-infrared channels to nowcast lightning occurrence in the next 60 minutes. The purple-to-orange shaded regions in the video below show GLM flash-extent density (i.e., flashes passing through a location).

A View of the Development of Geostationary Imagers through the lens of BAMS

May 14th, 2020 |

A collection of 60 BAMS covers spanning the years, to highlight the rapid advance of imaging from the geostationary orbit, is shown above (a version that loops more slowly can be seen here). The first cover is the first of BAMS, in January of 1920, while the second, from January of 1957 is the first time artificial ‘satellite’ was in a title of a BAMS article. The third image, from November of 1957, is a remarkable article on potential uses of satellites. This included both qualitative uses: (1) Clouds, (2) Cloud Movements, (3) Drift of Atmospheric Pollutants, (4) State of the Surface of the Sea (or of Large Lakes), (5) Visibility or Atmospheric Transparency to Light — and quantitative uses: (1) Albedo, (2) Temperature  of  a  Level  at  or  Near  the Tropopause, (3) Total Moisture Content., (4) Total  Ozone  Content, (5) Surface  (Ground-Air Interface) Temperature, and (6) Snow Cover. Early covers showcase rockets, balloons and high-altitude aircraft to prepare the way to human space travel (Gemini, Apollo, etc.), polar-orbiters (TIROS, NIMBUS, VHRR, NOAA, etc.) and finally geostationary orbit (ATS-1, ATS-3, SMS, GOES, Meteosat, INSAT, Himawari, etc.).

Reasons to look back at the BAMS covers:

Interactive web page, with links to the original “front matter”.

Montage of select BAMS covers

Montage of select BAMS covers

Note: All cover images are from the Bulletin of the American Meteorological Society.

Geostationary Satellite Matching Webapp

January 8th, 2019 |

A webapp that was developed to match images and hence learn about the ABI spectral bands. After navigating to the page, click and hold on an image, draw a line to the matching image. You can “right click” to open an image. Finally click on the yellow “check” box to verify your selections.

8 Spectral bands

A (30 sec) movie (mp4) showing how to use the matching webapp for 8 of the ABI spectral bands. (Click to play)

One option is to match 8 of the total 16 ABI spectral bands.

16 Spectral bands

Beginning to match two of the 16 ABI spectral bands.

Another option is to match all 16 ABI spectral bands.

Matching Visible bands to ground-based photos

In this case, ABI visible images match to photographs.

With this webapp, one matches the correct ABI visible spectral band with a photograph that was taken on the ground. A short (30-sec) animation on running the webapp: mp4 or mov.

How to use this webapp with your images

Check out this site for directions on how to build a web page to match pairs of images. This webapp is Copyright © 2018 by Tom Whittaker. Images from T. Schmit and NOAA. Inspired by Jordan Gerth’s ABI Matching game.

Questions and other Webapps

Webapps have been developed to demonstrate other concepts of remote sensing, such as pixel size and generating composites.

For any questions.

Transitory Solar Reflectance in GOES-R Series Imagery

March 5th, 2018 |

GOES-16 Visible (0.64 µm) animation, 1637-1732 UTC on 5 March 2018 (Click to enlarge)

Animations of GOES-16 Visible, near-Infrared and shortwave Infrared over North America shortly before the Vernal Equinox, and shortly after the Autumnal Equinox, (that is, when the Sun is overhead in the Southern Hemisphere) show bright spots that propagate quickly from west to east (these features were first noted by Frank Alsheimer of the National Weather Service). The animation above shows the visible imagery (0.64 µm) over the Continental United States on 5 March 2018 (Click here for a slower animation speed). Brightening over regions between 30 and 40 N between 1637 UTC and 1732 UTC is apparent. The animation below of the shortwave infrared (3.9 µm) shows slight warming (Click here for a slower animation), as might be expected with reflected solar energy. The brightening is also apparent in the Band 4 “Cirrus”  (1.37 µm) — in fact, a closer look at southern Colorado reveals the bright signature of sunlight reflecting off solar panels at the Alamosa Solar Generating Facility (Google maps).

GOES-16 Shortwave Infrared (3.9 µm) animation, 1637-1732 UTC on 5 March 2018 (Click to enlarge)

The increased reflectance can cause the ABI Clear Sky Mask to mis-characterize clear regions as cloudy (See the animation below; click here for a slower animation). Thus, Cloud properties (Cloud-top Height, Temperature, Pressure, etc.) can be identified in clear regions.

GOES-16 Clear Sky Mask (White: Clouds ; Black : No Clouds) from 1637 UTC – 1732 UTC on 5 March 2018 (Click to enlarge)

The bright spots in the visible, and warms spots in the shortwave infrared, occur when the Earth’s surface, the GOES Satellite and the Sun are aligned on one line. If you were within the bright spot with a powerful telescope trained on the Sun, you would see the GOES Satellite transecting the solar disk. The location of these bright spots changes with season: they appear in the Northern Hemisphere shortly before the (Northern Hemisphere) vernal equinox and shortly after the (Northern Hemisphere) autumnal equinox. Similarly, they appear in the Southern Hemisphere shortly before the (Southern Hemisphere) vernal equinox and shortly after the (Southern Hemisphere) autumnal equinox. On the Equinox, the bright spots are centered on the Equator.

This animation (courtesy Daniel Lindsey, NOAA/CIRA and Steve Miller, CIRA) shows where the reflection disk moves during the days around the Northern Hemisphere Autumnal Equinox; a similar animation for the Northern Hemisphere vernal equinox would show a disk starting at the North Pole and moving southward with time.

The animation below (from this link that is used for calibration exercises), shows the difference in reflectance (Bands 1-6) or Brightness Temperature (Bands 7-16) between 1657 and 1652 UTC on 3 and 5 March 2018. Two things are apparent: The centroid of the largest difference in solar reflectance has moved southward in those two days, as expected; the effect of this solar backscatter is most obvious in the visible, near-infrared and shortwave infrared channels (that is, bands 1-7 on the ABI).  The effect is most pronounced in clear skies.

Time Difference in each of the 16 ABI Channels (1657 – 1652 UTC) on 3 and on 5 March 2018 (Click to enlarge)

This reflectance feature is also detectable in legacy GOES Imagery. However, the great improvements in detection and calibration in the GOES-R Series ABI (and AHI on Himawari-8 and Himawari-9) and the better temporal resolution with the GOES-R Series allows for better visualization of the effect.

The feature also shows up in “True Color” Imagery, shown below (from this site). Geocolor imagery (shown here), from CIRA, also shows the brightening.

CIMSS Natural True Color Animation ending 1757 UTC on 5 March 2018 (Click to enlarge)

Thanks to Daniel Lindsey and Tim Schmit, NOAA/ASPB, Steve Miller, CIRA and Mat Gunshor/Scott Bachmeier, CIMSS, for contributions to this blog post.