Satellite detection of ice

February 25th, 2022 |

A colder-than-normal February over the western Great Lakes (through the 27th, Duluth is 10o F below normal; Marquette is 5o F below normal; Green Day is 2o F below normal; Cleveland is 1o F below normal) has fostered a growth in ice cover over the Lakes (This figure, from here, for example). How can satellites detect that ice, especially for a region where winter-time cloudiness is notorious? In general, there are two different ways to detect ice: Visible/Infrared imagery and Microwave imagery. The toggle above shows Advanced Technology Microwave Sounder (ATMS) ice detection (using MiRS algorithms and data from Suomi-NPP (1807 UTC) and NOAA-20 (1859 UTC)) over the Great Lakes on 25 February 2022. (These data come from the Direct Broadcast antenna at CIMSS, and are processed using CSPP to produce AWIPS-ready tiles that are available via an LDM feed). A big challenge with this field is the very large ATMS footprint. Note that these sea-ice concentration values are quantitative: values change based on how much ice is within the footprint but are also dependent on view angle.

VIIRS data (as shown at the VIIRS Today website, for example) can also be used to infer regions of ice in a qualitative sense, as shown below. The true color imagery shows possible ice features over the lakes. It’s a challenge, however, to use a single VIIRS image to distinguish (mostly) stationary ice from (usually) moving clouds. Multispectral VIIRS imagery means (via the use below of the 2.25 µm M11 band) ice features are cyan colored and can be qualitatively distinguished from clouds. Consider, for example, the color difference in the False Color image between the clouds over eastern Lake Superior (white in both True- and False-Color) and near-shore ice over southern and eastern Lake Superior (white in the True-Color and cyan in the False-Color). There are also VIIRS-based Ice Concentration products that can be computed in clear skies.

Suomi-NPP VIIRS True- and False-Color imagery over the Great Lakes, 1806 UTC on 25 February 2022 (Click to enlarge)

For cloudy regions, Advanced Baseline Imagery can be used to create estimates of Ice Surface Temperature and Ice Concentration. Quantitative estimates such as these give more information than the qualitative estimates shown above. These estimates at present are only created hourly; in partly-cloudy regions, that cadence is sufficient to give lake-wide quantitative estimates of ice coverage and temperature. CONUS imagery — with a 5-minute time cadence — and mesoscale sectors — with 1-minute cadences — can be used to monitor (in a qualitative sense) how ice is moving (as shown link, for example). High temporal-resolution imagery is important because ice sheets can change rapidly under strong winds, as shown in this tweet. At high latitudes, there are also ways of monitoring ice motion via polar orbiters (link).

GOES-16 Estimates of Ice Surface Temperatures, 1800 and 1900 UTC on 25 February 2022 (Click to enlarge)
GOES-16 Estimates of Ice Concentration, 1800 and 1900 UTC on 25 February 2022 (Click to enlarge)

A much higher-resolution method of viewing ice (again, in a qualitative, not quantitative sense) in regions of both clear and cloudy skies, day or night, is through the use of Synthetic Aperture Radar (SAR). Data are available for each Great Lake at this link. Imagery for each Lake over the past days is available there, albeit infrequently (usually around 0000 and 1200 UTC only) and over small domains. This qualitative imagery, however, is very high-resolution and gives very impressive details. Imagery over Lake Erie is shown below.

RCM estimates of Lake Erie Ice, 24-27 February 2022 (Click to — greatly! — enlarge)

VIIRS and ABI give both qualitative (false-color, visible imagery) and quantitative (ice concentration and ice temperature) depictions of ice over the lakes — or over coastal waters around Alaska. Microwave data also gives quantitative estimates (ice concentrations with large microwave footprints) and qualitative estimates (SAR data). Use all products to create a clear picture of ice coverage.

Ozone and the airmass RGB

December 13th, 2021 |
GOES_17 airmass RGB, 2200 UTC on 12 December 2021 (Click to enlarge)

A GOES-17 airmass RGB, above, shows a strong feature in the Gulf of Alaska. It’s common to associate the orange and purple regions within that polar feature (that is accompanied by cloud features consistent with very cold air aloft) with enhanced ozone. What products are available online to gauge the amount of ozone?

The OMPS instrument on board NOAA-20 (and on Suomi-NPP) senses in the ultraviolet (from 250-310 nm) to compute ozone concentration. (For more information on OMPS, refer to this document) The figure below, taken from this Finnish website, shows ozone concentration for the 24 hours ending at 0110 UTC on 13 December. A distinct maximum is apparent over the Gulf of Alaska. Note the northern terminus of the observations that are related to the time of year: there is little Sun north of 60 N. The data for this were downloaded from the Direct Broadcast site at GINA at the University of Alaska-Fairbanks. OMPS data are also available (from Suomi-NPP) at NASA Worldview.

To determine the time of the data in the image below, consult the NOAA-20 orbital paths here. This image (from that site) shows a NOAA-20 ascending overpass between 2235 and 2245 UTC over the Gulf of Alaska.

Daily Composite of Ozone concentration for the 24 hours ending 0111 UTC on 13 December 2021 (click to enlarge)

NOAA-20 also carries the Cross-track Infrared Sounder (CrIS) and Advanced Technology Microwave Sounder (ATMS) instruments that are used to create NUCAPS vertical profiles; one of the trace gases retrieved in this way is ozone. The distribution of ozone (with values in regions where it was dark) from NUCAPS is shown below (from this website maintained by SPoRT), and it corresponds roughly with the OMPS estimates shown above.

Gridded NUCAPS estimates of ozone, 2217 UTC on 12 December 2021 (Click to enlarge)

Conclusion: The assumption that upper-tropospheric ozone values are large in regions where the airmass RGB is tinted red or purple is a good assumption, especially if other structures in the RGB — such as cumulus cloud development in the cold air — reinforce the idea that an intrusion of stratospheric air is occurring. The strong storm that this lowered tropopause is supporting is accompanied by a moist feed of air moving into central California, as shown below by MIMIC total precipitable water fields.

Total Precipitable Water, 2200 UTC on 12 December 2021 (Click to enlarge)

Gridded NUCAPS fields are being tested within RealEarth, as shown below. They should be generally available soon.

RealEarth Gridded NUCAPS estimates of ozone, 2217 UTC on 12 December 2021 (Click to enlarge)

TROPICS Pathfinder view of super typhoon Mindulle

September 26th, 2021 |
TROPICS Pathfinder 205 GHz imagery, 0545 UTC on 26 September 2021 (Upper Left) and Himawari-8 Band 3 Visible (0.64 µm) Imagery, 0540 UTC on 26 September (Lower Right) (Click to enlarge)

Time-Resolved Observations of Precipitation structure and storm Intensity with a Constellation of Smallsats (TROPICS) Pathfinder imagery from 0545 UTC on 26 September, when super typhoon Mindulle was near peak intensity, is compared above to Himawari-8 visible(0.64 µm) imagery at about the same time. A separate image links small features in the Pathfinder image to small convective elements that are apparent in the Himawari imagery. Click here to view the TROPICS Pathfinder image with a NOAA-20 true-color image from 0426 UTC.

The pathfinder satellite that provided the microwave data used for the image above is the first in a series of a constellation of low-Earth orbiters; six additional satellites will be launched next year. These are very small satellites, with a size of 10 cm x 10 cm x 36 cm. They weigh in at 5.34 kg / 11.8 pounds! Pathfinder imagery was provided courtesy of the Science Team working with the data. Himawari-8 imagery are courtesy of JMA.


As noted above, NOAA-20 overflew Mindulle at about 0430 UTC. The Advanced Technology Microwave Sounder (ATMS) instrument on NOAA-20 sampled the storm, and imagery (88.2 GHz and 183.3 GHz) with a timestamp of 0435 UTC (from this archive) is shown below. The NOAA-20 orbits over the western Pacific on that day are shown here (from this site). Structures in the Pathfinder imagery at 0545 UTC can be identified in the 0435 UTC ATMS imagery below. A side-by-side comparison of the Pathfinder 205 GHz and NOAA-20 ATMS 183.3 GHz is shown at bottom.

NOAA-20 ATMS imagery from channel 16 (88.2 GHz) and channel 18 (183.3 GHz), 0435 UTC , 26 September 2021 (Click to enlarge)
NOAA-20 ATMS 183.3 GHz imagery, 0435 UTC on 26 September (left) and Pathfinder 205 GHz imagery, 0545 UTC on 26 September (right) (Click to enlarge)

Overnight views of Tropical Storm Fred from NOAA-20 and Suomi-NPP

August 16th, 2021 |
Suomi NPP Day Night Band Visible (0.7 µm) and I05 (infrared, 11.5 µm) at 0704 UTC on 16 August 2021 (Click to enlarge)

VIIRS Imagery from Suomi-NPP (the Day Night Band, at 0.7 µm and I05, at 11.5 µm, above, at 0704 UTC) and from NOAA-20 (the Day Night Band, at 0.7 µm and I05, at 11.5 µm, below, at 0754 UTC), show polar orbiting perspectives of Tropical Storm Fred. The Day Night Band shows little detail in the cloud-tops, given the lack of lunar illumination (the moon — about half-illuminated — was below the horizon during these overpasses). The region of coldest cloud tops seems to have decreased from the NPP to the NOAA-20 overpass. The NHC discussion at 0300 UTC suggested that the low-level circulation had emerged from underneath the cirrus canopy, and perhaps that’s detectable in the toggles above and below. The storm does exist under southwesterly shear (morning analysis from the CIMSS Tropical Website).

Suomi NPP Day Night Band Visible (0.7 µm) and I05 (infrared, 11.5 µm) at 0754 UTC on 16 August 2021 (Click to enlarge)

Microwave data from the Advanced Technology Microwave Sounder (ATMS) on board both Suomi-NPP and NOAA-20 can be used to compute a rain rate, as shown below. The Rain Rate snapshots show a decrease in intensity between 0704 (a maximum of about 0.9″/hour) and 0754 UTC (a maximum of 0.5″/hour). That information, combined with the lack of observed lightning in the Day Night Band, might help guide an analyst in their description of the storm strength. Consider, however, that NPP had a near-nadir view whereas NOAA-20 was an edge view.

ATMS Rain Rate derived from Suomi NPP (0704 UTC) and NOAA-20 (0754 UTC) using MIRS Rain Rate algorithm (Click to enlarge)

VIIRS imagery and the ATMS Rain Rate is available in AWIPS-ready tiles via the CIMSS ldm feed. Imagery from Polar Orbiters over Fred (and Grace, and TD #8 (now Henri) in the Atlantic) is also available from the AOML Direct Broadcast link here.

For more information on Tropical Storm Fred, refer to the National Hurricane Center. Fred is forecast to make landfall on the Florida panhandle on 16 August.