Ice over the Great Lakes

April 17th, 2015
Suomi-NPP Imagery:  Visible (0.64µm), Day Night Band (0.70µm) and near-IR (0.85µm) (click to enlarge)

Suomi-NPP Imagery: Visible (0.64µm), Day Night Band (0.70µm) and near-IR (0.86µm) images (click to enlarge)

Visible Imagery over the Great Lakes on Friday April 17th showed mostly open waters over the five lakes, with regions that could be ice confined to coastlines of Lakes Superior, Huron, Erie and Michigan. The animation above is of Suomi NPP VIIRS visible (0.64µm and 0.70µm) and near-infrared (0.86µm) data. Can you tell with certainty which of the white features over the lakes are clouds vs. ice?

Suomi-NPP Infrared Imagery (3.74 µm),  (click to enlarge)

Suomi-NPP Infrared Imagery (3.74 µm) (click to enlarge)

Infrared data can give clues. The 3.74 µm imagery, above, shows the brightness temperature. Note how the white regions over Lakes Superior, Michigan and Ontario are about the same temperature as the surrounding water. In contrast, white regions over Lakes Erie and Ontario are much darker (warmer) in the 3.74 µm than the surrounding water. This is testimony to the superior scattering abilities around 3.74 µm of water-based clouds compared to lake ice. More solar radiation scattered towards the satellite by the clouds means a warmer temperature is detected.

Suomi-NPP Imagery:  Toggle between Visible (0.64µm) and near-IR (1.61 µm) (click to enlarge)

Suomi-NPP Imagery: Visible (0.64µm) and near-IR (1.61 µm) (click to enlarge)

The 1.61 µm near-infrared channel is useful because ice strongly absorbs solar radiation at that wavelength, appearing dark. The toggle above, of visible (0.64) and near-infrared (1.61) neatly distinguishes between clouds and ice. Ice (dark in the 1.61 µm because it does not reflect; at that wavelength, it absorbs) is apparent over eastern Lake Superior, eastern and northern Lake Huron and some small bays in northern Lake Michigan. There is no ice apparent on Lakes Erie or Ontario: features there exhibit signatures which are white in both visible and at 1.61 µm.

Another method to aid in the discrimination of snow/ice vs supercooled water droplet clouds is the creation of Red/Green/Blue (RGB) products. The example below toggles between the 0.64 µm visible image and an RGB image (which uses the VIIRS 0.64 µm/1.61 µm/1.61 µm data as the R/G/B components) — snow cover and ice appear as darker shades of red on the RGB image (in contrast to supercooled water droplet clouds, which are brighter shades of white). The snow depth on the morning of 17 April was still 13 inches at Munising in the Upper Peninsula of Michigan.

Suomi NPP VIIRS 0.64 µm visible and false-color RGB images (click to enlarge)

Suomi NPP VIIRS 0.64 µm visible and false-color RGB images (click to enlarge)

On this day there was only 1 pass of the Landsat-8 satellite over any of the ice-covered portions of the Great Lakes; the 15-meter resolution panchromatic visible (0.59 µm) image below shows a very detailed view of the far western portion of the ice that was north of the Keweenaw Peninsula in Lake Superior (zoomed image).

Landsat-8 panchromatic visible (0.59 µm) image (click to enlarge)

Landsat-8 panchromatic visible (0.59 µm) image (click to enlarge)

Terra and Aqua both carry the MODIS sensor, and MODIS can detect radiation at 1.38 µm, a wavelength at which cirrus is highly reflective. A 1.38 µm image from the 17th, below, shows the horizontal extent of cirrus.

MODIS Imagery:  near-IR (1.38 µm) (click to enlarge)

MODIS Imagery: near-IR (1.38 µm) (click to enlarge)

Hawai’i demonstrates that the Water Vapor channel is an Infrared channel

April 6th, 2015
GOES-15 6.5 µm water vapor channel images (click to play animation)

GOES-15 6.5 µm water vapor channel images (click to play animation)

GOES-15 (GOES-West) 6.5 µm “water vapor channel” images (above; click image to play animation; also available as an MP4 movie file) revealed an interesting transition in the signal displayed by the 2 summits (Mauna Kea and Mauna Loa) on the Big Island of Hawai’i on 06 April 2015 — beginning as a pair of colder (darker blue color enhancement) areas during the nighttime hours, becoming a pair of warmer (brighter yellow color enhancement) areas as daytime heating warmed the land surfaces.

As was discussed in a previous blog post, the water vapor channel is essentially an Infrared (IR) channel that senses the mean temperature of a layer of moisture — usually a layer which is located in the middle troposphere. However, if the middle troposphere is dry, the water vapor detectors are able to “see” lower into the atmosphere and detect radiation from the lower atmosphere (or even high-elevation terrain features, such as Mauna Kea and Mauna Loa). A comparison of the 00 UTC and 12 UTC rawinsonde profiles from Hilo (below) showed that the middle troposphere was indeed quite dry, with the typical tropical moisture residing below the 700 hPa pressure level.

Hilo, Hawai'i rawinsonde data profiles (00, 12 UTC)

Hilo, Hawai’i rawinsonde data profiles (00, 12 UTC)

The altitude (and depth) of the layer being sensed by a water vapor channel is defined by its weighting function, which depends on (1) the temperature and moisture profile of the atmosphere, and (2) the satellite viewing angle or “zenith angle”. This site allows you to select a rawinsonde site of interest, and the GOES Imager (and Sounder) water vapor channel weighting functions are calculated and plotted. The GOES-15 Imager water vapor channel weighting functions for the 2 Hilo soundings are shown below (along with the weighting function for the US Standard Atmosphere). It can be seen that the peak of the weighting function response is at a lower altitude for both Hilo soundings than it would be for the US Standard Atmosphere, which in part allows the strong cold/warm thermal signatures of the two Big Island summits to be seen on the GOES-15 water vapor imagery.

Hilo, Hawai'i GOES-15 imager water vapor weighting functions, compared with the US Standard Atmosphere

Hilo, Hawai’i GOES-15 imager water vapor weighting functions, compared with the US Standard Atmosphere

Extratropical Cyclogenesis over the western Pacific

March 30th, 2015
Himawari-8 AHI 0.64 µm visible channel images (click to play animation)

Himawari-8 AHI 0.64 µm visible channel images (click to play animation)

The AHI Instrument on Himawari-8 has 16 different channels sensing the atmosphere. The instrument is still in Post-Launch Testing, a period when instrument performance is monitored and adjusted. Extratropical cyclogenesis that occurred east of Japan on 30 March was captured by the different channels.

The 0.64 µm visible imagery, above, is the highest-resolution channel on AHI, with nominal 0.5-km resolution at the subsatellite point. The imagery above — at 1.5 km resolution and every 10 minutes — shows the development of an extratropical cyclone east of the main island of Japan (visible at the left edge of the imagery). Thin cirrus is spreading north of the storm and convection is developing both in the cool air north of the surface circulation center and along the cold front that is just to the west of the cirrus shield associated with the warm conveyor belt. Northerly surface winds north of the system and southern surface winds south of the system speak to the strengthening of the frontal boundary along which the storm is developing.

Himawari-8 AHI 0.85 µm infrared channel images (click to play animation)

Himawari-8 AHI 0.85 µm infrared channel images (click to play animation)

The 0.85 µm imagery, above, is in the near-infrared part of the electromagnetic spectrum, at wavelengths just a bit longer than red visible light (which is at 0.7 µm). It does an excellent job highlighting the land/water contrast (because bodies of water strongly absorb 0.85 µm solar radiation and land and clouds reflects it). This channel also is sensitive to vegetation. The larger-scale view shows jetstream cirrus south and southwest of the developing storm and an occluded system decaying to the east of Kamchatka.

The 0.46 µm imagery, below, is in the visible part of the electromagnetic spectrum, and is quite sensitive to aerosols (Click here for a fact sheet on ABI’s 0.46 µm “Blue Band”; fact sheets for all ABI Bands will be here in the future). The smog and pollution that surrounds Tokyo is more apparent in this imagery. Smog is also indicated near Osaka and Nagoya. A toggle between 0.64 µm, 0.46 µm and 0.85 µm imagery, here, from 30 March 2015 at 0000 UTC allows a comparison of the imagery.

Himawari-8 AHI 0.46 µm visible channel images (click to play animation)

Himawari-8 AHI 0.46 µm visible channel images (click to play animation)

The 1.60 µm imagery on AHI is useful because it can distinguish between clouds with water droplets (that scatter and reflect solar 1.60 µm radiation very effectively) and clouds with ice crystals (that absorb 1.60 µm radiation). In a standard enhancement, clouds with ice crystals appear grey, clouds with water droplets appear white. In the animation below, the glaciated cirrus canopy of the warm conveyor belt is readily apparent. Note also how the convection developing along the warm front has glaciated by the end of the animation.

Himawari-8 AHI 1.60 µm infrared channel images (click to play animation)

Himawari-8 AHI 1.60 µm infrared channel images (click to play animation)

The 3.9 µm on Himawari-8 provide detailed information about the sea surface temperature if clouds are not present, as was the case over the Kuroshio Current just east of Japan on 30 October. The animation below shows little change over 2 hours, as expected, except along the north wall of the current. Brightness Temperatures drop 10 C across the temperature gradient at the north end of the current.

Himawari-8 AHI 3.90 µm infrared channel images (click to play animation)

Himawari-8 AHI 3.90 µm infrared channel images (click to play animation)

Great Lakes surface geographical outlines evident on water vapor imagery

February 23rd, 2015
GOES-13 6.5 µm water vapor channel images (click to play animation)

GOES-13 6.5 µm water vapor channel images (click to play animation)

A cold and dry arctic air mass (morning minimum temperatures) was in place over the Great Lakes region on 23 February 2015. This arctic air mass was sufficiently cold and dry throughout the atmospheric column to allow the outlines of portions of the surface geography of the Great Lakes to be seen on GOES-13 (GOES-East) 6.5 µm water vapor channel images (above; click image to play animation).

In addition to the commonly-used 4-km resolution 6.5 µm water vapor channel on the GOES Imager instrument, there are also three 10-km resolution water vapor channels on the GOES Sounder instrument (centered at 6.5 µm, 7.0 µm, and 7.4 µm). A 4-panel comparison of these water vapor channel images (below; click image to play animation) provides the visual indication that each water vapor channel is sensing radiation from different layers at different altitudes — for example, the surface geographical outlines of the Great Lakes are best seen with the Sounder 7.4 µm (bottom left panels) and the Imager 6.5 µm (bottom right panels) water vapor channels.

GOES-13 Sounder 6.5 µm, 7.0 µm, 7.4 µm, and Imager 6.5 µm water vapor channel images (click to play animation)

GOES-13 Sounder 6.5 µm, 7.0 µm, 7.4 µm, and Imager 6.5 µm water vapor channel images (click to play animation)

An inspection of GOES Sounder and Imager water vapor channel weighting function plots (below) helps to diagnose the altitude and depth of the layers being sensed by each of the individual water vapor channels at a variety of locations. For example, the air mass over Green Bay, Wisconsin was cold and very dry (with a Total Precipitable Water value of 0.87 mm or 0.03 inch), which shifted the altitude of the various water vapor channel weighting functions to very low altitudes; this allowed surface radiation from the contrasting land/water boundaries to “bleed up” through what little water vapor was present in the atmosphere, and be sensed by the GOES-13 water vapor detectors. In contrast, the air mass farther to the south over Lincoln, Illinois was a bit more more moist, especially in the middle/upper troposphere (with a Total Precipitable Water value of 4.20 mm or 0.17 inch) — this shifted the altitude of the water vapor channel weighting functions to much higher altitudes (to heights that were closer to those calculated using a temperature/moisture profile based on the US Standard Atmosphere).

GOES-13 Sounder and Imager water vapor channel weighting function plots for Green Bay WI, Lincoln IL, and the US Standard Atmosphere

GOES-13 Sounder and Imager water vapor channel weighting function plots for Green Bay WI, Lincoln IL, and the US Standard Atmosphere

In addition to the temperature and/or moisture profile of the atmospheric column, the other factor which controls the altitude and depth of the layer(s) being detected by a specific water vapor channel is the satellite viewing angle (or “zenith angle”); a larger satellite viewing angle will shift the altitude of the weighting function to higher levels in the atmosphere. Recall that the water vapor channel is essentially an Infrared (IR) channel — it generally senses the mean temperature of a layer of moisture or clouds located within the middle to upper troposphere. In this case, the sharp thermal contrast between the cold land surfaces surrounding the warmer Great Lakes was able to be seen, due to the lack of sufficient water vapor at higher levels of the atmosphere to attenuate or block the surface thermal signature.

The new generation of geostationary satellite Imager instruments (for example, the AHI on Himawari-8 and the ABI on GOES-R) feature 3 water vapor channels which are similar to those on the current GOES Sounder, but at much higher spatial and temporal resolutions.

On a separate — but equally interesting — topic: successive intrusions of arctic air over the region allowed a rapid growth of ice in the waters of Lake Michigan. A 15-meter resolution Landsat-8 0.59 µm panochromatic visible image viewed using the SSEC RealEarth web map server (below) showed a very detailed picture of ice floes along the western portion of the lake, as well as a patch of land-fast ice in the far southern end of the lake.

Landsat-8 0.59 µm panochromatic visible image (click to enlarge)

Landsat-8 0.59 µm panochromatic visible image (click to enlarge)

The motion of the band of ice floes along the western  edge of Lake Michigan was evident in 1-km resolution GOES-13 0.63 µm visible channel images (below; click image to play animation) — along the east coast of Wisconsin, southwesterly winds gusting to around 20 knots were acting to move the ice floes away from the western shoreline of Lake Michigan.

GOES-13 0.63 µm visible channel images (click to play animation)

GOES-13 0.63 µm visible channel images (click to play animation)