CIMSS Satellite Blog

MTSAT-1R + GOES-11 visible images (Animated GIF)

The Kasatochi Volcano (located in the Aleutian Islands of Alaska) experienced a series of eruptions on 07-08 August 2008. A comparison of visible channel images from the MTSAT-1R and GOES-11 satellites (above) shows the initial sequence of volcanic plumes from 2 very different satellite viewing angles — note that the third eruption plume (beginning around 05 UTC on 08 August) appeared much darker that the previous 2 plumes, suggesting a higher volcanic ash content.

GOES-11 “split window” IR difference images (below) displayed a volcanic ash signal (yellow to cyan color enhancement) — again, the volcanic ash signal appeared to be more well-defined after the third eruption (beginning around 05 UTC on 08 August).

GOES-11 split window IR difference images (Animated GIF)

About 4 days after the initial eruption, AWIPS images of the GOES-11 + GOES-12 Sounder 7.4 µm channel (below) revealed a signature of a portion of the volcanic plume (lighter blue color enhancement) that was drifting eastward across the northwestern and north-central US on 11-12 August. The GOES Sounder 7.4 µm channel was designed to be used primarily as a lower-tropospheric “water vapor” channel, but it happens to  also be sensitive to sulfur dioxide (SO2).  However, this volcanic plume was also evident on GOES-11 visible channel images (QuickTime animation), which suggests that the “SO2 plume” is an aerosol feature (possibly a plume consisting of supercooled water coated sulfate particles).

AWIPS images of GOES-11 + GOES-12 Sounder 7.4 µm channel (Animated GIF)

NOAA Air Resources Laboratory HYSPLIT model trajectories (below) suggested that the features seen arriving over eastern Wyoming around 00 UTC on 12 August could very well have been transported from the region of the Kasatochi eruption over the Aleutians. There were also a number of pilot reports of volcanic ash over the region during that time period (including an interesting report of “SULFUR SNOW” over northeastern Montana).

HYSPLIT model trajectories

The GOES-13 satellite had just recently been taken out of on-orbit storage for evaluation and testing in early August 2008. A sequence of GOES-13 Sounder IR difference  [7.4 µm (Band 10) minus 13.4 µm  (Band 5] images (below; courtesy of Tony Schreiner, CIMSS) showed a signal of the “volcanic SO2 plume” (darker black enhancement) as it moved eastward from Montana and Wyoming on 11 August to Minnesota and Iowa on 13 August. As the cloud shield cleared over southern Wisconsin on 13 August, Arctic High Spectral Resolution Lidar located at the University of Wisconsin - Madison indicated a layer of aerosol backscatter centered around 12 km, which could very well have been part of the Kasatochi volcanic plume.

GOES-13 Sounder IR difference product (Animated GIF)

In addition, another portion of the “volcanic SO2 plume” could be seen moving southward across Ontario on 12 August, then moving southeastward across New England on 13 August (these particular volcanic plume features were forecast with remarkable accuracy by an Environment Canada Lagrangian transport model). The large hazy feature seen in the northeastern part of the MODIS true color image from the SSEC MODIS Today site (below) was the leading edge of the Ontario “volcanic SO2 / aerosol plume” as it began to move southward over the Great Lakes region on 12 August. According to the NASA Earth Observatory News, this was one of the largest volcanic sulfur dioxide clouds scientists have observed since Chile’s Hudson volcano erupted in August 1991. In addition, this was the second Alaskan volcanic plume in as many months to be observed over the Lower 48 states — the Okmok volcanic plume was seen in mid-July 2008.

MODIS true color image

** 15 AUGUST UPDATE ** Additional lidar data obtained from the University of Wisconsin - Madison on 14-15 August (below) continued to show thin layers of aerosol backscatter with small depolarization ratios (cyan colors) in the upper troposphere that were possibly due to Kasatochi volcanic plumes.

Arctic High Spectral Resolution Lidar data

It is interesting to note the thin “tail” of aerosol backscatter (cyan colors) that extended downward from the main aerosol layer (located between altitudes of 11-12 km) to as low as the 8-9 km altitude range during the 04-08 UTC time period on 15 August. AWIPS images of the GOES Sounder + GOES Imager water vapor channels (below) indicated that strong subsidence was occurring over Wisconsin during that time — warmer water vapor brightness temperatures values (indicative of drier air) were depicted by the blue to yellow to orange colors (depending on which particular water vapor channel was being viewed).

So the Question of the Day is: could the lidar data be showing evidence that some of the volcanic aerosol plume aloft was being transported downward several km by the strong subsidence that was occurring within the middle to upper troposphere over Wisconsin on 15 August? The GOES Sounder total column ozone product showed a lobe of elevated ozone values, concurrent with a lowering of the dynamic tropopause (taken to be the pressure of the PV1.5 surface) to around the 300 hPa pressure level (around 9 km) over the Madison WI area, in agreement with the lidar filament seen extending down to the 8-9 km level — so perhaps a stratospheric intrusion may have helped to transport a portion of the volcanic aerosol plume downward. HYSPLIT back trajectories (EDAS | GDAS) indicated that the transport arriving over Madison WI on 15 August at the 8, 10, and 12 km altitudes all passed over Ontario and Hudson Bay during the preceeding 24 hours, where the thick aerosol feature was seen on GOES and MODIS imagery 2 days earlier.

AWIPS images of GOES Sounder+Imager water vapor channel data

On the morning of 17 July 2008, we received the following email from Ron Miller at the National Weather Service forecast office in Spokane, Washington:

Can you tell me what you think the “clouds” are over the PacNW this morning. You can only see them on the first few visible images of the morning. Looking out out window we can’t detect them. Given the flow pattern over the past 24-48 hours, I have a hard time believing that it’s wildfire smoke from CA. One possibility is Volcanic Ash from Okmok, since after it’s eruption, the ash cloud drifted southeast and essentially hung out in the Gulf of AK trough. Any ideas?

Excellent question, Ron — thanks for bringing this case to our attention! As it turns out, the Okmok volcano erupted in the Aleutian Islands on 12  July, as can be seen in a comparison of visible images from the MTSAT-1R and GOES-11 satellites (above | QuickTime animation; see also:  Google Earth AVHRR false color image | VISIT Meteorological Interpretation Blog | GOES-11 visible Java animation) – and 5 days later, GOES-11 (GOES-West) visible imagery (below)  did indeed reveal a portion of the volcanic plume  (actually, three separate thin plumes) drifting eastward over the Pacific Northwest region early in the day on 17 July. The thin volcanic plumes seen on the GOES visible imagery  — which were likely composed primarily of ice crystals and sulfur dioxide (SO2) — were  high-altitude features (verified to exist at an altitude around 11-12 km by CALIPSO), so they showed up on the visible imagery before  the lower-altitude smoke (from to wildfires that had been burning in northern California) which  became illuminated by the rising sun a bit later in time. Forward scattering was more favorable at the times of the earlier visible images, enhancing the appearance of the volcanic plume features  — then, as the sun angle increased into the mid-morning hours, the thin volcanic plumes became less apparent on the visible imagery (but the thick low-level smoke drifting northeastward across Oregon continued to remain  obvious).

The corresponding GOES-11 10.7-12.0 µm “split window IR difference” product did not show a clear signal of volcanic ash content in the plume (nor was there an obvious plume signal in either the GOES-11 10.7 µm IR window, 3.9 µm shortwave IR, or 6.5 µm water vapor imagery). However, an Aqua MODIS IR difference image (below) did show the signature of an SO2-rich plume (darker blue colors) stretching east-northeastward across the Pacific Ocean and  reaching western Washington and Oregon around 11:00 UTC (5am local time). The daily evolution of the SO2 plume can also be seen in an animation of Ozone Measuring Instrument (OMI) images. Compare the 3-plume structure seen on the MODIS IR difference image with the similar plume structure seen on an AIRS brightness temperature difference image.

NOAA Air Resources Laboratory HYSPLIT model backward trajectories (below) indicated  that high-altitude air parcels arriving over western Washington and far northwestern Oregon at 11:00 UTC on 17 July had likely been transported southward, then east-northeastward across the Pacific Ocean during the previous 72-hour period.

A comparison of AWIPS images of the MODIS visible, IR window, water vapor, and cirrus detection channels from 18:47 UTC (11:47 am local time) on 17 July (below) show that there was no obvious volcanic plume signature on either the visible or IR images at that time. However, a brighter “plume signal” did show up in the cirrus detection channel image (which seems to correspond roughly to the northernmost of two moist plumes on the water vapor image).

The MODIS cirrus detection channel is a daytime-only “near-IR” channel (centered at 1.6 µm) which is sensitive to particles that are efficient scatterers of light (such as ice crystals, volcanic ash particles, airborne dust or sand, etc). Two consecutive  MODIS cirrus detection images from the Terra (18:47 UTC) and Aqua (20:26 UTC) satellites (below) both show evidence of a subtle volcanic plume  signal (brighter white streaks), which appeared to be moving farther inland over the Pacific Northwest.

By the end of the day on 17 July, a favorable forward scattering angle (with the sun getting lower in the western sky) allowed the volcanic plume to again be seen on GOES visible channel imagery — but this time using the GOES-12 (GOES-East) satellite (QuickTime animation).  The plume had moved eastward along the US/Canada border during the day, drifting across parts of northern North Dakota and even far northwestern Minnesota by sunset!

** 18 July Update: The volcanic plume was again evident on early morning  GOES-11  visible imagery (Animated GIF), stretching from the Pacific Northwest states all the way to southern  Manitoba and southern Ontario in Canada. During the afternoon and early evening hours, there were numerous pilot reports from aircraft encountering the plumes over northern Oregon, southwestern Washington, and other parts of the Pacific Northwest region, prompting the issuance of a Volcanic Ash SIGMET advisory.

A 17:49 UTC AWIPS image of the MODIS “cirrus detection” channel suggested that the Volcanic Ash SIGMET advisory could perhaps have been extended a bit farther northeastward across Montana — note the slightly brighter “streaky” volcanic plume signature  heading northeastward across Idaho into Montana, beyond the eastern boundary of the SIGMET (outlined in yellow). In fact, the boundary of the Volcanic Ash SIGMET was indeed extended northeastward soon thereafter (Animated GIF).

Chile is one of the most volcanically active countries on Earth. The latest volcano to erupt is Chaiten, which had previously lain dormant for at least 1000 years. Chaiten is at approximately 42 degrees South latitude, 72 degrees west longitude, close to Golfo de Ancud. A series of eruptions, starting on May 2, has prompted the evacuation of Chaiten, a provincial capital with a population of 4000.

GOES-10 captured the plume of an eruption that started near 12 UTC on 6 May. Volcanic ash does not have an emissivity of 1; that is, it does not emit as a blackbody. The emissivity at 10.7 microns is smaller than the emissivity at 12 microns. The smaller signal received at 10.7 microns (relative to the assumed blackbody) is interpreted as a cooler emitting surface. If the blackbody temperatures at 10.7 and 12.0 microns are compared, then, values at 12.0 microns are warmer. A channel difference can be used to highlight the horizontal extent of the volcanic ash. In the loop shown, the bluest pixels correspond to a blackbody temperature difference of nearly 10 K. That is, the 12 micron blackbody temperature is 10 K warmer than the 11 micron blackbody temperature. The remnants of an older eruption are also noted near the Atlantic Coast of Argentina.

A sequence of GOES-10 imager and sounder IR difference products during the 02-08 May period (below) shows evidence of plumes from multiple eruptions of the Chaiten Volcano. The GOES-10 Imager can provide nearly continuous (15 minute) coverage of the evolving ash cloud, while the GOES-10 Sounder can provide details on the upper-level SO2 plumes once every four hours. The former is derived utilizing the 11.0 micrometer and 12.0 micrometer bands from the Imager. SO2 plumes are revealed by differencing the 7.4 micrometer and 13.3 micrometer bands from the Sounder.

Click here to see an animated gif every from every four hours — that is, each hour for when sounder data are available.

An AVHRR false color image from 05 May (below, viewed using Google Earth) revealed a long plume from the Chaiten volcano, which stretched eastward across Argentina and then southeastward over the South Atlantic Ocean.