Explosive eruption at Kilauea Summit — May 17 2018

An explosive eruption at the Kilauea summit occurred near dawn local time on May 17, 2018 according to the USGS Hawaiian Volcano Observatory.  The Washington Volcanic Ash Advisory Center indicates that the eruption reached 30,000 feet.  Meteorological remote sensing tools like satellites and weather radars can provide complimentary information about clouds produced by explosive volcanic eruptions.

Below is an annotated image and a short animation of Himawari-8 ash/dust/SO2 false color imagery between 1400 and 1510 UTC 17 May 2018.  At 1420 UTC an SO2 plume (indicated by the green-ish colors; more information on ash/dust/SO2 false color imagery can be found here) located east of the Kilauea summit emerges and is subsequently advected to the east by the end of the loop.  While an ash signature is not evident in the RGB imagery currently, one must not make the conclusion ash is not present, as complex underlying conditions, including meteorological clouds can mask the ash signature.  Other factors include, but are not limited to, ash grain size effects and dry and wet deposition.  Other sources of information, such as cameras and radar clearly indicated that ash was present (see below).

Himawari-8 false color image from 17 May 2018 at 1450 UTC. An area of mid to upper tropospheric SO2 from Kilauea (highlighted in white) is apparent.

Using the 12 UTC 17 May 2018 observed sounding from Hilo, Hawaii (below), one can use the easterly motion of the SO2 to deduce a plausible height of the SO2.  Assessing the wind profile from Hilo, a large depth of the middle troposphere exhibited 10-15 knot westerly winds (ranging from roughly 16,000 feet to 32,000 feet AGL).  This would suggest the SO2 moving to the east, away from the summit, must reside somewhere in that broad layer.  Because there is a lack of wind shear within this layer (changing wind speed and/or direction with height) the wind analysis only provides a large layer in which the SO2 likely resides in this case.

12 UTC 17 May 2018 Hilo, Hawaii Sounding (source: University of Wyoming)

Combining the satellite and radiosonde height analysis with National Weather Service radar data, one can more accurately pinpoint the height of eruptive column because radar energy is reflected from volcanic ash particles and hydrometeors (if present) back to the radar.  Below is an animation of the NWS South Shore, Hawaii radar elevation angles (1413 – 1420 UTC 17 May 2018).  The various tilts show moderate to strong radar reflectivity associated with eruptive column to near 20,000 feet and faint echoes exist as high as the 10.0 degree elevation scan (over 27,000 feet).  The radar analysis suggests the cloud likely reached 30,000 feet, which is in excellent agreement with the Washington VAAC assessment as well as the possible height range inferred from the above satellite/radiosonde analysis.


To gain additional insight into the eruption, a 3D radar animation from 1357 UTC to 1507 UTC is shown below.  The radar indicates that after achieving a maximum injection height of about 30,000 ft, the particulate cloud, as sensed by the radar, quickly decreases in height.  The decrease in height may be caused, in part, by the preferential sensitivity of the radar to larger particles.  However, the rapid decrease is consistent with a lack of ash signature in satellite data.  A more detailed analysis is needed in order to draw firm conclusions.

Meteorological Satellite Monitoring of Kilauea – May 15, 2018

Enhanced eruptive activity at Kilauea continued on May 15, 2018, with lava flows and volcanic gas and ash emissions.

Clouds have prevented satellite detection of SO2 and ash emissions for much the overnight hours Monday (May 14) night and early morning hours on Tuesday (May 15) (local time).   A S-NPP overpass at 1206 UTC 15 May 2018 provided a higher spatial resolution view of the situation.  Compared to the previous day, less SO2 was evident in S-NPP VIIRS false color imagery.  However, a faint SO2 signal can be seen extending from the Kilauea summit to the south-southwest and is highlighted in the ash/dust/SO2 RGB image below.  It should be noted, that a lack of SO2 signature in imagery does not mean no SO2 is present, rather, the concentration may be below the infrared-based detection limit, which is a strong function of the temperature difference between the SO2 and the background.

S-NPP VIIRS Ash/SO2 false color (RGB) image from 1206 UTC on 15 May 2018. The SO2 plume is outlined in white.

Another relevant type of RGB image can be made by substituting the 3.9 um channel for the 8.6 um channel.  This type of RGB is useful for seeing ash plumes and strong positive surface temperature anomalies associated with lava.  In the S-NPP VIIRS image below, the dark purple spots, highlighted by white arrows, are a result of very high surface temperatures.  The western most arrow is the Kilauea summit, the center arrow is the Pu‘u ‘Ō‘ō crater, and eastern most arrow is lava emerging from fissures in the Lower East Rift Zone.  The 3.9 µm channel, owing to the highly non-linear relationship between temperature and emitted radiation at this wavelength, is critical for detecting positive surface temperature anomalies.

S-NPP VIIRS Ash 3.9 µm false color (RGB) image from 1206 UTC on 15 May 2018. Elevated surface temperatures are indicated by white arrows and appear to be dark purple in the imagery.

An animation of Himawari-8 AHI ash/dustSO2 RGB imagery shows the abundance of meteorological clouds that were present overnight on Monday (May 14) and on Tuesday (May 15) morning.  With much coarser spatial resolution (compounded by the oblique viewing angle of Hawaii), the faint SO2 plume visible on the 1206 UTC S-NPP overpass is not unambiguously distinguishable with Himawari-8 imagery.  This serves as a good example of how geostationary and polar orbiting satellites can be used in tandem for volcanic hazard monitoring.

Lastly, of interest is the Himawari-8 Ash/SO2 RGB imagery loop (below) of a small ash plume from the Kilauea summit during the morning into the afternoon of 14 May 2018 (local time).  Ash emissions occurred in the 1820 – 1940 UTC period and were advected south-southwest.  Being a faint feature, use of the 10-minute refresh of Himawari-8 makes the plume more easily identifiable.  At 2100 UTC the plume is located from near Pahala, Hawaii to offshore areas east of the southern tip of the Big Island (faint pink-brown feature in the imagery).  The plume continued to slowly move south-southwest reaching offshore regions south-southwest of the southern tip of the Big Island by the end of the loop.  Identification of ash plumes in false color imagery is discussed in a separate post.  False color images, like the one shown below, are available in near real-time on the VOLcanic Cloud Analysis Toolkit website.

Meteorological Satellite Monitoring of Volcanic Ash Plumes from Kilauea Summit – May 15, 2018

The USGS Hawaiian Volcano Observatory raised the aviation alert level for Kilauea from Orange to Red on Tuesday, 15 May 2018.  The alert level was raised in response to an increase in the intensity of ash emissions from the summit crater.

False color imagery from the Himawari-8 satellite can be used to identify and track ash emissions.  One of the more relevant false color images is sometimes called the ash/dust RGB.  While interpretation of the ash/dust RGB is anything but straightforward, it is very useful for identifying ash and SO2 emissions under the right conditions.  SO2 emissions were discussed in a previous post.  In this post, 15 May 2018 ash emissions from Kilauea are highlighted.  A loop of Himawari-8 Ash/Dust RGB imagery from 1800 UTC 15 May to 0600 UTC 16 May 2018 is shown below.  There are two notable emissions of ash from Kilauea during this period.  The first ash emission, identifiable in loop below, occurs shortly after 1800 UTC.  A small, faint pink area emerges from the summit and is advected to the southwest, moving parallel to the Big Island coast from 1830 to 2130 UTC 15 May 2018, until the volcanic ash is obscured by higher meteorological clouds.  A second emission occurs in the 2100 – 2200 UTC timeframe, evidenced by the pink/gray color near the summit at 2110 UTC.  Meteorological clouds make the plume difficult to distinguish around 2200 UTC, but by 2300 UTC, the plume is again distinguishable as the pink/gray region south of the Kilauea summit extending over offshore regions.  The subtle ash feature slowly moves southwest, reaching the southern tip of the Big Island and offshore regions to the south by 0200 UTC 16 May 2018 and eventually becomes diffuse by 0400 UTC.

Volcanic ash tends to appear red/pink in this type of false color imagery largely because ash often absorbs radiation more strongly around 10-11 um than 12 um, while the opposite is true for meteorological clouds.  For Himawari-8, the 12-10 um brightness temperature difference (BTD) is placed on the red color gun.  Due to the aforementioned absorption effects, the 12 – 10 um BTD tends to be larger (compared to meteorological clouds) when ash is present.  The larger 12 – 10 um BTD results in a larger contribution from the red color gun in the RGB image.  It is important to keep in mind that there are numerous caveats associated with interpreting any satellite image, especially false color images that utilize several spectral channels, as the relationship between spectral channels is highly complicated.  False color images, like the one shown below, are available in near real-time on the VOLcanic Cloud Analysis Toolkit website.

Meteorological Satellite Monitoring of Kilauea – May 14, 2018

The eruption of the Kilauea volcano continued on May 14, 2018, resulting in lava flows and enhanced sulfur dioxide (SO2) emissions.  Clouds of SO2 can be hazardous to human health.  Geostationary (GEO) and low earth orbiting (LEO) meteorological satellites (e.g., GOES-15, Himawari-8, and S-NPP) can be useful for identifying positive surface temperature anomalies associated with the presence of lava and for identifying and characterizing clouds produced by volcanic eruptions.  GEO satellites provide more frequent images (every 10-15 minutes), while LEO satellites provide higher spatial resolution imagery and/or better coverage of the electromagnetic spectrum.  Thus, GEO and LEO meteorological satellites should be used in tandem for volcanic hazard monitoring (and most other satellite remote sensing problems).  The following satellite images from May 14, 2018 help illustrate how meteorological satellites can complement other sources  of information (e.g. ground and aircraft) for monitoring the ongoing eruption of Kilauea.

The S-NPP VIIRS image below is able to highlight some of the Kilauea SO2 emissions on May 14, 2018.  SO2 is a selective absorber/emitter of radiation, which means that only satellite measurements that capture radiation at certain wavelengths or frequencies will be sensitive to the presence of SO2.  One of the wavelength ranges that is sensitive to SO2 is centered around 8.6 um.  Satellite sensors such as the JPSS VIIRS and Himawari-8 AHI have such spectral bands.  In order to use the “8.6 um” channel for SO2 detection, you need to contrast the measurements with an appropriate channel that is not sensitive to SO2.  We typically do this by computing the brightness temperature difference (BTD) between 11 and 8.6 um channels.  Further value is added if the 11 and 8.6 um BTD is combined with other infrared measurements that are sensitive to ash in a false color (RGB) image.  As part of the VOLcanic Cloud Analysis Toolkit (VOLCAT), we utilize the 12-11 um BTD (red), 11 – 8.6 um (green), 11 um BT (blue) RGB, which is sometime called the ash/dust/SO2 RGB (the names of RGB’s can be confusing and are often somewhat arbitrary given that RGB’s are sensitive to so many different features).  In such an RGB, SO2 will appear green if present with little to no ash.  If a detectable amount of ash is also present, regions of ash + SO2 will appear yellow.  In reality, that interpretation greatly oversimplifies the problem, as there are numerous caveats and nuances.  For instance, not all SO2 is detectable using this method, especially if the SO2 is close to the ground and/or is present in low concentrations.  Meteorological clouds also limit detection.  Surface features and other non-volcanic cloud types can also sometimes look similar to SO2 in such false color imagery.  Nevertheless, a broad region of SO2 from the Kilauea eruption is identifiable in the VIIRS RGB image with good spatial detail.

S-NPP VIIRS Ash/SO2 false color (RGB) image from 1225 UTC on 14 May 2018. The SO2 plume is outlined in white.

Similar false color imagery can be created with measurements from the Himawari-8 AHI instrument.  The image below shows the same SO2 plume depicted by the VIIRS RGB image.  Since Himawari-8 is located at 140 East longitude, the view of Hawaii is at a very oblique angle (satellite view angles of near 75 degrees), thereby degrading the spatial resolution and introducing varying degrees of parallax.

Himawari-8 AHI Ash/SO2 false color (RGB) image from 1230 UTC on 14 May 2018. The SO2 plume is outlined in white.

In a single Himawari-8 image, it may be difficult to definitively pick out the SO2 plume due to the courser spatial resolution at larger viewing angles.  However, if one takes advantage of the high refresh rate of Himawari-8 (every 10 minutes) and creates an animation, the SO2 plume becomes easier to distinguish as one can see the motion of the plume from the East Rift Zone to the southwest (as the low level winds have been blowing from northeast to southwest on Monday).  The first loop shows the same domain as the images above and the second loop is more zoomed in.

At the start of the loops the SO2 plume is obscured by higher level cirrus clouds.  Around 0800 UTC 14 May 2018, the SO2 plume becomes evident to the southwest of the East Rift Zone (largely over the ocean).  Between roughly 8 and 12 UTC, the SO2 plume is clearly visible from the source region to near the southern tip of the Big Island.  Notice near the source region, around 12 UTC, stratus clouds are advected in from the northeast, which again obscure the SO2 plume.  These stratus clouds continue to push to the southwest, obscuring more of the plume.  Finally late in the loop, the stratus clouds dissipate and more SO2 is again seen in the imagery (generally due south of the Kilauea summit).

This example highlights how it is critical to use multiple satellite assets for volcanic cloud monitoring.  Specifically, the higher spatial resolution S-NPP VIIRS nicely complements Himawari-8 imagery, which is refreshed every 10 minutes.  Once the GOES-17 satellite is moved to the western position later in 2018, satellite-based volcanic cloud monitoring in Hawaii will be further improved.

When using satellite imagery to determine the location and movement of SO2 plumes it is important to remember the following:

  • SO2 can be obscured by meteorological clouds (this is seen early and again late in the above animations).
  • Non-related surface and cloud features can take on a similar appearance to SO2 in false color satellite imagery.
  • Infrared based SO2 detection can be ineffective when the SO2 is very close to the ground and/or is present in relatively low concentrations.
  • SO2, especially at large satellite viewing angles (like those for Hawaii using Himawari-8), can be difficult to unambiguously identify.  Use of higher spatial resolution LEO data, like from S-NPP VIIRS can add considerable value and enhance the value of GEO imagery.