Volcanic ash poses a potentially serious hazard to high altitude jet aircraft, and because ash clouds are difficult to see and can persist for thousands of kilometers, they must be monitored by the use of satellite remote sensing techniques. The two-band "split window" product, based on the brightness temperature difference between InfraRed (IR) window channels at 11 um (micrometer) and 12 um wavelengths has been used extensively for this purpose. However, within an hour or two after a volcanic eruption, the split window imagery is often not effective due to the high percentage of water droplets within the eruption cloud. In tropical regions, the deep moist atmosphere can also mask the presence of volcanic ash, or prevent clear discrimination of the ash from meteorological clouds. This is especially the case for eruptions of smaller volcanoes that emit ash into the lower and middle portion of the tropical atmosphere, such as the recently active Soufriere Hills volcano on the Caribbean Island of Montserrat. Despite their relatively low altitude, these ash clouds can be hazardous to jet aircraft during climb or descent. For these reasons, there is a need to improve the detection of volcanic ash using remotely sensed IR data from geostationary or polar satellites.
Preliminary work is underway at the Cooperative Institute for Research in the Atmophere (CIRA) and the National Environmental Satellite, Data, and Information Service (NESDIS) Office of Research and Applications to develop improved satellite techniques for volcanic ash detection. One method involves the incorporation of a Geostationary Operational Environmental Satellite (GOES) shortwave IR channel at 3.9 um into the split window product. At 3.9 um, there is some absorption and daytime reflectance in areas of volcanic ash that result in a very warm signal that causes the ash to stand out from the background scene. An experimental product has been developed that merges the split window image with a difference of the GOES 3.9 um and 11 um IR bands. The resulting product has been generated hourly for selected areas, and made available on an Internet Web site. The experimental technique is also undergoing operational evaluation at the NESDIS Satellite Analysis Branch at the Washington Volcanic Ash Advisory Center (VAAC), responsible for tracking volcanic ash over an extensive portion of the mid-latitudes.
Additional work has involved the use of Principle Component Imagery (PCI), a form of eigenvector statistical analysis, to screen GOES imager and sounder channels. PCI images help to reduce the amount of redundant information in the various GOES channels and highlight the less obvious meteorological features in the atmosphere. The first PCI contains information common to all channels being screened, such as clouds. The second and subsequent PCIs contain information not explained by previous PCI components, such as channel differences. Some examples of these images will be shown for known volcanic eruptions
The 8th Conference on Aviation, Range, and Aerospace Meteorology