Quantification of volcanogenic water vapor using the atmospheric infrared sounder (AIRS)

Volcanic eruption columns transport magmatic and entrained atmospheric water vapor to the relatively dry stratosphere. The water vapor signal also interferes with the signatures of other gases and aerosols in volcanic clouds, causing under- or over-estimations of the species. In spite of this, water vapor, the most pervasive volcanic gas has largely been ignored due to its natural abundance in the atmosphere, which makes its detection and retrieval difficult. A retrieval algorithm to quantifying water vapor in upper tropospheric/lower stratospheric volcanic clouds has been developed using the Atmospheric Infrared Sounder (AIRS). Using a radiative transfer forward model, we are able to simulate atmospheres with varying amounts of water vapor injected at varying heights. Using a series of fitting algorithms, the retrieval compares model spectra to AIRS radiance data, where differences in spectra are used to quantify the amount of water vapor in the volcanic plume. Total water mass is calculated by multiplying individual pixel area, cloud thickness, and the retrieved water density in the cloud.As the water vapor mass retrievals are dependent on user parameters, it is imperative that their impact on the retrievals is known. Five variables (number of atmospheric layers, maximum water density, cloud top height, cloud thickness, and atmospheric profiles) were examined to determine the sensitivity of the algorithms. These results suggest that the number of atmospheric layers and water density are intrinsically linked the retrievals are broadly independent of cloud top height changing thickness produces a general increasing trend, and using a different atmospheric profile is most significant between continental and marine climates. The impact of the addition of noise to the signal was also examined and shows that the retrieval can handle up to 10% and still produce results within 8% of the original value. Peculiarities with the retrievals themselves were also assessed and it was determined that the coefficient used in the DOAS_PROD retrieval is location specific. These inspections help to better understand the retrievals and suggest new avenues to explore for future work.The water vapor retrieval is applied to recent high latitude (above 50° N) eruptions including Shiveluch (29 March 2007), Okmok (12 July 2008), and Kasatochi (8 August 2008). Up to four days of imagery from the Atmospheric Infrared Sounder (AIRS) were downloaded for each eruption and investigated for eruption clouds. The eruption cloud from Shiveluch is detected in one image, immediately following the eruption, while the Okmok and Kasatochi clouds were observed in seven and ten images, respectively. The length of detection is dependent on the explosivity of the eruption, interaction with surficial or ground water, and the atmospheric conditions. A significant limitation of the water vapor retrieval is the need for a priori knowledge of an eruption. This is highlighted in images from 23 and 24 March 2007 where passive degassing was reported and a cloud was located in the vicinity of the volcano. However, passive degassing would not account for a cloud above 5km, the absolute minimum altitude at which water can be detected by this retrieval thus resulting in a false positive. The masses of retrieved water from Okmok and Kasatochi are on the same order of magnitude as the retrieved sulfur dioxide from the Ozone Mapping Instrument (OMI), consistent with known SO2:H2O ratios.