 |
| Titel |
DOAS evaluation of volcanic SO2 using a modeled background spectrum: Examples from the NOVAC stations at Nevado del Ruiz (Colombia) and Tungurahua (Ecuador) |
| VerfasserIn |
Peter Lübcke, Johannes Lampel, Nicole Bobrowski, Santiago Arellano, Bo Galle, Gustavo Garzón, Silvana Hidalgo, Leif Vogel, Simon Warnach, Ulrich Platt |
| Konferenz |
EGU General Assembly 2015
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 17 (2015) |
| Datensatznummer |
250102479
|
| Publikation (Nr.) |
 EGU/EGU2015-1803.pdf |
|
|
|
|
|
| Zusammenfassung |
| SO2 emission rates are monitored using Differential Optical Absorption Spectroscopy
(DOAS) in the UV at an increasing number of volcano observatories. The Network for
Observation of Volcanic and Atmospheric Change (NOVAC) has currently installed 80
scanning DOAS instruments at 30 volcanoes world-wide. One important question for the
evaluation of spectra using DOAS is the availability of background spectra that are not
influenced by volcanic gas emissions. An SO2 contaminated background spectrum would
lead to a negative offset of the retrieved SO2 column densities, and thus to an underestimation
of the volcanic SO2 emission rate. In NOVAC this problem is approached by performing a
scan, e.g. through a plane from one horizon to the other horizon, and defining the average of
the 20% spectra with the lowest SO2 content as the zero-baseline value, which is assumed to
be gas free. To verify this assumption we revisit the idea of evaluating spectra using the
DOAS method with a modeled background spectrum based on a high-resolution
solar atlas. One challenge when evaluating spectra with a modeled background
spectrum is properly accounting for instrumental effects that are usually removed when
calculating the measured optical density relative to a measured background spectrum. We
present our approach to handle these instrumental effects, showing that we gain a
similar fit quality to the method using a measured reference spectrum. For example,
wavelength dependent structures in the spectrum due to the spectrometer (e.g., quantum
efficiency of the detector and grating efficiency) were identified with help of a principal
component analysis of an SO2 free subset of the residual spectra. These structures were
included in a second iteration of the fit in order to improve the evaluation. We further
discuss influences like strong ozone absorption and the instrument temperature on the
quality of the SO2 fit using a modeled background spectrum. The new evaluation
scheme was applied to data from Nevado del Ruiz (Colombia) and Tungurahua
(Ecuador). We investigated how often and under which circumstances SO2 contaminated
background spectra occur and how big the effect on the SO2 emission rates is. At
Nevado del Ruiz, although characterized currently by a large emission and therefore
broad plume the NOVAC stations are installed at a distance of only 2-5 km from the
volcano. At this volcano up to 30% of the scans underestimate the SO2 emission
rate, with stronger underestimation occurring at low wind speeds (below 5 m/s). At
Tungurahua, where the stations are installed at more than 5 km distance from the volcano
preliminary results indicate that fewer scans are influenced by contaminated background
spectra. |
|
|
|
|
|
|