Volcano applications commonly involve sizeable departures from the reference pressure and temperature of COSPEC calibration cells. Analysis shows that COSPEC SO2 column abundances and derived mass emission rates are independent of pressure and temperature, and thus unaffected by elevation effects related to deviations from calibration cell reference state. However, path-length concentrations are pressure and temperature dependent. Since COSPEC path-length concentration data assume the reference pressure and temperature of calibration cells, they can lead to large errors when used to calculate SO2 mixing ratios of volcanic plumes. Correction factors for COSPEC path-length concentrations become significant (c.10%) at elevations of about 1 km (e.g. Kilauea volcano) and rise rapidly to c.80% at 6 km (e.g. Cotopaxi volcano). Calculating SO2 mixing ratios for volcanic plumes directly from COSPEC path-length concentrations always gives low results. Corrections can substantially increase mixing ratios; for example, corrections increase SO2 ppm concentrations reported for the Mount St Helens, Colima, and Erebus plumes by 25-50%. Several arguments suggest it would be advantageous to calibrate COSPEC measurements in column abundance units rather than path-length concentration units.