Open Access

Christian von Savigny, Anna Lange, Anne Hemkendreis, Christoph G. Hoffmann, Alexei Rozanov

• The idea of estimating stratospheric aerosol optical thickness from the twilight colours in historic paintings – particularly under conditions of volcanically enhanced stratospheric aerosol loading – is very tantalizing because it would provide information on the stratospheric aerosol loading over a period of several centuries. This idea has in fact been applied in a few studies in order to provide quantitative estimates of the aerosol optical depth after some of the major volcanic eruptions that occurred during the past 500 years. In this study we critically review this approach and come to the conclusion that the uncertainties in the estimated aerosol optical depths are so large that the values have to be considered questionable. We show that several auxiliary parameters – which are typically poorly known for historic eruptions – can have a similar effect on the red–green colour ratio as a change in optical depth typically associated with eruptions such as, for example, Tambora in 1815 or Krakatoa in 1883. Among the effects considered here, uncertainties in the aerosol particle size distribution have the largest impact on the colour ratios and hence the aerosol optical depth estimate. For solar zenith angles exceeding 80∘, uncertainties in the stratospheric ozone amount can also have a significant impact on the colour ratios. In addition, for solar zenith angles exceeding 90∘ the colour ratios exhibit a dramatic dependence on solar zenith angle, rendering the estimation of aerosol optical depth highly challenging. A quantitative determination of the aerosol optical depth may be possible for individual paintings for which all relevant parameters are sufficiently well constrained in order to reduce the related errors.