Abstract. This study demonstrates a method of retrieving the mass column loading and cloud-top pressure of a volcanic ash cloud, together with the effective radius and spread of the ash particle size distribution, as well as the cloud top pressure of any underlying water cloud, using an optimal estimation technique applied to Infrared Atmospheric Sounding Interferometer data. Two shapes of particle size distribution are considered, a log-normal and a gamma distribution. Results show that it is viable to retrieve a measure of the size distribution spread, namely the geometric standard deviation, when a log-normal distribution is assumed, whereas this is not the case for an assumed gamma distribution in terms of its effective variance. The volcanic conditions under which the method works well are discussed, as are its shortcomings. The method is applied to two volcanic eruptions: Eyjafjallajökull, Iceland using data from 6th May 2010 and Kasatochi, Alaska using data from 8th August 2008. The results show that the retrieved geometric standard deviation of these ash clouds is spatially variable, and is generally similar to what is assumed in many passive infrared remote sensing techniques. An abrupt change in the retrieved geometric standard deviation has been observed for the Eyjafjallajökull eruption along the trajectory of the ash cloud, and possible explanations for this are discussed.