This thesis describes an investigation of, and improvements to, the microphysical modelling of aerosols in the Oxford-Rutherford Appleton Laboratory Aerosol and Clouds retrieval (ORAC), which is used to obtain aerosol properties from measurements by the Advanced Along Track Scanning Radiometer (AATSR). Modelling decisions determine the light scattering properties of the aerosol classes which in turn alter the retrieved aerosol properties: aerosol optical depth, and effective radius. The maritime, mineral dust, urban, and biomass burning aerosol classes were first investigated, and then improvements implemented. Major additions to the scheme include the ability to model non-spherical dust as spheroids, soot as fractal aggregates, and to coat spherical particles with an extra layer of differing refractive index (whose thickness can be modified by ambient relative humidity where necessary). Output from aerosol retrievals containing these new models is presented. Modelling of marine aerosol was found to be adequate, but an improvement in the relative humidity assumptions led to an average 5 % increase in aerosol optical depth (AOD). Modelling of mineral dust aerosols has been dramatically altered by the addition of non-spherical dust and hygroscopic particles, leading to increases in measured AOD of over 100 % during dust events, compared to the previous model. Measurement of biomass burning aerosol has been tested with an `ageing' aerosol scheme, leading to increases in over-land measured AOD of 0.14 (~50 % increase). With such significant changes in AOD, representation of aerosol light scattering properties is seen to be important factor in the accuracy of the ORAC scheme. Finally, a method of optimising the placement of detectors in an aerosol measurement device is presented.