Characterizing atmospheric aerosols is key to understanding their influence on climate through their direct and indirect radiative forcing. This requires long-term global coverage, at high spatial (åisebox-0.5ex~km) and temporal (i̊sebox-0.5ex~days) resolution, which can only be provided by satellite remote sensing. Aerosol load and properties such as particle size, shape and chemical composition can be derived from multi-wavelength radiance and polarization measurements of sunlight that is scattered by the Earth’s atmosphere at different angles. The required polarimetric accuracy of rs̊ebox-0.5ex~10(̂-3) is very challenging, particularly since the instrument is located on a rapidly moving platform. Our Spectropolarimeter for Planetary EXploration (SPEX) is based on a novel, snapshot spectral modulator, with the intrinsic ability to measure polarization at high accuracy. It exhibits minimal instrumental polarization and is completely solid-state and passive. An athermal set of birefringent crystals in front of an analyzer encodes the incoming linear polarization into a sinusoidal modulation in the intensity spectrum. Moreover, a dual beam implementation yields redundancy that allows for a mutual correction in both the spectrally and spatially modulated data to increase the measurement accuracy. A partially polarized calibration stimulus has been developed, consisting of a carefully depolarized source followed by tilted glass plates to induce polarization in a controlled way. Preliminary calibration measurements show an accuracy of SPEX of well below 10-̂3), with a sensitivity limit of 2*10^4̂). We demonstrate the potential of the SPEX concept by presenting retrievals of aerosol properties based on clear sky measurements using a prototype satellite instrument and a dedicated ground-based SPEX. The retrieval algorithm, originally designed for POLDER data, performs iterative fitting of aerosol properties and surface albedo, where the initial guess is provided by a look-up table. The retrieved aerosol properties, including aerosol optical thickness, single scattering albedo, size distribution and complex refractive index, will be compared with the on-site AERONET sun-photometer, lidar, particle counter and sizer, and PM10 and PM2.5 monitoring instruments. Retrievals of the aerosol layer height based on polarization measurements in the O2A absorption band will be compared with lidar profiles. Furthermore, the possibility of enhancing the retrieval accuracy by replacing the look-up table with a neural network based initial guess will be discussed, using retrievals from simulated ground-based data.