In an effort to better represent land surface processes in environmental prediction systems (weather, hydrology, land surface), a new version of the Canadian Land Data Assimilation System (CaLDAS) is currently being developed at Environment Canada (EC). In its current operational state, CaLDAS assimilates observations from surface stations to specify initial conditions of snow, soil moisture, and surface temperature, using simple techniques mostly based on optimal interpolation. Our main focus in current and upcoming projects is to include space-based remote sensing observations in CaLDAS, using more sophisticated methods based on variational data assimilation or ensemble Kalman filtering. As part of a collaborative project between the Finnish Meteorological Institute (FMI) and EC, MODIS-derived fractional snow covered area and snow water equivalent retrievals from AMSR-E will soon be assimilated in CaLDAS. For soil moisture, a project to assimilate L-band brightness temperatures from the European SMOS sensor is also underway. And it is planned to include radiometric and radar observations from NASA's Soil Moisture Active and Passive (SMAP) mission. Progress has also been achieved to use MODIS data in order to specify vegetation leaf area index, a quantity that is essential for the assimilation of other land surface variables such as snow and soil moisture. An overview of the CaLDAS project will be presented at the conference.