Presented at the 2013 SMOS-Aquarius Science WorkshopPrevious high frequency microwave radiometers such as TOPEX, Jason-1, and Jason-2 have all utilized external brightness temperature references as calibration targets. Global mean brightness temperature measurements (or vicarious cold brightness temperatures) from the ocean are used to calibrate the instrument at one end of the temperature spectrum and hot targets such as the Amazonian rainforests are used at the other end of the temperature spectrum. The external calibration sources provide a wide dynamic range to back out gain and offset variations from the instrument. Even though the Aquarius mission science measurements are made over a very small dynamic range, it is important to correctly identify the contribution of individual drift mechanisms. SMOS and NASA's upcoming SMAP mission's science measurements cover a wide brightness temperature range, necessitating gain and offset corrections.The external calibration sources developed for high frequencies cannot be directly applied at L-band. This has required the development of L-band brightness temperature calibration models over Antarctica and depolarized regions of the Amazon. The Antarctic region is a very attractive stable calibration source due to the deep penetration depth at L-band and potential annual stability of less than 0.2K. We have developed an advanced ice microwave emission model to predict brightness temperatures at L-band. This model includes in-situ surface data from AWS (Antarctic Weather Stations) and is coupled with a heat transport model to determine ice temperature profile. The MEMLS radiative transfer model is used to determine L-band Tb, constrained by other ancillary measurements such as AMSR-E or WindSat. We have used Aquarius measurements to determine a large homogeneous and temporally stable region of Antarctica to use as a stable calibration reference. Measurements from the Antarctic L-band modeling have aided in the determination of a gain-based drift in Aquarius.At L-band forest regions are expected to be slightly transparent. We have developed an Amazon rainforest calibration model that selects depolarized heavily vegetated areas. WindSat and TMI data is used to constrain the Amazon model and further uncertainties due to canopy temperature are removed by comparing inter-channel differences only. The Amazon model coupled with the Antarctic and Ocean model successfully identified the offset-component of drift in SMOS.A brief description of the drift observed by Aquarius channels will be presented. A description of the on-Earth reference models at L-band will be given followed by a discussion of the results.