Validation of the Level 1c and Level 2 SMOS Products with Airborne and Ground-based Observations

In the Austral summer and winter periods of 2010, two large field campaigns (the Australian Airborne Cal/val Experiments for SMOS (AACES)) were undertaken for the validation of the Level 1c brightness temperature and Level 2 soil moisture products of the ESA-led Soil Moisture and Ocean Salinity (SMOS) mission, which constitutes the first of a series of future soil moisture satellite missions. SMOS is also the first polar-orbiting satellite operating at a frequency of 1.4GHz (L-band) and even more importantly is the first 2-dimensional interferometric radiometer used for Earth observation purposes. Consequently, extensive and detailed field campaigns for its product validation are required. The field site of this study lies within the Murrumbidgee River catchment of southern New South Wales, in south-eastern Australia. The advantage of this particular catchment is its wide variety of surface conditions, ranging from flat, semi-arid areas in the west (mainly used for dryland farming) to the alpine ranges in the east near Canberra. Moreover, the seasonal differences and sporadic rain events during the campaigns meant that a wide range of conditions were found, resulting in a data set ranging from hot and dry to wet and cold. A total of 50,000km was covered, using an airborne sensor platform, operating over the same microwave wavelength as SMOS. The data collected with the airborne platform included passive L-band microwave observations, as well as measurements in the thermal infrared and eight spectral bands. Moreover, the catchment contains over 60 permanent soil moisture monitoring stations which were supplemented by spatially high-resolution soil moisture measurements across 20 focus farms on selected overpass days. In this paper, we discuss the quality of both brightness temperature data sets, as well as compare the ground station data against the SMOS Level 2 soil moisture product. It is found that the overall brightness temperature observations of SMOS tend to be systematically ~10K warmer than the airborne observations, both during the winter and summer periods, as well as across different vegetation types, which suggests that there is a persistent bias in one of the data sets. Similarly, the SMOS soil moisture products consistently underestimate the soil moisture observed on the ground. Nevertheless, a significant improvement in the accuracy of the soil moisture product has been achieved between the latest two Level 2 versions and it is expected that the accuracy will again increase when the data from all validation campaigns have been assessed.