Mixed pixel retrieval of soil moisture from L-band passive microwave observations
2017-02-27T03:46:23Z (GMT) by
Soil moisture plays a key role in the water, energy, and carbon exchanges at the interface between the atmosphere and earth surface. Its spatial and temporal distributions at regional and global scales are required by many disciplines, including hydrology, meteorology, and agriculture. During the last three decades, passive microwave remote sensing has been widely acknowledged as the most promising technique to measure the spatial distribution of near surface (top few centimetre) soil moisture, due to its direct relationship to the soil dielectric constant, its ability to penetrate clouds, and its reduced sensitivity to vegetation canopy and surface roughness. Therefore, the first two space missions dedicated to soil moisture, the Europe Space Agency (ESA)’s Soil Moisture and Ocean Salinity (SMOS) mission and the National Aeronautics and Space Administration (NASA)’s Soil Moisture Active Passive (SMAP) mission, are based on L-band (~1.4 GHz) passive microwave observations every two to three days. Using radiative transfer models, brightness temperature observations are used to estimate water content of the top approximately five centimetres soil with a target accuracy of ~0.04 m3/m3. Based on the current level of antenna technology, the best spatial resolution that can be achieved at L-band by both the SMOS and SMAP radiometer approaches is approximately 40 km. At such a coarse scale, non-soil targets such as surface rock, urban areas, and standing water are present within many SMOS and SMAP pixels across the world, potentially confounding the radiometric observations, and in turn degrading the soil moisture retrieval if not accounted for their contribution. Consequently, the objective of thesis is to determine the impact of land surface heterogeneity conditions on L-band passive microwave satellite footprints using airborne passive microwave brightness temperature observations collected during five Australian airborne field campaigns conduced within the past eight years. Using the Polarimetric L-band Multi-beam Radiometer (PLMR) mounted on a scientific aircraft, brightness temperature of the SMOS and SMAP sized study areas were measured at viewing angles of 7°, 21.5°, and 38.5°. Due to the strong angular dependency of brightness temperature, the multi-angular PLMR observations need to be normalised to a reference angle. The angle 38.5° was chosen to closely replicate the fixed incidence angle of SMAP. In this thesis the Cumulative Distribution Function (CDF) based method is developed for incidence angle normalisation by matching the CDF of observations for each non-reference angle. Subsequently, the effects of surface rock, urban areas, and standing water were explored using the incidence-angle-normalised airborne brightness temperature observations and coincident ground sampling data. The brightness temperature difference between that of the mixed pixel and its soil only equivalent was defined as the non-soil targets induced brightness temperature contribution that will potentially lead to a soil moisture retrieval error if not accounted for. It was found that about 13% of SMOS and SMAP pixels on the world’s land mass may be adversely affected by surface rock, urban areas, or standing water. However, such pixels are not uniformly distributed or coincident, meaning that such factors may be particularly important in some parts of the world.