HomeEventsPhD Defence Qiang Wang

PhD Defence Qiang Wang

soil moisture estimation by synergetic use of aquarius active and passive l-band microwave observations

Qiang Wang is a PhD student in the department of Water Resources. His supervisor is prof.dr. Z. Su from the faculty of Geo-Information Science and Earth Observation.

Soil moisture is a key variable in the water cycle and it plays an important role in the exchanges of energy, water and gasses between land surface and atmosphere. The availability of soil moisture information leads to a better understanding of biology, hydrology, meteorology and climatology. The most suitable frequency band to retrieve soil moisture data is considered to be the L band, since it can partially penetrate vegetation and is marginally affected by clouds. Numerous studies show that both active and passive microwave observations are sensitive to soil moisture and can be used to retrieve soil moisture information. However, vegetation influence and roughness effect form the main obstacles for soil moisture retrieval in, respectively, the passive and active configuration. As passive and active microwave observations differ in sensitivity to the relevant factors, combined use of both these observations is beneficial when studying soil moisture.

This dissertation contributes to a better estimation of soil moisture through synergetic use of active and passive observations from Aquarius, which is the first satellite to have both an L-band radiometer and a scatterometer onboard. The Tibetan Plateau has been selected as study area since it covers a large area with different climates, including humid, semi-arid and arid regions from east to west. Moreover, large amounts of in-situ data have been recorded across this area since 2008, providing ancillary data for the validation of  soil moisture estimations. 

Aquarius observations are firstly analyzed for one of the Tibetan Plateau observatory sites, Maqu, in chapter 3. This confirms that both the Aquarius radiometer and scatterometer observations show a response to soil moisture variation across Maqu, especially when the soil moisture is less than 0.30 m3 m-3. Moreover, the Microwave Polarization Difference Index (MPDI) is investigated and shows that the derived vegetation optical depth (τ) is in line with the vegetation dynamics. However, even though the Radar Vegetation Index (RVI) might capture the seasonal dynamic change of vegetation, the accuracy is insufficient from a meaningful signal-to-noise point of view.

In chapter 4, a discrete electromagnetic model developed by the Tor Vergata University of Rome (hereafter, Tor Vergata-discrete electromagnetic model, TV-DEM) is used to simulate both active and passive L-band responses and then compared with Aquarius observations from a view angle of 28.7° over Maqu, using a single set of input parameters. Litter biomass, litter moisture, plant moisture and standard deviation of height variations (s) in the TV-DEM are calibrated by minimizing the difference between the observed and simulated emissivity and backscattering coefficient from the warm seasons of 2012 and 2013. The calibrated parameters are used to reproduce the brightness temperature and backscattering coefficient in the warm seasons of 2014 and 2015, to validate the model’s performance. Furthermore, the soil moisture retrieval based on the TV-DEM is carried out and compared with the current single channel algorithm (SCA) retrieved soil moisture. Results present an unbiased root means square difference (ubRMSD) of 0.021 and 0.026 m3 m-3, as well as a coefficient of determination of 0.76 and 0.79 (-), for TV-DEM based soil moisture retrieval and SCA retrieval, respectively, with respect to the in-situ measurements.

Chapter 5 follows up on the results of chapter 4 and introduces an algorithm for retrieving soil moisture at plateau scale, combining the use of Aquarius active and passive L-band observations. Look-Up-Tables (LUTs) are generated through forward modeling of the TV-DEM by varying LAI and soil moisture while keeping litter biomass, litter moisture, plant moisture and surface roughness the same as the calibrated parameters. By searching for the minimum squared difference between the emissivity and backscattering coefficient observed by Aquarius and the simulations included in the LUT, the corresponding soil moisture is derived. The soil moisture retrievals are assessed at footprint scale with respect to the in-situ measurements collected at three regional scale networks across the Tibetan Plateau. An inter-comparison is also conducted among the TV-DEM retrieval, passive-only Aquarius, Metop-A Advanced SCATterometer (ASCAT) soil moisture L2 product, and the soil moisture of global atmospheric reanalysis (ERA-Interim) generated by the European Center for Medium-Range Weather Forecasts (ECMWF) on a point-scale. Furthermore, the spatial distribution of these four soil moisture retrievals is verified, alongside complementary rainfall (Climate Hazards Group Infrared Precipitation with Station data (CHIRPS)) and evapotranspiration (Surface Energy Balance System (SEBS)) products.

In conclusion, this dissertation confirms that soil moisture retrieval through the synergetic use of passive and active observations in the TV-DEM framework is comparable with those by the passive only Aquarius operational product, the C-band ASCAT product and the re-analysis ECMWF soil moisture product. Moreover, TV-DEM soil moisture retrieval scheme can be applied at plateau scale and the TV-DEM retrieval can capture the spatial distribution of soil moisture at plateau scale, opening up new opportunities in general for hydrology, meteorology and climatology.