Advancing the understanding of active microwave remote sensing of soil moisture and vegetation

Abstract

Active microwave remote sensing techniques provide a means for the monitoring of biogeophysical variables over land, independent of weather and cloud conditions and daylight. Several missions are in place nowadays which provide regular observations of the entire Earth surface. Observations provided by these sensors have for example been used for the retrieval of soil moisture (SM) and vegetation optical depth, a measure of canopy water content, density and structure. Some of these datasets are available publicly and on an operational basis. Despite the long history of active microwave remote sensing research, going back into the 1960s, there is a constant need to extend the understanding of how active microwave sensors perceive the land surface on the respective spatial scale of the satellite observations.The aim of this thesis was to study the multi-angular backscatter signal (σo) observed by the Advanced Scatterometer (ASCAT) sensor over the land surface. The focus was on the backscatter dependence on the incidence angle (σ’), as this relationship is crucial for the separation of SM and vegetation effects on the observed signal. Thereby, I aimed at advancing the understanding of how SM and vegetation dynamics influence σo and σ’ on the relatively coarse spatial scale of the ASCAT footprint. The main objectives can be summed up as follows: i) increase the understanding of the ASCAT back scatterincidence-angle relationship, in particular the first derivative σ’, ii) investigate the potential of a regional adjustment of parameter values for SM and vegetation optical depth retrieval, iii) improve theunderstanding of structural effects of vegetation canopies on σ’, and iv) reassess the assumption that σ’ is not or only weakly affected by SM.The conducted research highlighted the great potential of the coarse-scale ASCAT sensor for the retrieval of biogeophysical variables such as SM and vegetation dynamics. One main new finding was that ASCAT is highly sensitive to the water uptake of deciduous broadleaf trees in early spring, allowingfor the monitoring of spring reactivation in deciduous forests across wide regions, potentially even ona global level. Thanks to the increasing temporal coverage, (ASCAT) back scatter time series may be exploited for the study of growing season shifts as a reaction to climate change. The study clearly showed that canopy structure can have large effects on ASCAT observations, even if the responsible vegetation type makes up only a small fraction of the entire footprint.The thesis also revealed potentials for improvements in existing retrieval algorithms, such as the benefits of a stronger vegetation correction for the retrieval of SM in temperate-climate, agricultural regions. Moreover, it was shown that despite the clear and dominant control of σ’ by vegetation dynamics, there are short-term secondary effects in σ’ caused by SM, which need to be taken into account when interpreting σ’ time series or applying σ’ as vegetation dynamics indicator.The results of the thesis show that future studies of ASCAT σo and its dependence on the incidenceangle should be set up as broad as possible in order to take into account the numerous variables andprocesses that ASCAT is sensitive to, including combined effects that might cancel each other out orreinforce each other. However, detailed studies of selected processes will always be necessary in orderto understand how individual components contribute to the signal.High resolution backscatter datasets, for example provided by Sentinel-1, and enhanced observation techniques, for example fore seen for the upcoming Metop-SG missions, ensure the availability of longbackscatter time series, and open up new possibilities for investigating, understanding and monitoring backscatter from land surfaces. This thesis shall contribute to these efforts and support the way forward by shedding light on topics that were not studied in detail previously.