Agricultural drought, characterised by continuous soil moisture (SM) deficiency due to insufficient surface water supply, is marked by a persistence of high land surface temperature (LST). With the intensification of global warming, the frequency and severity of extreme events have escalated in Australia’s dryland ecosystems, necessitating accurate quantification of their extent and duration. Emerging remote sensing techniques offer a valuable opportunity to observe the spatiotemporal dynamics of agricultural drought; however, these observations are often constrained by sensor design limitations and retrieval uncertainties. To address these challenges, this thesis explores the utilisation of multi-source data integration and algorithmic refinements to enhance the spatiotemporal predictability of agricultural drought. Specifically, the thesis addressed the following research questions:

1. How can we accurately integrate multi-source satellite LST data, accounting for their systematic discrepancies, to obtain high spatiotemporal resolution information on surface heating dynamics?
2. Can data-driven approaches, e.g. machine learning, combine multiples sources of data, including LST, effectively to infer fine-scale SM information? Given their “black box” nature, how do we quantify their spatial applicability?
3. Can the integration of SM, LST and vegetation information provide enhanced warning capability for agricultural drought, spanning from continental to regional scales?

About the Speaker

Yi Yu is a PhD student under an academic collaboration between the Australian National University (ANU) and the Commonwealth Scientific and Industrial Research Organisation (CSIRO). Before commencing his PhD, Yi received a B.Mgmt. in Land Resources Management from Southwest University in 2018 and an M.Sc. in Environment (Advanced) from ANU in 2020. He is interested in using data-driven methods to better understand fine-scale land-atmosphere interactions in the context of climate extremes (e.g., drought and heatwave). His PhD project explored the role of land surface temperature in improving spatiotemporal predictability of soil moisture-related agricultural drought. He also worked on an ANU-CSIRO Himawari-8 project which aims at developing best-practice geostationary data products for enhanced sub-daily monitoring of Australia’s ecosystems.