Improved estimation of temporal dynamics in the ASCAT backscatter-incidence angle relation using regularization

Frantzen, P. ORCID: https://orcid.org/0009-0005-6959-4788, Steele-Dunne, S. C., Quaife, T. ORCID: https://orcid.org/0000-0001-6896-4613, Vreugdenhil, M. ORCID: https://orcid.org/0000-0003-4913-0255, Hahn, S. ORCID: https://orcid.org/0000-0002-4855-6273 and Wagner, W. ORCID: https://orcid.org/0000-0001-7704-6857 (2025) Improved estimation of temporal dynamics in the ASCAT backscatter-incidence angle relation using regularization. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 18. pp. 13457-13471. ISSN 2151-1535 doi: 10.1109/JSTARS.2025.3572306

Abstract/Summary

The relation between microwave backscatter and incidence angle estimated from observations of the Advanced Scatterometer (ASCAT) onboard the Metop satellites contains valuable information on the dynamics of vegetation water content and structure. The relation between backscatter and incidence angle (parameterized using so-called slope and curvature parameter) has been related to vegetation water dynamics in studies on the North American Grasslands and the Cerrado Savannah. The current approach to estimate time series of the slope and curvature parameters involves a kernel smoother, weighing observations according to their temporal distance to the day of interest. While this approach provides a robust representation of backscatter-incidence angle relation over longer time scales, it does not accurately capture the timing of short-term changes. To further improve the correspondence between backscatter-incidence angle relation and vegetation water dynamics, the timing of short-term changes should be preserved in the estimation of slope and curvature. This would allow slope and curvature to be reconciled with independent estimates of biogeophysical variables, and allow us to isolate high-frequency variations due to, for example, intercepted precipitation or soil moisture. Here, an alternative method is introduced to estimate the ASCAT backscatter-incidence angle relation using temporally constrained least squares. While the proposed method yields similar performance to the kernel smoother in aggregated statistics, this method retains the timing of short-term changes.

Item Type	Article
URI	https://centaur.reading.ac.uk/id/eprint/122828
Identification Number/DOI	10.1109/JSTARS.2025.3572306
Refereed	Yes
Divisions	Science > School of Mathematical, Physical and Computational Sciences > National Centre for Earth Observation (NCEO) Science > School of Mathematical, Physical and Computational Sciences > Department of Meteorology
Publisher	IEEE
Download/View statistics	View download statistics for this item