Towards resolution enhancement of P-band brightness temperature data using passive-passive downscaling with L-band radiometer data

Date

2025

Authors

White-Murillo, L.F.
Walker, J.P.
Ye, N.
Hills, J.
Wu, X.
Zhou, L.
Xiong, Z.
Zhu, L.
Ng, B.
Moghaddam, M.

Editors

Advisors

Journal Title

Journal ISSN

Volume Title

Type:

Journal article

Citation

Remote Sensing of Environment, 2025; 323:114737-1-114737-16

Statement of Responsibility

Luisa F. White-Murillo, Jeffrey P. Walker, Nan Ye, James Hills, Xiaoling Wu, Lixiaozhou Zhou, Ziwei Xiong, Liujun Zhu, Brian Ng, Mahta Moghaddam, Simon Yueh

Conference Name

DOI

10.1016/j.rse.2025.114737

Abstract

Current space-borne L-band radiometers have been routinely providing global soil moisture maps every 2 to 3 days for more than 15 years. However, they can only estimate soil moisture in the top 5 cm of the soil, which limits its usefulness in applications that need deep layer soil moisture. Fortuitously, recent work has shown that there is a potential to retrieve root zone soil moisture (RZSM) when combining P-band radiometer data with Lband radiometer data at the same spatial resolution. Nevertheless, there are antenna size limitations that currently restrict L-band radiometer data to a resolution of 40 km. Accordingly, if the same antenna is used for a joint P-band (750 MHz) and L-band (1.4GHz) satellite mission, the footprint size at P-band will be double that attained at L-band, limiting its usefulness in applications and making joint interpretation difficult. It is therefore essential to downscale the P-band radiometer measurements to at least the same resolution as the L-band measurements. Consequently, this paper has explored three alternative passive-passive P-band downscaling algorithms using the spatial information that exists in the L-band radiometer data, to achieve P-band brightness temperature (Tb) information at the same resolution as the L-band passive measurements. Analysis was conducted using data from three airborne campaigns in south-eastern Australia, with resolutions ranging from 36 km to 200 m under a range of moisture conditions and spatial characteristics. The three alternate downscaling algorithms used in this paper, designated as Algorithm A, B, and C, were not only compared to determine which generated the best performance, but were also analysed according to different land cover types and seasons. The results demonstrated that the Smoothing Filter-based Intensity Modulation (SFIM) method, referred to as Algorithm A, outperformed the others, with a median root mean square error (RMSE) compared to the original observations of 3 K when downscaling to half of the original spatial resolution, increasing to around 8 K when downscaling to 36 times finer than the original resolution.

School/Discipline

Dissertation Note

Provenance

Description

Access Status

Rights

© 2025 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ).

License

Grant ID

http://purl.org/au-research/grants/arc/DP210100430
 http://purl.org/au-research/grants/arc/DP170102373
 http://purl.org/au-research/grants/arc/LE190100045
 http://purl.org/au-research/grants/arc/LE150100047
 http://purl.org/au-research/grants/arc/LE0882509
 http://purl.org/au-research/grants/arc/LE0453434

Published Version

https://doi.org/10.1016/j.rse.2025.114737

Call number

Persistent link to this record

https://hdl.handle.net/2440/147870