Assimilating ASCAT normalized backscatter and slope into the land surface model ISBA-A-gs using a Deep Neural Network as the observation operator: Case studies at ISMN stations in western Europe

Shan, X.; Steele-Dunne, S.C.; Hahn, Sebastian; Wagner, Wolfgang; Bonan, Bertrand; Albergel, Clement; Calvet, Jean Christophe; Ku, Ou

doi:10.1016/j.rse.2024.114167

Assimilating ASCAT normalized backscatter and slope into the land surface model ISBA-A-gs using a Deep Neural Network as the observation operator

Title

Assimilating ASCAT normalized backscatter and slope into the land surface model ISBA-A-gs using a Deep Neural Network as the observation operator: Case studies at ISMN stations in western Europe

Author

Shan, X. (TU Delft Water Resources; TU Delft Geoscience and Remote Sensing)
Steele-Dunne, S.C. (TU Delft Mathematical Geodesy and Positioning)
Hahn, Sebastian (Technische Universität Wien)
Wagner, Wolfgang (Technische Universität Wien)
Bonan, Bertrand (Université de Toulouse)
Albergel, Clement (Université de Toulouse)
Calvet, Jean Christophe (Université de Toulouse)
Ku, Ou (Netherlands eScience Center)

Department

Geoscience and Remote Sensing

Date

2024

Abstract

ASCAT normalized backscatter (σ₄₀^o) and slope (σ^′) contain valuable information about soil moisture and vegetation. While σ₄₀^o has been assimilated to constrain soil moisture, sometimes together with Leaf Area Index (LAI), this study is the first to assimilate σ^′ directly to constrain vegetation states. Here, we assimilate σ₄₀^o and slope σ^′ into the ISBA-A-gs LSM using the Simplified Extended Kalman Filter (SEKF) using a Deep Neural Network (DNN) as the observation operator. The performances of the data assimilation (DA) and open loop (OL) are evaluated against in-situ soil moisture observations from the International Soil Moisture Network (ISMN), and LAI observations from the Copernicus Global Land Service (CGLS). Given an accurate and physically plausible observation operator, along with well-defined model and observation errors, the data assimilation system should yield improved estimates of the model states. However, results show that the DA performance is neutral compared to the OL in terms of the median unbiased root mean square error (ubRMSE) and Pearson correlation coefficient (ρ) across all validation sites. In addition, an analysis of the residuals and innovations confirms that DA had limited or no impact. This poor performance is perplexing. Furthermore, given the growing interest in the use of machine-learning-based observation operators, it is essential to understand the role that the use of the DNN may be playing in this poor performance. While representativeness errors and error specification play some part, it is demonstrated that the key factor constraining the efficacy of the SEKF is the correct estimation of the Jacobians that control the degree to which the observations update the states in the SEKF. It is argued that the DNN relating model states to satellite observations must have physically-plausible and robust Jacobians for the DNN to be effective in a data assimilation framework.

Subject

ASCAT
Data assimilation
Land surface model
Machine learning
Radar
Soil moisture
Vegetation

To reference this document use:

http://resolver.tudelft.nl/uuid:512c3db6-2621-49b1-b347-203be8b3e7c7

DOI

https://doi.org/10.1016/j.rse.2024.114167

ISSN

0034-4257

Source

Remote Sensing of Environment: an interdisciplinary journal, 308

Part of collection

Institutional Repository