-
CIDR interpolation: an enhanced SSA-based temporal filling framework for restoring continuity in downscaled GRACE(-FO) TWSA products
Authors:
Yu Gao,
Wenyuan Zhang,
Junyang Gou,
Shubi Zhang,
Yang Liu,
Benedikt Soja
Abstract:
Global total water storage anomaly (TWSA) products derived from the Gravity Recovery and Climate Experiment (GRACE) and its Follow-On mission (GRACE-FO) are critical for hydrological research and water resource management. However, persistent data gaps hinder their application in long-term water cycle monitoring. We propose a Correlation-based Iterative and Decompose-Restore (CIDR) interpolation m…
▽ More
Global total water storage anomaly (TWSA) products derived from the Gravity Recovery and Climate Experiment (GRACE) and its Follow-On mission (GRACE-FO) are critical for hydrological research and water resource management. However, persistent data gaps hinder their application in long-term water cycle monitoring. We propose a Correlation-based Iterative and Decompose-Restore (CIDR) interpolation method to address the dual limitations of conventional Singular Spectrum Analysis (SSA) methods: instability caused by flawed iterative strategy and accuracy degradation due to inadequate denoising. CIDR solves the above issues with two key ways: (i) a correlation-driven iteration stopping rule ensuring the method's overall consistency, and (ii) component-specific interpolation with optimized denoising through a decompose-restore framework. The 0.5° downscaled GRACE(-FO) TWSA dataset was employed with artificial gaps replicating intra-mission and inter-mission scenarios to evaluate the method's performance. At the grid scale, the global average relative Root Mean Square (rRMS) and Nash-Sutcliffe Efficiency (NSE) improved by 17.5% and 42.6% compared to the SSA method for intra-mission gaps, and by 18.6% and 27.6% for inter-mission gap interpolation, respectively. In addition, the CIDR method shows a significant improvement over the SSA method within the global basins, with the global basin area-weighted average rRMS improved by 20.3% and 20% for both gaps, respectively. This research is expected to provide an advanced gap-filling method to meet the dual requirements of hydrological and climate research for high spatial resolution and long-term continuous TWSA products, which is universally applicable for other hydrological and geodetic time series.
△ Less
Submitted 11 April, 2025;
originally announced April 2025.
-
Uncertainties of Satellite-based Essential Climate Variables from Deep Learning
Authors:
Junyang Gou,
Arnt-Børre Salberg,
Mostafa Kiani Shahvandi,
Mohammad J. Tourian,
Ulrich Meyer,
Eva Boergens,
Anders U. Waldeland,
Isabella Velicogna,
Fredrik Dahl,
Adrian Jäggi,
Konrad Schindler,
Benedikt Soja
Abstract:
Accurate uncertainty information associated with essential climate variables (ECVs) is crucial for reliable climate modeling and understanding the spatiotemporal evolution of the Earth system. In recent years, geoscience and climate scientists have benefited from rapid progress in deep learning to advance the estimation of ECV products with improved accuracy. However, the quantification of uncerta…
▽ More
Accurate uncertainty information associated with essential climate variables (ECVs) is crucial for reliable climate modeling and understanding the spatiotemporal evolution of the Earth system. In recent years, geoscience and climate scientists have benefited from rapid progress in deep learning to advance the estimation of ECV products with improved accuracy. However, the quantification of uncertainties associated with the output of such deep learning models has yet to be thoroughly adopted. This survey explores the types of uncertainties associated with ECVs estimated from deep learning and the techniques to quantify them. The focus is on highlighting the importance of quantifying uncertainties inherent in ECV estimates, considering the dynamic and multifaceted nature of climate data. The survey starts by clarifying the definition of aleatoric and epistemic uncertainties and their roles in a typical satellite observation processing workflow, followed by bridging the gap between conventional statistical and deep learning views on uncertainties. Then, we comprehensively review the existing techniques for quantifying uncertainties associated with deep learning algorithms, focusing on their application in ECV studies. The specific need for modification to fit the requirements from both the Earth observation side and the deep learning side in such interdisciplinary tasks is discussed. Finally, we demonstrate our findings with two ECV examples, snow cover and terrestrial water storage, and provide our perspectives for future research.
△ Less
Submitted 23 December, 2024;
originally announced December 2024.
-
Downscaling GRACE-derived ocean bottom pressure anomalies using self-supervised data fusion
Authors:
Junyang Gou,
Lara Börger,
Michael Schindelegger,
Benedikt Soja
Abstract:
The gravimetry measurements from the Gravity Recovery and Climate Experiment (GRACE) and its follow-on (GRACE-FO) satellite mission provide an essential way to monitor changes in ocean bottom pressure ($p_b$), which is a critical variable in understanding ocean circulation. However, the coarse spatial resolution of the GRACE(-FO) fields blurs important spatial details, such as $p_b$ gradients. In…
▽ More
The gravimetry measurements from the Gravity Recovery and Climate Experiment (GRACE) and its follow-on (GRACE-FO) satellite mission provide an essential way to monitor changes in ocean bottom pressure ($p_b$), which is a critical variable in understanding ocean circulation. However, the coarse spatial resolution of the GRACE(-FO) fields blurs important spatial details, such as $p_b$ gradients. In this study, we employ a self-supervised deep learning algorithm to downscale global monthly $p_b$ anomalies derived from GRACE(-FO) observations to an equal-angle $0.25^\circ$ grid in the absence of high-resolution ground truth. The optimization process is realized by constraining the outputs to follow the large-scale mass conservation contained in the gravity field estimates while learning the spatial details from two ocean reanalysis products. The downscaled product agrees with GRACE(-FO) solutions over large ocean basins at the millimeter level in terms of equivalent water height and shows signs of outperforming them when evaluating short spatial scale variability. In particular, the downscaled $p_b$ product has more realistic signal content near the coast and exhibits better agreement with tide gauge measurements at around 80% of 465 globally distributed stations. Our method presents a novel way of combining the advantages of satellite measurements and ocean models at the product level, with potential downstream applications for studies of the large-scale ocean circulation, coastal sea level variability, and changes in global geodetic parameters.
△ Less
Submitted 8 April, 2024;
originally announced April 2024.
-
VGOS VLBI Intensives between MACGO12M and WETTZ13S for the rapid determination of UT1-UTC
Authors:
Matthias Schartner,
Leonid Petrov,
Christian Plötz,
Frank G. Lemoine,
Eusebio Terrazas,
Benedikt Soja
Abstract:
In this work, we present a status update and results of the designated research and development VLBI Intensive program VGOS-INT-S, observed between MACGO12M and WETTZ13S for the rapid determination of the Earth's phase of rotation, expressed via UT1-UTC. The main novelty of these sessions is the use of a special observation strategy, rapidly alternating between high- and low-elevation scans, enabl…
▽ More
In this work, we present a status update and results of the designated research and development VLBI Intensive program VGOS-INT-S, observed between MACGO12M and WETTZ13S for the rapid determination of the Earth's phase of rotation, expressed via UT1-UTC. The main novelty of these sessions is the use of a special observation strategy, rapidly alternating between high- and low-elevation scans, enabling an improved determination of delays caused by the neutral atmosphere. Since 2021, 25 Intensive sessions have been observed successfully. In early 2022, VGOS-INT-S was among the most accurate Intensive programs with an average formal error $σ_{UT1-UTC}$ of 3.1 $μ$s and a bias w.r.t. IERS C04 of 1.1 $μ$s. Later, the session performance decreased due to multiple technical difficulties.
△ Less
Submitted 16 February, 2023;
originally announced February 2023.
