Identifying presence of correlated errors using machine learning algorithms for the selective de-correlation of GRACE harmonic coefficients

Piretzidis, Dimitrios; Sra, Gurveer; Karantaidis, Georgios; Sideris, Michael G.; Kabirzadeh, Hojjat

doi:10.1093/gji/ggy272

Cited by 12 publications

(7 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Gravity data processing. The iGrav SG was installed in October 2011 at the University of Calgary 51 . Portability and sensitivity tests were conducted over the next six months, where an accuracy of better than 1 µGal was achieved after a successful reduction of environmental effects 39,40 .…”

Section: Methodsmentioning

confidence: 99%

Microgravity effect of inter-seismic crustal dilatation

Kabirzadeh

Kim

Najafabadi

et al. 2022

Commun Earth Environ

Self Cite

View full text Add to dashboard Cite

Changes in the density of the shallow crust has been previously related to co-seismic strain release during earthquakes, however, the influence of inter-seismic deformation on crustal density variations is poorly understood. Here we present gravity observations from the iGrav superconducting gravimeter in southern Vancouver Island, British Columbia, Canada which reveal a substantial gravity increase between July 2012 and April 2015. We identify a negative correlation between this gravity increase and crustal dilatation strain derived from horizontal GPS velocities. The overall increasing gravity trend is caused by the gravity increase during and immediately before and after episodic tremor and slip events, which is partially compensated by gravity decrease occurring between the events. We conclude that the observed gravity increase results from a density increase due to crustal compression and that this is mostly a result of inter-seismic strain accumulation during the subduction of the Juan de Fuca plate beneath the North American plate.

show abstract

Section: Methodsmentioning

confidence: 99%

Microgravity effect of inter-seismic crustal dilatation

Kabirzadeh

Kim

Najafabadi

et al. 2022

Commun Earth Environ

Self Cite

View full text Add to dashboard Cite

show abstract

“…In the following, we will compare the formal error StD (standard deviation) of the filtered regional mass anomaly of the sparse DDK and the conventional DDK filters (DDK2, DDK3 and DDK4). For the sparse DDK filter, there are four options to calculate Q xx in Equation ( 19), namely with Equation (11)/Equation ( 13), with Equation (11)/Equation ( 14), with Equation ( 12)/Equation ( 13), or with Equation ( 12)/Equation (14). For easy reference, we call the corresponding variance-covariance estimates Q xx "MAP with σ 1 ", "EPL with σ 1 ", "MAP with σ 2 " and "EPL with σ 2 ", respectively.…”

Section: Filtered Mass Anomalies Analysismentioning

confidence: 99%

“…They are different from empirical decorrelation filters [8][9][10], which also find wide applications. There are also many other filters, such as [11][12][13][14][15][16][17][18], but these are out of the scope of this study since we mainly focus on DDK-like filters in this work. Usually, two types of input are necessary to design a DDK filter, namely the noise covariances and the signal covariances.…”

Section: Introductionmentioning

confidence: 99%

Sparse DDK: A Data-Driven Decorrelation Filter for GRACE Level-2 Products

et al. 2022

View full text Add to dashboard Cite

High-frequency and correlated noise filtering is one of the important preprocessing steps for GRACE level-2 products before calculating mass anomaly. Decorrelation and denoising kernel (DDK) filters are usually considered as such optimal filters to solve this problem. In this work, a sparse DDK filter is proposed. This is achieved by replacing Tikhonov regularization in traditional DDK filters with weighted L1 norm regularization. The proposed sparse DDK filter adopts a time-varying error covariance matrix, while the equivalent signal covariance matrix is adaptively determined by the Gravity Recovery and Climate Experiment (GRACE) monthly solution. The covariance matrix of the sparse DDK filtered solution is also developed from the Bayesian and error-propagation perspectives, respectively. Furthermore, we also compare and discuss the properties of different filters. The proposed sparse DDK has all the advantages of traditional filters, such as time-varying, location inhomogeneity, and anisotropy, etc. In addition, the filtered solution is sparse; that is, some high-degree and high-order terms are strictly zeros. This sparsity is beneficial in the following sense: high-degree and high-order sparsity mean that the dominating noise in high-degree and high-order terms is completely suppressed, at a slight cost that the tiny signals of these terms are also discarded. The Center for Space Research (CSR) GRACE monthly solutions and their error covariance matrices, from January 2004 to December 2010, are used to test the performance of the proposed sparse DDK filter. The results show that the sparse DDK can effectively decorrelate and denoise these data.

show abstract

“…Although the de-correlation filter can help reduce the stripe errors, they should be selectively applied only to correlated harmonic coefficients to avoid signal over-filtering. Piretzidis et al (2018) therefore developed a method (DP method) that identifies the presence of correlated errors by utilizing the capabilities of machine learning algorithms, which are proven to be effective in Canada and Greenland by…”

Section: Data Sets and The Processingmentioning

confidence: 99%

Converse Trends of the Terrestrial and Ground Water Storage Changes in Canada and the United States

Wang¹,

Xiang²,

Steffen³

et al. 2019

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

We study the terrestrial water storage (TWS) and groundwater storage (GWS) changes in Canada and United States. We employ the separation approach from Wang et al. (2013) together with the improved GRACE data of Release 6 for a longer time span until December, 2016. The TWS signals from lake levels are derived from satellite altimetry data over the lakes while TWS signals due to soil moisture (SM) and snow water equivalent (SWE) changes from hydrology models. There are four significant trend anomalies in North America for both TWS and GWS changes. Two positive anomalies are found in Canada with their centers in the provinces of Saskatchewan and Quebec, respectively, due to increased precipitation and/or increased runoff in their surroundings. Two negative anomalies are shown in the United States with their centers in California and the northwest of Texas, respectively, which are due to decreased precipitation and, especially for California, high water usage for agriculture.

show abstract

Identifying presence of correlated errors using machine learning algorithms for the selective de-correlation of GRACE harmonic coefficients

Cited by 12 publications

References 37 publications

Microgravity effect of inter-seismic crustal dilatation

Microgravity effect of inter-seismic crustal dilatation

Sparse DDK: A Data-Driven Decorrelation Filter for GRACE Level-2 Products

Converse Trends of the Terrestrial and Ground Water Storage Changes in Canada and the United States

Contact Info

Product

Resources

About