An efficient low-pass-filtering algorithm to de-noise global GRACE data

doi:10.1016/j.rse.2022.113303

An efficient low-pass-filtering algorithm to de-noise global GRACE data

2022-12-01

19

PRE

AI

DOI：10.1016/j.rse.2022.113303

Original

Original request for help

Save

Abstract

En 中文

The global monthly equivalent water height (EWH) anomaly (dEWH), derived from the Gravity Recovery and Climate Experiment (GRACE) level-2 data, helps understand the terrestrial water storage variation. Unfortu-nately, the dEWH data consist of stripes or noise. Numerous de-noising approaches have been studied. However, the following issues exist, possible overly smoothed results, two-step (de-striping and Gaussian-based smoothing) procedures introducing additional uncertainty and error propagation, requiring multi-temporal datasets or prior knowledge, and possible computational inefficiency. Thus, after analyzing the spectrum of the dEWH data, a novel low-pass-filtering algorithm is proposed to remove the noise and resolve the issues. Furthermore, without the global in situ measurements, an alternative assessment method is studied based on the additive character-istics of the signal and noise in the GRACE data. The method consists of the residual analysis and root mean square (RMS) value of the de-noised signal. Then, the proposed algorithm was applied to de-noising the dEWH datasets between 2002 and 2015. De-noised results are satisfactory qualitatively and quantitatively. Compared with well-known two-step de-noising methods, data blurring does not occur after the proposed algorithm. The algorithm removes noise the most assessed by the residual analysis and preserves the signal the most evaluated by large RMS values.

Keywords:

Equivalent water height anomaly ( dEWH )

Fourier transform (FT)

Gravity recovery and climate experiment

(GRACE)

Low -pass -filtering algorithm

Residual analysis

Root mean square ( RMS )

An efficient low-pass-filtering algorithm to de-noise global GRACE data

Abstract

Journal

Organization