Modeling seasonal oscillations in GNSS time series with Complementary Ensemble Empirical Mode Decomposition

The global navigation satellite system (GNSS) position time series provides essential data for geodynamic and geophysical studies. Interpolation of the GNSS position time series is necessary because missing data will produce inaccurate conclusions made from the studies. The spatio-temporal correlations between GNSS reference stations cannot be considered when using traditional interpolation methods. This paper examines the use of machine learning models to reflect the spatio-temporal correlation among GNSS reference stations. To form the machine learning problem, the time series to be interpolated are treated as output values, and the time series from the remaining GNSS reference stations are used as input data. Specifically, three machine learning algorithms (i.e., the gradient boosting decision tree (GBDT), eXtreme gradient boosting (XGBoost), and random forest (RF)) are utilized to perform interpolation with the time series data from five GNSS reference stations in North China. The results of the interpolation of discrete points indicate that the three machine learning models achieve similar interpolation precision in the Up component, which is 45% better than the traditional cubic spline interpolation precision. The results of the interpolation of continuous missing data indicate that seasonal oscillations caused by thermal expansion effects in summer significantly affect the interpolation precision. Meanwhile, we improved the interpolation precision of the three models by adding data from five stations which have high correlation with the initial five GNSS reference stations. The interpolated time series for the North, East, and Up (NEU) are examined by principal component analysis (PCA), and the results show that the GBDT and RF models perform interpolation better than the XGBoost model.

Section: Interpolation Of Different Seasonal Datamentioning

confidence: 99%

Section: Interpolation Of Different Seasonal Datamentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Interpolation of GNSS Position Time Series Using GBDT, XGBoost, and RF Machine Learning Algorithms and Models Error Analysis

Li,

Lu,

et al. 2023

“…Among them, y(t) is the GNSS vertical time series, t is the observation epoch, a is the constant term, b is the annual average vertical velocity, c, d and e, f are the annual and the semi-annual term coefficients, respectively, g j is the step coordinate mutation caused by the replacement of the antenna or the coseismic deformation, T g j represents the epoch of step, H is the Heaviside step function with H being 0 before the mutation or 1 after the mutation, and ε is the noise. Previously studies show that it is too optimistic to estimate the deformation rate with its value biased and its error underestimated if the error is assumed to be 'white noise' only [25]. Thus, in this study, noise models such as 'white noise' model, 'white noise + flicker noise' model and 'white noise + power-law noise' model, are used to estimate the parameters, respectively, and the optimal noise model is selected according to the Bayesian information criterion.…”

Section: Gnss and Data Processingmentioning

confidence: 99%

Influence of South-to-North Water Diversion on Land Subsidence in North China Plain Revealed by Using Geodetic Measurements

Wang,

Ding,

Chen

et al. 2023

As a major grain-producing region in China, the North China Plain (NCP) faces serious challenges such as water shortage and land subsidence. In late 2014, the Central Route of the South-to-North Water Diversion Project (SNWD-C) began to provide NCP with water resources. However, the effectiveness of this supply in mitigating land subsidence remains a pivotal and yet unassessed aspect. In this paper, we utilized various geodetic datasets, including the Gravity Recovery and Climate Experiment (GRACE) and GRACE Follow On (GRACE-FO), Global Navigation Satellite System (GNSS) and leveling data, to conduct a spatial-temporal analysis of the equivalent water height (EWH) and vertical ground movement in the NCP. The results reveal a noteworthy decline in EWH from 2011 to 2015, followed by a slight increase with minor fluctuations from 2015 to 2020, demonstrating a strong correlation with the water resources supplied by the SNWD-C. The GRACE-derived surface deformation rate induced by hydrological loading is estimated to be <1 mm/yr. In comparison, GNSS-derived vertical ground movements exhibit considerable regional differences during the 2011–2020 period. Substantial surface subsidence is evident in the central and eastern NCP, contrasting with a gradual uplift in the front plain of the Taihang Mountains. Three-stage leveling results indicate that the rate of subsidence in the central and eastern plains is gradually increasing with the depression area expanding from 1960 to 2010. Based on these geodetic results, it can be inferred that the SNWD-C’s operation since 2014 has effectively mitigated the reduction in terrestrial water storage in the NCP. However, land subsidence in the NCP persists, as the subsidence rate does not turn around in sync with the change in EWH following the operation of SNWD-C. Consequently, it’s necessary to maintain and enforce existing policies, including controlling groundwater exploitation and water resources supply (e.g., SNWD-C) to curtail the exacerbation of land subsidence in the NCP. Additionally, continuous monitoring of land subsidence by GRACE, GNSS, leveling and other geodetic techniques is crucial to enable timely policy adjustments based on monitoring results.

“…The main purpose of these elevation-dependent models is to reduce the contribution of the low elevation angle observations, due to the fact that typically they have a large error contribution to the parameter estimation. With the help of this advance software, when analyzing the GNSS coordinates time series, it is possible to capture the temporal variability of different physical phenomena, such as deformations in the Earth's crust, geodynamic phenomena providing information related to understanding strain and velocity fields, mass transfer, annual and semi-annual seasonal signals and other likewise phenomena [25][26][27][28].…”

Section: Introductionmentioning

confidence: 99%

Comprehensive Analysis on GPS Carrier Phase under Various Cutoff Elevation Angles and Its Impact on Station Coordinates’ Repeatability

Nistor,

Suba,

Buda

et al. 2024

When processing the carrier phase, the global navigation satellite system (GNSS) grants the highest precision for geodetic measurements. The analysis centers (ACs) from the International GNSS Service (IGS) provide different data such as precise clock data, precise orbits, reference frame, ionosphere and troposphere data, as well as other geodetic products. Each individual AC has its own strategy for delivering the abovementioned products, with one of the key elements being the cutoff elevation angle. Typically, this angle is arbitrarily chosen using generic values without studying the impact of this choice on the obtained results, in particular when very precise positions are considered. This article addresses this issue. To this end, the article has two key sections, and the first is to evaluate the impact of using the two different cutoff elevation angles that are most widely used: (a) 3 degrees cutoff and (b) 10 degrees cutoff elevation angle. This analysis is completed in two major parts: (i) the analysis of the root mean square (RMS) for the carrier phase and (ii) the analysis of the station position in terms of repeatability. The second key section of the paper is a comprehensive carrier phase analysis conducted by adopting a new approach using a mean of the 25-point average RMS (A-RMS) and the single-point RMS and using an ionosphere-free linear combination. By using the ratio between the 25-point average RMS and the single-point RMS we can define the type of scatter that dominates the phase solution. The analyzed data span a one-year period. The tested GNSS stations belong to the EUREF Permanent Network (EPN) and the International GNSS Service (IGS). These comprise 55 GNSS stations, of which only 23 GNSS stations had more than 95% data availability for the entire year. The RMS and A-RMS are analyzed in conjunction with the precipitable water vapor (PWV), which shows clear signs of temporal correlation. Of the 23 GNSS stations, three stations show an increase of around 50% of the phase RMS when using a 3° cutoff elevation angle, and only four stations have a difference of 5% between the phase RMS when using both cutoff elevation angles. When using the A-RMS, there is an average improvement of 37% of the phase scatter for the 10° cutoff elevation angle, whereas for the 3° cutoff elevation angle, the improvement is around 33%. Based on studying this ratio, four stations indicate that the scatter is dominated by the stronger-than-usual dominance of long-period variations, whereas the others show short-term noise. In terms of station position repeatability, the weighted root mean square (WRMS) is used as an indicator, and the results between the differences of using a 3° and 10° cutoff elevation angle strategy show a difference of −0.16 mm for the North component, −0.21 mm for the East component and a value of −0.75 mm for the Up component, indicating the importance of using optimal cutoff angles.