scholarly journals Singular Spectrum Analysis of Nonstationary Tidal Currents Applied to ADCP Data from the Northeast Brazilian Shelf

2006 ◽  
Vol 23 (1) ◽  
pp. 138-151 ◽  
Author(s):  
Marcio L. Vianna ◽  
Viviane V. Menezes

Abstract The development of new tools for the analysis of nonstationary currents, including tidal currents, has been the subject of recent research. In this work a method for studies of nonstationary barotropic or baroclinic currents based on empirical orthogonal function (EOF) and singular spectrum analysis (SSA) is proposed. It represents a new alternative to other methods of analysis of tidal currents in strong interaction with nontidal forcing, for example, the continuous wavelet transform. The advantage of the SSA method resides in the fact that it is fast, easy to implement, efficient for short-time records, and is based on the covariance structure of the data. If significant tidal constituents occur in the measurements, these are determined by the method itself even with short-time-series records. This is in contrast to the harmonic analysis (HA), where a large table of tidal constituents stated a priori are fitted to the data, even if the presence of some of these are spurious and not justified physically. The method is first demonstrated in the analysis of a synthetic current time series and then applied to an hourly current ADCP profile dataset of 410 days from the northeast Brazilian shelf. In both cases the SSA results were compared to the classical HA and the neoclassical short-term HA (STHA). The description of the shelf area where the ADCP was placed, the deployment and data acquisition operations, and the quality control data analysis are included for completeness. Analysis of the full ADCP quality-controlled data was done after a separation of the subtidal from the tidal high-frequency bands, although this traditional separation is not strictly necessary and was only made to better compare with HA and STHA. Analysis of the tidal band obtained from the ADCP data showed that the extracted tidal ellipse constituents present coherent oscillations dominated by the annual and 57-day periods, and changes in the sense of rotation of the current vector from anticyclonic to cyclonic in the ellipses. The subtidal band variability is shown to be also dominated by an annual and a 57-day period component, both polarized along the isobaths, which is suggestive of a nonlinear interaction of the subtidal and the tidal variability.

2020 ◽  
Vol 14 (3) ◽  
pp. 295-302
Author(s):  
Chuandong Zhu ◽  
Wei Zhan ◽  
Jinzhao Liu ◽  
Ming Chen

AbstractThe mixture effect of the long-term variations is a main challenge in single channel singular spectrum analysis (SSA) for the reconstruction of the annual signal from GRACE data. In this paper, a nonlinear long-term variations deduction method is used to improve the accuracy of annual signal reconstructed from GRACE data using SSA. Our method can identify and eliminate the nonlinear long-term variations of the equivalent water height time series recovered from GRACE. Therefore the mixture effect of the long-term variations can be avoided in the annual modes of SSA. For the global terrestrial water recovered from GRACE, the peak to peak value of the annual signal is between 1.4 cm and 126.9 cm, with an average of 11.7 cm. After the long-term and the annual term have been deducted, the standard deviation of residual time series is between 0.9 cm and 9.9 cm, with an average of 2.1 cm. Compared with the traditional least squares fitting method, our method can reflect the dynamic change of the annual signal in global terrestrial water, more accurately with an uncertainty of between 0.3 cm and 2.9 cm.


2018 ◽  
Vol 17 (02) ◽  
pp. 1850017 ◽  
Author(s):  
Mahdi Kalantari ◽  
Masoud Yarmohammadi ◽  
Hossein Hassani ◽  
Emmanuel Sirimal Silva

Missing values in time series data is a well-known and important problem which many researchers have studied extensively in various fields. In this paper, a new nonparametric approach for missing value imputation in time series is proposed. The main novelty of this research is applying the [Formula: see text] norm-based version of Singular Spectrum Analysis (SSA), namely [Formula: see text]-SSA which is robust against outliers. The performance of the new imputation method has been compared with many other established methods. The comparison is done by applying them to various real and simulated time series. The obtained results confirm that the SSA-based methods, especially [Formula: see text]-SSA can provide better imputation in comparison to other methods.


1997 ◽  
Vol 4 (4) ◽  
pp. 251-254
Author(s):  
A. Pasini ◽  
V. Pelino ◽  
S. Potestà

Abstract. An analysis of time series of monthly mean temperatures ranging from 1895 to 1989 is performed through application of Singular Spectrum Analysis (SSA) to data of several places in the USA. A common dynamics in the reconstructed spaces is obtained, with the evidence of a non-trivial and structured coupling of two Brownian motions, resembling the so-called Lévy flights. The idea that these two correlated functions are related to the zonal and eddy components of the atmospheric motions is suggested.


Sign in / Sign up

Export Citation Format

Share Document