DATE | 2019-01-08 12:44:23 |
IDABSTRACT | 2014/222 |
CONTACT | hasan.yildiz@hgk.msb.gov.tr |
PRESENTATION | ORAL |
INVITED | 0 |
IDSESSION | 2 |
TITLE | TRENDS OBSERVED IN TURKISH COASTAL MEAN SEA LEVEL DATA
USING EMPIRICAL MODE DECOMPOSITION METHOD |
AUTHORS | Hasan Yildiz (1) |
AFFILIATIONS | - Geodesy Department, General Command of Mapping, Ankara, TURKEY
|
ABSTRACT |
Typically, 30 years of data are required in order to determine a linear mean sea level (MSL) trend with a standard error of the order of 0.5 mm/yr. Such requirements on the length of time series are caused from decimetric interannual to interdecadal variability in tide gauge MSL records. The degree to which one is able to model the interannual to decadal variability determines the accuracy of the long-term sea level trends. As Turkish coastal MSL records are relatively short (< 30 years) for long term trend estimation a method based on Empirical Mode Decomposition (EMD) is used to separate the long-term sea level trend from interannual and decadal oscillations. Results show that Turkish MSL time series has a nonlinear character and can not be described in terms of a linear increase alone but includes significant variations on interannual and decadal time scales. We also demonstrate that EMD method gives more accurate long-term trends in MSL time series than the traditional linear regression approach. |
PAGE | 28 |
STATE | 1 |