Reconstruction and Projection of Sea Levels Around the Korean Peninsula Using Cyclo-Stationary Empirical Orthogonal Function

Abstract

and COBESST2 data of the North-west Pacific Ocean showed the best agreement with TG observations. Although the TG data was not used in the reconstruction process, the reconstructed results showed better agreement with the TG observation than a previous study that used the TG data. The projection of sea level around the Korean Peninsula was conducted as well corresponding to greenhouse gas emission scenarios. The basic concept of the projection was similar to the reconstruction in using the LVs of satellite altimeter data as the basis function, but the projection employed the simulated SST data of HadGEM2-ES to extend the PCTs of the basis functions. The projection was conducted for four RCP scenarios (RCP 2.6, 4.5, 6.0 and 8.5) over 2006-2100. The results in reconstruction and projection were given as both mean and standard deviation for each month and each grid point through Monte Carlo simulations.

it is relatively short compared with the history of tide gauge (TG) observation. Many studies have been conducted to extend the spatial resolution of the satellite data into the time before satellite measurements by using both satellite data and TG data simultaneously. However, most of the reconstructions of sea level were conducted on a global scale, which brought on reducing accuracy at some local areas, where the signal is relatively weaker than other regions and the number of TG is not enough to represent the area. The sea level around the Korean Peninsula is also relatively low signal zone, and the number of TG is few before 1960. Therefore, in this study, new methods are proposed to reconstruct the past sea level and project the future sea level around the Korean Peninsula. Using CSEOFs (Cyclo-Stationary Empirical Orthogonal Function) loading vectors (LV) of satellite data as basis functions of the reconstruction, the principal component time series (PCT) of LV is extended over 1900-2014. The PCTs of sea surface temperature data and altimeter data are used as independent variables and depending variables for regression analysis, respectively. The regression analysis considering time lags is conducted to find the lags and regression coefficients. Using the regression results, the PCTs of satellite data were extended into the past. In this study, we conducted 13 reconstructions

Since the operating of satellite altimeter, the understanding of the sea level has been increased dramatically. However, the history of the satellite altimeter dates back to the 1990s