| Empirical Mode Decomposition(EMD)is a developed adaptive data analysis method, which offers not only a potentially viable method for nonlinear and nonstationary data analysis, but also presents a new statistical view of the data, through which we can find out the underlying processes of any given physical phenomenon. Through the learning about the adaptive data analysis method and mainly using the satellite altimeter observation data, this thesis study changing regularity at global sea level and mesoscale ocean.In the first part of this thesis we analyze the global sea level observed data from satellite altimeter and tide gauges over 19932009 based on one-dimensional EMD method. Using the intrinsic trend obtained by decomposition, we present observational results on sea level change. The result shows that the sea level rising rate has slowed down dramatically since 1998,although the global mean sea level kept rising since 1900. Analysing the satellite sea level data derived from TOPEX/Poseidon, Jason-1, and ERS-1/2, we found that the nonlinear sea level rising speed has decreased from the rate 3.13 mm/yr in 1993 to 2.37 mm/yr in 2009, which is lower than the expected minimum mean linear rate 2.89 mm/yr during 1990s. Compared to altimeter data, tide gauges data also shows that the sea level rising rate is decreasing during the same period. If global mean sea level keeps rising with this rate(2.37 mm/yr), considering the annual oscillation and seasonal oscillations, the projected sea level would rise no more than 4.03 cm±0.86 cm in the next 17 years, which is much less than 8 cm from the IPCC 4th Assessment Report.Using the global three-dimensional ocean temperature and salinity date derived from Ishii, it is now possible to describe the upper 700 m steric sea level change and estimate the steric contribution to sea level rise. We found the upper 700 m steric sea level rising rate is reduced significantly, which is quite consistent with the sea level rising rate from satellite results. Averaged contribution from thermal expansion to sea level rise has dropped from 45% before 1999 to 25% at the end of 2006. Since about 2003, thermal expansion increase has almost stopped, whereas the sea level continues to rise, although at a reduced rate compared to previous. It implied that the ocean mass component dominates recent years sea level rise, explaining most observed rate.In the second part of this thesis we analyze satellite sea level data derived from TOPEX/Poseidon, Jason-1, and ERS-1/2 and study the mesoscale ocean phenomenon based on one-dimensional and Bidimensional EMD method. Mesoscale ocean phenomena usually refers to the time scale between few days with several months and spatial scale in tens to hundreds kilometers, which contains large kinetic energy and becomes the main factor affecting the global ocean variability.This data analysis method can effectively filter out large scale signal, decompose different spatial scale components and distinguish mesoscale components from these. By using the three-dimensional correlation structure involved time and space, objectively estimates may be made of the zonal and meridional length scales and propagation speeds in the north Pacific. We found that field with different spatial scales have different propagation speed. Statistically, the smallest spatial scale field has the lowest speed and the field with larger spatial scale has faster speed. Compared to first baroclinic Rossby waves phase speed, zonal averaged propagation speeds of mesoscale field have significant change with latitude, decreasing from the equator to high latitude, which are roughly in agreement with linear quasi-geostrophic theory values. However, at the same spatial scale the propagation speed of field is much higher than first baroclinic Rossby waves phase speed . In addition, the propagation phase of mososcale field may be subject to mean flow. Therefore, observed movement of mesoscale fields cannot be completely explained by the linear theory.
|
PDF Full Text Request