Periodic Oscillation Analysis of GPS Water Vapor Time Series Using Combined Algorithm Based on EMD and WD

Water vapor plays a key role in the global hydrologic cycle and the climate change. In order to investigate the spatial and the temporal characteristic of the Precipitable Water Vapor (PWV), the data of Global Positioning System (GPS) stations at Taiwan were chosen to analyze. The water vapor time values were calculated by Software Bernese 5.0 for 100 GPS stations from year 2006 to year 2013, and the sample interval is one hour. By analyzing the corresponding relation between the tropospheric water vapor and the geographical environment, the conclusion that the water vapor depending upon the latitude, the topography and the climatic conditions was made. Then, the long and the short time series of GPS water vapor for Taiwan were decomposed by Empirical Mode Decomposition (EMD) and Wavelet Decomposition (WD). The result shows that there were annual, semiannual, diurnal, and semidiurnal periodic oscillation graph in each water vapor series for GPS stations at Taiwan. Combining with geographical climate factors to analyze the physical reasons of these oscillation periods, conclusions could be found as follows. The annual monsoon changes impact the annual cycle of the water vapor series. The semiannual oscillation is mainly due to the alternated control of Taiwan between summer monsoon and winter monsoon. The diurnal cycle of the water vapor series is caused by the sea-land breeze circulation and the mountain-valley wind circulation. The amplitude of semidiurnal cycle in the water vapor series, which is mainly caused by local thermal convection due to the solar radiation, is small.