Project 4

Sea ice parameters in the north and south polar regions are important components of the global climate system. The proposed research project last year was on studying the variations in the Arctic sea ice temperatures utilizing the data obtained from the Scanning Multichannel Microwave Radiometer (SMMR) on board the NASA Nimbus 7 satellite. Two major components of the El Nino Southern Oscillation (ENSO) signal, the quasibiennial (2.4-year; QB) and the quasiquadrennial (4.2-year; QQ) oscillations have been studied. A combination of singular value decomposition (SVD) and a modified Butterworth filtering technique has been used to extract the QB and QQ components in the data. First, SVD of gridded data over the nine years has produced the principle components (PC) of the data. It has been shown that the singular values decay fast in magnitude so that a truncated outer-product expansion can be used to generate a filtered data using only one-fifth of the total components. It has also been shown that the outer-product expansion gives a separation of spatial and temporal portions of PCs. Next, the temporal portion of the PCs has been decomposed by a modified Butterworth filtering technique. Finally, ice temperatures have been reconstructed using the band passed data corresponding to QQ (or QB) frequency in place of the temporal portion of the PCs The resulting filtered data represents ice temperatures attributed to QQ (or QB) component only.

In order to better understand both spatial and temporal variations of the QB and QQ components in the Arctic sea ice temperatures, a detail study and analysis on the reconstructed data described above have been carried out last year. In particular, time series at some hotspot pixels for both QB and QQ components have been plotted to show temporal oscillations. QB and QQ components of the sea ice temperatures at selected times have been displayed as image maps to show spatial variations. Movies of both QB and QQ components have been made in MATLAB to simulate these interannual variabilities. These results show evidence of ENSO signal in Arctic sea ice, corroborate observations made in sea ice concentrations in earlier studies. For QQ component strong activities are in the upper Hudson and Baffin Bays and the Greenland and Barents Seas. For the QB component most strong activities are in the Chukchi Sea. Finally, comparison of above results with sea ice concentration has also been carried out. In particular, movies of image maps for both QB and QQ components of the sea ice concentration have been made with the corresponding sea ice temperatures plotted as contours and superimposed on top of the maps.

The research described here is an ongoing project in collaboration with Dr. Per Gloerson at NASA Goddard Space Flight Center (GSFC) in Greenbelt, Maryland. The main objective of the proposed research project last year as described above has been accomplished. One of research projects in the Oceans and Ice Branch at GSFC is to identify and assess the role of long-period oscillations in climate trend observables. In an earlier study of SMMR sea surface temperature (SST) time series, which runs from 1978 - 1987, multiple-window Prolate Spheroid filters have been used in conjunction with discreet Fourier analysis (MWF) to obtain well-defined spectra of the SSTs. The MWF analysis method provides amplitude and phase information on the spectral line computed, as well as confidence levels in the results. Variations of periodicity in both space and time are analyzed with the use of Empirical Mode Decomposition method. The QB and QQ fluctuations in the SSTs are not only in the equatorial Pacific, but are also observed globally, with as yet unexplained spatial distributions. Recently, new SST data from 1985 - 1999 have become available in different spatial/temporal resolutions. These data are derived from the 5-channel Advanced Very High Resolution Radiometers (AVHRR) on board the NOAA -7, -9, -11 and -14 polar orbiting satellites. With the new AVHRR data, we can not only do a comparison of the analysis on the two SST data sets, but also construct a much-longer combined SST time series. The combined data will allow us to study SST oscillations on decadal and interdecadal time scales. For that purpose we propose our future study of the Year 2 NASA EPSCoR Preparation Grant as follows:

1. Download the single-day AVHRR SSTs and composite the data into 6-day composites with a spatial grids of $0.5\times0.5$ degrees to match the SMMR SST time series.

2. Analyze the data with the methods described above and compare the results with that of SMMR data during the overlap period.

3. Construct a combined SST time series to study the long-period SST oscillations both locally and globally.

Prof. Jun Yu, the Principal Investigator for this project, is currently an Associate Professor of Mathematics and Statistics at the University of Vermont. He has expertise in the areas of applied mathematics and data analysis with applications in geophysics.