This paper presents a statistical analysis of sea ice in the Greenland Sea (70-80N and 10W-10E) from January 1979 to December 2007. We define four variables from satellite images: ice extent, ice area, eastward ice extent at 75N, and the shape of the ice edge. We establish relationships between these ice variables and five climate variables: sea surface temperature (SST), air temperature at Jan Mayen Island (JM), sea surface pressure (SSP), North Atlantic Oscillation (NAO) and Arctic Oscillation (AO). The ice extent, ice area and eastward ice extent are highly correlated with SST and JM over the whole time period, while SSP, NAO and AO are only correlated with the ice variables over shorter (one year) periods. When considering the different shapes of the ice edge, an ANOVA model shows that 49\% of the variance in extent can be explained by the shape of the ice edge, and similar amounts for the eastward ice extent and ice area. Fourier analysis of the sea ice data shows a 10 year peak which is consistent with previous hypotheses in work done on the Greenland Sea. Finally, we show that the negative trend in the sea ice extent, ice area and eastward extent is statistically significant. This can be interpreted as yet another indicator of global warming that has been taking place for the last 30 years.