Comparison of Antarctic Cross Polarization Gradient Ratio Melt Quantities to Automated Surface Station Readings

A database containing images of the melt extent in Antarctica is currently being constructed using passive microwave satellite imagery taken during the melt seasons of 1987 through 2014. To determine if there is a difference between above freezing temperatures at the surface and melt values yielded by the Antarctic data analysis using these methods, melt amount is compared to in-situ temperature observations taken from surface automated weather stations during significant melt events as well as days with less melt between the 1987 and 2014 seasons. The cross-polarized gradient ratio (XPGR) has been used with two varying melt-classification thresholds to determine surface melt extent on the Greenland and Antarctica ice sheets. This method can detect the presence or absence of surface melting due to changes in microwave brightness temperatures at the 19GHz and 37GHz frequencies. Because the changes in microwave brightness are abrupt, this method is unable to discern differences between small amounts of melt and large amounts of melt. To verify the presence of surface melt in areas where the XPGR indicates melting, XPGR melt classification will be compared to surface temperature readings. The temperature values can be represented in an attribute table as being either above or below freezing at times specific to the satellite overpass times. The resulting binary values (above/below freezing) can then be matched up to the existing binary melt signatures (melt/no melt). It can then be determined if spatial patterns of temperatures at the surface match up to melt signatures in the XPGR.