Arctic Sea Ice Melt Onset Timing From Passive Microwave-Based and Surface Air Temperature-Based Methods

Melt onset (MO) on Arctic sea ice has been monitored using satellite-based passive microwave (PMW) observations since 1979. In this work, surface air temperatures from the International Arctic Buoy Programme/Polar Exchange at the Sea Surface and NASA's Atmospheric Infrared Sounder are used to derive MO date estimates using three threshold methods which are then compared with a record of PMW-based MO dates. Results from the PMW data indicate a shift toward increasingly early MO timing, with significant trends as large as -9.45days/decade in the E. Siberian Sea and -5.69days/decade for the entire Arctic, consistent with other studies highlighting the overall decline of Arctic sea ice. Results indicate that the surface air temperature-based MO date estimates produced using a -1 degrees C threshold are similar to 11days later than the PMW-based MO date estimates Arctic wide. A statistical comparison of the Polar Exchange at the Sea Surface and PMW-based MO dates indicate that despite the similar to 11-day bias, correlations between the MO date time series are generally good (0.6) for most of the Arctic Ocean while the Atmospheric Infrared Sounder and PMW-based MO dates are generally better at capturing the statistically similar long-term trends in MO dates for the Arctic and several Arctic subregions. Application of these results can contribute to the development of new methods to monitor the sea ice melting state. Plain Language Summary The time of year when sea ice first begins melting is an important indicator of ongoing climate change in the Arctic. In this work, satellite data are used to identify the day of the year that melting begins on sea ice in the Arctic, known as the melt onset date. The satellite-based melt onset dates are compared with melt onset dates estimated from surface air temperature data. The temperature data used in this research come from two sources: (1) drifting ocean buoys and (2) atmospheric temperatures measured by satellite. A statistical analysis is used to compare the satellite-based melt onset dates with the temperature-based melt onset estimates from each source. The results show that melting of Arctic sea ice is beginning earlier in the year, changing at a rate of 5.69days/decade since 1979. Results also show that the best method to estimate melt onset from the temperature data used in this study indicate that melt onset begins similar to 11days later than the satellite-based melt onset dates. Despite the differences between the temperature-based and satellite-based melt onset methods, these results can be used to develop improved methods to monitor summer sea ice melting in the future.