Improving Estimates of Cloud Radiative Forcing over Greenland

Abstract:

Multiple driving mechanisms conspire to increase melt extent and extreme melt events frequency in the Arctic: changing heat transport, shortwave radiation (SW), and longwave radiation (LW). Cloud Radiative Forcing (CRF) of Greenland's surface is amplified by a dry atmosphere and by albedo feedback, making its contribution to surface melt even more variable in time and space. Unfortunately accurate cloud observations and thus CRF estimates are hindered by Greenland's remoteness, harsh conditions, and low contrast between surface and cloud reflectance. In this study, cloud observations from satellites and reanalyses are ingested into and evaluated within a column radiative transfer model. An improved CRF dataset is obtained by correcting systematic discrepancies derived from sensitivity experiments.

First, we compare the surface radiation budgets from the Column Radiation Model (CRM) driven by different cloud datasets, with surface observations from Greenland Climate Network (GC-Net). In clear skies, CRM-estimated surface radiation driven by water vapor profiles from both AIRS and MODIS during May-Sept 2010-2012 are similar, stable, and reliable. For example, although AIRS water vapor path exceeds MODIS by 1.4 kg/m² on a daily average, the overall absolute difference in downwelling SW is < 4 W/m². CRM estimates are within 20 W/m² range of GC-Net downwelling SW.

After calibrating CRM in clear skies, the remaining differences between CRM and observed surface radiation are primarily attributable to differences in cloud observations. We estimate CRF using cloud products from MODIS and from MERRA. The SW radiative forcing of thin clouds is mainly controlled by cloud water path (CWP). As CWP increases from near 0 to 200 g/m², the net surface SW drops from over 100 W/m² to 30 W/m² almost linearly, beyond which it becomes relatively insensitive to CWP. The LW is dominated by cloud height. For clouds at all altitudes, the lower the clouds, the greater the LW forcing. By applying systematic tuning to minimize discrepancies between the CRM and observations, we construct an improved CRF dataset. According to our preliminary results, tuning CWP alone improves the SW CRF by over 20 W/m². With this CRF, studies on how much clouds contribute to Greenland surface melt in the long-term and in extreme melt years can be better quantified.