ABSTRACT
Melting snow and glacier ice in the Himalaya forms an important source of water for people downstream. Incoming longwave radiation (LWin) is an important energy source for melt, but there are only few measurements of LWin at high elevation. For the modelling of snow and glacier melt, the LWin is therefore often represented by parameterizations that were originally developed for lower elevation environments. With LWin measurements at eight stations in three catchments in the Himalaya, with elevations between 3,980 and 6,352 m.a.s.l., we test existing LWin parameterizations. We find that these parameterizations generally underestimate the LWin, especially in wet (monsoon) conditions, where clouds are abundant and locally formed. We present a new parameterization based only on near-surface temperature and relative humidity, both of which are easy and inexpensive to measure accurately. The new parameterization performs better than the parameterizations available in literature, in some cases halving the root-mean-squared error. The new parameterization is especially improving existing parameterizations in cloudy conditions. We also show that the choice of longwave parameterization strongly affects melt calculations of snow and ice.
