Testing different methods to model melt in HMA

We recently published a new paper, led by Maxime Litt, providing guidelines for glacier-ablation modelling in HMA environments.

The conventional Temperature index (TI) models for modelling glacier ablation require few input variables and rely on simple empirical relations. The approach is assumed to be reliable at lower elevations (below 3500 m above sea level, a.s.l) where air temperature relates well to the energy inputs driving melt. Using field meteorological observation in Langtang and Khumbu, we show that temperature relates poorly to a number of important mass-loss drivers in high-altitude, so that temperature indexes have to be handled with care.
At the high elevation glaciers in Mountain Asia (HMA), we observed that incoming shortwave radiation is the dominant energy input and the full surface energy balance model relates only partly to daily mean air temperature. During monsoon surface melt dominates ablation processes at lower elevations (between 4950 and 5380 m a.s.l.). As net shortwave radiation is the main energy input at the glacier surface, albedo and cloudiness play key roles while being highly variable in space and time. For these cases only, ablation can be calculated with a TI model. Sublimation and other wind-driven ablation processes are important for mass loss, and remain unresolved with such simple methods. Ablation modeled with a SEB can diverge from the observations, but a suitable value for surface roughness can solve the issue.

Cumulated ablation calculated with the surface lowering measurements (thick blue line), with the surface energy balance for changing z0 values (orange dashed and continuous lines), with the TI (red line) and ETI (clear blue line) with one fixed set of factors. The hourly wind speed is shown upside down (green curve). Periods of surface melt (Ts = 0) are highlighted in orange. Results from Mera Glacier, 5380 m a.s.l in 2014 and 2017 (a) from Yala Glacier, 5350 m a.s.l., in 2014, 2016 (b) and Mera Glacier, 6352 m a.s.l, in 2015 and 2016 (c).

Read the paper in detail here.