Gridded glacier datasets are essential for various glaciological and climatological research because they link glacier cover with the corresponding gridded meteorological variables. However, there are significant differences between the gridded data and the shapefile data in the total area calculations in the Randolph Glacier Inventory (RGI) 6.0 at global and regional scales. Here, we present a new global gridded glacier dataset based on the RGI 6.0 that eliminates the differences. The dataset is made by dividing the glacier polygons using cell boundaries and then recalculating the area of each polygon in the cell. Our dataset (1) exhibits a good agreement with the RGI area values for those regions in which gridded areas showed a generally good consistency with those in the shapefile data, and (2) reduces the errors existing in the current RGI gridded dataset. All data and code used in this study are freely available and we provide two examples to demonstrate the application of this new gridded dataset.

In this study we analyse a 15-year long time series of surface mass-balance (SMB) measurements performed between 2001 and 2016 in the ablation zone of the Morteratsch glacier complex (Engadine, Switzerland). For a better understanding of the SMB variability and its causes, multiple linear regressions analyses are performed with temperature and precipitation series from nearby meteorological stations. Up to 85% of the observed SMB variance can be explained by the mean May–June–July temperature and the total precipitation from October to March. A new method is presented where the contribution of each month's individual temperature and precipitation to the SMB can be examined in a total sample of 224 (16.8 million) combinations. More than 90% of the observed SMB can be explained with particular combinations, in which the May–June–July temperature is the most recurrent, followed by October temperature. The role of precipitation is less pronounced, but autumn, winter and spring precipitation are always more important than summer precipitation. Our results indicate that the length of the ice ablation season is of larger importance than its intensity to explain year-to-year variations. The widely used June–July–August temperature index may not always be the best option to describe SMB variability through statistical correlation.

An assessment of glacier shrinkage (reduction of area) for all of High Mountain Asia requires a complete compilation of measured rates of change and also a methodology for objective comparison of rates. I present a compilation from 155 publications reporting glacier area changes, and also a methodology that overcomes the main obstacles hindering comparison. Glacier areas are not always assigned uncertainties, and this problem is addressed with an error model derived from published estimates. The problem of discordant survey dates is addressed by interpolating measured areas to fixed dates at pentadal intervals. Interpolation error depends only incoherently on the time span between measurements, but strongly on glacier size: smaller glaciers, in addition to changing more rapidly on average, exhibit more variable rates of change. The overlapping boundaries of study regions are reconciled by mapping all of the information to a 0.5° geographical grid. When coupled with glacier area information from the Randolph Glacier Inventory, the widely observed inverse dependence of shrinkage rates on glacier size shows promise as a tool for treating incomplete spatial coverage. Over High Mountain Asia as a whole from 1960 to 2010, the unweighted average shrinkage rate is –0.57% a–1, but corrections for variable glacier size raise the average to –0.34% a–1, and filling unmeasured gridcells with rates based on size dependence alters the latter estimate to –0.40% a–1. The uncertainties in these rates are large. The Karakoram anomaly is found to be a zonal feature extending well to the east of the Karakoram proper.