Description: The dataset contains model output presented in the manuscript 'A long-term dataset of climatic mass balance, snow conditions and runoff in Svalbard (1957-2018)', which is considered for publication in The Cryosphere. The data are structured in 3-D arrays containing spatially distributed and annual mean values of the variables specified below. The spatial resolution is 1x1-km. Variables included in the dataset: ----------------------------- - Climatic mass balance - Air temperature - Precipitation - Runoff - Refreezing - Pore space (down to 14 m) - Subsurface temperature (at 14 m depth) - Snow disappearance date - Snow onset date

Description: The present dataset contains measurements of density and observations of stratigraphy in the subsurface snow/firn/ice done at Lomonosovfonna during 1997-2015. The variables are named according to the year, when the data was derived: "LF" for Lomonosovfonna and "NN" corresponds to the year, e.g. 97 - 1997, 07 - 2007. Most variables contain the following fields: rho* - density measurements: column 1 - depth of sample top, m; column 2 - depth of sample bottom, m; column 3 - density values, kg m^-3; rho*_reg - density measurements on a regular 1 cm spaced grid, gaps are filled by linear interpolation and extrapolated using "nearest neighbor" logic; column 1 - depth, m; column 2 - density values, kg m^-3; strat - stratigraphy: column 1 - depth of sample top, m; column 2 - depth of sample bottom, m; column 3 - stratigraphy coded as: 1 - snow, 2 - firn, 3 - ice lens; lat* - latitude, degrees to the North of equator; lon* - longitude, degrees to the East of the prime meridian; h* - elevation, m above the sea level. LF97: Six shallow cores were drilled in spring 1997 approximately along the centerline of Nordenskiöldbreen. A system of mass balance stakes at locations of the cores is maintained by Uppsala university. The density is approximated from visual stratigraphic description according to the classification suggested by Pohjola et al., 2002 (see Table 1 there). The density values are the result of averaging of density values ascribed to stratigraphic units identified in core pieces and weighted by their respective thicknesses. LF99: Three shallow cores were drilled in the end of April 1999 at the location where in 1997 a 120 m long ice core was drilled. The cores were drilled along a North-South oriented line with a spacing of 2.5 m between the neighboring cores. The fields in the LF99 variable are named accordingly: fields with the data from the southern location contain "S", northern - "N", central - "C". The field and laboratory work was done by Håkan Samuelsson. Methods and analysis are described in detail in the Master Thesis: "Distribution of melt layers on the ice field Lomonosovfonna, Spitsbergen", defended at Uppsala University in 2001. LF08: One shallow core was drilled by Sanja Forsström, Elisabeth Isaksson, Veijo Pohjola and Jim Hedfors within ca 100 from the location where in 1997 a 120 m long ice core was drilled. Density was measured using two different methods at the cold lab of the Norwegian Polar Institute in Tromso by Sanja Forsström and Tonu Martma. The structure field "rho" contains values calculated from measured geometrical dimensions of the core samples and weights. The fields "rhoDEPcorepieces", "rhoDEPionsamples" and "rhoDEPisotopesamples" contain density values measured using dielectric profiling in three sets of samples. LF12: The core was drilled by Veijo Pohjola and Rickard Pettersson on the 13 of April 2012. Field notes done by Sergey Marchenko. Cold lab operations were done by Sergey Marchenko and Elena Klimenko at the University Centre in Svalbard (UNIS) in Longyearbyen, Norway during 26 April - 3 May 2012. Density was measured in cylindrical and cuboid samples prepared using a band saw to ensure regular shape. The structure field "rho_c" contains density measurements done using cylindrical core pieces. The structure field "rho_rb" contains density measurements done using relatively long cuboid samples prepared from cylindrical core pieces using a band saw. The structure field "rho_rs" contains density measurements done using shorter cuboid samples prepared from the longer pieces using a band saw. The structure field "rho_reg" is based on the "rho_rs" structure field. LF13: The core was drilled by Christian Zdanowics, Dorothee Vallot and Veijo Pohjola in April 2013 and later analyzed by Carmen Vega in the cold lab of the Norwegian Polar Institute in Tromso, Norway. LF14: The core was drilled by Veijo Pohjola and Ward van Pelt in the end of March 2014. Field notes are done by Veijo Pohjola. Cold lab operations were done by Sergey Marchenko, William Kohler and Elisbeth Isaksson in the Norwegian Polar Institute facilities in Tromso, Norway, during 05-10 of October 2014. Density was measured in cylindrical or cuboid samples prepared using a band saw to ensure regular shape. LF15: the core was drilled by Veijo Pohjola and Ward van Pelt on the 15th of April 2015. Field notes are done by Veijo Pohjola. Cold lab operations were done by Sergey Marchenko, Glennda Villanflor and Elisbeth Isaksson in the Norwegian Polar Institute facilities in Tromso, Norway, during 02-05 of November 2015. Density was measured in cylindrical or cuboid samples prepared using a band saw to ensure regular shape.

Description: Compared to other Arctic ice masses, Svalbard glaciers are low-elevated with flat interior accumulation areas, resulting in a marked peak in their current hypsometry (area-elevation distribution) at ~450 m above sea level. Since summer melt consistently exceeds winter snowfall, these low-lying glaciers can only survive by refreezing a considerable fraction of surface melt and rain in the porous firn layer covering their accumulation zones. We use a high-resolution climate model to show that modest atmospheric warming in the mid-1980s forced the firn zone to retreat upward by ~100 m to coincide with the hypsometry peak. This led to a rapid areal reduction of firn cover available for refreezing, and strongly increased runoff from dark, bare ice areas, amplifying mass loss from all elevations. As the firn line fluctuates around the hypsometry peak in the current climate, Svalbard glaciers will continue to lose mass and show high sensitivity to temperature perturbations. The data set includes annual cumulative SMB and components statistically downscaled from the output of the Regional Atmospheric Climate Model RACMO2.3 to 500 m spatial resolution (1958-2018). SMB components include total precipitation (snowfall and rainfall), snowfall, runoff, melt, refreezing and retention (mm w.e. per year), as well as summer (June-July-August) 2 m air temperature (K). The data set also includes modelled (RACMO2.3; 1958-2018) and observed (MODIS; 2000-2018) bare ice area, and modelled ablation zone area (1958-2018; km2). The mask file includes longitude/latitude (ºN/ºW), land-sea, ice and sector masks from the Randolph Glacier Inventory version 6, and surface topography (m above sea level) from the S0 Terreng Digital Elevation Model (Norwegian Polar Institute) on the 500 m grid. Daily downscaled SMB and components are available from the authors upon request and without conditions (b.p.y.noel@uu.nl).