Local-scale spatial variability in firn properties in Southwest Greenland

Abstract

The increasing mass loss from the Greenland ice sheet is becoming growingly important to present-day global sea level rise. Meltwater export to the ocean may be amplified by near-surface ice layers in firn (i.e., snow that has survived at least one melt season) which prevent infiltration. Such ice layers have been documented at various sites in southwest Greenland, but little is known about the local-scale (less than 1 km) variability in firn and ice properties. Here we investigate the local variability in firn density and ice content (ice layer thickness fraction) by analyzing 45 shallow firn cores retrieved from 15 sites in southwest Greenland between 2012 and 2019. The core dataset contains groups of cores that were collected at the same site (1–632 m apart from each other) and in the same year (here referred to as same site/year cores). The two-sided Mann-Whitney U-tests suggest that most same site/year cores do not have statistically significant differences in median density or ice content over the upper 0–4.5, 4.5–9, and 9–13.5 m sections below the winter snow layer. That said, the same site/year cores still exhibit substantial density and ice content differences as well as variable ice layer stratigraphy (especially the distribution of centimeters-thick ice lenses) not explainable by measurement uncertainties. The local-scale variability in density and ice content (calculated from the same site/year cores) over the three 4.5-m sections is 4% and 19–29%, respectively. The variability is comparable to the regional gradient in our study area within an elevation difference of up to ∼200 m. The local variability is non-negligible particularly at mid- to high-elevation sites (>2,000 m a.s.l.), where firn structure changes such as ice layer formation are more recent phenomena compared to lower elevations. This study confirms that single cores can provide representative bulk statistics such as mean density and ice content of one site. However, these mean values mask substantial local variability which can compromise regional comparisons. Therefore, collection of same site/year cores and consideration of local variability are necessary for an in-depth analysis of regional variability.