Development and testing of a method for glacier surface properties detection and classification using high resolution multispectral remote sensing for Langjökull ice cap and neighboring glaciers

In Iceland, glaciers are mainly covered with snow during winter but in spring the surface snow cover starts to melt exposing increasingly more of the ice underneath. The snow and ice melt are governed by the absorption of shortwave radiation which increases as the surface albedo decreases. As the ic...

Full description

Bibliographic Details
Main Author: Salka Kolbeinsdóttir 1995-
Other Authors: Háskóli Íslands
Format: Thesis
Language:English
Published: 2021
Subjects:
Online Access:http://hdl.handle.net/1946/38583
Description
Summary:In Iceland, glaciers are mainly covered with snow during winter but in spring the surface snow cover starts to melt exposing increasingly more of the ice underneath. The snow and ice melt are governed by the absorption of shortwave radiation which increases as the surface albedo decreases. As the ice albedo is considerably lower than the snow albedo, the appearance of bare ice during the summer months results in increased surface melt. Different glacial surfaces such as snow, bare ice and debris covered snow or ice all have a distinct spectral signature which makes it possible to utilize multispectral remote sensing to distinguish each surface property and estimate their spatial distribution. Here, temporal evolution of the surfaces of the Langjökull glacier group were estimated using supervised image classification techniques on available Sentinel-2 satellite images with less than 35% cloud cover. The study period was from 2016-2020 from May till September. To mitigate the effect of data gaps due to clouds, a cloud detection and gap-filling method was applied. During the studied melting seasons, all glaciers lost 30% or more of their surface snow cover by area. The method allows the timing of the exposure of bare ice at each survey station on Langjökull to be estimated during each melt season. Correlation coefficient of 0.83 was found between mass balance measurements at each survey station and the number of days of bare ice exposure. The main limitation of this method is the frequent cloud cover that hinders observation of the surface. The results of the study show that valuable data on the evolution of the winter snow during the melt season on glaciers are gained from remote sensing without the need for expansive fieldwork. Íslenskir jöklar eru stóran hluta ársins huldir snjó en þegar vorar bráðnar snjórinn og jökulís sem undir liggur kemur í ljós. Bráðnun snævar og íss er háð gleypni yfirborðsins á orku sólarljóss (stuttbylgjugeislun) sem eykst þegar endurkaststuðull yfirborðsins lækkar. Endurkaststuðull ...