Remote Sensing Study
Detecting Change in Ice Presence with Correlations to Climatic Variations in Svalbard, Norway
Meteorological Data
Between August 1st and 10th, 2018, the data for the previous year was downloaded from the weather station directly onto a field laptop using the HOBOware software for OnSET loggers (Fig 2.1). The new data was then exported and combined with datasets of previous years using a Matlab code (Fig 2.2). Once combined, the master dataset was used to combine weather data. used to create graphs showing variations in different weather componenets (such as temperature, for example) over time.
Fig 2.2: Matlab script used to combine weather data.
Fig 2.1: Data from the loggers as seen in HOBOware.
Satellite Data
Along with this data, Level 1 LandSAT-8 images were also downloaded for the time intervals in consideration and processed for change detection in ENVI 5.3 using the workflow seen in (Fig 2.3).
Using GLOVIS and Earth.esa we downloaded July and August images from 2013 to 2018 with resolutions of 30m and 15m for the multispectral bands and the panchromatic band, respectively. During visual inspection of each image, we found that the most reliable data to be from July as August had too much cloud cover most years. We applied a spatial subset during the radiometric calibration step and corrected the data to reflectance, allowing us to use the gram-schimdt pan sharpening technique to improve the multi-spectral image resolution to 15m. This was followed by the use of the QUAC (QUick Atmospheric Correction) tool in ENVI after which we were ready to use the normalized difference snow index (NDSI) to highlight the snow and ice of the glaciers for each image.
Band math subtraction of the NDSI images simplified the interpretation of moraine and glacier migrations for the studied interval. The brightness contrast in the subtraction images allowed quantitative measurement of the displacement of individual features by using the cursor value and distance tools in ENVI (Fig 2.4). RGB composites were created by combining the subtraction images which allowed for the observation of the migrations through time. With the false-colour images, displacement could be measured constrained to specific years (Fig. 2.5)
Fig 2.4: Measurement of Total displacement
