For my Digital Surface Modeling course project, I developed a method to create digital elevation models (DEMs) of ice sheets using high-resolution altimetry data from the IceSat2 satellite. The goal was to overcome the limitations and costs associated with traditional ground-based or airborne surveys.
The method involved the following steps:
- Data acquisition: Downloading the IceSat2 ATL06 data files containing point cloud information (latitude, longitude, elevation).
- Data preprocessing: Removing outliers and merging data from multiple files.
- Grid creation: Defining a 70×30 pixel grid to represent the area of interest.
- Interpolation: Using bilinear interpolation to estimate elevation values for each grid cell based on surrounding data points.
- Change detection: Comparing 2019 and 2023 DEMs to identify areas of ice sheet melting, using the Mines method.
I implemented this process in Python due to its flexibility and applied it to data sets from the Greenland ice sheet for the first six months of 2019 and 2023.
The results showed an accelerating rate of melting on the Greenland ice sheet between the two time periods. In 2019, the ice sheet appeared relatively stable, with some melting at the edges but no significant overall elevation change. However, by 2023, more extensive melting was observed at the edges, accompanied by a slight decrease in overall elevation.
The difference map between the 2019 and 2023 DEMs revealed an average elevation decrease of 12.6 meters, indicating a negative mass balance and supporting the observation of accelerated melting.