Data Science using remote sensing data for predicting sea ice thickness

Jump to: navigation, search

Title: Data Science using remote sensing data for predicting sea ice thickness

Mentor : Dr. Satish Puri

In this project, we will use machine learning to predict the thickness of sea ice around polar region which is important for understanding climate change. The data consists of tuples, where each tuple has four variables x, y, z, t where x, y are co-ordinates for latitude, longitude, z is the elevation of the surface of the earth and t is time. The data is collected by a satellite and it is publicly available for science applications. The dataset size is in terabytes.

NASA's Ice, Cloud, and Land Elevation Satellite (ICESat‐2) was launched in 2018 to measure accurate surface heights for understanding changes in ice sheets and sea ice around the arctic. The satellite is used to collect elevation (height) data for different locations around the earth which is used to estimate the thickness of the sea ice. This is done using a special kind of lidar sensor.  Sea ice thickness is not only an important indicator of how the polar oceans are responding to a warming climate but also useful for forecasting of future changes in the polar regions.

Students will gain experience with all aspects of data science – data cleaning, data wrangling, data visualization, implementing and evaluating machine learning models for accuracy, etc.

Student Research Activities: The REU fellows will perform the following major tasks: - Survey state-of-the-art in the area of geo-spatial data science on remote sensing data. - Learn and practice parallel processing to handle Big Data. - Perform data cleaning and data visualization using Python packages. - Implement different machine learning models such as neural network and LSTM. - Evaluate the Gaussian Process Regression model for interpolating sea ice thickness.

Student Background: Students need to have basic computing knowledge and introductory programming skills in Python.