Our research is divided into four focus areas:
1) Data, 2) Data and Model, 3) Prediction, 4) Scalability.
Below we describe each focus area and the corresponding research projects.
In this focus area, we will investigate convergent data science approaches for fusion of heterogenous data , reducing noise from our data , annotation via advanced visualization , and advanced auto-annotation techniques with human in the loop . This focus area will prepare our dataset for further investigation of ML and physical models in Focus Area 2 and Focus Area 3.
Data & Model
In this Focus Area, we will develop convergent physics informed machine learning method for generating the 3D model of ice bed , tracking the internal layers of ice sheet and integrating them in ice dynamic model , and investigating causal relationship between ice sheets, sea ice and atmosphere.
In this Focus Area we will develop ML algorithms to forecast ice sheet mass loss on a century-scale, extract spatiotemporal patterns in surface mass balance, hydrology, and atmospheric drivers and spatial-temporal pattern mining for data-driven understanding of ice dynamics.
In this focus area we will utilize supercomputers to scale up all of our previous algorithms. Utilizing computational power enabled by advanced cyberinfrastructure is necessary to harness the vast and diverse data from polar regions.