Polar climate scientist - exploring the causes and consequences of sea ice change using climate models, remote sensing and machine learning.

Research Scientist at the Earth System Science Interdisciplinary Center (ESSIC) of the University of Maryland.

Current research

• Hybrid AI super-forecasting of future sea ice, selected for a 6 month free upgraded access to Google's Antigravity agentic AI.
• Constrained projections of future polar climate change and melt state using CMIP6/CMIP7 data and sea ice observations from NASA's ICESat-2 mission.
• Fusion of ICESat-2 (laser altimeter) and SMAP (L-band) data with Gaussian processes for enhanced sea ice thickness profiling, including ongoing Arctic sea ice thickness analysis
• Improved sea ice state estimation using hybrid machine-learning data assimilation frameworks and novel PREFIRE radiance observations.
• Observing System Simulation Experiments (OSSEs) to define sea ice and land ice altimetry requirements for a future NASA surface topography mission.
• High-resolution commercial satellite imagery for ICESat-2 sea ice surface classification (Co-I, led by E Buckley).

Science teams and service

• IPCC AR7 Working Group I, Chapter 4, contributing author
• Steering group member of the Sea Ice Model Intercomparison Project (SIMIP).
• NASA ICESat-2 Science Team (since 2023; data product lead 2021–2022)
• NASA PREFIRE Science Team (since 2026), SMAP Science Team (since 2024), STV Science Team (since 2022)

Strong proponent of open science — data, code, and papers are in the tabs above. I am also fascinated by the power of AI to accelerate science, and a bit impatient about how slowly academia is talking about it. Some recent thoughts in the blog.

See my CV, Google Scholar, GitHub, ORCID, and LinkedIn.

akpetty@umd.edu