GeoCryoAI Model for Permafrost Thaw and Carbon Flux in Alaska

1963-2022 data supports an AI-driven ensemble learning model, GeoCryoAI, designed to quantify permafrost thaw dynamics and greenhouse gas emissions in Alaska. The dataset includes pre-processed input data from in-situ measurements, remote sensing, and process models, resampled to a 1-km grid. It was developed by ORNL_CLOUD and contains model code, input data, sample results, and documentation.

Use Cases

• Calibrating and validating process-based models like SIBBORK-TTE for thaw subsidence using field-measured active layer thickness.
• Analyzing spatial patterns of CO2 and CH4 flux derived from TCFM-Arctic and assimilated chamber, eddy-covariance, and tower measurements.
• Training hybrid machine learning frameworks on standardized, 1-km grid data combining thaw depth, active layer thickness, and thaw subsidence.
• Studying the permafrost carbon feedback by integrating disparate data sources from networks like CALM, GTNP, ITEX, and AmeriFlux.

Strengths

• Integrates data from multiple authoritative sources including CALM, GTNP, ITEX, AmeriFlux, and NEON networks.
• Data is pre-processed and standardized on a consistent 1-km grid, resolving disparate spatiotemporal sampling.
• Covers a long temporal range from 1963 to 2022 for analyzing climate change impacts.

Limitations

• Key dataset metrics like row count, size, and specific column names are unknown across all platforms.
• Conflicting last updated dates exist between sources (2026-03 13 and 2022-12-31).
• License is ambiguously listed as 'other-license-specified' without explicit terms.

Provenance

Source

ORNL_CLOUD

Collection Method

Data assimilated from in-situ measurements (mechanical probing, borehole temperatures, flux towers), remote sensing platforms (UAVSAR, AVIRIS-NG), and process-based modeling products.

Time Range

1963-2022

Freshness

2026-03-13 02:34:55.443580

Geography

Alaska