Machine learning and aerogravity regression analyses produce the first continent-wide mapping of Antarctic sedimentary basins. The project, funded under the American Rescue Plan Act, tests the hypothesis that these basins correlate with locations of high ice flow velocity. Results and open-source code were disseminated by AMD_USAPDC in 2023.
Use Cases
- Train ensemble machine learning algorithms, like randomized decision trees, to identify unknown sedimentary basins from known basin locations and geophysical compilations.
- Calculate gravity residuals from aerogravity data to propose new sedimentary basin locations based on density inhomogeneities.
- Compare gravimetrically identified basins with independent magnetic source depth estimates from the Werner deconvolution method.
- Correlate mapped sedimentary basin locations with ice velocity data to test their influence on fast ice flow behavior.
- Use the basin mapping to provide improved basal parametrization conditions for ice sheet models predicting sea level rise.
Strengths
- First continent-wide mapping of onshore and offshore sedimentary basins over Antarctica.
- Combines multiple large-scale geophysical datasets and independent validation methods.
Limitations
- Specific row counts, column details, and sample sizes for the underlying data are unknown.
- The dataset's creation relies on regression and machine learning inference, introducing potential model-dependent biases.
Provenance
- Source
- NASA Earthdata, from the organization AMD_USAPDC.
- Collection Method
- Machine learning applied to geophysical compilations and aerogravity data regression analyses.
- Time Range
- null
- Freshness
- Last updated in August 2023.
- Geography
- Pan-Antarctic coverage.