- This repository uses NASA's Airborne Topographic Mapper (ATM) surface roughness product (ILATM2) to create an ice sheet-wide gridded roughness product, characterize its spatiotemporal patterns, and identify the major climatic controls affecting the observed patterns.
- See 'Repository Contents' for more info on the breakdown of acquisition, processing, and analyses workflows.
- See 'Background' for more info on surface roughness and the OIB ATM.
-
/AcquisitionAndPreprocessing - Script for downloading and pre-processing the ILATM2 product.
-
/VariographyAndKriging - Script for interpolating and mapping gridded roughness for each year.
-
/alongTrackAnalyses - Script for analyzing the spatial and temporal trends in the along-track data.
-
/FilterILATM2byAWS - Script to prepare ILATM2 for random forest regression.
-
/RandomForestRegressionForILATM2AndAWS - Script for random forest regression.
Surface roughness is a key factor in the turbulent heat flux of the Greenland Ice Sheet (GrIS), significantly impacting the energy input to the upper snow surface and, consequently, the ice sheet mass balance. Despite its importance, predictive models often oversimplify surface roughness, which is essential for an improved understanding of GrIS sensitivity to climate change. Surface roughness is oversimplified largely due to the scale-dependency of roughness and the lack of standardized roughness parameters. For additional reading on these complications, I recommend 'Roughness in the Earth Sciences' by Mark W. Smith.
The ILATM2 product is a single-scale roughness estimate - the root mean square fit of 30 x 80 meter segments of along-track elevation shots. ILATM2 documentation can be found here: https://nsidc.org/data/ilatm2/versions/2. Using the springtime (March, April, May) campaigns from 2009 to 2019, this repository includes scripts to produce and map yearly gridded products for meter-scale roughness, supplemented with along-track analyses. With daily climate variables from PROMICE and GC-NET automated weather stations (https://promice.org), these scripts also use machine learning to examine potential climatic drivers of observed roughness patterns.