Predicting Antarctic Net Snow Accumulation at the Kilometer Scale and Its Impact on Observed Height Changes.

Area of Science:

• Glaciology
• Atmospheric Science
• Remote Sensing

Background:

• Sub-grid-scale processes near ice sheet surfaces, like snow redistribution and sublimation, are crucial for net accumulation but are often oversimplified in large-scale models.
• Observational challenges at the ice sheet surface limit the accurate representation of these critical processes.

Purpose of the Study:

• To develop high-resolution (1 km) models predicting net snow accumulation variability on the Antarctic Ice Sheet.
• To quantify the impact of local surface processes on net accumulation compared to existing atmospheric models.

Main Methods:

• Training random forest (RF) models using atmospheric variables and topographic data at 1 km resolution.
• Utilizing in situ and airborne radar data for observed net snow accumulation to train and validate the RF models.

Main Results:

• Local net accumulation was found to deviate by up to 172% from the mean predicted by atmospheric models.
• A spatial correlation was observed between predicted net accumulation variability and satellite-derived surface-height changes.

Conclusions:

• Surface processes significantly influence Antarctic net snow accumulation, leading to substantial local deviations from larger-scale model predictions.
• The study underscores the need to incorporate detailed surface process dynamics to accurately model ice sheet mass balance and its response to climate variability.