Significant contribution of small icebergs to the freshwater budget in Greenland fjords

Soroush Rezvanbehbahani,C J Van Der Veen,Leigh A Stearns,Siddharth Shankar,Ramtin Keramati

doi:10.1038/s43247-020-00032-3

Abstract

Icebergs represent nearly half of the mass loss from the Greenland Ice Sheet and provide a distributed source of freshwater along fjords which can alter fjord circulation, nutrient levels, and ultimately the Meridional Overturning Circulation. Here we present analyses of high resolution optical satellite imagery using convolutional neural networks to accurately delineate iceberg edges in two East Greenland fjords. We find that a significant portion of icebergs in fjords are comprised of small icebergs that were not detected in previously-available coarser resolution satellite images. We show that the preponderance of small icebergs results in high freshwater delivery, as well as a short life span of icebergs in fjords. We conclude that an inability to identify small icebergs leads to inaccurate frequency-size distribution of icebergs in Greenland fjords, an underestimation of iceberg area (specifically for small icebergs), and an overestimation of iceberg life span.

Highlights

Icebergs represent nearly half of the mass loss from the Greenland Ice Sheet and provide a distributed source of freshwater along fjords which can alter fjord circulation, nutrient levels, and the Meridional Overturning Circulation
Because identifying iceberg edges is crucial for frequency–size distributions, we apply a weighted cross-entropy loss function that penalizes the false detection of edges five times greater than false detections of nonedge regions
The frequency–size distribution of icebergs in fjords is often expressed as power-law when iceberg formation is mostly fracturedominated and log-normal when dominated by iceberg melt[2,4]

Summary

Introduction

Icebergs represent nearly half of the mass loss from the Greenland Ice Sheet and provide a distributed source of freshwater along fjords which can alter fjord circulation, nutrient levels, and the Meridional Overturning Circulation. Thresholding methods can be applied on high-resolution Planet imagery, our analysis using global thresholding and Otsu thresholding methods shows that the frequency–size distribution of icebergs cannot be accurately obtained with generic thresholding methods (see Supplementary Fig. 3), highlighting the superiority of convolutional neural networks for iceberg detection. The total number of small icebergs in Planet imagery is nearly five times greater than those in Sentinel-2 (Fig. 2a, b).

Results

Conclusion