A recent study, undertaken at UCLouvain by Marie Cavitte and co-authors (including our very own Sarah Wauthy), was published in the Journal of Glaciology looking at ice core snowfall records… Now why do we care ? Let’s go back to climate basics!
The edge of the ice sheet is sensitive under global warming because it is in contact with warmer water masses due to increasing temperatures and water mass changes. To predict the future contributions of the Antarctic Ice Sheet to sea level rise, we need to understand the evolution of surface mass balance (the net accumulation of snow at the surface), which can dampen or amplify mass losses at the edge of the ice sheet.
Figure SPM.5 in AR6 report of the IPCC. Simulated (top panel) annual mean temperature change in °C and (bottom panel) precipitation change in %at global warming levels of 1.5°C, 2°C and 4°C (20-year mean global surface temperature change relative to 1850–1900). Note that, high positive percentage precipitation changes in initially low precipitation regions may correspond to small absolute changes.
In Antarctica, surface mass balance is dominated by snowfall, with very little losses due to the sublimation and removal of snow at the surface by surface winds. In a warming world, increasing precipitation is predicted (and already observed!) simply because the air contains more moisture (simple law of physics). Over the Antarctic Ice Sheet, the air temperature is sufficiently cold that precipitation always falls (for now) in the form of snow. According to what I just stated, we would expect an increase in recent records of snowfall.
This figure, from the latest AR6 summary for policy makers, shows the estimates of mass change rates (how fast the gain or loss of ice mass will occur) for the West Antarctic Ice Sheet for surface mass balance (SMB, in blue), ice discharge (ice flowing from the coast into the ocean, in green) and mass balance (SMB minus ice discharge, in red and gray). Note that here surface mass balance (SMB) comes from model simulation output, not observation-based, and according to the physics, predicts the expected increase in surface mass balance due to the increasing temperatures over 1980-2020. Figure credit: Fig 9.16 of Ch9 of the AR6 report of the IPCC.
What we use to measure past surface mass balance (=net snowfall) are ice cores. To measure snowfall over the past 50 years, we don’t need to drill very deep (especially in coastal zones where annual snowfall is high), and we can basically count years visually in the ice core retrieved, as each year includes one dark and one light band (corresponding to the high and low snowfall seasons), or by measuring the water isotopes which can easily identify annual cycles. So we would expect that ice cores that sample the past 50 years, a period during which global temperatures have increased, should show an increase in net snowfall. But that is not what we observe. The ice cores collected around the coast of Antarctica each show different trends: some show increasing snowfall, some decreasing snowfall, and some don’t show any significant changes at all… Which is quite surprising.
Drilling ice cores in the field as part of the Mass2Ant project. Some preliminary measurements are made before packing the ice core for transport. Photo credit: Marie Cavitte.
Map of the coastal ice core sites (red pins) and radar data surveys (black lines) used in the Cavitte et al., 2022, study
Radar waves sent through the snow are reflected due to changes in the electrical properties of the snow it passes through, which can be due to changes in density or in the chemical acidity of the snow. Whenever a change is sufficiently strong, the wave is reflected back to the surface and we observe a reflector in the radar data. Since the radar data is collected continuously over the surface of the ice, we get continuous reflectors throughout the snow column. Now what is interesting is that these reflectors are isochronal: each point along the same reflector has the same age. When the radar passes near an ice core site, the depths of the reflectors can be matched to the ice core depths and we can give ages to the reflectors. These dated radar reflectors are much more extensive in space than a point measurement like an ice core. And the reflectors themselves can be used to reconstruct the past surface mass balance over the entire regions surveyed.
Example of radar data collected along one of the lines used in the Cavitte et al. (2022) study with visible reflectors, a few are highlighted with colored lines. The vertical axis is depth from the surface in meters, the y axis represents distance along the radar line. The inset picture shows the equipped ski-doo during data collection. Figure credit: Marie Cavitte.
What we did in this study, recently published in the Journal of Glaciology, is that we compared the surface mass balance history reconstructed from the entire radar surveyed regions to that measured in the ice cores. We showed that the ice core measurements seem to underestimate the mean surface mass balance signals for the whole regions examined by 7-15 cm of water per year, which corresponds to 8−40% of the ice cores’ mean surface mass balance signal. This implies that ice core surface mass balance records need to be adjusted to represent large surface areas, and radar data is well-suited to quantify the adjustment necessary.
On the other hand, we also showed that the ice cores record the “right” temporal variability of the surface mass balance (the remainder of the climate signal when the mean is removed). And that’s good news because temporal variability is used to understand the variability of the atmosphere, such as e.g. the Southern Annular Mode, El Niño, to name a few, which influence regional climate.
Written by Marie Cavitte