The Zwally paper has been accepted but the stodgy Journal of Glaciology does not provide an online copy and there's no telling when it will appear, firewalled, in an issue. Reviewer comments are kept secret permanently.
A copy could like be obtained by emailing Zwally at his NASA profile page. His older papers are available for download at
https://www.researchgate.net/profile/H_Zwally2/publicationsMass gains of the Antarctic ice sheet exceed losses
H. Jay Zwally, Jun Li, John W. Robbins, Jack L. Saba, Donghui Yi, Anita C. Brenner
accepted Journal of Glaciology #15J071
Antarctica mass balance has gone back and forth for years by ±75 Gt, Grace gravity vs laser altimetry. The latter has serious issues with firn modelling. Zwally attributes the difference to how GRACE researchers account for changes in the height of the bedrock (isostatic rebound). We have not yet heard from the other side.
Glacial ice forms as snow compresses under its own weight. In this case, the gain in ice mass in parts of Antarctica has not come from snowfall in the modern era, but from heavier snows that fell about 10,000 years ago, says Zwally. That snow became ice and started to flow slowly towards the sea — but so slowly that the ice began to thicken.
In my view, the mass gain approach makes sense until it doesn't. The issue today is really warm ocean waters lapping at the underside of vast ice shelves. As these come to destabilize grounded ice up-glacier, any slight excess of snowfall over ablation will fall through the floor quantitatively speaking.
The Swally paper is also saying Greenland mass balance has been figured quite wrong, ie it is contributing more to sea level rise than thought. I'm recalling 1/3 which would have to go to 2/3 since thermal expansion is not likely to be misfigured. Yet Greenland, being so much smaller and experimentally accessible than Antarctica, is better studied by these very same two techniques. So how do we get to 100% error in Greenland measurements?