I have suggested in an earlier blog here that the rate of subsidence of the bowl in Bárðarbunga is slowing down as a result of slowing down of magma flow out of the magma reservoir beneath the volcano. Halldór Björnsson has suggested that the lower rate of subsidence is only apparent and due to the flow of the glacier into the bowl. Let us evaluate the possible rate of glacier flow down into the bowl. The bowl is now about 45 meters deep and it has a radius of approximately 5000 meters. Then the slope into the bowl is about 1:110 or about 0.5 degrees. Glaciers do not respond or only slowly in such a low slope. The attached diagram (from Haeberli 1995) shows the slope of glacier surfaces in Antarctica plotted versus response time in years on the horizontal axis. Low slopes, such as the 0.5 degrees in Bárðarbunga, do not even make the page. The response time of the lowest slope shown (5 degrees) is of the order one hundred years. On this basis I conclude that glacier flow is of little or no consequence as yet in modification of the subsidence bowl. Therefore I think it most likely that the changes in subsidence rate are due to changes in the magma reservoir at depth under the caldera.