Scatter plots over time to explore different regimes - do these show a quadratic scaling like Holland et al. (2008) which is used for ice sheet model melt parameterisations?
An aside that might affect this analysis: if the temperature at the ice front is changing, how related is that to the melt rate at the same time?