Well, this is terrifying. I'm attaching my volume anomaly graph, which I've been producing on and off for the last year. As before, this shows the anomaly (or residuals) after accounting for both the linear trend and the seasonal variation in ice volume. The anomaly decreased (became more negative) through October and November, which is highly unusual. It trended slightly down in January, meaning that the current volume is 2 thousand cubic km below what would have been expected based on the downward linear trend for January.
The graph doesn't have a key, so every year since 1979 is shown as a thin black line. Years starting in 2010 are in thicker colored lines, in rainbow order, so red is 2010 and the purplish color is 2016. The colors are changed from previous graphs because of the new year. January 2017 is the thick black line. The volume anomaly is in thousand of cubic km.
I have sometimes posted about the time series model I use to predict ice volume. The prediction from last January for this January's volume was 16.4 thousand cubic km. Since the actual value is 14.6 thousand cubic km, I think we can call that a miss. Most of the error was due to the November volume, which came in 900 cubic km below my prediction based on the October data. Looking forward, the model predicts a monthly volume of 20.8 thousand cubic km for April and 3.6 thousand cubic km for September. If these hold, that would be a record low maximum by over 1.5 thousand cubic km, and also a record low minimum, barely edging out the minimum in 2012.
This is a little unnerving, because the model is generally somewhat conservative. When observed values are lower than predicted, as they have been, future predicted values generally discount the low observations and continue to predict higher values. In order to currently predict a new record low, recent observations must be far below the predictions.
The predictions have a substantial uncertainty, especially for September since it's 8 months in advance. While this means that the actual ice volume could come in well larger than these predictions, the volume could also be substantially smaller than the predictions.