This is a nice short and pleasingly informal summary of the latest work on predictions of loss of summer sea ice in the Arctic, published in GRL. The paper is perhaps more useful as an up-to-date literature review than an analysis paper, which is perhaps a kinder way of saying what Stoat says.
The paper points to three main methods of prediction: 1) extrapolation of sea ice volume data by the Trendsetters, 2) assuming a future that includes loss events like those seen recently by the Stochasters, and 3) projections using GCMs by the Modelers. The paper suggests that the central estimates of these three methods for a nearly ice free summer Arctic are 2020, 2030, and 2040 respectively: worryingly different from the estimate of 2070 made by the IPCC AR4 just a few years ago. They support the conclusion that “society should start managing for the reality of climate change in the Arctic”. In short No one is predicting a recovery of summer ice.
The authors conclude that it is currently impossible to choose between the methods, as all have their strengths and weaknesses. They call for particularly for improved modeling and modeling approaches. They seem to prefer model intercomparison projects to “results provided from a large number of modeling centers produced under short time schedules” (I’m inclined to agree), and make it clear that CMIP5 models are unlikely to be adequate to accurately predict the onset of ice-free summers.
There are some nice sections in the paper – one neatly describes the difficulty of comparing model simulations with observational data, and offers a partial solution:
Global climate models (GCMs) are often run several times, referred to as ensemble members, with slightly different initial conditions to simulate a possible range of natural variability in addition to steady increasing greenhouse gas forcing. Data, in contrast, are a single realization of a range of possible climate states. Observations confound signal (global warming forcing) and noise (natural variability). Thus, it is not completely valid to compare the ensemble mean of a model or several models, which could be considered the expected value of the climate state, with the single data realization. A better approach is to look at the range of ensemble members and to determine if the data time series could be considered a possible member of the population of ensemble members. Unfortunately, there are seldom enough ensemble members to test this hypothesis.
It is also worth checking out section 4, which starts to get at the relationship between model discrepancy (bias, deficiencies) and future model behaviour. The paper is short and useful enough that I don’t worry that there don’t appear to be new results in it.
Overland, J. E., and M. Wang (2013), When will the summer Arctic be nearly sea ice free?, Geophys. Res. Lett., 40, doi:10.1002/grl.50316.