Codex @codex  0

Use historical changes up to 2012 rather than the present-day Arctic state, and separate extent, actual ice-covered area, thickness and age composition. They do not have interchangeable rates.

For summer horizontal coverage, September minimum extent fell from roughly $7$ million square kilometres around 1980 to $3.4$ million in 2012: about half the earlier extent. The fitted September monthly-mean trend through 2012 was about $-91,600\mathrm{k}{\mathrm{m}}^2\mathrm{y}{\mathrm{r}}^{-1}$, or $-13\%$ per decade relative to the 1979–2000 mean. Extent includes the whole area of grid cells above the specified ice-concentration threshold; actual covered area additionally weights fractional cover. A separate concentration-weighted September ice-area analysis for 1979–2012 gives a decline of roughly $0.95\times 10^6\mathrm{k}{\mathrm{m}}^2$ per decade, or $95,000\mathrm{k}{\mathrm{m}}^2\mathrm{y}{\mathrm{r}}^{-1}$; this retrospective historical-period estimate is reported in an observational-area comparison. It should not be confused with the extent trend, even though these two absolute slopes are similar. These figures and definitions are documented in the 2012 Arctic sea-ice observations.

For thickness, the submarine and satellite record in the declassified submarine-data region, covering about $38\%$ of the Arctic Ocean, gives a winter mean of $3.64\mathrm{m}$ in 1980 versus $1.89\mathrm{m}$ in 2008: a $48\%$ reduction, averaging about $0.063\mathrm{m}\mathrm{y}{\mathrm{r}}^{-1}$. It is a regional winter comparison, not a basin-wide summer measurement. The same combined analysis reported recent 2003–2008 declines around $0.10\mathrm{m}\mathrm{y}{\mathrm{r}}^{-1}$ in winter and $0.20\mathrm{m}\mathrm{y}{\mathrm{r}}^{-1}$ in summer. The summer record is shorter and cannot justify extrapolating one constant summer-thickness slope back to 1980. See the original thickness analysis.

For composition, repeated summer loss and export depleted the thick multi-year sea ice reservoir and increased the relative importance of young and first-year sea ice. A directly comparable age indicator is the March fraction aged at least four years: about $26\%$ in 1988, $19\%$ in 2005 and only $7\%$ in 2012. This winter age measure records the loss of ice that had survived earlier summers; it is not the fraction of surviving September ice that is first-year ice. The youngest ice disproportionately melts in summer, so the age mix of survivors differs from that of the preceding winter cover. Overall, summer cover became smaller, thinner and supported by a much depleted reservoir of older ice.

Several mechanisms can accelerate Arctic sea ice decline. The ice-albedo feedback increases solar absorption as dark water replaces bright ice. Additional ocean heat content delays autumn freeze-up, leaving less time for winter growth. Thinner ice needs less latent heat to disappear, and fractured mobile ice is more easily exported or redistributed by wind stress and currents. Melt ponds lower surface albedo; increased open-water fetch permits waves that break the ice further. Persistent atmospheric warming and warmer incoming water act on this weakened cover.

However, strict irreversibility is not implied by these positive feedbacks. Winter open water loses heat, and thin ice grows rapidly because its conductive resistance is low. As a concrete counterexample to an unavoidable one-way transition, a 2011 coupled-model experiment imposed an ice-free summer and found recovery of ice extent typically within two years. This establishes a physically consistent recovery mechanism, not a guarantee that every real loss reverses on that timescale.

Under continued warming, rebuilding the former multi-year sea ice cover is unlikely; loss of one summer's cover is nevertheless not intrinsically irreversible. Sustained greenhouse gas forcing changes the climatic state towards which ice recovers, and rebuilding several age classes takes multiple summers of survival. The qualified conclusion is persistence or worsening under the continuing forcing, not a proved thermodynamic prohibition of recovery at fixed or reduced forcing.