Fig. 6: Long-term effects of an temperature overshoot on the high-latitude water-, energy-, and carbon cycle.

Differences between simulations initialized with the soil carbon concentrations after (P2100) and before (NoOS) an temperature overshoot (OS) that persist under nontransient atmospheric conditions (PACT1.5). Shown are relative differences in total soil water content (blue line; left y-axis), inputs of mineral nitrogen (yellow line; left y-axis) and net primary productivity (green line; left y-axis) as well as absolute differences in total terrestrial carbon (soil organic matter and vegetation biomass; brown line; right y-axis) and MJJASO temperatures at a depth of 1 m (red line; right y-axis). With respect to soil water, nitrogen inputs, productivity, and temperature the figure shows the average over the permafrost-affected regions, while the terrestrial carbon is an accumulated value. The simulations were initialized with different carbon pools and different states of the vegetation, but they started from the same initial conditions with respect to the system’s physical state (NoOS). Hence, the differences indicate only the indirect effects of the OS. All lines show 100-year-running means. It should be noted that in individual grid cells, the simulations may not have reached a steady state within the investigated time-span of 2500 years (see Fig. 4), but when averaged over the entire permafrost region the differences between the simulations stabilize after roughly 1000 years, providing an order-of-magnitude estimate of the consequences of a temporary warming that are irreversible under nontransient conditions.