Reversing proves a tricky manoeuvre

Rubber bands show "hysteresis" behaviour: they stretch and then ping back to their original state, but the ping-back follows a different route to the stretch. On Earth ice sheets exhibit classic hysteresis behaviour over geological timescales; as temperature rises they progressively melt, but if temperature then decreases again the ice sheet will not gradually reform. Instead temperature has to revert back to cold conditions before the ice sheet starts building again.

But what about other Earth systems such as ocean currents, clouds and rainfall? Do they respond symmetrically to change or will they too exhibit hysteresis? To investigate these questions Olivier Boucher, from the Laboratoire de Météorologie Dynamique at the Centre National de la Recherche Scientifique in Paris, and colleagues from the Met Office Hadley Centre, UK, used an Earth-system model to examine how various systems behaved when CO₂ was increased fourfold and then smoothly decreased again.

The scientists found that some systems went smoothly into reverse but others showed a distinct time lag, which induces a hysteresis behaviour. "The greatest time lags we see are for ocean heat content, ocean carbon content, soil carbon, and vegetation carbon," explains Boucher. "This was expected. However there are some variables for which the lag wasn't expected, such as surface ocean nutrients. There is also a lag for drought and permafrost extent because these depend on the deep soil which responds slowly to changes."

Some of these time lags were very long. For example the Met Office simulation showed that ocean heat content – and sea-level rise – took up to 120 years of decreasing CO₂ before it even began to respond. "This tells us that sea-level rise will not go away even if you could reduce CO₂," says Boucher. Other variables, like ocean carbon content and soil carbon content were quicker to reverse, but still stayed static for 40 to 50 years.

Many of the behaviours also depended upon location. Temperature and rainfall changes over land decreased reasonably quickly as CO₂ decreased, but over the oceans there were significant time lags – one or two decades – and different patterns on the way down.

Low-level clouds and ocean stratification in the Southern Ocean showed the most surprising responses, with hysteresis effects on timescales of decades – much longer than expected. These findings are published in Environmental Research Letters (ERL).

Boucher and his colleagues don't feel this idealized experiment has many direct policy implications (we are unlikely to be able to reduce atmospheric concentrations of CO₂ in this way in any case) but they say it adds to the body of work calling for early CO₂ emissions reductions.

It does also tell us that the climate system is yet more complicated than we assumed. "Some potential hysteresis effects are not included in climate models at present," Boucher told environmentalresearchweb. "Our findings justify even more complex models with proper treatment of ice sheets, permafrost thawing (and the consequences on CO₂ and methane emissions), and also groundwater recharge."