Several earlier posts have discussed the questionable data yielded by unrealistic experiments that slam some environmental change into an ecosystem in a bottle rather than phasing anticipated changes in at a more realistic pace. More on that.

They found that the rate of manmade climate change will not exceed background climate variability over the next few decades. What’s more, the predicted rates of change are too subtle to be mimicked experimentally at present.

The team says this is because phytoplankton need time to respond, and therefore adapt, to any change in climate. "For adaptation to change to occur, the change must be sustained," explained Boyd.

Most studies to date into the effects of climate change on organisms either use results from global model simulations or perform experiments in which the environmental properties of seawater containing phytoplankton are altered rapidly. . .

"We have to reassess our current way of doing experiments to look at climate change – where we alter, overnight, all the environmental properties in seawater to which phytoplankton has been added," said Boyd. "This tells us little about how cells might respond to gradual change due to climate change."

Instead, determining the resilience of different species or groups to an envelope of conditions is a better technique, he added. There are even clues in geological records that could help, and these could be linked with better-designed contemporary experiments.

It's a general problem. Systems - from sea ecologies to human societies - have a suite of responses for coping with natural variability. It is only when change exceeds natural variability and is sustained that those existing coping mechanisms will be challenged, and might then result in adaptation.

Experimental design needs to somehow deal with this truth to yield useful data. Easier said than done.