Credit: M. GODFREY

Thresholds are comforting for decision-makers. There is no controversy when a high-jumper makes the bar, in contrast to a figure-skater who wins based on form and execution. When the skater doesn't make the grade, there is endless debate about whether the judges were too harsh and what revisions are needed in scoring procedures.

In personal health, as long as we are alive we can be pretty sure we haven't crossed a threshold of dire consequence. But in many cases, identifying and waiting for thresholds also allows misbehaviour that might be better nipped in the bud. Humans don't die of the first cigarette they inhale, but the slow cumulative effects of smoking can hasten the journey towards one's ultimate personal threshold.

Ecologists believe there are numerous thresholds in nature (Nature 413, 591–596; 2001). As we see anthropogenic changes in the Earth system, we need to decide whether we want to allow human activities to disrupt Earth's life-support processes, or whether to begin now to sustain something that is pleasant and potentially more healthful for humans and the other species that share this planet with us. Ongoing changes in global chemistry should alarm us about threats to the persistence of life on Earth, whether or not we cross a catastrophic threshold anytime soon.

Rockström et al. (Nature 461, 472–475; 2009) guess that an acceptable human impact on the global nitrogen cycle should not exceed 25 per cent of the current anthropogenic transfer of nitrogen from the atmosphere to the land surface. This threshold for nitrogen seems arbitrary and might just as easily have been set at 10 per cent or 50 per cent. Since nitrogen can also be denitrified by soil bacteria and ecosystem remediation is theoretically possible, greater human impacts might potentially be tolerated with proper management (Proc. Natl Acad. Sci. USA 106, 203–208; 2009).

But is a threshold really a good idea at all? In areas of excess nitrogen deposition from the atmosphere — for example, in pastures in Great Britain — species decline linearly as a function of increasing nitrogen inputs to the land (Science 303, 1876 –1879; 2004). Some experimental studies with nitrogen fertilizer show a greater loss of species at low levels of excess nitrogen deposition, with diminishing incremental effects thereafter (Nature 451, 712–715; 2008). Waiting to cross the threshold allows much needless environmental degradation.

Rockström et al. set a lenient limit for acceptable human perturbation of the global phosphorus cycle, suggesting it should not exceed ten times the background weathering of phosphorus. But if we cross a threshold for phosphorus that leads to deep oceanic anoxia, we risk a truly dire situation. And lower levels of phosphorus input have well-documented effects on fresh water, which led regulators to set limits on the phosphorus content of detergents nearly 40 years ago.

Moreover, the background value for phosphorus is difficult to estimate. Rivers now carry an estimated 22 × 1012 grams of phosphorus per year (gP yr−1) to the sea, but an unknown fraction of that is derived from human activities (Treatise on Geochemistry Vol. 8, 585–643; Elsevier, 2005). Not all phosphorus in rivers is reactive; most is bound to iron and aluminium minerals and is rapidly deposited in marine sediments. The current human contribution to reactive phosphorus in river waters (about 1 × 1012 gP yr−1; Amer. J. Sci. 282, 401– 450; 1982) is probably equivalent to the natural background flux, doubling the total reactive phosphorus load and causing difficulties in coastal waters. The total background flux is probably greater than 11 × 1012 gP yr−1, so the suggested tolerable boundary for the human impact would exceed 110 × 1012 gP yr−1, enough to deplete known phosphorus reserves in less than 200 years and certainly not sustainable.

Unfortunately, policymakers face difficult decisions, and management based on thresholds, although attractive in its simplicity, allows pernicious, slow and diffuse degradation to persist nearly indefinitely. Through the Holocene, atmospheric CO2 was nearly constant; nature mitigated the effects of humans. The human impact on the carbon cycle now exceeds the natural buffering capacity of the Earth system, leading to cumulative changes in the environment for life in every corner of the planet. When these changes are more rapid than evolution, extinctions mount and the ability of the planet to support life is diminished (Nature 427, 145–148; 2004). Setting boundaries is fine, but waiting to act until we approach these limits merely allows us to continue with our bad habits until it's too late to change them.

Comment on this article at the Climate Feedback blog