There has been a flurry of recent commentary concerning Amazon drought – some of it good, some of it not so good. The good stuff has revolved around a recently-completed interesting field experiment that was run out of the Woods Hole Research Center (not to be confused with the Woods Hole Oceanographic Institution), where they have been examining rainforest responses to drought – basically by using a very large rainproof tent to divert precipitation at ground level (the trees don’t get covered up). As one might expect, a rainforest without rain does not do well! But exactly what happens when and how the biosphere responds are poorly understood. This 6 year long field experiment may provide a lot of good new data on plant strategies for dealing with drought which will be used to improve the models and our understanding of the system.
The not-so-good part comes when this experiment is linked too directly to the ongoing drought in the southern Amazon. In the experiment, older tree mortality increased markedly after the third year of no rain at all (with around 1 in 10 trees dying). Since parts of the Amazon are now entering a second year of drought (possibly related to a persistent northward excursion of the ITCZ), the assumption in the Independent story (with the headline ‘One year to save the Amazon’) was that trees will start dying forest-wide next year should the drought continue.
This is incorrect for a number of reasons. Firstly, drought conditions are not the same as no rain at all – the rainfall deficit in the middle of the Amazon is significant, but not close to 100%! Secondly, the rainfall deficits are quite regionally variable, so a forest-wide response is highly unlikely. Also, the trees won’t all die in just one more year and could recover, depending on yearly variation in climate.
While this particular article is exaggerated, there are, however, some issues that should provoke genuine concern. Worries about the effects of the prolonged drought (and other natural and human-related disturbances) in the Amazon are indeed widespread and are partly related to the idea that there may be a ‘tipping point’ for the rainforest (see this recent article for some background). This idea is exemplified in a study last year (Hutrya et al, 2005) which looked at the sharp transition between forest and savannah and related that to the coupling of drought incidence and wild fires with the forest ecosystem. Modelling work has suggested that the Amazon may have two vegetation/regional climate equilibria due to vegetation and climate tending to reinforce each other if one is pushed in a particular direction (Oyama and Nobre, 2003). The two alternative states could be one rainforested and wet like today, the other mainly savannah and dry in the Eastern Amazon. Thus there is a fear that too much drought or disturbance could flip parts of the forest into a more savannah-like state. However, there is a great deal of uncertainty in where these thresholds may lie and how likely they are to be crossed, and the rate at which change will occur. Models go from predicting severe and rapid change (Cox et al, 2004), to relatively mild changes (Friedlingstein et al (2003)). Locally these responses can be dramatic, but of course, these changes also have big implications for total carbon cycle feedback and so have global consequences as well.
Part of that uncertainty is related to the very responses that are being monitored in the WHRC experiment and so while I would hesitate to make a direct link, indirectly these results may have big consequences for what we think may happen to the Amazon in the future.
Special thanks to Nancy Kiang for taking the time to discuss this with me.
Update: WHRC comments on the articles below.