Response to Lewis et al.: The uncertain response of tropical forests to global change

I thank Lewis et al. for their letter [1] in response to my recent review article in TREE [2], but argue here that the response of tropical forests to global and regional anthropogenic change, hereafter anthropogenic change, is highly uncertain despite evidence marshaled elsewhere [1,3,4]. Here, I pose six questions to pinpoint sources of this uncertainty. First, are old-growth tropical forests carbon sinks? Repeated censuses of small tropical forest plots detect increasing aboveground biomass (AGB), which suggests that they are [3]. By contrast, models of latitudinal atmospheric CO2 concentration gradients that incorporate ocean processes identify a large CO2 source in tropical and southern lands and a larger CO2 sink in tropical oceans (https://www.icdc7.com/proceedings/abstracts/ jacobson1FF224Oral.pdf). One of these assessments is thus incorrect. Second, are apparent AGB increases real? Many sources of artifact have been debated [1–4], whereas others have not. Specifically, many plots were first censused by one investigator and re-censused by others using ‘progressively refined’ protocols [3]; thus it is difficult to compare such results accurately. I would like to see an analysis that is limited to censuses conducted by the same lead personnel using the same protocols within each plot. Such an analysis has yet to be presented. Third, if AGB increases are real, would they necessarily imply a response to anthropogenic change? Assuming that tropical forests were in equilibrium before anthropogenic change, AGB increases become evidence for increases in some limiting resource (e.g. CO2, nitrogen or light), which is attributed to anthropogenic change [3,4]. However, the assumption is certainly false for forests that are recovering from disturbances, where AGB will increase because the forest is regenerating. Fourth, is disturbance widespread? Hurricanes, downbursts associated with convectional storms, drought and/or fire disturb most tropical forests repeatedly during each tree generation [2,5,6]. Anthropogenic disturbance is also likely to have been a common occurrence for some time; for example, up to 100 000 000 people inhabited the Americas in 1491, when neotropical forests were widely managed for useful trees or converted to agriculture [7]. Thus, diverse disturbances disrupt forests repeatedly, even on continental scales. Fifth, do tree-plot censuses distinguish between responses to anthropogenic change and recovery after