Document Type


Publication Date


Publication Source



In a recent report on the patterns of tree species richness in eastern and western North America, Marks et al. (2016) claimed to have identified an operational indicator of environmental harshness (maximum tree height) and concluded that environmental stressors that limit tree height also act as ecological filters on species richness. Marks et al. (2017) attributed the positive association between species richness and maximum tree height to both the direct effects of environmental harshness on species richness and the indirect effects of environmental harshness on species richness as mediated by maximum tree height.

This finding overlooked the fact that many environmental stressors such as cyclonic disturbance affect tree height and tree species diversity in different directions. In a study of elevational patterns in Taiwan, Chi et al. (2015) reported sharply contrasting relationships between tree species diversity and canopy tree height in sites that were subject to tropical cyclone disturbance vs. those that were not. In the mountains of southeastern China beyond the reach of tropical cyclones, both tree species richness and canopy tree height decreased with increasing elevation (Zheng et al. 2004, Chi et al. 2015), supporting the harshness hypothesis (Marks et al. 2016, 2017). In contrast, in Taiwan, where tropical cyclones occur annually, tree species richness decreased but maximum tree height increased with increasing elevation, the opposite of the predictions of the harshness hypothesis (Fig. 1). We attributed the contrasting elevational patterns and associations between tree diversity and canopy tree height in Taiwan to topographic mediation of tropical cyclone disturbance. The shorter tree stature in lower elevations was attributed to more severe tropical cyclone damage (Chi et al. 2015). Although tropical cyclones limit tree height, tree mortality is very low, possibly a result of both evolutional and ecological responses of these forest ecosystems through the long-term interaction between cyclones and the forest ecosystems (Lin et al. 2011). As an example, multiple category three tropical cyclones on the Saffir-Simpson scale (Simpson and Riehl 1981) caused <2% tree mortality in low-elevation evergreen broadleaf forest in northeastern Taiwan in 1994, a record year of tropical cyclone frequency and intensity (Lin et al. 2011). However, taller trees were selectively killed and defoliation was severe, both of which contributed to the low stature of the forest even though the mean annual temperature (18°C) and precipitation (3800) mm are high (Lin et al. 2011). Cyclone disturbance limits vertical development of trees but does not lead to their elimination. Thus, more tropical cyclone disturbance at lower elevations overrides climatic controls on elevational patterns of tree height but does not change the elevational pattern of tree species richness. We suggest that there is an important difference between actual maximum tree height and potential maximum tree height in the presence of disturbance (e.g., tropical cyclone).



Document Version

Published Version


This open-access article carries a Creative Commons Attribution License (CC-BY). Permission documentation is on file. DOI:


John Wiley & Sons