Ecological contingency in species shifts: downslope shifts of woody species under warming climate and land-use change

A predicted impact of a warming climate is an upslope shift of montane plant species. These upslope shifts may be amplified by land-use changes or attenuated by forest recoveries at low elevations where historical disturbances were ceased allowing for plant regrowth. Consequently, species may shift downslope back to low elevations where they had been previously harvested. The cessation-driven downslope shifts are hypothesized to dampen or even reverse climate-driven upslope shifts. We tested this hypothesis by a 20 year (1989-2009) forest inventory dataset from five mountainous areas in eastern China. In our study region, intense deforestation occurred mostly at low elevations until 1970, but was then ceased to facilitate natural forest recovery. Based on the analyses of 30 216 woody plants in 609 plots, we found that: (1) forest recovery occurred over the 20 year survey period, and increment rates of both recruitment and basal area increased up to 2004. However, in the last period (2004-2009), increment rates of basal area leveled off and recruitment was close to zero; (2) forest recovery was faster at lower elevations, as indicated by the higher increment rates there; (3) despite rising regional temperatures, the mean elevations of study species showed a downslope shift over the 20 years; and (4) the contribution of forest recovery to elevational shifts was supported by the fact that the species shifts were positively related to elevational changes in the recruitment increment, e.g. the negative (or downslope) shifts occurred in association with higher increments at lower elevations. These results suggest that, the cessation of disturbances and consequent lowland forest recovery had greater effects on the species distributions than did warming climate. In mountain systems that are being allowed to recover from historical disturbances, the effects of forest recovery on species distributions should be explicitly accounted for when assessing and predicting climate change impacts.