Modelling tree height to assess climatic conditions at tree lines in the Bolivian Andes

One of the major problems in assessing growth conditions of trees at upper tropical tree lines is that the vast majority of tree lines have been influenced or destroyed by human activities. Confronted with this situation, we modelled tree height of Polylepis trees in the Bolivian Andes correlating field data from 167 Polylepis stands and a climatic model at a resolution of 925 m × 925 m, across a wide range of elevations and tree growth conditions. Overall, mean annual temperature showed the strongest relationship to tree height (r2 = 0.73), closely followed by a temperature index summing up temperatures above 6.5 °C (r2 = 0.71) and elevation (r2 = 0.68). Mean annual precipitation (r2 = 0.32), humidity balance (r2 = 0.27), potential evaporation (r2 = 0.26), and solar radiation (r2 = 0.18) all showed much weaker relationships to tree height. However, when we divided the Bolivian Andes into three contrasting bioclimatic regions and used residual tree height (against temperature) as a variable, in the humid NE Bolivian Andes solar radiation was significantly positively related to residuals of tree height, in semi-arid SE Bolivia humidity balance and mean annual precipitation had positive and solar radiation negative influence, and in the arid W cordillera mean annual precipitation had positive influence. Our modelling approach thus confirms the overall predominance of temperatures in determining tree height growth. This temperature-determined pattern is regionally modified by positive effects of increasing water availability in arid regions and of increasing solar radiation in very humid and cloudy regions. This suggests that tree growth is primarily limited by temperatures (ca. 70% of the variance explained), but that water and light availability also play a role (an additional 5%). Accordingly, it appears that there may not be one constant factor limiting tree growth at high elevations, but rather that different tree lines may be affected by different combinations of climatic factors which in turn lead to different ecophysiological effects.