Abstract ID: 3.10432 | Accepted as Poster | Talk/Oral | TBA | TBA

The elevation of the climatic upper treeline can be predicted with astonishing accuracy based on climate information alone, hinting at a common cause for treelines across climate zones. However, although the overall accuracy of climate-based treeline predictions is very high, our validation of such a prediction (the TREELIM model by Paulsen & Körner (2014), implemetned based on CHELSA clmate data by Karger et al. (2019)) showed clear geographical patterns in the model bias. Thereby a lack of observed trees at the predicted treeline might be due to factors lowering tree cover (e.g. disturbances, lagged responses to climate change) and does not necessarily indicate an inaccurate model prediction. However, in some regions trees and forest were observed far above the predicted treeline, which is, by definition, not possible. Regions showing this pattern included very wet regions and those with very low annual thermal seasonality (Patagonia, the Canadian Coast Mountains) and regionswith very high solar radiation inputs (parts of the Tropical Andes, including the Bolivian Altiplano). This phenomenon still awaits an explanation, but we will present a number of suggestive correlations and speculations, hoping to encourage discussion on the limitations of models and climate data and on the causes of global treeline.