Looking backwards : using tree rings to evaluate long-term growth patterns

Abstract

To improve our understanding of the ecology of tropical forest trees, it is essential to obtain information on tree growth over periods of decades to centuries. Using tree-ring analysis such data can be derived as this technique allows reconstructing the growth history over the entire lifespan of a tree. This PhD thesis reports on the use of tree-ring analysis for reconstructing long-term growth patterns of Bolivian forest trees. A large number of discs was collected from five tree species, and tree rings were identified and measured. First, I evaluated whether those trees that are currently large have had faster rates of diameter growth when they were small compared to present-day small trees. This ‘juvenile selection effect’ would imply that fast-growers have a higher probability to reach the forest canopy and become reproductive. For three out of five species, I could indeed detect a juvenile selection effect. Thus, for these three species fast growth of small trees may be essential to reach the canopy. The above finding has consequences for growth models that are often used to simulate tropical timber yield. If fast-growing trees have a higher chance to reach the size at which they can be harvested, it is important to include these higher growth rates in such models. I found that this was indeed the case: simulated timber yield was higher when this juvenile selection effect was taken into account. Nevertheless, even with the higher growth rates, the recuperation of timber volume during one logging cycle remained low. Only 20-33% of the timber volume harvested at the first harvest could be obtained at second harvest after 20 years. Reaching the forest canopy for tropical trees implies large investments in height. Such investments may imply that less carbohydrates are invested in diameter growth and stem volume. I collected stem discs at various heights of juvenile trees to reconstruct height, diameter and volume growth. I found that individual trees differed very strongly in growth rates (height, stem volume, diameter). These growth rates were related: the faster height growers grew relatively little in stem volume, and vice versa. This suggests that juvenile trees in favourable (light) conditions invest relatively more in stability and crown development than in height growth Tree-ring analysis allows evaluating whether diameter growth rates have increased or decreased over time. For four out of five species I found that growth rates of small trees increased over the last two centuries. This pattern is consistent with what would be expected due to CO2-fertilization, but other causes cannot be ruled out. I then checked for indications of changes in forest dynamics over time. To this end, I calculated the frequency of releases (periods of fast growth) and autocorrelation strength. No indications for a change in forest dynamics were found. In this thesis I showed that tree-ring analysis can strongly contribute to improving our understanding of long-term ecological processes in tropical forests. This knowledge is much needed in assessing the response of tropical forests to the predicted climate changes in the future.