Feedback loop dominance analysis of two tree mortality models: relationship between structure and behavior.

Tree mortality is the least understood process of a tree's lifecycle. Two hypotheses on how mortality progresses in a tree are proposed in the literature: Manion's gradual decline hypothesis and Bossel's sudden death hypothesis. Bossel formulated a mechanism in his model, BAUMTOD, as the cause of sudden death phenomena. BAUMTOD, however, cannot be used to generate a causal understanding of Manion's hypothesis. Therefore, we postulated a causal mechanism for the gradual decline pattern advocated by Manion and modified the BAUMTOD accordingly. The modified model is called BAUMTOD-M. The suggested mechanism concerns the internal imbalance of respiration demand and available photosynthate supply. We then employed a novel approach to analysis of structure-behavior relationships in mechanistic models, called eigenvalue elasticity analysis (EEA), to pinpoint how the simulated tree responds to various stresses. Specifically, we applied EEA to BAUMTOD and BAUMTOD-M to study the likely structural causes behind tree mortality according to Bossel's and Manion's hypotheses, respectively. The analyses of both models suggest that, in the absence of a significant amount of stress, a tree functions as an integrated organism. Growth rates in foliage, feeder roots and respiring permanent biomass are synchronized; however, this coordinated structure is lost if a tree is seriously affected by stress factors. The analyses further suggest that the inability to supply respiration demand plays a crucial role at the onset of mortality in both models. Differences in assumed mechanisms regarding the impact of this inability result in different paths to mortality in each model. This may mean either (1) Bossel's hypothesis is a special case of Manion's hypothesis, or (2) there are truly different mechanisms at work in various observed mortality cases. More data and research are needed to clarify these points.