ALBERT

Description: Large uncertainties characterize forest development under global climate change. Although recent studies have found widespread increased tree mortality, the patterns and processes associated with tree death remain poorly understood, thus restricting accurate mortality predictions. Yet, projections of future forest dynamics depend critically on robust mortality models, preferably based on empirical data rather than theoretical, not well-constrained assumptions. We developed parsimonious mortality models for individual beech ( Fagus sylvatica ) trees and evaluated their potential for incorporation in Dynamic Vegetation Models (DVMs). We used inventory data from nearly 19′000 trees from unmanaged forests in Switzerland, Germany and Ukraine, representing the largest dataset used to date for calibrating such models. Tree death was modelled as a function of size and growth, i.e., stem diameter ( dbh ) and relative basal area increment ( relBAI ), using generalized logistic regression accounting for unequal re-measurement intervals. To explain the spatial and temporal variability in mortality patterns, we considered a large set of environmental and stand characteristics. Validation with independent datasets was performed to assess model generality. Our results demonstrate strong variability in beech mortality that was independent of environmental or stand characteristics. Mortality patterns in Swiss and German strict forest reserves were dominated by competition processes as indicated by J-shaped mortality over tree size and growth. The Ukrainian primeval beech forest was additionally characterized by windthrow and a U-shaped size-mortality function. Unlike the mortality model based on Ukrainian data, the Swiss and German models achieved good discrimination and acceptable transferability when validated against each other. We thus recommend these two models to be incorporated and examined in DVMs. Their mortality predictions respond to climate change via tree growth, which is sufficient to capture the adverse effects of water availability and competition on the mortality probability of beech under current conditions. This article is protected by copyright. All rights reserved.

Description: Dynamic Vegetation Models (DVMs) are designed to be suitable for simulating forest succession and species range dynamics under current and future conditions based on mathematical representations of the three key processes regeneration, growth and mortality. However, mortality formulations in DVMs are typically coarse and often lack an empirical basis, which increases the uncertainty of projections of future forest dynamics and hinders their use for developing adaptation strategies to climate change. Thus, sound tree mortality models are highly needed. We developed parsimonious, species-specific mortality models for 18 European tree species using 〉 90 000 records from inventories in Swiss and German strict forest reserves along a considerable environmental gradient. We comprehensively evaluated model performance and incorporated the new mortality functions in the dynamic forest model ForClim. Tree mortality was successfully predicted by tree size and growth. Only a few species required additional covariates in their final model to consider aspects of stand structure or climate. The relationships between mortality and its predictors reflect the indirect influences of resource availability and tree vitality, which are further shaped by species-specific attributes such as maximum longevity and shade tolerance. Considering that the behavior of the models was biologically meaningful, and that their performance was reasonably high and not impacted by changes in the sampling design, we suggest that the mortality algorithms developed here are suitable for implementation and evaluation in DVMs. In the DVM ForClim, the new mortality functions resulted in simulations of stand basal area and species composition that were generally close to historical observations. However, ForClim performance was poorer than when using the original, coarse mortality formulation. The difficulties of simulating stand structure and species composition, which were most evident for Fagus sylvatica L. and in long-term simulations, resulted from feedbacks between simulated growth and mortality as well as from extrapolation to very small and very large trees. Growth and mortality processes and their species-specific differences should thus be revisited jointly, with a particular focus on small and very large trees in relation to their shade tolerance. This article is protected by copyright. All rights reserved.

Description: Abstract Dynamic vegetation models (DVMs) are important tools to understand and predict the functioning and dynamics of terrestrial ecosystems under changing environmental conditions. In these models, uncertainty in the description of demographic processes, in particular tree mortality, is a persistent problem. Current mortality formulations lack realism and are insufficiently constrained by empirical evidence. It has been suggested that empirically estimated mortality submodels would enhance DVM performance, but due to the many processes and interactions within a DVM, the claim has rarely been tested.Here, we compare the performance of three alternative growth‐dependent tree mortality submodels in the DVM ForClim: (1) a mortality function with theoretical foundation (ForClim v3.3); (2) a mortality function with parameters directly estimated based on forest inventory data; and (3) the same function, but with parameters estimated using an inverse approach through Bayesian calibration (BC). Time series of inventory data from 30 ecologically distinct Swiss natural forest reserves collected over 35+ years, including the main tree species of Central Europe, were used for the calibration and subsequent validation of the mortality functions and the DVM.The recalibration resulted in mortality parameters that differed from the direct empirical estimates, particularly for the relationship between tree size and mortality. The calibrated parameters outperformed the direct estimates, and to a lesser extent the original mortality function, for predicting decadal‐scale forest dynamics at both calibration and validation sites. The same pattern was observed regarding the plausibility of their long‐term projections under contrasting environmental conditions. Our results demonstrate that inverse calibration may be useful even when direct empirical estimates of DVM parameters are available, as structural model deficiencies or data problems can result in discrepancies between direct and inverse estimates. Thus, we interpret the good performance of the inversely calibrated model for long‐term projections (which were not a calibration target) as evidence that the calibration did not compensate for model errors. Rather, we surmise that the discrepancy was mainly caused by a lack of representativeness of the mortality data. Our results underline the potential for learning more about elusive processes, such as tree mortality or recruitment, through data integration in DVMs.