A methodological review on identification of tree mortality and their applications

Abstract

About 30% of global land areas are covered by forests, which account for nearly 45% of terrestrial ecosystem carbon stock. Thus, the dynamics of forest growth and mortality are critical for regulating the global cycling of matter and energy. In the past years, climate change and its associated disturbance events have profoundly accelerated forest mortality and degradation, leading to negative impacts on ecosystem functions in global forests. Currently, the key approaches to study forest or tree mortality include manipulative experiments, in-situ observations, and ecological modeling. These approaches have been applied from individual, ecosystem, regional to global scales. However, almost all of these approaches have to identify forest or tree mortality based on certain criteria. Unfortunately, due to the complexity of the mortality process, a unified definition or identical criterion for forest and tree mortality is still lack. Thus, it is necessary to summarize which methods have been used to define and identify forest and tree mortality in ecological studies. In this synthesis, we provide a methodological review of different approaches which have been widely used in field investigation, field long-term monitoring, remote-sensing studies, and ecological modeling. For field investigation, tree mortality can be directly identified via tree vigor and defoliation conditions. Two inter-related hypotheses, including hydraulic failure and carbon starvation, are proposed to be the major mechanisms associated with drought-induced mortality. Despite the strong debate over the contribution of carbon starvation during drought-induced tree mortality, the bulk of manipulative experiments and field observations has highlighted the importance of understanding the carbon dynamics and interaction of hydraulic failure and carbon starvation during drought mortality. Forest inventory plots and remote sensing are the two powerful tools to monitor forest and tree mortality on the regional scale. Many countries have established forest inventory plots that allow detailed measurements of demographic process, including mortality. Satellite data are available to map forest mortality by detecting changes in forest vegetation productivity at broad spatiotemporal scales. In global carbon-cycle models, the tree mortality has been represented by diverse algorithms. In general, the mortality algorithms in current models are relatively simple. For example, tree mortality in CLM4.5 is parameterized as a constant proportion of total biomass. In summary, this review synthesizes the available methods for identifying forest and tree mortality based on the published literatures. These methods have been developed for different purposes and their applications are scale dependent. There is an urgent need of explicitly simulating tree mortality in Earth system models. Thus, this study calls for a better understanding on the underlying mechanisms of tree mortality and more research efforts on improving mortality algorithms in global carbon-cycle models.