Assessing the performance of theoretical and empirical tree mortality models using tree-ring series of Norway spruce

摘要: Abstract Successional dynamics of forests under current and changed climate are often investigated using gap models, a subset forest succession models that simulate establishment, growth, mortality trees. However, the submodels largely based on theoretical assumptions, have not been tested in detail. In present study, we compared performance range functions (TMFs) commonly used with several empirical (EMFs) were derived logistic regression from growth patterns tree-ring series as predictor variables. Data dead living Norway spruce ( Picea abies (L.) Karst.) trees subalpine at three study sites Switzerland to this end. Three four EMFs consistently performed better all sites, while TMFs worse than remaining functions. At one site, these correctly classified 71–78% (48–72% for TMFs) 73% (49–64%) 44–54% (21–25%) predicted die within 15 years prior death. 0–2% (7–10%) 5% (19–31%) more 60 last measured year. We conclude that, unless parameters optimized individual species, appropriate predict time tree death, spite their widespread use. A substantial change simulated is be expected if currently implemented replaced by species-specific EMFs.