Survival functions for boreal tree species in northwestern North America

Logistic survival probability models were developed for seven tree species in northwestern North America using Permanent Sample Plot (PSP) data from: six Canadian provincial and territorial governments, the government of Alaska (USA), and four forestry companies; for a total of 1,250,257 trees within 11,673 PSPs. The survival probability of: white/Engelmann spruce (Picea glauca (Moench) Voss/Picea engelmannii Parry ex Engelm.), black spruce (Picea mariana (Mill.) BSP.), lodgepole pine (Pinus contorta Douglas ex Loudon), jack pine (Pinus banksiana Lamb.), trembling aspen (Populus tremuloides Michx.), balsam poplar (Populus balsamifera L.), and balsam fir (Abies balsamea (L.) Mill.) was modeled using tree size (dbh), competition estimates (basal area of the larger trees by species group), climate normal, tagging limits and the time elapsed between consecutive measurements. Survival increased nonlinearly with tree size and the effect of competition on tree survival was related to the shade-tolerance of the species and to stand composition, with shade intolerant conifer species (i.e. lodgepole and jack pine) being more negatively affected by competition compared to shade intolerant deciduous species (trembling aspen and balsam poplar) and shade tolerant spruce species (white and black spruce). Competition from larger spruce (Picea spp.), fir (Abies spp.) and deciduous species (e.g. Populus spp. and Betula spp.) had stronger influences on survival than pine species (Pinus spp.). Intraspecific competition also had significant effects on survival of the majority of the species. Climate Moisture Index provided better results than other climate variables for most species and survival probability increased with increasing values of CMI (i.e. relatively cooler and wetter climate), while for pine species survival probability decreased with increasing CMI and showed higher levels of survival on warmer and drier sites for the range of conditions included in our data. The predictive equations developed in this study could be used to improve the predictive ability of existing growth and yield models such as the Mixedwood Growth Model.