Integrating climatic response in competition dependent tree-level growth models for northern hardwoods.

摘要

With increased rates of climate change, it is imperative for forest managers to have access to models that can take into account the expected effects of climate change on tree growth. To this end, growth function are sometimes used that include climatic variables such as mean annual temperature or precipitation averaged over decades. Such growth models are usually relatively easy to develop but they do not take into account the fact that tree diameter growth on a given year is determined not by climatic conditions that prevailed up to 30 years before but mainly by climatic conditions that prevailed during the current and previous year. Our objective is determine if including climatic variables obtained from dendroclimatic response function will lead to growth models having a better fit to data than versions with 30-year average climatic conditions, or no climate at all. Growth models were developed for Betula alleghaniensis, Acer saccharum, Acer rubrum and Fagus grandifolia using data from south-eastern Quebec. Three types of growth function were compared. A first set of growth function was developed in which the potential growth of a tree was modeled as a function of tree size and site characteristics (vegetation type and drainage) to be further modified as a non-linear function of plot basal area. The effect of climate was not explicitly accounted for in this fort set of growth function, therefore they will be referred to as Climate-implicit models. A second set of growth function was developed in which we explicitly accounted for the effect of climate by incorporating 30-year mean annual temperature and precipitation in the growth function. In a third type of growth function, also climate-explicit, we incorporated the most significant recent climatic variables identified using climatic response function developed for each species based on dendrochronological and climatic data. The three types of models were compared based on the Akaike information criterion (AIC). Our results showed that Climate-explicit growth models with climatic variables obtained from response function analysis outperformed other growth models for three out of four species (B. alleghaniensis, A. saccharum and F. grandifolia). Incorporating climate in the form of 30-year average climatic conditions brought some improvement over a non-climatic function for A. rubrum, but this was not the case for other species. Accounting for growth dependency on climate by including recent monthly climatic variables provided by response function could be a potentially useful approach for the development of a new lineage of tree growth models dealing with climate change.