Long-term forest development prediction requires climate-sensitive models of tree growth, mortality, and recruitment. This is especially true in the Iberian Peninsula, where rising aridity and extreme climate events can significantly impact tree growth, survival, and regeneration. Trees and forests adapt to their local growing conditions, so differences in growth between sites do not directly reflect the effects of climate change over time. This study introduces a method to separate the impacts of site conditions and climate variability on forest dynamics. The ratio of mean annual temperature or precipitation to the long-term average at each location was used to represent temporal effects. Site effects were characterized by the long-term average climate metrics at each site. The models were calibrated for Catalan Pinus sylvestris using an optimization-based approach. Climate data and periodic measurements from sample plots in the Spanish national forest inventory served as inputs. Precipitation was the most influential climatic factor for annual diameter growth and survival probability. The interaction between temperature and precipitation was also important; increased precipitation generally enhanced growth and survival, especially during warmer years and in warmer sites. The impact of yearly precipitation variations was more pronounced than site-specific conditions. Conversely, for tree survival, site effects were more influential than yearly climate fluctuations. For two trees of equal diameter growing with similar competition, survival was highest in moist, warm sites and lowest in dry, cool sites. The set of individual-tree models sensitive to climate over one year significantly improves forest management planning in Spain amid climate change. • Temporal and site effects of climatic variables on tree growth are decoupled. • Precipitation is the most influential climatic variable in Catalan Pinus sylvestris. • Precipitation interacts with temperature. • Optimization was used to obtain annual models from periodical growth data.
Pukkala et al. (Wed,) studied this question.