Optimization and validation of universal response functions for interior spruce (Picea glauca, Picea engelmannii, and their hybrids)

Ecological models, including various forms of population transfer and response functions, have been widely employed to predict the performance of forest trees under future climate conditions. However, the accuracy of model predictions remains a challenge due to the lack of systematic model optimization and validation processes. In this study, we explored a series of candidate universal response functions (URFs) based on comprehensive provenance trials for interior spruce, a tree hybrid complex of great ecological and economic importance in British Columbia. We first developed URFs with various numbers of climate variables and their combinations, then evaluated and validated these models with conventional statistical methods. Although these models showed high predictive power and performed well in statistical validation, the map visualizations of these models varied substantially. Therefore, we further conducted a pixel-wise spatial validation based on site productivity data to identify the model with the best performance. Finally, we analyzed the impact of climate change on interior spruce based on the final model. Our results revealed that: 1) a minimum number of independent climate variables is required to generate a reliable and informative ecological model; 2) spatial validation is critical to ensure the credibility of ecological models; 3) among-population variation resulting from local adaptation is relatively weak, indicating a potential for modest productivity gains from selecting well-adapted populations for stand establishment; and 4) most of the current interior spruce habitat is projected to be vulnerable to anthropogenic climate change, largely due to temperature rise, with a major upslope shift predicted.

Automated summary

This study developed a series of universal response functions (URFs) based on comprehensive provenance trials for interior spruce. The results showed that a minimum number of climate variables is required to generate a reliable and informative ecological model. Spatial validation is critical to ensure the credibility of ecological models. The current interior spruce habitat is projected to be vulnerable to anthropogenic climate change, largely due to temperature rise, with a major upslope shift predicted.