Fig. 3: Productivity–climate associations.
Season–provenance diagrams: climatic sensitivity traits (CStraits), represented by t-value scores of the generalized additive mixed models (GAMM) linking annual NPP and annual seasonal temperature, soil moisture and snowfall fluctuations, by population and common garden site a Mont-Laurier (n = 42 independent populations) and b Chibougamau (n = 45 independent populations). A positive influence of a climate variable on NPP is denoted in blue; a negative influence is denoted in red. The period of analysis was 1976–2016. Studied populations (rows) are ordered by increasing mean annual growing degree days above 5 °C at provenance (GDD5p; color scale at left from warm to cold; legend at far right) origin. Seasons in capital letters represent the current year of ring formation; seasons in lower case represent climate variables during the year preceding ring formation. Significant t-value scores at the 5% level (two sided) are represented by dots. Vertical-bar plots: Tests for genetic differentiation and clinal variations in t-value scores using generalized additive models (GAM). A high adjR2 value denotes a high goodness-of-fit between NPP and the explanatory variables GDD5p, mean annual precipitation (MAPp), admixture proportions along Western, Central, and Eastern genetic clusters (gWest, gCenter, gEast), and spatial factors (sp) represented by the distance-based Moran’s eigenvector maps. All shown variables are significant at the two-sided 5% level.
