Relating climate, drought and radial growth in broadleaf Mediterranean tree and shrub species: a new approach to quantify climate-growth relationships

Abstract

The quantification of climate–growth relationships is a fundamental step in tree-ring sciences. This allows the assessment of functional responses to climate warming, particularly in biodiversity and climate-change hotspots including the Mediterranean Basin. In this region, broadleaf tree and shrub species of pre-Mediterranean, subtropical origin, have to withstand increased aridification trends. However, they have not been widely studied to assess their long-term growth responses to climate and drought. Since these species evolved under less seasonal and wetter conditions than strictly Mediterranean species, we hypothesized that their growth would mainly respond to higher precipitation and water availability from spring to early summer. Here, we quantified climate–growth relationships in five of these broadleaf species showing different leaf phenology and wood type (Pistacia terebinthus L., Pistacia lentiscus L., Arbutus unedo L., Celtis australis L., and Laurus nobilis L.) by using dendrochronology. We calculated Pearson correlations between crossdated, indexed, mean ring width series of each species (chronologies) and monthly climate variables (mean temperature, total precipitation). We also calculated correlations between the species’ chronologies and a drought index on 7-day scales. Lastly, we compared the correlation analyses with “climwin” analyses based on an information-theoretic approach and subjected to cross-validation and randomization tests. As expected, the growth of all species was enhanced in response to wet and cool conditions during spring and early summer. In some species (P. lentiscus, A. unedo, C. australis,) high prior-winter precipitation also enhanced growth. Growth of most species strongly responded to 9-month droughts and the correlations peaked from May to July, except in L. nobilis which showed moderate responses. The “climwin” analyses refined the correlation analyses by (i) showing the higher explanatory power of precipitation (30%) vs. temperature (7%) models, (ii) selecting the most influential climate windows with June as the median month, and (iii) providing significant support to the precipitation model in the case of P. terebinthus confirming that the radial growth of this species is a robust proxy of hydroclimate variability. We argue that “climwin” and similar frameworks based on information-theoretic approaches should be applied by dendroecologists to critically assess and quantify climate–growth relationships in woody plants with dendrochronological potential.