Wood specific gravity estimation based on wood anatomical traits: Inference of key ecological characteristics in fossil assemblages

Specific gravity (SG) is a key functional trait that has been linked to plant life history traits (e.g. growth rate, life spans) and ecological properties (e.g. cavitation resistance, water transport capacity). SG estimation can thus provide important information about the ecological characteristics of fossil assemblages. Since SG is also related to wood productivity, it can also inform about forest biomass and variation in terrestrial carbon stores over time. Here, we develop a straightforward method to predict SG based on wood anatomical traits that are frequently measured when fossil samples are described. We modeled the relationship of SG as a function of fiber and vessel traits in three extant Mexican communities (135 spp) and selected the best predictive equations. Fiber wall to lumen ratio was the variable most highly related with SG and was present in all the selected prediction models; this variable, together with vessel diameter, was part of the best model. In the validation datasets these models provided good estimates of SG both, at individual sample (overall R2 = 0.63, varying from 0.44 to 0.84 in different localities) and average community levels. We applied these models to predict SG in three Mexican and one Peruvian fossil assemblages that ranged from Eocene to Miocene in age. To determine the reliability of the estimates, we compared them to observed SG from other extant communities and geographic regions. The estimates of two of the Mexican paleofloras were relatively high, most likely due to deficient preservation. In the other two fossil assemblages, El Cien Formation, Mexico and El Bosque Petrificado Piedra Chamana, Peru, the results were consistent with their inferred paleoenvironment and with the current understanding of the vegetation type they represented. The Peruvian paleoflora had relatively low SG suggesting comparatively higher growth rates and wood productivity, and relatively milder conditions compared to El Cien Formation paleoflora.

► We modeled specific gravity as function of wood anatomical traits in extant sites.
► We built predictive models to predict specific gravity in fossil assemblages.
► Predictive models provided good results in the validation datasets.
► In fossils the reliability of the estimates is related to the quality of preservation.
► El Cien and Piedra Chamana paleofloras showed divergent ecological strategies.