Trait-based approaches to paleoenvironmental reconstructions: What is in the community-weighted trait mean?

Trait-based approaches to paleoenvironmental reconstructions offer a powerful framework for inferring past climates and environments. While analytical methods have advanced empirically, their evolutionary-theoretical basis is rarely explicitly incorporated into methodological developments. Here, we integrate trait-based palaeoecological modelling with the Law of Energy Equivalence that follows from the Red Queen's hypothesis to refine how community-weighted trait means are computed. Using traits from present-day large herbivorous mammalian communities, we test four methods of weighting community trait means – by species, relative abundance, biomass, and energy intake to predict annual precipitation, mean annual temperature, and terrestrial net primary productivity at the continental scale. Our results show that energy intake-weighted traits provide the most accurate predictions in most environments in line with theoretical expectations, closely followed by species-weighted mean traits, while relative abundance-weighted traits perform best in climatically extreme sites. Refining trait-based methodologies for paleoenvironmental reconstructions contributes to a broader understanding of ecosystem dynamics across time.