Leaf Traits Explain Shrubs' and Herbs' Functional Composition Shifts in a Climate-Impacted Tropical Montane Grassland

Abstract

Atmospheric models predict increases in temperature worldwide. However, such predictions are not always accurate in high-elevation ecosystems, due to their specific environmental characteristics (e.g., influence of fog in vegetation via foliar water uptake, high atmospheric demand). We propose that leaf morphological traits usually employed in paleoclimate reconstruction (i.e., toothed leaves presence, leaf size categories) could also be used to predict future climate change effects on montane ecosystems. By correlating such leaf traits with mean annual temperature for 90 Southern Hemisphere localities (including a tropical montane grassland in Brazil) we evaluated whether temperature increases (3°C) will result in functional composition shifts in our tropical montane grassland site. Additionally, using a dataset of 63 plant species for foliar water uptake and nine species for leaf water potential, stomatal conductance and turgor loss point, we evaluated if those physiological traits explain the mechanisms behind compositional change in the tropical montane grassland, and how those physiological traits are associated with the presence/absence of toothed leaves. We found that in the tropical montane grassland, an increase of 3°C in temperature by the year 2100 will likely increase the proportion of species with entire leaves from 33.3% to 70.4%. As leaves with entire margins tend to exhibit lower transpiration surfaces, this change could increase the proportion of water-saving strategist species in the vegetation. We also showed that plants with toothed leaves tend to have higher leaf water potentials/higher stomatal conductance, making them possibly more vulnerable to future atmospheric droughts. In summary, an increase in temperature will likely cause an increase in the proportion of drought-resistant species in the tropical montane grasslands. Therefore, models combining leaf traits usually used in paleoclimate reconstruction and physiological traits can be useful for predicting vegetation responses to climate change.