Contrasting responses of pollen and fruit to whole-tree heating in two tropical savannah species

The Brazilian Cerrado - the world's most biodiverse savannah - is rapidly warming, potentially threatening thousands of species as well as large-scale water supply and carbon storage services. Predicting possible impacts requires a better understanding of plant physiological responses to increasing temperatures, yet direct investigations of wild and tropical species are lacking. Sexual reproduction is considered particularly sensitive to temperature stress and high-temperature related reductions in yield are well documented in crop species. Here we report the first in situ experimental heating of entire, mature individuals of tropical woody species aiming to investigate the impacts of warming on reproductive success. We expected elevated temperatures to reduce both pollen viability and fruit set.

Two common and widespread Cerrado species - Byrsonima pachyphylla (Malpighiaceae) and Davilla elliptica (Dilleniaceae) - were heated by 2-3°C during the daytime throughout reproductive development. Viability of pollen produced (analysed through differential staining) and fruit set (calculated as a proportion of hand-pollinated flowers) were compared between heated and unheated individuals. Pollen viability values were also compared with those of the previous year (prior to heating).

Pollen viability unexpectedly increased in both species, indicating a high temperature resilience of pollen development. In contrast, fruit set (in B. pachyphylla) declined under heating as anticipated. Results suggest that high temperatures constrain reproduction between pollination and early fruit development, while divergent responses of pollen and fruit may indicate greater sensitivity of female than male development, unusual in studied species.

Given already low natural levels of fruit set, observed reductions in B. pachyphylla under predicted levels of warming - perhaps replicated in other Cerrado species - have serious implications for future recruitment, potential migration to more favourable environments, species persistence, interacting species, and community structure. Our results therefore emphasise the urgent need for systematic research into heating impacts on tropical diversity.