Soil surface temperature, an important driver of terrestrial biogeochemical processes, depends strongly on soil albedo, which can be significantly modified by factors such as plant cover. In sparsely vegetated lands, the soil surface can be colonized by photosynthetic microbes that build biocrust communities. Here we use concurrent physical, biochemical and microbiological analyses to show that mature biocrusts can increase surface soil temperature by as much as 10 °C through the accumulation of large quantities of a secondary metabolite, the microbial sunscreen scytonemin, produced by a group of late-successional cyanobacteria. Scytonemin accumulation decreases soil albedo significantly. Such localized warming has apparent and immediate consequences for the soil microbiome, inducing the replacement of thermosensitive bacterial species with more thermotolerant forms. These results reveal that not only vegetation but also microorganisms are a factor in modifying terrestrial albedo, potentially impacting biosphere feedbacks on past and future climate, and call for a direct assessment of such effects at larger scales.

Endolithic microbial communities are prominent features of intertidal marine habitats, where they colonize a variety of substrates, contributing to their erosion. Almost 2 centuries worth of naturalistic studies focused on a few true-boring (euendolithic) phototrophs, but substrate preference has received little attention. The Isla de Mona (Puerto Rico) intertidal zone offers a unique setting to investigate substrate specificity of endolithic communities since various phosphate rock, limestone and dolostone outcrops occur there. High-throughput 16S rDNA genetic sampling, enhanced by targeted cultivation, revealed that, while euendolithic cyanobacteria were dominant operational taxonomic units (OTUs), the communities were invariably of high diversity, well beyond that reported in traditional studies and implying an unexpected metabolic complexity potentially contributed by secondary colonizers. While the overall community composition did not show differences traceable to the nature of the mineral substrate, we detected specialization among particular euendolithic cyanobacterial clades towards the type of substrate they excavate but only at the OTU phylogenetic level, implying that close relatives have specialized recurrently into particular substrates. The cationic mineral component was determinant in this preference, suggesting the existence in nature of alternatives to the boring mechanism described in culture that is based exclusively on transcellular calcium transport.