Lost rain, not added heat, raised a soil fungal marker

Nine years of an imposed climate

A patch of semiarid gypsum shrubland near Sorbas, in the province of Almería, has spent nine years living in the climate that models project for the Mediterranean later this century. María del Mar Alguacil, José Ignacio Querejeta and colleagues at CEBAS-CSIC in Murcia laid the plots out in May 2011 at 400 metres above sea level. Hexagonal open-top chambers of transparent methacrylate, hung about 3 cm clear of the ground so air could still move through, raised mean annual air temperature by roughly 2 degrees Celsius. Passive rainout shelters intercepted about 30 per cent of the rain, and a third set of plots carried both devices at once. Both dried the topsoil to a similar degree — warming by 16 per cent on average, the shelters by 14 — but only the chambers changed the temperature.

The site is severe: about 275 mm of rain a year, a mean of 17 degrees Celsius, and soils more than half gypsum with very little organic carbon. Leaf nitrogen-to-phosphorus ratios across the six resident shrubs run from 21.7 to 57.7 depending on species, and anything above 20 means phosphorus is throttling growth. That shortage underwrites the whole study: in this soil, a fungus that can find phosphorus is worth what it costs to feed.

In June 2020 the team took 47 soil cores in a single sampling, 0 to 10 cm deep, one per plot and all beneath shrub canopies: 11 warmed plots, 9 with shelters, 10 with both, and 17 controls picked at random because sampling every available control was not affordable. The results appeared online on 2 April 2026 in New Phytologist and in print in the June issue, volume 250, issue 6, pages 3961 to 3975. The full text is open at PubMed Central; the DOI is registered and correct, but the publisher's own site turns automated requests away.

The marker followed the rain

Fungal biomass was never weighed. It was measured as the neutral lipid fatty acid 16:1 omega 5, a storage lipid that in soil occurs almost only in arbuscular mycorrhizal fungi. Measured that way it reached nearly twice the control value wherever rainfall had been cut — in the shelter-only plots as well as the combined ones — and barely moved under warming alone. This is the firmest number in the paper: an adjusted coefficient of determination of 0.116 at P = 0.008.

The plants were faring badly. Warming with rainfall reduction produced large drops in photosynthesis, in leaf dry mass and in leaf number per unit shoot length; warming alone also cut photosynthesis and growth sharply, while rainfall reduction alone had only minor, non-significant effects on either. That leaves a tension the design cannot settle: the treatment that lifted the fungal marker is the one under which nothing measurable went badly wrong for the shrubs.

Nor do the two halves of the study come from the same moment. Gas exchange was read in early May 2018, one plant per plot and in three of the six target shrubs only — Helianthemum squamatum, H. syriacum and Gypsophila struthium — because the others carry leaves too small and fragile to measure reliably; shoot growth was recorded in November 2018, and the soil two years after that. The authors call the gap an important limitation themselves, and add that their design cannot fully resolve causality between the fungal and the plant variables.

Composition shifted too. Glomeraceae virtual taxa per sample rose with warming, from 21 plus or minus 2.7 in the controls to 34 plus or minus 3.7 under warming plus rainfall reduction, and the lineage's share of reads went from 68 plus or minus 5 to 81 plus or minus 6 per cent. It did not take the soil over, though: it already held it, and the treatments merely deepened a dominance that was there from the start.

What the numbers will not carry

Title and summary both say biomass and diversity increased markedly; in the results, only biomass is solid. Virtual-taxon richness and the Shannon-Wiener index were, in the authors' words, highly variable within the climate treatments, with adjusted coefficients of 0.005 and 0.014 — effectively no explained variance — and the Glomeraceae taxon-count effect sits at 0.027. Phylogenetic diversity actually fell in drier soil, a PERMANOVA found no significant difference in composition between control and combined plots at P = 0.084, and of the four lineages that thinned as the topsoil dried, Claroideoglomeraceae stopped at P = 0.055, just the wrong side of the 0.05 threshold. The 78 virtual taxa are not 78 species either, but sequence-defined units matched against the MaarjAM database, ten of which matched nothing already in it.

No roots were examined at all, because digging for sparse fine roots would have wrecked a nine-year experiment. Colonisation inside roots is therefore unknown, and the authors concede they cannot rule out a shift in the split between soil-exploring and root-colonising mycelium. Their framing — that climatically stressed plants invested more carbon in their partners — is an interpretation rather than a measurement, since no carbon flux was measured and the links to leaf phosphorus are correlations. None of this means that drought suits fungi: the vegetation suffered, and what rose was a proxy in the soil beneath it. And none of it touches foraging, because arbuscular mycorrhizal fungi are Glomeromycota, microscopic root symbionts that produce no fruit bodies at all.

The setting does a great deal of work here, and it does not transfer to a Polish forest or a temperate grassland. How much the place matters shows in a sister experiment on gypsum soils near Aranjuez in central Spain, established in February 2011 by an overlapping team with the same chambers and shelters: there, in the rhizosphere of Helianthemum squamatum, the relative abundance of ectomycorrhizal sequences had fallen by more than 66 per cent by June 2014, under every one of the three climate treatments. A different site, the same imposed drought, the opposite direction for the other symbiosis.