Soil Steps

Research

Soil Steps

Linking soil food web ecology to carbon modelling in grazed grasslands.

Soil Steps tests whether soil biology, not just bulk carbon measurement, explains why grazing based carbon claims vary so widely, then builds that biology into better carbon and risk models.
~33%of land worldwide is grassland
Mostgrassland carbon sits below ground
3+journal papers planned from this PhD

Why current carbon claims fall short

Grasslands are increasingly promoted as a nature based climate solution, particularly through regenerative grazing and soil carbon credit markets. But the evidence on how grazing affects soil carbon is mixed. Moderate grazing can stimulate plant productivity and carbon inputs, while heavy grazing can strip carbon out through compaction and biomass removal, and the outcome depends heavily on context.

Current carbon accounting frameworks mostly rely on bulk soil carbon measurements and simplified turnover models. They rarely capture the biological processes, particularly soil microbial and faunal communities, that actually determine whether carbon is stabilised in soil or lost again through respiration. That gap is what Soil Steps addresses.

The research questions

What Soil Steps measures

Soil, vegetation, microbial, and faunal data are collected seasonally across grassland sites spanning low, moderate, and intensive grazing management. Soil sampling covers physical properties (texture, bulk density, moisture, aggregate stability), chemical properties (pH, CEC, total and inorganic nitrogen, phosphorus, C:N ratio), and carbon fractionation that distinguishes labile from stabilised pools: permanganate oxidisable carbon (POXC), particulate organic matter, and mineral associated organic matter (MAOM).

Vegetation productivity is estimated from aboveground biomass, litter, and root biomass. Microbial activity is assessed through microbial biomass carbon and nitrogen, CO2 respiration, and potentially mineralizable nitrogen. Soil food web structure is assessed by extracting nematodes (the Baermann funnel method) and classifying them into functional trophic groups: bacterivores, fungivores, herbivores, omnivores, and predators, using established faunal indices and metabolic footprints to estimate carbon and energy flow through trophic pathways.

Analysis uses linear mixed effects models and redundancy analysis in R to relate grazing intensity to soil and biological variables, and structural equation modelling to test the full mechanistic cascade from grazing, through vegetation inputs and soil chemistry, to microbial activity, nematode trophic structure, and carbon and nitrogen cycling, as one connected system rather than isolated pairwise relationships.

Why it matters beyond academia

This research sits close to George's day job. As investment in regenerative grazing and soil carbon credit projects accelerates, carbon claims are often made without mechanistic biological validation, a real source of both scientific and financial risk.

A microbial and food web informed carbon model can flag where projected carbon gains are likely overstated (transition risk), where microbial metabolism is vulnerable to drought or warming (physical climate risk), and where declining soil biological complexity signals reduced ecosystem resilience. That is directly relevant to nature related financial disclosure requirements, including TNFD, that companies are now expected to report against.

Dissemination

Findings are intended for academic, policy, and farmer facing audiences. At least three peer reviewed journal articles are planned across the project, targeting journals including Soil Biology & Biochemistry and Global Change Biology, alongside conference presentations, annual summaries for participating farms and industry partners, and a final policy facing report on grazing management and soil carbon.

Supervision and funding

Supervised by William Stiles and Sarah Watson-Jones, IBERS, Aberystwyth University. This research is supported by Regeno, with additional external sponsorship from Soil Steps Ltd.