The shortcut in question is loss-on-ignition, or LOI, a technique beloved by monitoring agencies because it is fast, cheap and requires nothing more exotic than a furnace and a balance. A dried sediment sample is heated to 450 degrees Celsius for four hours, and the mass it loses is taken as a proxy for its organic matter content. Because organic carbon typically makes up only a fraction of that organic matter, the LOI value must then be converted into a total organic carbon, or TOC, estimate using a regression equation. The trouble, as the new study demonstrates, is that this conversion is not universal. It shifts with sediment composition, mineralogy and analytical protocol, meaning that a formula calibrated in one setting can quietly mislead in another.

The stakes are far from academic. Continental shelves cover roughly nine percent of the global seafloor, yet the top metre of subtidal, non-vegetated marine sediments is estimated to contain more than 2.3 trillion tonnes of organic carbon and to accumulate between 126 and 350 million tonnes each year. Fine-grained shelf deposits, including the Western Irish Sea mud belt, are recognised as hotspots of carbon accumulation, where silts and clays settling from suspension in low-energy waters carry organic matter down into long-term storage. Ireland’s Exclusive Economic Zone spans roughly 450,000 square kilometres, about ten times its land area, but its seabed carbon inventory remains poorly constrained despite years of mapping and sampling effort.

To test whether a single conversion equation could really serve such a varied environment, the team assembled an unusually rich dataset spanning surveys conducted between 2010 and 2023, drawing on the national INFOMAR mapping programme, Marine Institute monitoring cruises and the QUEST research project. Grab samples, box cores, gravity cores and vibrocores collected from the RV Celtic Voyager, RV Keary and RV Celtic Explorer yielded sediments ranging from coastal shallows less than five metres deep to the central basin at around 115 metres. Grain size was measured by laser diffraction, iron concentrations by portable X-ray fluorescence, and TOC by combustion elemental analysis after careful acid removal of carbonate carbon. Crucially, a subset of samples was analysed by both LOI and direct TOC methods, allowing the researchers to calibrate one against the other.

The paired measurements correlated strongly across the whole dataset, with a coefficient of determination of 0.948, and the baseline Western Irish Sea model took the form TOC equals 0.41 times LOI minus 0.18. That might sound like a success story, and in aggregate it is. But the residuals told a subtler tale. The pooled equation slightly over-predicted carbon in sandy, low-carbon samples and under-predicted it in carbon-rich muds, hinting that a single line was being forced through what were really two different populations of sediment.

Statistical tests made the split explicit. Chow tests, which probe whether a relationship changes abruptly across a threshold, identified significant structural breaks in the LOI-to-TOC relationship at 0.99 percent iron, 19.8 percent mud content and 29.6 metres water depth. Analysis of covariance confirmed that the regression slopes genuinely differed between the groups. The high-mud slope of 0.42 was nearly double the low-mud slope of 0.24, meaning that a given loss-on-ignition value corresponds to almost twice as much organic carbon in fine, deeper sediments as it does in shallow, sand-dominated ones. Mean TOC-to-LOI ratios reinforced the pattern, reaching 0.34 in fine sediments against just 0.20 in coarse ones.

The physical explanation lies in how minerals shelter organic matter from decay. Fine-grained matrices rich in clay minerals and iron or manganese oxides can bind organic molecules and protect them from microbial attack, allowing more carbon to persist per unit of organic matter. Iron phases such as oxyhydroxides and sulfides co-accumulate with organic carbon in quiet, fine-grained or oxygen-poor settings, which is why the iron threshold emerged as the strongest independent statistical split. The researchers are careful, however, to frame iron and water depth as proxies for depositional environment rather than proven causal controls, since all three variables co-vary with the fundamental grain-size regime of the seafloor.

Splitting the dataset at the 19.8 percent mud threshold produced group-specific models that improved fit markedly for high-mud sediments, reaching a coefficient of determination of 0.97, while the low-mud model, at 0.72, retained greater uncertainty. That asymmetry matters. In coarse sediments, where organic content is low and absolute LOI values are small, estimates are acutely sensitive to minor inorganic mass losses, structural water released from clays, salt effects and seasonal inputs of organic material. Prediction uncertainty is greatest in exactly these low-carbon sands, particularly for samples near the classification threshold, and the authors caution that extrapolating the fitted intercepts to near-zero LOI values is unreliable.

The study is equally candid about its own limits. The paired LOI-TOC subset is smaller than the full metadata set, group sizes are uneven at 18 high-mud and 59 low-mud samples, and the thresholds were chosen through exploratory analysis without independent validation. LOI itself, even when carefully performed, cannot match the specificity of direct combustion analysis, since ignition at 450 degrees Celsius drives off not only organic matter but also clay structural water and other volatile components. The authors stress that their models convert LOI to estimated TOC only; they do not by themselves quantify carbon stocks, accumulation rates, burial rates or climate-mitigation potential, all of which require dry bulk density, sediment thickness, spatial data, chronological constraints and rigorous propagation of uncertainty.

What the study does offer is a practical blueprint. Rather than treating LOI as a poor substitute for elemental analysis, the researchers position it as a way to extend the spatial reach of existing datasets, unlocking carbon estimates from archived samples where LOI was measured but TOC was not, provided paired calibration, sediment-class metadata and uncertainty estimates accompany the conversion. Just as importantly, the work demonstrates how routine geotechnical measurements, grain-size data and monitoring samples from mapping programmes, regulators, ports, offshore energy developers and commercial surveys can be woven together into national carbon inventories, reducing duplicated sampling and strengthening the evidence base on which marine spatial planning, protected area design and blue carbon accounting increasingly depend.

Subject of Research: Region-specific loss-on-ignition to total organic carbon conversion models for estimating blue carbon in Western Irish Sea marine sediments

Article Title: Region-specific LOI–TOC conversion models for marine sediment blue carbon assessment in the Western Irish Sea

Article References: Grey, A., Mhaoldomhnaigh, C. N., Chatting, M., Walsh, P., Lee, A., Kelleher, B., & Coughlan, M. (2026). Region-specific LOI–TOC conversion models for marine sediment blue carbon assessment in the Western Irish Sea. Environmental Monitoring and Assessment, 198(10), Article 1078. https://doi.org/10.1007/s10661-026-15918-6

Image Credits: AI Generated

DOI: 10.1007/s10661-026-15918-6

Keywords: blue carbon, marine sediments, loss-on-ignition, total organic carbon, Western Irish Sea, continental shelf, sediment classification, mud content, iron geochemistry, carbon stock assessment, seabed mapping, organic matter preservation