Comparison of pyrogenic carbon abundance in coarse-textured soil by hydrogen pyrolysis, NMR and dichromate oxidation and MIR-PLSR

• Globally distributed coarse-textured soils analysed for pyrogenic carbon by four techniques. • HyPy, and NMR abundances strongly correlated, but NMR abundances ∼7× higher. • Dichromate oxidation results variable and not correlated to other techniques. • MIR-PLSR predicts pyrogenic carbon abundance determined by HyPy and NMR. Soil pyrogenic carbon (PyC) is of considerable significance to the global carbon cycle as a carbon pool which is resistant to mineralization and thus offers opportunities to facilitate net carbon sequestration. Quantifying the size and dynamics of the soil PyC pool is hampered by the large number of techniques that yield a wide range of abundances even when applied to the same sample. We used hydrogen pyrolysis to quantify stable polycyclic aromatic carbon (SPAC) of pyrogenic origin (PyC SPAC ) in a globally distributed set of coarse-textured soils, in which the percentage of particles finer than 53 µm ranged from 0.1 to 24.1 % (mean = 7.2 ± 5.8 % 1σ). PyC SPAC values ranged from 0 to 0.37 % (mean = 0.08 ± 0.06 %). We compared the PyC SPAC values with estimates derived from nuclear magnetic resonance spectroscopy (PyC NMR ) and found a strong correlation between the two (r = 0.90). However, the PyC NMR estimates were ∼7 times higher than PyC SPAC values, attributed partly to NMR measuring a wider range of pyrogenic molecules but also likely due to the inclusion of aromatic ‘resistant’ soil carbon of non-pyrogenic origin. In contrast, there was little correspondence between either PyC SPAC or PyC NMR and abundances determined by dichromate oxidation (PyC OREC ). Partial least squares modelling of the mid-infrared (MIR) spectra was able to predict both PyC SPAC and PyC NMR values with high confidence (r = 0.77 and 0.94 respectively). The study suggests that, with appropriate scaling factors, PyC SPAC and PyC NMR can be directly compared, and both can be predicted by MIR.