Introduction
The turnover of carbon (C) in soil and sediments is partly constrained by its chemical composition, as well as environmental and physical factors (Hemingway et al. Reference Hemingway, Rothman, Rosengard and Galy2017). The susceptibility of organic matter to decay can be evaluated from its thermal recalcitrance, which is greater at higher temperatures as more energy is required for its decomposition (Garnett et al. Reference Garnett, Pereira, Taylor, Murray and Ascough2023; Plante et al. Reference Plante, Beaupré, Roberts and Baisden2013; Rosenheim et al. Reference Rosenheim, Day, Domack, Schrum, Benthien and Hayes2008). Thermal recalcitrance can be measured by ramped pyrolysis oxidation (RPO; e.g., Rosenheim et al. Reference Rosenheim, Day, Domack, Schrum, Benthien and Hayes2008) or ramped oxidation (RO; Bolandini et al. Reference Bolandini, Maria, Haghipour, Wacker, Hemingway, Eglinton and Bröder2025). RO involves progressive heating of a sample in a stream of oxygen (sometimes in a mixture with other gases like nitrogen), and the simultaneous measurement of the carbon dioxide (CO2) evolved (Plante et al. Reference Plante, Fernández, Haddix, Steinweg and Conant2011). The resulting thermograms (plots of CO2 concentration versus temperature) can be used to calculate activation energy distributions, representing the energy required to break C bonds (Hemingway et al. Reference Hemingway, Rothman, Rosengard and Galy2017). The thermograms can also be used to define organic carbon (OC) thermal reactivity pools as distinct temperature windows. By collecting the CO2 evolved from these defined temperature ranges, and subsequently analyzing it for 14C content, an estimate of the age of the OC reactivity pools can be obtained (Rosenheim et al. Reference Rosenheim, Day, Domack, Schrum, Benthien and Hayes2008).
The 14C analysis of thermally-resolved RO fractions (RO-14C) has great potential for the investigation of saltmarsh soils because they accumulate differently-aged C, sourced from both autochthonous (in-situ photosynthesising vegetation) and allochthonous (externally derived) sources with different lability (Houston et al. Reference Houston, Garnett and Austin2024; Komada et al. Reference Komada, Bravo, Brinkmann, Lu, Wong and Shields2022; Saintilan et al. Reference Saintilan, Rogers, Mazumder and Woodroffe2013). Although RO offers the potential to thermally fractionate samples sufficiently to isolate sole C sources for 14C dating, most research shows that the thermally resolved C pools are composed of mixed ages (e.g., Rosenheim et al. Reference Rosenheim, Day, Domack, Schrum, Benthien and Hayes2008; Stoner et al. Reference Stoner, Schrumpf, Hoyt, Sierra, Doetterl, Galy and Trumbore2023). Moreover, the composition of thermally defined OC reactivity pools can be compared to that of respired CO2 (e.g. from incubation studies), to assess thermal reactivity versus bioavailability of soil OC, although studies which do this are rare (Houston et al. Reference Houston, Garnett and Austin2025; Peltre et al. Reference Peltre, Fernández, Craine and Plante2013; Plante et al. Reference Plante, Fernández, Haddix, Steinweg and Conant2011).
In a previous study, we found that young (14C-enriched) OC was preferentially respired over old (14C-depleted) OC during aerobic incubations of saltmarsh soils, although a portion of the CO2 respired came from an aged, allochthonous source (Houston et al. Reference Houston, Garnett and Austin2024). Characterization of the OC by RO-14C determined that the young respired CO2 was likely from a predominantly thermally labile (i.e. low RO temperature) OC pool (Houston et al. Reference Houston, Garnett and Austin2025). For the surface soil samples, the respired CO2 was 14C-enriched relative to any of the RO fractions, implying that there was variability in the age and source of OC even within the most thermally labile OC pool (150–325 °C; Houston et al. Reference Houston, Garnett and Austin2025). These initial results suggested that ramped oxidation-14C (RO-14C) analysis of finer intervals of the low temperature fraction than reported by Houston et al. (Reference Houston, Garnett and Austin2025) might yield improved insights into the turnover of young OC in these systems.
Here, we apply RO-14C analysis of fine-scale temperature intervals (as low as 25 °C) to partition into fine fractions the most thermally labile (150–325 °C) component of the salt marsh soils investigated in our earlier study (Houston et al. Reference Houston, Garnett and Austin2025). Our aim is to test the usefulness of fine-scale ramped oxidation 14C measurements for generating new and detailed insights into soil carbon cycling.
Methods
Sample collection and handling
The field site and sample collection protocols have been previously described (Houston et al. Reference Houston, Garnett and Austin2024, Reference Houston, Garnett and Austin2025). In brief, three soil cores (T1, T2, T3) were extracted from the lower marsh zone of the Skinflats saltmarsh, Scotland, United Kingdom (56° 3’34.04”N, 3°43’59.16”W) using a 20 cm deep corer. A full 20 cm depth of soil profile could not be obtained at each coring location due to the open bottom of the corer resulting in the loss of some of the deeper soil during extraction, and therefore the total profile investigated at each site was from the surface to a depth of 19 cm (T1), 16 cm (T2), and 20 cm (T3). Each core was ca. 30 m apart. Once retrieved, the cores were wrapped in foil and stored in a cool box on ice and transported back to the laboratory where they were placed in a cool (4 °C), dark room. Within 24 hrs, each core was subdivided into 1 cm thick slices at 0–1 cm, 5–6 cm, and the basal 1 cm (18–19 cm, 15–16 cm,19–20 cm; for T1, T2, T3, respectively). Each slice was then split into sub-samples for aerobic laboratory incubations from which the respired CO2 was collected for 13C and 14C analysis (Houston et al. Reference Houston, Garnett and Austin2024), and for all soil analyses including the RO procedures (this study and Houston et al. Reference Houston, Garnett and Austin2025).
Ramped oxidation
The RO procedure for this study followed published guidance (Garnett et al. Reference Garnett, Pereira, Taylor, Murray and Ascough2023, Reference Garnett, Harman, Murdoch, Taylor and Ascough2025) and preparation of the samples for analysis has been previously reported (Houston et al. Reference Houston, Garnett and Austin2025). Briefly, samples were dried and homogenised at the University of St Andrews and sent to the NEIF Radiocarbon Laboratory for RO-14C. The samples were not decarbonated to avoid potential losses from the thermally labile OC pool (Bao et al. Reference Bao, McNichol, Hemingway, Gaylord and Eglinton2019a; He et al. Reference He, Yang and Cui2025; Houston et al. Reference Houston, Garnett and Austin2025) and to enable comparison to the incubation soils (Houston et al. Reference Houston, Garnett and Austin2024) which were not decarbonated as the acid-treatment to do so would affect respiration processes.
Thermograms were obtained by performing RO on an aliquot of the prepared sample material and have already been reported in Houston et al. (Reference Houston, Garnett and Austin2025). We used these thermograms to determine the finest-scale temperature windows for which sufficient CO2 would be evolved for collection for 14C analysis (> 3 mL CO2) for all samples, given a maximum sample size of ca. 600 mg that could be contained within the sample holder of the ramped combustion vessel. The defined temperature ranges were 150–250 °C, 250–275 °C, 275–300 °C, 300–325 °C.
We repeated the ramped combustion (for this study), this time collecting the CO2 at the defined fine-scale temperature intervals. We used the same equipment and identical settings as used for the original thermograms and coarse-scale RO-14C sample collections (Houston et al. Reference Houston, Garnett and Austin2025), i.e. a carrier gas of high purity oxygen (ca. 30 mL/min, 99.9995% purity, BOC, UK) and a ramp rate of 5 °C per minute. Complete oxidation of evolved compounds was achieved with a Pt catalyst in a second furnace set to a constant 950 °C. The thermally evolved CO2 was collected in foil gas bags and cryogenically recovered as pure CO2 before being graphitised (Slota et al. Reference Slota, Jull, Linick and Toolin1987) and analyzed for 14C. The 14C measurement was carried out at the SUERC Accelerator Mass Spectrometry (AMS) Laboratory (Ascough et al. Reference Ascough, Bompard, Garnett, Gulliver, Murray, Newton and Taylor2024), as in Houston et al. (Reference Houston, Garnett and Austin2024, Reference Houston, Garnett and Austin2025). The 13C content (δ1 3C-VPDB) was determined from an aliquot of the recovered CO2 by isotope ratio mass spectrometry (Thermo-Fisher Delta V, Germany) and used to correct the 14C results for isotopic fractionation by normalising to a δ13C of –25‰ (Stuiver and Polach Reference Stuiver and Polach1977). The 14C activities were background-corrected based on the results for 14C-dead standards processed using the same RO-14C method (anthracite; Garnett et al. Reference Garnett, Harman, Murdoch, Taylor and Ascough2025) and reported as %Modern (F14C x 100; Reimer et al. Reference Reimer, Brown and Reimer2004).
Known age radiocarbon standards were processed alongside the samples to confirm quality assurance. We also performed mass balance checks by comparing the summed CO2 yields from the fine RO-14C fractions, against that yielded from the previously reported (Houston et al. Reference Houston, Garnett and Austin2025) 150–325 °C RO fraction of the same sample material (Table A1). Equivalent isotope mass balance checks were also undertaken using the measured 14C and 13C contents (Tables A2 and A3).
Data analysis
RStudio V4.2.2 (R Core Team, 2022) was used for data analysis. Statistical analyses were undertaken in Sigmaplot V12.5 (Mann-Whitney-U and Kruskal-Wallis for data which did not meet assumptions of normality, and t-tests and ANOVAs for data which did).
Results
Mass balance checks
We summed the %C yield from our fine-scale RO thermal fractions and compared this to the %C of the 150–325 °C RO fraction in Houston et al. (Reference Houston, Garnett and Austin2025) (Table A1). The C yields ranged from 76–109 % and were not significantly different between the two studies (Mann-Whittney-U test; p = 0.48. Table A1). The weighted 14C and 13C contents of the fine-scale RO thermal fractions (Eq. A1 and A2) were not significantly different to the 150–325 °C RO fraction in Houston et al. (Reference Houston, Garnett and Austin2025) either (Tables A2 and A3. Mann-Whittney-U test; p = 0.33 and 0.99, respectively). When combined, the thermal fractions for this fine-scale RO study (150–250 °C, 250–275 °C, 275–300 °C, 300–325 °C) were therefore comparable to the lowest temperature (150–325 °C) RO fraction in the earlier investigation of the same soil samples (Houston et al. Reference Houston, Garnett and Austin2025).
Radiocarbon of fine scale ramped oxidation fractions
The new, fine-scale RO-14C contents ranged from 85.91-98.41 % Modern for 150–250 °C, 86.35–96.98 % Modern for 250–275 °C, 78.48–93.47 % Modern for 275-300 °C, and 67.91–87.87 % Modern for 300–325 °C (Figure 1; Table 1). RO14C activities were lower at higher temperatures for both sets of measurements (Figure 1). 14C content decreased significantly with increasing temperature for the RO temperature fractions between 150–325 °C (Kruskal-Wallis, p < 0.001), overlapping with the previously published RO-14C contents for the same temperature interval (81.50–97.53 % Modern. Figure 1; Houston et al. Reference Houston, Garnett and Austin2025).
Thermograms (red lines, right-hand y-axis) overlaying the 14C content of the CO2 evolved from each temperature fraction (black bars, left-hand y-axis) for each sample. The gray bars represent the 14C content of the CO2 evolved from the RO experiments in Houston et al. (Reference Houston, Garnett and Austin2024). The horizontal green dashed lines represent the 14C content of the CO2 respired from the aerobic incubation experiments of Houston et al. (Reference Houston, Garnett and Austin2024).
Figure 1. Long description
The image contains nine panels, each showing a combination of bar and line graphs. Each panel represents data at different depths: 0.5 cm, 5.5 cm, and 18.5 cm for T1; 0.5 cm, 5.5 cm, and 15.5 cm for T2; and 0.5 cm, 5.5 cm, and 19.5 cm for T3. The x-axis of each graph represents temperature in degrees Celsius, ranging from 0 to 800 degrees Celsius. The left y-axis represents the percent modern carbon-14 content, ranging from 0 to 100 percent. The right y-axis represents the normalized carbon dioxide concentration relative to peak carbon dioxide, ranging from 0.00 to 1.00. The red lines indicate the normalized carbon dioxide concentration, while the black and gray bars represent the percent modern carbon-14 content. The horizontal green dashed lines represent the carbon-14 content of the carbon dioxide respired from aerobic incubation experiments. Each panel shows distinct patterns of carbon dioxide release and modern carbon content across different temperatures and depths.
Radiocarbon concentration (% Modern) of each RO temperature fraction and the respired CO2 from Houston et al. (Reference Houston, Garnett and Austin2024). 14Crange is the difference between the maximum and minimum 14C contents for each RO sample. Errors are reported to one standard deviation from the mean. δ1 3C results are presented in Table A3.
Table 1. Long description
A table comparing radiocarbon concentrations at different temperatures and depths. The table has 10 rows and 6 columns. The columns are labeled as follows: 150-250 degrees Celsius, 250-275 degrees Celsius, 275-300 degrees Celsius, 300-325 degrees Celsius, 14C range, and Respired CO2 (Houston et al. 2024). The rows are labeled with different depths and temperatures: T1 0.5 cm, T1 5.5 cm, T1 18.5 cm, T2 0.5 cm, T2 5.5 cm, T2 15.5 cm, T3 0.5 cm, T3 5.5 cm, T3 15.5 cm, and T3 19.5 cm. Each cell contains values representing the percentage of modern carbon-14 and their respective errors. The table shows the variation in radiocarbon concentrations across different temperature fractions and depths, with values ranging from 67.91 to 98.41 percent modern carbon-14. The 14C range column indicates the difference between the maximum and minimum carbon-14 contents for each sample. The Respired CO2 column provides additional data from Houston et al. 2024.
For the new fine-scale RO-14C measurements, there was no significant change in RO-14C (Figure 1; Table 1) with depth (0.5 cm, 5.5 cm, and the deepest samples in each core) for any of the temperature fractions (ANOVA; p = 0.29, 0.82, 0.96, 0.89, for 150–250 °C, 250–275 °C, 275–300 °C, 300–325 °C, respectively). Likewise, the range of 14C contents for each sample (14Crange, Table 1) did not change significantly with depth (ANOVA, p = 0.22).
Comparison of fine-scale ramped oxidation and incubation results
For the 0.5 cm depth samples, each of the RO-14C contents from this study and Houston et al. (Reference Houston, Garnett and Austin2025) were 14C-depleted relative to both CO2 evolved from the incubated soils (Figure 1; Table 1. Houston et al. Reference Houston, Garnett and Austin2024) and the atmosphere when the samples were collected in 2022 (Hua et al. Reference Hua, Turnbull, Santos, Rakowski, Ancapichún, Pol-Holz, Hammer, Lehman, Levin, Miller, Palmer and Turney2022). For the deeper samples, the 14C content of the respired CO2 from Houston et al. (Reference Houston, Garnett and Austin2024) was much closer to the RO-14C contents measured in this study, aside from T1 18.5 cm and T3 19.5 cm, which were 14C-depleted relative to the fine-scale temperature fractions (Figure 1; Table 1).
Discussion
14C content of ramped oxidation fractions
The 14C content of each RO fraction (RO-14C) was <100 % Modern (Table 1), therefore, each sample was most likely predominantly composed of carbon sequestered from the atmosphere before the 1963 14C bomb-spike (pre-bomb, Hajdas et al. Reference Hajdas, Ascough, Garnett, Fallon, Pearson, Quarta, Spalding, Yamaguchi and Yoneda2021; extrapolating the record of Hua et al. Reference Hua, Turnbull, Santos, Rakowski, Ancapichún, Pol-Holz, Hammer, Lehman, Levin, Miller, Palmer and Turney2022 suggests that the 14C activity of atmospheric CO2 was ca. 100 % Modern at the time of sample collection in July 2022). While lower % Modern 14C likely therefore corresponds to older ages, it is also likely that each thermal fraction is composed of a complex mixture of OC sources with different 14C contents, including post-bomb (> 100 % Modern) sources.
Our sampling site, Skinflats, is an accreting saltmarsh with photosynthesising vegetation which fix modern atmospheric C (Miller et al. Reference Miller, Smeaton, Yang and Austin2023; Smeaton et al. Reference Smeaton, Garrett, Koot, Ladd, Miller, McMahon, Foster, Barlow, Blake, Gehrels, Skov and Austin2024), and therefore we would expect some post-bomb C contributions in at least one of the thermal fractions in the surface (0.5 cm) samples. There could also be post-bomb 14C contributions in deeper soils due to wetland plant roots which can extend to 40 cm depth in some wetland soils (Bernal et al. Reference Bernal, Megonigal and Mozdzer2017). In Houston et al. (Reference Houston, Garnett and Austin2025) we hypothesised that due to the lower 14C contents at higher temperatures, post-bomb 14C was likely to have been present in the most thermally labile fraction (150–325 °C) but obscured by pre-bomb C also contained within this temperature window. However, the 14C results for all fine-scale RO-14C fractions in the current study returned only pre-bomb (<100 % Modern) 14C activities (Figure 1; Table 1) suggesting that either the post-bomb C is not present or is still obscured by the presence of older C even in these fine-scale fractions.
As saltmarshes are periodically inundated, it is possible that some saltmarsh vegetation fixes C from seawater dissolved inorganic carbon (DIC) which is 14C-depleted relative to the atmosphere, due to the marine reservoir effect (Hajdas et al. Reference Hajdas, Ascough, Garnett, Fallon, Pearson, Quarta, Spalding, Yamaguchi and Yoneda2021). A previous study on saltmarsh soils found that they accumulate C with post-bomb 14C contents (Capooci et al. Reference Capooci, Seyfferth, Tobias, Wozniak, Hedgpeth, Bowen, Biddle, McFarlane and Vargas2024), and saltmarsh macrofossils have been shown to follow the atmospheric 14C bomb curve (Marshall et al. Reference Marshall, Gehrels, Garnett, Freeman, Maden and Xu2007). However, we did not make any site-specific measurements of vegetation 14C content so we cannot determine if any of the autochthonous OC accumulating at the Skinflats saltmarsh is derived from a seawater DIC source. Another possible explanation for the lack of observed post-bomb 14C contents in any of the RO fractions or the respired CO2 (Figure 1; Table 1) is if saltmarsh vegetation fixed CO2 produced from old carbon in the soil, as has been found for some terrestrial peatlands (Jungner et al. Reference Jungner, Sonninen, Possnert and Tolonen1995). It has also been suggested that some fungi can fix C from old, 14C-depleted CH4 in the soil (Kilian et al. Reference Kilian, van der Plicht and van Geel1995). However, these processes have not been observed for saltmarsh soils in any previous studies, and we cannot specifically attribute our lack of post-bomb 14C contents (Figure 1; Table 1) to any of these processes.
There is likely to be heterogeneity in the OC thermal reactivity pools measured in this study and these may not be narrow enough to thermally separate the post-bomb C accumulating at Skinflats saltmarsh. Another possible explanation is that the thermal lability of “young” and “old” C overlap, e.g., two separate compounds of different ages, which have similar thermal lability. If two different compounds have the same thermal lability they will remain inseparable without sample pre-treatment, e.g., physical/chemical fractionation prior to ramped heating (e.g., Stoner et al. Reference Stoner, Schrumpf, Hoyt, Sierra, Doetterl, Galy and Trumbore2023). Indeed, physical/chemical fractionation has been shown to separate fractions with different properties, such as particulate and mineral-associated OC which can be differently aged but have the same thermal reactivity (Stoner et al. Reference Stoner, Schrumpf, Hoyt, Sierra, Doetterl, Galy and Trumbore2023).
OC source attribution of ramped oxidation fractions
Given the differences in 13C contents between the RO thermal fractions (Table A3), it is likely that the fractions were composed of different proportions of OM compounds. For the temperature range studied here (150–325 °C), the 13C-depleted OC has the greatest thermal lability, which is consistent with a greater contribution from autochthonous/terrestrial OC sources (Miller et al. Reference Miller, Smeaton, Yang and Austin2023; Smeaton et al. Reference Smeaton, Garrett, Koot, Ladd, Miller, McMahon, Foster, Barlow, Blake, Gehrels, Skov and Austin2024). Autochthonous and terrestrial OC have overlapping 13C contents at Skinflats saltmarsh (Miller et al. Reference Miller, Smeaton, Yang and Austin2023; Smeaton et al. Reference Smeaton, Garrett, Koot, Ladd, Miller, McMahon, Foster, Barlow, Blake, Gehrels, Skov and Austin2024), so further information is required to infer OC provenance.
As we did not measure the relative abundances of any OM compounds in our samples, we cannot attribute specific contributions from OM sources to the RO thermal fractions. However, for the temperature ranges at which CO2 was collected from in this study (150–325 °C), the contribution of lignin relative to lipid compounds has been shown to increase with temperature for agricultural soils (Sanderman and Grandy Reference Sanderman and Grandy2020), and work on saltmarsh soils has posited that lipid compounds usually have shorter soil residence times than lignin compounds (Luk et al. Reference Luk, Todd-Brown, Eagle, McNichol, Sanderman, Gosselin and Spivak2021). Lignin compounds tend to be 13C enriched relative to lipids (Luk et al. Reference Luk, Todd-Brown, Eagle, McNichol, Sanderman, Gosselin and Spivak2021), so greater contributions from lignin compounds to higher temperature RO fractions would cause 13C contents to increase with temperature, which we see for the saltmarsh soils studied here (Table A3).
The 13C-enrichment of the higher temperature fractions could also be caused by greater proportions of microbially processed OC (Soldatova et al. Reference Soldatova, Krasilnikov and Kuzyakov2024), which have previously been shown to contribute in greater amounts at higher temperature RO fractions than less processed OC (Sanderman and Grandy Reference Sanderman and Grandy2020). The 13C-enrichment with increasing thermal recalcitrance is consistent with previous work on the same samples (Houston et al. Reference Houston, Garnett and Austin2025), and most soil and sediment studies which have found 13C contents to be greater at higher temperatures (e.g., Sanderman and Grandy, Reference Sanderman and Grandy2020; Stoner et al. Reference Stoner, Schrumpf, Hoyt, Sierra, Doetterl, Galy and Trumbore2023).
Comparison of fine-scale ramped oxidation and incubation 14C results
For each of the surface (0.5 cm) samples, the respired CO2 (Houston et al. Reference Houston, Garnett and Austin2024) was 14C-enriched relative to any of the thermal fractions measured in this study (Figure 1; Table 1). Therefore, the fine-scale temperature intervals for these samples did not thermally isolate the youngest (assuming there was some post-bomb C present in this accreting marsh), nor the most biologically available OC (14CO2 from the incubation experiments in Houston et al. (Reference Houston, Garnett and Austin2024)).
The respired CO2 from the deeper samples was much closer in 14C content to the fine-scale RO thermal fractions measured in this study than the surface (0.5 cm) samples (Figure 1; Table 1), apart from T1 18.5 cm and T3 19.5 cm which were deemed to be derived from an IC pool in Houston et al. (Reference Houston, Garnett and Austin2024), which follows different decomposition pathways to OC (Van Dam et al. 2021). Only the respired CO2 from the surface samples was 14C-enriched compared to any of the thermal fractions (Figure 1), implying that the most biologically available OC (in aerobic conditions) accumulating at Skinflats saltmarsh gets respired/consumed during burial. In Houston et al. (Reference Houston, Garnett and Austin2024), it was found that respiration processes were significantly more active for the surface (0.5 cm depth) than the deeper samples. The faster respiration rate in the surface soil may therefore reflect a component of relatively labile and young (14C-enriched) OC in the 0.5 cm layer, which is rapidly depleted. Thus, this young component contributes insignificantly to deeper layers resulting in the closer agreement in the 14C content of the respired CO2 and RO-14C fractions (Figure 1).
In most cases, 14CO2 respired from the incubation experiments (Houston et al. Reference Houston, Garnett and Austin2024) did not reflect the older C stored in these soils, which are resolved by the RO fractions in this study and Houston et al. (Reference Houston, Garnett and Austin2025) (Figure 1; Table 1). If we had only performed the incubation study and had not measured the RO fractions, we would not have known that there was old, thermally labile OC in these saltmarsh soils as the respired CO2 was, although pre-bomb, much younger than the bulk soil (Houston et al. Reference Houston, Garnett and Austin2024). Conversely, if we had not done the incubation study and only measured the RO-14C fractions, we would have expected the respired CO2 from the 0.5 cm samples to be older than was the case, as the RO fractions were all 14C-depleted relative to the respired CO2 at this depth (Figure 1), even with the fine-scale fractions. It may be possible to resolve differently aged OC pools of the same thermal lability by physically/chemically fractionating samples prior to analysis to separate particulate and mineral associated carbon pools (Stoner et al. Reference Stoner, Schrumpf, Hoyt, Sierra, Doetterl, Galy and Trumbore2023), but it is possible that some heterogeneity within each thermal fraction would still remain. Furthermore, thermal analyses are only a proxy for degradation, and they do not show directly what microbes are accessing and respiring in soils - an assessment of respired CO2/CH4 is required to gain insights into this. A combined incubation and RO approach may generate new insights into carbon cycling in complex systems by measuring the 14C content (age) of thermally and biologically separated C pools. For example, depositional environments with contributions from allochthonous sources are likely to have a wide range of C ages accumulating in their soils.
Thermal lability does not necessarily correspond to biological lability or soil residence time
Whilst thermal lability is a useful proxy for estimating soil carbon residence time (thermally recalcitrant C residing for longer in the soil than thermally labile C), the two are not always closely related. The T1 and T2 surface samples were 14C-enriched at 250–275 °C relative to 150–250 °C (Figure 1), demonstrating that 14C content is not always directly correlated with thermal lability, although most studies have found age and thermally lability to be correlated (Bao et al. Reference Bao, Zhao, McNichol, Wu, Guo, Haghipour and Eglinton2019b; Rosengard et al. Reference Rosengard, Mauro, Moura, Spencer, Johnson, McNichol, Boehman and Galy2025; Rosenheim et al. Reference Rosenheim, Day, Domack, Schrum, Benthien and Hayes2008). Plant OM can be derived from a range of compounds of differing labilities and have the same “age.” For example, lignin compounds would be evolved at higher temperatures than lipids and polysaccharides but have the same 14C content if the carbon was fixed from the atmosphere at the same time (e.g., Sanderman and Grandy Reference Sanderman and Grandy2020). Likewise, fungi have been found to decompose old C (Hobbie et al. Reference Hobbie, Weber, Trappe and Van Klinken2002; Newsham et al. Reference Newsham, Garnett, Robinson and Cox2018), forming new biomass which can differ in age and thermal lability from the source material, a potential pathway for old, 14C-depleted OC having a thermally labile signature. Therefore, although young C tends to be the predominant component of the thermally labile C pools, it can also contribute to thermally recalcitrant fractions. Conversely, old C can be thermally labile, if stored in an environment with low productivity and slow turnover rates like peatlands or permafrost soils (Estop-Aragonés et al. Reference Estop-Aragonés, Olefeldt, Abbott, Chanton, Czimczik, Dean, Egan, Gandois, Garnett, Hartley, Hoyt, Lupascu, Natali, O’Donnell, Raymond, Tanentzap, Tank, Schuur, Turetsky and Anthony2020; Houston et al. Reference Houston, Garnett and Austin2025), or if processed from an old C source in the soil by fungi (Hobbie et al. Reference Hobbie, Weber, Trappe and Van Klinken2002; Newsham et al. Reference Newsham, Garnett, Robinson and Cox2018).
Although we demonstrate a broad relationship between thermal lability and biological availability (Figure 1), there are key differences between the two. The availability of OC for decomposition depends on a range of (a)biotic factors, including redox conditions and hydrodynamics (Spivak et al. Reference Spivak, Sanderman, Bowen, Canuel and Hopkinson2019). Whilst a thermally labile source may be predominantly respired in oxic conditions (i.e., the experimental conditions tested in this study), a more thermally recalcitrant pool may be utilised in anoxic conditions. For our experiments we found a broad relationship between the thermal reactivity and bioavailability of saltmarsh soil OC in oxic conditions, but the correlation is not exact (Figure 1) and may differ in different experimental conditions, as well as in-situ. Therefore, great care should be taken when assessing the vulnerability of soil carbon to loss from thermal proxies like ramped oxidation alone. Where possible, comparing thermal and biological indices (like soil incubations) may provide additional insights into the vulnerability of soil carbon pools. Accurately assessing which components of the soil organic carbon pool are being returned to the atmosphere, and under which scenarios (e.g., saltmarsh drainage and degradation switching soils from anoxic to oxic), is crucial to help inform climate change mitigation efforts.
Our results demonstrate the importance of saltmarshes as stores of differently-aged, thermally-labile OC, likely due to their low-oxygen soils which slow microbial decomposition of OC (McTigue et al. Reference McTigue, Walker and Currin2021), as well as the deposition and accumulation of aged, allochthonous C during tidal inundation. Our finding that pre-bomb OC with variable 14C contents contributes to the most thermally labile OC pool of the Skinflats saltmarsh implies that degradation could result in accelerated losses of thermally labile OC from different-aged sources. However, because biological turnover of OC is related to physical and environmental conditions as well as thermal recalcitrance, it is possible that degradation could also result in accelerated losses from the older (14C-depleted), more thermally recalcitrant OC pools.
Conclusions
We provide the finest-scale isotopic measurements to date for any RO-14C study, demonstrating that even within the most thermally labile RO fraction of a previous study (Houston et al. Reference Houston, Garnett and Austin2025), the OC was derived from multiple sources with different ages (14C contents) and that age and thermal lability are not necessarily always tightly correlated. These findings demonstrate the utility of such fine-scale RO measurements for understanding carbon turnover for all sample types and provide crucial insights into the turnover of OC in saltmarsh soils. However, we also show that even these fine-scale RO temperature fractions were unable to thermally separate the youngest nor the most biologically available OC in our saltmarsh soil samples, and that thermal methods may remain limited in this regard. A combined aerobic incubation 14CO2 (Houston et al. Reference Houston, Garnett and Austin2024) and RO-14C approach reveals greater insights into carbon cycling for samples which contain a wide range of soil C sources. Combining thermal and biological separation may be a powerful avenue for exploring carbon cycling in soils and sediments.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/RDC.2026.10245
Acknowledgments
We thank the NERC SUPER DTP for funding the PhD through which this research was undertaken (NE/S007342/1). We acknowledge support from the National Environmental Isotope Facility in funding the 14C measurements for this study under grant NE/S011587/1 (allocation numbers 2594.1022, 2709.1023, 2863.1024). WENA and AH also acknowledge support provided by the HORIZON-CL5-2023-D1-02-02 grant C-BLUES, Innovation to advance the evidence base for reporting of Blue Carbon inventories and greenhouse gas fluxes in coastal wetlands. Thanks are extended to Chloe Bates for assisting with sample collection.
Open access
