Carbon stocks of coastal seagrass in Southeast Asia may be far lower than anticipated when accounting for black carbon

Valuing sedimentary ‘blue carbon’ stocks of seagrass meadows requires exclusion of allochthonous recalcitrant forms of carbon, such as black carbon (BC). Regression models constructed across a Southeast Asian tropical estuary predicted that carbon stocks within the sandy meadows of coastal embayments would support a modest but not insignificant amount of BC. We tested the prediction across three coastal meadows of the same region: one patchy meadow located close to a major urban centre and two continuous meadows contained in separate open embayments of a rural marine park; all differed in fetch and species. The BC/total organic carbon (TOC) fractions in the urban and rural meadows with small canopies were more than double the predicted amounts, 28 ± 1.6% and 36 ± 1.5% (±95% confidence intervals), respectively. The fraction in the rural large-canopy meadow remained comparable to the other two meadows, 26 ± 4.9% (±95% confidence intervals) but was half the amount predicted, likely owing to confounding of the model. The relatively high BC/TOC fractions were explained by variability across sites of BC atmospheric supply, an increase in loss of seagrass litter close to the exposed edges of meadows and sediment resuspension across the dispersed patchy meadow.


Introduction 28
The effect of anthropogenic emissions of CO2 on the climate highlights the importance 29 of quantifying and managing existing sedimentary organic carbon stocks. Lately, focus has 30 turned to 'blue carbon' reservoirsseagrass, saltmarsh, mangrove and macro-algae ecosystems 31 (1,2). Together, these store half the ocean's total organic carbon (TOC), despite occupying 32 less than 2% of the area (3). This disproportionate contribution results from their high rates of 33 primary production and ability to enhance and stabilise accumulation of deposited litter and 34 detritus and prevent it from remineralising, ostensibly back to CO2 (4, 5). Of the four, seagrass 35 ecosystems are best placed to augment organic carbon stocks. Seagrass canopy can trap soils 36 eroded from adjacent landscapes, which can account for as much as 50% of total sedimentary 37 organic stocks (6) that would otherwise be remineralised across the continental shelf (7). 38 is determined by the accumulated BC within the soil profile and is tempered by the edaphic 51 organic content of the soil itself (8). Taken together, such regressions give first-order predictive 52 estimates of meadow BC/TOC. 53 The accuracy of estuarine model's BC/TOC coastal predictions around this region, 54 while moderate, may be confounded by failure to take into account the relatively efficient 55 retention of seagrass litter within estuaries. In more exposed coastal systems, most of the litter 56 across a meadow or close to its exposed boundary can be lost to shelf waters (10-12). It is also 57 possible that increased turbulence closer to the exposed edges or across more patchy seagrass 58 seascapes may lead to localised sediment resuspension. This may further exacerbate the 59 remineralisation of more labile organic fractions, thereby amplifying the importance of BC to 60 the TOC stock. 61 This study examines the above estuarine predictions and likely controlling factors (i.e. 62 litter loss and sediment resuspension) across coastal meadows in Southeast Asia (Sabah, 63 third meadow, chosen for its sparse, patchy configuration, was located near the University of development across the urban environs of Sepanggar Bay (15). The area has a relatively 78 moderate wind fetch and canopy size. It sustains meadows comprising Cymodocea serrulata, 79 Halodule uninervis and Halophila ovalis and isolated small stands of E. acoroides, which 80 assisted in reducing confounding comparisons with the rural systems (supplemental Figure S1). The meadow canopies exhibited clear height differences, attributable to species 87 characteristics and relative wind fetch (supplemental Figure S1), in the order of LL >> BB < 88 OD (Table 1). Differences in canopy coverage between meadows LL and BB in the relatively 89 pristine marine park and OD in the relatively polluted Sepanggar Bay were as expected (LL ≈ 90 BB >> OD, Table 1). Furthermore, significant differences in both height and coverage between 91 transect pairs was only apparent for OD, reflecting its sparse, patchy configuration. 92 The most striking difference in sediment characteristics was between LL, where the 93 longer Enhalus canopy species supported higher silt-clay fractions and a significantly smaller 94 mean particle size, and BB and OD, which were sandier (Table 1). However, these 95 dissimilarities do not reflect inter-meadow variability between transect pairs, which showed 96 significant differences only for OD (Table 2). 97 Differences in both TOC and BC content between meadows (LL << OD > BB) were 98 inverted compared to expectations regarding percentage of slit-clay content (i.e. LL >> OD < 99 BB; Table 1). Interestingly, it was noted that the mean particle sizes within OD's patchy stands 100 All rights reserved. No reuse allowed without permission.
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint . http://dx.doi.org/10.1101/493262 doi: bioRxiv preprint first posted online Dec. 13, 2018; The mechanism by which BC/TOC fractions have seemingly been elevated within 137 coastal embayments can best be distinguished via comparisons of intra-meadow transects in 138 relation to sedimentary parameters, supported by patterns consistent with the relative canopy 139 sizes. As expected, the more exposed outer transects for both LL and BB supported lower 140 sedimentary TOC. Considering that the transect pairs have nearly identical sedimentary 141 parameters, falling rates of resuspension and subsequent oxidation closer to the meadows' 142 exposed edges is not a likely explanation. Neither do the contributions of BC from adjacent 143 bare sediments explain the data patterns. The net deposition of these sediments is likely to be 144 greater across the inner transects, as turbulence is increasingly attenuated (17). This should 145 reduce rather than increase the TOC of the inner transects. Thus, a model that describes 146 increases in BC with TOC, converging towards a positive TOC intercept close to its origin is 147 more consistent with additional organic carbon not associated with BC such as soil washout. 148 How much litter and soil make up these meadow's sediments is unknown and beyond the scope 149 All rights reserved. No reuse allowed without permission.
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint . http://dx.doi.org/10.1101/493262 doi: bioRxiv preprint first posted online Dec. 13, 2018; the exposed edge due to greater attenuation of currents by the canopy. This contention is further 151 supported, though a weak inference, by a larger difference and increasing significance in the 152 TOC variance for the larger canopy species of LL. 153 In contrast, the intra-transect comparison in the patchy OD meadow shows little 154 evidence of attenuation in TOC or BC across the meadow. Nevertheless, there is evidence of 155 resuspension, as implied by the increased particle size and decreased silt-clay fraction away 156 from the exposed edge. Additionally, particle size increased within patches over adjacent 157 unvegetated areas. This pattern suggests that a sparse, patchy configuration allows sufficient 158 space for the seagrass leaves to add to turbulence and resuspension (17) The study demonstrated that allochthonous recalcitrant BC may represent a major 171 fraction of the sediment TOC content of coastal seagrass meadows, and its underestimation 172 may lead to significant overestimates of the carbon sink services of such meadows. Within 173 rural contiguous meadows, the size of seagrass canopy species appears to explain a higher than 174 All rights reserved. No reuse allowed without permission.
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint . http://dx.doi.org/10.1101/493262 doi: bioRxiv preprint first posted online Dec. 13, 2018; development pressure, BC content is largely independent of canopy parameters and TOC 176 variability, and this meadow supports the highest and most consistent BC/TOC fractions. 177 Clearly, more work is needed within Southeast Asian BC hotspots, both rural and urban, to 178 understand the full extent of BC stock bias. We thus recommend moving away from simple 179 mass balance estimation approaches towards one that includes stability and origin via (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint . http://dx.doi.org/10.1101/493262 doi: bioRxiv preprint first posted online Dec. 13, 2018; outside the meadows within the estuary, at OD, and at LL and BB combined, respectively. The 194 box plot shows the median, 25% and 75% quartiles, 95% confidence limits and outliers. Data TOC and BC means were found for LL (P < 0.001, P < 0.001) and BB (P < 0.07, P < 0.001). 202 A power analysis (80%) indicated that determining TOC between BB transects would require 203 another two or three samples to resolve differences in the mean from a possible Type 1 error. 204 (b) Ordinary least squares regressions for BC with TOC across the three meadows. The 205 (which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.