DOI: 1.138/NGEO2778 Hydrologic control of carbon cycling and aged carbon discharge in the Congo River basin Enno Schefuß, Timothy I. Eglinton, Charlotte L. Spencer-Jones, Jürgen Rullkötter, Ricardo De Pol-Holz, Helen M. Talbot, Pieter M. Grootes, Ralph R. Schneider Sources of n-c 24 alcohol We are aware that the n-c 24 alcohol has been reported to be produced by freshwater phytoplankton 41,42, in addition to vascular land plants. However, based on the strong resemblance of its compound-specific D values with the D values of the n-c 29 alkane (Fig. 2c), we conclude that it must mainly derive from leaf waxes in samples from GeoB6518-1. Moreover, even if the n-c 24 alcohol would partly be derived from freshwater phytoplankton, the strong correlation of its 14 C content with that of the leaf wax-derived n-alkanes (Fig. 3) would indicate a similar sequestration and release of phytoplankton-derived OM in the Cuvette Congolaise. AminoBHPs as indicators of methanotrophs and wetland environments Bacteriohopanepolyols (BHPs) are pentacyclic triterpenoid lipids with a polyfunctionalised side chain (see Figure S2 for examples). Previous studies demonstrated that bacteria, which oxidise methane aerobically (methanotrophs), are prolific sources of several BHPs containing an amine functionality at the terminal C-35 NATURE GEOSCIENCE www.nature.com/naturegeoscience 1
DOI: 1.138/NGEO2778 position (collectively termed aminobhps) 4,43. However, only methanotrophic bacteria from the phylum Gammaproteobacteria were found to produce the hexafunctionalised BHP, 35-aminobacteriohopane-3,31,32,33,34-pentol (aminopentol) 23,4,43. Methanotrophs from the phyla Alphaproteobacteria and Verucomicrobia as well as Gammaproteobacteria are also known sources of other, less highly functionalised aminobhps including 35-aminobacteriohopane-32,33,34-triol (aminotriol) and 35- aminobacteriohopane-31,32,33,34-tetrol (aminotetrol) 23,4,43. However, as these compounds can have other sources they are considered less specific, in particular if aminopentol is absent 23. Only one none-methanotroph, a species of Desulfovibrio (sulfate reducing bacteria, Deltaproteobacteria), has thus far been shown to produce aminopentol, and then only at trace levels 44, rendering aminopentol a highly diagnostic marker for aerobic methanotrophy (for further information see review in ref 23). Previous studies of BHPs from a range of source environments from the Congo region (soils, wetland, estuary and deep-sea fan sediments) reported a very high relative abundance of aminopentol in the sediments, whilst the same compound was absent or only present at very low levels in all but one of the soils 22. This difference in distribution allowed a link to be inferred between the wetland environment, the estuary, and the BHPs deposited in the deep-sea fan whilst excluding soils as a primary source 23. Furthermore, compound-specific stable carbon isotope analysis of the BHPs in the deepsea fan sediments (ODP Site 175) yielded 13 C values for corresponding BHP lipids that were insufficiently 13 C-depleted compared to bulk organic carbon to indicate a marine source for the primary methane. Therefore, a terrestrial and more specifically a wetland source for these compounds was considered most likely 22,23. 2 NATURE GEOSCIENCE www.nature.com/naturegeoscience
DOI: 1.138/NGEO2778 SUPPLEMENTARY INFORMATION Methanotroph-derived hopanoids in GeoB6518-1 All sediment samples from GeoB6518-1 contained at least 2 biohopanoids with an amine functionality at the C-35 position: aminotriol and aminopentol (Figure S2, Table S3 and S4). Almost all of the samples also contained low levels of aminotetrol. Comparison of the distribution of the three aminobhps in GeoB6518-1 with previously published data from the Congo region (22 soils, 6 wetland sediments and 1 estuary sediment 23 ) indicated that the compounds found in this study correlate most closely with the estuary and wetlands sediments (Figure S2). Overall the distribution of aminobhps, and the high relative abundance of aminopentol in particular is similar to that reported for a lower-resolution study of another core from the Congo deep-sea fan (ODP site 175) spanning the last 1.2 Ma 23. A recent study of shelf and deep-sea fan sediments from the Amazon also reported a wetland source for aminobhps 45. This study proposed that the BHP distributions document export of BHP-containing material from terrestrial wetland regions but also noted that subsequent re-working (of older deposits) of this material, both under recent and past climate conditions would also contribute to material deposited in the marine sedimentary archives 45. In the Congo system, similar aridification-triggered mobilization of aged BHPs originally produced and accumulated during the Early Holocene could cause a dampening effect on declining BHP levels that would result from the reduction in methane-producing land cover during the Mid- to Late Holocene as observed here. NATURE GEOSCIENCE www.nature.com/naturegeoscience 3
DOI: 1.138/NGEO2778 Influence of pre-aged fractions on lipid proxy records All analysed OM fractions from GeoB6518-1 contain isochronous and aged portions (see main text). Despite relatively high estimates of soil-derived OM 18 and the increasing mean age of terrestrial OM fractions towards the present (Fig. 2a), terrestrial OM fractions in the uppermost sample contain radiocarbon derived from atmospheric nuclear weapons testing (visible in 14 C initial values, Fig. S3); this can only be explained by a substantial portion of fresh material that is rapidly transported to the Congo deepsea fan. Similarly, Congo River runoff variations detected by oxygen isotope analyses of planktonic foraminifera in GeoB6518-1 resemble rainfall intensity changes inferred from hydrogen stable isotope compositions of plant waxes in the same samples 21 (Fig. S3). Despite their relatively high mean apparent age, plant waxes thus record even abrupt changes in Congo rainfall in coherence with runoff variations following intrabasin rainfall changes with a time lag of only 1-2 months 46. Since the apparent age of the OM fractions (Fig. 2a), including the plant waxes, is generated by variable mixing of fresh and aged portions, it is not feasible to adjust the depositional ages of the lipidbased climatic records to their respective ages. We thus infer that relative contributions of aged material will not lead to time lags in the palaeo-climatic records as recently suggested 47. Instead, the variable admixture of aged portions will lead to an attenuation or muting of detected palaeo-climatic changes, in particular, when aged material with climatic signals reflecting earlier periods is released. The changes in lipid-based palaeoclimatic records would therefore reflect climatic changes that were of even larger amplitude and more abrupt than those preserved in the sedimentary record. Discrepancies to other studies 47 may in part be caused by differences in depositional setting (lacustrine versus marine) and associated differences in transport mode of plant waxes. 4 NATURE GEOSCIENCE www.nature.com/naturegeoscience
DOI: 1.138/NGEO2778 SUPPLEMENTARY INFORMATION References (continued from Methods) 41 Volkman, J. K., Barrett, S. M. & Blackburn, S. I. Eustigmatophyte microalgae are potential sources of C 29 sterols, C 22 -C 28 n-alcohols and C 28 -C 32 n-alkyl diols in freshwater environments. Org. Geochem. 3, 37-318 (1999). 42 Volkman, J. K. et al. Microalgal biomarkers: A review of recent research developments. Org. Geochem. 29, 1163-1179 (1998). 43 Cvejic, J. H., Bodrossy, L., Kovacs, K. L. & Rohmer, M. Bacterial triterpenoids of the hopane series from the methanotrophic bacteria Methylocaldum spp.: phylogenetic implications and first evidence for an unsaturated aminobacteriohopanepolyol. FEMS Microbiol. Lett. 182, 361-365 (2). 44 Blumenberg, M., Hoppert, M., Kruger, M., Dreier, A. & Thiel, V. Novel findings on hopanoid occurrences among sulfate reducing bacteria: Is there a direct link to nitrogen fixation? Org. Geochem. 49, 1-5 (212). 45 Wagner, T. et al. Microbial biomarkers support organic carbon transport from methane-rich Amazon wetlands to the shelf and deep sea fan during recent and glacial climate conditions. Org. Geochem.67, 85-98 (214). 46 Todd, M. C. & Washington, R. Climate variability in central equatorial Africa: Influence from the Atlantic sector. Geophys. Res. Lett. 31, L2322 (24). 47 Douglas, P. M. J. et al. Pre-aged plant waxes in tropical lake sediments and their influence on the chronology of molecular paleoclimate proxy records. Geochim. Cosmochim. Acta 141, 346-364 (214). 48 Lisiecki, L. E. & Raymo, M. E. A Pliocene-Pleistocene stack of 57 globally distributed benthic delta O-18 records. Paleoceanography 2, PA13 (25). NATURE GEOSCIENCE www.nature.com/naturegeoscience 5
DOI: 1.138/NGEO2778-1 14 C initial ( ) -2-3 -4 n-c24 alcohol n-alkanes -5 2 4 6 8 Carbon Preference Index Figure S1: Absence of fossil OM influences on OM ages. Initial radiocarbon offsets of n-alkanes (grey dots) and n-c 24 alcohol (green dots) versus their respective Carbon-Preference-Indices (CPI), error bars show analytical uncertainty propagated with 14 Catm uncertainty. 6 NATURE GEOSCIENCE www.nature.com/naturegeoscience
DOI: 1.138/NGEO2778 SUPPLEMENTARY INFORMATION GeoB 6518-1 (n=22) Wetlands (n=6) Congo soils (n=22) Congo River estuary (n=1) 2 1 1 9 8 3 7 % Aminopentol 4 6 % Aminotetrol 5 5 NH 2 NH 2 6 4 7 3 8 2 9 1 1 1 2 3 4 5 6 7 8 9 1 % Aminotriol NH 2 Figure S2: BHP compositions of Congo wetland deposits, soils, and sediments. Ternary plot of the relative abundance of aminobhp compounds (aminotriol, aminotetrol and aminopentol) present in core GeoB6518-1 (22 samples; this study), the Congo estuarine sediment, 6 wetland sediments and 22 soils from the Congo (data from ref 23). NATURE GEOSCIENCE www.nature.com/naturegeoscience 7
DOI: 1.138/NGEO2778 a -1 14 C ( ) initial -2-3 n-alkanes -4-17 b -1.2 -.8 -.4.4 18 O ( VPDB) D ( VSMOW) -16-15 -14-13 -12 c d.8 1 8 6 4 2 Linear sedimentation rate -3-1 (cm 1 yrs ) 4 8 12 16 2 Calendar age (years before present) Figure S3: Independence of OM ages, proxy records and sedimentation. Comparison of initial radiocarbon contents of n-alkanes in GeoB6518-1 (a) with Congo River runoff reconstructions based on oxygen isotope compositions of planktonic foraminifera (b, black), compound-specific plant-wax based 21 reconstruction of rainfall intensity in the Congo basin (c, blue) and linear sedimentation rate of GeoB6518-1 (d). Error bars in a) reflect analytical uncertainty. 8 NATURE GEOSCIENCE www.nature.com/naturegeoscience