Introduction

Globally, soils store more than twice as much carbon as the atmosphere1,2. Since the soil carbon cycle is sensitive to climate change and human activities1,3,4, future warming, shifts in precipitation patterns and land use might perturb the soil-carbon storage and subsequently result in positive feedbacks on global warming via CO2 release into the atmosphere1,5. Soil carbon storage is regulated by carbon influx (fixation through net primary production; NPP) and efflux. The latter is controlled by microbial respiration, soil erosion and fire emissions2,5. These processes determine τsoil defined6 as:

$${{{{\rm{\tau }}}}}_{{\mbox{soil}}}=\frac{{{\mbox{C}}}_{{\mbox{soil}}}}{{\mbox{f}}}$$
(1)

where Csoil is the soil carbon stock (in kgC m−2) and f either the carbon influx (NPP) or the efflux (in kgC m−2 yr−1). Under steady state conditions influx and efflux are equal7. Turnover times are critical components in carbon cycling for constraining the time scales of carbon exchange between different reservoirs. τsoil depends on soil temperature3,4,8 and moisture content3,4 but also on chemical properties9,10,11 and soil fertility9,11. Temperature effects on τsoil are widely observed across the globe4 while hydroclimate may exert strong control in low to mid latitudes where it may override temperature effects4,12,13. However, the key controls on τsoil and their interactions are still debated3,10,12. This forms a major open question in tropical and subtropical regions where combined effects of future warming and precipitation changes may be amplified or attenuated depending on whether warming will be accompanied by drier or wetter conditions12. One compromising factor of understanding turnover times and their environmental controls is that our knowledge mostly relies on short-term observations of years to decades (e.g. ref. 12). The geological record is a unique and important means to gain information about centennial to millennial time scales. Characterized by global warming, hydroclimate change and rising atmospheric CO214,15 the last deglaciation (~18,000–11,000 yrs before present (BP), henceforth referred to as 18–8 kyrs BP) is a promising analogue to investigate climate–soil-carbon turnover interactions over several millennia. Unfortunately, proxy data constraining deglacial changes in soil carbon storage and τsoil in the tropics and subtropics are very scarce and existing data provide qualitative estimates only16. The aim of this study is to provide quantitative glacial-to-Holocene reconstructions of τsoil in the (sub-)tropics and to identify the major environmental controls.

We investigate how τsoil changed in the Nile River catchment during the last 18 kyrs. With a length of 6650 km the Nile River is the longest river in the world. Spanning 35° of latitude (4 °S to 31 °N) in northeastern Africa and draining a catchment of nearly 3 million km2 it extends over several vegetation zones (rainforest in the headwaters, savannah, Sahara Desert and the Mediterranean zone at the coast; Fig. 1). Mediterranean sediments supplied by the Nile River load form a powerful recorder of climate change integrating over this vast catchment area and thus being representative of the entire northern African tropics and subtropics. During the last deglaciation the northern African climate warmed17,18 and humid conditions during the African Humid Period (AHP, 14.8–5.5 kyrs BP)19 allowed for plants and permanent water bodies to persist in the nowadays barren, hyperarid Sahara Desert (Green Sahara)20. The different timing of changes in temperature17,18,21 and hydroclimate21,22,23 around the AHP allows for disentangling temperature and precipitation effects on τsoil.

Fig. 1: Map of the study area.
figure 1

African vegetation zones are drawn after ref. 25. The Nile River catchment is marked by the blue shading. The red star indicates the study site GeoB7702-3.

Full size image

Given the absence of proxies for NPP and carbon stock size paleo τsoil cannot be calculated based on Eq. (1). Instead, we investigate the response of τsoil to these climatic changes using compound-specific radiocarbon dating (CSRA) of terrigenous biomarkers, i.e. long chain n-alkanoic acids and long chain n-alkanes preserved in marine sediment core GeoB7702-3, which was retrieved in the eastern Mediterranean from the continental margin off the Sinai Peninsula (Fig. 1). Both compounds are constituents of epicuticular leaf waxes and specific biomarkers for higher land plants24. In marine sedimentary archives they serve as recorders of terrestrial environmental change23,24,25. At the time of deposition in marine sediments these refractory lipids are commonly pre-aged due to intermediate storage (e.g. in soils) and land-ocean transport26,27. The degree of pre-aging (or the age at the time of deposition) is a measure for terrestrial residence times of these compounds which is commonly used to trace changes in terrestrial carbon cycling16,26,28,29. Their age at the time of deposition can be determined by radiocarbon dating26. However, since soil carbon is a complex mixture of various compounds which all possess different turnover times30, the ages of leaf-wax lipids only represent a small fraction of the soil organic matter and do not represent τsoil31. Ages of leaf-wax lipids generally exceed the calculated mean τsoil by a multiple31. Analyzing the 14C-ages of n-alkanoic acids in particulate organic matter from a global sample set comprising coastal sediments near river mouths, riverbeds and banks as well as suspension load, ref. 31 identified globally constant offsets between 14C-ages of n-alkanoic acids and τsoil (see methods for more details). This allows to calculate catchment-integrating mean τsoil (in yrs) from the 14C-ages of n-alkanoic acids in marine sedimentary archives and to monitor changes in the carbon cycle within a river catchment through time.

Here, we deduce past mean τsoil for the Nile River catchment from the 14C-ages of leaf-wax biomarkers at the time of deposition at site GeoB7702-3. To calculate the age at the time of deposition of the long chain n-alkanoic acids and long chain n-alkanes we use the “reservoir age offset” notation32 (given in 14C years; see methods) between the biomarkers and the atmosphere at the time of deposition (Table 1 and Fig. 2c).

Table 1 Reservoir age offsets of leaf-wax lipids, catchment-weighted mean soil carbon turnover times (τsoil), soil mean carbon ages and climate variables for the Nile River watershed over the past 18 kyrs
Full size table
Fig. 2: Environmental changes in the Nile-River delta region during the past 18 kyrs.
figure 2

a Ice-core CO2-contents given as indicator for atmospheric CO2 concentrations (gray dots: data points; black line: spline-smoothed record)15. b Atmospheric Δ14C contents according to IntCal2052. c Reservoir age offsets (R) between the n-alkanoic acids and the atmosphere at the time of deposition at site GeoB7702-3 (this study). τsoil deduced from R of n-alkanoic acids. Error bars indicate the standard deviation. d Sea surface temperature reconstruction for the eastern Mediterranean based on the TEX86 proxy from core GeoB7702-317. e Hydrogen isotopic composition of precipitation (δDp) calculated from the δD of n-alkanoic acids from core GeoB7702-3 as proxy for rainfall amount23. f Oxygen isotopic compositions of the planktic foraminifera species Globigerinoides ruber (δ18OG.ruber) in core MS27PT (Fig. 1) indicating salinity changes in the eastern Mediterranean associated with freshwater runoff from the Nile River35. g Aminopentol abundances in core GeoB7702-3 used as proxy for the extent of methane-producing wetlands in the catchment (this study). AU: arbitrary units; dw: dry weight of extracted sediment. Additional abundance profiles from the suite of amino-bacteriohopanepolyols are given in Supplementary Fig. 1. h Concentrations of n-alkanoic acids (Σn-C26:0, n-C28:0, n-C30:0, n-C32:0) reporting on the land-ocean transport of terrigenous organic matter23. i Global rate of sea-level change over the last 20 kyrs33. The blue bars mark the timing of the African Humid Period (AHP) and Green Sahara and their optimum19,20. LGM: Last Glacial Maximum, HS1: Heinrich Stadial 1, B/A: Bølling/Allerød interstadial, YD: Younger Dryas stadial.

Full size image

Results and discussion

Environmental signals in the compound-specific radiocarbon data

The reservoir age offsets of n-alkanoic acids and n-alkanes in core GeoB7702-3 range between approximately 0 and 8700 14C yrs. It is striking that glacial reservoir age offsets (7800–8700 14C yrs at 18 kyrs BP) are substantially higher than those during the Holocene (0–3400 14C yrs; between ~2–11.5 kyrs BP). This implies a drastic reduction of turnover times of soil carbon during the deglaciation. However, before converting the reservoir age offsets into mean τsoil three factors that may introduce biases need to be considered.

First, sea level rose by up to 120 m over the deglaciation33 and coastal erosion during shelf flooding led to the deposition of pre-aged organic matter on continental margins28,34. Such processes may mask hinterland signals in the reservoir age offsets of leaf-wax lipids in marine sediments. However, biases from coastal erosion during retrogradation of the Nile Delta are unlikely as the concentration profile of n-alkanoic acids in core GeoB7702-3 differs from the global rate of sea-level change33 (Fig. 2h, i) but resembles the oxygen isotopic composition of planktic foraminifera Globigerinoides ruber (δ18OG.ruber) off the Nile River delta, a proxy for freshwater discharge from the Nile River35 (Fig. 2f). Hence, the export of organic matter was primarily controlled by river runoff23.

Second, in addition to mineral soils peatlands need to be considered as source of pre-aged organic matter31. Anaerobic conditions in wetlands hamper degradation of organic matter leading to its preservation in peat over millennia36. During wetland contraction, erosion and fluvial export of this pre-aged material29 could thus bias the calculations of mean τsoil of mineral soils31. This might be relevant to the Nile River catchment since wetlands occur along the basin today37. To constrain wetland dynamics we analyzed a suite of amino-bacteriohopanepolyols (amino-BHPs; Supplementary Fig. 1) which are specific markers for methane oxidizing bacteria in wetlands38 and thus indicative of the relative extension and contraction of methane producing landcover29. Low concentrations of amino-BHPs imply that between 18–11 kyrs BP methane producing permanently flooded wetlands were barely present in the catchment (Fig. 2g and Supplementary Fig. 1) rendering it unlikely that the decrease in the reservoir age offset stems from wetland dynamics. High concentrations of amino-BHPs suggest that wetlands expanded later, i.e. between 11-8 kyrs BP, which probably occurred in response to maximal rainfall and river runoff associated with the AHP-optimum (Fig. 2e, f, g). Contributions of pre-aged organic matter mobilized from wetland contraction at the end of the AHP were probably minor as reservoir age offsets remain constant when amino-BHP concentrations decline in our core (Fig. 2c, g).

Third, river dynamics including morphology and runoff are known controls on the ages of organic matter discharged into the ocean39,40. Increased fluvial runoff may strengthen riverbank erosion and export of relatively old material from deeper soil horizons potentially overprinting signals from τsoil40. Although the Nile-River runoff increased in response to intensified rainfall during the AHP22,35 considerable biases from deep-soil erosion are unlikely given the decrease in reservoir age offsets of n-alkanoic acids and n-alkanes at these times (Table 1, Supplementary Fig. 2). However, intensified Nile River runoff35 may have increased the transport velocity hampering aging of organic matter during land-ocean transit39. This speed-up would have led to smaller ages of plant waxes in core GeoB7702-3 and would be congruent with the observed decrease in our reservoir age offsets. Although signals of the transport efficiency in our data cannot be fully ruled out we consider a predominant control of river dynamics and morphology on ages of discharged organic matter unlikely for the following reasons. River runoff decreased after 7 kyrs BP (Fig. 2f) while the reservoir age offsets of leaf-wax biomarkers remained relatively constant (Table 1; Fig. 2c). The second argument is the similarity between the ages of n-alkanoic acids and n-alkanes (Table 1 and Supplementary Fig. 2). As elaborated in ref. 23, n-alkanoic acids reflect a local signal from the Nile delta region while the n-alkanes provide a catchment-integrating signal23. The extensive Nile catchment is characterized by multiple fluvial environments that differ in geomorphology, flow regime and sedimentary processes41,42. If such morphologic characteristics exerted substantial control on the ages of organic matter in the fluvial load39, n-alkanoic acids and n-alkanes would show different ages and trends which is not the case (Supplementary Fig. 2).

τ soil during the past 18 kyrs

Excluding these potential biases, we conclude that reservoir age offsets of the leaf-wax biomarkers in core GeoB7702-3 can be used to calculate mean τsoil (see methods). For n-alkanes the relationship to mean τsoil is unknown31 which is why we focus on the n-alkanoic acids. Despite the local origin of the n-alkanoic acids23 catchment-wide inferences on changes in τsoil are justified given the strong similarity with the reservoir age offsets of the n-alkanes that provide catchment integrating signals23 (Supplementary Fig. 2).

During the last 10 kyrs, τsoil was 9–22 yrs (average 16 yrs) and 218 yrs during the glacial, meaning that τsoil was reduced by an order of magnitude across the deglaciation (Table 1 and Fig. 2c). τsoil is regulated by the efflux rates of carbon. Degradation of organic matter via microbial respiration constitutes the majority of the total efflux and contributions of lateral fluxes are minor13. As such, the substantial reduction in mean τsoil attests to a substantial increase in microbial respiration rates over the deglaciation.

It is well constrained that microbial respiration accelerates in response to warming and increased soil moisture3,7,12. Both, temperature17,18,21 and rainfall amount21,22,23 increased in the Nile River catchment during the deglaciation (Fig. 2d, e). To investigate the relationship of τsoil to temperature and rainfall amount we fit the natural logarithm of τsoil to proxy-based temperature estimates from the eastern Mediterranean17 and to the hydrogen isotopic composition of paleo precipitation (δDp) in the Nile delta23 (Fig. 3). As mean annual air temperature estimates covering the past 18 kyrs are not available for the Nile River catchment, we use the TEX86-based temperature record from GeoB7702-3 interpreted to reflect sea surface temperature (SST) in the eastern Mediterranean17. We assume that SST and surface air temperatures in the Nile delta region developed similarly due to heat exchange between the sea surface and the overlying air. As for δDp, we use a record based on the δD of n-alkanoic acids in core GeoB7702-323. δDp is generally controlled by several factors including changes in the moisture source, temperature, evapotranspiration and rainfall amount43. In northern Africa and the Mediterranean realm δDp predominantly reflects the amount of rainfall44.

Fig. 3: Power–law relationships between τsoil and temperature and rainfall.
figure 3

a Correlation with temperature estimates based upon the TEX86-proxy (TTEX86) from core GeoB7702-3. TTEX86 are adopted from ref. 17 and interpreted to reflect sea surface temperature17. b Correlation with the hydrogen isotopic composition of precipitation (δDp) which serves as proxy for rainfall amount. δDp is calculated from the hydrogen isotopic composition of n-alkanoic acids (n-C26:0 and n-C28:0 homologues) from core GeoB7702-323 and given relative to the Vienna Standard Mean Ocean Water (VSMOW). In a and b error bars represent the standard deviation (SD). The gray shadings represent the 95% confidence intervals (CI) and the error of the slope therefore contains 2σ. The p-values for the regressions are <0.05. The temperature sensitivity expressed as the Q10-value, i.e. the factor by which τsoil decreases per 10 °C temperature change7,45, can be deduced from the slope of the regression line in a using Eq. (2) leading to Q10 = 10.7 (7.0–16.3, 95% CI).

Full size image

We find that τsoil is strongly negatively correlated with temperature (R2 = 0.82; Fig. 3a). A negative correlation of τsoil with δDp also exists but it is weaker (R2 = 0.59; Fig. 3b). This indicates that temperature was a critical control on microbial respiration rates over the past 18 kyrs (Fig. 3a) while precipitation effects were relatively small. The slope of the correlation in Fig. 3a is a measure for the temperature sensitivity of τsoil during the past 18 kyrs. The temperature sensitivity of soil respiration and τsoil is commonly expressed as the Q10 value, the factor determining the shift in τsoil per 10 °C change in temperature7,45. Q10 is defined as:

$${{\mbox{Q}}}_{10}={{\mbox{e}}}^{10{\mbox{a}}}$$
(2)

where a is the slope of the regression in the temperature-ln(τsoil) plot (Fig. 3a). Accordingly, we obtain a Q10 of 10.7 (7.0–16.3, 95% confidence interval) for the last 18 kyrs. Note that TEX86-based temperatures from core GeoB7702-3 suggest a warming of 10 °C across the deglaciation (Fig. 2d), which is higher than what is typically proposed from other temperature records from the eastern Mediterranean as well as from climate models (3–8°C)14,17,46. For a smaller amplitude in deglacial warming, the slope of the regression line in Fig. 3a would be steeper which would lead to even higher Q10 values. Furthermore, our regressions in Fig. 3 are only based on the mean values and uncertainties in y. If with a different regression algorithm also the uncertainties in variables in x direction were considered then the slopes in the regressions would get even steeper.

For modern conditions, Q10 values of 1–13 have been reported but mean values commonly are about 2–3 in most biomes31,47. Our Q10 estimate of 10.7 (7.0–16.3) is at the top of the range substantially exceeding the modern average. Field observations revealed that Q10 is spatially and temporally variable and that Q10 itself is inversely correlated to temperature47,48. That is why ecosystems in colder regions and higher latitudes have relatively high Q10 compared to lower latitudes and warm settings49. These observations potentially explain why we find rather high Q10 for cold glacial and deglacial climates. The dependency of Q10 to climate and environmental conditions also indicate that there might not be the rather simple linear relationship between temperature and ln(τsoil)48 which is suggested by the Q10 concept, but that the relation between both variables is more complex. If so, our finding of a deglacial (sub-)tropical Q10 at the upper end of the observed modern range may also point to a limitation of the Q10 concept.

Implications for the global carbon cycle

The high glacial τsoil indicate that the carbon exchange between northeastern African soils and the atmosphere was much slower than during the Holocene owing to lower respiration rates during a colder climate. A higher τsoil agrees with previous estimates of a lower glacial global NPP50 which is congruent with a lower carbon efflux from soils assuming equilibrium conditions (Eq. (1)). When discussing turnover times of organic carbon in soils and the implications of changes in carbon storage and turnover time for the global carbon cycle one has to acknowledge that soil organic matter is a complex mixture of fast-cycling labile fractions which degrade within years to decades and slow-cycling refractory compounds that decompose on centennial to millennial time scales30,51. The assumption that τsoil determined by the ratio of NPP over carbon stock size (Eq. (1)) is representative of the entire soil carbon pool oversimplifies soil carbon dynamics as the calculation is actually biased towards the fast cycling pool. This becomes evident when comparing turnover times calculated after Eq. (1) with radiocarbon dates of bulk soil organic matter (the so-called soil mean carbon ages51). If soil organic matter was homogenous τsoil and soil mean carbon ages would match. But in reality τsoil calculated after Eq. (1) underestimates soil mean carbon ages51. This discrepancy is because slow-cycling compounds accumulate in soils owing to their long residence times and dominate the soil organic carbon pool30,51. By contrast, τsoil based on NPP and carbon stock size is biased towards the fast cycling pool as the majority of organic compounds introduced into soils by NPP degrades quickly on years to decades51. To investigate the response of soil carbon dynamics to climate change soil mean carbon ages should be considered next to τsoil, in particular because the slow-cycling pool is more vulnerable to climate change than fast cycling compounds8,31. It is documented that for a given change in temperature the change in turnover rates is greater for a slow-cycling compounds than for the fast-cycling ones8. Given its size, the slow-cycling pool is thus critical for potential positive climate feedbacks from soil carbon dynamics in a warming world8. According to refs. 14,31, 14C-ages of n-alkanoic acids off rivers have constant offsets not only with mean τsoil but also with soil mean carbon ages (integrated over 0–100 cm soil depth; Methods)31,51. Calculating soil mean carbon ages from our reservoir age offsets of n-alkanoic acids (see methods) reveals that during the last glacial soil organic carbon was up to more than ten thousands of years old (14,000 yrs at 18 kyrs BP; Table 1) which is by an order of magnitude older than during the Holocene (1000 yrs; Table 1). The rejuvenation of soil organic matter accompanying the reduced τsoil implies a massive mobilization of pre-aged organic carbon from soils during the deglaciation once the climate warmed. Today, respiration constitutes the majority of the total efflux (>90%)13 and assuming this relation was similar in the past, the decrease in our estimated τsoil and soil mean carbon ages almost entirely reflects increased efflux of aged CO2 into the atmosphere. Accordingly, the reduction of τsoil and soil mean carbon ages by an order of magnitude implies an increase in soil-to-atmosphere CO2 flux of a similar size (Eq. 1). This forms a positive feedback to global warming.

During the last deglaciation atmospheric CO2 rose by about 80–90 ppm15 while the atmospheric radiocarbon content (∆14C) declined concurrently52 (Fig. 2a, b). To explain these changes oceanic outgassing of old, 14C-depleted CO2 (ref. 53) together with contributions from release of aged CO2 from thawing permafrost soils in the Northern Hemisphere have been invoked28,34,50. Our findings suggest that, if widespread across the tropics and sub-tropics, the loss of pre-aged carbon from (sub-)tropical soils due to amplified respiration rates may have formed an additional terrestrial source of old CO2 to the atmosphere (Fig. 2a, b) next to the permafrost domain. There is evidence for accelerated soil-carbon turnover in the Ganga-Brahmaputra River catchment as inferred from reservoir age offsets of long-chain n-alkanoic acids from the Bengal Fan16. We calculate τsoil from these data and find that the range of values and the magnitude of deglacial changes (τsoil falls from ~200 to ~20 yrs; Table 2) are very similar to the results from the Nile River catchment. Thus, given the similarities between datasets from (sub-)tropical river catchments from two continents it is likely that drops in τsoil by one order of magnitude during Termination I were a common feature across the (sub-)tropics. Interestingly, the radiocarbon data from the Bengal Fan are correlated with rainfall indicating that variability of the Indian summer monsoon played an important role in this positive soil-carbon-climate feedback16. However, the results from the Nile River catchment do not confirm the critical involvement of hydroclimate but suggest a direct response of soil respiration rates to warming.

Table 2 τsoil and soil mean carbon ages for the Ganga–Brahmaputra river catchment during the past 17 kyrs
Full size table

Dynamic global vegetation models (DGVM) allow for investigating the effect of the decreasing τsoil on the global carbon cycle and atmospheric CO2. We revisit the analysis performed using the Lund Potsdam Jena DGVM (LPJ DGVM)54 and calculate the differences in τsoil, soil respiration (Rh) and soil carbon between the Last Glacial Maximum (LGM; 21 kyrs BP) and pre-industrial conditions (PI; 1 kyr BP). The results are shown in Fig. 4a–e. Details of the simulation are described in the methods and ref. 54.

Fig. 4: Recalculation of results from the Lund Potsdam Jena Dynamic Global Vegetation Model (LPJ DGVM) over the last 21 kyrs.
figure 4

These LPJ results are from simulations identical to those that have been forced by the Hadley center climate model as discussed in ref. 54. Relative changes between the LGM and pre-industrial conditions (PI, here: 1 kyr BP) are shown. a τsoil calculated based on the carbon influx (net primary production (NPP)). b τsoil based on the carbon efflux (Rh), where Rh is the heterotrophic respiration. Large positive anomalies (red) occur on shelf areas inundated during deglacial sea-level rise, while the areas with large negative anomalies (blue) were covered by large continental ice sheets during the LGM. Calculating τsoil from net primary production (NPP) reveals similar results as the calculation from respiration fluxes (Rh) indicating that NPP and Rh are in equilibrium. c Relative changes in NPP. d Relative changes in Rh. e Absolute changes in soil carbon (Csoil).

Full size image

As described in ref. 54, the model simulates a total increase in the global terrestrial carbon pools of 820 PgC between the LGM and PI54. This agrees well with the median of 850 PgC estimated by a recent multi-proxy approach55 showing that the simulated global patterns agree with other studies. When subtracting the effect of CO2 fertilization, the model suggests a reduction of the global land carbon stock by 200–250 PgC for areas unaffected by rising sea level or ice retreat for PI relative to the LGM54. This represents the summed-up change in vegetation and soil carbon caused by temperature and precipitation variability and is attributed to higher global turnover rates at PI54. However, we find pronounced discrepancies between our data-based reconstruction of the change in τsoil (decrease by 200 yrs, Table 1) and the simulated values for the wider (sub-)tropics (Fig. 4a, b). The model indicates marginal change in τsoil of less than 50 yrs. Substantial changes of similar magnitude as in our reconstruction are simulated only in the northern high latitudes (Fig. 4a, b). Considering the relationships in Eq. (1), the underestimation of changes in (sub-) tropical τsoil translates into underestimated, simulated changes in microbial respiration rates, respectively CO2 efflux. The discrepancies between our data-based estimates of τsoil and the LPJ DGVM simulations suggest that the climate feedback from amplified (sub-)tropical soil respiration due to the deglacial warming is underestimated in models.

The temperature sensitivity of τsoil is the key parameter for estimating changes in the soil carbon content in response to warming. Some models operate with constant Q10 values, typically 2 (refs. 5,56), but in DGVMs the relationship between temperature and soil respiration is typically described with a rather complex equation. For example, the dependency embedded in the LPJ DGVM57 is, when plotted as relative loss of soil carbon content versus temperature change, dependent on the baseline temperature T0, from which the anomalies are calculated. Results are for a T0 of 10, 20, or 30 °C similar to a Q10 of 3, 2.3, or smaller than 2.0, respectively (Fig. 5). This pronounced difference to our data-based estimate of Q10 = 10.7 (7.0–16.3) probably explains at least in parts why the simulated changes in τsoil between LGM and PI are substantially smaller than in our data-based reconstructions. However, since the data from the Indian subcontinent point to a stronger influence of precipitation on τsoil there16, but the simulated τsoil in the LPJ DGVM results are not different between Africa and India (Fig. 4a, b) some other substantial shortcomings possibly exist in the model, which we cannot identify here. Discrepancies between simulations and data-based estimates of modern τsoil, respectively terrestrial ecosystem respiration have also been documented previously4,58 and have been attributed to inaccurate parameterizations of Q10 (ref. 58).

Fig. 5: Transferring our temperature-dependent soil carbon turnover time into the Q10 concept.
figure 5

The relative carbon loss ratio (f/f0, where f0 is the efflux at ΔT = 0) as function of temperature anomaly is plotted for different Q10, including results based on recent data by Eglinton et al.31. In addition the output of this soil carbon loss rate for the equation used in the LPJ DGVM is plotted for anomalies for three different temperature baselines (Eq. (23) in ref. 57.

Full size image

Our study provides data-based evidence for a reduction in mean soil carbon turnover time and soil mean carbon ages by an order of magnitude in (sub-)tropical Africa during the last deglaciation. These results suggest that carbon sequestration via vegetation and soils was slower but the efficiency of the soils to remove carbon from the atmosphere and to protect it from biogeochemical cycling was higher. We conclude that microbial respiration rates amplified in direct response to rising temperature and that the release of pre-aged CO2 from (sub-)tropical soils into the atmosphere may have contributed to rising atmospheric CO2 and declining atmospheric ∆14C, a mechanism that has not received much of attention so far. However, for a thorough assessment of the impacts on the global carbon cycle more data-based reconstructions across the (sub-)tropics are needed to obtain a comprehensive view on the timing and magnitude of changes in τsoil and to evaluate the role of soil-carbon feedbacks outside the permafrost domain during the deglaciation. Moreover, the disagreement between our data and the LPJ DGVM simulations stresses that more research on temperature sensitivity of soil carbon turnover under different settings and different changing climatic boundary conditions is necessary to bring reconstructions and models in closer agreement.

Methods

Core material and chronology

Gravity core GeoB7702-3 was retrieved onboard RV Meteor at the continental slope off the Sinai Peninsula during cruise M52/2 in 200259. Due to the anticlockwise surface circulation in the eastern Mediterranean the fluvial load of the Nile River is transported eastward along the coast so that terrigenous biomarkers in core GeoB7702-3 serve as recorders of environmental change in the Nile River watershed17,22,23. Prior to sample preparation, the core was stored at 4°C. The sample set for bacteriohopanepolyol (BHP) quantification comprised 21 samples. Samples for compound-specific radiocarbon analysis (CSRA) were taken from 9 selected horizons (~2 cm thickness). Age depth modeling is based upon 24 radiocarbon dates of planktic foraminifera and was previously published in ref. 17 and updated by ref. 23.

Lipid extraction

Samples were freeze-dried and homogenized with a mortar. Samples for CSRA (ca. 100–120 g) were extracted with Dichloromethane (DCM):Methanol (MeOH) 9:1 (v/v) using a Soxhlet-apparatus (60 °C, 48 h) and were processed without internal standards. The samples were hydrolyzed with 0.1 N potassium hydroxide (KOH) in MeOH:H2O 9:1 (v/v) at 80 °C for two hours. Neutral compounds were extracted with n-hexane, acids with DCM after acidifying the saponified solution with hydrochloric acid (HCl). Hydrocarbons were separated from polar compounds by column-chromatography using deactivated SiO2. The hydrocarbons were eluted with n-hexane, polar compounds with DCM:MeOH 1:1 (v/v). The fatty acids were derivatized to fatty acid methyl esters (FAME). The methylation was performed with MeOH of known Δ14C, together with HCl at 50°C. Air in the headspace of the sample-tube was replaced by nitrogen gas (N2). FAMEs were recovered with n-hexane and were subsequently cleaned-up with column chromatography using deactivated SiO2 and NaSO4. FAMEs were eluted with DCM:Hexane 2:1 (v/v).

Freeze-dried sediment samples dedicated for BHP analysis (ca. 3–6 g) were extracted using a modified Bligh and Dyer extraction60. The sediment samples were ultrasonically extracted (10 min) with a solvent mixture containing MeOH, DCM and phosphate buffer (2:1:0.8, v:v:v). After centrifugation, the solvent was collected, combined and the residues re-extracted twice. The combined solvent layers were added to separatory funnels and separated from the aqueous layer by the addition of DCM and Milli-Q water. After the layers separated, the bottom layer (DCM) was drawn off and collected, while the remaining aqueous layer was washed twice with DCM. The combined DCM layers were dried under a continuous flow of N2. Aliquots of the total lipid extracts (TLEs) were obtained and DGTS (1,2-dipalmitoyl-sn-glycero-3-O-4′-(N,N,N-trimethyl)-homoserine, Avanti Polar Lipids) was added as an internal standard before ultra-high performance liquid chromatography – ultra high resolution mass spectrometry (UHPLC-HRMS) analysis.

UHPLC-HRMS analysis of non-derivatized BHPs

Non-derivatized BHPs were quantified by injecting 1% of the TLE with 2 ng internal standard (DGTS) dissolved in MeOH:DCM (9:1, v:v) on a Dionex Ultimate 3000RS ultra-high performance liquid chromatography (UHPLC) system connected to a Bruker maXis Plus Ultra-High Resolution quadrupole time-of-flight tandem mass spectrometer (UHR-qTOF-MS) equipped with an ESI ion source operating in positive mode (Bruker Daltonik, Bremen, Germany). The non-derivatized BHP analysis was performed according to ref. 61 with a column temperature of 30 °C and a modified separation method. Briefly, separation was achieved on an Acquity BEH C18 column (2.1 × 150 mm, 1.7 μm particle size, Waters, Eschborn, Germany) and a solvent system consisting of eluent A of MeOH:H2O (85:15) and eluent B MeOH:isopropanol (1:1) with both containing 0.12 % (v/v) formic acid and 0.04 % (v/v) aqueous ammonia. Compounds were eluted with 5% B for 3 min, followed by a linear gradient to 60% B at 12 min and then to 100% B at 50 min and holding at 100% B until 80 min. The column was then equilibrated for 20 min leading to a total run time of 100 min. The flow rate was held constant at 0.2 ml min−1. Mass spectra were acquired in positive ion monitoring of m/z 50 to 2000 and data-dependent fragmentation of the most abundant ions (dynamically selected, typically 3–8) for a total cycle time of 2 s and dynamic exclusion (activation after 5 spectra, release after 15 s). Ion source settings and parameters for detection and fragmentation of BHPs were optimized while infusing extracts. Every analytical run was mass-calibrated by loop-injection of Agilent ESI-L tune mix and lock mass calibration (m/z 922.0098, added in ESI source) of each mass spectrum, leading to typical mass deviations of <1–3 ppm.

BHPs were identified based on the exact mass of the protonated or ammoniated molecular ion, relative retention time and MS2 fragmentation similar to ref. 61. Extracted ion chromatograms (EIC) of the most abundant molecular ion (10 mDa mass accuracy window) were used to (semi-)quantify individual BHPs by peak integration. MS variability and ion suppression was controlled by the peak area of the DGTS internal standard. As no authentic standards were available for BHP quantification, abundances are reported based on peak areas of the individual BHPs normalized to the dry weight of the extracted sediments (i.e., in arbitrary units (AU)/µg dw).

Purification of leaf-wax lipids

For CSRA the target FAMEs and n-alkanes were purified using preparative capillary gas chromatography62. The purification was performed on an Agilent 7890B gas chromatograph (GC), equipped with a temperature programmable cooled injection-system (CIS, Gerstel) and connected to a preparative fraction collector (PFC, Gerstel). Separation was performed on a Restek Rxi-1ms fused silica capillary column (30 m, 0.53 mm i.d., 1.5 µm film thickness). All samples were injected repeatedly with 5µL per injection from a concentration of 1 µg/µl (FAMEs) and 500 µg/µl (n-alkanes) using n-hexane. The injector was operated in solvent vent mode (vent: 100 ml/min, 0 psi until 0.12 min). The CIS temperature program was: 60°C (0.05 min), 12 °C/s to 320 °C (5 min), 12 °C/s to 340 °C (5 min). The GC temperature program was set: 60 °C (2 min), 20 °C/min to 150 °C, 8 °C/min to 320 °C (40 min). Helium was used as carrier gas (4.0 ml/min). The transfer line and PFC were heated at 320°C while the traps for collection were maintained at room temperature. The backflush system of the PFC was constantly switched off. The traps were rinsed with n-hexane to recover the purified compounds. Splits (0.1%) were analyzed by GC-FID to check for potential contaminants and to quantify the purified target compounds for CSRA.

CSRA

The isotopic ratio (14C/12C) of the FAMEs and n-alkanes was determined by Accelerator Mass Spectrometry (AMS). The measurements were carried out on the Ionplus MICADAS-system equipped with a gas-ion source63,64,65 at the Alfred Wegener Institute Helmholtz Centre for Polar and Marine Research, Bremerhaven. CSRA was performed according to the protocols described in ref. 66. In short, the purified individual target compounds were transferred into tin capsules and packed. As for FAMEs, the n-C26:0 and n-C28:0 homologues were prepared individually except for two samples for which the homologues had to be combined in order to achieve adequate sample size (Supplementary Table 1). For n-alkanes we combined the n-C29, n-C31 and n-C33 homologues to obtain enough material for dating. Samples were combusted via the Elementar vario ISOTOPE EA (Elemental Analyzer) and the produced CO2 was directly transferred into the coupled MICADAS. Radiocarbon contents of the samples were analyzed along with reference standards (oxalic acid II; NIST 4990c) and blanks (phthalic anhydride; Sigma-Aldrich 320064) and in-house reference sediments. In order to account for 13C isotopic fractionation, the 14C/12C by convention is normalized to a δ13C value of −25‰ PDB, the postulated mean value of terrestrial wood67. Blank correction and standard normalization were performed via the BATS software68. The AMS results are reported as “fraction modern carbon” (F14C) and Δ14C as defined in ref. 67.

Assessment of procedure blanks and correction

To correct for carbon introduced during sample processing, procedure blanks were assessed by isolating n-alkanoic acids from a modern and a fossil standard material according to the methods described above. Leaves of a corn plant, collected in 2019, were used as modern standard (F14C: 1.0096 ± 0.0024) while “Rekord” coal-briquette (lignite from Lusatia, Eastern Germany) served as fossil standard (F14C: 0.0019 ± 0.0002). For the coal, asphaltene precipitation was performed additionally using DCM:MeOH 97:3 (v/v) and pentane. The mass and the F14C of the procedure blank were assessed using a Bayesian approach according to ref. 69. The blank had a mass of 3.079 ± 0.433 µgC with an F14C of 0.529 ± 0.072. Blank-correction of the samples and error propagation was performed using mass balance. The blank corrected F14C-values of FAMEs were further corrected for the methyl-group, which had been added during the derivatization process, using isotopic mass balance.

14C-ages of the lipids at the time of deposition

The age of the compounds at the time of deposition can be calculated using the “reservoir age offset” (R)32 which describes the age offset (in 14C years) between two carbon reservoirs at a given time32. In our case it was calculated from the ratio of the radiocarbon contents of the sample and the atmosphere at the time of deposition in marine sediments (Eq. (3)).

$${\mbox{R}}=8033 \times {{\mathrm{ln}}}\left(\frac{{{{\mbox{F}}}^{14}{\mbox{C}}}_{{\mbox{initial}}}}{{{{\mbox{F}}}^{14}{\mbox{C}}}_{{\mbox{atm}}}}\right)$$
(3)

where F14Cinitial is the F14C-value the sample had at the time of deposition at site GeoB7702-3 and F14Catm is the radiocarbon content of the atmosphere. F14Cinitial can be calculated by correcting the measured F14C-value of the sample (F14Csample) for the decay that has taken place since the deposition (Eq. (4)).

$${{{\mbox{F}}}^{14}{\mbox{C}}}_{{\mbox{initial}}}={{\mbox{F}}}^{14}{{\mbox{C}}}_{{\mbox{sample}}} \times {{\mbox{e}}}^{{{{\rm{\lambda }}}}t}$$
(4)

where t is the time of deposition and λ the decay constant of radiocarbon67. The time of deposition was inferred from radiocarbon dates of planktic foraminifera (core chronology)23. F14Catm values were adopted from IntCal2052. In case of samples for which the F14C values of the n-C26:0 and n-C28:0 homologues had been measured separately, we calculated R from the abundance-weighted mean of the F14C values in order to keep comparability with samples for which the two homologues had been combined prior to AMS measurement (Supplementary Table 1).

Calculation of τ soil and soil mean carbon ages

The authors of ref. 31 discovered a linear relationship between the 14C-ages of long-chain n-alkanoic acids and catchment-weighted mean τsoil in a global dataset of near-coastal sediments, suspended coastal sediments near river mouths, riverbeds and banks as well as suspension load (Eq. (5)).

$${{\mbox{Age}}}_{n-{\mbox{alkanoic}} \; {\mbox{acid}}}{\mbox{=}}40.1\times{{{{\rm{\tau }}}}}_{{\mbox{soil}}}$$
(5)

where the agen-alkanoic acid is given in 14C years31. Under the premise that this relationship has remained constant since the last glacial, we calculated τsoil from Eq. (5) using the reservoir age offsets at the respective time of deposition at site GeoB7702-3.

In ref. 31, constant offsets between n-alkanoic acids and soil mean carbon ages have been reported (Eq. (6)).

$${{\mbox{Age}}}_{n-{\mbox{alkanoic}} \; {\mbox{acid}}}=0.62 \times {{\mbox{soil}}} \; {{\mbox{age}}}$$
(6)

The soil mean carbon age here is defined as the age integrated over the top 100-cm depth31,51. Agen-alkanoic acid is the 14C-age31.

Using Eqs. (5), and (6), we calculated paleo-τsoil and soil mean carbon ages from the reservoir ages offsets (R; Eq. (3)) of n-alkanoic acids in core GeoB77023.

The sample set of ref. 31 covers a broad range of latitude (73 °N to 38 °S) and consequently represents different biomes and climate zones from tropical rainforest to arctic tundra. It reflects broad ranges of annual air temperature (−16 to 27 °C) and mean annual precipitation (amount 230–2200 mm/yr)31. The range of 14C-ages from n-alkanoic acids covered by the dataset is recent to >10,000 yrs31. The ages at the time of deposition calculated for the n-alkanoic acids in core GeoB7702-3 are within that range (348 ± 240 to 8723 ± 212 yrs; Table 1 and Supplementary Table 1). Thus, our inferred τsoil are within the calibrated range. Since the relationship between τsoil and the ages of n-alkanes at the time of deposition is unknown, we cannot convert our n-alkane age into τsoil.

Dynamic global vegetation model simulation

Temperature and soil moisture effects have been implemented in dynamical global vegetation models for decades45,57. For this study, we revisited the analysis performed by refs. 54,70 using the LPJ DGVM and investigate changes in τsoil, net primary production (NPP), soil respiration (Rh) and soil carbon between the Last Glacial Maximum (LGM; 21 kyrs BP) and pre-industrial (PI, 1 kyrs BP; Fig. 4). The global land carbon cycle was transiently simulated across Termination I subtracting the effect of CO2 fertilization and restricting the analysis to areas unaffected by rising sea level or continental ice retreat54. For this study, τsoil is calculated according to Eq. (1) using the simulated soil–carbon stock and the simulated NPP and Rh, respectively.