Determination of geochronology and sedimentation rates of shallow lakes in the middle Yangtze reaches using 210 Pb, 137 Cs and spheroidal carbonaceous particles

: Accurate chronologies for recent sediments of shallow lakes in the Yangtze floodplain are critical to calibrate proxy records for reconstructing environmental changes during the past century. This study presents the results of detailed 210 Pb analysis from eight lake sediment cores collected from the middle Yangtze reaches, southeast China. Unsupported 210 Pb activities generally declined exponentially with mass depth in the eight cores. The chronologies and sedimentation rates for the sediment cores were calculated using different 210 Pb-based mathematical models. The 137 Cs chronomarker (i.e. the 1963 fallout peak) and the spheroidal carbonaceous particle (SCP) chronomarker (i.e. the start of the rapid increase in 1970 AD) were selected to validate the 210 Pb dating. Sedimentation rates derived from different models were validated using historical data including lake area, arable land area, sediment discharge and reservoir volume in the Yangtze basin. The SCP-corrected CRS (constant rate of supply) model performs better than other models, based on validation using historical documents in the Yangtze basin. The 137 Cs chronomarker might be erroneous due to catchment-driven 137 Cs inputs from soil erosion and post-depositional diffusion. Both SCPs and 137 Cs are susceptible to inputs from catchment soil erosion, but SCPs show no apparent degradation and post-depositional changes in lake sediments. The SCP profile provides a relatively reliable chronomarker, which can be used for validating 210 Pb chronologies in these floodplain lakes. Generally, sedimentation rates in the eight lakes were less than 0.2 g cm -2 yr -1 before the 1930s, and then increased to a peak in the 1960s. Afterwards, sedimentation rates decreased and remained low after the 1980s.


Introduction
Reliably dated lacustrine sediments are crucial archives of past environmental changes, particularly in rapidly developing countries where long-term monitoring data are limited. For dating recent sediments spanning the past 100-150 years, 210 Pb is the most widely used radionuclide with a half-life of 22.26 years (Appleby, 2001;Sanchez-Cabeza and Ruiz-Fernández, 2012). 210 Pb has two main sources in lake sediments, including 'supported' 210 Pb ( 210 Pbsup) produced in situ by the decay of its parent radionuclide 226 Ra and 'excess' 210 Pb ( 210 Pbex) supplied by the decay of atmospheric 222 Rn (Appleby, 2001). After dry/wet deposition, rapid sedimentation processes remove this radionuclide from the water column into sediments via adsorption onto suspended particles (Krishnaswamy et al., 1971;Mabit et al., 2014). 210 Pbex decay rates are predictable based on the half-life, and hence 210 Pbex has become widely used for geochronology determination (Appleby, 2001(Appleby, , 2008Sanchez-Cabeza and Ruiz-Fernández, 2012).
In stable environments with uniform sediment accumulation rates, chronologies based on appropriate dating models are very reliable (Appleby, 2001). For example, in lakes where the sedimentation rate and 210 Pbex flux are both constant, the constant flux prerequisites of the commonly used models (e.g. constant initial concentration of 210 Pbex and constant sedimentation rate) are not always fulfilled in nature, and hence, different models may produce inconsistent dating results and be more suited to particular environmental circumstances Córdoba et al., 2017;Lan et al., 2018;Mabit et al., 2014;Tylmann et al., 2013Tylmann et al., , 2016von Gunten et al., 2009).
Since it is not known a priori which of the models produces the best result, a systematic comparison of different numerical models is vital for establishing accurate chronologies (von Gunten et al., 2009). In addition, the 210 Pb geochronology should be validated using at least one independent tracer which separately provides a clear timestratigraphic horizon (Smith, 2001;Appleby, 2008). The artificial radionuclide 137 Cs and spheroidal carbonaceous particle (SCP) profiles are widely used to validate 210 Pb chronologies (Appleby, 2001;Rose, 2001). Records of the 137 Cs fallout peak from the atmospheric testing of thermo-nuclear weapons can be used to identify the 1963 depth (Ritchie and McHenry, 1990). SCPs are a component of fly-ash formed by the incomplete high-temperature combustion of fossil fuels, and the characteristics of inert composition, efficient extraction and easy identification make them useful indicators of industrial contamination (Rose, 2001). By comparison with the history of regional industrial activities (e.g. coal consumption), changes in the SCP profile can be used as independent chronomarkers for dating recent sediments (Boyle et al., 1999;Rose, 2001;Wu et al., 2005).
The Yangtze floodplain is a particularly interesting region to study environmental changes during the past century using 210 Pb dating (Boyle et al., 1999;Xiang et al., 2002;Xu et al., 2017;Xue and Yao, 2011). This region is characterized by more than Paleolimnological studies have revealed that enhanced human activities in the Yangtze basin (e.g. large-scale deforestation, land reclamation and dam constructions) during the mid-20 th century have altered the erosional input from the watershed, subsequently changing sedimentation rates as well as the flux of 210 Pbex arriving at lake sediment (Xu et al., 2017;Xue and Yao, 2011). For example, in Taibai Lake, sedimentation rate doubled between 1945 and 1953 due to intensive land reclamation at the lake margins (Liu et al., 2012). Previous 210 Pb chronologies of lacustrine sediments in the Yangtze floodplain lakes were mainly derived from the CRS model (Boyle et al., 1999;Du et al., 2001;Xiang et al., 2002;Xu et al., 2017). One dating method alone or single model calculation may result in an over-or under-estimation of the sediment age (Tylmann et al., 2016). In order to validate the robustness of the chronologies, it has been strongly recommended that multiple dating models should be calculated and calibrated with independent chronomarkers Liu et al., 2009;Smith, 2001;von Gunten et al., 2009).
This study describes the establishment of chronologies for short sediment cores collected from eight shallow lakes of the middle Yangtze reaches. We applied three different dating approaches, i.e. 210 Pb, 137 Cs and SCP. In order to reveal the range of possible dating results, the 210 Pb chronologies were calculated using five models, i.e. the CIC, CFCS, CRS, 137 Cs-corrected CRS and SCP-corrected CRS models.
Sedimentation rates of the eight lakes were calculated and compared to historical documents of arable land area, lake area and sediment discharge from the Yangtze basin.
We expected that multiple dating approaches would help to establish reliable chronologies for these shallow floodplain lakes, in comparison with one dating method alone or single model calculation.

Study area
The Yangtze River ranks third globally in terms of length (6300 km), ninth in catchment area (1.8 ×10 6 km 2 ), fifth in water discharge (~900 km 3 per year), and fourth in sediment flux (~500 million tonnes per year, up until 1980) (Chen et al., 2001). The river is generally divided into upper (from the source to Yichang), middle (from Yichang to Hukou) and lower (below Hukou) Yangtze reaches. The Yangtze basin is characterized by a monsoonal climate that features cold, dry winters and hot, wet summers. In the middle Yangtze reaches, mean annual precipitation and mean annual temperature are ~1200 mm and 16-18 °C, respectively. At Yichang and Hankou Stations, annual runoff has been quite stable since 1950; however, sediment discharge has declined markedly since the mid-1980s (Bulletin of Yangtze River sediment, [2003][2004][2005][2006][2007][2008][2009][2010][2011][2012][2013][2014][2015]. The middle Yangtze reaches are characterized by alluvial floodplains, which encompass a number of ecologically and economically valuable lakes and wetlands (Xie et al., 2017). Due to the unique biodiversity, this region has been designated by the World Wide Fund for Nature (WWF) as one of the Global 200 priority ecoregions for conservation (Olson and Dinerstein, 1998). Increasing human disturbances from rapid population growth and economic development in the catchment have caused environmental degradation of these lakes, including surface area shrinkage due to agricultural land reclamation, biodiversity loss and water quality deterioration (Cui et al., 2013;Xie et al., 2017).

Sampling
Sediment cores were collected from eight lakes between 2011 and 2016 using a gravity corer. Each sediment core was macroscopically described, including lithology, colour and sedimentary structures, and then sectioned at 1-cm intervals in the field ( Fig.   1 and Table 1). The eight lakes include Dongting (DT), Honghu (HH), Luhu (LH), Futou (FT), Shahu (SH), Yanxi (YX), Wanghu (WH) and Poyang (PY) lakes. The sediment samples were transported to the laboratory and stored at 4 C prior to analyses. Sediment cores from HH and WH have previously been studied for 210 Pb and SCPs, and these cores were collected between 1992 and 1993 (Boyle et al., 1999). In order to retrieve recent environmental change in the two lakes, new cores were collected for further analysis.

Analytical procedures
Radioactive isotopes 210 Pb, 226 Ra and 137 Cs were counted at 2-cm intervals on a gamma spectrometer (Ortec HPGe GWL). The 137 Cs was determined via its 662 keV photopeak, while 210 Pb was determined via gamma emission at 46.5 keV and 226 Ra at 295 and 352 keV γ-rays emitted by its daughter isotope 214 Pb, following three weeks of storage in sealed containers to allow radioactive equilibration. Concentrations of 210 Pbex were calculated by subtracting the 226 Ra-supported 210 Pb concentrations ( 210 Pbsup) from the total 210 Pb activities ( 210 Pbtotal) (Appleby, 2001).
Particle size spectra of sediments were measured at 1-cm intervals using a Malvern automated laser optical particle-size analyser (Mastersizer-2000) after the removal of carbonates by 10% HCl and organic matter by 30% H2O2 with sodium hexametaphosphate as a dispersing agent. The measurement error was within 5%. Clay (<4 μm) content and median grain size were shown in this study. SCP in the sediments was extracted at 2-cm intervals using HNO3, HF and HCl according to Rose (1994).
Particles larger than 8 μm were identified and counted using an Olympus CX21 microscope at 400 magnification (Rose, 2008). SCP concentrations are expressed as grains per gram dry mass of sediment (grains g -1 ). SCPs in LH were not analysed.

Numeric models
Because the appropriate choice of the calculation model is not known a priori, the CIC, CFCS, CRS, 137 Cs-corrected CRS and SCP-corrected CRS models were chosen to calculate chronologies and sedimentation rates (Appleby, 2001;Sanchez-Cabeza and Ruiz-Fernández, 2012).
The CIC model assumes that sediments have a constant initial 210 Pb concentration regardless of accumulation rate (Shukla and Joshi, 1989). The supply of 210 Pb to lake sediment must vary directly in proportion to the sedimentation rate. Under such a scenario, the sediment age t at a depth x can be calculated using the formula where λ is the 210 Pb radioactive decay constant (0.03114 yr -1 ), and C(x) and C(0) are the 210 Pbex activity at a depth x and at the surface of the core, respectively.
The CFCS model assumes that both 210 Pb flux and sedimentation rate are constant (Krishnaswamy et al., 1971). The 210 Pbex activity versus depth relation is closely approximated by an exponential relation, in which case a plot of the logarithm of the 210 Pbex against cumulative mass depth will be closely approximated by a straight line.
The mean sedimentation rate can be calculated from the slope of the regression line.
When the 210 Pbex activities are plotted against the cumulative mass depth, the resultant least squares best fit slope is used to calculate the mass accumulation rate (ω in g cm -2 yr -1 ): The cumulative mass depth M(x) is calculated using the formula where ρ(ξ) is the dry bulk density at depth ξ.
Using the CFCS model, the sediment age t at a depth x can be calculated using the equation: Where sedimentation rates have varied, the plot of 210 Pbex versus depth will deviate significantly from an exponential relation and one of the more complex models must be used. The CRS model assumes that in these cases the flux of 210 Pbex to the sediment surface remains constant regardless of any such variations . This method relies on the ratio of the cumulative residual 210 Pbex beneath a given depth to the total 210 Pbex activity in the sediment column. It is therefore imperative that the entire 210 Pb inventory of a core is measured or estimated as accurately as possible.
The inventory is calculated as where C(m) is the mass specific 210 Pbex concentration at mass depth m, and A(x) is the 210 Pbex inventory beneath a given depth x. A(0) is the total 210 Pbex activity in the sediment column. The sediment age at a depth x can be calculated using the formula A variety of causes such as flood events and land-use change within floodplain may cause variations in the initial 210 Pbex concentration, probably resulting in erroneous dates in the models mentioned above. Where there are discrepancies, changes in the 210 Pb supply rate can be calculated using independent dates by chronostratigraphic markers (Appleby, 2001). If x1 is the depth of one reference point (e.g. the 137 Cs fallout peak and the start of rapid increase in the SCP profile) in the core with known age t1, the mean 210 Pb flux (P) above the reference point is where ∆A is the 210 Pb inventory above the depth x1.
Corrected 210 Pb dates and sedimentation rates for intermediate depths can be derived using chronostratigraphic markers by applying the principles of the CRS model with the relevant 210 Pb flux. The sediment age at a depth x above the depth x1 is determined by solving the equation where ∆ ( 1 , ) is the 210 Pb inventory between x1 and x. For example, using 137 Cs activities as independent chronostratigraphic markers, the presence of a subsurface maximum in the 137 Cs profile will in most cases identify the 1963 depth (x1).
The date beneath the given depth x1 can be calculated using the formula where T1 and Ax1 are the chronostratigraphic date at the depth x1 and the 210 Pbex inventory beneath the given depth x1, respectively.

Historical document collections
The studied eight lakes are located in Hunan, Hubei and Jiangxi provinces, respectively (Table 1). In order to validate changes in sedimentation rates of the studied lakes, continuous annual data for suspended sediment concentration  at Datong Station and total capacity of reservoirs in the Yangtze basin  were sourced from Dai et al. (2016). Discontinuous data for arable land area  and lake area  in Hunan, Hubei and Jiangxi provinces were collected from published documents (Cao et al., 2014;Cui et al., 2013). In order to provide context for the SCP profiles, continuous annual data for coal consumption in Hunan, Hubei and Jiangxi provinces  were collected from compendium of statistics for China's coal industry (China National Coal Association, 2011). The sequential t-test analysis of regime-shifts (STARS) algorithm was used to determine the shifts with the respective mean values and to evaluate the confidence levels for coal consumption data in each province and SCP concentration data of each core (Rodionov, 2004), using a cut-off length of 10 points (p < 0.05).

Particle size
Lake sediments in the study area are characterized by fine particles, with mean percentage of clay accounting for 40-60% in the eight cores (Fig. 2). Median grain size in lake sediments mainly ranged from 2 to 8 μm, except in the upper strata of HH and PY. There were some differences in the particle size spectra among the eight cores.
Median grain size in DT and FT showed fluctuating trends in the lower strata, and subsequently increased in the upper 10 cm of sediment. In HH, LH and WH, median grain size increased broadly from the lower to upper section. Median grain size in SH and YX fluctuated below 20 cm and then remained relatively stable in the upper 20 cm.
Median grain size in PY increased generally in the lower strata, followed by a decrease in the upper 10 cm of sediment.

Caesium-137 profiles
The distributions of 137 Cs in the eight sediment cores are presented in Fig. 3. To facilitate comparison of the data from the eight cores, mass depth (g cm -2 ) was used in order to reduce the influence of different rates of compaction. The 137 Cs activities in 6 out of the 8 sediment cores generally increased from 0 Bq kg -1 in the bottom layers to the peak values (ranging from 11 to 28 Bq kg -1 ), followed by declining trends (Fig. 3).
The generally low activities (< 30 Bq kg -1 ) of 137 Cs probably resulted from relatively greater losses from the lakes via the outflow, due to the high solubility of 137 Cs. 137 Cs was above 4 Bq kg -1 at the bottom of the profile in DT, indicating that shorter timespan of sedimentary material was recovered (Fig. 3a). The 137 Cs activities in PY were very low and below the detection limit (Fig. 3h), probably resulting from its fast hydraulic flushing of the lake (mean retention time of ca. 52 days; Wang and Dou, 1998 inputs. In contrast, the total 137 Cs inventory for HH, SH and WH was lower than the reference values, probably due to the removal of 137 Cs via lake outflows.

Lead-210 profiles
Supported 226 Ra concentrations were virtually constant in each core (Fig. 4). 210 Pbtotal activity reached a radioactive equilibrium with 226 Ra activity at the bottom layer of HH, LH and SH cores, but not at that of DT and PY cores (Fig. 4). In FT, YX and WH, 4), probably due to an efficiency calibration error with the gamma spectrometers. In order to remove the discrepancy, a correction value was calculated by subtracting the mean supported 226 Ra activity from the mean asymptotic activity of 210 Pbtotal at the bottom samples of FT, YX and WH, respectively. The correction value was subtracted from the original 210 Pbex activity at each level to obtain the corrected 210 Pbex activities in FT, YX and WH, respectively (Fig. 5). The activities of 210 Pbex in the surface sediments varied considerably, from 191 to 515 Bq kg -1 , almost by a factor of 2 ( Table   2). The maximum concentration was found in a subsurface layer in 7 out of the 8 sediment cores (

SCP concentration maxima in this study are similar to those observed in other
Yangtze floodplain sites, such as the Jianghan Plain (4000-15000 grains g -1 ; Boyle et al., 1999), Taihu Lake (600-9000 grains g -1 ; Cao et al., 2013;Rose et al., 2004), Longgan Lake (3000-4000 grains g -1 ; Wu et al., 2005) and Sanliqi Lake (8000 grains g -1 ; Chen et al., 2014). The variations among the sediment cores are probably influenced by SCP emission from the local industries (Boyle et al., 1999) and other site specific processes such as sediment focusing (Rose, 2001). Generally, the SCP profiles showed similar trends (Fig. 6). Low concentrations in the lower strata of the profiles were followed by rapid increases before reaching a peak. Above the peak, the SCP concentrations fluctuated around 3000-8000 grains g -1 , followed by a decrease to the sediment surface. Coal consumption in Hunan, Hubei and Jiangxi provinces increased slowly before reaching a peak around 1960, and then decreased and fluctuated between 1961 and 1969 (Fig. 7). For each province, coal consumption increased rapidly from 1970, except a transient slump around 2000 (Fig. 7). The trough around 2000 was still higher than the historical peak in 1960. The results of regime shift detection applied for coal consumption data are presented in Fig. 7, indicating three to four significant shifts (p < 0.05). The first shift happened in 1970 in Hubei, in 1969 in both Jiangxi and Hunan, respectively. In the seven studied cores, low SCP concentrations in the lower strata should be consistent with a low level of coal consumption between the 1950s and 1960s.
The start of the rapid increase in SCP concentrations was identified by the regime shift detection (Fig. 6), which should correspond to the major expansion in coal consumption since 1970. Hence, the date of the depth, above which the SCP concentrations started to increase rapidly, should be around 1970 (Figs. 6 and 7).

Chronologies and sedimentation rates
We applied the CFCS, CIC, CRS, 137 Cs-corrected CRS and SCP-corrected CRS models to the eight cores. Chronologies provided by alternative models differed, but the tendencies observed were similar for each core (Fig. 8 8c).
Sedimentation rates estimated by the CRS, SCP-corrected CRS and 137 Cs-corrected CRS models are showed in Fig.9. As a result of the discrepancies among multiple models, we obtained different changes in sedimentation rates. Generally, the sedimentation rates (SARs) increased gradually to a maximum (0.5-2.5 g cm -2 yr -1 ), and then declined in each core. The timing of maximum SARs was different for alternative models, e.g., in the 1950s using the 137 Cs-corrected CRS model, in the 1960s based on the SCP-corrected CRS model, and in the 1970s using the CRS model.

Alternative 210 Pb dating models
The chronologies estimated by alternative models showed significant discrepancy in the same core, as also observed in former studies (Blais et al., 1995;von Gunten et al., 2009). The CIC model resulted in large time inversions, especially for the lower part of the profiles, which is consistent with previous studies because the 210 Pb profile is nonmonotonic in these lakes (Baskaran et al., 2014;von Gunten et al., 2009). During the mid-20th century, widespread deforestation in the upstream regions and land reclamation of lake margin regions for farming have increased soil erosion and so sedimentation rates in the Yangtze floodplain lakes (Xiang et al., 2002;Xu et al., 2017). 210 Pbex fallout can be rapidly absorbed by the clay minerals and organic matter in the surface soil and its redistribution will occur in association with erosion and sediment transport processes (Gaspar et al., 2013). Recycled 210 Pbex fallout on catchments via surface runoff delivery and sediment focusing probably contribute to enhanced 210 Pb supply rates in these floodplain lakes (Appleby, 2008;Tylmann et al., 2016). Since most lakes have a large catchment area/surface area ratio (Table 1), the erosional input from the watershed could contribute significant proportions of 210 Pb inventories.
Alterations in initial concentrations of 210 Pb and sedimentation rates indicate the failure of the assumption of a constant sedimentation rate and constant initial concentration of 210 Pb (Abril and Brunskill, 2014), and hence the CIC and CFCS models failed to yield a reliable chronology.
The CRS model allows changes in the initial concentrations and the sedimentation rates at different layers, assuming constant 210 Pbex flux to the sediment surface . Usually, the CRS model yields more realistic results than the CIC model (Appleby, 2008). In this study, the unconstrained CRS model-yielded ages were younger than the 137 Cs and SCP chronomarkers (1963 and1970 AD) at the corresponding horizon of each core (Fig. 8). Underestimation of sediment ages by the unconstrained CRS model was also reported in previous studies (Tylmann et al., 2016).
These relatively young sediment ages probably resulted from the declining 210 Pbex inventory in the upper strata, because the influx of sediment and recycled 210 Pbex from the catchment decreased after reservoir construction and soil conservation measures were instated in the upper catchment after the 1970s (Dai et al., 2016; see the details in section 4.2). In the upper strata of the eight cores (above 10 g cm -2 ), the 210 Pbex activities declined rapidly with depth (Fig. 5), indicating that SARs were lower in the upper strata than in the lower layers (cf. Mabit et al., 2014). In cases where there is significant discrepancy between the CRS model chronologies and the 137 Cs and SCP chronomarkers, corrected 210 Pb dates can be calculated by applying the CRS model in a stepwise way to different sections of the core using the chronostratigraphic dates as reference points (Appleby, 2008).
In each lake (except PY), 137 Cs exhibited a sharp peak, suggesting that substantial mixing of the sediments had not occurred and the 137 Cs peak can be an independent stratigraphic chronomarker. In the cores from HH, FT, SH and WH, the 137 Cs-corrected CRS model overestimated sediment ages in the lower layers (stratigraphically below the 1963 fallout peak), in comparison with the unconstrained CRS and SCP-corrected CRS models (Fig. 8). The results might be linked to delayed delivery of 137 Cs from the watershed. Besides direct atmospheric fallout, 137 Cs is easily transported in dissolved form from upstream soils into lakes by surface runoff (Putyrskaya et al., 2009), which is another important source for 137 Cs in lake sediments (Ritchie and McHenry, 1990;Walling and He, 1997). The peak of the total 137 Cs in the lake stratigraphies probably reached several years after the maximum atmospheric fallout, because the continuous input of 137 Cs from the catchment into the lake, which would be exacerbated by soil erosion, dominates over radioactive decay before the peak is reached (Putyrskaya et al., 2009). Hence, the impact of 137 Cs delivery from the watershed is likely to be important.
In addition, the 137 Cs profile can be impacted by vertical diffusion because caesium is an element susceptible to dissolution (Xiang et al., 2002;Putyrskaya et al., 2009).
Apart from the initial 137 Cs peak, the secondary peak was not observed in the 137 Cs profiles (Fig. 3), indicating that the chronomarker of the 1986 Chernobyl accident is not available in our studied lakes. The peak identified as the 1986 fallout was observed in some lakes of the Yangtze floodplain (e.g. Gucheng Lake; Xiang et al., 2002), but may result from sediment mixing and upward migration of caesium. The distribution of Chernobyl fallout was very uneven and mainly controlled by the path of the radioactive plume and its coincidence with rainfall (Appleby, 2001). Former studies have revealed that the 1986 Chernobyl fallout was mainly observed at sites in Europe and the western former USSR (Ritchie and McHenry, 1990). The 137 Cs profiles of 18 lakes in East Asia and the annual 137 Cs deposition flux for Tokyo indicate that the Chernobyl fallout is not significant in the East Asian region (Zhang et al., 2012).
SCPs are inert to changes in water and sediment chemistry (Rose, 2001), and they could not be transported in dissolved form as for 137 Cs. However, SCPs may enter lakes with soil particles when catchment erosion occurs. After deposition SCPs show no apparent degradation and post-depositional changes in lake sediments Rose and Appleby, 2005). Hence, whilst soil erosion processes may deliver soil-derived SCP particles to lakes, the magnitude of changes in sediment flux was always < 10-fold, setting boundaries on the possible SCP contributions from soil erosion (Fig. 10). It is likely that a good proportion of eroded soil was old and so SCP-free. In contrast, concentrations of SCP particles increased in the order of 10 3 , making it unlikely that soil erosion processes could be responsible for altering the down-core patterns in SCP concentrations (Fig. 6). Since SCPs are solely anthropogenic in origin, their temporal trends are seen to be in good agreement with known historical coal consumption in Hunan, Hubei and Jiangxi provinces (Fig. 7). SCPs displayed slight decreases in surface sediments, probably due to the deployment of clean coal technologies, such as dust removal (Chang et al., 2016). In accordance with the results of previous studies in both HH and WH (Boyle et al., 1999), the start of rapid increase in SCPs in 1970 AD in this study provided a clear chronomarker for dating recent sediment. Hence, the accuracy of the chronologies for the sediment cores can be improved by validating with the SCP chronomarker in 1970 AD.

Sedimentation rate changes
Before the 1930s, most lakes were hydrologically connected with the Yangtze River, and human disturbance was weak. Under such a scenario, sedimentation rates in the Yangtze floodplain lakes were generally less than 0.2 g cm -2 yr -1 before the 1930s (Fig.   10a), in agreement with the results of previous studies (Chen et al., 2011;Liu et al., 2012;Xiang et al., 2002;Xu et al., 2017;Xue and Yao, 2011;Yi et al., 2004;Zeng et al., 2018;Zhang et al., 2010). Between the 1930s and the 1950s, there was an extension of arable land to meet the food demands of the rising population in the Yangtze basin (Cao et al., 2014;Fig. 10b). Firstly, floodplain lakes were reclaimed for agricultural land (Cui et al., 2013;Fig. 10c). Meanwhile, intensive reclamation and deforestation accelerated surface soil erosion, subsequently increasing sediment loading into the Yangtze floodplain lakes (Xiang et al., 2002), as suggested by rising sedimentation rates in Wushan, Shijiuhu, Guchenghu and Nanyihu lakes during this period (Zhang et al., 2010;Xue and Yao, 2011). Meanwhile, increasing summer precipitation in the early-1950s probably promoted the transport of sediment into lakes (Xue and Yao, 2011).
Accordingly, sedimentation rates derived from the SCP-corrected model reached a peak in the 1960s for most of lakes in this study, such as HH, FT, SH and PY (Fig. 10a).
After 1970, the total reservoir capacity in the Yangtze basin increased rapidly (Fig.   10d), and dam trapping caused a decrease in suspended sediment in the Yangtze River Moreover, surface soil erosion and sediment discharge declined further after 1986, as a consequence of the implementation of water and soil conservation strategies in the Yangtze basin (Dai et al., 2016). Based on the SCP-corrected CRS model, sedimentation rates in the studied lakes (except DT and LH) remained relatively stable after the 1980s, with mean values ranging from 0.09 to 0.58 g cm -2 yr -1 (Fig. 10a). Taken together, the SCP-corrected CRS model provided results that matched historical changes in arable land and sediment discharge more closely.
DT had relatively higher sedimentation rates than other lakes (Fig. 10a), probably linked to a high influx of sediment from the catchment (Du et al., 2001). Yang et al. (2007) estimated that average annual deposition of sediment was ca. 211 Mt in the Yangtze basin for the period 1956-2004, and more than half (~113 Mt) of sediment deposited in DT which is located in the upstream position among our lakes (Fig.1).
Annual deposition of sediment in DT was high, but declined gradually after the 1950s (Yang et al., 2007), whereas SARs in DT increased gradually until 2003. It is notable that water surface area of DT decreased from ~4000 km 2 in the 1950s to ~2500 km 2 in 1998, with an average decrease rate of ~30 km 2 per year from 1950 (Zhao et al., 2005).  , 2003-2015), probably accounting for a recent reduction in sedimentation rate.

Conclusions
Multiple 210 Pb-based dating models were tested to establish an accurate chronological framework for recent lake sediment in the Yangtze floodplain. Two independent time markers were used to validate the 210 Pb chronologies, including 137 Cs and SCP. Five models, i.e. CIC, CFCS, CRS, 137 Cs-corrected CRS and SCP-corrected CRS, were calculated. Compared with regional historical documents, sedimentation rates derived from the SCP-corrected CRS model matched historical changes in the Yangtze basin more closely. Due to land reclamation and deforestation, sedimentation rates generally increased between the 1930s and the 1950s, and reached a peak in the 1960s. Reservoir construction and soil conservation measures caused a decrease in sedimentation rates after the 1970s. The SCP-corrected CRS model seems to be a suitable technique for calculating chronologies and sedimentation rates in these shallow and rapidly-flushed Yangtze floodplain lakes and other similar lacustrine systems within floodplain landscapes.           Table 1 Sampling locations and detailed information of the eight studied lakes. Most of the data is sourced from Wang and Dou (1998).

Lake
Sampling site Sampling date Location Water depth (m)