lacustrine sedimentological and geochemical …...lacustrine sedimentological and geochemical...

Lacustrine sedimentological and geochemical records for the last170 years of climate and environmental changes in southeastern China

KANDASAMY SELVARAJ, BAO ZHI LIN, JIANN-YUH LOU, WEI LAN XIA, XIANG TONG HUANG AND CHEN-TUNG A.CHEN

Selvaraj, K., Lin, B. Z., Lou, J.-Y., Xia, W. L., Huang, X. T. & Chen C.-T. A.: Lacustrine sedimentological andgeochemical records for the last 170 years of climate and environmental changes in southeastern China. Boreas .10.1111/bor.12143. ISSN 0300-9483.

Reconstruction of modern climate and environmental changes in east Asia using inland natural climate archivescan provide valuable insights on decadal–multidecadal climate and environmental patterns that are probablyrelated to both natural and anthropogenic forcing. Here we investigated an 89-cm-long sediment core (TH1) fromTian Lake, southeastern China, for sedimentological, physical and geochemical parameters in order to under-stand climate and environmental changes for the latest two centuries. 137Cs- and 210Pb-based age models showthat the fine sand–coarse silt-dominated core contains ~170 years (c. AD 1842–2011) of continuous sedimenta-tion. Sediments with fine sands, low MS values, high water content, high TOC content and a high C:N ratio fromc. AD 1842 to 1897 suggest intense hydrological conditions and strong runoff in the catchment, probably becauseof a humid climate. From AD 1897 to 1990, sediments with very fine sand and coarse silt, high MS values, lowwater content and unchanged TOC and C:N ratios indicate normal hydrological conditions and in-lake algae-derived organic matter. During this interval, the chemical weathering indicators show stronger weathering condi-tions compared with sediments deposited during AD 1842–1897, supporting the dominance of weathered surfacesoil input in the earlier interval and physical erosion dominance in the later period, respectively. Since AD 1990,the continuous decrease of geochemical proxies suggests human-interacted Earth surface processes in the catch-ment of Tian Lake. A PCA revealed four dominant geochemical controlling factors – detrital input, trophic status,grain size and early diagenesis –, accounting for 26, 20, 18 and 16% of total variance, respectively. This study forthe first time provides lacustrine geochemical evidence for the most recent two centuries of climate and environ-mental changes in coastal southeastern China, a region that is currently undergoing an inversion of critical zone,i.e. an overturning of its soil profile, owing to swift modernization.

Kandasamy Selvaraj ([email protected]) and Bao Zhi Lin, State Key Laboratory of Marine EnvironmentalScience, Xiamen University, Xiamen 361102, China; Jiann-Yuh Lou, Department of Marine Sciences, Chinese NavalAcademy, Kaohsiung 81345, Taiwan; Wei Lan Xia, Key Laboratory of Lake Sedimentation and Environment, Nan-jing Institute of Geography and Limnology, Chinese Academy of Sciences, Nanjing 210008, China; Xiang TongHuang, State Key Laboratory of Marine Geology, Tongji University, Shanghai 200092, China; Chen-Tung A. Chen,Department of Oceanography, National Sun Yat-sen University, Kaohsiung 80424, Taiwan; received 17th December2014, accepted 16th August 2015.

Lake sediments have been widely regarded as animportant natural archive for the reconstruction ofpast climatic and environmental changes over centen-nial–millennial time scales in both humid and aridparts of China (Yancheva et al. 2007; Chen et al. 2010,2013; An et al. 2012; Selvaraj et al. 2012). However,similar studies on shorter, decadal–multidecadal timescales focusing on the last millennium and part thereofusing lake sediments from this region are rare. Thismay hamper our understanding of potential futureenvironmental changes, given the drastically changingland-use pattern of China during the last two decadesthat might have a great impact on the sedimentationpattern of inland aquatic systems in the near future.Investigation of the hydrodynamic characteristics anderosion-weathering processes of a lake basin using mul-ti-proxy analyses is common in palaeoenvironmentalresearch (Selvaraj et al. 2007; Nahm et al. 2010; Chenet al. 2013). For example, grain-size variations of lakesediments are generally used as a reliable proxy forpast changes in precipitation and eolian activity (e.g.Yan et al. 2011; Qiang et al. 2014). Likewise, sedimentmagnetic susceptibility (e.g. Warrier & Shankar 2009)

and colour intensities (e.g. Moy et al. 2002) are used toinfer past climate changes. In addition, sedimentaryinorganic (Ti, Al, Si, K, Fe, Mg, Ca and other traceelements) and organic geochemical proxies (TC, TOCand C:N ratio) reflect information about physical ero-sion and chemical weathering processes as well as vege-tation types in the catchment (Meyers 2004; Selvarajet al. 2007, 2011, 2012; Giguet-Covex et al. 2011; Liuet al. 2014). Hence, a combination of diverse geochem-ical parameters is useful to extract information aboutpast climate and environmental changes dependingupon the sediment accumulation rate as well as thelength of sediment core under investigation.

A substantial increase in climate and environmentalreconstructions using lake sediment records hasrecently been achieved in China (e.g. Yancheva et al.2007; Chen et al. 2010; An et al. 2012). However, mostresearchers have mainly focused upon longer-termclimate changes, with only a few studies addressingenvironmental changes at shorter, centennial–decadal,time scales, especially in southeastern China, includingTaiwan Island (Lou & Chen 1997; Selvaraj et al. 2007,2011, 2012; An et al. 2011), with a complete lack of

DOI 10.1111/bor.12143 © 2015 Collegium Boreas. Published by John Wiley & Sons Ltd

bs_bs_banner

geochemical investigation on island lakes along the SEcoast of China. There are around 6500 islands alongthe 18 000 km coastline of China, of which 93% areconsidered as not suitable for human settlementbecause of an acute shortage of fresh water. Nonethe-less, Yushan Island, located in Fujian Province(Fig. 1A), is one island that has a lot of fresh water tosupport human habitation, with a current populationof ~6500. Tian Lake, a relatively pristine lake and animportant source of fresh water in this island, hasnever been investigated either in terms of water chem-istry or sediment geochemistry (Fig. 1B). Here, weused an integrated sedimentological, physical and geo-chemical approach to infer the modern climate andenvironmental changes of Tian Lake using a short sed-iment core collected from the lake in May 2012. Weaimed to investigate (i) the hydrological changes in thecatchment surrounding the lake; (ii) the trophic statuswithin the lake; (iii) the factors that drive the sedimentinput; and (iv) the processes that control the geochemi-cal composition of sediments.

Study area

In China, the subtropical region falls between ~22 and34°N latitude and is bordered by the Qingling Moun-tain and Huaihe River in the north, the Leizhou Penin-sula in the south and extends to the HengduanMountain in the west (Wu 1980). Fujian Province,located at the southeastern boundary of the subtropi-cal area (Fig. 1A), consists mostly of mountains withundulated topography and faces the East China Sea tothe southeast. Tian Lake (latitude 26°56052″N, longi-tude 120°20056″E) is located on Yushan Island of Fud-ing City, northeast of Fujian Province and YushanIsland’s southeast border is the East China Sea(Fig. 1A). Covering an area of 21 km2, Yushan Island

is the seventh largest island in the province and has acoastline of ~30 km surrounded by the sea. The localbedrock of Fuding City consists of various rock types,including Carboniferous, Jurassic, Cretaceous andQuaternary ages, of which Jurassic and Cretaceousrocks are most developed, covering an area of about1200 km2, and accounting for ~78% of the total landarea. Yushan Island is underlain mainly by Jurassicgranites, dark grey intermediate acidic dacite and tufflava (source: Compiling Team for Annals of FudingCounty). The morphology of Yushan Island is largelydefined by an abrasion landmass covered mostly withboth coarse- and fine-grained granites (Fig. 2A–C).

Yushan Island is strongly influenced by subtropicaloceanic monsoons and thus has a mild and moist cli-mate, with mildly cold winters and humid summers.The rainfall is controlled dominantly by the East Asianmonsoon, and has an annual mean of ~1300 mm. Themean annual evaporation is 1270 mm. The annualmean temperature is 15.1 °C and the relative humidityis ~100%. The average number of foggy days per monthis 6.2 days and the soil is mostly fertile red clay type,which is suitable for grass growth. Therefore, YushanIsland has widespread grassland (Fig. 1B), covering anarea of around 667 km2. The dominant species ofgrass is Miscanthus floridulus. Affected by the topogra-phy, the land hydrology in this island is mostlyrecharged from both groundwater and rainfall.Oceanographically, Yushan Island is a region of regu-lar semidiurnal tide with mean annual macro-tidalvariation of 5.01–5.83 m and mean neap tide varyingbetween 3.37 and 4.10 m. Yushan Island containsthree natural lakes, namely Large Tian Lake, SmallTian Lake and Jiuzhuxiachao Lake, with a combinedgross water storage capacity of up to 1.7 million m3.Tian Lake has an average altitude of 306 m a.s.l., withminimum and maximum altitudes of 303 and 310 m

A B

Fig. 1. A. Map showing the location of Tian Lake on Yushan Island (black dot), southeastern China. B. A panoramic view of Tian Lake. Redarrow shows the location of the sediment core TH1 investigated in the present study.

2 Kandasamy Selvaraj et al. BOREAS

a.s.l., respectively. As the lake direction is the sameas the rock crevice direction, Tian Lake has beencategorized as a tectonic fissure lake. Small and LargeTian lakes have areas of 0.1 and 0.7 km2, respectively,with maximum depths of around 5 and 18.5 m. AsLarge Tian Lake has recently been dammed artificiallyfor drinking and agricultural purposes, we selected thepristine Small Tian Lake, which is surrounded by>100–200 m high hills (Fig. 1B), for sediment coring,aiming to investigate climate and environmentalchanges for the last ~170 years.

Material and methods

Four short (<1 m long), push sediment cores weremanually collected from chest-deep waters in TianLake during May 2012, using 4-cm diameter transpar-ent plastic tubes. Amongst them, the longest sedimentcore (89 cm), named as TH1, was chosen for this study(Fig. 4). The core was subsampled at 1-cm intervals,except for the top depth of 1–2 cm and bottom depthof 86–87.5 cm, for sedimentological (grain size), physi-cal (magnetic susceptibility and red, green and blue(RGB) colour intensities) and organic (total carbon(TC), total organic carbon (TOC), total nitrogen(TN)), as well as inorganic (major and trace elements)geochemical analyses.

Prior to the detailed laboratory work, all subsampleswere freeze-dried for 24 h to achieve a constant weightand were homogenized. A low-frequency (0.47 kHz)magnetic susceptibility analysis was performed using aBartington MS-2B sensor attached to a MS-2 suscepti-bility meter, and each sample was measured in tripli-cate. Grain-size analysis was carried out with a Coulter

LS-100 laser particle size analyser, as detailed in Jan-itzky (1987). Briefly, samples were sieved to removegrains larger than 1000 lm and then treated with HCland H2O2 to remove carbonate and organic matter,respectively. In the final stage, sodium hexametaphos-phate was added to prevent the flocculation of clayminerals and the samples received a short period ofultrasonic agitation. RGB colour intensities of sampleswere measured by digital colour scale. Grain size, mag-netic susceptibility and colour intensities were mea-sured using instruments located at the Department ofMarine Sciences, Chinese Naval Academy, Taiwan.

Forty-nine bulk sediment subsamples from every 2-cm interval were ground to � 200 micrometer in anagate mortar for geochemical analysis. Sediments wereacidified with 1 N HCl to remove carbonates and thenrinsed three times with distilled water, dried at 50 °Cand homogenized using a mortar and pestle. Precisely20 mg of sediment was weighed into a tin boat, whichwas then folded into a pellet. These pellets were mea-sured for TOC and TN using a Vario EL III elementalanalyzer (Selvaraj et al. 2007). The standard referencematerials used were BCSS (2.24% C, 0.24% N) andNIST2704 (3.34% C, 0.22% N). C:N ratios were calcu-lated as molar proportions. Selected major (Al, Ca, Fe,Mg, Mn, Na, K, P, S, Si and Ti) and trace (Ba, Rb, Srand Zr) elements were determined using an X-ray fluo-rescence (XRF) spectrometer equipped with an Rhtube located at the State Key Laboratory of MarineGeology, Tongji University, China. Details of the XRFmethod have been described in Selvaraj & Chen (2006)and Selvaraj et al. (2010). Activities of 210Pb and 137Csradionuclides in subsamples of core TH1 were mea-sured with EG and G Ortec gamma spectrometry using

A B

C

Fig. 2. A. Grey granite rocks on the mountain surrounding Tian Lake. B. Pink granite boulders, with a broken surface, on the mountainslope. C. Hammered rock pieces of coarse-grained intermediate dacite along the mountain slope.

BOREAS Climate and environmental changes, SE China 3

a germanium detector located at the Institute of Lim-nology and Geography, Chinese Academy of Sciences,Nanjing, China (Yao et al. 2008).

Results and discussion

Chronology: 137Cs dating

The radioisotope caesium-137 (137Cs–t1/2 = 30.1 years)is a fission product and was initially introduced intothe environment in significant quantities as a result ofatmospheric nuclear weapons tests conducted in theearly 1950s (Nahm et al. 2010). Widespread globaldispersal of 137Cs to the environment began withhigh-yield thermonuclear tests in November 1952 � 2(Perkins & Thomas 1980). Also, the maximum fall-outof 137Cs has been observed in the Northern Hemi-sphere during 1963, and in the Southern Hemisphereduring 1964 (Longmore et al. 1983). The fall-out of137Cs from the 1986 Chernobyl accident was mainlyobserved at sites in Europe and has been used as a timemarker for the sediment deposited during that year(Wieland et al. 1993; Fan et al. 2010).

The profiles of 137Cs in coastal sediments of Chinagenerally have two to three distinct chronologicalmarkers (Chen et al. 2009; Sun et al. 2011). As formany places in the Northern Hemisphere, the timemarkers include the onset of measurable amounts insoils in 1954 and the peak fall-out in 1963. In someareas of China, there may also be another smaller peakof 137Cs fall-out that probably derives from the Cher-nobyl fall-out in 1986 (Xiang 1998; Zhang 2005). How-ever, in some situations, the 1954 time marker may givea false result because of the detection limit of 137Cs ,which has a value of ~0.5 Bq kg�1 (Campbell 1983).

Figure 3 shows the variations in 137Cs activity withdepth for our sediment core TH1. The 137Cs pattern is

very similar to the theoretical fall-out pattern, suggest-ing that such processes might not have occurred bychance (Grousset et al. 1999). The first peak, whichhad the highest 137Cs activity (8.33 Bq kg�1), mayrecord the nuclear weapons fall-out peak that occurredin 1963, which appeared at 25–26 cm in core TH1. Sec-ondary small peaks at around 13–14 cm (~6.05 Bqkg�1) are likely to have been caused by the Chernobylfall-out. As sampling was carried out in May 2012, weassumed that the surface layer of core TH1 corre-sponds to the sediment accumulated in AD 2011. Theaverage sedimentation rate (SR) deduced from the1963 time marker is 0.53 cm a�1. The SR derived fromthe 1986 time marker is 0.54 cm a�1. It is apparentfrom Table 1 that the SR given by the 1986 marker isthe same as the SR determined from the 1963 peak137Cs activity.

210Pb dating

Lead-210 (210Pb–t1/2 = 22.3 years) is a naturalradioactive by-product of the 238U decay series that issupplied to the environment via atmospheric precipita-tion (e.g. Selvaraj et al. 2010). The 210Pb activity oflake sediments has two components, a supported com-ponent (210Pbs) derived from 222Rn decay within the

137Cs, 226Ra, 210Pb (Bq kg–1)

0 5 10 100 200 300 400 500

Dept

h (c

m)

0

10

20

30

40

50

60

70

80

90

137CS

210Pbex

226Ra

210Pbtotal

210Pbex (Bq kg–1)

0 100 200 300 400 500

Dep

th (c

m)

10

15

20

25

30

35

40

45

y = 112.33 - 16.51InxR2 = 0.84SR = 0.51 cm a–1

Fig. 3. Down-core variations in 137Cs, 226Ra, excess 210Pb (210Pbex) and total 210Pb activities in sediment core TH1 from Tian Lake. The pro-file of 210Pbex is plotted on a logarithmic scale for the depth of 13–41 cm, giving a sedimentation rate of 0.51 cm a�1.

Table 1. Comparative results of sedimentation rates calculated using137Cs and 210Pb dating methods in sediment core TH1 from TianLake, southeast China.

Time ofpeak activity

Depth(cm)

Time-span(year)

Sedimentationrate (cm a�1)

137Cs1986 13.5 1986–2011 0.541963 25.5 1963–2011 0.53

210Pbex 13–41 1932–1985 0.51


sediment column, and an unsupported (or excess) com-ponent (210Pbex) derived from the atmospheric fall-outof 210Pb. 210Pbs can, for most purposes, be approxi-mated by the 226Ra concentration. In the absence of210Pb fall-out, 210Pb and 226Ra would be in radioactiveequilibrium. 210Pbex is determined by subtracting226Ra from the total 210Pb concentration (Appleby &Oldfield 1983). The unsupported 210Pb concentrationin each sediment layer declines with its age in accor-dance with the usual radioactive decay law. This lawcan be used to calculate the age of the sediment pro-vided that the initial unsupported 210Pb concentrationwhen deposited on the lake bottom can be estimated(Appleby & Oldfield 1983).

Figure 3 shows the activity profiles of 210Pbex,210Pbt

and 226Ra in core TH1. The profile of excess 210Pb(210Pbex) shows a tendency to decrease with depth to~41 cm, but generally homogenous activity occurs at41–86 cm intervals, which is supported by the decay of226Ra. The profile of 210Pbex activity at the depth of13–41 cm shows an exponential decrease with minorfluctuations (Fig. 3). The age of sediments of depth mis thus calculated using the equation: t = k/ln (Α0/A),where t is the age of sediments of depth m; Α0 repre-sents the total excess 210Pb in the sediment column; Ais the cumulative residual 210Pb in the depth m in thesediment and k is the 210Pb radioactive decay constant(0.03114 a�1). The SR calculated from 210Pbex is0.51 cm a�1 and this value is very similar to the SR(0.53 cm a�1) derived from the 137Cs dating (Table 1).

Figure 3 also shows that 210Pbex profile is nonlinearfrom the surface to the bottom. This nonlinearity sug-gests that either sediment accumulation rates havevaried slightly over time, or that a number of other fac-tors are involved, such as migration of 210Pb throughinterstitial waters near the sediment–water interface(Koide et al. 1973), mixing of near-surface sedimentsby physical (Petit 1974) or biological (Robbins et al.1977) processes, and postdepositional redistribution ofsediments either discontinuously through slumping(Edgington & Robbins 1976) or more or less continu-ously by sediment erosion. The disequilibrium between210Pbt and

226Ra activity at the bottom of the core maymean that the lowermost appearance of 210Pb has notbeen reached in our sampling core.

Considering the uncertainty of 210Pb dating becauseof the above-mentioned factors and the close similarityof the SRs obtained between the 137Cs and 210Pb meth-ods (Table 1), we adopted 137Cs dating to calculate theage of core TH1, given that the peak 137Cs activity inour sediment core is consistent with the peak 137Csactivity of the Northern Hemisphere reported in anumber of previous studies (Zhang et al. 2005; Arnaudet al. 2006; Liu & Fan 2011). If we assume that the sur-face layer of core TH1 corresponds to AD 2011, thenthe bottom of the core (89 cm) corresponds to aboutAD 1842 (AD 1836), according to a linear extrapola-

tion using the SR of 0.53 cm a�1 (0.51 cm a�1) derivedfrom the 137Cs (210Pb) method.

Lithology and magnetic susceptibility (v)

Variations in the grain size of lake sediments may beinfluenced by changes in lake level, palaeoenvironmen-tal zones of deposition, and transport energy related tovariations in rainfall and/or wind (Chen et al. 2004;Conroy et al. 2008). As the relative humidity inYushan Island is ~100%, the grain size of Tian Lakemay be controlled by overland flows from the catch-ment. Magnetic susceptibility is widely used to provideinformation on the concentration of ferromagneticminerals present in soils and sediments (Dearing1999). Magnetic properties of soil and sedimentsdepend not only on minerals inherited from parentrocks, but also on pedogenic processes such as weath-ering and biogenic/authigenic formation of minerals(Ortega-Guerrero et al. 2004).

Lithologically, core TH1 consists of find sand, siltysand and sandy silt (Fig. 4). Mean contents of sand,silt and clay are about 72, 25 and 4%, respectively,throughout the core. The mean grain size ranges from34 to 327 lm with a mean value of 129 lm, indicatingthe dominance of fine–very fine sand and coarse silt.The median grain size varies between 35 and 618 lmwith a mean value of 170 lm (Fig. 4). The trend ofmedian grain size is similar to the mean grain size.There is not much change in the composition of sedi-ment in the bottom 15 cm and top 13 cm. In the 67–13 cm interval, which roughly equals the time intervalfrom AD 1885 to 1986 (AD 1880–1985), the grain sizefluctuated significantly, with two peaks at ~29.5–41.5 cm and ~51.5–55.5 cm. The average v value inour core TH1 is 3.2 SI. v is close to zero from the baseof the core to ~70 cm. The middle section (70–13 cm)of the record is characterized by fluctuating v values(0.8–11.5 SI), whereas the upper portion of the recordshows a decreasing trend towards the core top.

TC, TN, TOC and C:N ratio

The atomic ratio of total organic carbon to total nitro-gen (C:N ratio) of organic matter is little influenced bydecomposition and can be a reliable proxy for organicmatter sources (algae-derived or land-derived), albeitwith some small early diagenetic alterations (Meyers2004). Algae generally have atomic C:N ratios <10,whereas land plants commonly have ratios >20 (Mey-ers & Ishiwatari 1993). The TOC concentration canprovide important information about variations inorganic matter derived from land plants, which isbrought into the lake via runoff when precipitation ishigh (Meyers et al. 2006; Selvaraj et al. 2007). There-fore, a combination of high TOC content and highC:N ratios is probably related to higher inputs of soil


material accompanied by organic matter from thecatchment during enhanced rainfall. By contrast, asmall amount of land plant material was deliveredwhen rainfall was low and during which the staticwater column may also favour aquatic productivityand thus sediments in general have low C:N ratios.

The TOC content in core TH1 ranges from 0.5 to10.9%, with an average TOC content of 2.3% (Fig. 5).In the bottom ~10 cm, the TOC content is high andranges from 4.3 to 10.9%, with large oscillations com-pared with the top 78.5 cm of the core, where the TOCcontent is comparatively less, ranging from 0.5 to3.5%. TOC shows an increasing trend from 78.5 cm,corresponding to AD 1863 (AD 1857). TN content(0.09–0.53%) in the sediment core varies in concertwith TOC toward the top of the core. C:N ratios arehigh in the bottom ~10 cm of the core, whereas in thetop 78.5 cm of the core they remain mostly at ~8–10

with fewer fluctuations. The TN shows a strong linearcorrelation with TOC (R2 = 0.88, p < 0.0001; Fig. 6)with a positive intercept on the TN axis for sedimentsfrom the upper 78.5 cm, suggesting the presence of asmall amount of inorganic N (~0.03%) in TN.TOC-rich sediments from the singular indicate largerproportions of land-derived organic matter, as alsoevidenced by higher C:N ratios in these sediments(Fig. 5), although they contain proportional contentof inorganic N (~0.17%; Fig. 6), probably because ofthe presence of clay-associated N in terrestrial materi-als in the bottom part.

Proxies of chemical weathering

As soils and rocks in the catchment represent majorsources of terrigenous input into the lake (Bertrandet al. 2005), elemental concentrations in sediments are

TC (%)

0 1 2 3 4 5 8 12

TOC (%)0 1 2 3 5 10

C:N ratio3 4 5 6 7 8 9 10 16

Age

(yea

r AD

)

2011

1992

1972

1954

1935

1916

1897

1878

1859

1842

TN (%)

0.0 0.2 0.4 0.6

Dep

th (c

m)

0

10

20

30

40

50

60

70

80

90

Fig. 5. Total carbon (TC), total organic carbon (TOC), total nitrogen (TN) and the TOC to TN (C:N) ratio vs. depth and age in sedimentcore TH1.

Fig. 4. Litho-log of core TH1 along with down-core distributions of clay (<4 lm), silt (4–64 lm) and sand (>64 lm) contents (all in %) aswell as mean grain size (Ms, lm), median grain size (Md, lm) and magnetic susceptibility (v in SI) against year (in AD). The two horizontalshaded bars mark the dominance of medium-size sand in core TH1 during AD 1906–1913 and AD 1935–1953, suggesting a stronger erosionin the catchment, probably as a result of the influence of typhoons in SE China.


helpful to reconstruct the intensity of chemical weath-ering, an important process in the global hydro-geo-chemical cycle of elements (Gaillardet et al. 1999;Warrier & Shankar 2009; Gupta et al. 2011). As thisprocess is strongly influenced by rainfall and tempera-ture (Ollier 1984; Gaillardet et al. 1999), humid tropi-cal and subtropical climates with denser vegetation ingeneral enhance the weathering of source rocks (Sel-varaj & Chen 2006). Conversely, in arid or polar cli-mates, weathering intensity is weak to non-existentowing to either low precipitation or water locked in iceand thus erosion predominates over weathering withfew to no new secondary mineral formation (Schoen-born & Fedo 2011).

The degree of weathering can be obtained by calcu-lation of the chemical index of alteration (CIA):CIA = 100Al2O3/(Al2O3+CaO*+Na2O+K2O) (Nesbitt& Young 1982), and the plagioclase index of alteration(PIA): PIA = 100[(Al2O3–K2O)/(Al2O3+CaO*+Na2O–K2O)] (Fedo et al. 1995), where CaO* is the amount ofCaO incorporated in the silicate fraction. These indicesindicate the degree of decomposition of total feldspar(both alkali and plagioclase) and plagioclase feldspar,respectively, to secondary clay products, where CIA(PIA) values of about 50 indicate unweathered bedrock(plagioclase), and values of 75–100 indicate completeconversion of feldspars to aluminous clay minerals (in-tense chemical weathering; see Fedo et al. 1995). Inten-sely weathered rock thus yields kaolinite or gibbsitewith CIA/PIA values of 100 (Fedo et al. 1996, 1997).Similar to CIA and PIA, ratios such as Rb/Sr, K/Aland Al/Na are also good proxies of chemical weather-ing intensity (Selvaraj & Chen 2006; Clift et al. 2008;Selvaraj et al. 2010), as increasing weathering rapidlyleaches Sr (Na) compared with Rb (Al) (Nesbitt &Young 1982) and, therefore, the Rb/Sr and Al/Naratios increase in the weathered product (Ma et al.2000). K/Al has been used as a proxy for the illite/kaolinite ratio (Yarincik & Murray 2000), as illite rep-resents the dominance of physical weathering in tem-perate to arid climates, whereas kaolinite indicatesintense weathering in tropical, humid climates (e.g.Bonatti & Gartner 1973).

As shown in Fig. 7, the CIA values for core TH1range from 79 to 86 with an average of 84. Such highCIA values indicate that most primary feldspar miner-als in the catchment have been converted to aluminousclay minerals (Fedo et al. 1996). Consistent with the

TOC (wt. %)0 2 4 6 8 10 12

TN (w

t. %

)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

R2 = 0.70, p = 0.0197 (red)R2 = 0.88, p < 0.0001 (blue)

Fig. 6. Bi-plot showing the relationship between TN and TOC. Redsquares are sediments from the base of the core to 78.5 cm, whereasblue squares represent sediments from 78.5 cm to the core top.

1900-1986

1986-2011

1842-1900

Plagioclase index of alteration-PIA89 90 91 92 93 94 95 96 97

1842

1859

1878

1897

1916

1935

1954

1972

1992

2011

K/Al0.160.200.240.28

Rb/Sr1.2 1.4 1.6 1.8 2.0 2.2

Age

(yea

r AD

)

Chemical index of alteration-CIA78 79 80 81 82 83 84 85 86 87

Dep

th (c

m)

0

10

20

30

40

50

60

70

80

90

Al/Na30 40 50 60 70 80 90 100

Fig. 7. Age vs. chemical index of alteration (CIA), the ratio of K to Al (K/Al), the ratio of Al to Na (Al/Na), plagioclase index of alteration(PIA) and the ratio of Rb to Sr (Rb/Sr) in sediment core TH1. The horizontal shaded bar represents the interval of AD 1900–1986.


CIA values, the PIA values for core TH1 vary from 89to 96 with an average of 94, indicating that all primaryplagioclase feldspars have been converted to secondaryclay minerals. High Rb/Sr ratios (1.21–2.05) with amean value of 1.84, which is higher than the Rb/Srratios of different shales (0.80–0.88; Selvaraj & Chen,2006), supporting the idea that source rocks in thecatchment might have been intensively weatheredbecause of favourable climatic conditions (high rainfalland high temperature) in tropical and subtropicalregions (Selvaraj & Chen 2006). Rb/Sr ratios in illiteminerals are usually >1, with a maximum value of 6.46(Chaudhuri & Brookings 1979), suggesting that thehigh illite content in the lake sediments might beresponsible for the high Rb/Sr ratios. Values for K/Alrange from 0.16 to 0.27, averaging 0.21. The variationsin K/Al follow CIA, with the ratio increasing duringwarm periods and decreasing during cold periods. Thedegree of chemical weathering of lake sediments ishigh, as shown by their very high Al/Na ratios (32–95with an average value of 61), resulting from theextreme dissolution of plagioclase (Fig. 7). This weath-ering condition is comparable to lake sediments fromthe sub-alpine region of nearby Taiwan Island, whichshow extreme weathering because of the high precipita-tion and dense vegetation (Selvaraj & Chen 2006; Sel-varaj et al. 2007). This similarity also suggests regionaluniformity of chemical weathering of rocks in subtrop-ical Taiwan and SE China, although the catchment ofTian Lake is devoid of meta-sedimentary rocks, whichare dominantly present in Taiwan Island.

Principal component analysis (PCA)

Geochemical factor analyses have been used for deter-mining the factors that control the sources of sediment(Selvaraj et al. 2010) and processes that control thegeochemical composition of coastal and river sedi-ments (Chen & Selvaraj 2008). Therefore, to disentan-gle significant grouping of factors and relationships topossible geochemical controlling processes, all of thesediment geochemical data, including grain size, mag-netic susceptibility and colour intensities, were sub-jected to factor analysis using SPSS software (version17.0, Chicago, IL, USA). The factors were extractedusing the Varimax rotation scheme with Kaiser nor-malization.

The results reveal that four factors account for 80%of the total variance (eigenvalue >2, Table 2). Factor 1accounts for 26% of the variance and Ti, Al, Mg, K,Pb, Ba, Rb and Sr are positively loaded in this factor.The strong positive correlations of Ti (r = 0.85), Ba(r = 0.79), Pb (r = 0.75), Rb (r = 0.74), Sr (r = 0.62),K (r = 0.79) and Mg (r = 0.83) with Al indicate thatthey are associated with aluminous clay minerals suchas kaolinite, illite and/or smectite Table 3. In general,all these elements showed a similar pattern over the

whole core (Fig. 8). Ti, Rb, Ba, Sr, K and Mg are ingeneral incorporated into clay minerals during chemi-cal weathering, whereas Ca and Na tend to be leached(Nesbitt et al. 1980; Fedo et al. 1996). These elementsare mainly derived from the catchment and thusrelated to detrital input into the lake.

Variables such as intensities of red, green and bluecolours, and magnetic susceptibility are positivelyloaded and Ca and S are negatively loaded in factor2, which accounts for 20% of the total variance. Therelationships between colour and the physical andchemical characteristics of soil, such as iron, organicmatter, moisture content, mineralogy and structure,have been studied (Krishnamurti & Satyanarayana1971; Davey et al. 1975). It is evident from Fig. 9that Ca and S may be responsible for the low valuesfor colour intensities as well as for magnetic suscepti-bility in the bottom 10 cm of the core. The covari-ance of Ca, S, RGB and magnetic susceptibilitysuggests that similar environmental conditions causedtheir accumulations/variations. High C:N ratios aregenerally associated with land plants and the increasein sediment TOC in the bottom part of the core mayreflect an increase in land-derived organic matter inthe lake (Fig. 5). When land-derived organicsincrease, lakes become anoxic, resulting in high TOCpreservation in sediment and thus reducing sedimentmagnetic susceptibility values. Such an interpretationreceives support from the S profile in our study.

Table 2. Varimax rotated factor matrix (n = 49). Significant load-ings (>0.6) are in bold.

Variable Factor 1 Factor 2 Factor 3 Factor 4

Si �0.17 0.12 0.13 �0.90Ti 0.78 0.23 �0.15 �0.31Al 0.90 0.23 �0.05 �0.24Fe �0.28 0.14 �0.22 0.86Mg 0.69 0.10 �0.12 �0.52Ca �0.08 �0.75 0.23 0.08Na 0.06 �0.24 0.23 0.71K 0.71 0.21 �0.05 �0.46P �0.16 �0.51 0.11 0.53S 0.09 �0.69 0.20 0.21Mn �0.43 �0.03 �0.22 0.60Pb 0.83 0.30 �0.14 0.14Ba 0.94 �0.08 0.07 0.03Zr 0.40 0.41 �0.17 0.69Rb 0.83 0.38 �0.17 0.26Sr 0.83 0.03 �0.03 0.37MGS 0.04 �0.13 0.95 �0.11Md 0.13 0.04 0.91 �0.06Clay 0.22 0.49 �0.73 �0.01Silt 0.24 0.24 �0.90 0.02Sand �0.24 �0.28 0.89 �0.02Red 0.36 0.85 �0.22 0.12Green 0.39 0.83 �0.21 0.05Blue 0.43 0.80 �0.20 �0.10MS 0.04 0.83 0.08 �0.01Eigenvalue 6.49 5.09 4.42 4.12Variance (%) 26 20 18 16


Therefore, factor 2 may be interpreted as a represen-tation of the trophic status.

Factor 3 explains about 18% of the total varianceand has negative loadings for clay and silt contentsand positive loadings for sand content, mean grain sizeand median grain size. As variations in the grain-sizedata are primarily attributed to catchment runoff vari-ability, factor 3 can be interpreted as representing rain-fall variations in the study area. Two peak values ofmean and median grain sizes around AD 1906–1913and AD 1935–1953 (Fig. 4) perhaps seem to be attrib-uted to the influence of typhoons in the study area,given that heavy downpours associated with tropicalstorms can transport coarser sediments, although acausal link needs to be clarified in future studies.

Factor 4 accounts for 16% of the total variance andshows positive loadings with Fe, Na, Zr and Mn anda negative loading with Si, probably suggesting the

influence of Fe-Mn oxyhydroxides in the sediments.Figure 10 shows enrichment of Fe and Mn in the sur-face sediment, which may be attributed to the redox-related diagenetic recycling of iron and manganese insediments (Boyle et al. 1999; Boyle 2001). Mn and Fein sediments are dissolved under reducing conditions,migrate upward and accumulate in the oxidized surfacesediments. Si is not only associated with many alumi-nosilicate minerals, but is also related to diatom pro-ductivity as a component in their frustules (Peinerud2000). Based on poor correlations with K, Ti and Rb(Table 3), Si behaviour seems to be controlled by pri-mary productivity. The negative relationship betweenSi and Zr found here substantiates a biogenic source ofSi (Table 3).

Consistently, the depth profile of the Si:Al ratio(Fig. 11), an approximate indicator of biogenic silica,shows that the Si:Al ratio of the lake sediments ismostly less than the mean ratio of 4.13 obtained fromeight rock samples investigated in the catchment(range: 3.07–4.66). The plot also includes the Si:Alratio of two soil samples (2.01 and 2.32) investigatedfrom the catchment with a mean ratio of 2.17. TheSi:Al ratio in sediments depends upon the Si:Al ratioof the settling suspended particles. If the suspendedmatter is dominated by detrital particles (i.e. physicallyeroded soil or bedrock particles), the Si:Al ratio will be

Ti (p

pm)

3500

4000

4500

5000

5500

6000

6500

Mg

(%)

0.250.300.350.400.450.500.550.60

K (%

)

1.6

1.8

2.0

2.2

2.4

2.6

2.8

Rb

(ppm

)

60

80

100

120

140

160

180

Ba

(ppm

)

400450500550600650700750

Pb (p

pm)

20

25

30

35

40

45

50

Sr (p

pm)

40

50

60

70

80

90

Depth (cm)0 10 20 30 40 50 60 70 80 90

Al (

%)

6

8

10

12

14

1900-1986 1842-19001986-2011

Fig. 8. Down-core variations in elements of Factor 1, representingthe detrital input. The vertical shaded bar represents the interval ofAD 1900–1986.

S (p

pm)

0

1000

2000

3000

4000

5000

6000

Ca

(ppm

)

1500

2000

2500

3000

3500

4000

4500

Blu

e

0

50

100

150

200

250

300

Red

10

20

30

40

50

60

Depth (cm)0 10 20 30 40 50 60 70 80 90

MS

(SI)

02468101214

Gre

en

40

80

120

160

200

2401986-2011

1900-1986 1842-1900

Fig. 9. Down-core variations in elements of Factor 2, representinglake trophic levels and/or pedogenesis. The vertical shaded bar repre-sents the interval of AD 1900–1986.


Table

3.Correlation

coefficients(r)betw

eendifferentpa

irsof

variable(n

=49

).Boldvalues

aresign

ifican

tat

the0.01

level.MGS=meangrainsize;M

d=mediangrainsize;M

S=bu

lkmagneticsusceptibility.

Ti

Al

Fe

Mg

Ca

Na

KP

SMn

Pb

Ba

Zr

Rb

SrMGS

Md

Clay

Silt

Sand

Red

Green

Blue

MS

Si0.11

0.01

�0.77

0.27

�0.03

�0.63

0.29

�0.52

�0.31

�0.49

�0.18

�0.16

�0.61

�0.31

�0.39

0.19

0.13

�0.07

�0.14

0.13

�0.10

�0.05

0.07

0.14

Ti

0.85

�0.35

0.89

�0.31

�0.23

0.80

�0.33

�0.23

�0.36

0.64

0.66

0.20

0.65

0.44

�0.10

0.01

0.37

0.35

�0.36

0.46

0.51

0.58

0.21

Al

�0.40

0.83

�0.33

�0.20

0.79

�0.37

�0.15

�0.52

0.75

0.79

0.28

0.74

0.62

�0.01

0.11

0.33

0.29

�0.30

0.49

0.55

0.61

0.19

Fe

�0.49

�0.20

0.40

�0.47

0.44

�0.08

0.79

�0.05

�0.26

0.61

0.11

0.02

�0.32

� 0.27

0.10

0.15

�0.15

0.11

0.03

�0.13

0.02

Mg

�0.31

�0.37

0.89

�0.32

�0.27

�0.40

0.46

0.62

�0.02

0.51

0.24

�0.04

0.02

0.22

0.26

�0.26

0.25

0.32

0.41

0.10

Ca

0.32

�0.41

0.38

0.69

0.07

�0.32

0.03

�0.39

�0.44

0.09

0.31

0.21

�0.44

�0.34

0.36

�0.62

�0.60

�0.55

�0.47

Na

�0.28

0.55

0.45

0.27

�0.04

0.10

0.26

0.06

0.22

0.15

0.12

�0.19

�0.22

0.22

�0.11

�0.14

�0.20

�0.05

K�0

.38

�0.33

�0.45

0.48

0.69

0.07

0.62

0.31

�0.02

0.03

0.25

0.20

�0.21

0.30

0.37

0.45

0.24

P0.33

0.40

�0.33

�0.08

�0.02

�0.23

�0.06

0.13

0.01

�0.41

�0.25

0.28

�0.42

�0.45

�0.49

�0.45

S�0

.15

�0.09

0.08

�0.20

�0.26

0.14

0.26

0.17

�0.37

�0.26

0.28

�0.45

�0.46

�0.45

�0.47

Mn

�0.31

�0.32

0.27

�0.14

�0.18

�0.27

�0.25

0.00

0.07

�0.06

�0.16

�0.21

�0.30

�0.02

Pb

0.70

0.63

0.87

0.82

�0.15

0.00

0.42

0.41

�0.42

0.59

0.59

0.57

0.21

Ba

0.34

0.77

0.81

0.08

0.15

0.10

0.12

�0.12

0.22

0.26

0.30

0.06

Zr

0.75

0.62

�0.29

�0.15

0.32

0.33

�0.33

0.57

0.51

0.37

0.26

Rb

0.81

�0.22

�0.07

0.45

0.43

�0.44

0.63

0.63

0.59

0.32

Sr�0

.04

0.08

0.26

0.27

�0.27

0.39

0.39

0.37

0.08

MGS

0.96

�0.70

�0.85

0.84

�0.33

�0.30

�0.26

�0.08

Md

�0.54

�0.71

0.70

� 0.14

�0.11

�0.07

0.07

Clay

0.87

�0.90

0.66

0.68

0.70

0.40

Silt

�1.00

0.50

0.51

0.50

0.15

Sand

�0.53

�0.54

�0.54

�0.19

Red

0.99

0.95

0.68

Green

0.98

0.68

Blue

0.67


close to the ratio in the soil or bedrock within thecatchment. If we assume that the soil in the catchmentis free of biogenic silica, then the depth profile of theSi:Al ratio suggests the presence of biogenic silica(5–10%), especially in the sediments deposited from68.5–86 cm (20 cm) and sediments deposited from12.5–3.5 cm in core TH1 (Fig. 11). However, the totalCaO content of the sediments measured using XRFwas always <0.6% (Fig. 11). This amount is very lowcompared with the mean content of CaO (1.99%) ineight rock samples analysed from the catchment ofSmall Tian Lake, suggesting that the sediments of thislake contain insignificant amounts of biogenic materi-als in the form of carbonate and thus are dominantlydetrital in nature. The negative relationship of Si withFe and Mn therefore indicates the dilution effect ofauthigenic Fe-Mn hydroxides on detrital silica ratherthan biogenic silica (Boyle et al. 1998). Thus, factor 4may be regarded as representing the dilution of detritalinput by Fe-Mn enrichment resulting from early diage-netic reactions.

Thus, when arranged all four factors depth-wise(Fig. 12), changing PCA scores reveals the importanceof relative components to the overall geochemicalrecord. PCA1 shows larger variation at a depth of 86–56 cm, corresponding to AD 1848–1905 (AD 1842–

1901) and then decreases slightly to a depth of13.5 cm. PCA1 decreases significantly over the top13 cm. PCA2 increases from the bottom to a depth of

Si:Al1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

Dep

th (c

m)

0

10

20

30

40

50

60

70

80

90

Total CaO%0.2 0.3 0.4 0.5 0.6

Mea

n - 4

.13

Ran

ge o

f roc

ks -

3.07

-4.6

6

Ran

ge in

soi

ls -

2.01

-2.3

2M

ean

- 2.1

7Fig. 11. Down-core variations in Si:Al ratio and CaO content, show-ing the presence of biogenic materials in the form of silica in coreTH1. The vertical grey bars represent the ranges of the Si:Al ratiosof eight rock samples and two soil samples investigated in the studyarea.

Si (%

)

24

26

28

30

32

34

36

Fe (%

)

0

2

4

6

8

10

Na

(ppm

)

0.120.140.160.180.200.220.240.260.28

Depth (cm)0 10 20 30 40 50 60 70 80 90

Mn

(ppm

)

100200300400500600700800900

Zr (p

pm)

100

200

300

400

500

600

1986-2011 1900-1986 1842-1900

Fig. 10. Down-core variations in elements of Factor 3, representingrunoff variability. The vertical shaded bar represents the interval ofAD 1900–1986.

Age (year AD)

Det

rital

inpu

tPC

A1

–3

–2

–1

0

1

2

PCA

2Tr

ophi

c st

atus

–3–2–10123

Gra

in s

ize

PCA

3

–2

–1

0

1

2

3

Depth (cm)0 10 20 30 40 50 60 70 80 90

PCA

4Ea

rly d

iage

nesi

s

–2

–1

0

1

2

3

2011

1992

1935

1916

1842

1972

1954

1897

1878

1859

1986-2011 1900-1986 1842-1900

Fig. 12. Age vs. the four main components (PCA1, PCA2, PCA3and PCA4) extracted by the principal component analysis (PCA).The vertical shaded bar represents the interval of AD 1900–1986.


around 33 cm and then decreases slightly to the top.PCA3 shows a decreasing trend from the bottom to adepth of 63.5 cm and then varies relatively signifi-cantly. PCA4 decreases from the bottom to a depth of66.5 cm and then increases to the top.

Palaeolimnological interpretation

The bottom layer of core TH1 is dated at AD1842 (AD 1836), thus enabling us to reconstructthe last 170 years of climate and environmentalchanges in southeastern China. Three periods canbe distinguished: two short, erosion-dominated ter-rigenous input intervals at 1842 to 1900 (89–58 cm)and 1986 to 2011 (13–0 cm), and a relatively longperiod (1900–1986) of weathering-dominated terrige-nous input (Fig. 8).

During the period from AD 1842 to 1900, the min-eral load was generally low and punctuated by extremeevents of relatively short duration, probably related toextreme cold fluctuations that occurred during thistime (Figs 7, 8). It has been suggested that southeast-ern China experienced a very severe cold winter in AD1892, when many lakes (e.g. Tai Lake and DongtingLake) and rivers (e.g. low reaches of Huaihe, Huangpuand Yangtze Rivers) became frozen for a long dura-tion, as recorded in Chinese historical documents(Zheng et al. 2012). After AD 1891, the mineral loadsshow an abruptly decreasing trend, which may corre-spond to this severe cold event. In this interval, somepeaks of TOC, TN and the C:N ratio are consistentwith mineral loading, indicating that the organicsource of sediment was mainly terrestrial material(Fig. 5). Down-core variations in CIA, PIA, K/Al, Rb/Sr and Al/Na also show relatively lower values(Fig. 7), indicating relatively weak chemical weather-ing of rocks in the catchment during this interval. Dur-ing AD 1842 to 1876 (71–86 cm), the relatively highvalue and stable level of grain size and high water con-tent suggest wetter than normal hydrological condi-tions during the formation of sediment, probablyrelated to high precipitation. High values of TOC(>4%), TN (>0.2%) and C:N ratios (>10) further sup-port high levels of runoff in the catchment as a resultof high rainfall and thus delivery of terrestrial materi-als into the lake. The v-value is close to zero from thebase of the core to 71 cm, which is consistent with highwater content and TOC values of sediments, suggest-ing that the organic matter is diamagnetic in nature(Figs 4, 5). High concentrations of S and Ca in thebottom 15 cm suggest a sub-oxic/anoxic environmentwithin the lake probably because of high water levels,supporting the theory of heavy rainfall during AD1842–1876 (Fig. 9). Therefore, the climate was gener-ally cold from AD 1842 to 1900, although the climatefrom AD 1842 to 1876 was more humid than duringAD 1876–1900. Such an interpretation of less humid

conditions from AD 1876–1900 is roughly consistentwith a higher number of typhoons observed in c. AD1870–1930 in Taiwan when low solar activity prevailed(Hung 2013).

Overall, from AD 1900 to 1986, the levels of ele-ments related to detrital input show a highly weatherednature of sediments, although they decreased up-coreslightly (Figs 7, 8). All chemical weathering indicesindicate greater chemical weathering of rocks in thecatchment compared to sediments that accumulatedduring AD 1842–1900, probably because of a warmand humid climate (Fig. 7). However, distinct varia-tions in grain size suggest that the depositional envi-ronment was relatively turbulent (Fig. 4). For instance,two medium-size sand-dominated layers in the middleof core TH1 correspond to AD 1906–1913 and AD1935–1953 (137Cs dating; Fig. 4) and AD 1902–1910and AD 1931–1951 (210Pb dating), suggesting that thehydrodynamic conditions in the catchment were prob-ably wetter with high surface runoff during these inter-vals. This may imply the influence of typhoons aroundthese time intervals (Fig. 4). Given that there is noprominent river input into the lake and that the lake issurrounded by hills (~100–200 m a.s.l.), the depositionof medium-size sand-dominated layers in these inter-vals might have contributed to greater hill slope ero-sion in the catchment. Consistent with this inference,the reconstruction of temperatures for the last500 years from historical records in China revealedthat the warmest interval was from the 1920s to the1940s (Wang et al. 1991). Likewise, the historicalinventory contains 41 typhoon strikes during the 26-year interval between AD 1884 and 1909 (Liu et al.2001). The position of the medium-size sand-domi-nated layers in core TH1 roughly corresponds to thewarmest epoch and the period of intense typhoonstrikes, indicating that our sedimentary proxy records,in particular the grain-size proxies, may respond to theinfluence of stronger typhoons in SE China and thusmay be an ideal natural archive for the reconstructionof decadal–centennial scale climate changes in SEChina. The relatively low C:N ratios with few fluctua-tions suggest an algal source of organic matter andthus a lacustrine environment during this interval. Theamounts of Fe and Mn increase as a result of precipita-tion of Fe-Mn oxyhydroxides, resulting in increases inv and RGB intensity (Figs 9, 10). Low values of Caand S also substantiate the existence of a warm andhumid climate during this interval (Fig. 9).

In the last ~25 years (AD 1986–2011), the mineralload reduced abruptly and then slightly increased(Fig. 8). Chemical weathering proxies also show trendssimilar to the mineral loads (Fig. 7). The mean grainsize shows smaller fluctuations, probably indicatingonly one source of sediment. The median grain size isalso relatively constant, indicating a stable depositionalenvironment (Fig. 4). These changes may be attributed


to changes in land use patterns over the last twodecades, an interpretation that is roughly consistentwith the expansion of the urban construction area innearby Fuzhou City to 64.2 km2 during the AD 1988–2000 period (Bai & Chen 2013). High concentrationsof Fe and Mn, however, suggest a well-oxygenated sed-iment–water interface probably because of low rainfall,resulting in a lower lake water level. When the lakelevel is low, it is expected that the core location becomecloser to the catchment, resulting in either the accumu-lation of larger grain sizes at the core site, or the preci-pitation of Fe-Mn oxyhydorxides, both might havediluted the detrital input (Figs 7, 8, 10).

Conclusions

Sedimentological, physical, and organic and inorganicgeochemical records of a sediment core from TianLake in SE China have provided unprecedentedrecords of climatic and environmental changes andrelated catchment processes for southeastern coastalChina over the last ~170 years. Our study revealed thatthe climate was mostly wet during this time period.Although generally wet, the geochemical parametersindicate the dominant input of physically eroded mate-rials during c. AD 1842–1900, whereas chemicallyweathered surface soil input seems to have dominatedover a relatively longer period from c. AD 1900–1986.The situation was completely different since AD 1986until the end of the core dating period in 2012, duringwhich time geochemical parameters suggest the influ-ence of land use changes on the sedimentation patternand/or dilution of detrital elements by early diageneticenrichments of Fe and Mn. Given that most coastalestuaries and shallow coastal systems have alreadybeen altered because of intense economic developmentalong the coastal zone of SE China, the wide variety ofphysical and geochemical data provided here will beuseful for the future assessment of climate and environ-mental, including man-made, changes to the coastaland inland aquatic systems of SE China.

Acknowledgements. – This work was supported by the NationalScience Foundation of China (grant no. 41273083), Shanhai Fund ofXiamen University, China (grant no. 2013SH012) and Open Fund ofTongji University, China (grant no. MGK1201). We thank S. J. Kao,James Ho and T. Y. Lee for their help during the fieldwork and assis-tance in collecting push sediment cores and B. J. Wang for C and Nanalyses. We gratefully acknowledge Phil Meyers and an anonymousreviewer for their helpful comments and critique of the originalmanuscript.

References

An, Z. S., Clemens, S. C., Shen, J., Qiang, X. K., Jin, Z. D., Sun, Y.B., Prell, W. L., Luo, J. J., Wang, S. M., Xu, H., Cai, Y. J., Zhou,W. J., Liu, X. D., Liu, W. G., Shi, Z. G., Yan, L. B., Xiao, X. Y.,Chang, H., Wu, F., Ai, L. & Lu, F. Y. 2011: Glacial-interglacialIndian summer monsoon dynamics. Science 333, 719–723.

An, Z. S., Colman, S. M., Zhou, W. J., Li, X. Q., Brown, E. T., Ti-mothy Jull, A. J., Cai, Y. J., Huang, Y. S., Lu, X. F., Chang, H.,Song, Y. G., Sun, Y. B., Xu, H., Liu, W. G., Jin, Z. D., Liu, X. D.,Cheng, P., Liu, Y., Ai, L., Li, X. Z., Liu, X. J., Yan, L. B., Shi, Z.G., Wang, X. L., Wu, F., Qiang, X. K., Dong, J. B., Lu, F. Y. &Xu, X. W. 2012: Interplay between the Westerlies and Asian mon-soon recorded in Lake Qinghai sediments since 32 ka. ScientificReports 2, 619, doi:10.1038/srep00619.

Appleby, P. & Oldfield, F. 1983: The assessment of 210Pb data fromsites with varying sediment accumulation rates. Paleolimnology 15,29–35.

Arnaud, F., Magand, O., Chapron, E., Bertrand, S., Bo€es, X., Char-let, F. & M�eli�eres, M. 2006: Radionuclide dating (210Pb, 137Cs,241Am) of recent lake sediments in a highly active geodynamic set-ting (Lakes Puyehue and Icalma—Chilean Lake District). Scienceof the Total Environment 366, 837–850.

Bai, L. Y. & Chen, Z. Q. 2013: A study by urban construction landexpansion mode in Fuzhou City. Applied Mechanics and Materials316–317, 219–222.

Bertrand, S., Xavier, B., Julie, C., Charlet, F., Urrutia, R., Espinoza,C., Lepoint, G., Charlier, B. & Fagel, N. 2005: Temporal evolutionof sediment supply in Lago Puyehue (Southern Chile) during thelast 600 yr and its climatic significance. Quaternary Research 64,163–175.

Bonatti, E. & Gartner, S. 1973: Caribbean climate during PleistoceneIce Ages. Nature 244, 563–565.

Boyle, J. F. 2001: Inorganic geochemical methods in palaeolimnol-ogy. In Last, W. M. & Smol, J. P. (eds.): Tracking EnvironmentalChanges Using Lake Sediments, Vol. 2, Physical and GeochemicalMethods. Developments in Paleoenvironmental Research, 83–141.Kluwer Academic Publishers, Dordrecht.

Boyle, J. F., Mackay, A. W., Rose, N. L., Flower, R. J. & Appleby, P.G. 1998: Sediment heavy metal record in Lake Baikal. Journal ofPaleolimnology 20, 135–150.

Boyle, J. F., Rose, N. L., Bennion, H., Yang, H. & Appleby, P. G.1999: Environmental impacts in the Jianghan plain: evidence fromlake sediments.Water Air and Soil Pollution 122, 21–40.

Campbell, B. 1983: Application of environmental caesium-137 forthe determination of sedimentation rates in reservoirs and lakesand related catchment studies in developing countries. Radioiso-topes in Sediment Studies, IAEA 298, 7–30.

Chaudhuri, S. & Brookings, D. G. 1979: The Rb-Sr systematics inacid-leached clay minerals. Chemical Geology 24, 231–242.

Chen, C. T. & Selvaraj, K. 2008: Evaluation of elemental enrich-ments in surface sediments off southwestern Taiwan. Environmen-tal Geology 54, 1333–1346.

Chen, F. H., Chen, J. H., Holmes, J., Boomer, I., Austin, P., Gates, J.B., Wang, N. L., Brooks, S. J. & Zhang, J. W. 2010: Moisturechanges over the last millennium in arid central Asia: a review,synthesis and comparison with monsoon region. QuaternaryScience Reviews 29, 1055–1068.

Chen, J. A., Wan, G., Zhang, D. D., Zhang, F. & Huang, R. 2004:Environmental records of lacustrine sediments in different timescales: sediment grain size as an example. Science in China SeriesD: Earth Sciences 47, 954–960.

Chen, S. Y., Wang, S. M., Chen, Y. Y., Zhang, E. L., Chen, Y. J. &Shu, Y. X. 2009: Vertical distribution and chronological implica-tion of 210Pb and 137Cs in sediments of Dongping Lake, ShandongProvince. Quaternary Sciences 29, 981–987 (in Chinese, with Eng-lish abstract).

Chen, Y. Y., Chen, S. Y., Liu, J. Z., Yao, M., Sun, W. B. &Zhang, Q. 2013: Environmental evolution and hydrodynamicprocess of Dongping Lake in Shandong Province, China, overthe last 150 years. Environmental Earth Sciences 68, 69–75.

Clift, P. D., Hodges, K. V., Heslop, D., Hannigan, R., Long, H.V. & Calves, G. 2008: Correlation of Himalayan exhumationrates and Asian monsoon intensity. Nature Geoscience 1, 875–880.

Conroy, J. L., Overpeck, J. T., Cole, J. E., Shanahan, T. M. & Stei-nitz-Kannan, M. 2008: Holocene changes in eastern tropical Paci-fic climate inferred from a Gal�apagos lake sediment record.Quaternary Science Reviews 27, 1166–1180.


Davey, B. G., Russell, J. D. & Wilson, M. J. 1975: Iron oxide and clayminerals and their relation to colours of red and yellow podozolicsoils near Sydney, Australia. Geoderma 14, 125–138.

Dearing, J. A. 1999: Magnetic susceptibility. In Walden, J., Oldfield,F. & Smith, J. (eds.): Environmental Magnetism, a Practical Guide,35–62. Quaternary Research Association, London.

Edgington, D. N. & Robbins, J. A. 1976: Records of lead depositionin Lake Michigan sediments since 1800. Environmental Science andTechnology 10, 266–274.

Fan, Y. X., Chen, F. H., Wei, G. X., Madsen, D. B., Oviatt, C. G.,Zhao, H., Chun, X., Yang, L. P., Fan, T. L. & Li, G. Q. 2010:Potential water sources for Late Quaternary Megalake Jilantai-Hetao, China, inferred from mollusk shell 87Sr/86Sr ratios. Journalof Paleolimnology 43, 577–587.

Fedo, C. M., Eriksson, K. A. & Krogstad, E. J. 1996: Geochemistryof shales from the Archean (~3.0 Ga) Buhwa Greenstone Belt,Zimbabwe: implications for provenance and source-area weather-ing. Geochimica et Cosmochimica Acta 60, 1751–1763.

Fedo, C. M., Nesbitt, H. M. & Young, G. M. 1995: Unravelling theeffects of potassium metasomatism in sedimentary rocks and pale-osols, with implications for paleoweathering conditions and prove-nance. Geology 23, 921–924.

Fedo, C. M., Young, G. M., Nesbitt, H. W. & Hanchar, J. M. 1997:Potassic and sodic metasomatism in the southern province of theCanadian Shield: evidence from the Paleoproterozoic Serpent For-mation, Huronian Supergroup, Canada. Precambrian Research 84,17–36.

Gaillardet, J., Dupre, B., Louvat, P. & Allegre, C. J. 1999: Global sili-cate weathering and CO2 consumption rates deduced from thechemistry of large rivers. Chemical Geology 159, 3–30.

Giguet-Covex, C., Arnaud, F., Poulenard, J., Disnar, J. R., Del-hon, C., Francus, P., David, F., Enters, D., Rey, P. J. & Delan-noy, J. J. 2011: Changes in erosion patterns during theHolocene in a currently treeless subalpine catchment inferredfrom lake sediment geochemistry (Lake Anterne, 2063 m asl,NW French Alps): the role of climate and human activities.The Holocene 21, 651–665.

Grousset, F., Jouanneau, J., Castaing, P., Lavaux, G. & Latouche, C.1999: A 70 year record of contamination from industrial activityalong the Garonne River and its tributaries (SW France). Estuar-ine, Coastal and Shelf Science 48, 401–414.

Gupta, H., Chakrapani, G. J., Selvaraj, K. & Kao, S. J. 2011: The flu-vial geochemistry, contributions of silicate, carbonate and saline-alkaline components to chemical weathering flux and controllingparameters: Narmada River (Deccan Traps), India. Geochimica etCosmochimica Acta 75, 800–824.

Hung, C. W. 2013: A 300-year typhoon record in Taiwan and therelationship with solar activity. Terrestrial, Atmospheric and OceanSciences 24, 737–743.

Janitzky, P. 1987: Particle size analysis. In Singer, M. & Janitzky, P.(eds.): Field and Laboratory Procedures Used in Soil Chrono-Se-quence Study. USGS Bulletin 1648, 11–16.

Koide, M., Bruland, K. W. & Goldberg, E. D. 1973: 228Th/232Th and210Pb geochronologies in marine and lake sediments. Geochimicaet Cosmochimica Acta 37, 1171–1187.

Krishnamurti, G. S. R. & Satyanarayana, K. V. S. 1971: Influence ofchemical characteristics in the developing of soil colour. Geoderma5, 243–248.

Liu, J. B., Chen, J. H., Selvaraj, K., Xu, Q. H., Wang, Z. L. & Chen,F. H. 2014: Chemical weathering over the last 1200 years recordedin the sediments of Gonghai Lake, Lvliang Mountains, NorthChina: a high-resolution proxy of past climate. Boreas 43, 914–923.

Liu, K-b., Shen, C. & Louie, K-s. 2001: A 1000-year history oftyphoon landfalls in Guangdong, southern China, reconstructedfrom Chinese historical documentary records. Annals of the Associ-ation of American Geographers 91, 453–464.

Liu, M. & Fan, D. 2011: Geochemical records in the subaqueousYangtze River delta and their responses to human activities in thepast 60 years. Chinese Science Bulletin 56, 552–561.

Longmore, M., O’Leary, B. & Rose, C. 1983: Caesium-137 pro-files in the sediments of a partial-meromictic lake on Great

Sandy Island (Fraser Island), Queensland, Australia. Hydrobi-ologia 103, 21–27.

Lou, J. Y. & Chen, C. T. A. 1997: Paleoclimatological and paleoenvi-ronmental records since 4000 a B.P. in the sediments of alpine lakesin Taiwan. Science in China D Series: Earth Sciences 40, 424–431.

Ma, Y., Liu, C. & Huo, R. 2000: Strontium isotope systematic duringchemical weathering of granitoids: importance of relative mineralweathering rates. Journal of Goldschmidt Conference Abstracts 5,657.

Meyers, P. A. 2004: Application of elemental and isotopic sourceidentification of sedimentary organic matter. Chemical Geology114, 289–302.

Meyers, P. A. & Ishiwatari, R. 1993: Lacustrine organic geochemistry- an overview of indicators of organic matter sources and diagene-sis in lake sediments. Organic Geochemistry 20, 867–900.

Meyers, P. A., Bernasconi, S. M. & Forster, A. 2006: Origins andaccumulation of organic matter in expanded Albian to Santonianblack shale sequence on the Demerara Rise, South American mar-gin. Organic Geochemistry 37, 1816–1830.

Moy, M. C., Seltzer, G. O., Rodbell, D. T. & Anderson, D. M. 2002:Variability of El Nino/Southern Oscillation activity at millennialtimescales during the Holocene epoch. Nature 420, 162–165.

Nahm, W. H., Lee, G. H., Yang, D. Y., Kim, J. Y., Kashiwaya, K.,Yamamoto, M. & Sakaguchi, A. 2010: A 60-year record of rainfallfrom the sediments of Jinheung Pond, Jeongeup, Korea. Journal ofPaleolimnology 43, 489–498.

Nesbitt, H. W. & Young, G. M. 1982: Early Proterozoic climates andplate motions inferred from major element chemistry of lutites.Nature 299, 715–717.

Nesbitt, H. W., Markovics, G. & Price, R. C. 1980: Chemicalprocesses affecting alkalis and alkaline earths during continen-tal weathering. Geochimica et Cosmochimica Acta 44, 1659–1666.

Ollier, C. 1984:Weathering. 270 pp. Longman Press, London.Ortega-Guerrero, B., Sedow, S., Solleiro-Rebolledo, E. & Soler, A.2004: Magnetic mineralogy in Barranca Tlalpan exposure pale-osols, Tlaxcala, Mexico. Revista Mexicana de Ciencias Geologicas21, 120–132.

Peinerud, E. K. 2000: Interpretation of Si concentrations in lake sedi-ments: three case studies. Environmental Geology 40, 1637–1653.

Perkins, R. W. & Thomas, C. W. 1980: Worldwide fallout. InHanson,W. C. (ed.): Transuranic Elements in the Environment, 53–82. Tech-nical Information Center, U. S. Department of Energy, Washing-ton, D. C.

Petit, D. 1974: 210Pb et isotopes stables du plomb dans des sedimentslacustres. Earth and Planetary Science Letters 23, 199–205.

Qiang, M. R., Liu, Y. Y., Jin, Y. X., Song, L., Huang, X. T. & Chen,F. H. 2014: Holocene record of eolian activity from GenggahaiLake, northeastern Qinghai-Tibetan Plateau, China. GeophysicalResearch Letters 41, 589–595.

Robbins, J. A., Krezoski, J. R. & Mozley, S. 1977: Radioactivity insediments of the Great Lakes: post-depositional redistribution bydeposit-feeding organisms. Earth and Planetary Science Letters 36,325–333.

Schoenborn, W. A. & Fedo, C. M. 2011: Provenance and pale-oweathering reconstruction of the Neopreoterozoic Johnnie For-mation, southeastern California. Chemical Geology 285, 231–255.

Selvaraj, K. & Chen, C. T. A. 2006: Moderate chemical weatheringof subtropical Taiwan: constraints from solid–phase geochemistryof sediments and sedimentary rocks. Journal of Geology 114, 101–116.

Selvaraj, K., Chen, C. T. A. & Lou, J. Y. 2007: Holocene East Asianmonsoon variability: links to solar and tropical Pacific forcing.Geophysical Research Letters 34, L01703, doi:10.1029/2006GL028155.

Selvaraj, K., Chen, C. T. A., Lou, Y. J. & Kotlia, B. S. 2011: Holo-cene weak summer East Asian monsoon intervals in Taiwan andplausible mechanisms. Quaternary International 229, 57–66.

Selvaraj, K., Parthiban, G., Chen, C. T. A. & Lou, J. Y. 2010: An-thropogenic effects on sediment quality offshore southwestern Tai-wan: assessing the sediment core geochemical record. ContinentalShelf Research 30, 1200–1210.


Selvaraj, K., Wei, K. Y., Liu, K. K. & Kao, S. J. 2012: Late Holocenemonsoon climate of northeastern Taiwan inferred from elemental(C, N) and isotopic (d13C, d15N) data in lake sediments. Quater-nary Science Reviews 37, 48–60.

Sun, Q., Liu, D., Liu, T., Di, B. & Wu, F. 2011: Temporal and spatialdistribution of trace metals in sediments from the northern YellowSea coast, China: implications for regional anthropogenic pro-cesses. Environmental Earth Sciences 66, 697–705.

Wang, R., Wang, S. & Fraedrich, K. 1991: An approach of recon-struction of temperature on seasonal basis using historical docu-ments from China. International Journal of Climatology 11, 381–392.

Warrier, A. K. & Shankar, R. 2009: Geochemical evidence for theuse of magnetic susceptibility as a paleorainfall proxy in the trop-ics. Chemical Geology 265, 553–562.

Wieland, E., Santschi, P., H€ohener, P. & Sturm, M. 1993: Scavengingof Chernobyl 137Cs and natural 210Pb in Lake Sempach, Switzer-land. Geochimica et Cosmochimica Acta 57, 2959–2979.

Wu, Z. Y. 1980: Vegetation of China. 1385 pp. Science Press, Beijing(in Chinese).

Xiang, L. 1998: Dating sediments on several lakes inferred fromradionuclide profiles. Journal of Environmental Sciences 10,56–63.

Yan, H., Sun, L. G., Oppo, D. W., Wang, Y. H., Liu, Z. H., Xie, Z.Q., Liu, X. D. & Cheng, W. H. 2011: South China Sea hydrologicalchanges and Pacific Walker Circulation variations over the lastmillennium. Nature Communications 2, 1–5.

Yancheva, G., Nowaczyk, N. R., Mingram, J., Dulski, P., Schettler,G., Negendank, J. W., Liu, J. Q., Sigman, D. M., Peterson, L. C. &Haug, G. H. 2007: Influence of the intertropical convergence zoneon the East Asian monsoon. Nature 445, 74–77.

Yao, S. C., Li, S. J. & Zheng, H. C. 2008: 210Pb and 137Cs dating ofsediments from Zigetang Lake, Tibetan Plateau. Journal of Radio-analytical and Nuclear Chemistry 278, 55–58.

Yarincik, K. M. & Murray, R. W. 2000: Climatically sensitive eolianand hemipelagic deposition in the Cariaco Basin, Venezuela, overthe last 578 000 years: result from Al/Ti and K/Al. Paleoceanogra-phy 15, 210–228.

Zhang, X. B. 2005: Discussion on interpretations of 137Cs depth dis-tribution profiles of lake deposits. Journal of Mountain Science 23,294–299 (in Chinese with English abstract).

Zhang, Y., Peng, B., Chen, J. & Lu, J. 2005: Evaluation of sedimentaccumulation in Dianchi Lake, using 137Cs dating. Acta Geograph-ica Sinica 60, 71–78.

Zheng, J. Y., Ding, L. L., Hao, Z. X. & Ge, Q. S. 2012: Extreme coldwinters in southern China during AD 1650–2000. Boreas 41, 1–12.


lacustrine sedimentological and geochemical …...lacustrine sedimentological and geochemical...

Documents