a r t i c l e
i n f o
Article history:
Received 23 September 2014
Received in revised form 20 March 2015
Accepted 24 March 2015
Available online 31 March 2015
Keywords:
Typhoon
East Asian summer monsoon
ENSO
Precipitation
Vegetation history
Paleolimnology
a b s t r a c t
We used multi-decadal pollen and diatom records from sediment core TFL-1 from Tsuifong Lake to reconstruct the vegetation dynamics and hydroclimate in northeastern Taiwan during the past 3500 cal BP.
Coarse grained sediments in association with higher percentages of wetland pollen (Cyperaceae) and
upper conifer pollen (Tsuga and Pinus) in the lower part of the core indicate low lake levels and a relatively cold/dry climate between 3500 and 2030 cal BP, reecting a decline of the East Asian summer monsoon (EASM). Muddy sediments coupled with reduction of wetland pollen represent the rise of lake
levels, implying the re-strengthening of the EASM during the past 2000 years. Paleotemperature was
inferred from the variation of pollen origin from the upper and lower mountain forest, indicating the global temperature anomalies of the Medieval Warm Period (MWP) and the Little Ice Age (LIA). In comparison to the main climate forces in the North Pacic, we suggest that the long-term climatic trend in
Taiwan was controlled by variations in EASM intensity, while increased precipitation over the past
2000 years may also be linked to warmer sea surface temperature (SST) of the western Pacic Warm
Pool (WPWP) and increased El Nio-Southern Oscillation (ENSO) events, which increased typhoon intensity. Higher diatom-inferred pH during 29302030 cal BP and the LIA suggest strong hydrological disturbances, reecting more typhoons passing over Taiwan. The frequent typhoon events could be linked by
an abrupt shift of typhoon track, due to the reduction of the WPWP and expansion of the Northwestern
Pacic High, which move typhoons in a more westerly direction.
2015 Elsevier Ltd. All rights reserved.
1. Introduction
The East Asian summer monsoon (EASM) and typhoon systems
of the West Pacic have signicant effects on the regions hydroclimate. The summer monsoon precipitation is an important factor
for food production and maintaining ecosystem biodiversity. Its
anomaly often induces extreme ooding or droughts in the worlds
most populated area (Zhang et al., 2011). Typhoons are one of the
most extreme natural hazards and a major cause of forest ecosystem disturbances (Ito, 2010), catastrophic landslides (Tsou et al.,
2011), and organic carbon transportation at tectonically active
margins (Kao and Liu, 1997; Goldsmith et al., 2008) across the
regions. A clear understanding of historical EASM and typhoon
Corresponding author. Tel.: +886 75252000x5161.
E-mail address: yuanpin.chang@mail.nsysu.edu.tw (Y.-P. Chang).
http://dx.doi.org/10.1016/j.jseaes.2015.03.037
1367-9120/ 2015 Elsevier Ltd. All rights reserved.
448
L.-C. Wang et al. / Journal of Asian Earth Sciences 114 (2015) 447456
2. Study site
The TFL (24300 N, 121360 E, 1840 m above sea level (a.s.l.)) is a
subalpine, hydrological-closed, seepage lake in northeastern
Taiwan (Fig. 1). The lake is recharged mainly by precipitation, runoff and groundwater input from small catchments. The climatic
condition of the area is mainly affected by subtropical monsoon
system with a mean annual temperature of 12.5 C (ranges from
3.3 to 17.9 C), and a precipitation of 362 mm (ranges from 117
to 788 mm) based on records from the Taipinsan weather station
(24300 N, 121310 E, 1810 m a.s.l.). Bedrock components of this lake
Fig. 1. Important climate features of Taiwan and the location of Tsuifong Lake (TFL). Locations (black squares) are denoted as follow: (1) TFL; (2) marine sediment core
17,940; (3) Hani peat bog; (4) Lake Qinghai; (5) marine sediment core MD982141. EASM = East Asian summer monsoon.
L.-C. Wang et al. / Journal of Asian Earth Sciences 114 (2015) 447456
449
Fig. 2. This diagram denotes TFL-1 core sampling site (solid star) and vegetation distributions within dry and wet seasons in TFL basin. Two pie charts represent the water
coverages, rock fragments and the four vegetation zones during dry and wet periods, respectively. The vegetation maps are redrawn according to the TFL survey in April 2006
(dry season) and in August 2006 (wet season) by Mao (2006).
basin are mainly compose of slate, embedded with some metamorphic sandstone and conglomerate layers (Integrated Geological
Data Inquiry System, Central Geological Survey, MOEA, Taiwan).
The vegetation type on northeastern Taiwan is strongly inuenced by altitudinal changes. Trees growing in the upper mountain
needle-leaf forest (corresponding to upper mountain forest, UMF)
at height above 2200 m a.s.l. are dominated by Tsuga chinensis
var. formosana and Pinus taiwanensis. At altitudes between 1200
and 2200 m a.s.l., Chamaecyparis formosensis, T. chinensis var. formosana, Cyclobalanopsis morii, Alnus formosana, Trochodendron aralioides, and Hydrangea angustipetala are the main components of
montane mixed needle-broad-leaf forest (corresponding to mountain forest, MF). At altitudes below 1200 m a.s.l., Machilus thunbergii, Castanopsis cuspidata var. carlesii, Engelhardtia roxburgiana,
Lagerstroemia subcostata, Mallotus japonicus, Myrica rubra, Morus
australis, and Ardisia sieboldii are dominated in lower mountain
evergreen broad-leaved forest (corresponding to the lower mountain forest, LMF). Vegetation type in the study area is assigned to be
mountain mixed needle-broad-leaf forest (MF), although the primary forest is reclaimed by yellow cypress (C. formosensis and
Cryptomeria japonica) during AD 19261982.
The TFL basin is covered mainly by lake water, rock fragments,
moss, shrubs, Poaceae and aquatic plants during the dry seasons.
Otherwise, the rock fragments and moss areas are totally submerged with higher lake levels during the wet seasons (Mao,
2006) (Fig. 2). In the shrub zone, domination of Pourthiaea villosa
var. parvifolia is less impacted by uctuated seasonal water table.
Miscanthus sinensis and Haloragis micrantha are the main species
of the Poaceae zone, which is presents at lake borders during wet
seasons. Carex phacota, Festuca ovina, Polygonum thunbergii,
Polygonum praetermissum are characteristic species of the aquatic
zone, and are submerged more than 50% of the duration of wet seasons (Fig. 2). Moss and rock fragment areas are represented merely
at lake borders during the dry periods.
3. Materials and methods
3.1. Field coring and chronological framework
The sediment core, TFL-1, was collected from the central region
of the TFL at November 2004. A continuous 168-cm long lacustrine
record was retrieved by using a 2 m long stainless gravity corer.
Sediments were subsampled at 1 cm interval and freeze-dried
before further analysis. The lithology of TFL-1 is mainly composed
450
L.-C. Wang et al. / Journal of Asian Earth Sciences 114 (2015) 447456
examined under microscope as well as the required number of diatom specimen was counted (300 grains) (Friel et al., 2014).
Taxonomic identication and ecological interpretation are based
primarily on Freshwater Diatom Flora of Taiwan (Wu et al.,
2011) and collecting references (Charles, 1985; Krammer and
Lange-Bertalot, 1986; Wang et al., 2010). All diatom taxa were
subgrouped according to the inferring pH values (alkaliphilous,
circumneutral, acidophilous).
Based on a previous study examining diatom inferred pH
(DI-pH) in the TFL (Wang et al., 2013), we estimated the DI-pH of
core TFL-1 by using a transfer function derived from the combined
pH training set of European Diatom Database Initiative (EDDI)
(Juggins, 2001). The EDDI transfer function for pH reconstruction
is developed by using locally-weighted weighted averaging
method (LWWA). Validation of this approach had been undertaken
before proceeding. The web verication program of EDDI-set
(http://craticula.ncl.ac.uk/Eddi/jsp/verify.jsp) was executed to verify taxonomy against an EDDI transfer function. Our result prove
that the sum of fossil diatom at each level is more than 75%, suggesting the high reliability of reconstructed DI-pH of this study
(Birks and Birks, 1998).
3.3. Numerical methods for pollen data
Numerical analyses for pollen data were based on the terrestrial
pollen percentages by excluding the wetland pollen taxa. In order
to reduce the effect causing by rare taxa which has less inuence
on explaining varied vegetation, only pollen taxa values larger than
1% were included. The pollen-based biostratigraphy zones were
identied using CONISS cluster analysis within the computer programs Tilia and Tiliagraph (Grimm, 1993). Pollen and diatom diagrams were then constructed using the Tilia, Tiliagraph and C2
programs (version 1.5 for Windows) (Grimm, 1993; Juggins, 2005).
Multivariable analysis was conducted to examine the links
between important environmental variables with the pollen data.
Multivariable analysis is undertaken with the program CANOCO
version 4.5 (ter Braak and Smilauer, 2002). Pollen percentages
were square-root-transformed to stabilize variance and to smooth
the impact of dominate taxa (Jantz and Behling, 2012). Detrended
correspondence analysis (DCA) was carried out to dissolve the
underlying linearity of the pollen data in this study. The rst axis
of DCA with shorter gradient length of 0.96 standard deviations
was obtained, indicating that the principle component analysis
(PCA) was appropriate to explore the relationship between pollen
assemblages and environmental variables (Ammann et al., 2000;
ter Braak and Smilauer, 2002). Several proxies were picked, including percentages of wetland pollen and fern spores, published TOC,
TN, C/N ratio, d13Corg and, d15N (Selvaraj et al., 2012), as supplementary variables for PCA analysis to enrich the environmental
interpretation (Abdi and Williams, 2010). The values of all supplementary variables were denoted with log10(x + 1) and transformed
to standardize the variances (Cao et al., 2014).
4. Results
4.1. Chronology
Seven radiocarbon dates were used to establish the age-depth
model based on Bayesian analysis for the core TFL-1 (Fig. 3). The
radiocarbon date of plant remains in the surface sediment represented a negative age of 800 BP, which reected the inuence
of modern nuclear weapon testing. Thus, the chronology of surface
sediment was assumed as present time (54 cal BP) (Selvaraj et al.,
2012). The Bayesian age-depth analysis yielded a nearly linear
interpolation model (Fig. 3), covering the period from 3500 cal
Fig. 3. Age-depth model of core TFL-1 bases on seven AMS dates and surface
sample (54 cal BP) by using Bayesian software Bacon program for R (Blaauw and
Christen, 2011). Radiocarbon dates are calibrated with 2r errors of condence level.
Gray shades indicate models 95% probability intervals used in this study.
L.-C. Wang et al. / Journal of Asian Earth Sciences 114 (2015) 447456
451
Fig. 4. Lithology and percent pollen diagram with biostratigraphic zones determined with constrained cluster analysis (CONISS) for core TFL-1. Only taxa with percentage
larger than 1% are plotted. Percentages of all taxa are based on the sum of terrestrial pollen. Secondary y-axis presents calibrated age. UMF = upper mountain forest,
MF = mountain forest, LMF = lower mountain forest.
Fig. 5. Percentages of diatom taxa respect to pollen zones in TFL-1. Only the main taxa with >2% in percentage are plotted.
(27%) was seen more frequently among the terrestrial herbs. The
percentages of the wetland pollen Cyperaceae (913%) were lower
than in TFL3 (Fig. 4). The diatom concentration increased gradually
and reached a maximum at the end of the zone. The acidophilus
diatom Aulacoseira distance (6084%) was the most abundant species (Fig. 5). The average estimated DI-pH increased slightly to 5.57
(Fig. 6).
The pollen concentration in zone TFL5 (400 cm, 560 to 54 cal
BP) remained high. UMF pollen showed an increased trend with an
452
L.-C. Wang et al. / Journal of Asian Earth Sciences 114 (2015) 447456
Fig. 6. Fluctuations of PCA derived rst axis based on pollen data (pollen-PC1), diatom-inferred pH (DI-pH), total organic carbon (TOC), C/N ratio, d13C, and d15N with
respecting to pollen zones (TFL). Dashed lines represent the mean value of each proxy.
L.-C. Wang et al. / Journal of Asian Earth Sciences 114 (2015) 447456
453
Fig. 7. This diagram presents PCA results for 41 samples of pollen data. Red arrows on the left denote the 30 pollen taxa with a minimum value of >1% along with the rst axis
(PC1) and the second axis (PC2). Please note that only the names of major pollen taxa (>2% in at least one samples for each species) are shown. Sediment samples and
supplementary data (black arrows) along with two PCA axes are also presented. (For interpretation of the references to color in this gure legend, the reader is referred to the
web version of this article.)
454
L.-C. Wang et al. / Journal of Asian Earth Sciences 114 (2015) 447456
Fig. 8. Summary for the sedimentary proxies inferred paleotemperature (ratio of upper montane forest and lower montane forest, UMF/LMF, pollen-PC2), precipitation
(pollen-PC1), and typhoon activities (DI-pH) in TFL comparing to the proxies of major climate forces of East Asia (Fig. 1), including East Asian summer monsoon (EASM), sea
surface temperature (SST) of western Pacic Warm Pool (WPWP) (Stott et al., 2004), and El Nio-Southern Oscillation (ENSO) intensity (Moy et al., 2002). Records associated
with EASM intensity are sea surface salinity (SSS) of marine sediment core 17,940 (Wang et al., 1999), d13C of plant cellulose of Hani peat bog (Hong et al., 2005), and sediment
redness of Lake Qinghai (Ji et al., 2005). Thick red lines are smooth curves by using locally weighted scatterplot smoothing method in C2 program version 1.5 for Windows
(Juggins, 2005). Please note that the x-axis is reversed for UMF/LMF, 17,940 SSS, and Hani peat d13C. Right column shows the late Holocene climate conditions in associating
with periods of global temperature changes, including the MWP, and the LIA inferred by multi-proxies records of core TFL-1 (Chen et al., 2015; Yin et al., 2010). (For
interpretation of the references to color in this gure legend, the reader is referred to the web version of this article.)
South China Sea, indicating events of enhanced precipitation during the past 2000 years. During this period, the smaller d13C of peat
cellulose at Hani peat bog (Hong et al., 2005) and increased sediment redness at Lake Qinghai (Ji et al., 2005) also indicated the
stronger EASM activity or the moister climate in the monsoon
region. The reverse trend in EASM during the past 2000 years
might be linked to the movement of the position of ITCZ and
ENSO variability, which were forced by changes in austral summer
insolation and increases of greenhouse gas (Zhao et al., 2013).
Typhoon activities strongly affect regional and local hydrological patterns in East Asia. The paleo-precipitation record seems
positively correlated to the variations in the SST of WPWP and
ENSO intensity, except for the periods with frequent typhoon
activities, which were inferred by higher DI-pH (Fig. 8). SSTs of
WPWP and ENSO intensities control the intensity, frequency, and
tracks of typhoons originating from WPWP (Wang and Chan,
2002). Based on recent observational data, the warm SST anomalies
over the equatorial western and central Pacic tend to induce
typhoon enhancement, namely, the weakening of the subtropical
high, strengthening of the EASM, and enhanced low-level vorticity
(Tu et al., 2009). Therefore, we suggest that the increase in precipitation in northeastern Taiwan during the past 2000 years can also
be associated with enhanced typhoon activities in the western
Pacic due to the warmer SST of WPWP and coupled with intensied ENSO events.
Based on grain size records in lake sediments deposited in ood
events from the Ilan Plain in NE Taiwan, Chen et al. (2012) stated
that super typhoons chiey occurred in Taiwan during the La
Nia-like stage because typhoon tracks are highly inuenced by
the ENSO. Yin et al. (2010) also found that more typhoons affected
southern China during La Nia years even though the amount of
heavy rain days decreased. According to recent meteorological
records, the westward shift in typhoon tracks is relative to the
shrinking of WPWP and expansion of the Northwestern Pacic
High, which pushes typhoons westerly toward southern China
and Taiwan (Ho et al., 2004). Therefore, the frequent typhoons
and abrupt precipitation anomalies in NE Taiwan during
Acknowledgments
The authors appreciate the nancial support provided by a
Grant from the National Science Council (NSC 100-2611-M-110011-, 101-2611-M-110-003-, 102-2917-I-564-058), Ministry of
Science and Technology (MOST 103-2811-M-194-010), the Open
L.-C. Wang et al. / Journal of Asian Earth Sciences 114 (2015) 447456
455
456
L.-C. Wang et al. / Journal of Asian Earth Sciences 114 (2015) 447456
Wang, B., Chan, J.C.L., 2002. How strong ENSO events affect tropical storm activity
over the western North Pacic. J. Clim. 15, 16431658. http://dx.doi.org/
10.1175/1520-0442(2002)015<1643:HSEEAT>2.0.CO;2.
Wang, L.-C., Behling, H., Chen, Y.-M., Huang, M.-S., Chen, A.C.-T., Lou, J.-Y., Chang, Y.-P.,
Li, H.-C., 2014a. Holocene monsoonal climate changes tracked by multiproxy
approach from a lacustrine sediment core of the subalpine Retreat Lake in Taiwan.
Quatern. Int. 333, 6976. http://dx.doi.org/10.1016/j.quaint.2014.02.014.
Wang, L.-C., Behling, H., Lee, T.-Q., Li, H.-C., Huh, C.-A., Shiau, L.-J., Chang, Y.-P.,
2014b. Late Holocene environmental reconstructions and their implications on
ood events, typhoon, and agricultural activities in NE Taiwan. Clim. Past 10,
18571869. http://dx.doi.org/10.5194/cp-10-1857-2014.
Wang, L.-C., Behling, H., Lee, T.-Q., Li, H.-C., Huh, C.-A., Shiau, L.-J., Chen, S.-H., Wu, J.T., 2013. Increased precipitation during the little ice age in northern Taiwan
inferred from diatoms and geochemistry in a sediment core from a subalpine
lake. J. Paleolimnol. 49, 619631. http://dx.doi.org/10.1007/s10933-013-9679-9.
Wang, L.-C., Lee, T.-Q., Chen, S.-H., Wu, J.-T., 2010. Diatoms in Liyu Lake, Eastern
Taiwan. Taiwania 55, 228242.
Wang, L.-C., Wu, J.-T., Lee, T.-Q., Lee, P.-F., Chen, S.-H., 2011. Climate changes
inferred from integrated multi-site pollen data in northern Taiwan. J. Asian
Earth Sci. 40, 11641170. http://dx.doi.org/10.1016/j.jseaes.2010.06.003.
Wang, L., Sarnthein, M., Erlenkeuser, H., Grootes, P.M., Grimalt, J.O., Pelejero, C.,
Linck, G., 1999. Holocene variations in Asian monsoon moisture: a bidecadal
sediment record from the South China Sea. Geophys. Res. Lett. 26, 28892892.
Wenske, D., Bse, M., Frechen, M., Lthgens, C., 2011. Late Holocene mobilisation of
loess-like sediments in Hohuan Shan, high mountains of Taiwan. Quatern. Int.
234, 174181. http://dx.doi.org/10.1016/j.quaint.2009.10.034.
Wu, J.-T., Babu, B., Chou, C.-L., Saraswathi, S.J., 2011. Freshwater Diatom Flora of
Taiwan. Biodiversity Research Center, Academia Sinica, Taipei, Taiwan, p. 747.
Yin, Y., Gemmer, M., Luo, Y., Wang, Y., 2010. Tropical cyclones and heavy rainfall in
Fujian Province, China. Quatern. Int. 226, 122128. http://dx.doi.org/10.1016/
j.quaint.2010.03.015.
Yuan, D., Cheng, H., Edwards, R.L., Dykoski, C.A., Kelly, M.J., Zhang, M., Qing, J., Lin, Y.,
Wang, Y., Wu, J., Dorale, J.A., An, Z., Cai, Y., 2004. Timing, duration, and
transitions of the last interglacial Asian monsoon. Science 304, 575578. http://
dx.doi.org/10.1126/science.1091220.
Zhang, J., Chen, F., Holmes, J.A., Li, H., Guo, X., Wang, J., Li, S., L, Y., Zhao, Y., Qiang,
M., 2011. Holocene monsoon climate documented by oxygen and carbon
isotopes from lake sediments and peat bogs in China: a review and synthesis.
Quatern. Sci. Rev. 30, 19731987. http://dx.doi.org/10.1016/j.quascirev.2011.04.
023.
Zhao, C., Chang, Y.-P., Chen, M.-T., Liu, Z., 2013. Possible reverse trend in Asian
summer monsoon strength during the late Holocene. J. Asian Earth Sci. 69, 102
112. http://dx.doi.org/10.1016/j.jseaes.2012.09.028.
Zhao, Y., Yu, Z., Chen, F., Ito, E., Zhao, C., 2007. Holocene vegetation and climate
history at Hurleg Lake in the Qaidam Basin, northwest China. Rev. Palaeobot.
Palynol. 145, 275288.