[The Journal of Geology, 2006, volume 114, p. 101116] 2006 by The University of Chicago. All rights reserved. 0022-1376/2006/11401-0006$15.

00
101
Moderate Chemical Weathering of Subtropical Taiwan:
Constraints from Solid-Phase Geochemistry
of Sediments and Sedimentary Rocks
Kandasamy Selvaraj and Chen-Tung Arthur Chen
Institute of Marine Geology and Chemistry, National Sun Yat-Sen University, Kaohsiung 804, Taiwan
(e-mail: kselva8@yahoo.com)
AB S T R ACT
The well-known earthquake-and-storm-triggered extremely high physical weathering rate in Taiwan is consistent
with present geochemical studies of sediments from different subenvironments (offshore, coastal, river, and lake) and
sedimentary rocks of different geological ages, indicating a moderate chemical weathering condition. Major and trace
element concentrations normalized to the average upper crust of Yangtze Craton show that the sediments and the
average composition of sedimentary rocks of Taiwan are depleted in Ca, Mg, Na, and Sr, enriched in Rb and Zr, and
unchanged with respect to K, indicating their moderately altered nature. The mean chemical index of alteration (CIA;
7175) and plagioclase index of alteration (PIA; 8186) values of coastal and offshore sediments reveal the sediments
derivation from sedimentary rocks by moderate silicate chemical weathering processes. The mean CIA value (62) of
sedimentary rocks of Taiwan is similar to that for Chinese sediment (61), further conrming the above inference.
A-CN-K, (A-K)-C-N, and A-CNK-FM plots ( ; ; ; ; ; ) also

ApAl O CpCaO NpNa O K pK O F pFeO MpMgO

2 3 2 2 T
conrm that the sediments and sedimentary rocks in Taiwan have undergone moderate silicate weathering, an
interpretation consistent with CIA and PIA values. The plots also indicate the presence of illite, chlorite, and a
subordinate amount of unaltered feldspars in sediments and sedimentary rocks, which are indicative of the physically
weathered and/or moderately chemically altered nature of sediments. The dominance of illite, chlorite, and unaltered
feldspars as inferred from geochemical data suggests that the immature nature of sediments and sedimentary rocks
is probably a result of low residence times in the source region or river basin and quick removal of materials from
the soil prole by steep, mountainous rivers (physical weathering dominates). Elemental ratios such as Rb/Sr, K/Rb,
molar K/Na, and Al/Na are close to crustal values. Average shale and river particulates such as those from the Yellow
River also indicate moderate chemical weathering conditions for sediments and sedimentary rocks, except for high
alpine lake sediments, where the prevailing extreme chemical weathering condition over erosion is clearly differ-
entiated by higher CIA (8084) and PIA (9296) values and by their positions on triangular plots. These inferences
have also been illustratively corroborated by scatter plots of data such as Rb/Sr versus molar K/Na, and Al/Na versus
CIA. Additional evidence from published sources noted here also favors moderate chemical weathering conditions
for Taiwan. Geochemical variation of offshore, coastal, and river sediments is mainly controlled by nonsteady state
weathering dominated by erosion. Steady state weathering, however, seems to produce highly weathered sediments
in the alpine region of Taiwan.
Online enhancement: appendix table.
Introduction
Disproportionately high physical and chemical de-
nudation rates occur in the oceanic island of Tai-
wan under the inuence of copious orographic rain-
fall associated with the Asian monsoon, frequent
typhoon activities, and related storm-triggered
landslides. The average erosion rate estimated for
Manuscript received October 4, 2004; accepted July 12, 2005.
the entire island is 5 mm/yr (WRPC 1973). The
average physical denudation rate for Taiwan (1365
mg/cm
2
/yr) is probably the highest in the world, as
is the chemical denudation rate (50 mg/cm
2
/yr),
which is 27% of the physical denudation rate (Li
1976). Recent estimates of maximum erosion rates
(36 mm/yr) and the 30-yr average maximum sed-
iment erosion rate (3.9 mm/yr; Dadson et al. 2003)
102 K . S E LVA R A J A N D C . - T. A . C H E N
have demonstrated that across Taiwan, cumulative
seismic moment release (earthquakes of magnitude
greater than ) correlated linearly with de- M p5.0
w
cadal erosion rates over a sixfold range during a
period between 1900 and 1998. Dadson et al. (2003)
concluded that storm runoff is a rst-order control
on erosion rates in Taiwan, and the modern erosion
rates are not controlled by relief and precipitation.
Other recent erosion rates estimated fromriver wa-
ter and sediment discharges of the Eastern Central
Range of Taiwan show a wide range of values (2.2
8.3 mm/yr) resulting from the inuence of variable
intensity of storms rather than from lithology, tec-
tonic environment, or climate (Fuller et al. 2003).
Li (1976), however, inferred that the great change
in the physical denudation rate is related to relief
and/or bedrock geology. Indeed, Milliman and Sy-
vitski (1992) projected that mountainous rivers
draining in South Asia and Oceania have much
greater sediment yields (two- to threefold) than
other mountainous rivers of the world, owing to
the inuence of human activity, climate, and ge-
ology. Chen et al. (2004) listed 13 rivers in Taiwan
among the top 20 worldwide in terms of sediment
yield. Previous studies have documented highly
variable physical and chemical denudation rates for
this island, but no clear consistency between them
has been established because relationships among
climate, physical erosion, and chemical weathering
remain poorly quantied since long-term weath-
ering rates are difcult to measure (Riebe et al.
2001). The purpose of this article is to determine
if the very high, widely variable physical and chem-
ical denudation rates reported from this oceanic
island show any covariance with the intensity
of chemical weathering, especially silicate
weathering.
Geochemical studies have contributed apprecia-
bly to the understanding of the growth of the con-
tinents through time (Taylor and McLennan 1985).
Numerous factors including source area composi-
tion, source area weathering conditions, hydraulic
sorting, adsorption, diagenesis, and metamorphism
affect the composition of siliclastic sediments and
sedimentary rocks over a wide scale (Fedo et al.
1996). Silicate weathering strongly affects the
major-element geochemistry and mineralogy of sil-
iclastic sediments (e.g., Nesbitt and Young 1982;
Johnsson et al. 1988; McLennan 1993), where larger
cations (Al
2
O
3
, Ba, Rb) remain xed in the weath-
ering prole preferentially over smaller cations (Ca,
Na, Sr), which are selectively leached (Nesbitt et
al. 1980). These chemical signatures are ultimately
transferred to the sedimentary record (e.g., Nesbitt
and Young 1982; Wronkiewicz and Condie 1987),
thus providing a useful tool for monitoring source
area weathering conditions. Silicate weathering in-
dexes such as chemical index of alteration (CIA),
plagioclase index of alteration (PIA), and chemical
index of weathering (CIW) are therefore widely used
to interpret the weathering history of modern and
ancient sediments (Harnois 1988; Fedo et al. 1996;
Colin et al. 1999; Tripathi and Rajamani 1999). For
example, high CIA values reect removal of labile
cations (Ca
2
, Na

, K

) relative to stable residual

constituents (Al
3
, Ti
4
) during weathering (Nesbitt
and Young 1982). Conversely, low CIA values in-
dicate the near absence of chemical alteration and
might reect cool and arid conditions (Fedo et al.
1995).
Chemical weathering rate (based on dissolved
loads) includes dissolution of elements from both
carbonate and aluminosilicate source rocks as well
as atmospheric and modern anthropogenic inputs
(Gaillardet et al. 1999; Sarin 2001; Han and Liu
2004). Though silicate weathering is Earths long-
term sink for atmospheric CO
2
(Berner et al. 1983),
in most natural environments, ionic contribution
from carbonate rocks generally dominates the dis-
solved phase (Han and Liu 2004), with silicate con-
tributing smaller amounts of dissolved solids
(!15%; Sarin 2001). This is especially true where
Cenozoic rocks dominate, such as in Taiwan.
Hence, the chemical weathering rate, especially sil-
icate weathering, of this island may be lower than
previously projected (e.g., of total 97.09%2.41%
chemical weathering; Lai 2003). Sediments pro-
duced by landslides and typhoon activities from
rough, mountainous terrain and rapidly transported
to the continental margin by uvial systems of Tai-
wan would be expected to display minimal chem-
ical weathering. That is, silt and mud would be
expected to contain a comparatively lowproportion
of aluminous clay minerals and a commensurately
higher proportion of primary minerals (feldspars).
In order to evaluate the chemical weathering in-
tensity and provide accurate conditions of silicate
weathering of Taiwan rocks, a wide representative
range of geochemical data of offshore, coastal, river,
and lake sediments of Taiwan (see table A1, avail-
able in the online edition or from the Journal of
Geology ofce), along with published geochemical
results of sediments from Taiwan Strait (Chao and
Chen 2003) and sedimentary rocks (Lan et al. 2002)
of different geological ages of Taiwan, has been sub-
jected to calculation of weathering indexes (CIA,
PIA, and CIW). A variety of triangular and scatter
plots have also been constructed from the geo-
chemical data to accomplish our aim. To facilitate
the interpretation of sediments and sedimentary
Journal of Geology S U B T R O P I C A L TA I WA N W E AT H E R I N G 103
Figure 1. Map of Taiwan showing the study area
(dashed line). Representative surface and core sediment
samples collected from different subenvironments in
southern Taiwanoffshore (open circle), coastal (lled
circle), river (star), and lake (open triangle)have been
used for this study. The open square represents Kao-
hsiung City.
rocks, elemental values of representative river par-
ticulates such as Kaoping (Lai 2003), Yellow and
Yangtze (both from Li et al. 1984), and Amazon
(Martin and Meybeck 1979), as well as loess (Li et
al. 1984) and other important reference composi-
tions such as upper continental crust (UCC), post-
Archean Australian shale (both from Taylor and
McLennan 1985), average shale (Turekian and
Wedepohl 1961), and North American shale com-
posite (Gromet et al. 1984), have also been included
for comparison.
Methodology
A total of 106 sediment samples were collected
from different subenvironments (12 offshore sur-
face sediments, eight coastal surface sediments,
and core sediment consisting of 16 subsamples, all
from off southwestern Taiwan; one bed sediment
sample fromKaoping River; and core sediment con-
sisting of 69 subsamples from high alpine, anoxic
Great Ghost Lake) in southern Taiwan (g. 1), with
the help of appropriate sampling devices and tech-
niques. Before chemical analysis, the samples were
freeze-dried and homogenized, and the bulk sedi-
ment of each sample was nely ground (!200 mesh)
in an agate mortar. Major (Si, Ti, Al, Fe, Mn, Mg,
Ca, Na, K, and P) and trace (Ba, Rb, Sr, and Zr)
elements were determined using an x-ray uo-
rescence (XRF) spectrometer (Rigaku RIX 2000)
equipped with an Rh tube at the Institute of Marine
Geology and Chemistry, National Sun Yat-SenUni-
versity, Taiwan. Details of the XRF method are de-
scribed by Chen et al. (2001). The accuracy of the
analytical method was established using six inter-
nationally recognized standard reference materials:
MAG-1, BCSS-1, PACS-1, MESS-1, NIES-2, and
GBW07314. Based on these standards, the accuracy
and precision of the analysis were within 1% for
major elements such as Al, Ca, Fe, K, Mg, Na, Si,
and Ti and within 5% for Mn, P, Rb, Sr, and Zr.
The precision and accuracy for Ba was within
10%.
A LECO CHN-932 elemental analyzer was em-
ployed to determine carbon content at 950C. After
the samples were repeatedly rinsed with 1 N HCl
to remove inorganic carbon (IC), total organic car-
bon was determined. The amount of IC was esti-
mated by the difference between measured total
carbon and organic carbon. The detection limit of
IC is 0.01%. In the samples that were not measured
for inorganic carbon, the mean value of each sam-
ples particular subenvironment was used for the
calculation of CaO in only the silicate fraction
( ). The CIAand PIAvalues of sedimentary and

CaO
metasedimentary rocks as well as reference com-
positions such as UCC, average shale, post-Archean
Australian shale, and North American shale com-
posite were calculated without correcting CaO for
carbonates and phosphates. In most of the sedi-
mentary rocks of Taiwan, the CaO value is less
than 2% (see table A1), with the exception of ar-
gillite (2.53%); the calculated values are, therefore,
slightly lower than actual CIA and PIAvalues. Loss
on ignition (LOI) was calculated as a percentage of
dry weight after the samples were ignited at 550C
for 1 h (Dean 1974).
Table 1. Summary Statistics of Geochemical Compositions of Sediments and Sedimentary Rocks of Taiwan
Major oxides (wt%)
Total
Trace elements (ppm)
SiO
2
TiO
2
Al
2
O
3
Fe
2
O
3
a
MnO MgO CaO Na
2
O K
2
O P
2
O
5
LOI
b
Ba Rb Sr Zr
Offshore sediments (np12):
Minimum 54.33 .69 14.60 4.72 .05 1.74 .84 .97 2.41 .11 4.40 96.47 381 84 99 178
Maximum 64.38 .88 17.61 8.58 .36 2.75 4.62 1.43 3.25 .21 9.52 100.01 490 150 187 1432
Mean 60.43 .75 15.65 5.79 .12 2.05 2.47 1.16 2.76 .14 6.91 98.23 425 117 137 342
SD 2.84 .06 .94 1.07 .11 .33 1.18 .14 .26 .03 1.43 1.15 37.8 19.0 22.6 349.2
Coastal sediments (np8):
Minimum 51.55 .52 13.70 4.26 .05 1.36 1.52 .90 2.17 .06 4.23 98.29 356 85 113 164
Maximum 72.73 .81 16.57 6.09 .05 3.47 3.58 2.45 3.52 .14 7.19 99.08 465 111 143 287
Mean 63.69 .62 14.48 5.10 .05 2.07 2.25 1.39 2.73 .10 5.14 98.72 396 97 124 207
SD 7.32 .09 .88 .68 .00 .84 .73 .64 .49 .03 1.12 .30 34.5 7.9 9.6 46.7
Offshore core sediment (np16):
Minimum 66.32 .81 14.56 5.76 .05 1.64 1.12 .99 2.72 .09 2.86 98.80 436 41 121 178
Maximum 68.63 .84 15.34 6.19 .07 1.77 1.35 1.27 2.86 .10 3.89 100.80 472 48 132 198
Mean 67.30 .82 15.01 5.93 .06 1.73 1.26 1.17 2.77 .09 3.37 99.53 459 45 126 189
SD .65 .01 .21 .10 .00 .03 .06 .07 .04 .00 .31 .49 10.4 1.9 3.5 4.7
Core sediment, Great Ghost Lake (np69):
Minimum 50.12 .64 15.38 4.56 .02 .74 .06 .28 2.41 .21 9.60 100.22 465 101 52 161
Maximum 58.74 .82 23.08 5.55 .03 1.24 .16 .59 3.31 .43 23.80 100.54 686 228 70 254
Mean 53.78 .76 17.04 5.05 .03 .87 .12 .39 2.64 .33 19.38 100.38 537 167 63 212
SD 1.71 .03 1.44 .19 .00 .09 .02 .07 .19 .04 3.05 .07 45.3 26.9 3.6 20.5
Bed sediment, Kaoping River (np1):
72.36 .54 12.39 4.75 .04 1.29 1.66 1.31 2.07 .09 2.40 98.91 492 93 129 209
Core tops, Taiwan Strait (np3):
c
Minimum 73.12 .29 5.55 2.97 .02 .78 1.45 .65 1.33 .06 1.39 100.27 335 53 88 47
Maximum 86.60 .41 12.29 5.51 .06 1.62 3.86 1.37 2.37 .11 2.48 100.79 413 97 188 57
Mean 78.97 .36 8.72 4.41 .04 1.26 2.28 .96 1.82 .09 2.10 100.56 371 70 139 52
SD 6.91 .06 3.39 1.31 .02 .43 1.37 .37 .52 .03 .62 .27 39.3 23.6 50.1 5.0
Sedimentary rocks of Taiwan (np14):
d
Minimum 58.75 .10 6.65 .71 .01 .05 .22 .00 .32 .02 1.25 99.59 110 7 22 48
Maximum 86.50 .90 17.97 7.25 .17 3.00 2.53 3.08 3.84 .13 7.50 101.11 1034 169 291 593
Mean 70.65 .59 13.37 4.49 .09 1.61 1.27 1.70 2.74 .08 3.67 100.26 507 109 134 212
SD 9.81 .23 4.12 2.27 .05 .94 .62 .85 1.13 .03 1.90 .41 229 51 78 136.3
a
Total Fe expressed as Fe
2
O
3
.
b
on ignition. LOI ploss
c
Analysis from Chao and Chen 2003.
d
Analysis from Lan et al. 2002.
Journal of Geology S U B T R O P I C A L TA I WA N W E AT H E R I N G 105
Figure 2. Spider plot showing comparison between av-
erage compositions of sediments from different suben-
vironments (offshore, coastal, river, and lake) and the av-
erage composition of sedimentary rocks of Taiwan, both
normalized to the average upper crustal (UC) values of
Yangtze Craton (YC; Gao et al. 1998). Note the similarity
in the behavior of most of the elements between sedi-
ments and the average compositionof sedimentaryrocks.
The similarity suggests that most of the recent sediments
are physically eroded and/or moderately chemically al-
tered products of sedimentary rocks. The depletion of Ca,
Mg, Na, Ba, and Sr relative to the upper crust of Yangtze
Craton could be attributed to their mobility during
weathering.
Sediment Composition
The average composition of sedimentary rocks of
Taiwan (table 1) has been calculated based on the
data of Lan et al. (2002). The calculation involves
14 samples consisting of four widely varying sed-
imentary and metasedimentary rocks, such as phyl-
lite (one sample), sandstones (three), argillites (two),
and metapelites (eight), with the assumption that
these four rock types will represent approximately
average values for sedimentary rocks in this small
area (0.024%) of the earths surface. Based on Sr-
Nd-O isotopic geochemistry of Taiwan granitoids
and metapelites, Lan et al. (1995) suggested that
cover sediments of Taiwan received recycled con-
tinental crustal material from South China and the
basement rocks of Taiwan. It has recently been
shown, based on the Nd isotopic composition of
sedimentary and metasedimentary rocks of Tai-
wan, that these rocks are recycled materials of the
UCC of the South China region (Lan et al. 2002).
South China consists of the Yangtze Craton in the
west and the Cathaysian fold in the east (Jahn et
al. 1990; Lan et al. 1995). Because we are not aware
of published results for the UCC of South China,
in this study, sediment and sedimentary rock com-
positions of Taiwan were compared with the upper
crust of Yangtze Craton (Gao et al. 1998). The av-
erage major (Si, Al, Ti, Fe, Ca, Mg, Na, and K) and
trace (Ba, Rb, Sr, and Zr) element compositions of
sediments from different subenvironments (off-
shore, coastal, river, and lake) and the average com-
position of sedimentary rocks (table 1) normalized
with upper crustal values of Yangtze Craton (Gao
et al. 1998) are shown in gure 2. The sediments
and sedimentary rocks have remarkably similar
patterns. The similarity suggests that most of the
sediments are dominantly physically eroded and/
or moderately chemically modied. All the sedi-
ments and average composition of sedimentary
rocks are invariably depleted in Ca, Mg, Na, Ba,
and Sr, enriched in Rb and Zr, and unchanged with
respect to K. The degree of elemental variation is
more prominent in lake sediments because of their
highly altered nature. The slight depletion of Ti,
Fe, and Mg shown by the average composition of
sedimentary rocks of Taiwan is similar to UCCnor-
malized values for these elements in the Huanghe
sediments (Yang et al. 2004). Like sedimentary
rocks, coastal and river sediments are depleted in
Al, Ti, and Fe, whereas slight richness of these el-
ements is evident in offshore and lake sediments.
Silicate Weathering: Geochemical Indicators
and Triangular Plots
Geochemical processes such as weathering and soil
formation are dominated by alteration of feldspars
(and volcanic glass), which accounts for 70% of the
upper crust if the relatively inert quartz is dis-
counted (Nesbitt and Young 1982, 1984). Feldspars
are by far the most abundant labile minerals, and
consequently, the key process during silicate
weathering of the earths upper crust is the deg-
radation of feldspars by aggressive soil solutions so
that the proportion of alumina to alkalis typically
increases in the weathered product. A good mea-
sure of the degree of weathering can be obtained
by calculation of the CIA using molecular propor-
tions:

CIA p100(Al O /Al O CaO Na O
2 3 2 3 2
, where is the amount of CaO incorpo-

K O) CaO
2
rated in the silicate fraction of the rock. The re-
106 K . S E LVA R A J A N D C . - T. A . C H E N
sultant CIA gives a measure of the proportion of
secondary aluminous clay minerals to primary sil-
icate minerals such as feldspars (Nesbitt and Young
1982; Young et al. 1998).
The calculated CIA values for sediments and sed-
imentary rocks including river particulates, loess,
and important reference compositions presented in
table 2 demonstrate that the surface sediments off
southwestern Taiwan (water depth 105537 m)
have moderate CIA values of 6677, with an av-
erage of 74. Values for the near coastal sediments
(water depth !100 m) from the same region are
slightly lower but comparable, with CIA ranging
from 59 to 78 and a mean of 73. Sixteen subsamples
of core sediment off southwestern Taiwan (water
depth 294 m) exhibit a narrow range of CIA values
(7376) with a mean of 75. Three surface sediments
(top 015 cm) of three cores from Taiwan Strait, off
central Taiwan (data from Chao and Chen 2003),
also show comparable CIA values (6972). The CIA
values of surface and core sediments have means
ranging from 71 to 75 and fall within the range of
values for average shale (7075), indicating that the
sediments are possibly the product of sedimentary
and metasedimentary rocks that have undergone in-
termediate chemical weathering. To deduce the sil-
icate weathering trends, Nesbitt and Young (1982)
and Nesbitt et al. (1996) used A-CN-K (Al
2
O
3
-
-K
2
O) and A-CNK-FM (Al
2
O
3
-

CaO Na O
2
- ) ternary plots.

CaO Na OK O FeO MgO

2 2 T
The data of sediments and sedimentary rocks of Tai-
wan plotted in Al
2
O
3
- -K
2
O composi-

CaO Na O
2
tional space (g. 3) fall on a trend parallel to the A-
join, and the trend approaches the Al
2
O
3
-K
2
O CN
join at about CIA80. The CIAvalues of many weath-
ering proles and sediments are linear, subparallel
to the A-CN join in the A-CN-K plot (e.g., Nesbitt
and Young 1984; Fedo et al. 1995; Selvaraj et al.
2004). All the samples that display moderate sili-
cate weathering are thus plotted between CIA 60
and 80 (scale shown on the right side of the diagram
for comparison), except for the sediments fromhigh
alpine Great Ghost Lake of southern Taiwan. The
lake sediments that show a narrow range of higher
CIA values between 80 and 84 ( ; table mean p82
2), therefore, fall very near to apex A (g. 3) as ad-
vanced weathering leads to an aluminum-richcom-
position (e.g., Nesbitt and Young 1984; Nath et al.
2000). This indicates extreme chemical weathering
in the high altitude region of this island. The lake
is one of the wettest places in Taiwan (annual
mm; Lou et al. 1997); the presence rainfall p4200
of thick vegetation on the gently sloping shore (24
27 inclination) surrounding the lake acts as an
ideal site for soil development, and the large
amount of humus material (mean total organic
; ) available in this region carbon p11.92% n p69
is probably responsible for the relatively low pH in
soil solutions of lake catchments caused by pro-
duction of organic acids. It has been shown by Colin
et al. (1999) that increasing vegetation cover in the
ood plains of the Irrawaddy River during the sum-
mer monsoon reinforcement favors soil develop-
ment. This inference indicates that acidic byprod-
ucts of vegetation promote silicate weathering. The
lowest CIAvalue of loess (46; table 2) indicates low-
input acids to soils as a result of a cool, dry climate.
Most of our sedimentary rocks plot between CIA
60 and 70 in gure 3, close to values for the ref-
erence compositions such as the upper crusts of
Yangtze Craton (56) and Central East China (54).
The weathering trend (arrow 2 in g. 3) connecting
recent sediments and sedimentary rocks falls in a
single line, suggesting that the sediments are the
products of moderately weathered sedimentary
rocks. The weathering trend of sediments and sed-
imentary rocks shows slight enrichment of K
2
O
compared with particulate matters of the major riv-
ers Yellow, Yangtze, Brahmaputra, and Amazon and
with upper crustal compositions, indicating the
presence of considerable unaltered K-feldspar in the
samples. In gure 3, arrow 2 intersects the feldspar
join (Pl-Ks) at point x and may represent the original
parent rock composition, which is also slightly rich
in K-feldspar when compared with upper crustal
compositions. Indeed, values for the bed sediment
of the Kaoping River and its suspended particulate
matter fall very close to those for river particulates
of the Yellow River and average world values,
which is evidence for the moderate silicate weath-
ering in the river basin.
The CIA values of sediments from different sub-
environments are higher than the calculated CIA
values of sedimentary rocks (argillite, metapelite,
phyllite, and sandstone) in Taiwan (Lan et al. 2002),
which all have mean CIA in the narrow range be-
tween 61 and 64 (table 2). These values are strikingly
similar to the mean CIA (61) of Chinese sediment
(Yang et al. 2004), suggesting that sedimentary rocks
of Taiwan have experienced silicate weathering of
intensity similar to that of Chinese sediments. This
further substantiates that the sediments and sedi-
mentary rocks are the product of rocks that expe-
rienced moderate chemical weathering in the source
regions. The calculated CIAvalues of sediments and
sedimentary rocks have been classied according to
the scheme suggested by Fedo et al. (1995) and
shown in table 2. Nearly all the sediments and sed-
imentary rocks, including metapelites, invariably
show intermediate intensity of chemical weather-
Table 2. Chemical Index of Alteration (CIA), Plagioclase Index of Alteration (PIA), and the Mean Rb/Sr and K/Rb Ratios of Sediments, Sedimentary Rocks,
Riverine Particulates of Taiwan and China, and Important Reference Compositions
Location
CIA PIA
Mean
Rb/Sr
Mean
K/Rb
Weathering
intensity
based on
mean CIA Data source Range Mean Range Mean
Offshore sediments, southwestern Taiwan
a
6677 74 7190 84 .89 197 Intermediate This study
Coastal sediments, southwestern Taiwan
a
5978 73 6490 83 .78 235 Intermediate This study
Core sediments, southwestern Taiwan
a
7376 75 8388 86 .36 508 Intermediate This study
Lake core sediments, Taiwan
a
8084 82 9296 94 2.67 134 Extreme This study
Bed sediment, Kaoping River
b
63 66 .72 185 Intermediate This study
Core tops, Taiwan Strait
a
6972 71 7982 81 .55 220 Intermediate Chao and Chen 2003
Earliest Cenozoic to Oligocene meta-
sedimentary to sedimentary rocks of
Taiwan:
Phyllite, Early Cenozoic (Chulai)
b
61 65 1.48 237 Intermediate Lan et al. 2002
Sandstone, Eocene (Meichi)
b
63 71 .88 215 Intermediate Lan et al. 2002
Sandstone, Oligocene (Tsuku)
b
5967 63 6275 69 1.33 186 Intermediate Lan et al. 2002
Argillite, Oligocene (Kankou)
b
63 69 1.71 191 Intermediate Lan et al. 2002
Argillite, Oligocene (Tatungshan)
b
64 69 1.20 201 Intermediate Lan et al. 2002
Metapelites, Jurassic (Tienhsiang)
b
4578 62 4486 67 .80 256 Intermediate Lan et al. 2002
River particulates:
Kaoping, Taiwan
b
59 62 Incipient Lai 2003
Yangtze, China
a
74 80 .77 182 Intermediate Li et al. 1984
Yellow, China
a
61 64 .39 233 Intermediate Li et al. 1984
Amazon
b
73 77 Intermediate Martin and Meybeck 1979
Brahmaputra
a
67 73 Intermediate Galy and France-Lanord 2001
Ganges
a
75 87 Intermediate Galy and France-Lanord 2001
River particulate matter
b
65 68 Intermediate Martin and Meybeck 1979
Shale compositions:
Average shale
b
65 71 .82 190 Intermediate Turekian and Wedepohl 1961
Post-Archean Australian shale
b
69 77 .80 192 Intermediate Taylor and McLennan 1985
North American shale composite
b
57 60 .88 264 Incipient Gromet et al. 1984
Loess and crustal compositions:
Loess
a
46 46 .43 196 Equal to UCC Li et al. 1984
Upper continental crust (UCC)
b
46 45 .32 252 Taylor and McLennan 1985
Mean crust
b
46 45 Bowen 1979
Yangtze Craton, China (upper crust)
a
56 58 .32 248 Gao et al. 1998
Central East China (upper crust)
a
54 56 .31 262 Gao et al. 1998
East China (total crust)
a
50 50 .24 272 Gao et al. 1998
Note. Weathering intensity has been given based on the threefold classication of Fedo et al. (1995); weathering intensity for reference compositions such as UCC and mean
crust are not given.
a
Indicates the CaO values of only the silicate fraction.
b
Represents the total CaO used for CIA and PIA calculations.
108 K . S E LVA R A J A N D C . - T. A . C H E N
Figure 3. Major element composition of sediments and
sedimentary rocks, river particulates, loess, and other ref-
erence compositions plotted as molar proportions on an
Al
2
O
3
-( )-K
2
O(A-CN-K) diagram. Arrow 1

CaO Na O
2
represents the weathering trend of river particulates,
arrow 2 indicates the weathering trend of sediments
and sedimentary rocks of Taiwan, and arrow 3 marks
the limit of weathering. The diagram also represents
the elds of idealized minerals: ; Pl pplagioclase
; ; ; Ks pK-feldspars Bi pbiotite Smpsmectite Mu p
; ; ; ; muscovite Il pillite Ka pkaolinite Gi pgibbsite
; . Arrow 2 through the sedi- Ch pchlorite Gt pgarnet
ments and sedimentary rocks intersects the feldspar join
(Pl-Ks) at point x, which may give an indication of the
original composition of the source material. The relation
between chemical index of alteration (CIA) scale (Nesbitt
and Young 1982) and the triangle is shown at right. Most
of the sediments and sedimentary rocks fall between CIA
60 and 80, indicating intermediate intensity of silicate
weathering. Note that the lake sediments fall above CIA
80, suggesting their extremely altered nature. The reason
for the parallelism between the weathering trends of ma-
jor river particulates (arrow 1) and sediments and sedi-
mentary rocks of Taiwan (arrow 2) could be explained as
follows: most of the worlds major rivers are perhaps
draining in the terrains consisting of signicant amounts
of granite, granodiorite, gneiss, and related geologically
older rocks. In Taiwan island, however, the percentage
of these rocks is very low, and most of the rocks are
geologically younger sedimentary and metasedimentary
clans. This probably causes the chemical variability ob-
served between the two weathering trends, which in turn
suggests the slight but signicant differences between
the composition of the original source rocks of sediments
and sedimentary rocks of Taiwan and the composition
of the UCC. The values of A coordinates are given in
parentheses for comparison as well as to avoid confusion
from overlapping points.
ing. CIW ( ; Har- 100Al O / [Al O CaONa O]
2 3 2 3 2
nois 1988) ranges from 70 to 91, with a mean of 81,
for coastal and offshore sediments. These values are
close to those for post-Archean shales, with mod-
erate losses of Ca, Na, and Sr fromsource area weath-
ering (Condie 1993), and they correspond well with
CIA values. The above interpretation is consistent
with the mineralogy of coastal and offshore sedi-
ments, which all show the dominance of quartz,
feldspars, illite, and chlorite composition, with mi-
nor amounts of calcite and kaolinite (Chen 1997).
The CIA values are also consistent with sediment
compositions obtained from ODP site 1144 of the
northern South China Sea (Boulay et al. 2003). The
site is believed to be partly sourced by Taiwans ter-
restrial materials.
Monitoring plagioclase weathering alone also
yields additional information for silicate weather-
ing of sediments and sedimentary rocks, and the
PIA can be calculated as 100(Al
2
O
3
-K
2
O)/(Al O
2 3
-K
2
O) in molar proportions. It can

CaO Na O
2
also be represented on a triangle showing molecular
proportions of Al
2
O
3
(minus Al associated with K),
CaO, and Na
2
O (see the PIA diagram of Fedo et al.
1995). The PIA values for coastal and offshore sed-
iments, in general, fall between 80 and 90 (g. 4),
with a mean of 85, except for a few samples that
show low PIA values (as low as 64; table 2) and
indicate the presence of small amounts of plagio-
clase feldspars in sediments. Sediments display low
CaO values and seem to be totally depleted in the
Journal of Geology S U B T R O P I C A L TA I WA N W E AT H E R I N G 109
Figure 4. Triangular plot (after Fedo et al. 1995) show-
ing molar proportions of Al
2
O
3
(minus that associated
with K), CaO, and Na
2
O. The diagram shows that the
plagioclase index of alteration (PIA) for most of the sed-
iments falls between 80 and 90 (scale shown at right)
because of the presence of small amounts of plagioclase
feldspar resulting from moderate silicate weathering.
Lake sediments fall very close to the Al
2
O
3
-K
2
O apex,
which substantiates that they are the products of highly
weathered sedimentary rocks. ; An panorthite Al p
. Symbols and other abbreviations are as in gure albite
3.
Figure 5. Macs ternary plot of Nesbitt and Young
(1984). A is molar proportions of Al
2
O
3
, CNK represents
, and FM identies .

CaO Na OK O FeO MgO

2 2 T
Symbols are as in gure 3. Also shown are the elds of
idealized minerals (see g. 3 for details; ). Fs pfeldspars
anorthite component (g. 4). The sedimentary
rocks, on the other hand, seem to contain a higher
proportion of anorthite than sediments. Lake sed-
iments show very high PIA values (9296; table 2)
and plot near the Al
2
O
3
apex on the triangle, in-
dicating their highly aluminous character. They are
also virtually depleted in Ca, but all have small
amounts of Na
2
O (sodic feldspar), which is more
prominent in offshore and coastal sediments. These
differences are well matched with a high degree of
weathering for the lake sediments and moderate
weathering for other sediments. Weathering has
proceeded to a stage at which signicant amounts
of Ca and Na have been removed from the sedi-
ments because of copious rainfall (source for or-
ganic and inorganic acids) in this island. Weather-
ing studies show that Ca, Na, and Sr are rapidly
lost during chemical weathering, and the amount
of these elements lost is proportional to the degree
of weathering (e.g., Wronkiewicz and Condie 1987).
The consistency of CIA and PIA values among
the sediments, sedimentary rocks, and metasedi-
mentary rocks clearly suggests that, in general, the
degree of silicate weathering in this island is a pro-
cess operating on a moderate scale over a long pe-
riod (from Jurassic to recent, even though our study
has some gaps in sampling representation for a few
periods such as Miocene, Pliocene, and Pleisto-
cene). The Pleistocene sediments (ODP site 1144)
of the northern South China Sea have CIA values
of 7379 (Boulay et al. 2003). Among the sedimen-
tary rocks, metapelites have a wide range of CIA
values (4578), the lowest value being equal to
those of UCC and loess. The mean value (62;
), however, is consistent with those of the n p8
other sedimentary rocks (6164) but lower than
those of the sediments discussed above. The CIA
and PIA values of average composition of sedimen-
tary rocks are about 20% higher than upper crust
values of Yangtze Craton (g. 2).
To illustrate the approximate mineralogical com-
position of sediments and sedimentary rocks, major
element data of all the samples have been plotted
in the macs triangle (g. 5; Nesbitt and Young
1984; Nesbitt and Wilson 1992). It portrays molec-
ular proportions of Al
2
O
3
, ,

CaO Na OK O
2 2
and (A-CNK-FM). Most of the sedi- FeO MgO
T
ments, sedimentary rocks, and river particulates
(Kaoping, Yellow, and Yangtze) plot within the
compositional triangle of feldspars, garnet, and
chlorite. This implies the approximate mac min-
erals composition of samples; however, x-ray dif-
fractogram (XRD) patterns of recent sediments
110 K . S E LVA R A J A N D C . - T. A . C H E N
have not shown any presence of garnet. Therefore,
the coastal and offshore sediments plot away from
the feldspar-garnet join but close to the feldspar-
chlorite join, suggesting that sediments are essen-
tially composed of feldspars and chlorite, domi-
nated by the latter. This inference is compatible
with the XRD patterns of our coastal and offshore
sediments, all showing prominent feldspar (d p
) and chlorite ( and 3.51) peaks (Chen 3.18 d p7.16
1997). The presence of illite and chlorite inferred
from geochemical data is consistent with the ear-
lier clay mineral investigation by Chen (1973), who
found the illite-chlorite dominant suite over the
continental shelf from the East China Sea to the
southern part of Taiwan Strait. The close associa-
tion of sediments with sedimentary rocks (g. 5)
indicates that the former is mostly the product of
the latter. The lake sediments again plot toward
higher Al than do other sediments and sedimentary
rocks and well above the line of the feldspar-
chlorite join, further conrming their higher sili-
cate weathering. Bed sediment of the Kaoping River
plots close to river particulates of the Yellow River
and average world values, as in the A-CN-K plot;
this denotes moderate silicate weathering in the
river basin of the Kaoping and the similarity in the
mineralogical composition of source rocks of the
Kaoping and Yellow rivers. This nding, however,
contradicts some previous reports (e.g., Yang 2001;
Lai 2003) that suggested high chemical/silicate
weathering rates based on dissolved river ux, prob-
ably because the dissolved ux also includes the
carbonate dissolved elements. The river particu-
lates of the Amazon and Yangtze rivers, however,
plot close to the sediments, with lower values re-
sulting from the incorporation of total CaO instead
of CaO in only the silicate fraction.
Large-Ion Lithophile Elements: Rb, Sr, K, and Na
The increase in chemical weathering intensity rap-
idly leaches Sr compared to Rb (Nesbitt and Young
1982); therefore, the Rb/Sr ratio increases with in-
creasing CIA (Ma et al. 2000). Likewise, with the
increase in chemical weathering intensity, K will
normally show depletion against Rb (Wronkiewicz
and Condie 1989), thus leading to a lower K/Rb
ratio. Rb has been considered to be primarily xed
in weathering residues and less reactive than Ca,
Na, and Sr (Nesbitt et al. 1980). The Rb/Sr ratios
of sediments and sedimentary rocks can thus be
used to monitor the degree of source rock weath-
ering (McLennan et al. 1993). The mean Rb/Sr ra-
tios (0.89 and 0.78) of offshore and coastal surface
sediments of southwestern Taiwan are consistent
with the Rb/Sr ratios of different shale composi-
tions (0.800.88; table 2), corroborating that the de-
gree of source rock weathering was moderate. The
lower mean Rb/Sr ratios of core sediments off
southwestern Taiwan (0.36; ) and Taiwan n p16
Strait (0.55; ) are closer to the Rb/Sr ratios of n p3
different upper crustal and loess compositions
(0.310.43). The mean Rb/Sr ratios of sandstones
and metapelites are 0.88 and 0.80, respectively, in-
dicating that the recent sediments fromcoastal and
offshore regions are mainly derived from these sed-
imentary rocks. This inference is supported by the
similar Rb/Sr ratio (0.71) for bed sediment of the
Kaoping River. High Rb/Sr ratios (1.743.77) of lake
sediments and their higher mean value of 2.67
( ) further support their derivation from in- n p69
tensively weathered rocks such as phyllite and ar-
gillite (mean Rb/Sr ratios 1.48 and 1.71), which are
dominant in the entire drainage basin (90.3 km
2
) of
the lake. Rb/Sr ratios in the illite minerals are usu-
ally 11, with a maximum value of 6.46 (Chaudhuri
and Brookings 1979), suggesting that the higher il-
lite content in lake sediments might be responsible
for higher Rb/Sr ratios. The moderate chemical
weathering of sediments is also indicated by K/Rb
ratios (table 2) of the investigated sediments that
are not depleted in K (g. 2), and mean K/Rb ratios
of coastal (235), offshore (197), and river (185) sed-
iments are consistent with K/Rb ratios of Taiwans
sedimentary rocks (220), upper crusts of Yangtze
Craton (248), and loess (196).
Molar ratios of K/Na of sediments and sedimen-
tary rocks are correlated well with Rb/Sr ratios as
well as CIA and PIA values (g. 6). Both ratios are
moderate and consistent with different shale com-
positions and bed sediment of the Kaoping River,
demonstrating moderate silicate weathering. The
diagram shows the presence of minor plagioclase
(Na and Sr) and K-feldspar (K and Rb) in recent sed-
iments. The close association of sediments with
loess and Yellow River particulates suggests com-
positional similarity and also that the TaiwanStrait
might have been sourced considerably from the
loess plateau either by the Yellow River or from
dust storm input. High K/Na and Rb/Sr ratios of
lake sediments indicate stronger chemical weath-
ering at higher altitudes as well as preferential dis-
solution of plagioclase (Na and Sr) relative to K-
feldspar during the silicate weathering process
(Yang et al. 2004). The scatter plot of Al/Na ratio
versus CIA (g. 7) illustrates the interrelation be-
tween both indexes, which reects the silicate
weathering intensity. The diagram shows that the
degree of chemical weathering of sediments and
sedimentary rocks of Taiwan is moderate, as in-
Journal of Geology S U B T R O P I C A L TA I WA N W E AT H E R I N G 111
Figure 6. Diagram showing variations in Rb/Sr versus
molar K/Na in sediments and sedimentary rocks of Tai-
wan and other reference compositions. Note the two sep-
arate elds; low and intermediate values of chemical and
plagioclase indexes of alteration (CIA and PIA) for sedi-
ments and sedimentary rocks in eld 1, having low Rb/
Sr and K/Na ratios compared to eld 2, encircle the high
CIA and PIA samples of Great Ghost Lake (GGL), which
have higher Rb/Sr and K/Na ratios. See gure 3 for ex-
planation of symbols and abbreviations.
Figure 7. Scatter plot of Al/Na ratio versus chemical
index of alteration (CIA). Note the interrelation between
both indexes, which reects the silicate weathering in-
tensity. Symbols and abbreviations are as in gure 3.
dicated by their low Al/Na ratios (013.09), which
reveal the presence of small amounts of Na
2
O (pla-
gioclase) in the samples. The degree of chemical
weathering of lake sediments is high, as shown by
their very high Al/Na ratios (20.7541.42) resulting
from extreme dissolution of plagioclase. These in-
ferences are consistent with their positions on the
other triangular and scatter plots.
Several lines of evidence have been drawn as ad-
ditional support for the conclusions arrived at in
this study. The XRD patterns of our offshore and
coastal surface sediments show quartz-feldspar-
chlorite-illite-calcite-kaolinite association (Chen
1997). Consequently, illite and chlorite are the
most abundant and ubiquitous clay minerals in the
rock formations on the Hengchun Peninsula in
southern Taiwan (Lin and Wang 2001). Both illite
and chlorite may be derived either from the deg-
radation of muscovite and biotite from metamor-
phic formations or from the erosion of sedimentary
rocks (Chamley 1989). Micas are essential minerals
of rocks such as schist and phyllite. Chlorite is con-
sidered a primary mineral of low-grade metamor-
phic rocks. The basement complex of Taiwan, Tan-
anao schist and the associated choritoid rocks, is
considered the probable source of chlorite in these
sediments. Accordingly, illite, chlorite, and quartz
are treated as products of physical erosion or mod-
erate chemical weathering (Chamely 1989). Colin
et al. (1999) identied from the core sediments of
the Bay of Bengal and Andaman Sea that the sed-
iments deposited during the last glacial maximum
are characterized by a decrease in the smectite/
( ) ratio resulting from increased illite chlorite
physical weathering. If the source rocks experi-
enced intense chemical weathering, feldspars pres-
ent in the source rock would be altered totally as
aluminous clay (Fedo et al. 1996). Repeated cycles
of weathering and abrasion during transport even-
tually result in destruction of feldspars and for-
mation of clay minerals (Nesbitt and Young 1996).
As long as feldspars persist, the sediments will re-
main compositionally immature (Nesbitt and
Young 1996). It has also been mentioned that the
absence of feldspar in sediments and sedimentary
rocks is a characteristic feature of supermature sed-
imentary rocks (Medaris et al. 2003). Maturity
would have been further enhanced by additional
weathering during uvial transport in a warm, hu-
mid climate (Johnsson et al. 1988) and by prefer-
ential destruction of labile minerals and lithic frag-
ments in the high-energy uvial and shallow
marine environments (Odom et al. 1976). The pres-
ence of feldspars in sediments of this study, inferred
from gures 4 and 6 and the illite-chlorite associ-
ation shown in gure 5, suggests their derivation
by physical weathering (a key process in Taiwan)
and/or moderate chemical weathering.
Direct evidence of chemical weathering was
112 K . S E LVA R A J A N D C . - T. A . C H E N
Figure 8. Bivariate plot of CIAand mean grain size (mm)
for sediments of Taiwan. Symbols are as in gure 3.
found in the record of clay minerals in sedimentary
basins by France-Lanord and Derry (1997). To es-
timate CO
2
consumption from silicate weathering,
they selected the Himalayan-derived sediments of
the late Pleistocenemid-Miocene age recovered
from the distal Bengal Fan on ODP Leg 116. Before
7 Ma and after 1 Ma, the clay mineral assemblage
in the Bengal Fan was dominantly illite and chlorite
(I-C assemblage), reecting moderate weathering in
the Ganges-Brahmaputra system. Between 7 and 1
Ma, clays in the Fan were dominantly pedogenic
smectite and kaolinite (S-K assemblage), reecting
more intense weathering. The I-C sediments have
lost little or no K
2
O or MgO relative to the source
rocks but have lost about half of their Na
2
O and
CaO, while the S-Ksediments have lost some MgO,
about half of their K
2
O and CaO, and much of their
Na
2
O. Therefore, the secondary mineral assem-
blage of I-C in the sediments of southwestern Tai-
wan inferred from gures 5 and 6 and their un-
changed K and slightly depleted Mg contents (g.
2) corroborate that these sediments are the products
of physically weathered rocks and/or intermediate
chemical weathering. The clay mineral composi-
tion of present-day Ganges River sediment is es-
sentially illite and chlorite, however, indicating
rapid mechanical weathering in the source area
(Singh et al. 2003).
Mean grain sizes (determined with a Coulter LS
particle size analyzer) of offshore surface (6.4715.08
mm; ) and core (5.138.55 mm; ) sedi- n p12 n p13
ments fall within ne and very ne silt classes of
siliclastic sediments. Coarser coastal sediments
have a wide range of mean sizes, from 65 to 452 mm,
and fall in very ne to medium sand classes. Fine
and medium silt sizes (7.3027.4 mm; ) are n p37
the characteristics of highly weathered lake sedi-
ments. Silt and sand dominance suggests the pres-
ence of a relatively higher amount of feldspars than
secondary clay minerals. The insignicant correla-
tion ( ; ; ) between mean r p0.465 P p0.0001 n p69
grain size and CIA of studied sediments (g. 8) sup-
ports the above inference and indicates that the
chemical variability seems to be less likely to be
grain-size controlled because of nonsteady state
weathering.
Consequences of Erosion and
Chemical Weathering
Tectonism and climate generally determine the rel-
ative rates of erosion and chemical weathering
(McLennan and Taylor 1991). Likewise, the relative
rates of chemical weathering and erosion chiey
control the composition of siliclastic sediments
(Nesbitt et al. 1997). Balanced rates of chemical
weathering and erosion result in steady state
weathering, which produces compositionally sim-
ilar sediments over a long period. Nonsteady state
weathering, however, occurs where climate and
tectonism vary greatly, altering the rates of chem-
ical weathering and erosion and resulting in pro-
duction of chemically diverse sediments. The ex-
treme height of Taiwans mountains (4 km) and the
monsoon climate (average precipitation is 2500
mm/yr with an average of four typhoons per year)
result in very rapid physical denudation and fast
transport of sediments to the ocean. In such a
weathering-limited regime, soils are thin because
of a high rate of mechanical denudation (physical
formation). The residence time of weathering 1 soil
secondary products of weathering is much lower
because storage within the high-standing islands
riverine systems is thought to be minimal (Lyons
et al. 2002); rates of chemical weathering are low,
similar to those of the Himalayas (France-Lanord
and Derry 1997), but high physical weathering rates
continuously create new mineral surfaces that are
responsible for enhanced river chemical uxes
(Gaillardet et al. 1999). Rivers of Taiwan, therefore,
show high elemental uxes that are highly related
to huge storm-induced, physically eroded sedi-
ments transported in short mountainous rivers
(erosion effect) rather than true alteration of feld-
spars to clay minerals, i.e., silicate weathering.
Rivers draining in low-altitude areas may have a
better chance of segregation of elements in weath-
ering relative to high-altitude systems, as con-
cluded by Zhang and Liu (2002). They observed that
Journal of Geology S U B T R O P I C A L TA I WA N W E AT H E R I N G 113
the segregation factor ([Ca K Na Mg]/[Fe
) in suspended matter of Chinese rivers increases Al]
with sediment yields in regions dominated by phys-
ical weathering. In contrast, the segregation factor
decreases with sediment yield in regions dominated
by chemical and biological weathering because the
reduced water column turbidity allows chemical
and biological reactions to process into depth over
the drainage basin. In addition, an increase in the
height/length (H/L) ratio of the river course cor-
responds to a reduced segregation of elements in
weathering. Therefore, the very high H/L ratios of
Taiwan rivers and their huge suspended load (400
t/yr; Dadson et al. 2003) are also probably not con-
ducive to chemical weathering. Further, extensive
erosion results in reduction of soil particle reten-
tion in the weathering crust, coupled with incom-
plete chemical reaction caused by reduction of
water-particle contact and segregation of elements
(e.g., Al and Si; Zhang et al. 2003).
The ratio between the rates of physical and
chemical denudation in Taiwan ranges from 10 to
40 mg/cm
2
/yr, approximately. In general, the areas
of lower physical denudation rate show higher
chemical denudation rates and vice versa (see g.
1 of Li 1976). Similarly, the ratios of physical and
chemical denudation rates of Chinese rivers lie
mostly between 2 and 10 mg/cm
2
/yr, excluding
loess areas, where the ratio is high, at 90 mg/cm
2
/
yr. The average physical erosion rate in the Yellow
Basin was about kg/km
2
/yr, while the av-
6
1.4 #10
erage chemical erosion rate was kg/km
2
/
3
25 #10
yr. In comparison, the average physical and chem-
ical erosion rates in the Yangtze Basin were
and kg/km
2
/yr, respectively
6 3
0.29 #10 104 #10
(Li et al. 1984). This implies that intense physical
weathering cannot naturally result in a high silicate
weathering rate, mainly because the strong erosion
prevents the newly formed sediment from accu-
mulating in the catchments area or in the soil pro-
le. As a consequence, the source rocks and sedi-
ments have less time to react with the weathering
agents, leading to moderately weathered material.
Hence, because of the combination of active tec-
tonic and climatic regimes (high relief, high rain-
fall, and storm-induced landslides), the continental
rocks of Taiwan are eroding rapidly at 37 mm/yr
(Dadson et al. 2004) through processes of uvial
bedrock incision (Hartshorn et al. 2002), landslid-
ing (Hovius et al. 2000), and debris ows (Lin et al.
2004). Large volumes of such physically eroded de-
tritus are responsible for high dissolved elemental
uxes (mass/volume ratio) in the uvial system.
Chemical weathering has a smaller effect, and
physical denudation and subsequent erosion are
more important in controlling sediment composi-
tions of the coastal and offshore sediments studied.
These sediments are produced through nonsteady
state weathering dominated by hill-slope mechan-
ical erosion. The possibilities of highly weathered
zones of soil prole are expected to be slight on the
slopes of steep mountains, since the prole mate-
rials are eroded before chemical weathering can
produce the appropriate mineralogy or soil zona-
tion (Nesbitt et al. 1997). This nonsteady state
weathering is likely to be a characteristic feature
of tectonically active high-standing islands of Asia
and Oceania, such as Taiwan, the Philippines, In-
donesia, New Zealand, and Papua New Guinea,
where the entire spectrum of weathering zones de-
veloped on bedrock is susceptible to rapid erosion.
Tectonism, moderate relief of lake catchments, and
very high rainfall in the alpine region of Taiwan
result in more vegetation, which stabilizes soils
and aids in the production of organic acids that
decompose most of the feldspars. Deep weathering
proles are therefore possible in high-altitude areas
of Taiwan where the dominance of chemical weath-
ering rates resulting from high annual acid input
to mineral zones of soils over erosion produces
highly weathered sediments.
Conclusions
In spite of huge orographic rainfall in this moun-
tainous island, chemical weathering is not an in-
tense process, except in the sediments of alpine
Great Ghost Lake, where the CIA and PIA values
show extreme weathering conditions. The CIA
and PIA values between the upper crust of Yangtze
Craton and the sedimentary rocks of Taiwan and
sedimentary rocks and recent sediments increase
by just 20% in each stage. Moreover, sedimentary
rocks are recycled continental crustal materials,
their derivatives are recent sediments, and the
chemical weathering intensity of sediments is
more or less equal to the weathering conditions
of rivers such as Changjiang and Brahmaputra.
Our conclusions in this study emphasize the need
for additional data in this area and the need to
recheck the chemical weathering rates of this oro-
gen with more precise and appropriate techniques.
The dissolved ux data of almost all the rivers in
Taiwan show a very high concentration of nitrate,
clearly indicating the inuence of anthropogenic
input in the dissolved loads rather than true
weathering uxes from dissolution of rock-
forming minerals. Anthropogenic input appears to
be the major source of riverine nutrients in the
Kaoping River (Yang 2001). For example, anthro-
114 K . S E LVA R A J A N D C . - T. A . C H E N
pogenic inputs of total dissolved nitrogen and
phosphate are about and kg/
4 3
56 #10 48 #10
d, respectively. This condition may hold for most
of the west-owing rivers of Taiwan, where the
total population of the country is residing. Mod-
erate silicate weathering is further supported by
the fact that the concentrations of dissolved silica
in all but six rivers of Taiwan (C.-T. A. Chen, un-
pub. data) are lower than the reported values of
this parameter in 60 large rivers in the world (Gail-
lardet et al. 1999). Clearly there is still a need for
additional information and long-term geochemi-
cal evidence about chemical weathering of this
orogen. Therefore, our future investigation will
focus on long core sediments of the northwest
Pacic eastern side of the island, where huge
amounts of weathered materials are directly trans-
ported by 11 fast-owing rivers from the steep
slopes of the Eastern Range of Taiwan (Fuller et
al. 2003), with an aim to reconstruct the long-term
weathering history of Taiwan.
A C K N O W L E D G M E N T S
Financial support from the Republic of China Na-
tional Science Council (NSC 93-2611-M-110-009,
93-2811-M-110-006, and 93-2621-Z-110-004) is
greatly appreciated. We thank B. N. Nath for help-
ful comments and N. R. Rao for ne-tuning the
revised manuscript. We gratefully acknowledge W.
Nesbitt and an anonymous reviewer for their help-
ful comments and critiques of the original manu-
script.
R E F E R E N C E S C I T E D
Berner, R. A.; Lasaga, A. C.; and Garrels, R. M. 1983. The
carbonate-silicate geochemical cycle and its effect on
atmospheric carbon dioxide over the last 100 million
years. Am. J. Sci. 205:641683.
Boulay, S.; Colin, C.; Trentesaux, A.; Pluquet, F.; Bertaux,
J.; Blamart, D.; Buehring, C.; and Wang, P. 2003. Min-
eralogy and sedimentology of Pleistocene sediment in
the South China Sea (ODP Site 1144). In Prell, W. L.;
Wang, P.; Blum, P.; Rea, D. K.; and Clemens, S. C.,
eds. Proc. ODP, Sci. Results 184:121.
Bowen, H. T. M. 1979. Environmental chemistry of the
elements. London, Academic Press, 333 p.
Chamley, H. 1989. Clay sedimentology. Berlin, Springer,
623 p.
Chao, H.-J., and Chen, J.-C. 2003. Grain size, mineralog-
ical and chemical characteristics of cored sediments
from offshore Hsinchu and their geological implica-
tions. Acta Oceanogr. Taiwan. 41:6196 (in Chinese
with English abstract).
Chaudhuri, S., and Brookings, D. G. 1979. The Rb-Sr sys-
tematics in acid-leached clay minerals. Chem. Geol.
24:231242.
Chen, C.-T. A. 1997. Research report on phytoplankton,
benthos, shery resources and sediment quality inand
around the slag disposal area of the China Steel Com-
pany, off southwest Taiwan. No. 31. Institute of Ma-
rine Geology and Chemistry, National Sun Yat-Sen
University, Taiwan, 322 p. (in Chinese).
Chen, C.-T. A.; Liu, J. T.; and Tsuang, B.-J. 2004. Island-
based catchment: the Taiwan example. Reg. Environ.
Change 4:3948.
Chen, C.-T. A.; Wann, J.-K.; and Lou, J.-Y. 2001. Aeolian
ux of metals in Taiwan in the past 2600 years. Che-
mosphere 43:287294.
Chen, P.-Y. 1973. Clay minerals distribution in the sea-
bottom sediments neighbouring Taiwan island and
northern South China Sea. Acta Oceanogr. Taiwan. 3:
2564.
Colin, C.; Turpin, L.; Bertaux, J.; Desprairies, A.; and Kis-
sel, C. 1999. Erosional history of the Himalayan and
Burman ranges during the last two glacial-interglacial
cycles. Earth Planet. Sci. Lett. 171:647660.
Condie, K. C. 1993. Chemical composition and evolution
of the upper continental crust: contrasting results
from surface samples and shales. Chem. Geol. 104:1
37.
Dadson, S. J.; Hovius, N.; Chen, H.; Dade, W. B.; Heish,
M.-L.; Willett, S. D.; Hu, J.-C.; et al. 2003. Links be-
tween erosion, runoff variability and seismicity in the
Taiwan orogen. Nature 426:648651.
Dadson, S. J.; Hovius, N.; Chen, H.; Dade, W. B.; Lin,
J.-C.; Hsu, M.-L.; Lin, C.-W.; et al. 2004. Earthquake-
triggered increase in sediment delivery from an active
mountain belt. Geology 32:733736.
Dean, W. E., Jr. 1974. Determination of carbonate and
organic matter in calcareous sediments and sedimen-
tary rocks by loss on ignition: comparison with other
methods. J. Sediment. Petrol. 44:242248.
Fedo, C. M.; Eriksson, K. A.; and Krogstad, E. J. 1996.
Geochemistry of shales from the Archean (3.0 Ga)
Buhwa Greenstone Belt, Zimbabwe: implications for
provenance and source-area weathering. Geochim.
Cosmochim. Acta 60:17511763.
Fedo, C. M.; Nesbitt, H. W.; and Young, G. M. 1995.
Unraveling the effects of potassium metasomatismin
sedimentary rocks and paleosols, with implications
for paleoweathering conditions and provenance. Ge-
ology 23:921924.
France-Lanord, C., and Derry, L. A. 1997. Organic carbon
burial forcing of the carbon cycle fromHimalayan ero-
sion. Nature 390:6567.
Fuller, C. W.; Willett, S. D.; Hovius, N.; and Slingerland,
R. 2003. Erosion rates for Taiwan mountain basins:
newdeterminations fromsuspended sediment records
and a stochastic model of their temporal variation. J.
Geol. 111:7187.
Journal of Geology S U B T R O P I C A L TA I WA N W E AT H E R I N G 115
Gaillardet, J.; Dupre, B.; Louvat, P.; and Alle` gre, C. J.
1999. Global silicate weathering and CO
2
consump-
tion rates deduced from the chemistry of large rivers.
Chem. Geol. 159:330.
Galy, A., and France-Lanord, C. 2001. Higher erosion
rates in the Himalaya: geochemical constraints on riv-
erine uxes. Geology 29:2326.
Gao, S.; Luo, T.-C.; Zhang, B.-R.; Zhang, H.-F.; Han, Y.-
W.; Zhao, Z.-D.; and Hu, Y.-K. 1998. Chemical com-
position of the continental crust as revealed by studies
in East China. Geochim. Cosmochim. Acta 62:1959
1975.
Gromet, L. P.; Dymek, R. F.; Haskin, L. A.; and Korotev,
R. L. 1984. The North American Shale Composite:
its compilation, major and trace element character-
istics. Geochim. Cosmochim. Acta 48:24692482.
Han, G., and Liu, C.-Q. 2004. Water geochemistry con-
trolled by carbonate dissolution: a study of the river
waters draining karst-dominated terrain, Guizhou
Province, China. Chem. Geol. 204:121.
Harnois, L. 1988. The CIW index: a new chemical index
of weathering. Sed. Geol. 55:319322.
Hartshorn, K.; Hovius, N.; Dade, W. B.; and Slingerland,
R. L. 2002. Climate-driven bedrock incision in an ac-
tive mountain belt. Science 300:20362038.
Hovius, N.; Stark, C. P.; Chu, H.-T.; and Lin, J.-C. 2000.
Supply and removal of sediment in a landslide-
dominated mountain belt: Central Range, Taiwan. J.
Geol. 108:7389.
Jahn, B.-M.; Zhou, X. H.; and Li, J. L. 1990. Formation
and tectonic evolution of southeastern China and Tai-
wan: isotopic and geochemical constraints. Tectono-
physics 183:145160.
Johnsson, M. J.; Stallard, R. F.; and Meade, R. H. 1988.
First-cycle quartz arenites in the Orinoco River Basin,
Venezuela and Colombia. J. Geol. 96:263277.
Lai, Y. C. 2003. Geochemistry of major and trace ele-
ments in the Kaoping River: weathering and human
inuences. M.S. thesis, Institute of Marine Geology
and Chemistry, National Sun Yat-Sen University, Tai-
wan, 104 p. (in Chinese with English abstract).
Lan, C.-Y.; Lee, C.-S.; Shen, J. J.-S.; Lu, C.-Y.; Mertzman,
S. A.; and Wu, T.-W. 2002. Nd-Sr isotopic composition
and geochemistry of sediments from Taiwan and their
implications. West. Pac. Earth Sci. 2:205222.
Lan, C.-Y.; Lee, T.; Jahn, B.-M.; and Yui, T.-F. 1995. Tai-
wan as a witness of repeated mantle inputs: Sr-Nd-O
isotopic geochemistry of Taiwan granitoids and me-
tapelites. Chem. Geol. 124:287303.
Li, Y.-H. 1976. Denudation of Taiwan island since the
Pliocene epoch. Geology 4:105107.
Li, Y.-H.; Teraoka, H.; Yang, T.-S.; and Chen, J.-S. 1984.
The elemental composition of suspended particles
from the Yellow and Yangtze rivers. Geochim. Cos-
mochim. Acta 48:15611564.
Lin, C. W.; Shieh, C. L.; Yuan, B. D.; Shieh, Y. C.; Liu,
S. H.; and Lee, S. Y. 2004. Impact of the Chi-Chi earth-
quake on the occurrence of landslides and debris
ows: example from the Chenyulan River watershed.
Eng. Geol. 71:4961.
Lin, S. B., and Wang, Y. R. 2001. Clay minerals in the
rock formations on the Hengchun Peninsula, southern
Taiwan, and their tectonic implications. West. Pac.
Earth Sci. 1:157174.
Lou, Y.-J.; Chen, C.-T. A.; and Wann, J.-K. 1997. Paleo-
climatological records of the Great Ghost Lake in Tai-
wan. Sci. China Ser. D 40:284292.
Lyons, W. B.; Nezat, C. A.; Carey, A. E.; and Hicks, D.
M. 2002. Organic carbon uxes to the ocean from
high-standing islands. Geology 30:443446.
Ma, Y.; Liu, C.; and Huo, R. 2000. Strontium isotope
systematics during chemical weathering of granitoids:
importance of relative mineral weathering rates. J.
Goldschmidt Conf. Abstr. 5:657.
Martin, J.-M., and Meybeck, M. 1979. Elemental mass-
balance of material carried by major world rivers. Mar.
Chem. 7:173206.
McLennan, S. M. 1993. Weathering and global denuda-
tion. J. Geol. 101:295303.
McLennan, S. M.; Hemming, S.; McDaniel, D. K.; and
Hanson, G. N. 1993. Geochemical approaches to sed-
imentation, provenance, and tectonics. In Johnsson,
M. J., and Basu, A., eds. Processes controlling the com-
position of clastic sediments. Geol. Soc. Am. Spec.
Pap. 284:2140.
McLennan, S. M., and Taylor, S. R. 1991. Sedimentary
rocks and crustal evolution: tectonic setting and sec-
ular trends. J. Geol. 91:122.
Medaris, L. G., Jr.; Singer, B. S.; Dott, R. H., Jr.; Naymark,
A.; Johnson, C. M.; and Schott, R. C. 2003. Late Pa-
leoproterozoic climate, tectonics, and metamorphism
in the southern Lake Superior region and ProtoNorth
America: evidence from Baraboo interval quartzites.
J. Geol. 111:243257.
Milliman, J. D., and Syvitski, J. P. M. 1992. Geomorphic/
tectonic control of sediment discharge to the ocean:
the importance of small mountainous rivers. J. Geol.
100:525544.
Nath, B. N; Kunzendorf, H.; and Plu ger, W. L. 2000. In-
uence of provenance, weathering, and sedimentary
processes on the elemental ratios of the ne-grained
fraction of the bedload sediments from the Vembanad
Lake and the adjoining continental shelf, southwest
coast of India. J. Sed. Res. 70:10811094.
Nesbitt, H. W.; Fedo, C. M.; and Young, G. M. 1997.
Quartz and feldspar stability, steady and non-steady-
state weathering, and petrogenesis of siliclastic sands
and muds. J. Geol. 105:173191.
Nesbitt, H. W.; Markovics, G.; and Price, R. C. 1980.
Chemical processes affecting alkalis and alkaline
earths during continental weathering. Geochim. Cos-
mochim. Acta 44:16591666.
Nesbitt, H. W., and Wilson, R. E. 1992. Recent chemical
weathering of basalts. Am. J. Sci. 292:740777.
Nesbitt, H. W., and Young, G. M. 1982. Early Proterozoic
climates and plate motions inferred from major ele-
ment chemistry of lutites. Nature 299:715717.
. 1984. Prediction of some weathering trends of
plutonic and volcanic rocks based on thermodynamic
116 K . S E LVA R A J A N D C . - T. A . C H E N
and kinetic considerations. Geochim. Cosmochim.
Acta 48:15231534.
. 1996. Petrogenesis of sediments in the absence
of chemical weathering: effects of abrasion and sorting
on bulk composition and mineralogy. Sedimentology
43:341358.
Nesbitt, H. W.; Young, G. M.; McLennan, S. M.; and
Keays, R. R. 1996. Effects of chemical weathering and
sorting on the petrogenesis of siliclastic sediments,
with implications for provenance studies. J. Geol. 104:
525542.
Odom, I. E.; Doe, T. W.; and Dott, R. H., Jr. 1976. Nature
of feldspar-grain size relations in some quartz-rich
sandstones. J. Sediment. Petrol. 46:862870.
Riebe, C. S.; Kirchner, J. W.; Granger, D. E.; and Finkel,
R. C. 2001. Strong tectonic and weak climatic control
of long-term weathering rates. Geology 29:511514.
Sarin, M. M. 2001. Biogeochemistry of Himalayan rivers
as an agent of climate change. Curr. Sci. 81:1446
1450.
Selvaraj, K.; Ram Mohan, V.; and Szefer, P. 2004. Eval-
uation of metal contamination in coastal sediments
of the Bay of Bengal, India: geochemical and statistical
approaches. Mar. Pollut. Bull. 49:174185.
Singh, I. B.; Jaiswal, M.; Singhvi, A. K.; and Singh, B. K.
2003. Rapid subsidence of western Ganga plain during
late Pleistocene: evidence from optical dating of sub-
surface samples. Curr. Sci. 84:451454.
Taylor, S. R., and McLennan, S. M. 1985. The continental
crust: its composition and evolution. London, Black-
well, 312 p.
Tripathi, J. K., and Rajamani, V. 1999. Geochemistry of
the loessic sediments on Delhi Ridge, eastern Thar
Desert, Rajasthan: implications for exogenic pro-
cesses. Chem. Geol. 155:265278.
Turekian, K. K., and Wedepohl, K. H. 1961. Distribution
of the elements in some major units of the earths
crust. Geol. Soc. Am. Bull. 72:175192.
Water Resources Planning Commission. 1973. Prelimi-
nary estimation of river sedimentation in Taiwan. Tai-
pei, Ministry of Economic Affairs.
Wronkiewicz, D. J., and Condie, K. C. 1987. Geochem-
istry of Archean shales from the Witwatersrand Su-
pergroup, South Africa: source-area weathering and
provenance. Geochim. Cosmochim. Acta 51:2401
2416.
. 1989. Geochemistry and provenance of sedi-
ments fromthe Pongola Supergroup, SouthAfrica: evi-
dence for a 3.0-Ga-old continental craton. Geochim.
Cosmochim. Acta 53:15371549.
Yang, G. I. 2001. Denudation and transport of terrestrial
materials fromthe Kaoping River watershed. M.S. the-
sis, Institute of Marine Geology and Chemistry, Na-
tional Sun Yat-Sen University, Taiwan, 128 p. (in Chi-
nese with English abstract).
Yang, S. Y.; Jung, H.-S.; and Li, C. X. 2004. Two unique
weathering regimes in the Changjiang and Huanghe
drainage basins: geochemical evidence from river sed-
iments. Sediment. Geol. 164:1934.
Young, G. M.; von Brunn, V.; Gold, D. J. C.; and Minter,
W. E. L. 1998. Earths oldest reported glaciation: phys-
ical and chemical evidence from the Archean Mozaan
Group (2.9 Ga) of South Africa. J. Geol. 106:523538.
Zhang, J., and Liu, C. L. 2002. Riverine composition and
estuarine geochemistry of particulate metals inChina:
weathering features, anthropogenic impact and chem-
ical uxes. Estuarine Coastal Shelf Sci. 54:10511070.
Zhang, J.; Ren, J. L.; Liu, S. M.; Zhang, Z. F.; Wu, Y.;
Xiong, H.; and Chen, H. T. 2003. Dissolved aluminum
and silica in the Changjiang (Yangtze River): impact
of weathering in subcontinental scale. Global Biogeo-
chem. Cycles 17:10771087.