Embed Size (px)
Citation preview
ORIGINAL PAPER
Quantitative climate reconstruction linking meteorological,limnological and XRF core scanner datasets: the LakeSanabria case study, NW Spain
S. Giralt • M. T. Rico-Herrero • J. C. Vega •
B. L. Valero-Garces
Received: 22 September 2010 / Accepted: 18 February 2011 / Published online: 1 March 2011
� Springer Science+Business Media B.V. 2011
Abstract Monthly limnological monitoring in Lake
Sanabria (Spain) since 1986 provided a unique oppor-
tunity to test relationships among climate, hydrology
and lake dynamics and how they are recorded in the
lake sediments. Four datasets were employed: (1)
meteorological (monthly maximum and minimum air
temperature and total precipitation), (2) limnological
(Secchi disk, water temperature, conductivity, pH,
dissolved oxygen, nitrate, silicon, total and reactive
phosphorus, and total chlorophylls and chlorophyll a),
(3) hydrological (Tera River water input and output),
and (4) XRF core scanner measurements carried out in
short cores. Linear models between the different
dataset variables allowed us to characterize the climate
signal transmission from one to the other and cross-
correlation analyses permitted us to identify the
different response times (if any) between them.
Principal Component Analyses (PCA) of the limno-
logical and geochemical datasets allowed us to identify
the main processes that link lake dynamics, primarily
nutrient supply and organic productivity, with some
sedimentological processes, e.g. organic matter and
phosphorus accumulation. Sediment chronology was
established by gamma spectrometry (210Pb). Water
input to Lake Sanabria is controlled mostly by the Tera
River input and is linked directly to precipitation.
Response of the Lake Sanabria water budget to climate
oscillations is immediate, as the strongest correlation
between these two datasets occurs with no lag time.
PCA of the limnological dataset indicated that most of
the variance is related to nutrient input, and compar-
ison with the Tera River water discharge shows that
nutrient input was controlled mainly by oscillations in
the hydrological balance. The lag time between the
hydrological and limnological datasets is 1 month.
The PCA of the XRF core scanner dataset showed that
the principal process that controls the chemical com-
position of the Lake Sanabria sediments is related to
sediment and nutrient delivery from the Tera River and
organic productivity. Comparison of the nutrient input
reconstructed using the limnological dataset and the
XRF core scanner data indicated that the sediments act
as a low-pass filter, smoothing the climate signal. It
was, however, possible to establish the link between
these datasets, and obtain a quantitative reconstruction
of precipitation for the 1959–2005 AD period that
captures the regional variability. This quantitative
S. Giralt (&)
Institute of Earth Sciences Jaume Almera (ICTJA-CSIC),
Lluıs Sole i Sabarıs s/n, 08028 Barcelona, Spain
e-mail: [email protected]
M. T. Rico-Herrero � B. L. Valero-Garces
Pyrenean Institute of Ecology (IPE-CSIC), Av.
Montanana 1005, 50059 Zaragoza, Spain
e-mail: [email protected]
B. L. Valero-Garces
e-mail: [email protected]
J. C. Vega
Laboratory of Limnology of the Lago de Sanabria Natural
Park, Eras 1, 49300 Puebla de Sanabria, Zamora, Spain
e-mail: [email protected]
123
J Paleolimnol (2011) 46:487–502
DOI 10.1007/s10933-011-9509-x
precipitation reconstruction suggests it is possible to
obtain accurate climate reconstructions using non-
laminated sediments.
Keywords Quantitative precipitation
reconstruction � Lacustrine sediments � Iberian
Peninsula � Statistical modeling � Multiproxy
approach
Introduction
The sediment record of lakes is one of the best
continental sensors for qualitative reconstruction of
past environmental and climate changes. Using a
variety of techniques and proxies, recent lacustrine
sediments have been used to quantitatively infer, with
high-temporal resolution, past precipitation (Kalugin
et al. 2007; Nichols et al. 2009; Tonello et al. 2009)
and temperature (Pla and Catalan 2005; Francis et al.
2006; Blass et al. 2007; Larocque and Finsinger
2008), as well as to qualitatively derive land manage-
ment (Hyodo et al. 2008; Djamali et al. 2009;
Striewski et al. 2009) and lake environmental or
limnological conditions such as pH (Schwalb and
Dean 2002; Battarbee et al. 2005), salinity (Chen et al.
2010), or quantify nutrient concentrations (Bennion
et al. 2005; Bigler and Hall 2002). The reliability of
paleoenvironmental reconstructions improves if they
are validated with instrumental data (Fritz 1990;
Kattel et al. 2008). Although quantitative climate and
environmental reconstructions constitute essential
robust data for Global Climate Models (GCM), few
quantitative high-resolution records are available
(Trachsel et al. 2008). Scarcity of such records is
due, in part, to the lack of good, well-calibrated proxy
datasets of high resolution and spatial distribution, as
well as a lack of long-term ([10 years) lake monitor-
ing data to validate environmental reconstructions. In
most regions there are few long, reliable limnological
and/or meteorological datasets, which precludes con-
version of qualitative climate reconstructions into
quantitative ones.
The Iberian Peninsula (IP) is an excellent site to
conduct quantitative climate reconstructions owing to
its location between the Eurosiberian and Mediterra-
nean biogeographic/climatic regions (Carrion et al.
2010). Due to its geographic position, climate across
most of the IP is controlled by the North Atlantic
Oscillation (NAO), the position of the Azores High
and the Intertropical Convergence Zone (ITCZ),
though other climate modes such as the El Nino
Southern Oscillation (ENSO) also exert some influ-
ence (Rodo et al. 1997; Cacho et al. 2001; Romero-
Viana et al. 2008; Martın-Puertas et al. 2008).
There have been several attempts to reconstruct
climate variability using lake sediment records, with
variable degrees of success (Battarbee et al. 2002,
2005; Catalan et al. 2009; Pla and Catalan 2005;
Trachsel et al. 2008; Tonello et al. 2009). Only a few
tried to calibrate the sediment record using instru-
mental or monitoring data. Recently, the link between
measured physical–chemical variables and sedi-
ments has been established using marine sediments
(Abrantes et al. 2009).
In Spain, Lakes Sanabria and Estany Redon have
some of the longest monitoring records. Physical and
biological variables in the lakes were measured
monthly since 1986 (23 years) and 1984 (25 years),
respectively (De Hoyos 1996, http://www.ceab.csic.
es/obser1.htm). In the Spanish Central Range there
are also monthly data for Penalara Lake (Sierra de
Guadarrama, Madrid) since 1995 (14 years) and
several years between 1989–1993 (Toro et al. 2006)
and Cimera Lake (Sierra de Gredos, Avila) for the
1996–1999 and 2006–2008 periods (Granados and
Toro 2000). The aim of this study was to explore how
the climate signal is transferred from the atmosphere
to the lake, and ultimately to the sediments. Estab-
lishment of such links allowed us to infer quantita-
tively the pattern of precipitation over 46 years
(1959–2005 AD) using the chemical composition of
recent sediments in Sanabria Lake.
Site description
Lake Sanabria (Zamora) is located on the northwestern
Iberian Peninsula (IP) (42�070 N–06�430 W), at
1,000 m a.s.l., within a protected area called ‘‘Parque
Natural del Lago de Sanabria y Alrededores’’
(Fig. 1). The basement of the catchment consists of
granitic rocks (gneiss and granodiorite) that origi-
nated during the Variscan Orogeny (Martınez-Garcıa
1973; Vega and Aldasoro 1994), and of Quaternary
deposits of glacial origin. The lake is located in a
valley bounded by a terminal moraine that was
deposited as ice retreated at the end of the Last
488 J Paleolimnol (2011) 46:487–502
123
Glacial Maximum (Vega et al. 1991; Cowton et al.
2009). The long axis of the lake runs W–E, coincid-
ing with the longitudinal direction of the pre-glacial
fluvial valley. From the climate perspective, the lake
lies on the boundary between the relatively maritime
north coast of Iberia and the arid central Meseta
altiplano. Monthly surveys since 1986 indicate that
river discharge, lake productivity and phytoplankton
succession are greatly influenced by the North
Atlantic Oscillation (NAO), highlighting the strong
Atlantic influence on the entire area (De Hoyos 1996;
Luque and Julia 2002).
Lake Sanabria is the largest glacial lake (368 ha) in
the IP (Aldasoro et al. 1991; Vega et al. 1992; De
Hoyos 1996). The lake is approximately 3,160 m long
by 1,530 m wide and has a volume of 96 hm3 as
determined by topographic survey (Vega et al. 2005).
The lake bathymetry displays two subbasins. The
western subbasin has a maximum depth of 46 m,
whereas the eastern subbasin has a zmax of 51 m. The
topographic high that separates the subbasins is
located 20 m below the water surface. Transverse
sections show a typical glacial morphology, with steep
north and south basin margins and a flat bottom. The
east and west margins show gentle gradients. The lake
lies in an exorheic basin (127.3 km2). The Tera River
is the main source of water and sediment input to the
lake. The hydrologic contribution of this river to Lake
Sanabria ranges between 60 and 100% (mean 85%) of
the total water input, depending on the month and year.
Upstream there are small dams used to produce
electric power, to regulate flow, and provide water to
the lake during summer. Nevertheless, their capacity is
limited and they are repeatedly emptied within the
year, depending on water demands downstream. They
discharge through an overflow channel during strong
storms. Their influence on the hydrology of the Tera
River can therefore, be considered negligible. A
hydroelectric power plant on the Tera River is located
at Moncabril, approximately 1.5 km upstream of Lake
Sanabria (Fig. 1). Furthermore, there are small, semi-
permanent input streams located around the lake.
Fig. 1 a Location of Lake Sanabria on the Iberian Peninsula.
b Bathymetric map of the lake and location of the core sites
and of the monthly limnological sampling locations (red filleddots). The bathymetric contour interval is 1.70 m (modified
from Vega et al. 2005). The locations of the old village of
Ribadelago, the new Ribadelago de Franco village constructed
after the 1959 AD flood that destroyed the old village, and the
Moncabril waste treatment plant are also shown. The distance
between the waste treatment plant and the new Ribadelago de
Franco village is about 1.5 km
J Paleolimnol (2011) 46:487–502 489
123
The most important one is Seoane Creek, located in
the NW, and those found along the south shore of the
lake. These small streams and creeks are mostly active
during autumn, when most of the annual precipitation
in this area occurs. Water input into the lake is strongly
linked to precipitation in the watershed and limnolog-
ical studies proved that the lake is highly sensitive to
the rainfall regime (De Hoyos 1996). Lake Sanabria is
oligotrophic and warm monomictic, remaining ther-
mally stratified from March/April to mid November.
The lake water has pH values between 6 and 7.3,
conductivity is *13 lS/cm and Total Dissolved
Solids (TDS) are between 7.5 and 13 mg/l (De Hoyos
1996).
Lake Sanabria experienced a catastrophic event
when the dam on the Vega de Tera Reservoir, located
upstream of the Tera River, failed on the night of 9
January 1959. Intense rainfall and poor dam con-
struction provoked the collapse and a flash flood that
killed 144 inhabitants of Ribadelago, a small village
located between the lake and the dam. Rapid input of
about 8 9 103 m3 of water, sediment and debris into
the lake, raised the lake water level about 2.6 m and
left a clear signal (a clastic layer) in the sediment
record (Luque 2003). The village was reconstructed
downstream of the Tera River and it is known as
Ribadelago de Franco (Fig. 1).
Materials and methods
Sediment coring campaigns were carried out in May
2004 and June 2007. We retrieved 5 Kullenberg
cores, up to 8.9 m long, and 13 short gravity cores,
from 29 to 69 cm long, in the two subbasins. All
cores except one were immediately sealed and stored
in a cold room at ?4�C until they were opened for
sampling in the laboratory. One short core (SAN07-
2 M) was extruded and sampled in the field. The
other short gravity cores were transported and stored
vertically to prevent disturbance of the sediment-
water interface. This paper focuses on two short
gravity cores, SAN07-1 M (33 cm) and SAN07-2 M
(64 cm), collected in 2007 from the deepest part of
the east subbasin (Fig. 1).
Magnetic susceptibility and lightness were mea-
sured every 0.5 cm using a GEOTEK multi-sensor
core logger at the Limnological Research Center
(LRC), University of Minnesota (USA). Grain size
was measured on SAN07-2 M at 1-cm intervals with
a LS-13-320 Beckman Coulter Counter in the
Department of Stratigraphy, Paleontology and Marine
Geosciences, University of Barcelona (Spain). X-Ray
Fluorescence (XRF) core scanner data from SAN07-
1 M allowed us to do a high-resolution characteriza-
tion of the chemical composition of the uppermost
sediment infill. XRF measurements were carried out
every mm for the uppermost 13 cm using the ITRAX
XRF core scanner of the Large Lake Observatory
(USA). The XRF settings were 45 kV, 30 mA and
30 s of time exposure with Mo and Cr tubes to obtain
statistically significant results for the maximum
number of chemical elements possible. Results are
expressed as element intensities in counts per second
(cps).
The Total Carbon (TC) content of the two studied
cores was determined every cm using a LECO SC-
144DR available at the Pyrenean Institute of Ecology
(IPE-CSIC), Spain. Results are expressed in percent
TC with respect to the total sample weight.
The SAN07-2 M core was sliced into cm sections
in the field immediately after recovery, and samples
were stored in sterilized, pre-weighed containers. The210Pb activity was measured by gamma spectrometry
in 20 samples between 0 and 64 cm depth. Measure-
ments were carried out at the St. Croix Watershed
Research Station, Science Museum of Minnesota
(USA). Results are expressed in pico-Curies per gram
dry sediment (pCi/g). The age model was obtained
using the CIC (constant initial concentration) model,
which assumes that the initial excess 210Pb activity in
surface sediments is constant and there is no mixing
of sediments (Robbins and Edgington 1975).
Since January 1986, the lake has been monitored
monthly for Secchi disk, water temperature, conduc-
tivity, pH, dissolved oxygen, nutrients (nitrates, sili-
con, total phosphorus, reactive phosphorus, total
chlorophyll and chlorophyll a) and hydrological (Tera
River water input and output, expressed as cubic
meters per month, m3/month) variables (De Hoyos
(1996). Table 1 shows the range of values for these
variables for the period 1986–2005 AD. Conductivity,
pH, temperature and dissolved oxygen were measured
in situ using a WTW multi-parameter liquid analyzer.
Secchi disk measurements were carried out using a
19-cm-diameter black and white disk. Water samples
were obtained using a 5-L PVC Niskin bottle.
Determination of nutrient concentrations generally
490 J Paleolimnol (2011) 46:487–502
123
followed standard procedures (APHA 1989), though
some methods were modified slightly owing to low
chemical concentrations. Phosphates were determined
using a modified version of the method of Murphy and
Riley (1962), by using a spectrophotometer with a
cuvette containing a 100-mm light path. Nitrates
(Morris and Riley 1963) and chlorophyll (SCOR-
UNESCO 1966) contents were measured using stan-
dard methods, whereas silicon was determined using
the Dienert and Wandenbulcke (1923) method. Mea-
surements were generally carried out between the
second and the third week of every month in the center
of the deepest subbasin at 0, 2.5, 5, 10, 15, 20, 25, 35
and 45 m water depth for nutrients and every 2.5 m
from 0 to 50 for oxygen and temperature. Total and
reactive phosphorus concentrations are different
before and after 1990 owing to the installation of a
wastewater treatment plant on the Tera River just
above where it enters the lake. Total phosphorus
content was not measured systematically before
January 1992. Hence, the limnological dataset
employed in this study ranged from 1992 to 2005
AD, whereas the hydrological dataset comprised data
from January 1986 to December 2005. All variables
were measured on the same day.
Monthly maximum and minimum air temperature
(Celsius degrees), and monthly precipitation (millime-
ters) from the Ribadelago, Paramio de Sanabria and
Puebla de Sanabria meteorological stations cover the
period 1950–2007 and were used to explore relation-
ships between rainfall, Lake Sanabria hydrological
balance, and limnological variables. Meteorological
variables were checked for inconsistencies and quality
control followed the recommended procedure of Brunet
et al. (2006). The most complete period of instrumental
information, i.e. limnological and meteorological data,
spans from January 1992 to December 2007.
We used statistical approaches such as ordination
analyses (Principal Component Analyses), factor
analysis, time series (auto- and cross-correlation
functions) and linear models (lm) using the R software
package (R Development Core Team 2010) to inves-
tigate relationships between the chemical composition
of the sediment core and the limnological dataset.
Factor Analysis (FA) allowed us to define linear
combinations of variables (factors) that represent
underlying fundamental quantities of which the
observed variables are expressions. The idea behind
FA is that a small number of factors might explain aTa
ble
1M
axim
um
and
min
imu
mv
alu
eso
fth
em
on
thly
lim
no
log
ical
par
amet
ers
of
Lak
eS
anab
ria
for
the
Jan
uar
y1
98
6–
Mar
ch2
00
8st
ud
ied
per
iod
Par
amet
erW
ater
dep
th
Ox
yg
en
con
ten
t
Wat
er
tem
per
atu
re
pH
To
tal
ph
osp
ho
rus
Rea
ctiv
e
ph
osp
ho
rus
Nit
rate
sS
ilic
on
To
tal
Ch
loro
ph
yll
Ch
loro
ph
yll
-aS
ecch
i
dis
k
Co
nd
uct
ivit
y
Un
its
mm
g/l
�Clg
P/l
lg
P-P
O4/l
lgN
-NO
3/l
mg
/llg
/ll
g/l
ml
S/c
m
Min
imu
mv
alu
e0
1.5
3.4
5.2
10
.10
.10
.90
.06
0.0
13
.58
.6
Max
imu
mv
alu
e5
01
2.3
25
.28
.41
82
6.0
22
1.0
2.4
9.3
88
.31
12
.21
9.0
J Paleolimnol (2011) 46:487–502 491
123
large number of measurements (Venables and Ripley
2002). Linear regression models belong to classical
statistics and are the basis for much of the statistical
practice. These statistical models assume that there is a
plausible physical explanation for a trend in a variable
that can be be modeled in a deterministic manner.
These models are part of a larger multivariate statistical
family that has been used widely to develop transfer
functions to reconstruct meteorological variables
quantitatively from biological and inorganic variables
(Guiot and de Vernal 2007 and references therein).
They have been used extensively in paleoclimatic and
paleoenvironmental reconstructions (Johnson and
Ingram 2004; Mangini et al. 2005; Danis et al.
2006; Blass et al. 2007; Guiot and de Vernal 2007).
Stratigraphically-constrained cluster analysis was
employed to define homogenous geochemical zones.
Results
Relationships between Lake Sanabria water input
and meteorological variables
Because the Tera River contributes 60–100% (mean
85%) of the total water input to Lake Sanabria,
depending on the month and year, this implies that
secondary streams and creeks on the south and NW
shores have a limited influence on the lake water
budget. Initially, the Tera River water discharge was
considered as the only water input to the lake.
Figure 2 displays the Tera River discharge entering
Lake Sanabria, total annual precipitation and the
maximum air temperature for the period between
January 1986 and December 2004. The Tera River
water discharge and precipitation show a similar
pattern, with maximum values between October and
April and minimum ones between May and September.
Maximum temperature values are inversely related to
both precipitation and Tera River water discharge.
Maximum temperatures are usually recorded between
May and October, whereas minimum values occur
during the rest of the year.
A linear model (lm) for the 1986–2004 period
shows that monthly meteorological variables rainfall
and temperature are strong predictors of annual
oscillations of Tera River water discharge. The two
meteorological variables explain [76% of the total
variance of the Tera River water discharge:
water discharge ¼ 8234454þ 39199 � precipitation
� 337594 � temperature
ð1Þ
where water discharge is expressed in m3/month,
precipitation in mm and temperature in degrees
Celsius. The best correlation between Tera River
water discharge and the meteorological variables
occurs when no lag time is used, indicating that
regional climate fluctuations have an immediate
effect on the total amount of water entering the lake.
Precipitation is the variable that contributes most
to oscillations in river discharge (r = 0.85, p \0.001), though temperature also has an effect
(r = -0.55, p \ 0.001). The lowest Tera river water
discharge values usually occur during the summer
months, when the temperature (precipitation) reach
Fig. 2 Monthly
instrumental maximum
temperature (blue solidline), monthly precipitation
(green solid line) and
monthly Tera River water
discharge (red solid line)
for the period between
January 1986 and
December 2004
492 J Paleolimnol (2011) 46:487–502
123
its highest (lowest) values, indicating that evapora-
tion in the catchment also plays a role in the Lake
Sanabria water budget.
As precipitation is the instrumental variable that
explains the largest variability of the Tera River
water discharge to Lake Sanabria, water discharge
was modeled only using this meteorological variable:
water discharge ¼ 3852606þ 44518 � precipitation
ð2aÞThis equation explains 71% of the total variance of
Tera River water discharge, which indicates that
precipitation values can be inferred from water
discharge with accuracy. Reorganizing Eq. 2a, we
can write:
precipitation ¼ �86:54þ 2:25� 10�5
� water discharge ð2bÞ
Relationships between the Lake Sanabria limnological
parameters
Relationships between mean monthly limnological
variables in the eastern subbasin were explored using
Principal Component Analysis (PCA). The first two
eigenvectors (EV) explain 45% of the total variance.
The first one accounts for[24%, whereas the second
one explains 20.8% (Fig. 3).
The first EV is associated with dissolved oxygen,
reactive phosphorus and pH at the positive end, and
by water temperature, conductivity, and to lesser
extent, nitrates, chlorophyll a and total chlorophyll at
the negative end. On the other hand, the second EV is
related to Secchi disk and water temperature at the
positive end, and by pH, chlorophyll a and total
chlorophyll at the negative end (Fig. 3).
Lithology
SAN07-1 M and SAN07-2 M consist of homoge-
nous, massive, dark brown silty clays. Two erosion
surfaces were identified in core SAN07-1 M by
changes in sediment color, magnetic susceptibility
and total carbon (TC) values (Fig. 4). This interme-
diate layer, between 18.5 and 7 cm core depth and
identified in Fig. 4 by a gray band, is composed of
coarser silts, with higher magnetic susceptibility and
lower TC content. Mean grain size between 22 and
11 cm in the SAN07-2 M core ranges from about
30–60 lm, whereas sediments located above and
beneath have a mean grain size of about 20 lm. The
coarse layer is visible in both cores (gray bands of
Fig. 4) and displays a coarsening upwards sequence
with lighter, less organic sediment towards the top.
This distinctive layer corresponds to sediments
deposited during the Tera River dam failure in 1959
AD and is recognizable all over the lake (Luque
2003).
Chronological model
The 210Pb activity profile in SAN07-2 M shows a
general decreasing trend from the top of the core
(18.2 pCi/g) to 30 cm depth (0.43 pCi/g), below
which it is relatively constant to the base of the core
(Fig. 5). The 210Pb fluctuation between 11 and 22 cm
depth corresponds to the coarser, lighter-color sedi-
ment layer deposited during the Tera River dam
collapse. Lower 210Pb activities were a consequence
of dilution of the 210Pb by rapid accumulation of
coarser sediment after the dam collapse. Sedimenta-
tion of this layer is considered instantaneous, and it
was used as a chronological marker. There is
considerable excess 210Pb in the 10–11 cm interval,
which suggests gradual deposition of sediments over
time following the rapid deposition of the coarser
dam-related sediments. Application of the CIC model
Fig. 3 Plot of the plane defined by the first two eigenvectors
obtained by Principal Component Analysis (PCA) of the
monthly instrumental limnological dataset for the 1992–2005
AD period. Numbers refer to the years. Recent years are
indicated by large numbers
J Paleolimnol (2011) 46:487–502 493
123
gives an date of 1961.4 ± 5.7 AD at 13 cm depth,
which is statistically indistinguishable from the date
of dam failure in 1959. Therefore, the uppermost
11 cm of core SAN07-2 M were deposited from 1959
to 2005 AD.
Geochemical composition of the sediments
XRF core scanner analyses of SAN07-1 M evaluated
17 chemical elements (Al, Si, P, S, K, Ca, Ti, Mn, Fe,
Ni, Zn, As, Se, Rb, Sr, Zr and Pb).The ratio between
the incoherent (Compton) and coherent (Raleigh)
X-ray scatter intensities was used as an indicator of
organic matter content (Fig. 6), as has been done for
other lake cores (Saez et al. 2009). The excellent
covariance between the measured TC and this ratio,
and the higher measurement resolution of the latter,
allowed us to use the inc/coh ratio in the statistical
model as indicator of the organic matter content.
Stratigraphically-constrained cluster analysis of
the uppermost 13 cm of XRF core scanner data from
core SAN07-1 M revealed that there are two geo-
chemically homogenous zones. Zone A extends from
13 to 6.35 cm depth and the inc/coh ratio is very low.
Zone B runs from 6.35 cm to the core top, where
there is a progressive upward increase in the inc/coh
ratio and a decrease in lithogenic elements.
Total Carbon (TC) profiles in sediments deposited
after the Tera River dam failure display the same
trends in both cores (Fig. 4). Above 6.35 cm in
SAN07-1 M and above 11 cm in SAN07-2 M, the
TC content generally rises, despite minor fluctua-
tions. Previous works carried out in other cores from
Lake Sanabria showed that the total inorganic carbon
(TIC) content is negligible (Luque 2003). Therefore,
for Lake Sanabria, TC represents the total organic
carbon (TOC) content.
The TC content, grain size, magnetic susceptibility
and lithological features of the deposits suggest that
the sediments of zone A, 22-11 cm in core SAN07-
2 M and 18.5-6.35 cm in core SAN07-1 M, were
deposited after the main flooding event, but nonethe-
less during a period of unusual sedimentation in the
lake caused by the Vega de Tera dam collapse. Zone
B corresponds to sediments deposited in Lake
Sanabria after a return to baseline lake sedimentation
Fig. 4 Main lithological features of cores SAN07-1 M and
SAN07-2 M. Core photographs, lightness values and magnetic
susceptibility correspond to the SAN07-1 M core, whereas the
mean and median grain size curves were obtained from core
SAN07-2 M. Total Carbon (TC) profiles were obtained in both
cores. The gray bands identify sediments deposited due to the
Vega de Tera dam collapse in 1959 AD
494 J Paleolimnol (2011) 46:487–502
123
Fig. 5 Right. Unsupported 210Pb profile obtained by gamma
spectrometry on SAN07-2 M sediments. Left. Age model
derived form the application of the CIC (Constant Initial
Concentration) model. Horizontal error bars represent 95%
confidence interval. In both cases, the gray, thick vertical baridentifies the sediments correspondent to the Tera River dam
failure
Fig. 6 XRF-core scanner
profiles of the main
chemical elements in
SAN07-1 M sediments.
Stratigraphically-
constrained cluster analysis
allowed us to identify the
sediments related to the
Tera River dam collapse.
Distances are expressed as
squared Euclidean distances
J Paleolimnol (2011) 46:487–502 495
123
conditions, following the catastrophic event. Princi-
pal Component Analysis (PCA) of the XRF dataset
was only conducted on samples from Zone B (Fig. 7).
The PCA results of the XRF dataset show that the
first two EVs account for more than 53% of the total
variance. The first EV explains 36.1%, whereas the
second one only accounts for 17.1%. The first eigen-
vector is mainly tied to the inc/coh ratio, an indicator of
the organic matter content as explained above, and to a
lesser extent, to S, P and Se at the positive end and K,
Ti, Si, Ni and Fe at the negative end. The second
eigenvector (EV2) is controlled by Pb, Zr and Sr, and to
a lesser extent by Ca, Rb, Mn and S at the positive end
and by Al and As at the negative end (Fig. 7). Factor
Analysis (FA) of this dataset found that only EV1 is
significant (w2 = 755.85 on 44 degrees of freedom,
p \ 0.0001).
Discussion
Chronological model
The chronological model for core SAN07-2 M has an
associated dating uncertainty (standard deviation)
that ranges between 2.46 and 4.66 years for the
period 2005–1959 AD. This uncertainty confounds
age control when the chronological model from core
SAN07-2 M is transferred to core SAN07-1 M. In
spite of this problem, good agreement between the
TC profiles in the two cores permits good strati-
graphic correlation between the profiles and justifies
accurate projection of 210Pb modeled ages from dated
core SAN07-2 M to depths in core SAN07-1 M,
which was analyzed by XRF (Fig. 8). Sediments
corresponding to the Tera River dam collapse are
found in core SAN07-1 M between about 18 and
7 cm depth and the entire interval corresponds to
1959 AD (Fig. 4). Consequently, the uppermost
6.5 cm of the sediment in core SAN07-1 M contains
the environmental history from 1959 to 2005 AD.
Environmental interpretation of the limnological
eigenvectors
We focused on the first EV because it explains the
largest percent of the total variance (24.2%). Oppo-
sition of dissolved oxygen and reactive phosphorus,
to water temperature and conductivity, and to lesser
extent nitrates, in the EV1 of the PCA, is thought to
be related to periodic oscillations in nutrient input to
the lake linked to water input (Fig. 9). During
summer, when the water input to the lake is low
and the lake water column is thermally stratified,
which implies warm epilimnetic waters, nitrogen-fixing
Fig. 7 Plot of the first two eigenvectors obtained by the
Principal Component Analysis (PCA) of the topmost 6 cm in
the SAN07-1 M XRF-core scanner dataset. Numbers represent
the location of the samples with respect to their core depth.
Large numbers indicate samples located close to the bottom of
the core, whereas small numbers point to samples near the top
Fig. 8 Comparison of Total Carbon (TC) variations in cores
SAN07-1 M (light dashed line) and SAN07-2 M (dark solidline) for the period 1960–2006. Note that despite the different
sampling intervals applied to both cores, the TC contents
display the same long-term trend
496 J Paleolimnol (2011) 46:487–502
123
cyanobacteria proliferate in the metalimnion. As
summer progresses, Secchi disk measurements
increase and reach their maximum values during
autumn, due to water column stability, feeding
activity of zooplankton (rotifers) and to the low
productivity of phytoplankton in the epilimnion. The
major water input to the lake, however, occurs during
autumn and winter. During this period, cold, well-
oxygenated water enters the lake, along with
allochthonous suspended material and nutrients. Fur-
thermore, the thermocline disappears, provoking
mixing of the lake water column. Mixing oxygenates
the bottom waters and recirculates phosphorus that
accumulated in deeper water layers during the period
of stratification. The input of external nutrients and
regeneration of available phosphorus triggers the
growth of diatoms during this period (De Hoyos
1996; Luque 2003).
The first EV of the PCA on the monthly limno-
logical dataset can therefore, be interpreted as an
indicator of wet/dry season conditions (pplimno).
Positive values of the EV1 indicate wet conditions
and are linked to the rainy season (spring and
autumn), whereas negative values of this eigenvector
mark dry-season conditions related to summer. A plot
of the x-coordinates of every sample, calculated from
the vector plane defined by the first two eigenvectors
of the PCA, with respect to their core depth, allowed
us to qualitatively reconstruct wet/dry-season condi-
tions (pplimno) through time.
Periodic oscillations of the pplimno are clearly
linked to fluctuations of the Tera River water inputs
to the lake (Fig. 9). These calculated wet/dry-season
conditions do not track the Tera River water input
exactly. The best correlation between the two vari-
ables is obtained with a one-month lag time
(r = 0.66, p \ 0.001).
Good correlation between the instrumental Tera
River water discharge and the reconstructed wet/dry-
season conditions allowed us to model the first
variable as a function of the second one. The equation
that defines the relationship between these two
variables, taking into account the one-month response
(lag) time is:
water discharge ¼ 9633827þ 3026090 � pplim no
ð3Þ
where water discharge is expressed as m3/month and
pplimno does not have units.
In this case, the equation explains 43.3% of the
total variance (p \ 0.0001), suggesting that the lake
water column smoothes the meteorological signal due
to its buffer effect.
Environmental interpretation
of the sedimentological eigenvectors
Similar to the limnological data set, we only focused
on the environmental significance of the first EV
because it explains the largest variance of the PCA.
The opposite behavior of organic matter (inc/coh
ratio) versus K, Ti, Si, Ni and Fe content (Fig. 7) can
be interpreted in terms of oscillations in primary
productivity (ppxrf). Higher inc/coh values indicate
Fig. 9 Comparison of the first EV obtained in the PCA
applied to the limnological dataset (dark solid line) with Tera
River water discharge to Lake Sanabria (light solid line) for the
period 1992–2006. Note the excellent agreement between both
curves when a one-month lag is applied to the first dataset
J Paleolimnol (2011) 46:487–502 497
123
higher organic matter in the sediments. Accumulation
of organic matter in the sediments is a consequence
of high biotic activity in the lake water column,
mainly chlorophyceans and diatoms, because they are
the most dominant phytoplankton in the lake. On the
other hand, terrigenous input to the lake is directly
related to the Tera River water input. Autumn rainfall
provokes an increase in catchment runoff and mobi-
lization of topsoil particles that are delivered to the
Tera River at the same time that Tera River water
discharge to Lake Sanabria increases. Therefore,
positive values of the ppxrf correspond to periods of
enhanced primary productivity and organic matter
accumulation, whereas negative ones represent peri-
ods of low primary productivity.
Comparison of the first limnological
and sedimentological eigenvectors
The first eigenvector of the PCA applied to the
limnological dataset (pplimno) responds to changes in
the wet/dry-season conditions, whereas the first
eigenvector from the XRF core scanner dataset
(ppxrf) responds to oscillations in lake primary
productivity. Therefore, they must show the same
temporal responses, as they must be related. During
wet-season conditions, higher input of nutrient-rich
and well-oxygenated waters increases lake primary
productivity. Conversely, during the dry season, there
is lower input of nutrients, waters are poorly
oxygenated, and primary productivity is lower.
Figure 10 compares the first eigenvector obtained
in the PCA applied to the limnological dataset with
that obtained in the PCA applied to the XRF core
scanner dataset for the period 1992–2005 AD. In
spite of differences in smoothness, due to the
different temporal sampling intervals, and a temporal
offset, the two curves show similar patterns. The
pplimno displays an annual pattern linked to the
seasonality of rainfall. The ppxrf shows a pluriannual
oscillation pattern, possibly related to two factors: (1)
low sedimentation rate in Lake Sanabria and (2) the
‘‘low-pass filter’’ nature of most non-varved lakes,
filtering out the high-frequency climate oscillations.
Varved lakes are known to capture inter-annual
climate varibility because of their limnological
(seasonally stratified, anoxic hypolimnion), sedi-
mentological (non-bioturbated sediments, marked
seasonality, etc.) and geomorphological (deep and
wind-protected lakes) peculiarities. These features
are lacking in almost all non-varved sediments. This
probably accounts for the fact that this inter-annual
climate variability is smoothed out and therefore, not
recorded in the sediments. This ‘‘low-pass filter’’
nature of non-varved lakes has been recognized in
other water bodies such as Lake Gallocanta (Rodo
et al. 1997). Another explanation is that the sampling
step in the XRF measurements was too large to
capture higher-frequency environmental events.
The temporal offset between 1992 and 2005 may
be related to uncertainties in the age model. Despite
the excellent agreement between the TC curves in the
Fig. 10 Comparison of the
first eigenvector obtained in
the PCA applied to the
limnological dataset (darkthin solid line) with that
obtained in the PCA applied
to the XRF core scanner
dataset (light thick solidline)
498 J Paleolimnol (2011) 46:487–502
123
two cores, which allowed projection of 210Pb dates
from one core to the other, small correlation errors
may have been made. Furthermore, the 210Pb age
model plot (Fig. 5) shows that age uncertainties
become larger with core depth, which could have led
to the observed temporal offset. Both factors have a
positive feedback. Nevertheless, the maximum tem-
poral offset identified for the 13 years of the studied
period (1992–2005 AD) is about 1 year, i.e. *1.3
mm of sediment thickness.
This long-term fit between the pplimno and ppxrf
curves also permits us to model the first curve as a
function of the second one. The equation that relates
both reconstructions is:
pplim no ¼ 0:23þ 0:92 � ppxrf ð4Þ
where both variables do not have units. The percent
variance explained by Eq. (4) is low (9%, p \ 0.001),
but this could be related to the low-pass filter nature
of the XRF core scanner dataset. When comparing
the pplimno values derived from Eq. (3) with those
obtained by applying Eq. (4), there is a notable
difference in the range of the values. The values of
the pplimno curve from Eq. (3) range from -4 to 4,
whereas those from reconstructed pplimno using
the Eq. (4) vary from -1 to 1, suggesting that the
depositional processes in the lake integrate the
climate signal, but smooth out high-frequency
climate events (Fig. 10).
Quantitative reconstruction of the precipitation
of the Lake Sanabria region
Equations (2a, 2b, 3) and (4) show that it is possible
to trace how the climate signal from the atmosphere
is transferred to the lake water mass, and from there
to the sediments. This should allow us to invert such
relations and reconstruct quantitatively precipitation
oscillations using the first score of the PCA carried
out with XRF core scanner values. Substituting (4) in
(3) and the subsequent Eq. in (2b) permits us to
obtain an equation relating the instrumental precip-
itation values to the ppxrf. This general equation is:
precipitation ¼ 145:27þ 62:63 � ppfrx ð5Þ
where ppfrx does not have units and precipitation is
expressed in mm. Figure 11 shows precipitation
inferred using the EV1 from the PCA on the XRF core
scanner dataset (ppfrx) for the period 1959–2005 AD.
The comparison between reconstructed and instru-
mental rainfall shows similar patterns and although the
reconstructed precipitation is lower, most of the rainier
months are reflected in the reconstructed curve. The
reconstructed precipitation follows the long-term
pattern of the measured rainfall, smoothing out its
high-frequency components, but capturing the annual
variability. This is apparent if the reconstructed
precipitation is compared with the 5-month-smoothed
instrumental precipitation (Fig. 11).
Fig. 11 Quantitative reconstructed precipitation curve obtained
from the XRF core scanner values for the 1959–2007 AD period
(dark thick line). This quantitative reconstruction is compared
with the raw (dashed thin line) and 5-month-smoothed
(continuous thin line) instrumental meteorological rainfall
curves for the 1959–2005 AD period. The periods with no
instrumental meteorological data are highlighted with the grey
bands
J Paleolimnol (2011) 46:487–502 499
123
Conclusions
The [20-years monitoring dataset for Sanabria Lake
provided a unique opportunity to investigate the
relationship between recent depositional dynamics in
the lake and climate. The comparison of limnolog-
ical, hydrological and meteorological datasets with
the chemical composition of short cores from Lake
Sanabria illustrates the potential and pitfalls of
climate reconstructions based on high-resolution
sediment analyses. The Tera River water discharge
to Lake Sanabria is directly controlled by rainfall, and
to lesser extent by temperature. The best correlation
between the hydrological and meterological datasets
is obtained when no lag is introduced, indicating that
the hydrologic response of Lake Sanabria to rainfall
and temperature is immediate. PCA of the limnolog-
ical dataset showed that the largest variance is linked
to changes in nutrient input. Comparison of the first
eigenvector (EV) of this PCA with the Tera River
water discharge for the period 1992–2005 AD
demonstrates that the pattern of the wet/dry-season
conditions follows fluctuations in the hydrological
balance, and therefore, oscillations in precipitation.
The lag time between these two datasets is around
1 month. The PCA of the XRF core scanner
measurements showed that the main process control-
ling the chemical composition of the Lake Sanabria
sediments is related to primary productivity, and
therefore, to river water input. Comparison of the
precipitation reconstructed using the limnological
data with that obtained from the XRF core scanner
dataset indicated that the sediments act as a low-pass
filter, smoothing the climate (rainfall) signal.
Our analyses demonstrate that the climate signal is
transmitted from the atmosphere to the lake water
mass and, from there to the sediments. The climate
signal, however, is smoothed out by the sedimenta-
tion process and uncertainties in the age model cause
temporal offsets between measured and reconstructed
rainfall. This study highlights that two main limita-
tions might hamper application of the proposed
method: (1) lack of long and reliable limnological
datasets, and (2) lack of a robust core chronology that
minimizes temporal uncertainties related to the
reconstructed meteorological variable. In spite of
these problems, it was possible to establish links
between these datasets, and to obtain a quantitative
reconstruction of rainfall during the period
1959–2005 AD, which captured the main temporal
variability. This quantitative precipitation reconstruc-
tion suggests that it may be possible to obtain
accurate climate reconstructions using non-laminated
sediments.
Acknowledgments The Spanish Ministry of Science and
Innovation funded the research at Lake Sanabria through
the projects LIMNOCLIBER (REN2003-09130-C02-02/CLI),
CALIBRE (CGL2006-13327-C04/CLI), IBERLIMNO (CGL2004-
20236-E) and CONSOLIDER-GRACCIE (CSD2007-00067). The
Limnological Research Center and the Large Lakes Observatory
(University of Minnesota, USA) are acknowledged for technical
assistance with the ITRAX XRF-core scanner and the GEOTEK,
and Daniel R. Engstrom from St. Croix Watershed Research
Station (Science Museum of Minnesota, USA) for radiometric
dating of the SAN07-2 M core. We thank Javier Sigro and
Manola Brunet (Centre for Climate Change, University Rovira i
Virgili, Tarragona, Spain) for providing the precipitation and
temperature data. We are grateful to the ‘‘Parque Natural del Lago
de Sanabria y Alrededores’’ of the Consejerıa de Medio Ambiente
y Ordenacion del Territorio de la Junta de Castilla y Leon
(Environmental Council of Castilla and Leon Autonomous
Spanish Region) and owner of the Laboratorio de Limnologıa
del Parque Natural (Laboratory of Limnology of the Natural Park)
for the field and administrative facilities. Dr. Mark Brenner, Dr.
Pere Anadon and an anonymous referee are acknowledged for
their exhaustive reviews and very constructive comments that
greatly helped to improve this manuscript.
References
Abrantes F, Lopes C, Rodrigues T, Gil I, Witt L, Grimalt J,
Harris I (2009) Proxy calibration to instrumental data set:
implications for paleoceanographic reconstructions. Geo-
chem Geophy Geosy 10. doi: 10.1029/2009GC002604
Aldasoro JJ, De Hoyos C, Vega JC (1991) El lago de Sanabria
(Estudio Limnologico). Publicaciones Monograficas de la
Red de Espacios Naturales de Castilla y Leon. Consejerıa
de Medio Ambiente y Ordenacion del Territorio, 2. Di-
reccion General del Medio Ambiente, Valladolid (Spain)
APHA, AWWA, WPCF (1989) Standard methods for the
examination of water and wastewater, 17th edn. USA
Battarbee RW, Grytnes JA, Thompson R, Appleby PG, Catalan
J, Korhola A, Birks HJB, Heegaard E, Lami A (2002)
Comparing palaeolimnological and instrumental evidence
of climate change for remote mountain lakes over the last
200 years. J Paleolimnol 28:161–179
Battarbee RW, Monteith DT, Juggins S, Evans CD, Jenkins A,
Simpson GL (2005) Reconstructing pre-acidification pH
for an acidified Scottish loch: a comparison of palaeo-
limnological and modelling approaches. Environ Pollut
137:135–149
Bennion H, Johnes P, Ferrier R, Phillips G, Haworth E (2005)
A comparison of diatom phosphorous transfer functions
and export coefficient models as tools for reconstructing
lake nutrient histories. Freshw Biol 50:1651–1670
500 J Paleolimnol (2011) 46:487–502
123
Bigler C, Hall RI (2002) Diatoms as indicators of climatic and
limnological change in Swedish Lapland: a 100-lake
calibration set and its validation for paleoecological
reconstructions. J Paleolimnol 27:97–115
Blass A, Bigler C, Grosjean M, Sturm M (2007) Decadal-scale
autumn temperature reconstruction back to AD 1580
inferred from the varved sediments of Lake Silvaplana
(southeastern Swiss Alps). Quat Res 68:184–195
Brunet M, Saladie O, Jones P, Sigro J, Aguilar E, Moberg A,
Walther A, Lister D, Lopez D, Almarza C (2006) The
development of a new daily adjusted temperature dataset
for Spain (1850–2003). Int J Climatol 26:1777–1802
Cacho I, Grimalt JO, Canals M, Sbaffi L, Shackleton NJ,
Schonfeld J, Zahn R (2001) Variability of the western
Mediterranean Sea surface temperature during the last
25,000 years and its connection with the Northern Hemi-
sphere climatic changes. Paleoceanography 16:40–52
Carrion JS, Fernandez S, Gonzalez-Samperiz P, Gil-Romera G,
Badal E, Carrion-Marco Y, Lopez-Merino L, Lopez-Saez
JA, Fierro E, Burjachs F (2010) Expected trends and
surprises in the Lateglacial and Holocene vegetation his-
tory of the Iberian Peninsula and Balearic Islands. Rev
Palaeobot Palynol 162:458–475
Catalan J, Pla S, Garcia J, Camarero L (2009) Climate and CO2
saturation in an alpine lake throughout the Holocene.
Limnol Oceanogr 54:2542–2552
Chen FH, Chen JH, Holmes J, Boomer I, Austin P, Gates JB,
Wang NL, Brooks SJ, Zhang JW (2010) Moisture changes
over the last millennium in arid central Asia: a review,
synthesis and comparison with monsoon region. Quat Sci
Rev 29:1055–1068
Cowton T, Hughes PD, Gibbard PL (2009) Paleoglaciation of
Parque Natural Lago de Sanabria, northwest Spain. Geo-
morphology 108:282–291
Danis PA, Masson-Delmotte V, Stievenard M, Guillemin MT,
Daux V, Naveau Ph, von Grafenstein U (2006) Recon-
struction of past precipitation d18O using tree-ring cellu-
lose d18O and d13C: a calibration study near Lac
d’Annecy, France. Earth Planet Sci Lett 243:439–448
De Hoyos C (1996) Limnologıa del Lago de Sanabria: vari-
abilidad interanual del fitoplancton. Unpublished Ph.D.
Thesis. Universidad de Salamanca, 438 pp
Dienert F, Wandenbulcke F (1923) Sur le dosage de la silice
dans les eaux. C R Acad Sci 176:1478–1480
Djamali M, de Beaulieu JL, Andrieu-Poniel V, Berberian M,
Miller NF, Gandouin E, Lahijani H, Shah-Hosseini M,
Ponel P, Salimian M, Guiter F (2009) A late Holocene
pollen record from Lake Almalou in NW Iran: evidence
for changing land-use in relation to some historical events
during the last 3,700 years. J Archaeol Sci 36:1364–1375
Francis DR, Wolfe AP, Walker IR, Miller GH (2006) Inter-
glacial and Holocene temperature reconstructions based
on midge remains in sediments of two lakes from Baffin
Island, Nunavut, Arctic Canada. Palaeogeogr Palaeocli-
matol Palaeoecol 236:107–124
Fritz SC (1990) Twentieth-century salinity and water-level
fluctuations in Devils Lake, North Dakota: test of a dia-
tom-based transfer function. Limnol Oceanogr 35:
1771–1781
Granados I, Toro M (2000) Recent warming in a high mountain
lake (Laguna Cimera, Central Spain) inferred by means of
fossils chironomids. In: Lami A, Cameron N, Korhola A
(eds) Paleolimnology and ecosystem dynamics at remote
European Alpine lakes. J Limnol 59(Suppl 1):109–119
Guiot J, de Vernal A (2007) Tranfer functions methods for
quantitative paleoceanography based on microfossils. Dev
Mar Geol 1:523–563
Hyodo F, Tsugeki N, Azuma J, Urabe J, Nakanishi M, Wada E
(2008) Changes in stable isotopes, lignin-derived phenols,
and fossil pigments in sediments of Lake Biwa, Japan:
implications for anthropogenic effects over the last
100 years. Sci Total Environ 403:139–147
Johnson KR, Ingram BL (2004) Spatial and temporal vari-
ability in the stable isotope systematics of modern pre-
cipitation in China: implications for paleoclimate
reconstructions. Earth Planet Sci Lett 220:365–377
Kalugin I, Daryin A, Smolyaninova L, Andreev A, Diekmann
B, Khlystov O (2007) 800-yr long records of annual air
temperature and precipitation over southern Siberia
inferred from Teletskoye Lake sediments. Quat Res 67:
400–410
Kattel GR, Battarbee RW, Mackay AW, Birks HJB (2008)
Recent ecological change in a remote Scottish mountain
loch: an evaluation of a Cladocera-based temperature
transfer-function. Palaeogeogr Palaeoclimatol Palaeoecol
259:51–76
Larocque I, Finsinger W (2008) Late-glacial chironomid-based
temperature reconstructions for Lago Piccolo di Avigliana
in the southwestern Alps (Italy). Palaeogeogr Palaeocli-
matol Palaeoecol 257:207–223
Luque JA (2003) El Lago de Sanabria: un sensor de las osci-
laciones climaticas del Atlantico Norte durante los ulti-
mos 6.000 anos. El registro de la Pequena Edad del Hielo
(LIA), los eventos Dansgaard-Oeschger (D-O), y el ciclo
de Bond. Unpublished Ph.D. Thesis. University of Bar-
celona, 446 pp
Luque JA, Julia R (2002) Lake sediment response to land-use
and climate change during the last 1000 years inthe oli-
gotrphic Lake Sanabria (northwest of Iberian Peninsula).
Sediment Geol 148:343–355
Mangini A, Spotl C, Verdes P (2005) Reconstruction of tem-
perature in the Central Alps during the past 2000 yr from a
d18O stalagmite record. Earth Planet Sci Lett 235:741–751
Martınez-Garcıa E (1973) Deformacion y metamorfismo en la
zona de Sanabria. Studia Geologica Salmanticensia 5:
7–106
Martın-Puertas C, Valero-Garces BL, Mata MP, Gonzalez-
Samperiz P, Bao R, Moreno A, Stefanova V (2008) Arid
and humid phases in southern Spain during the last
4000 years: the Zonar Lake record, Cordoba. Holocene 18:
907–921
Morris AW, Riley JP (1963) The determination of nitrate in
sea-water. Anal Chim Acta 29:272–279
Murphy J, Riley IP (1962) A modified single solution method
for the determination of phosphate in natural waters. Anal
Chim Acta 27:31–36
Nichols JE, Walcott M, Bradley R, Pilcher J, Huang Y (2009)
Quantitative assessment of precipitation seasonality and
summer surface wetness using ombrotrophic sediments
from an Arctic Norwegian peatland. Quat Res 72:443–451
Pla S, Catalan J (2005) Chrysophyte cysts from lake sediments
reveal the submillennial winter/spring climate variability
J Paleolimnol (2011) 46:487–502 501
123
in the northwestern Mediterranean region throughout the
Holocene. Clim Dyn 24:263–268
R Development Core Team (2010) R: a language and envi-
ronment for statistical computing. R Foundation for Sta-
tistical Computing, Vienna
Robbins A, Edgington DN (1975) Determination of recent
sedimentation rates in Lake Michigan using Pb-210 and
Cs-137. Geochim Cosmochim Acta 39:285–304
Rodo X, Baert E, Comın FA (1997) Variations in seasonal
rainfall in Southern Europe during the present century:
relationships with the North Atlantic Oscillation and the
El Nino-Southern Oscillation. Clim Dyn 13:275–284
Romero-Viana L, Julia R, Camacho A, Vicente E, Miracle MR
(2008) Climate signal in varve thickness: Lake La Cruz
(Spain), a case study. J Paleolimnol 40:703–714
Saez A, Valero-Garces BL, Giralt S, Moreno A, Bao R, Pueyo
JJ, Hernandez A, Casas D (2009) Glacial to Holocene
climate changes in the SE Pacific. The Raraku Lake
sedimentary record (Easter Island, 27�S). Quat Sci Rev
28:2743–2759
Schwalb A, Dean WE (2002) Reconstruction of hydrological
changes and response to effective moisture variations
from North-Central USA lake sediments. Quat Sci Rev
21:1541–1554
SCOR-UNESCO (1966) Determination of photosynthetic pig-
ments in sea water. UNESCO Monographs on Oceano-
graphic Methodology 1. Paris
Striewski B, Mayr C, Flenley J, Naumann R, Turner G, Lucke
A (2009) Multi-proxy evidence of late Holocene human-
induced environmental changes at Lake Pupuke, Auck-
land (New Zealand). Quat Int 202:69–93
Tonello MS, Mancini MV, Seppa H (2009) Quantitative
reconstruction of Holocene precipitation changes in
southern Patagonia. Quat Res 72:410–420
Toro M, Granados I, Robles S, Montes C (2006) High moun-
tain lakes of the Central Range (Iberian Peninsula):
regional limnology & environmental changes. Limnetica
25:217–252
Trachsel M, Eggenberger U, Grosjean M, Blass A, Sturm M
(2008) Mineralogy-based quantitative precipitation and
temperature reconstructions from annually laminated lake
sediments (Swiss Alps) since AD 1580. Geophys Res Lett
35:L13707. doi:10.1029/2008GL034121
Vega JC, Aldasoro JJ (1994) Geologıa de Sanabria. In: Mo-
nografıas de la Red de Espacios Naturales de Castilla y
Leon. Consejerıa de Medio Ambiente y Ordenacion del
Territorio (Junta de Castilla y Leon), Valladolid
Vega JC, De Hoyos C, Aldasoro JJ (1991) Estudio del sistema
de lagunas de las sierras Segundera y Cabrera. In: Mo-
nografıas de la Red de Espacios Naturales de Castilla y
Leon, Consejerıa de Medio Ambiente y Ordenacion del
Territorio (Junta de Castilla y Leon), Valladolid
Vega JC, De Hoyos C, Aldaroso JJ (1992) The Sanabria Lake:
the largest natural freshwater lake in Spain. Limnetica
8:49–57
Vega JC, De Hoyos C, Aldasoro JJ, Fraile H (2005) Nuevos
datos morfometricos para el lago de Sanabria. Limnetica
24:115–122
Venables WN, Ripley BD (2002) Modern applied statistics
with S. Springer Science ? Business Media, LLC, New
York
502 J Paleolimnol (2011) 46:487–502
123
