namely photosynthesis and heterotrophy. Firstly, as
light levels decrease due to cloud cover, the rate of
photosynthesis by the coral’s symbiotic zooxanthellae decreases, and skeletal d
13 C decreases. The
reverse occurs when light levels increase. Secondly,
zooplankton have a low d
13 C value relative to coral.
During upwelling events in the Red Sea, nutrient and
zooplankton level increases have been linked to decreases in coral skeletal d
13 C values. Other upwelling
tracers include cadmium (Cd) and barium (Ba) concentrations, and D
14 C. Cadmium and barium are
trace elements whose concentrations are greater in
deep water than in surface water. During upwelling
events, deep water is driven to the surface and cadmium/calcium (Cd/Ca) and barium/ calcium (Ba/Ca)
ratios in the surface water, and consequently in the
coral skeleton, increase (Table 1). Although SST also
influences Ba/Ca ratios, most of the variation in Ba/
Ca ratios in corals is due to nutrient fluxes and upwelling. D
14 C is also an excellent tracer for detecting
upwelling and changes in seawater circulation (D
14
C
is the per mil deviation of the ratio of
14 C/
12
C relative to a nineteenth century wood standard). For
example, in the eastern equatorial Pacific Ocean, the
D
14 C value of deep water tends to be very low relative to the D
14 C of surface water. Here, increased
upwelling or increases in the proportion of deep
water contributing to surface water results in a decrease in the D
14 C of the coral skeleton. Manganese
(Mn) is a trace element whose concentration is
highest in surface waters and decreases with depth.
Therefore, during upwelling events, Mn/Ca ratios
decrease. The ratio of Mn/Ca can also record prolonged and sustained changes in winds. In at least
one case, Mn/Ca ratios from a Tarawa Atoll coral
increased during El Nin ˜ o events as a result of strong
and prolonged wind reversals that had remobilized
manganese from the lagoon sediments.
Other Proxy Indicators
Other environmental parameters that can be inferred
from coral skeleton structure and composition are
river outflow (fluorescence bands) and pH (boron
isotope levels) (Table 1). Large pulses in river outflow can result in an ultraviolet-sensitive fluorescent
band in the coral record. New evidence strongly
suggests that the fluorescent patterns in coral skeletal
records are due to changes in skeletal density, not
terrestrially derived humics as previously thought.
Variations in salinity associated with fresh water
discharge pulses from rivers appear to cause changes
in coral skeletal growth density which can be observed in the skeletal fluorescence pattern. In the case
of boron, d
11 B levels in sea water increase as pH
increases. Changes in the pH at the site of coral
calcification seem to reflect changes in productivity
of the symbiotic zooxanthellae. As photosynthesis
increases, pH increases, and d
11 B levels in the coral
skeleton increase.
Coral Records: What has been
Learned About Climate From Corals?
To date, there are over 100 sites where coral cores
have been recovered and analyzed (Figure 3). Of
these, at least 22 have records that exceed 120 years
in length. In most cases, d
13 C and d
18 O have been
measured at annual-to-subannual resolution. d
18
O as
a SST and/or SSS proxy is the best understood and
most widely reported of all the coral proxy measurements. In a few cores, other isotopic, trace, and
minor elements, or skeletal density and growth
measurements have also been made.
Typically, coral-derived paleoclimate records are
studied on three timescales: seasonal, interannual-todecadal, and long-term trends. The seasonal variation refers to one warm and one cool phase each
year. An abrupt shift in the proxy’s long-term mean
often indicates a decadal modulation in the data.
Long-term trends are usually associated with a
gradual increase or decrease in the measured proxy
over the course of several decades or centuries. The
following sections explore some of the seasonal,
decadal and long-term trends in coral-derived
paleoclimate records and some of the limitations
associated with interpreting coral proxy records.
Seasonal Variation in Coral Climate Records
Seasonal variation accounts for the single largest
percentage of the variance in most coral isotope,
trace, and minor element records. In regions such as
Japan and the Gala ´ pagos with distinct SST seasonality, the annual periodicity in d
18
O, Sr/Ca, Mg/Ca,
and U/Ca is pronounced. In regions heavily affected
by monsoonal rains such as Tarawa Atoll, the seasonal variation in d
13 C is regular and pronounced.
D
14
C also has an annual periodicity in coral from
regions with a strong seasonal upwelling regime such
as is seen in the Gala ´ pagos (see next section). The
strength and duration of the upwelling season is reflected in the length and degree of D
14
C decrease in
the coral record.
Interannual-to-decadal Variation in Climate and El
Nin ˜ o Southern Oscillation (ENSO)
The second largest component of the variance in
Pacific coral isotope records is associated with the
interannual-to-decadal variation in the El Nin ˜ o
426 PAST CLIMATE FROM CORALS
light levels decrease due to cloud cover, the rate of
photosynthesis by the coral’s symbiotic zooxanthellae decreases, and skeletal d
13 C decreases. The
reverse occurs when light levels increase. Secondly,
zooplankton have a low d
13 C value relative to coral.
During upwelling events in the Red Sea, nutrient and
zooplankton level increases have been linked to decreases in coral skeletal d
13 C values. Other upwelling
tracers include cadmium (Cd) and barium (Ba) concentrations, and D
14 C. Cadmium and barium are
trace elements whose concentrations are greater in
deep water than in surface water. During upwelling
events, deep water is driven to the surface and cadmium/calcium (Cd/Ca) and barium/ calcium (Ba/Ca)
ratios in the surface water, and consequently in the
coral skeleton, increase (Table 1). Although SST also
influences Ba/Ca ratios, most of the variation in Ba/
Ca ratios in corals is due to nutrient fluxes and upwelling. D
14 C is also an excellent tracer for detecting
upwelling and changes in seawater circulation (D
14
C
is the per mil deviation of the ratio of
14 C/
12
C relative to a nineteenth century wood standard). For
example, in the eastern equatorial Pacific Ocean, the
D
14 C value of deep water tends to be very low relative to the D
14 C of surface water. Here, increased
upwelling or increases in the proportion of deep
water contributing to surface water results in a decrease in the D
14 C of the coral skeleton. Manganese
(Mn) is a trace element whose concentration is
highest in surface waters and decreases with depth.
Therefore, during upwelling events, Mn/Ca ratios
decrease. The ratio of Mn/Ca can also record prolonged and sustained changes in winds. In at least
one case, Mn/Ca ratios from a Tarawa Atoll coral
increased during El Nin ˜ o events as a result of strong
and prolonged wind reversals that had remobilized
manganese from the lagoon sediments.
Other Proxy Indicators
Other environmental parameters that can be inferred
from coral skeleton structure and composition are
river outflow (fluorescence bands) and pH (boron
isotope levels) (Table 1). Large pulses in river outflow can result in an ultraviolet-sensitive fluorescent
band in the coral record. New evidence strongly
suggests that the fluorescent patterns in coral skeletal
records are due to changes in skeletal density, not
terrestrially derived humics as previously thought.
Variations in salinity associated with fresh water
discharge pulses from rivers appear to cause changes
in coral skeletal growth density which can be observed in the skeletal fluorescence pattern. In the case
of boron, d
11 B levels in sea water increase as pH
increases. Changes in the pH at the site of coral
calcification seem to reflect changes in productivity
of the symbiotic zooxanthellae. As photosynthesis
increases, pH increases, and d
11 B levels in the coral
skeleton increase.
Coral Records: What has been
Learned About Climate From Corals?
To date, there are over 100 sites where coral cores
have been recovered and analyzed (Figure 3). Of
these, at least 22 have records that exceed 120 years
in length. In most cases, d
13 C and d
18 O have been
measured at annual-to-subannual resolution. d
18
O as
a SST and/or SSS proxy is the best understood and
most widely reported of all the coral proxy measurements. In a few cores, other isotopic, trace, and
minor elements, or skeletal density and growth
measurements have also been made.
Typically, coral-derived paleoclimate records are
studied on three timescales: seasonal, interannual-todecadal, and long-term trends. The seasonal variation refers to one warm and one cool phase each
year. An abrupt shift in the proxy’s long-term mean
often indicates a decadal modulation in the data.
Long-term trends are usually associated with a
gradual increase or decrease in the measured proxy
over the course of several decades or centuries. The
following sections explore some of the seasonal,
decadal and long-term trends in coral-derived
paleoclimate records and some of the limitations
associated with interpreting coral proxy records.
Seasonal Variation in Coral Climate Records
Seasonal variation accounts for the single largest
percentage of the variance in most coral isotope,
trace, and minor element records. In regions such as
Japan and the Gala ´ pagos with distinct SST seasonality, the annual periodicity in d
18
O, Sr/Ca, Mg/Ca,
and U/Ca is pronounced. In regions heavily affected
by monsoonal rains such as Tarawa Atoll, the seasonal variation in d
13 C is regular and pronounced.
D
14
C also has an annual periodicity in coral from
regions with a strong seasonal upwelling regime such
as is seen in the Gala ´ pagos (see next section). The
strength and duration of the upwelling season is reflected in the length and degree of D
14
C decrease in
the coral record.
Interannual-to-decadal Variation in Climate and El
Nin ˜ o Southern Oscillation (ENSO)
The second largest component of the variance in
Pacific coral isotope records is associated with the
interannual-to-decadal variation in the El Nin ˜ o
426 PAST CLIMATE FROM CORALS
