ENSO Signals in the Coral 18O Record from Micronesia

As expected from seasonal variations in SST and SSS (Fig. 1), annual periodicity was clearly observed in S18Oc profiles of the Chuuk and Pohnpei corals (Fig. 4). We employed a coral chronology based on peak matching between skeletal S18Oc and instrumental SST with simple linear interpolation; the minima and maxima in the annual S18Oc curves were designated as the highest and lowest annual SSTs (Fig. 7). Comparison of the monthly S18Oc profile of Chuuk coral CHU99-01 with times series of environmental variables (1980-1999) revealed a modal change in annual S18Oc variation relating to the ENSO cycle (Fig. 7). During non-El Niño periods, a combination of warm SSTs and high rainfall in the wet season (June-November) causes coral S18Oc values to decrease, whereas the cooler dry season (December-May) produces higher S18Oc values. On the other hand, pronounced increases in coral S18Oc are evident during El Niño maxima,

Figure 6: Time-series of instrumental SSTs (gray line) and estimated SST based on coral S18Oc records (black line with open circles) from Ishigaki Island, southern Japan. A 10-day mean SST record at Ishigaki Port conducted by the JMA Ishigaki-jima Meteorological Observatory is shown for comparison with the coral d 8Oc record. SSTs were estimated using equations (1), (2), and (3) for IU96-07, IS91-06, and IY99-01, respectively. The slope of the 818Oc-SST relationship is shown in each profile. The coral time-series is plotted using linear interpolation between the minima and the maxima of the coral 8T8Oc estimates of SST.

Figure 6: Time-series of instrumental SSTs (gray line) and estimated SST based on coral S18Oc records (black line with open circles) from Ishigaki Island, southern Japan. A 10-day mean SST record at Ishigaki Port conducted by the JMA Ishigaki-jima Meteorological Observatory is shown for comparison with the coral d 8Oc record. SSTs were estimated using equations (1), (2), and (3) for IU96-07, IS91-06, and IY99-01, respectively. The slope of the 818Oc-SST relationship is shown in each profile. The coral time-series is plotted using linear interpolation between the minima and the maxima of the coral 8T8Oc estimates of SST.

corresponding to cooler SSTs and drought conditions around Chuuk Atoll. Similar climate variations are recorded by the skeletal S18Oc of corals living within the WPWP, particularly those from the north coast of Papua New Guinea (Tudhope et al., 1995, 2001; McGregor and Gagan, 2004), North Sulawesi, Indonesia (Moore et al., 1997; Hughen et al., 1999), and Fiji (Le Bec et al., 2000). In detail, the amplitude of the S18Oc anomaly corresponds well to El Nino intensity, as indicated by SOI values; the unusually strong El Nino event of 1982-1983, which had widespread ecological and economic impact, shows the largest anomaly in the S18Oc profile (Fig. 7).

Coral S18Oc values of CHU99-01 showed moderate correlation with both SST and SSS, reflecting that SST and SSS in the region fluctuate with

rs/

/ VA.

V

1981 1983 1985 1987 1989 1991 1993 1995 1997 1999

1981 1983 1985 1987 1989 1991 1993 1995 1997 1999

Figure 7: Comparison of skeletal S18Oc of the Chuuk coral and instrumental records of climate for the period between 1980-99. Rainfall was observed at a NOAA station (http://lumahai.soest.hawaii.edu/Enso/data/ftp/ STATION-RAIN2/chuuk-sum.txt), and the SOI is a standardized sea-level pressure difference (http://ingrid.ldgo.columbia.edu/SOURCES/.Indices/.soi/ .standardized/). See text for other data sources. No filter was applied to isotope data. Double arrows indicate the distinct increase in S18Oc during El Nino years, while single arrows indicate compression of S18Oc curves in the year following the El Nino peak. Note that 818Oc and SSS is reverse-plotted so that El Nino-related anomalies vary in the same direction as the other parameters. Time-series of monthly rainfall and SOI are smoothed by a 3-month and 5-month moving average, respectively. El Nino events are shown by shading according to the classification by the NCEP/CPC (http: //www.cpc.ncep.noaa.gov/products/analysis_monitoring/ensostuff/ensoyears. html).

comparable amplitude with respect to the coral S18Oc signal (Fig. 8). Although when all S18Oc data were plotted against SST, the slope of the regression line was different from that of the temperature dependency relationship for coral aragonite (0.13%o-0.18%o °C_1; Fig. 8); the discrepancy can be attributed to changes of 818OW in seawater during the ENSO cycle. Coral S18Oc values corresponding to El Nino events showed a shift toward higher 818O values owing to 18O enrichment of the seawater.

OQH

i , i

SgSPaiP o

» o

-5.6 -

e

o^Sf? o°

• Strong El Nino

\ 0

-5.2

r = 0.67

Figure 8: Crossplots of coral 818Oc and monthly observations of SSS (A) and SST (B) at Chuuk Atoll. Monthly coral S18Oc time-series were resampled using AnalySeries software (Paillard et al., 1996). Months categorized as reflecting strong El Nino conditions by the NCEP/CPC are plotted as closed circles. SSS - 41.7+1.30S180c, R2 - 0.35, P<0.01; SST - 16.9-2.11 S180c, R2 - 0.45, P<0.01. The gray bar in panel B represents S180c-SST slope for coral aragonite (-0.13% to -0.18% °C-1).

The intra-annual S18Oc fluctuations were well recorded by high-resolution microsampling, so the temporal evolution of each El Nino event can be discussed using coral records. El Nino events are usually characterized by initial cooling of SSTs in the tropical western Pacific, followed by lower than average rainfall. A distinct positive shift in S18Oc in the year of a strong El Nino event can be attributed to this initial cooling. The skeletal S18Oc values for the year following a strong El Nino event do not vary much, suggesting the persistence of cool and/or high-salinity conditions for almost a year (Fig. 7). Although the lower SST can be attributed primarily to the eastward migration of the WPWP during the El Nino event, the shallowing of the therm-ocline brought about by relaxation of the trade winds across the Pacific basin is also important for surface cooling. High-salinity conditions in the year following a strong El Nino event also appear to be caused by the shallowing of the thermocline, along with the direct effect of the decrease in rainfall locally. Because rainfall and SST data quickly recovered to normal levels just after the El Nino peak, salt advection by the North Equatorial Counter Current and/or subsurface mixing due to transformation of a "barrier layer'' (Lukas and Lindstrom, 1991) seems to account for the positive S180c anomalies in the years following the El Nino peak.

The combination of a positive shift in S180c followed by compression of the S18Oc curve, signals an El Nino event throughout the region because these features were also seen in the Pohnpei Island coral S180c record (Fig. 4). The coral S18Oc record also shows a potential for quantitative examination of the individual characteristics and temporal evolution of each past El Nino event. Such an examination would improve our understanding of the ENSO cycle, although additional measurements of the skeletal Sr/Ca ratio, an independent proxy for SST, are desirable for a more precise reconstruction. Further S18Oc measurements on longer cores would allow us to explore changes in the frequency of El Nino events over the past several centuries and the effect of global warming on the ENSO cycle.

Was this article helpful?

0 0

Post a comment