coefficient: NA

fraction: NA

rank: NA

scope: climate

coefficient: NA

fraction: NA

rank: NA

scope: climate

rank: NA

scope: climate

coefficient: NA

fraction: NA

rank: NA

scope: climate

rank: NA

scope: climate

direction: decrease

interpDirection: decrease

scope: climate

seasonality: subannual

variable: temperature

variableDetail: sea surface

variableGroup: temperature and precipitation

scope: climate

scope: climate

basis: Hetzinger 2006 calibration and general d18O to SST slope for corals used (-0.2)

coefficient: -0.2

direction: negative

fraction: 0.3364

inferredMaterial: seawater

integrationTime: 1

integrationTimeBasis: Chronology was developed on the basis of the seasonal cycle in coral ?18O, and by counting the well-developed annual density bands. The coral ?18O-record extends from 1918 to 2004. The measured skeletal d18O minimum (maximum) was assigned to the month September (February) which represents the average seasonal SST maximum (minimum) at the study site. Linear interpolation was used in order to obtain monthly resolution for statistical analysis.

integrationTimeUncertainty: ~1?2 months in any given year, no errors given for annual chronology

integrationTimeUncertaintyType: chronological

integrationTimeUnits: month

mathematicalRelation: linear

rank: 1

scope: isotope

seasonality: subannual

variable: temperature

variableGroup: Temperature

variableGroupDirection: negative

variableGroupOriginal: T_water

basis: However, the magnitude of the multidecadal variations in coral d18O is larger than expected based on SST alone. The standard deviation of the smoothed SST index shown in Figure 1B is 0.12i¿½C (based on SST from the HadISSTv.1.1 database; Rayner et al., 2003), while the standard deviation of coral !18O is 0.06?, which would correspond to 0.3 i¿½C, assuming well-established relationships. Hence, ~50% of the amplitude is due to variations in d18Oseawater. The d18Oseawater contribution is estimated by calculating the d18Oresidual (i.e., by subtracting the SST component from measured coral d18O; for details see GSA Data Repository). Figure 1C compares the d18Oresidual with a regional precipitation index computed from weather stations in the southeastern Caribbean, which includes the study site. The d18Oresidual and precipitation are highly correlated at low frequencies, implying that multidecadal d18Oseawater variations are primarily atmosphere-driven (r = ?0.46; 1923?1998; and r = ?0.67 for detrended values).

coefficient: NA

direction: negative

fraction: NA

inferredMaterial: seawater

integrationTime: 1

integrationTimeBasis: Chronology was developed on the basis of the seasonal cycle in coral ?18O, and by counting the well-developed annual density bands. The coral ?18O-record extends from 1918 to 2004. The measured skeletal d18O minimum (maximum) was assigned to the month September (February) which represents the average seasonal SST maximum (minimum) at the study site. Linear interpolation was used in order to obtain monthly resolution for statistical analysis.

integrationTimeUncertainty: ~1?2 months in any given year, no errors given for annual chronology

integrationTimeUncertaintyType: chronological

integrationTimeUnits: month

mathematicalRelation: linear

rank: 2

scope: isotope

variable: seawaterIsotope

variableGroup: EffectiveMoisture

variableGroupDirection: negative

variableGroupOriginal: d18O_seawater

coefficient: NA

fraction: NA

rank: NA

scope: isotope