rank: NA

scope: climate

basis: The mean annual cycles of d18O and the Sr/Ca ratios suggest temperature amplitudes of only 40-50% of the 8.4i¿½C in recent temperatures. The results indicated by the proxies appear unrealistically low, and we believe that the data sets do not capture the full annual amplitude, despite a sampling resolution of more than 12 samples per annual skeletal growth increment.

direction: decrease

interpDirection: decrease

scope: climate

seasonality: subannual

variable: temperature

variableDetail: sea surface

variableGroup: Temperature

variableGroupDirection: negative

variableGroupOriginal: T

variableOriginal: temperature

scope: climate

scope: climate

basis: This coral was not directly calibrated since the top is not modern. They used other calibrations (Porites), and not from Diploria labyrinthiformis. Using more recent Goodkin calibration would be applicable to this coral. "we used calibrations from the Atlantic and Pacific Oceans. For estimating temperatures from coral skeletal d18O we used the gradient of the temperature to N18O equation of Leder et al. (1996) obtained from Florida corals (4.52i¿½C per mil in d18O).

coefficient: NA

direction: negative

fraction: NA

inferredMaterial: seawater

integrationTime: ~15 samples/year

integrationTimeBasis: The top and base of the core section were U/Th-dated (measured at the Institut fur Umweltphysik, Heidelberg, Germany), with a precision of �15 yr. The midpoint of the top dated sample is 14 density band couplets above the profile for geochemical measurements, and the midpoint of the bottom dated sample is seven band couplets below the profile (Fig. 2). In other words, the age model within the chronology is based on the annual banding. According to band couplet counts and the number of annual cycles in the d18O, the U/Th-dated sections differ in age by 104 yr, while the difference is 130 yr according to their radiometric dates (base: 1499, top: 1629 AD). To remove the surplus of 26 yr we adjusted the radiometric ages by adding 13 yr to the bottom and subtracting 13 yr from the top sample age. Therefore, individual years in our record are correct with respect to their positions within the time series, but they do not necessarily match calendar years. Years in the isotopic record are defined by cyclic variations in d18O, which are assumed to be annual, with each cycle spanning ca. 2.5-4 mm core length. Subannual dating is based on the annual d18O cycles, where the most positive d18O value in each cycle is assigned to the coldest average month, which, for Bermuda, is March. This introduces a non-cumulative error of approximately �1 month. The ages of other sampling points were linearly interpolated according to their spatial distances along the sampling profile.

integrationTimeUncertainty: 15 years interannual and 1 month subannual

integrationTimeUncertaintyType: U-Th dating uncertainty

integrationTimeUnits: Varies needs to be interpolated

mathematicalRelation: linear

rank: 1

scope: isotope

seasonality: subannual

variable: temperature

variableGroup: Temperature

variableGroupDirection: negative

variableGroupOriginal: T_water

coefficient: NA

fraction: NA

rank: NA

scope: isotope

coefficient: NA

fraction: NA

rank: NA

scope: isotope

rank: NA

scope: climate

direction: decrease

interpDirection: decrease

scope: climate

seasonality: subannual

variable: temperature

variableDetail: sea@surface

variableDetailOriginal: sea_surface

variableGroup: Temperature

variableGroupDirection: negative

variableGroupOriginal: T

variableOriginal: temperature

scope: climate

scope: climate

coefficient: NA

fraction: NA

inferredMaterial: seawater

rank: NA

scope: isotope

coefficient: NA

fraction: NA

rank: NA

scope: isotope

coefficient: NA

fraction: NA

rank: NA

scope: isotope