Figure 1.
Locations of study sites in the Similan Islands, Andaman Sea, Thailand.
Symbols show locations of temperature loggers and growth rate experiments (▴) and of bioerosion experiments (▪). (Figure inset courtesy A. Buschmann, AWI. Data source: http://www.ngdc.noaa.gov/mgg/shorelines/gshhs.html).
Figure 2.
Experimental design of Porites lutea growth rate experiment.
For the experiment in each case 20 coral nubbins were collected from 10 mother colonies at a sheltered east site in 7(close to E 8.1, cf. Fig. 1) and transplanted to the west and east sides of 5 islands at 2 depths (7 m and 20 m) from February 2007 to February 2008 along the Similan Islands (cf. Fig. 1).
Figure 3.
Temperature anomalies at Similan island Ko Miang.
Study sites E 4.1 and W 4.1, period March 2007 to November 2008. Anomalies were calculated relative to mode values (gap in data due to exceeded storage capacity).
Figure 4.
Temperature data for all Similan Islands sites.
Central tendency box plots (median with 25th and 75th percentile and non-outlier range) with extremes (dots) (A, B) and scatter plots with standard deviation (SD, whiskers) for the study periods [Feb 2007 to Feb 2008 (2007) and Nov 2007 to Nov 2008 (2008)], sides (east, E and west, W) and depths (7 and 20 m), featuring overall monthly minimum, mode and maximum temperatures (A), monthly temperature ranges (B), and negative cumulative temperature anomalies (calculated as degree days, °C d; calculation details see methods) (C).
Table 1.
Differences in monthly temperature values at Similan Islands between depths and sides.
Figure 5.
Coral growth rates (mass change of calcium carbonate, CaCO3) of Porites lutea nubbins.
Nubbins were transplanted from sheltered east (E) 7 m to 7 and 20 m depth at E and west (W) sides of Similan Islands (Central tendency box plots: median with 25th and 75th percentile and non-outlier range, extremes: dots) between February 2007 and February 2008 (number of replicates: E 7 m: 27, E 20 m: 33, W 7 m: 13, W 20 m: 19).
Table 2.
Analysis of covariance (ANCOVA) for coral growth experiment.
Figure 6.
Calcium carbonate (CaCO3) mass change accretion (>0) and bioerosion (<0).
Measured mass change (white columns) and calculated mass change (grey columns, corrected for accretion due to fouling organisms) of CaCO3 blocks at east (E) and west (W) side of Similan island Ko Miang. Error bars: ±1 SE of mean. Results from 12 months exposure (February 2007 to February 2008) (A), and 21 months exposure (February 2007 to November 2008) (B) normalized to 1 year (number of replicates in each case: 9).
Table 3.
Analysis of covariance (ANCOVA) for bioerosion experiment.
Figure 7.
Coral growth and bioerosion as functions of temperature anomaly in west (W) and east (E), 7 m and 20 m depth.
Coral growth rate of Porites lutea nubbins at all study sites (n = 20, df = 1, F-value = 4.4, r2 = 0.46, p = 0.039) along the Similan island chain after 12 months exposure (February 2007 to February 2008; see Fig. 4) (A), and bioerosion (calculated mass change in calcium carbonate (CaCO3) corrected for accretion due to fouling organisms) on CaCO3 blocks at the central Similan island Ko Miang (see Fig. 1) after 12 months (grey depiction; February 2007 to February 2008; n = 4, degrees of freedom = 1, F-value = 14.4, r2 = 0.30, p<0.001) and 21 months exposure normalized to 1 year (February 2007 to November 2008; n = 4, degrees of freedom = 1, F-value = 17.9, r2 = 0.58, p<0.001) (see Fig. 5; <0, bioerosion) (B) as functions (GLM) of the cumulative negative temperature anomaly (calculated as °C d) of the respective time period. All values are plotted as mean ± SE.