COMPUTER-SIMULATIONS SHOWING THE LIKELY EFFECTS OF CALIX ARCHITECTURE AND OTHER FACTORS ON RETRIEVAL OF DENSITY INFORMATION FROM CORAL SKELETONS

A computer model was built to simulate slices of skeleton taken from a massive colony of the hard coral Parites solida. A square wave density pattern was imposed upon the model of the skeleton by varying the thickness of lines representing thecal walls separating coral calices. Densitometric measurements were simulated and provided profiles of relative density along the "slices". Variables introduced into the overall computer model simulated aspects of real skeletal slices and real densitometric measurements. Skeletal architecture seriously interfered with the recovery of density information from the simulated skeletal slices. The recovered density profiles had, at best, ≈70% of variance in common with the initial square wave density pattern. Variance in common was usually <60%. Similar values were obtained from comparing different but equivalent regions of the same recovered density profiles. Densitometry on real skeletal slices is likely to provide density profiles with even less variance in common with the actual density pattern. Skeletal architecture introduced likely errors of up to 4 months in dating of peaks and troughs associated with the annual density cycle. It also created spurious peaks and troughs which could affect inter-annual dating of the density profile. Results obtained using the model suggest that skeletal slices should be cut more thickly than has been usual of late, and that densitometer beams should have a diameter about equal to the width of the narrowest density band to be measured. © 1990.