Table 2 Results of generalised least squares regression of log₁₀ subsampled marine genus biodiversity (quorum=0.4) on the δ¹⁸O palaeotemperature proxy of Prokoph et al.⁴⁰ and estimated sea level¹⁸.

Environment	Region	Independent variable	N	GLS			OLS (untransformed)
				Phi	Int.	Slope	Int.	Slope	R ²
Subsampled.
Marine	Global	Prokoph δ¹⁸O	12	0.68	0.26	−0.05 (0.227)	0.32	−0.03 (0.539)	-0.06
Marine	Global	Prokoph δ¹⁸O	12	0.88	0.25	−0.05 (0.188)	0.31	−0.03 (0.581)	-0.17
		+phase^†				−0.09 (0.503)		0.003 (0.972)
Marine	Global	Sea level	12	0.47	0.32	0.001 (0.159)	0.33	0.002 (0.114)	0.15
Marine	Global	Sea level	12	−0.01	0.46	0.005** (0.002)	0.46	0.005** (0.002)	0.61
		+phase^†				−0.25** (0.006)		−0.249** (0.006)
Counted.
Marine	Global	Prokoph δ¹⁸O	18	0.57	0.64	0.047 (0.500)	0.67	0.067 (0.373)	-0.01
Marine	Global	Prokoph δ¹⁸O	18	0.52	0.87	0.066 (0.331)	0.91	0.093 (0.184)	0.19
		+phase^†				−0.302 (0.155)		−0.311** (0.042)
Marine	Global	Sea level	18	0.72	0.60	0.004 (0.172)	0.62	<0.000 (0.978)	-0.06
Marine	Global	Sea level	18	0.54	0.89	0.005* (0.068)	0.96	0.006** (0.047)	0.30
		+phase^†				−0.439* (0.050)		−0.545** (0.008)

‘GLS’ denotes generalised least squares regression incorporating a first-order autoregressive covariance model; OLS denotes ordinary least squares regression; ‘fd’ indicates that first-differencing was applied to the input data series; phi is the serial correlation coefficient⁷⁰; ‘Int.’ is the y-intercept; R² is the adjusted R² of ordinary least squares regression, and comparable R² values cannot be computed for generalised least squares. ; **significant at alpha=0.05 (significance).
^†including a binary variable indicating the Jurassic–Early Cretaceous marine radiation as “1” and the Late Cretaceous–Cenozoic marine radiation as ‘2’

Quick links

Search