Tables
Table 1. Summary of the linear model fits of guild abundances
in 2018 as a function of the guild abundances recorded in 1982. We used
the same abundance estimation methods in both periods of time. Slope is
the estimated slope by the unconstrained slope model and 95% confidence
intervals in parenthesis. The column labeled n indicates the number of
species included in the analyses and in parenthesis we show the number
of species identified as outliers and excluded from the analysis. See
Table S2 for detailed list of outlier species. LRT = Likelihood Ratio
Test of a linear model with both, intercept and slope constrained to 0
and 1 respectively (null model) versus a linear model with constrained
intercept at 0 and unconstrained slope (alternative model). P-value is
the significance of the LRT with a value smaller than a
Bonferroni-corrected alpha level of 0.004 (=0.05/14) of rejecting the
null hypothesis. The column “Conclusion” states the decision outcome
of the test. If we failed to reject the null model then the conclusion
yielded a No change outcome. Alternative if the null model was rejected,
a slope larger than 1 denoted an increase and smaller the 1 denoted a
decrease in the abundance between the two time periods. Finally, if the
confidence interval of the slope overlapped with 0 the conclusion was
stated as No relationship. * and ** denote significantly higher number
of outliers than expected by chance testing at a 0.05 and 0.01 level
respectively.