Spatially explicit comparison of abundances:
We additionally compared the local spatial distribution between time
periods computing Kernel Density Estimates (KDEs, 25x25m kernel size)
for the 111 species with >10 observations, using the raw
detection points data from the spot maps (Calenge 2006). We performed
first an overall comparison of the community spatial diversity, and
second, a species-by-species spatial distribution comparison. First, to
compare the community’s alpha-diversity between time periods, we
standardized the KDEs for every species between 0 and 1 by dividing the
density in each pixel by the maximum density for that species. We then
added pixel by pixel densities of all species as an analogous diversity
metric to Terborgh et al’s (1990) Figure 3. Second, we compared KDEs for
each of the 111 species separately. We computed Pearson’s correlation
coefficient using the Raster package (Hijmans 2020) for both community
diversity and single species. For single species comparisons we tested
whether the mean and median correlation coefficients were, on average,
significantly higher than expected by chance using a non-parametric
bootstrap test (Supplementary Information; Manly 2006).
Finally, we used a multi-model approach to determine if body mass,
foraging strata, guild, and sociality explained single species’ spatial
distribution similarity between time periods using a generalized least
squares (gls) accounting for phylogenetic non independence, assuming a
Brownian motion model evolution (Paradis & Schliep 2019). We fitted the
data with all possible combinations of the independent variables using
the MuMIn package (Barton 2018) with the Bayesian Information Criterion
(BIC) as the model rank statistic. To propagate uncertainty in
phylogenetic estimation, we repeated the model selection procedure using
100 randomly sampled trees from the ones used in previous analyses. We
report the proportion of times that each candidate model is deemed as
best using a delta BIC of at least 3 points (Taper & Ponciano 2016).