Fig 1.
The current extent of rainforest and locations of study sites within five subregions are shown.
Table 1.
Numbers of shared species between subregions (below the diagonal) and the total numbers of species for each subregion with number of unique species to the subregion in parentheses (in the diagonal).
All values are based on observed sample data from: Spec Uplands (SU), Atherton Uplands (AU), Bellenden Ker Uplands (BK), Carbine Uplands (CU) & Windsor Uplands (WU).
Fig 2.
Species richness of flightless ground beetles in the Wet Tropics recorded previously (charcoal bars) and this study (light grey bars).
SU = Spec Uplands; AU = Atherton Uplands; BK = Bellenden Ker Uplands; CU = Carbine Uplands & WU = Windsor Uplands. A new species was recorded at Spec Uplands (Castelnaudia sp.1).
Fig 3.
Changes in flightless ground beetle (a) observed species richness (±SE, n = 3) and (b) abundance (±SE, n = 3) with elevation across elevational gradients. SU = Spec Uplands; AU = Atherton Uplands; BK = Bellenden Ker Uplands; CU = Carbine Uplands & WU = Windsor Uplands.
Fig 4.
Community composition of flightless ground beetles in the Wet Tropics as described by NMDS ordination.
Ordination based on Bray Curtis similarities derived from square-root transformed abundance data from pitfall traps (stress = 0.22). Overlaid vectors are those of the five predictor variables indicated by the information theoretic approach as most likely (P value <0.001) to explain variation in the community composition of flightless ground beetles. Seas.P = precipitation seasonality; Ann.P = annual precipitation; Hist.veg = historical vegetation stability; Latitude = plot latitude; Ave.T = annual mean temperature. SU = Spec Uplands; AU = Atherton Uplands; BK = Bellenden Ker Uplands; CU = Carbine Uplands & WU = Windsor Uplands. All of the samples collected at 100 m a.s.l. at AU and CU were excluded as no beetles were collected.
Table 2.
Summary results of the information theoretic approach for relationships between predictor variables and observed species richness of flightless ground beetles, showing (a) summed Akaike weights from observed data, (b) mean summed Akaike weights from randomised data, (c) standard deviation (SD) of summed Akaike weights, standardised effect size ((a-b)/c), and P values calculated from 999 null models generated by permutation.
Hist.veg = historical vegetation stability; Treefall = disturbance from treefalls; Ave.T = annual mean temperature; Seas.P = precipitation seasonality; Seas.T = temperature seasonality; Litter = fine-litter standing crop; AWC = available water capacity; TWI = topographic wetness index; Ann.P = annual precipitation; Latitude = plot latitude; Aspect = plot aspect; Habitat = habitat heterogeneity of forest floor. Predictor variables displayed in bold were significant.
Fig 5.
Correlations between observed species richness and (A) historical vegetation stability, (B) disturbance from treefalls, (C) annual mean temperature and (D) precipitation seasonality, indicated by the information theoretic approach (see Table 2). SU = Spec Uplands; AU = Atherton Uplands; BK = Bellenden Ker Uplands; CU = Carbine Uplands & WU = Windsor Uplands.
Table 3.
Summary results of the information theoretic approach for relationships between predictor variables and community composition of flightless beetles, showing (a) summed Akaike weights from observed data, (b) mean summed Akaike weights from randomised data, (c) standard deviation (SD) of summed Akaike weights, standardised effect size ((a-b)/c), and P values calculated from 999 null models generated by permutation.
Seas.P = precipitation seasonality; Ann.P = annual precipitation; Hist.veg = historical vegetation stability; Latitude = plot latitude; Ave.T = annual mean temperature; Litter = fine-litter standing crop; AWC = available water capacity; Seas.T = temperature seasonality; Treefall = disturbance from treefalls; TWI = topographic wetness index; Aspect = plot aspect; Habitat = habitat heterogeneity of forest floor. Predictor variables displayed in bold were significant.