SUPPLEMENTARY FIGURES

Transcription

1 SUPPLEMENTARY FIGURES 1 Supplementary Figure 1 The Australian Wet Tropics study region. a, Australia, with the state of Queensland shown in pale grey. b, Queensland, with the study region shown as a coloured digital elevation model. c, the study region, with the elevation shown, and the boundary of the Wet Tropics Bioregion (within which lies the majority of tropical rainforest) shown with a blue line.

2 2 Supplementary Figure 2 Projected median dispersal distances. a, plant fruit group (PFG) 1 at 2010, b, PFG1 at 2100, c, PFG2 at 2010, d, PFG2 at 2100, e, PFG3 at 2010, f, PFG3 at 2100, g, PFG4 at 2010, h, PFG4 at 2100, i, PFG6 at 2010, j, PFG6 at 2100, k, PFG7 at 2010, l, PFG7 at (for PFG5, see Fig. 1).

3 3 a b Supplementary Figure 3 Projected change in dispersal within 100 m elevation bands of the Australian Wet Tropics bioregion. a, change in median dispersal distance and; b, change in long distance dispersal (99 th percentile) for each 100 m elevation band (Fig. S1c) from 2010 to 2100 under climate change. Elevation bands that show significant decreases (D) and increases (I) in dispersal distance over time are labelled (P < 0.001; n = 224,733 grid cells; assessed through permutation test). Data are boxplots (with outliers shown) for all grid cells within each elevation band.

4 4 Supplementary Figure 4 Sensitivity of Median Dispersal Distance (MDD 50 th percentile) and Long Distance Dispersal (LDD 99 th percentile) for all plant fruit groups combined, averaged across the Australian Wet Tropics bioregion, to independent variation in each of five major parameter sets (abundance, proportion fruit eaten, passive dispersal, gut passage time, displacement). For each parameter set, all parameters within that set were varied in a uniform random manner in the range ± 50 % of the estimated value (Tables S1, S2), then dispersal kernels calculated (repeated 100 times for each parameter set). The grey line in a and b shows the value for the original parameters, while the boxplots for each parameter set are summaries of the 100 replicates of the sensitivity test (with outliers shown).

5 5 Supplementary Figure 5 The methodological framework applied to generate projected seed dispersal kernels for fleshy fruited plant fruit groups in the Australian Wet Tropics. The categories in brackets show the relevant biological, spatial, and habitat factors considered in each step, specifically being: each of the 45 frugivorous vertebrate species (species); each of the 11 disperser functional groups (DFG); each of the 7 plant fruit groups (PFG); each location across the Australian Wet Tropics (location), and; both contiguous and fragmented habitat types (habitat).

6 6 Supplementary Figure 6 Differentiation between contiguous and fragmented habitat in the Australian Wet Tropics. Dispersal kernels for each plant functional group were developed for both contiguous, and fragmented habitats. Fragmented habitat was defined as grid cells for which the proportion of unmodified habitat in the surrounding area (1 km radius) was < 75 %. All other habitat was considered to be contiguous. Note that dispersal kernels are from each grid cell, being either fragmented or contiguous habitat.

7 SUPPLEMENTARY TABLES 7 Supplementary Table 1 Summary of the projected median dispersal distance (MDD: 50 th percentile) and long dispersal distance (LDD: 99 th percentile) for each plant fruit group (PFG) over the Wet Tropics Bioregion (Fig. S1c) under current and future climates. The minimum (min) and maximum (max) values across all the grid cells in the Wet Tropics Bioregion are given, as well and the mean and standard deviation (sd) across all 224,733 grid cells. PFG1 PFG2 PFG3 PFG4 PFG5 PFG6 PFG7 MDD current min (2010) mean (sd) (±49.6) (±55.1) 43.4 (±5.6) 70.1 (±32.2) 57.6 (±10.7) 34.9 (±11.3) 28.2 (±5.9) max future min (2100) mean (sd) 73.1 (±47.3) 98.0 (±54.3) 23.3 (±7.1) 46.2 (±27.0) 30.7 (±14.1) 23.9 (±9.2) 20.4 (±5.7) max change min ( ) mean (sd) (±35.1) (±36.3) (±5.8) (±25.0) -27 (±10.8) (±8.8) -7.8 (±4.3) max LDD current min (2010) mean (sd) (±199.7) (±15.4) (±200) (±13.7) (±208.4) (±15.3) (±45) max future min (2100) mean (sd) (437.7) (23.9) (455.1) (21.6) (377.6) (23.2) (11.2) max change min ( ) mean (sd) -996 (327.9) (13.0) (335.1) (11.7) (293.2) (12.3) (44.2) max

8 8 Supplementary Table 2 Sample size, test AUC score (area under the receiver operating characteristic curve) and environmental variable contributions (in grey shading) for MaxEnt models 39 of the habitat suitability of 45 frugivorous vertebrate species in the Australian Wet Tropics. Also shown are the parameters for the negative exponential function for estimating the abundance of each species from their predicted habitat suitability 40. Variable contribution to the MaxEnt species distribution model Scientific name Common name DFG code Training sample number AUC Test Annual precipitation Annual evaporation Precipation:evaporation Annual mean radiation Annual mean temp. Max. temp. warmest period Min. temp. coldest period Isothermality C3 microphyll GI C3 mesophyll GI C3 megaphyll GI C4 megaphyll GI Abundance coefficient Ailuroedus melanotus Spotted catbird v Amblyornis newtonianus Golden bowerbird v Aplornis metallica Metallic starling * Casuarius casuarius Southern cassowary L Coracina lineata Barred cuckoo-shrike v Coracina novaehollandiae Black-faced cuckoo-shrike * Coracina papuensis White-bellied cuckoo-shrike * Cracticus quoyi Black butcherbird v Dicaeum hirundinaceum Mistletoebird v Dicrurus bracteatus Spangled drongo v Dromaius novaehollandiae Emu L Ducula bicolor Pied imperial-pigeon * Eudynamys orientalis Common koel * Hypsiprymnodon moschatus Musky rat-kangaroo * Lalage leucomela Varied triller v Lichenostomus flavus Yellow honeyeater * Lichenostomus frenatus Bridled honeyeater v Lichenostomus versicolor Varied honeyeater * Lopholaimus antarcticus Topknot pigeon v Meliphaga gracilis Graceful honeyeater v Meliphaga lewinii Lewin's honeyeater v Meliphaga notata Yellow-spotted honeyeater v Myzomela obscura Dusky honeyeater v Nyctimene robinsoni Eastern tube-nosed bat * Oriolus flavocinctus Yellow oriole * Oriolus sagittatus Olive-backed oriole * Philemon buceroides Helmeted friarbird * Philemon corniculatus Noisy friarbird * Pteropus alecto Black flying-fox * Pteropus conspicillatus Spectacled flying-fox * Pteropus scapulatus Little Red flying-fox * Ptilinopus magnificus Wompoo fruit-dove v Ptilinopus regina Rose-crowned fruit-dove * Ptilinopus superbus Superb fruit-dove v Ptilonorhynchus maculatus Spotted bowerbird * Ptilonorhynchus nuchalis Great bowerbird * Ptilonorhynchus violaceus Satin Bowerbird v Ptiloris victoriae Victoria's riflebird v Scenopoeetes dentirostris Tooth-billed bowerbird v Scythrops novaehollandiae Channel-billed cuckoo * Sphecotheres viridis Figbird * Strepera graculina Pied currawong v Syconycteris australis Common blossom bat * Xanthotis macleayanus Macleay's honeyeater v Zosterops lateralis Silvereye v Max. abundance DFG = disperser functional group; Temp. = Temperature; Max. = Maximum; Min. = Minimum; GI = plant growth index Abundance parameters for negative exponential function derived from: v = VanDerWal et al. (2009); * = mean of nearest taxonomic group; L = literature AUC Test = AUC for test set from 10-fold cross-validation

9 9 Supplementary Table 3 Parameters for the displacement, gut passage time and proportion of crop eaten for each of the 11 frugivorous disperser functional groups (DFG) in the Australian Wet Tropics. The parameters c and z are from negative exponential models of DFG displacement, while parameters k and θ are from gamma distributions of seed gut passage time. Displacement Contiguous habitat Fragmented habitat Gut passage time Proportion of plant fruit group crop eaten per unit abundance Disperser functional group DFG code c z c z k ϴ PFG 1 PFG 2 PFG 3 PFG 4 PFG 5 PFG 6 PFG 7 Large terrestrial 1 9.E Large within forest Occasional Small within forest Terrestrial within forest Throughout landscape facultative throughout landscape frugivore Wide-ranging small seeds Wide-ranging large seeds Wide-ranging rapid gut Wide-ranging slow gut

10 Supplementary Table 4 Sensitivity of Median Dispersal Distance (MDD 50 th percentile) for each plant fruit group (and all combined), averaged across the Australian Wet Tropics bioregion, to independent variation (100 replicates) in each of the five major parameter sets (abundance, proportion fruit eaten, passive dispersal, gut passage time, displacement). 10 PFG1 PFG2 PFG3 PFG4 PFG5 PFG6 PFG7 All PFG abundance min (sd) 16.9 (±1.6) 14.6 (±1.6) 9.8 (±0.6) 13.4 (±1.2) 11.6 (±1.1) 18.7 (±1.4) 11.2 (±1.3) 8.2 (±0.2) mean (sd) 48.9 (±2.4) 52.7 (±6.0) 32.2 (±1.2) 42.0 (±1.9) 44.4 (±1.2) 30.8 (±0.5) 29.4 (±0.5) 15.5 (±0.3) max (sd) 95.9 (±9.4) (±12.8) 73.8 (±3.7) 89.8 (±4.1) 77.7 (±5.3) 87.9 (±1.5) 69.3 (±4.1) 28.7 (±1.0) proportion fruit min (sd) 16.8 (±1.1) 14.7 (±1.1) 9.6 (±0.7) 13.3 (±1.0) 11.4 (±1.0) 18.5 (±0.6) 11.4 (±1.3) 8.2 (±0.2) eaten mean (sd) 49.6 (±4.0) 55.2 (±11.3) 32.2 (±1.7) 41.8 (±2.8) 44.2 (±1.0) 30.8 (±0.6) 29.6 (±0.4) 15.5 (±0.2) max (sd) 97.7 (±16.8) (±21.6) 74.2 (±5.8) 90.0 (±6.4) 78.1 (±7.6) 88.1 (±1.7) 68.5 (±5.5) 28.9 (±1.0) passive min (sd) 10.9 (±2.0) 8.8 (±1.5) 12.0 (±2.2) 10.5 (±1.9) 10.7 (±2.0) 25.5 (±0.7) 21.7 (±2.1) 9.0 (±1.1) dispersal mean (sd) 31.5 (±2.3) 24.1 (±2.4) 37.9 (±0.4) 31.9 (±1.7) 41.7 (±1.3) 34.9 (±0.1) 32.9 (±0.0) 15.5 (±0.8) max (sd) 55.5 (±1.4) 47.2 (±3.0) 85.0 (±0.0) 69.0 (±1.3) 71.9 (±0.7) 91.0 (±0.0) 76.0 (±0.0) 28.2 (±0.7) gut passage min (sd) 10.7 (±0.4) 9.0 (±0.1) 11.5 (±0.5) 10.0 (±0.0) 10.6 (±0.5) 25.1 (±5.9) 20.1 (±4.4) 8.9 (±0.6) time mean (sd) 31.1 (±3.4) 22.8 (±2.5) 38.5 (±5.3) 31.1 (±3.4) 41.3 (±6.6) 34.9 (±6.0) 32.1 (±6.9) 15.3 (±1.2) max (sd) 56.5 (±7.9) 45.4 (±5.8) 85.7 (±16.9) 68.6 (±11.8) 78.5 (±16.4) 92.1 (±21.7) 80.1 (±23.1) 28.8 (±4.0) displacement min (sd) 10.7 (±0.5) 9.0 (±0.1) 11.5 (±0.5) 10.0 (±0.0) 10.6 (±0.5) 24.8 (±6.2) 20.3 (±4.1) 8.9 (±0.6) mean (sd) 30.7 (±3.5) 23.1 (±2.5) 38.1 (±5.5) 31.1 (±3.4) 40.1 (±6.8) 34.8 (±6.8) 33.0 (±7.3) 15.1 (±1.2) max (sd) 55.0 (±8.0) 45.0 (±5.5) 84.5 (±16.3) 67.9 (±10.4) 76.6 (±16.4) 91.3 (±21.7) 78.8 (±22.0) 28.5 (±3.6)

11 11 Supplementary Table 5 The effect of maintaining the current (2010) abundances of the vertebrate frugivore species within each disperser functional group (DFG) to 2100 under climate change on both long and median seed dispersal distances for each plant fruit group (PFG), and the weighted mean of all PFG s. Values are the difference in dispersal distance (m) in 2100 when the abundances within each DFG are maintained at 2010 levels. Large positive values indicate increased dispersal distances (shaded darker blue), while large negative values indicate decreased dispersal distances (shaded darker red) when abundances of each DFG are maintained at 2010 levels to PFG1 PFG2 PFG3 PFG4 PFG5 PFG6 PFG7 All Dispersers Distance (small fig) (large fig) (small multi) (large multi) (small drupe) (med. drupe) (large drupe) DFG1 Median (large terrestrial) Long DFG2 Median (large, within forest) Long DFG3 Median (occasional) Long DFG4 Median (small within forest) Long DFG5 Median (terrestrial within forest) Long DFG6 Median (throughout landscape facultative) Long DFG7 Median (throughout landscape frugivore) Long DFG8 Median (wide-ranging, small seeds) Long DFG9 Median (wide-ranging, large seeds) Long DFG10 Median (wide-ranging, rapid gut) Long DFG11 Median (wide-ranging, slow gut) Long