Research ArticleECOLOGY

Multiple macroevolutionary routes to becoming a biodiversity hotspot

See allHide authors and affiliations

Science Advances  06 Feb 2019:
Vol. 5, no. 2, eaau8067
DOI: 10.1126/sciadv.aau8067
  • Fig. 1 Hotspots are poor in ancient lineages and sometimes rich in recent lineages.

    Residuals from linear models predicting cell-specific richness of (A) ancient and (B) recent lineages in hotspots (H; shown in red) and non-hotspot regions (N; shown in blue). Positive residuals indicate a regional excess of ancient/recent lineages, and negative residuals indicate a deficit. The bottom panels show the median of the difference between a random hotspot point and a random non-hotspot point and the 95% confidence interval around that median calculated with a Wilcoxon rank sum test.

  • Fig. 2 Contrasting rates of in situ cladogenesis in hotspots compared to surrounding non-hotspot regions.

    (A) In situ cladogenesis rates between 2 to 26 Ma ago within non-hotspots were subtracted from rates within hotspots in each of six biogeographic realms and divided by the overall SD to allow for comparison across realms. Solid lines indicate median differences (Δ) ± 90% confidence interval. Intervals overlapping the dotted line indicate a lack of statistically significant differences at α = 0.10. (B) Differences for each 2-Ma time bin.

  • Fig. 3 Source-sink dynamics of hotspots and their surrounding regions.

    (A) Dispersal rates between 2 and 26 Ma ago from non-hotspots to hotspots (N → H) were subtracted from hotspot to non-hotspots (H → N) rates within each realm. Lines and shaded areas presented as in Fig. 2.

  • Fig. 4 Hotspots are more spatially complex and have more energy than surrounding regions.

    Standardized differences in the average cell values of environmental variables between hotspots and non-hotspot cells within each of six biogeographic realms based on the coefficients of univariate spatially autocorrelated linear regressions. Blank areas indicate nonsignificant differences. CCV, climate change velocity.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/5/2/eaau8067/DC1

    Supplementary Text

    Fig. S1. Global maps of the mammal and bird hotspots in this study (shown in red).

    Fig. S2. DR estimates are correlated across the pseudoposterior distribution and also correlate with BAMM estimates.

    Fig. S3. Age of colonization in hotspots and non-hotspots.

    Fig. S4. Empirically estimated in situ cladogenetic rates in hotspots and non-hotspots differ from rates estimated in “control areas” with similar size and spatial structure to the real hotspots.

    Fig. S5. Empirically estimated dispersal rates from hotspots to non-hotspots (H → N) and from non-hotspots to hotspots (N → H) differ from rates estimated in control areas with similar size and spatial structure to the real hotspots.

    Fig. S6. Similar differences in contiguity of hotspot and non-hotspot cells across biogeographic realms.

    Fig. S7. Species richness-based hotspots and narrow ranged species-based hotspots are poor in ancient lineages and sometimes rich in recent lineages.

    Fig. S8. Contrasting macroevolutionary routes in species richness-based hotspots and non-hotspots and in narrow ranged species-based hotspots and non-hotspots.

    Fig. S9. Example of simulating control hotspots.

    Table S1. DR and BAMM produce consistent differences between hotspot and non-hotspot regions.

    Table S2. Total size and proportion of hotspot cells across biogeographic realms.

    Table S3. Mean of median distances (kilometer) of each cell to every neighboring cell of the same class with a radius of 1000 km is shown for hotspots and non-hotspots for mammals and birds.

    Table S4. Overlap of WE-based hotspots with SR- and NRS-based hotspots.

    Table S5. Model fit of BioGeoBEARS in six biogeographic realms.

  • Supplementary Materials

    This PDF file includes:

    • Supplementary Text
    • Fig. S1. Global maps of the mammal and bird hotspots in this study (shown in red).
    • Fig. S2. DR estimates are correlated across the pseudoposterior distribution and also correlate with BAMM estimates.
    • Fig. S3. Age of colonization in hotspots and non-hotspots.
    • Fig. S4. Empirically estimated in situ cladogenetic rates in hotspots and non-hotspots differ from rates estimated in “control areas” with similar size and spatial structure to the real hotspots.
    • Fig. S5. Empirically estimated dispersal rates from hotspots to non-hotspots (H → N) and from non-hotspots to hotspots (N → H) differ from rates estimated in control areas with similar size and spatial structure to the real hotspots.
    • Fig. S6. Similar differences in contiguity of hotspot and non-hotspot cells across biogeographic realms.
    • Fig. S7. Species richness-based hotspots and narrow ranged species-based hotspots are poor in ancient lineages and sometimes rich in recent lineages.
    • Fig. S8. Contrasting macroevolutionary routes in species richness-based hotspots and non-hotspots and in narrow ranged species-based hotspots and non-hotspots.
    • Fig. S9. Example of simulating control hotspots.
    • Table S1. DR and BAMM produce consistent differences between hotspot and non-hotspot regions.
    • Table S2. Total size and proportion of hotspot cells across biogeographic realms.
    • Table S3. Mean of median distances (kilometer) of each cell to every neighboring cell of the same class with a radius of 1000 km is shown for hotspots and non-hotspots for mammals and birds.
    • Table S4. Overlap of WE-based hotspots with SR- and NRS-based hotspots.
    • Table S5. Model fit of BioGeoBEARS in six biogeographic realms.

    Download PDF

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article