Research ArticleCLIMATOLOGY

A multimillennial climatic context for the megafaunal extinctions in Madagascar and Mascarene Islands

See allHide authors and affiliations

Science Advances  16 Oct 2020:
Vol. 6, no. 42, eabb2459
DOI: 10.1126/sciadv.abb2459
  • Fig. 1 Climatology and proxy locations.

    Spatial pattern of December, January, and February (DJF) mean precipitation amount (millimeters per day) (shaded) between 1980 and 2015 from the European Centre for Medium-Range Weather Forecasts Reanalysis fifth-generation (ERA5) precipitation data. The climatological mean location of the southern boundary of the tropical rain belt (Sb-TRB) is indicated by the 4 mm/day isohyet (black). The precipitation amount of less than 4 mm/day is masked for clarity. The mean latitudinal positions of the northernmost and the southernmost Sb-TRB (4 mm/day isohyet) between 55°E and 75°E (dashed vertical lines) and 10°S to 25°S are shown by the red and blue isohyets, respectively (see Materials and Methods). Locations of Rodrigues (star) and other proxy records (yellow circles with numbers) (also see fig. S1) are marked as follows: 1, Anjohibe and Anjokipoty caves (10, 2629); 2, Lake Mitsinjo (25); 3, Lake Ihotry (20); 4, Lake Andolonomby (23); and 5, Ste-Luce peat bog cores (12).

  • Fig. 2 Changes in Sb-TRB positions and regional precipitation.

    (A) Spatial pattern of December through January (DJF) precipitation anomalies for the anomalous northward minus southward positions of the southern boundary of the tropical rainfall belt (Sb-TRB). Locations of Rodrigues (star) and other proxy records (yellow circles with numbers) are marked as in Fig. 1 (also see fig. S1). (B) Year-to-year comparisons between the mean latitudinal positions of DJF Sb-TRB (black) and ERA5 precipitation (z-scored) averaged over northern Madagascar [12°S to 20°S, 45°E to 50°E (blue)] and Rodrigues [17°S to 22°S, 60°E to 66°E (red)] (see Materials and Methods). (C) Same as (B) but for 1900 and 2009 from ERA-20C data. The correlation coefficients (r) with 95% confidence interval (brackets) among three time series are listed below (B) and (C). The mean northward and southward latitudinal positions of Sb-TRB (~17.6°S and 21.2°S) are defined by ±1σ departures (dashed lines) from the mean latitudinal position of Sb-TRB (~19.4°S) (solid line), respectively, in (B) and (C). (D) Same as (C) but for a 30-year low pass–filtered time series (see Materials and Methods). Shaded gray bar in (B) highlights the anomalous northward excursion of Sb-TRB and negative precipitation anomalies during the La Niña event of 1998 to 1999. Vertical bar in (D) highlights a sustained northward shift in Sb-TRB and changes in precipitation over Rodrigues and northern Madagascar.

  • Fig. 3 The new speleothem stable isotope records from Rodrigues and Madagascar (this study).

    (A) δ13C profiles of Rodrigues speleothems LAVI-15-7 (light green) and LAVI-4 (dark green) and smoothed Mg/Ca spectrum intensity ratio of LAVI-15-7 (black) together with 230Th dates and 2σ error bars (color-coded by sample names). VPDB, Vienna Pee Dee belemnite. (B) δ18O profiles of LAVI-15-7 (dark blue) and LAVI-4 (light blue). (C and D) δ13C (light brown) and δ18O (dark brown) profiles of a Madagascar speleothem ABC-1with 230Th dates and 2σ error bars. The mean isotopic values of each record are marked by dashed lines.

  • Fig. 4 Comparison between Rodrigues and Madagascar speleothem stable isotope records.

    (A to D) δ13C and δ18O profiles of LAVI-15-7 (Rodrigues, this study) (A), ABC-1 (Madagascar, this study) (B), AB2 [Madagascar, (26)] (C), and AB3 [Madagascar, (10)] (D) (see fig. S1 for proxy locations). All records are 5-year interpolated. Linear trends (least-squares fits of speleothem time series) that are significantly different from zero at 99% confidence level are shown with thick solid lines. Statistically insignificant linear trends are delineated by dotted lines. Plum and purple bars delineate the aragonite portions of stalagmites ABC-1 and AB2, respectively. Shaded bars on the top (with varying saturation intensity reflecting uncertainties) mark the timelines of permanent human settlements on Mauritius [red; (2, 3)], Rodrigues [blue; (5)], and Madagascar [green; (16, 18)]. The vertical color bars with varying saturation intensity show correlative features in all δ18O time series. The correlation coefficients (r) with 95% confidence interval and P values for each pair of δ18O profiles are reported on the figure. Statistical significance for each paired δ18O time series was assessed after accounting for autocorrelation and reduced degrees of freedom due to interpolation (see Materials and Methods).

  • Fig. 5 Timelines of climatic/biotic changes and human settlements.

    (A) The horizontal bars with varying saturation intensity show timelines of human settlements on the Mascarene Islands (red and blue), Madagascar (green), and the Malagasy megafaunal extinction (orange). Yellow and cyan circles mark the initial dates of human settlement after (16) and (18), respectively. The dotted line shows the temporal range of human settlements on Madagascar based on contentious evidence with the purple square representing the earliest date (17). (B) The z score–transformed (interpolated to 5 years) composite δ13C (black) and δ18O (cyan) speleothem profiles from Rodrigues (this study). Periods of drier conditions are identified by z score values of >1 (vertical brown bars with varying saturation intensity). The megadroughts are identified by the z score values >1 lasting 20 consecutive years or more (asterisks). Drying trends over the past 2 and 1.5 ka are shown by red and yellow arrows. (C) Proxy-inferred drying trends and droughts from southeast (SE) (12), northwest (NW) (25), and southwest (SW) Madagascar (20, 23) are indicated by shaded bars (fig. S1). (D) The z score δ18O profiles of ANJ94-5 (blue) (29) and 60-year smoothed ABC-1 (purple, this study) from Madagascar. Bars indicate the aragonitic portions in AB2 (green) (26), ABC-1 (purple, this study), and ANJ94-5 (blue) (29). (E) Published 14C dates of extinct subfossils from Madagascar for body masses between 10 and 150 kg (orange circles) and >150 kg (green circles) (9, 13). Vertical orange and green lines mark the upper 50% confidence intervals (CIs) for island-wide reported 4C dates for 10- to 150-kg and >150-kg body-mass groups (9). (F) Pace of megafaunal decline represented by maximum likelihood χ2 statistics showing subfossil extinct-to-extant odds ratio, after (11). Estimated changes in the effective population size (purple curve) after (45). Shaded orange bar is the same as in (A).

Supplementary Materials

  • Supplementary Materials

    A multimillennial climatic context for the megafaunal extinctions in Madagascar and Mascarene Islands

    Hanying Li, Ashish Sinha, Aurèle Anquetil André, Christoph Spötl, Hubert B. Vonhof, Arnaud Meunier, Gayatri Kathayat, Pengzhen Duan, Ny Riavo G. Voarintsoa, Youfeng Ning, Jayant Biswas, Peng Hu, Xianglei Li, Lijuan Sha, Jingyao Zhao, R. Lawrence Edwards, Hai Cheng

    Download Supplement

    The PDF file includes:

    • Figs. S1 to S10

    Other Supplementary Material for this manuscript includes the following:

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article