Research ArticleCLIMATOLOGY

Climate models predict increasing temperature variability in poor countries

See allHide authors and affiliations

Science Advances  02 May 2018:
Vol. 4, no. 5, eaar5809
DOI: 10.1126/sciadv.aar5809
  • Fig. 1 Relative changes of SD of monthly temperature anomalies until the end of the 21st century.

    (A) Boreal summer [June, July, and August (JJA)], (B) austral summer [December, January, and February (DJF)], and (C) the whole year, averaged over 37 climate models. In hatched areas, at least 30 of the 37 models agree on the sign of change from 1850 to 2100 expressed in the Kendall τ value [for example, see fig. S1A for the model agreement in (C)].

  • Fig. 2 Changes in variability of the surface energy balance until the end of the 21st century.

    The horizontal axes in (A), (B), (D), and (E) show changes in the mean, whereas all other axes and the color bar show changes in SD. Each dot represents the difference between the periods 1875–1904 and 2055–2084 at a particular land grid cell in the Southern Hemisphere without Antarctica (A to C) and the Northern Hemisphere without Greenland (D to F) for the multimodel average. All figures show summer conditions (DJF in the Southern Hemisphere, JJA in the Northern Hemisphere). The blue coordinate system in (C) and (F) defines the index shown in Fig. 3, with negative values indicating larger increases in LH variability and positive values indicating larger radiative variability.

  • Fig. 3 Index related to the changes in the SD of LH versus downwelling net radiation.

    See main text and Fig. 2 (C and F). Negative values indicate that changes are driven by soil drying, whereas positive values indicate a larger contribution of atmospheric variability. Ice sheets and locations where changes in temperature SD are smaller than 0.1K have been masked out (white areas). Each hemisphere shows summer conditions (JJA in the North, DJF in the South). All changes are calculated from the multimodel average and the period 1875–1904 versus 2055–2084.

  • Fig. 4 Trends in temperature variability in the period 1958–2017 in observations and models.

    (A) Observations from HadCRUT4, (B) observations from GISTEMP, and (C) CMIP5 models. (A) and (B) show Kendall τ values, whereas (C) shows the number of models agreeing on the sign of the Kendall τ value. The window size for removing the annual cycle was 15 years, and the window to compute the Kendall τ value for the change in SD was 20 years.

  • Fig. 5 Relative change in SD of monthly temperature anomalies until the end of the 21st century versus per capita GDP and greenhouse gas emissions.

    (A) Per capita GDP and (B) per capita greenhouse gas emissions (without land use and forestry) between 1990 and 2013 in different countries. The red line marks zero change in temperature variability. The blue lines mark half of the world population (A) and half of the total worldwide emissions (B).

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/4/5/eaar5809/DC1

    fig. S1. Model agreement on changes in SD in monthly anomalies for all months of the year.

    fig. S2. Absolute and relative changes in SD of monthly temperature anomalies until the end of the 21st century.

    fig. S3. Relative change in SD of monthly temperature anomalies per global mean warming.

    fig. S4. Changes in temperature gradients and SD due to advection.

    fig. S5. Temporal correlation between monthly anomalies in latent heat flux and downwelling net radiation.

    fig. S6. Model evaluation of temperature SD between 1980 and 2010.

    fig. S7. Greenhouse gas emissions per year and person between 1990 and 2013 versus relative change in temperature SD in different countries (using all seasons).

    fig. S8. Documentation of our time series analysis method.

  • Supplementary Materials

    This PDF file includes:

    • fig. S1. Model agreement on changes in SD in monthly anomalies for all months of the year.
    • fig. S2. Absolute and relative changes in SD of monthly temperature anomalies until the end of the 21st century.
    • fig. S3. Relative change in SD of monthly temperature anomalies per global mean warming.
    • fig. S4. Changes in temperature gradients and SD due to advection.
    • fig. S5. Temporal correlation between monthly anomalies in latent heat flux and downwelling net radiation.
    • fig. S6. Model evaluation of temperature SD between 1980 and 2010.
    • fig. S7. Greenhouse gas emissions per year and person between 1990 and 2013 versus relative change in temperature SD in different countries (using all seasons).
    • fig. S8. Documentation of our time series analysis method.

    Download PDF

    Files in this Data Supplement:

Navigate This Article