Research ArticleATMOSPHERIC SCIENCE

Projected increases in intensity, frequency, and terrestrial carbon costs of compound drought and aridity events

See allHide authors and affiliations

Science Advances  23 Jan 2019:
Vol. 5, no. 1, eaau5740
DOI: 10.1126/sciadv.aau5740
  • Fig. 1 Coupling of VPD and SM and their impacts on carbon uptake across the 66 flux tower sites.

    (A) Mean probability of each percentile bin of VPD and SM. Mean anomalies of GPP (B), TER (C), and NEP (D) for each percentile bin of VPD and SM. (E) Spearman correlation coefficient of VPD and SM. PMF (F) and anomalies of GPP, TER, and NEP (G) above 90th percentiles of VPD and below 10th percentiles of SM. Mean value is shown as a cross in the boxplots. At each site, anomalies of GPP, TER, and NEP were calculated as the difference of daily values and the mean daily values in the warm season.

  • Fig. 2 PMF in CMIP5 models.

    Model mean PMF of concurrent extreme VPD (above 90th percentile VPD) and SM (below 10th percentile SM) in historical simulations (1871–1970) (A) and in future simulations (2001–2100) (B and C). Thresholds used to define future extreme events are based on 2001–2100 (B) and 1871–1970 (C) (note the larger range of values expressed in this panel).

  • Fig. 3 Anomalies of GPP, TER, and NEP due to extreme high VPD and low SM in CMIP5 models.

    Extreme VPD (above 90th percentile VPD) and SM (below 10th percentile SM) in historical simulations (1871–1970) (A to C and J) and in future simulations (2001–2100) (D to I and K and L). Thresholds used to define future extreme events are based on 2001–2100 (D to F and K) and 1871–1970 (G to I and L). Anomalies of GPP, TER, and NEP were calculated as the difference between monthly values and the mean monthly values in the warm season for historical and future simulations individually.

  • Fig. 4 Additional effects of extreme high VPD and low SM on NEP in CMIP5 models.

    NEP anomaly due to extreme VPD (above 90th percentile VPD) or SM (below 10th percentile SM) in historical simulations (1871–1970) (A, B, and G) and in future simulations (2001–2100) (C to F and H and I). Thresholds used to define future extreme events are based on 2001–2100 (C, D, and H) and 1871–1970 (E, F, and I). The additional effect of extreme VPD (or SM) was calculated as the difference of NEP anomaly between compound extreme events and extreme SM (or VPD) alone.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/5/1/eaau5740/DC1

    Fig. S1. Frequency of compound extreme events and impacts on carbon uptake for different extreme levels.

    Fig. S2. GPP and TER anomalies as a function of VPD and soil water content.

    Fig. S3. Coupling of temperature and SM and their impacts on carbon uptake across the 66 flux tower sites.

    Fig. S4. Coupling of VPD and temperature across the 66 flux tower sites.

    Fig. S5. Distribution of VPD and SM in historical simulations (1871–1970).

    Fig. S6. Distribution of VPD and SM in future simulations (2001–2100).

    Fig. S7. Spearman correlation coefficient between VPD and SM in CMIP5 models.

    Fig. S8. SD of SM variability (SMv) in CMIP5 models.

    Fig. S9. Difference in PMF between historical and future simulations.

    Fig. S10. Comparison of VPD and SM thresholds between historical and future simulations.

    Fig. S11. Difference in anomalies of GPP, TER, and NEP between historical and future simulations.

    Fig. S12. Anomalies of GPP, TER, and NEP due to extreme high VPD in CMIP5 models.

    Fig. S13. Anomalies of GPP, TER, and NEP due to extreme low SM in CMIP5 models.

    Fig. S14. Additional effects of extreme high VPD and low SM on TER in CMIP5 models.

    Fig. S15. Additional effects of extreme high VPD and low SM on GPP in CMIP5 models.

    Fig. S16. Anomalies of GPP, TER, and NEP due to extreme high VPD and low SM in CMIP5 models.

    Fig. S17. PMF in CMIP5 models.

    Table S1. Information on the 66 flux tower sites from the FLUXNET2015 Dataset.

    Table S2. Model ensembles from the CMIP5 experiments.

    Table S3. A list of possible bivariate copula families.

  • Supplementary Materials

    This PDF file includes:

    • Fig. S1. Frequency of compound extreme events and impacts on carbon uptake for different extreme levels.
    • Fig. S2. GPP and TER anomalies as a function of VPD and soil water content.
    • Fig. S3. Coupling of temperature and SM and their impacts on carbon uptake across the 66 flux tower sites.
    • Fig. S4. Coupling of VPD and temperature across the 66 flux tower sites.
    • Fig. S5. Distribution of VPD and SM in historical simulations (1871–1970).
    • Fig. S6. Distribution of VPD and SM in future simulations (2001–2100).
    • Fig. S7. Spearman correlation coefficient between VPD and SM in CMIP5 models.
    • Fig. S8. SD of SM variability (SMv) in CMIP5 models.
    • Fig. S9. Difference in PMF between historical and future simulations.
    • Fig. S10. Comparison of VPD and SM thresholds between historical and future simulations.
    • Fig. S11. Difference in anomalies of GPP, TER, and NEP between historical and future simulations.
    • Fig. S12. Anomalies of GPP, TER, and NEP due to extreme high VPD in CMIP5 models.
    • Fig. S13. Anomalies of GPP, TER, and NEP due to extreme low SM in CMIP5 models.
    • Fig. S14. Additional effects of extreme high VPD and low SM on TER in CMIP5 models.
    • Fig. S15. Additional effects of extreme high VPD and low SM on GPP in CMIP5 models.
    • Fig. S16. Anomalies of GPP, TER, and NEP due to extreme high VPD and low SM in CMIP5 models.
    • Fig. S17. PMF in CMIP5 models.
    • Table S1. Information on the 66 flux tower sites from the FLUXNET2015 Dataset.
    • Table S2. Model ensembles from the CMIP5 experiments.
    • Table S3. A list of possible bivariate copula families.

    Download PDF

    Files in this Data Supplement:

Navigate This Article