According to the Global Precipitation Climatology Projection (GPCP; Adler et al. 2018), over 70% of the nondesert regions over Iran (the climatological precipitation > 0.5 mm day−1) have experienced unprecedented low annual precipitation in 2021 since 1979 (Figs. 1a,b). Cumulative precipitation deficits from October 2020 through March 2021 initiated the 2020/21 drought. Over Iran, a majority of crops (corn, cotton, millet, rice, sorghum, soybean, and sunflower) are planted in April and harvested in October/November (USDA 2022), which have been adversely affected by the 2020/21 drought. According to the ERA5 data (Hersbach et al. 2020), this drying was induced by the strong Arabian subtropical high, which expanded over 2021 (Fig. 1c) and were associated with the positive–negative (north–south) dipole pattern of midtropospheric pressure anomalies over the surrounding regions of Iran (Bárdossy et al. 2021).
(a) Spatial distribution of record-breaking precipitation over Iran and (b) time series of the GPCP annual precipitation anomaly percentages over Iran (regional average over the grid cells colored in orange) relative to the 1981–2010 climatology. In (a), grids colored in gray represent desert areas (30-yr average of annual mean precipitation ≤ 0.5 mm day−1) and “1” indicates the grid cell with the lowest annual precipitation in 2021 since 1979. (c) Spatial distribution of geopotential height anomalies (shading) at 850 hPa in 2021. In (c), green and black colored lines depict contour lines of 1,525 gpm for 2021 and the 1981–2010 climatology, respectively. (d) Time series of monthly SPI12 values (bar graph) and precipitation anomalies (green line; in mm day−1) during 2017–21. (e) Time series of the monthly relative online search activity volumes using the search topic “drought” during 2017–21.
Citation: Bulletin of the American Meteorological Society 103, 12; 10.1175/BAMS-D-22-0149.1
The Google Trends data show that relative online search activities with a search topic, “drought” [defined as “drought awareness” in Kam et al. (2019)], have been low in 2021 compared with those during the 2018 Iran drought (Figs. 1d,e). This indicates a lack of the national-level awareness about the ongoing drought, which possibly triggered water protests in the two major provinces, Khuzestan and Esfahan, and thus causing several casualties and hundreds injured by security forces (Iran International 2022). Previous studies found limited indication of the detection of increased drying trends (Orlowsky and Seneviratne 2013; Spinoni et al. 2019; Vaghefi et al. 2019) and the attribution of recent meteorological droughts to anthropogenic impacts remains elusive (Barlow and Hoell 2015). In this respect, assessing human contribution to 2020/21-like Iran meteorological droughts will provide important implications.
Here, we investigate the anthropogenic contribution to the likelihood of 2021-like meteorological droughts (persistence, severity, and intensity) by quantifying the contribution of anthropogenic greenhouse gas and aerosol forcing, and natural (solar + volcanic) forcing from the CMIP6 multimodel individual forcing simulations. The findings of this study will have policy-makers and local stakeholders on the alert about the need to proactively adapt to climate change for water and food security.
Data and methods
To calculate the meteorological drought index, first, we computed the regional averages of monthly precipitation over Iran, excluding desert regions (the climatological precipitation ≤ 0.5 mm day−1). We used the CPC Merged Analysis of Precipitation (CMAP; Xie and Arkin 1997), Global Precipitation Climatology Centre (GPCC; Schneider et al. 2011), Climate Research Unit TS v.4.06 (CRU; Harris et al. 2020), GPCP, and ERA5 precipitation data (Fig. ES1a in the online supplemental material; https://doi.org/10.1175/BAMS-D-22-0149.2). In addition, we also computed the regional averages of 30 stations over Iran during 1956–70 (see Fig. ES1b) to compare the climatology (mean and standard deviation) of precipitation before 1979. We used the observational and reanalysis precipitation data over 1979–2021 considering large uncertainties before 1979 due to the rapidly increased station numbers (see Figs. ES1c–f). We computed the 12-month standardized precipitation index (SPI-12) using the regional averages of observed monthly precipitation.
To assess the human contribution to 2020/21-like droughts, we defined the characteristics of 2020/21-like droughts using the observed threshold values of drought persistence, severity, and intensity in 2021. These threshold values include the consecutive 10 months (March–December) with SPI-12 ≤ –0.5, –12 of drought severity (the sum of the negative SPI-12 value over the 10 months), and –2 of drought intensity (the minimum value). These characteristics of Iran droughts have confined our assessment to attribution, rather than associated drought mechanisms like ENSO, because of time-varying impacts at the seasonal scale (Nazemosadat and Ghasemi 2004). Furthermore, CMIP6 models show a limited ability to reproduce observed SST teleconnections (Barlow et al. 2021). Multiple observational data show a strong consensus on the detection of three historical droughts over 1979–2021 (Fig. ES1a). These severe three droughts (2008, 2010, and 2021) have been previously reported (Barlow et al. 2016). Despite strong consensus on the drought persistence, the multiple observational data showed a wide range of the drought severity [–17.0 (CRU) through –9.1 (ERA5)] and intensity [–2.8 (GPCP) through –1.5 (ERA5)]. The threshold values for the severity and intensity of the 2020/21 drought are –12.0 and –2.0 (the averages from the four observational and reanalysis datasets) to have sufficient samples from the CMIP6 runs.
Next, we computed the regional averages of monthly precipitation from multiple historical (H)-forcing experiment runs of the 13 CMIP6 models (Eyring et al. 2016). These 13 CMIP6 models were selected based on the availability of multiple ensemble members of each forcing experiment run [at least two ensemble members for H-, anthropogenic greenhouse gas–only (G)-, natural-only (N)-, and anthropogenic aerosol–only (A)-forcing experiment runs]. The raw CMIP6 data were regridded through the bilinear interpolation at the 2.5° (250-km) resolution before analysis.
We evaluated models for the seasonality of simulated precipitation based on the percentage of each monthly precipitation to the annual total precipitation (Fig. ES2b). Then, we calculated temporal correlation coefficients with observed seasonality and found 13 CMIP6 models reproduced well the seasonality of observed precipitation (Fig. ES2c). The H-forcing experiment runs (ended in 2014) were extended up to 2020 using the corresponding shared socioeconomic pathway (SSP) 2-4.5 scenario runs (O’Neill et al. 2020). We chose the SSP2-4.5 scenarios runs for consistency because G-, N-, and A-forcing experiment runs were extended for 2015–20 under the SSP2-4.5 scenarios (Gillett et al. 2016). For H-forcing experiment runs, GISS-E2-1-G and IPSL-CM6A-LR were not extended with the SSP2-4.5 scenario runs because their G-, A-, and N-forcing experiment runs were ended in 2014. The detailed information on the CMIP6 models used in this study is shown in Table ES1.
Using 66 ensemble H-forcing experiment runs of the 13 CMIP6 models, we computed the SPI-12 values over 1979–2021. First, we summed monthly precipitation over the preceding 12 months (t – 11 to t) and then computed two parameters (α and β) of the beta distribution function for each ensemble H-forcing run. Next, we used the fitted beta distribution function to compute the percentages of the 12-month cumulative precipitation through the maximum likelihood estimation method following Thom (1958). Last, we used the estimated two parameters of the beta distribution from the corresponding ensemble H-forcing runs to compute the percentages of the 12-month cumulative precipitation, which can clarify the relative impact of G, A, and N forcing on the SPI estimates (Kam et al. 2021). We used the inverse standard normal distribution function to standardize the precipitation percentiles.
To quantify the human contribution to the persistence, severity, and intensity of the 2020/21-like Iran droughts, we counted the numbers of the years when SPI-12 ≤ –0.5 from February through December over the recent 10-yr segments (2011–20) of H-, G-, N-, and A-forcing runs of the 13 CMIP6 models. We also computed the sum and minimum of the SPI-12 values over these 11 months of the corresponding year. The total sample size for each experiment is 612 (66 ensemble runs × 10 years) because the eight ensemble runs of the GISS-E2-1-G and IPSL-CM6A-LR models end in 2014.
To construct the multimodel distribution of the persistence, severity, and intensity, we randomly sampled 66 ensembles from H-, G-, N-, and A-forcing experiment runs with replacement (bootstrapping sampling) based on the ensemble numbers of the 13 CMIP6 models (see Table ES1) 1,000 times. Then, we counted the numbers of the years (i.e., a kind of order statistics) when the 10 months (March–December) with SPI-12 ≤ –0.5 (persistence) from the 1,000 bootstrapping samples. We counted the events with the sum and minimum of the SPI-12 values are –12.0 (severity) and –2.0 (intensity) or below, respectively, over these 10 months from the bootstrapping samples. Therefore, the total sample size for the multimodel distribution construction is 612,000 (1,000 sets of the 612 ensemble years). Last, we computed the occurrence probability ratio (PR) of 2020/21-like droughts (PH/PN, PG/PN, and PA/PN) for persistence, severity, and intensity. Last, we constructed the 95th-percentile range for PH/PN, PG/PN, and PA/PN, from the 25th- and 975th-rank PR values in the 1,000 bootstrapping samples.
Results
The 2021 Iran meteorological drought started in February, was exacerbated rapidly in March and April (ΔSPI-12 ≤ –0.5), and reached the SPI value below –2.0 in December (Fig. 1d). The monthly precipitation deficits for 2020/21 showed that the cumulative precipitation deficits from October 2020 initiated the 2020/21 drought (Fig. ES2d). While the SPI-12 values from multiple observational data show a wide range, the 10-month drought persistence (March–December) exhibited strong consensus among the observational and reanalysis precipitation data (not shown).
CMIP6 multimodel 10-yr samples (out of 612 years) show that there are 104, 49, 80, and 50 drought years for H-, N-, G-, and A-forcing runs, respectively, which have the SPI-12 values below –0.5 from March through December. The drought years with the severity (intensity) ≤ –12.0 (–2.0) are 89, 39, 80, and 28 (57, 28, 63, and 10) from H-, N-, G-, and A-forcing runs, respectively. The histograms of the drought persistence, severity, and intensity from the CMIP6 models clearly show that the H- and G-forcing runs have longer lasting droughts (SPI-12 ≤ –0.5) with intensified the severity and intensity (Figs. 2a–c).
Probability mass functions (PMFs) of (a) the persistence, (b) the severity, and (c) intensity values of simulated drought years from the CMIP6 models. Orange, red, blue, and green lines depict the PMFs from the H-, G-, N-, and A-forcing runs, respectively. Black lines in (a)–(c) depict the observed threshold values in 2020/21. (d) Circles and bars respectively depict the median and 95th-percentile range (2.5th–97.5th percentiles) of PR values obtained from the 1,000-bootstrapping, 10-yr segment samples. See text for details.
Citation: Bulletin of the American Meteorological Society 103, 12; 10.1175/BAMS-D-22-0149.1
The PR values of PH/PN, PG/PN, and PA/PN for the persistence of 2020/21-like droughts (the 95th-percentile range) are obtained as 2.2 (1.68–2.68), 1.9 (1.5–2.5), and 1.0 (0.8–1.4) (Fig. 2d). This result indicates that the human contribution has likely increased the occurrence probability of 2021-like Iran meteorological droughts. For the severity of 2020/21-like droughts, the values of PH/PN, PG/PN, and PA/PN are found as 2.3 (1.8–3.0), 2.05 (1.57–2.75), and 0.73 (0.48–1.1). For the intensity, the values of PH/PN, PG/PN, and PA/PN are found as 2.0 (1.5–2.9), 2.24 (1.6–3.2), and 0.4 (0.2–0.6). The results support the robust human contribution to the increased 2020/21-like droughts. The anthropogenic greenhouse gas forcing was found to increase the probability of more severe meteorological droughts while the anthropogenic aerosol forcing offset the GHG-induced change.
In summary, the 2021 Iran meteorological drought initiated from precipitation deficits in October 2020 and persisted the drought conditions until December 2021, recording the second-highest severity (first-highest severity in 2008) over nondesert regions in Iran since 1979. When analyzing the CMIP6 multimodel individual forcing simulations, the human contribution has increased the probability exceeding the observed drought persistence, severity, and intensity in 2021 by at least 50%, particularly via the anthropogenic greenhouse gas increases, which overwhelms a minor offsetting influence of anthropogenic aerosols. Although we used different observations and evaluated models, it should be noted that large uncertainties remain in precipitation observations and climate model performances (Barlow et al. 2016; Bárdossy et al. 2021; Yazdandoost et al. 2021). Particularly, decreased participant stations over Iran in the recent decade raise a concern about uncertainties in the observational data (Figs. ES1d,f) because some stations have been no longer maintained (Harris et al. 2020). In addition, physical processes associated with the persistence of Iran droughts needs to be further evaluated across the seasonal scales, including anomalous large-scale circulations and SST teleconnections, via the observational and reanalysis data (Barlow and Hoell 2015; Barlow et al. 2021). While exceptional droughts have occurred not only in Iran but also in other Middle East countries (Kaniewski et al. 2012), human disturbance over Iran (dam operation failure, groundwater deletion, and a rapid economic growth) is another challenge in future drought mitigation.
Acknowledgments.
We thank the CMIP6 project and the National Oceanic and Atmospheric Administration’s Central Library for making available the CMIP6 data and observational precipitation from stations. We also thank the NOAA PSL, CRU, and ECMWF for providing gridded observational and reanalysis precipitation datasets and Dr. Yeon-Hee Kim for sharing some of the CMIP6 data that were not available at the time of writing from the CMIP6 data archive. This study is supported by the National Research Foundation of Korea (NRF-2021R1A2C1093866).
References
Adler, R. F., and Coauthors , 2018: The Global Precipitation Climatology Project (GPCP) monthly analysis (new version 2.3) and a review of 2017 global precipitation. Atmosphere, 9, 138, https://doi.org/10.3390/atmos9040138.
Bárdossy, A., E. Modiri, F. Anwar, and G. Pegram , 2021: Gridded daily precipitation data for Iran: A comparison of different methods. J. Hydrol. Reg. Stud., 38, 100958, https://doi.org/10.1016/j.ejrh.2021.100958.
Barlow, M., and A. Hoell , 2015: Drought in the Middle East and central–southwest Asia during winter 2013/14. Bull. Amer. Meteor. Soc., 96, S71– S76, https://doi.org/10.1175/BAMS-D-15-00127.1.
Barlow, M., B. Zaitchik, S. Paz, E. Black, J. Evans, and A. Hoell , 2016: A review of drought in the Middle East and southwest Asia. J. Climate, 29, 8547– 8574, https://doi.org/10.1175/JCLI-D-13-00692.1.
Barlow, M., A. Hoell, and L. Agel , 2021: An evaluation of CMIP6 historical simulations of the cold season teleconnection between tropical Indo-Pacific sea surface temperatures and precipitation in southwest Asia, the coastal Middle East, and northern Pakistan and India. J. Climate, 34, 6905– 6926, https://doi.org/10.1175/JCLI-D-19-1026.1.
Eyring, V., S. Bony, G. A. Meehl, C. A. Senior, B. Stevens, R. J. Stouffer, and K. E. Taylor , 2016: Overview of the Coupled Model Intercomparison Project phase 6 (CMIP6) experimental design and organization. Geosci. Model Dev., 9, 1937– 1958, https://doi.org/10.5194/gmd-9-1937-2016.
Gillett, N. P. , and Coauthors, 2016: The Detection and Attribution Model Intercomparison Project (DAMIP v1.0) contribution to CMIP6. Geosci. Model Dev., 9, 3685– 3697, https://doi.org/10.5194/gmd-9-3685-2016.
Harris, I., T. J. Osborn, P. Jones, and D. Lister , 2020: Version 4 of the CRU TS monthly high-resolution gridded multivariate climate dataset. Sci. Data, 7, 109, https://doi.org/10.1038/s41597-020-0453-3.
Hersbach, H., and Coauthors , 2020: The ERA5 global reanalysis. Quart. J. Roy. Meteor. Soc., 146, 1999– 2049, https://doi.org/10.1002/qj.3803.
Iran International, 2022: Iran’s capital faces serious water shortage in coming months. Accessed 25 June 2022, www.iranintl.com/en/202204040494.
Kam, J., K. Stowers, and S. Kim , 2019: Monitoring of drought awareness from Google Trends: A case study of the 2011–17 California drought. Wea. Climate Soc., 11, 419– 429, https://doi.org/10.1175/WCAS-D-18-0085.1.
Kam, J., S.-K. Min, P. Wolski, and J.-S. Kug , 2021: CMIP6 model-based assessment of anthropogenic influence on the long sustained western Cape drought over 2015–19. Bull. Amer. Meteor. Soc., 102, S45– S50, https://doi.org/10.1175/BAMS-D-20-0159.1.
Kaniewski, D., E. Van Campo, and H. Weiss , 2012: Drought is a recurring challenge in the Middle East. Proc. Natl. Acad. Sci. USA, 109, 3862– 3867, https://doi.org/10.1073/pnas.1116304109.
Nazemosadat, M. J., and A. R. Ghasemi , 2004: Quantifying the ENSO-related shifts in the intensity and probability of drought and wet periods in Iran. J. Climate, 17, 4005– 4018, https://doi.org/10.1175/1520-0442(2004)017<4005:QTESIT>2.0.CO;2.
O’Neill , and Coauthors, 2016: The Scenario Model Intercomparison Project (ScenarioMIP) for CMIP6. Geosci. Model Dev., 9, 3461– 3482, https://doi.org/10.5194/gmd-9-3461-2016.
Orlowsky, B., and S. I. Seneviratne , 2013: Elusive drought: Uncertainty in observed trends and short- and long-term CMIP5 projections. Hydrol. Earth Syst. Sci., 17, 1765– 1781, https://doi.org/10.5194/hess-17-1765-2013.
Schneider, U., A. Becker, P. Finger, A. Meyer-Christoffer, B. Rudolf, and M. Ziese , 2011: GPCC full data reanalysis version 6.0 at 0.5°: Monthly land-surface precipitation from rain-gauges built on GTS-based and historic data. DWD, accessed 12 June 2022, https://doi.org/10.5676/DWD_GPCC/FD_M_V6_050.
Spinoni, J., and Coauthors , 2019: A new global database of meteorological drought events from 1951 to 2016. J. Hydrol. Reg. Stud., 22, 100593, https://doi.org/10.1016/j.ejrh.2019.100593.
Tasnim News Agency, 2021: The drought crisis covered 95% of Sistan and Baluchistan/we had the least rainfall (translated to English by Google Chrome). Accessed 13 July 2022, https://tn.ai/2626877.
Thom, H. C. S. , 1958: A note on the gamma distribution. Mon. Wea. Rev., 86, 117– 122, https://doi.org/10.1175/1520-0493(1958)086<0117:ANOTGD>2.0.CO;2.
United States Department of Agriculture, 2022: Crop calendars for Middle East (Iran, Iraq, Syria) and Turkey. Accessed 23 Aug 2022, https://ipad.fas.usda.gov/rssiws/al/crop_calendar/metu.aspx.
Vaghefi, S. A., M. Keykhai, F. Jahanbakhshi, J. Sheikholeslami, A. Ahmadi, H. Yang, and K. C. Abbaspour , 2019: The future of extreme climate in Iran. Sci. Rep., 9, 1464, https://doi.org/10.1038/s41598-018-38071-8.
Xie, P., and P. A. Arkin , 1997: Global precipitation: A 17-year monthly analysis based on gauge observations, satellite estimates, and numerical model outputs. Bull. Amer. Meteor. Soc., 78, 2539– 2558, https://doi.org/10.1175/1520-0477(1997)078<2539:GPAYMA>2.0.CO;2.
Yazdandoost, F., S. Moradian, A. Izadi, and A. Aghakouchak , 2021: Evaluation of CMIP6 precipitation simulations across different climatic zones: Uncertainty and model intercomparison. Atmos. Res., 250, 105369, https://doi.org/10.1016/j.atmosres.2020.105369.
