Supplementary Figure 1 A figure of changing surface air temperature and top-1m soil moisture: (A) Annual mean surface air temperature, and (B) top

Size: px

Start display at page:

Download "Supplementary Figure 1 A figure of changing surface air temperature and top-1m soil moisture: (A) Annual mean surface air temperature, and (B) top"

Winifred Knight
5 years ago
Views:

Supplementary Figure 1 A figure of changing surface air temperature and top-1m soil moisture: (A) Annual mean surface air temperature, and (B) top 1-m soil moisture averaged over

1 Supplementary Figure 1 A figure of changing surface air temperature and top-1m soil moisture: (A) Annual mean surface air temperature, and (B) top 1-m soil moisture averaged over California from CESM1. Surface air temperature increases by about 5 degrees over the period 1920 to 2080, while no significant change in the top 1-m soil moisture is projected. 1

2 Supplementary Figure 2 Seasonal cycle of precipitation and surface air temperature: Seasonal cycle of precipitation (top) and surface air temperature (bottom) averaged over California for the period of (solid) and that of (dashed) from CESM1. 2

3 Supplementary Figure 3 A figure of 30-year running variance and multivariate correlation using different observational rainfall datasets: (A) 30-year running variance of annual mean precipitation over California (similar to Figure 1D) from four different observational precipitation datasets including CRU, GPCC, UDEL, and CPCUnified and (B) running multivariate correlation between California water year rainfall and ENSO cycle using four observational precipitation datasets (similar to Figure 1G). 3

4 Supplementary Figure 4 A figure of the multivariate correlation using CMIP5: The multivariate correlations between the Niño-3.4, Niño-3.4(Y+1) and water year precipitation in California within a running 30- year window during the NDJ season. Statistical significance of the multivariate correlations is given as a color bar below, showing the number of the ensemble members that are significant with 90% confidence based on the F-test. 4

1G, (B) linear correlation between the GAR index and Niño-3.

5 Supplementary Figure 5 A figure of linear correlation of the GAR index: (A) Same figure as Fig.1G, (B) linear correlation between the GAR index and Niño-3.4, and (C) Niño-3.4(Y+1). Black line is from the CESM1, and red line is from the 20CR. 5

6 Supplementary Figure 6 A figure of the multivariate linear correlation using an individual member of CESM1: Same as Figure 1F except showing all 30 ensemble members of CESM1. 6

7 Supplementary Figure 7 A figure of relationship between SST and rainfall over Niño-3 region using the CESM1: Relationship between eastern equatorial Pacific (Niño-3 region: 5 S 5 N, 150 W 90 W) total rainfall and SST in the first 50 years ( ) and the last 50 years ( ) in the large ensemble of the CESM1. Total numbers of the extreme (moderate) El Niño in each period are shown in the upper left corner in red (green). Black points indicate normal or La Niña conditions. 7

8 Supplementary Figure 8 A figure of relationship between rainfall anomalies and SST anomalies over Niño-4 region using the CESM1: Relationship between rainfall anomalies and SST anomalies averaged over Niño-4 region (5 S 5 N, 160 E 150 W) in the first 50 years ( ) and the last 50 years ( ) in the large ensemble of the CESM1. Both are detrended using 2 nd order linear fit. Total numbers of the extreme, moderate, and weak La Niña in each period are shown in the upper left corner in red, green, and purple, respectively. Black points indicate normal or El Niño conditions. 8

9 Supplementary Figure 9 Histogram of the meridional SST gradient and the vertical temperature gradient using the CESM1: (A) Histogram of the meridional SST gradient in the eastern equatorial Pacific for all 30 members of CESM1, which is defined as the average SST over the east off-equatorial region (5 N 10 N, 150 E 90 W) minus the average over the eastern equatorial region (2.5 S 2.5 N, 155 E 120 W). (B) Histogram of the upper ocean vertical temperature gradient in the Niño-4 region, defined as the averaged of temperature in the top 50m minus temperature at 60m. 9

10 Supplementary Figure 10 Scatter plot of the extreme ENSO event change and that of the extreme water cycle extremes in California: Scatter plot of the extreme ENSO event change from the first 50 ( ) and the last 50 ( ) years and that of the extreme water cycles extremes in California including both wet and dry events. 25 models in the CMIP5 are marked with different symbols shown above and individual members of CESM1 are with gray dots and numbers. 10

11 Supplementary Figure 11 A histogram showing spatial correlation of ENSO: Spatial correlation of regressed SST with Niño3.4 index during boreal winter (December January February) from all 38 coupled climate models in CMIP5 (red bars) and the large ensembles of the CESM1 (gray dots). 11

12 Supplementary Figure 12 A histogram showing spatial correlation of ENSO precursor: Spatial correlation of regressed SST with Niño-3.4 index with 1-year lag (Y+1), which indicates the ENSO precursor pattern, during boreal winter (December January February) from 38 coupled climate models in CMIP5 (red bars) and the large ensembles of the CESM1 (gray dots). 12

13 Supplementary Figure 13 A histogram of annual mean precipitation over California: Blue bar represents the median of 30 members of the CESM1 with the minimum and maximum values. Four different observational precipitation datasets are shown in black bars including CRU; 1, GPCC; 2, UDEL; 3, and CPCUnified; 4,5. 13

14 Supplementary Figure 14 A histogram of root mean squared error (RMSE) of rainfall over California: Root mean squared error (RMSE) of the modeled rainfall over California compared to CRU dataset. RMSE is computed for each month and annual mean of the RMSE is plotted. All modeled and observed precipitation are interpolated into the same resolution of 0.9 latitude 1.25 longitude. 14

15 Supplementary Figure 15 A histogram of spatial correlation of the modeled rainfall over California: Spatial correlation of the modeled rainfall over California compared to CRU dataset. Spatial correlation is computed for each month and its annual mean is plotted. All modeled and observed precipitation are interpolated into the same resolution of 0.9 latitude 1.25 longitude. 15

16 Supplementary Figure 16 Three sub-regions in California: Three sub-regions in California divided by the latitudinal lines of 39 N and 36 N (red lines). 16

17 Supplementary Table 1 List of the coupled climate models from the CMIP5 examined and tested in the present study. Model name ACCESS1-0 ACCESS1-3 BCC-CSM-1.1 BCC-CSM-1.1m BNU-ESM CMCC-CESM CMCC-CM CMCC-CMS CanESM2 CCSM4 CNRM-CM5 GFDL-CM3 GFDL-ESM2G GFDL-ESM2M CSIRO-Mk3.6.0 FGOALS-g2 FIO-ESM Had-GEM2 Had-GEM2-CC Had-GEM2-ES INMCM4 Institute CSIRO (Commonwealth Scientific and Industrial Research Organisation, Australia), and BOM (Bureau of Meteorology, Australia) Beijing Climate Center, China Meteorological Administration College of Global Change and Earth System Science, Beijing Normal University Centro Euro-Mediterraneo per I Cambiamenti Climatici Canadian Centre for Climate Modeling and Analysis National Center for Atmospheric Research Centre National de Recherches Meteorologiques / Centre Europeen de Recherche et Formation Avancees en Calcul Scientifique, France NOAA, Geophysical Fluid Dynamics Laboratory Commonwealth Scientific and Industrial Research Organization/Queensland Climate Change Centre of Excellence (CSIRO-QCCCE) Institute of Atmospheric Physics, Chinese Academy of Sciences The First Institute of Oceanography, SOA, China Met Office Hadley Centre (additional HadGEM2-ES realizations contributed by Instituto Nacional de Pesquisas Espaciais) Institute for Numerical Mathematics IPSL-CM5A-LR IPSL-CM5A-MR IPSL-CM5B-LR MIROC5 MIROC-ESM MIROC-ESM- CHEM MPI-ESM-LR MPI-ESM-MR MRI-CGCM3 Institute Pierre-Simon Laplace Atmosphere and Ocean Research Institute (The University of Tokyo), National Institute for Environmental Studies, and Japan Agency for Marine-Earth Science and Technology Japan Agency for Marine-Earth Science and Technology, Atmosphere and Ocean Research Institute (The University of Tokyo), and National Institute for Environmental Studies Max Planck Institute for Meteorology (MPI-M) Meteorological Research Institute 17

18 NorESM1-M NorESM1-ME CESM1-BGC CESM1-CAM5- FV2 Norwegian Climate Centre (NCC) National Science Foundation, Department of Energy, National Center for Atmospheric Research 18

19 Supplementary Table 2 Skewness of the precipitation over Niño-3 region following ref. 5 and 6. Models with skewness larger than 1.0 is highlighted. Totally 25 models are selected and used for analysis. Model name Skewness of precipitation over Niño-3 region ACCESS ACCESS BCC-CSM1-1-M 2.72 BCC-CSM BNU-ESM 1.65 CANESM CCSM CESM1-BGC 1.82 CMCC-CESM 3.38 CMCC-CM 4.32 CMCC-CMS 3.83 CSIRO-Mk FGOALS-g FIO-ESM 0.50 GFDL-CM GFDL-ESM2G 1.32 GFDL-ESM2M 2.48 GISS-E2-H-CC 1.76 GISS-E2-H 2.40 GISS-E2-R-CC 0.69 GISS-E2-R 0.61 HadGEM2-AO 0.95 HadGEM2-CC 1.28 HadGEM2-ES 0.90 INMCM IPSL-CM5A-LR 0.63 IPSL-CM5A-MR 0.51 IPSL-CM5B-LR 1.81 MIROC MIROC-ESM-CHEM 0.06 MIROC-ESM 0.12 MPI-ESM-LR 3.98 MPI-ESM-MR 1.19 MRI-CGCM MRI-ESM NorESM1-ME 1.96 NorESM1-M

20 CESM1-CAM5-1-FV CESM ( ) 20

21 Supplementary note 1 Running variance and correlation with multiple observational datasets Using four different observational precipitation datasets including CRU 1, GPCC 2, UDEL 3, and CPCUnified 4,5 with the observed ENSO cycle based on the Hadley SST dataset 8 is shown in Supplementary Fig. 3. It is known that modulation of Californian rainfall by ENSO cycle is affected by the low frequency variability Pacific Decadal Oscillation (PDO) 9. However, increasing tendency in both running variance and ENSO correlation is also observed. Supplementary note 2 Linear correlation and multivariate linear correlation of individual member of the CESM1 Increase in the multivariate linear correlation (Fig. 1G) raises a question whether this increase is cause by strengthened relation to Niño-3.4 or Niño-3.4 (Y+1). Linear correlation only using a single index is also performed. Linear correlation between the GAR index and Niño-3.4 (Y) is shown in Supplementary Fig. 5B, and Niño-3.4 (Y+1) is shown in Supplementary Fig. 5C. It is noted that both Niño-3.4 (Y) and Niño-3.4 (Y+1) contribute to strengthen relation to ENSO cycle. Also, it is noted that there are some members exhibiting even more rapid increase in the correlation in Supplementary Fig. 6, despite the fact the increase in correlation by the ensemble mean of the CESM1 is weaker that that of observation in Fig. 1F. Supplementary note 3 Extreme El Niño and La Niña in the large ensemble of the CESM1 Extreme El Niño is defined as when the boreal winter rainfall averaged over equatorial Pacific (Niño-3 region: 5 S 5 N, 150 W 90 W) is over 5mm day -1, and moderate El Niño is defined as events with SST anomalies greater than 0.5 standard deviation of the control period and that are not extreme El Niño following ref. 7. Extreme La Niña is defined as events with SST anomalies over Niño-4 region (5 S 5 N, 160 E 150 W) is smaller than standard deviation of the control period ( ; historical 21

22 simulation period in our case) following ref. 6. Moderate (weak) La Niña is defined as those smaller than -1.0 (-0.5) standard deviation but greater than (-1.0) standard deviation. Supplementary Figure 7 shows the extreme El Niño events (red dot) increase more than four times, i.e., from 45 in the first 50 years ( ) to 200 in the last 50 years ( ), while the moderate El Niño (green dot) decrease slightly. Also, the extreme (moderate) La Niña events increase more than five times (almost double), i.e. from 11 (149) in the first 50 years to 63 (250) in the last 50 years, while weak La Niña occurrence decreases by around 30%, i.e., from 305 in the first 50 years to 221 in the last 50 years (Supplementary Figure 8). This is consistent with the previous studies based on the multiple climate models in CMIP5 6,7. Changes in both extreme El Niño and La Niña events are statistically significant using the Bootstrap method 10. The 1,500 DJF samples from the 30 members of CESM1 in the early period ( ) are re-sampled randomly with replacement to construct 10,000 realizations. The standard deviation of the extreme El Niño/La Niña frequency in the inter-realization is about 6.8/6.2 events per 1500 years, much smaller than the difference between the early ( ) and the later periods ( ) at 155/52 per 1500 years (Supplementary Figures 7 and 8). Also, the physical mechanisms to cause the increase of extreme El Niño/La Niña is demonstrated in Supplementary Figure 9, consistently with the previous studies 6,7. First, the weakening of the meridional SST gradient due to faster warming in the equatorial Pacific than the off-equatorial Pacific leads to a large increase in convective activity in the eastern equatorial Pacific region, i.e. a large increase in potentially extreme El Niño events (Supplementary Figure 9A). Vertical temperature gradient in the Niño4 region shows the thermocline shoals, i.e., becomes shallower because greenhouse gas forcing warms the upper ocean first, indicating an enhanced efficiency of the Ekman pumping, which is critical for extreme La Niña (Supplementary Figure 9B). Supplementary note 4 Performance of the CESM1 The performance of the simulated ENSO and the ENSO precursor is evaluated with the spatial correlation of SST with the Niño-3.4 (Y) and the Niño-3.4 (Y+1) index (Supplementary Figures 11 and 12). The spatial pattern of the ENSO (Supplementary 22

23 Figure 11) is quite robustly simulated by all CMIP5 models 11. On the other hand, there exists considerable diversity in simulating the ENSO precursor (Supplementary Figure 12). Both the ENSO and the ENSO precursor affect California rainfall 12,13. Therefore, it is important to have both are captured by the models. We analyzed the rainfall climatology computed in the historical period ( ) from all 30 members of the CESM1 and 39 climate models in CMIP5. Supplementary Figure 13 shows annual mean precipitation (mm day -1 ) averaged over California from models and 4 different observational precipitation datasets including Climate Research Unit (CRU) 1, Global Precipitation Climatology Centre (GPCC) 2, University of Delaware (UDEL) 3, and Climate Prediction Center s unified precipitation datasets (CPCUnified) 4,5. Observation is in the range of mm day -1. The median value of the CESM1 is 1.87, which is slightly wetter compared to the observation. However, the root mean squared error (RMSE; Supplementary Figure 14) and the spatial correlation (SCORR; Supplementary Figure 15) show robust performance of the CESM1. By ranking, the CESM is within top 10 in terms of the RMSE and the 1 st in terms of the SCORR. 23

24 Supplementary References 1 Harris, I., Jones, P. D., Osborn, T. J. & Lister, D. H. Updated high-resolution grids of monthly climatic observations - the CRU TS3.10 Dataset. International Journal of Climatology 34, , doi: /joc.3711 (2014). 2 Schneider, U. et al. GPCC's new land surface precipitation climatology based on quality-controlled in situ data and its role in quantifying the global water cycle. Theoretical and Applied Climatology 115, 15-40, doi: /s x (2014). 3 Legates, D. R. & Willmott, C. J. MEAN SEASONAL AND SPATIAL VARIABILITY IN GAUGE-CORRECTED, GLOBAL PRECIPITATION. International Journal of Climatology 10, , doi: /joc (1990). 4 Xie, P. et al. A Gauge-based analysis of daily precipitation over East Asia. Journal of Hydrometeorology 8, , doi: /jhm583.1 (2007). 5 Chen, M. et al. Assessing objective techniques for gauge-based analyses of global daily precipitation. Journal of Geophysical Research-Atmospheres 113, doi: /2007jd (2008). 6 Cai, W. et al. Increased frequency of extreme La Nina events under greenhouse warming. Nature Climate Change 5, , doi: /nclimate2492 (2015). 7 Cai, W. J. et al. Increasing frequency of extreme El Nino events due to greenhouse warming. Nature Climate Change 4, , doi: /nclimate2100 (2014). 8 Rayner, N. A. et al. Improved analyses of changes and uncertainties in sea surface temperature measured in situ sice the mid-nineteenth century: The HadSST2 dataset. Journal of Climate 19, , doi: /jcli (2006). 9 Mo, K. C. Interdecadal Modulation of the Impact of ENSO on Precipitation and Temperature over the United States. Journal of Climate 23, , doi: /2010jcli (2010). 10 Austin, P. C. & Tu, J. V. Bootstrap methods for developing predictive models. American Statistician 58, , doi: / (2004). 11 Taschetto, A. S. et al. Cold Tongue and Warm Pool ENSO Events in CMIP5: Mean State and Future Projections. Journal of Climate 27, , doi: /jcli-d (2014). 12 Mo, K. C. & Higgins, R. W. Tropical influences on California precipitation. Journal of Climate 11, (1998). 13 Wang, S. Y., Hipps, L., Gillies, R. R. & Yoon, J. H. Probable causes of the abnormal ridge accompanying the California drought: ENSO precursor and anthropogenic warming footprint. Geophysical Research Letters 41, , doi: /2014gl (2014). 24

Supplement of Insignificant effect of climate change on winter haze pollution in Beijing

Supplement of Insignificant effect of climate change on winter haze pollution in Beijing Supplement of Atmos. Chem. Phys., 18, 17489 17496, 2018 https://doi.org/10.5194/acp-18-17489-2018-supplement Author(s) 2018. This work is distributed under the Creative Commons Attribution 4.0 License.