CESM2 (WACCM6): Stratospheric Evaluation

CESM2 (WACCM6): Stratospheric Evaluation D. Kinnison, M. Mills, R. Garcia, D. Marsh, A. Gettelman, F. Vitt, S. Glanville, C. Bardeen, S. Tilmes, A. Smith, J-F. Lamarque, L. Emmons, A. Conley, J. Richter, J. Bacmeister, Fei Wu, and William Randel (NCAR) Whole Atmosphere Working Group, Boulder, 13 February 2018

Evaluation of CESM2 (first cut) Chemistry Climate Model Initiative (CCMI): Status Examine Total Column Ozone (TCO) 2011: vs OMI, SBUV Examine T, HCl, ClO, HNO 3, and O 3 : vs Aura MLS Examine SH Temperature and NH SSWs: vs MERRA Compare AOA: CESM1 vs CESM2 Examine H 2 O Tape Recorder: vs Aura MLS Examine Temperature Trends: vs SSU and MSU Model Versions: CESM1 (WACCM-CCMI), tag34 version (released to community) CESM2 (WACCM), F-case and SD, #262a

CESM1-WACCM4 CCMI Simulations (~45 publications; 10 in review) Scenario Period Ocean RCP Members REF-C1 1955-2014 Data - 5 REFC1-fODS1960 1955-2014 Data - 5 REF-C1-PI 1850-1960 Data - 1 REF-C1SD / MERRA 1979-2014 Data - 1 REF-C2 1960-2100 Interactive RCP6.0 3 SEN-C2 2001-2100 Interactive RCP4.5 1 SEN-C2 2001-2100 Interactive RCP8.5 3 SEN-C2-fGHG1960 1960-2100 Interactive RCP6.0 3 SEN-C2-fODS1960 1960-2100 Interactive RCP6.0 3 SEN-C2-fODS2000 2000-2100 Interactive RCP6.0 3 SEN-C2-nVSL 1955-2100 Interactive RCP6.0 1 SEN-C2-fEmis 1960-2100 Interactive RCP6.0 1 SEN-C2-fCO21960 1960-2100 Interactive RCP6.0 1 SEN-C2-fN2O1960 1960-2100 Interactive RCP6.0 1 SEN-C2-fCH41960 1960-2100 Interactive RCP6.0 1 SEN-C2-CH4RCP85 2000-2100 Interactive Mixed 1

Column Physics Chemistry CCMI vs CESM2 Chemistry (Major Changes!) CESM1 (WACCM4) CCMI CESM2 (WACCM6) Horizontal Resolution 1.9 x2.5 0.95 x1.25 Vertical Layers 26/66/88 32/70/88 Boundary Layer HB CLUBB Shallow Convection Hack CLUBB Deep Convection ZM ZM Macrophysics R&K CLUBB Microphysics R&K MG 2.0 Radiation CAMRT RRTMG Aerosols Bulk MAM4 QBO Nudged to Observations Interactive (70L) Chemical Mechanism 180 species 228 Species, with SOAs Chemical rates JPL-11 JPL-15 Sulfate SAD Prescribed (CCMI) Interactive (MAM) ICE SAD Bulk Scheme MG 2.0 Solar Variability Lean ETF (SPEs) CMIP6 ETF (SPEs, MEEs) GHG abundances Meinshausen, 2011 (RCPs) Meinshausen, 2016 (SSP) Halogens WMO, 2010 Meinshausen, 2016; Future?

First we will Examine the representation of: Total Column Ozone (TCO); Polar de-nitrification (de-no Y ); Polar dehydration (de-h 2 O); Heterogeneous chlorine activation (ClOx); Local Polar ozone depletion. Polar decent (using N 2 O); The model results will be from the SD CESM2 (WACCM) / MERRA #262 simulation.

Latitude, degree Latitude, degree 90 60 30 0-30 -60 CCMI (2011) - 2deg -90 90 60 30 0-30 -60 CESM2 (2011) - 1deg -90 90 60 Aura OMI (2011) DU 480. 460. 440. 420. 400. 380. 360. 340. 320. 300. 280. 260. 240. 220. 200. 180. 160. 140. 120. 100. Total Column Ozone (TCO), SD configuration Comments: Overall, both model versions compare well to TOZ observations for year 2011. Tropics: CESM2 is more consistent with OMI. There is larger regions of low (260DU) TOZ. (Black circles) Polar NH (60N) Spring maximum: CESM2 is more consistent with OMI. (Green Box) Polar SH (60S) Spring maximum: CESM2 is more consistent with OMI. (Blue circle) Latitude, degree 30 0-30 -60-90 Day-of-Year Polar SH (>70S) Winter: CESM2 has less TCO (~10DU) than CCMI? A region where OMI doesn t observe. (red arrows)

Pressure, hpa Pressure, hpa Aura MLS Aura MLS Day-of-Year Aura MLS binned up by Chuck Bardeen, NCAR Temperature (K) H 2 O (ppmv) CESM2 (WACCM) CESM2 (WACCM) Day-of-Year 81S, 2011 SD Configuration 289. 283. 277. 271. 265. 259. 253. 247. 241. 235. 229. 223. 217. 211. 205. 199. 193. 187. 181. 175. 9.1 8.7 8.3 7.9 7.5 7.1 6.7 6.3 5.9 5.5 5.1 4.7 4.3 3.9 3.5 3.1 2.7 2.3 1.9 1.5 Model over dehydrates by ~0.4ppmv

Pressure, hpa Pressure, hpa Aura MLS Aura MLS Day-of-Year Temperature (K) HNO 3 (g) (ppbv) CESM2 (WACCM) CESM2 (WACCM) Day-of-Year 81S, 2011 SD Configuration The de-noy is underestimated by ~0.5ppbv. This comparison can be improved by changing the #part. cm -3 from 0.01 to 0.005. 289. 283. 277. 271. 265. 259. 253. 247. 241. 235. 229. 223. 217. 211. 205. 199. 193. 187. 181. 175. 16.0 14.0 12.0 9.0 8.0 7.0 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 0.5 0.0

Pressure, hpa Pressure, hpa Aura MLS Aura MLS Day-of-Year Temperature (K) HCl (g) (ppbv) CESM2 (WACCM) CESM2 (WACCM) Day-of-Year 81S, 2011 SD Configuration All CCMs tend to underestimate HCl depletion in winter. However by spring, HCl => ClOx and good ozone depletion is derived. 289. 283. 277. 271. 265. 259. 253. 247. 241. 235. 229. 223. 217. 211. 205. 199. 193. 187. 181. 175. 3.8 3.6 3.4 3.2 3.0 2.8 2.6 2.4 2.2 2.0 1.8 1.6 1.4 1.2 0.8 0.6 0.4 0.2 0.0

Pressure, hpa Pressure, hpa Aura MLS Aura MLS Day-of-Year Temperature (K) ClO (ppbv) CESM2 (WACCM) CESM2 (WACCM) Day-of-Year 81S, 2011 SD Configuration The model does an excellent job of het. activation of LS ClO in Spring. The US gas-phase chemistry is also well represented. 289. 283. 277. 271. 265. 259. 253. 247. 241. 235. 229. 223. 217. 211. 205. 199. 193. 187. 181. 175. 1.14 8 2 0.96 0.90 0.84 0.78 0.72 0.66 0.60 0.54 0.48 0.42 0.36 0.30 0.24 8 2 0.06 0.00

Pressure, hpa Pressure, hpa Aura MLS Aura MLS Day-of-Year Temperature (K) O 3 (ppmv) CESM2 (WACCM) CESM2 (WACCM) Day-of-Year 81S, 2011 SD Configuration In Winter, the model underestimates O 3 consistent with the low TCO. In Spring, O 3 is well represented, within 0.3ppmv of obs. 289. 283. 277. 271. 265. 259. 253. 247. 241. 235. 229. 223. 217. 211. 205. 199. 193. 187. 181. 175. 5.7 5.4 5.1 4.8 4.5 4.2 3.9 3.6 3.3 3.0 2.7 2.4 2.1 1.8 1.5 1.2 0.9 0.6 0.3 0.0

Pressure, hpa Pressure, hpa 1 10 100 Aura MLS Aura MLS Day-of-Year Temperature (K) N 2 O (ppmv) 1 10 100 CESM2 (WACCM) CESM2 (WACCM) Day-of-Year 81S, 2011 SD Configuration The decent rate of N 2 O in SH Polar Winter / Spring is well represented by the nudging of CESM2 to MERRA. 289. 283. 277. 271. 265. 259. 253. 247. 241. 235. 229. 223. 217. 211. 205. 199. 193. 187. 181. 175. 285. 270. 255. 240. 225. 210. 195. 180. 165. 150. 135. 120. 105. 90. 75. 60. 45. 30. 15. 0.

Next we will Examine the representation of: SH Polar temperature biases and NH SSWs; Representation of the TCO; Tropical H 2 O tape recorder; Stratosphere mean AOA; Global temperature trends. The model results will be from the Fcase #262a simulation (1975-2014) and the previously discussed SD CESM2 (WACCM) / MERRA #262 simulation.

CESM2: Temperature Bias Reduced in the SH Polar Region (90S-60S) 262a FR Configuration, CESM2, #262 MERRA WACCM results within ± 1 degrees in Sept. and October. 262a-MERRA 262a-MERRA Figures Courtesy of R. Garcia, NCAR CESM2 includes a new Ridge scheme that excites Orographic GW (Julio Bacmeister, NCAR).

Northern Hemisphere Sudden Stratospheric Warming (SSWs) FR Configuration, CESM2, #262 SSW definition: zonal wind reversal at 10hPa, 60 N. Promising agreement between MERRA and CESM2 Fcase. Possibly too few events in February; however, more realizations are needed to properly characterize this diagnostic. Figure Courtesy of R. Garcia, NCAR

CESM1 (WACCM-CCMI) *** Southern Hemisphere 400 TCO [63S-90S] - October Observations CESM1 (WA-Fcase), CCMI (5) 350 Total Column Ozone (DU) 300 250 200 Dashed line is ~1980 TOZ Blue line is ensemble mean Symbols are individual realizations 150 1950 1960 1970 1980 1990 2000 2010 2020 Year Similar variability between CESM1 (WACCM) and the observations post 1990.

CESM2-SD *** Southern Hemisphere 400 350 TCO [63S-90S] - October Observations CESM1 (WA-Fcase), CCMI (5) CESM2 (WA-SD), #262 Total Column Ozone (DU) 300 250 200 Dashed line is ~1980 TOZ Blue line is ensemble mean Symbols are individual realizations 150 1950 1960 1970 1980 1990 2000 2010 2020 Year For the 2005-2014 period, CESM2 (WACCM-SD) compares well with observations.

CESM2 vs CCMI *** Southern Hemisphere 400 350 TCO [63S-90S] - October Observations CESM1 (WA-Fcase), CCMI (5) CESM2 (WA-SD), #262 CESM2 (WA-Fcase), #262a Total Column Ozone (DU) 300 250 200 Dashed line is ~1980 TOZ Blue line is ensemble mean Symbols are individual realizations 150 1950 1960 1970 1980 1990 2000 2010 2020 Year We do need to examine the pre-1975 period to make sure we are not biased during low Chlorine period.

CESM2 vs CCMI *** Northern Hemisphere 540 500 Observations CESM1 (WA-Fcase), CCMI (5) CESM2 (WA-SD), #262 TCO [63N-90N] - March Total Column Ozone (DU) 460 420 380 340 Dashed line is ~1980 TOZ Blue line is ensemble mean Symbols are individual realizations 300 1950 1960 1970 1980 1990 2000 2010 2020 Year Extreme depletion events in the observations for several years during the 1990s and year 2011.

CESM2 vs CCMI *** Northern Hemisphere 540 500 TCO [63N-90N] - March Observations CESM2 (WA-Fcase), #262a CESM1 (WA-Fcase), CCMI (5) CESM2 (WA-SD), #262 Total Column Ozone (DU) 460 420 380 340 Dashed line is ~1980 TOZ Blue line is ensemble mean Symbols are individual realizations 300 1950 1960 1970 1980 1990 2000 2010 2020 Year The pre-1980 period is in good agreement. The higher chlorine loading period has a few extreme depletion years.

CESM2 at Halley (75.35ºS, 26.39ºW) Total Column Ozone (DU) 400 350 300 250 200 150 100 50 Halley Obs September 1960 1980 2000 2020 400 350 300 250 200 150 100 50 October CESM1 Fcase (3) CESM2 SD CESM2 Fcase 1960 1980 2000 2020 450 400 350 300 250 200 150 100 50 November 1960 1980 2000 2020 Total Column Ozone (DU) 450 400 350 300 250 200 150 100 Halley Obs December 1960 1980 2000 2020 Year 400 350 300 250 200 150 January 1960 1980 2000 2020 Year 350 325 300 275 250 225 200 February 1960 1980 2000 2020 Year The CESM2 (WACCM) better represents both December and January TCO compared to CESM1 (WACCM-CCMI).

Polar Temperature (K), 78ºS, zm Aura MLS, V4 Conclusions: CESM2 (WACCM) overall is in good agreement in both magnitude and seasonality with observations. CESM2 (WACCM) - Fcase There does seem to be an enhanced warm peak in the model near hpa that is not seen in the observations?

Polar H 2 O (ppmv), 78ºS, zm Aura MLS, V4 Conclusions: CESM2 (WACCM) vertical Structure is consistent with observations. CESM2 (WACCM) - Fcase The model is biased ~0.4 ppmv low in the upper stratosphere. The model minimum H 2 O in the lower stratosphere vortex is consistent with observations.

Polar ClO (ppbv), 78ºS, zm Aura MLS, V4 Conclusions: CESM2 (WACCM) vertical Structure is consistent with observations. CESM2 (WACCM) - Fcase Model is biased ~0.05 pptv low in the Upper Stratosphere. Model LS vortex ClO is in good agreement with observations. Aura MLS does show minimum values of 0.05 pptv that is not derived by the model.

Polar O 3 (ppmv), 78ºS, zm Aura MLS, V4 Conclusions: Overall, CESM2 (WACCM) shows good seasonality compared to observations. CESM2 (WACCM) - Fcase The model underestimates the minimum ozone in the LS, Spring, by ~0.3 ppmv (similar to SD). The peak ozone between 5-10 hpa is underestimated by the model by ~0.5ppmv?

H 2 O Tape Recorder *** 10ºS-10ºN *** 2005-2012 Very good representation of the cold phase of the TR. White lines are Aura-MLS, v4 observations WACCM-SD (MERRA) has ~0.5ppmv more entry level H 2 O. SD and FR have ~0.5ppmv more H 2 O during the wet phase of the TR Figure Courtesy of Sasha Glanville, NCAR.

Mean AOA 5.0 Annual Mean AOA *** 1hPa *** 23S-23N CESM1 (WACCM-CCMI) Fcase (3) CESM2 (WACCM), Fcase, #262a 4.5 Mean Age (years) 4.0 3.5 Blue line is ensemble mean 3.0 1960 1970 1980 1990 2000 2010 2020 Years Comment: The mean AOA is ~ 0.5 years older in CESM2 (WACCM) vs CESM1 (WACCM-CCMI). The trend in AOA between the two model versions is similar over the hindcast period.

0.001 Annual Mean AOA *** 23S-23N Mean AOA Climatologies Pressure, hpa 0.010 00 00 CESM1 (WACCM-CCMI) Fcase CESM2 (WACCM), Fcase, #262a Comment: Both the tropical average and global average mean AOA profiles are similar. Pressure, hpa 00 00 10-year Climatology (years: 2003-2012) 0 1 2 3 4 5 6 Mean Age (years) 0.001 0.010 00 00 Annual Mean AOA *** 90S-90N CESM1 (WACCM-CCMI) Fcase CESM2 (WACCM), Fcase, #262a The mean AOA is profile is similar between the two model version up to ~20hPa. More work is needed to understand why the mean AOA has increased in CESM2 (WACCM). Rolando Garcia will be doing a detailed TEM analysis of both model versions ASAP. 00 00 10-year Climatology (years: 2003-2012) 0 1 2 3 4 5 6 Mean Age (years)

Temperature Trend *** CESM1 (WACCM-CCMI) The CESM1 (WACCM-CCMI) Fcase compares well to observations. Here 5- realizations are show. This simulation used CCMI aerosol loading and mean radius data. See Randel et al., JGR, 2017 for details.

Temperature Trend *** CESM2 (WACCM) The CESM2 (WACCM) Fcase also compares well to observations. This simulation use aerosol loading and mean radius information from the interactive MAM approach (Mills et al., JGR, 2016). Here only one realization is shown. Figure Courtesy of Fei Wu and Bill Randel, NCAR.

General Conclusions for CESM2 (WACCM) Free-Running (Fcase): CESM2 Total Column Ozone (TOZ) is within the range of previous CCMI results and observations. The tropical H 2 O tape recorder is well represented (best I ve seen from a model simulation). Mean AOA is older in FR by ~0.5 years (relative to CCMI) a detailed TEM analysis will be conducted by R. Garcia in the next few weeks. Global temperature trends are well represented compared to SSU and MSU observations. Volcanic periods also match obs (i.e., using the new interactive MAM approach)! Specified-Dynamics (year 2011) CESM2 TCO is consistent with Aura OMI observations for year 2011. CESM2 has less TCO (relative to CCMI) from May to August at high SH latitudes. The cause of this difference is unknown. The low CESM2 TCO results are consistent with the high latitude evolution of O 3 relative to Aura MLS O 3. This ozone bias is most likely a dynamical issue since chemistry is not very active during polar winter. Comparison to Aura MLS for HNO 3 (g), HCl(g), H 2 O(g), ClO, O 3, and N 2 O are overall very well represented by CESM2 (WACCM).