Lab Sea Ice Analysis of Coupled Simulations Atmosphere
Lab Sea Ice: Analysis of Coupled Simulations Atmosphere Perspective Gettelman, Neale, Bacmeister et al.
Summary of Simulations (from Holland)
Summary • Good: 202, 205, 211 • Ugly: 210, 215, 217, 218 • 210: yr 60, 215 yr 50, 217 yr 30 -40, 218 yr 50 • Bad: Too early? 219. Guessing it should go bad. • Apologizing for lack of figures (no time to add, could do that) • All are available in atmospheric diagnostics for coupled runs.
What went ‘bad’: Atmosphere perspective • Atms: in 217 see warmer everywhere but cold Lab sea/ N Atlantic. Same in 218 (early). 215 is colder generally NH but also lab sea/N. Atlantic. Also in 210 (hard to tell exactly from diags). • So runs that go ‘bad’ have cold N. Atlantic SSTs before they go bad • 217 (early: 20 -39 v. 202) • Clearly colder N. Atlantic than 202, warmer everywhere else • BUT: Atmosphere has MORE downward SW flux into that region (FSDS)? LW similar (FLDS, FLNS) • Big RESSURF difference. RESSURF = FSNS + FLNS + LHFLX + SHFLX • Due to SHFLX and LHFLX. Loosing less heat at surface, so net flux is higher. Heat into ocean? • SAME FOR 215 • If SST decreasing some other process is cooling SST? Not atmospheric Rad flux • Turbulent fluxes not in RESSURF. • Surface stress: Decreasing Surface Westerlies in 215 and 217. Might be a turbulent effect. • But: if look at first 10 -30 years: do not see especially cold SSTs. • So it evolves. • But: wind stress is lower, from the beginning? • 210 is not as bad (for RESSURF, also LH +SH flux), Surf stress not available. • 210 also not that different in TS. • Does not look like a local rad forced signal. SSTs go down. • Turbulent fluxes might be an issue. Lower from beginning of runs? 215, 217, 218. 219 and 210 not as low.
LWP and Precipitation • LWP & Precip Is it different in runs that went off? • I. e. : more lwp and more precip = less FSDS and freshening • NO: generally LESS precip in runs that went off. LWP ambiguous • Could LESS Precip drive cooler temps? • Most of these differences are AFTER year 30: not much before that is different between good and bad runs: so likely a response? • LESS LWP in 217 than 202. JJA and DJF. Slightly less precip • 215: more LWP than 202 to E. JJA. Less over Lab sea DJF. Less Precip • 210 v. 205: less precip in N. Atlantic. More LWP to E. again (less in Lab sea)
219 (215 + ice albedo change) • TS, & RESSURF anomalies v. 211: look more like ‘bad’ runs. • Hard to tell exactly: but 219 TS COLDER than 218. (based on obs comparisons) • Wind stress in 219 a little stronger than 218. • So why did this run not go off? Is it the wind stress? • 219 looks more like the ‘bad’ runs: 211, 217, 218
E-P (Evap – Precip) • Is there a difference in P-E between good and bad runs? • P-E < 0 in N. Atlantic • 218 – 211: more negative in 218: already bad. (But 218 v. 202 same: early on) • 217 – 202 : slightly more negative. (note: lab sea a little more positive DJF in 202, 211 than 217). Could be 30%? • 210: not much difference in P-E v. 205 (Similar to 219? ) Early and mid run before freeze. • 219: less negative P-E (but still a bit lower than good runs) • Consistent with lower P? But only half of this is P. • Not really consistent for good and bad simulations. • Do not see effects in first few years: maybe a result of the changes.
Surface RH? Winds? • From Vince: Is the surface relative humidity or temperature too large, or is it more a problem of weak surface winds? I. e. is the problem scalar vertical mixing or momentum vertical mixing? In the areas of weak wind, is there too much turbulence or not enough? • Seems like weaker wind in bad runs • Don’t have surface RH in diagnostics • Where the surface winds are too weak, is the wind aloft too weak? • Not clear: U 200 maybe < 2 m/s weaker (i. e. small on 20 m/s). Within 1 contour in N. Atlantic. 1 st level wind is <1 -2 m/s weaker (on 7 m/s). • Zonal average zonal wind is NOT weaker. (similar).
Summary • There are some consistent differences in the atmosphere in runs that go ‘bad’ (runaway lab sea ice): 210, 215, 217, 218 • Evolution is that SSTs get cold along with different surface LH and SH fluxes • Wind stress: lower Westerly wind speeds in ‘bad’ simulations (from beginning). 210 and 219 are ‘not as weak’ for winds (210 goes bad ‘later’) • Surface radiative fluxes in N. Atlantic do not change much between simulations. I. e. : this does NOT look like a cloud issue (they don’t even respond as the surface does) • Surface wind stress in N. Atlantic is weaker in ’bad’ runs • 219 (215+darker sea ice albedos): looks like a ‘bad’ run, but if it does not go bad, maybe the ice albedo helps?
Questions. Comments? • No Ocean diagnostics posted for most of these runs. • Would be nice to see them. • What is going on with history of salinity in N. Atlantic region? What is happening to ocean mixing? • Is there a logical chain for how the ice grows • From atmosphere: see signals of cooling and surface flux changes before freeze over. Also in 219: not frozen yet. • From atmosphere side: Not clouds, maybe turbulent heat flux? • Will the ice albedo fix things? • Is it a spin-up, adjustment issues? I. e. Sea ice ‘settles down’ • Would like to startup 217 for example from end of 202 simulation.
- Slides: 10