Proj1.5 Notebook

Tue Dec 20 11:47:07 GMT 2016

Spinup jobs (non-interactive ice/defined GrIs topog). Ice tile fractions are initialised as 10km of ice (i.e., snow at ice-density).

• xmvlf: CTL run. Present day SST/SICE.
• xmvlg: Perturbed. RCP4.5 SST/SICE from Jonathan.

Script: Annual 1.5m temp timeseries

import pylab as py
import sys
# Add to PYTHONPATH to make global:
sys.path.append('/DUM/DUM/Projects/FAMOUS/lib')

import umfile
import umfield

expid='xmvlf'
expid='xmvlg'

fpref='/DUM/DUM/Data/plane-01/DUM/FAMOUS/'+expid+'/'+expid+'a#pa00000'

marr=['ja','fb','mr','ar','my','jn','jl','ag','sp','ot','nv','dc']
stash=3236

#nyr=100

nyr=99 # As final year is missing Dec
yr1=1958

nmn=len(marr)


t15a=py.zeros(nyr,dtype='f8')

for yy in range(nyr):
 year=yr1+yy
 print year
 inf1=fpref+str(year)
 if yy == 0:
   infile=inf1+marr[0]+'+'
   hvals=umfield.get_pphead(infile)
   nx=hvals.ihead[0][18]
   ny=hvals.ihead[0][17]

 mdum=py.zeros([nmn,ny,nx],dtype='f8')
 for mm in range(nmn):
   infile=inf1+marr[mm]+'+'
   tarray=umfield.get_fieldob(infile,stash=stash)
   mdum[mm,:,:]=tarray.data

 t15a[yy]=mdum.mean()

py.save('t15m_'+expid+'_'+str(yr1),t15a)

py.figure(2)
py.clf()
py.plot(t15a)
py.title('t1.5 '+expid)
py.show()


#Tue Dec 20 17:06:37 GMT 2016

import pylab as py
import time
import sys

toc=273.15
ctl=py.load('t15m_xmvlf_1958.npy')
prt=py.load('t15m_xmvlg_1958.npy')

nt=ctl.size
dtime=py.arange(nt)+1958

tsr=(time.strftime("%d/%m/%Y %H:%M"))
fname=(sys.argv[0])
py.figure(10)
py.clf()
fig,ax=py.subplots()
ax.plot(dtime,ctl-toc,'r',linewidth=2,label='CTL (xmvlf)')
ax.plot(dtime,prt-toc,'b',linewidth=2,label='PRT (xmvlg)')
ax.legend()
ax.grid()
fig.text(0.8, 0.45, tsr,
         fontsize=40, color='gray',
                  ha='right', va='bottom', alpha=0.5)
fig.text(0.8, 0.55, fname,
         fontsize=40, color='gray',
                  ha='right', va='bottom', alpha=0.5)
py.xlim([1950,2060])
py.ylim([2,9])
py.title('Global Average 1.5m Temperature')
py.ylabel('Temp (C)')

py.show()

py.savefig('temp1.5m_both.png')
#Tue Dec 20 17:54:23 GMT 2016

Thu Dec 22 09:47:56 GMT 2016

Found a problem with the SST infilling (for coastal tiling). Land mask was set to zero, and this was seeping through as a valid (fractional) surface temperature ... leading to cold spots in the Indian ocean. Tropical location had large weighting on global mean.

import pylab as py
import umfile
from copy import copy

# This does "MORE" infilling, wrt coastal tiling
# Originally from .../Projects/FAMOUS/Boundary

def box_av(inb,sw=0):
  nn=inb.size
  if nn != 9:
    print 'error'
    raise
  old=inb[1,1]
  inb=inb.reshape(9)
  ick=py.where(inb > -1000)[0]
  if len(ick) > 0:
    res=inb[ick].mean()
  else:
    if sw == 0:
       res=0
    else:
       res=old
  return res

def fill_in(inarr,sw=0):
  print 'fill',sw
  nn=inarr.shape
  nx=nn[1]
  ny=nn[0]
  nx2=nx+2
  temparr=py.zeros([ny,nx2],dtype=inarr.dtype)
  res=py.zeros([ny,nx],dtype=inarr.dtype)
  temparr[:,1:nx+1]=inarr
  temparr[:,0]=inarr[:,nx-1]
  temparr[:,nx2-1]=inarr[:,0]
  #print temparr[27,:]
  for yy in range (1,ny-1):
   for xx in range(1,nx2-1):
     if temparr[yy,xx] < -1000:
       bdat=temparr[yy-1:yy+2,xx-1:xx+2]
       res[yy,xx-1]=box_av(bdat,sw=sw)
     else:
       res[yy,xx-1]=temparr[yy,xx]
  res[0,:]=inarr[0,:]
  # Bodge
  res[ny-1,:]=inarr[ny-1,:]
  return res

# Suck in SST climatologies and pop them out again to cope with coastal tiling

infile='/DUM/DUM/Data/acacia-11/FAMOUS/ancil/qrclim.sst'
fbtype,idtype,rdtype,bits,bytes=umfile.ftype(infile)

fixhdr=umfile.get_fixhdr(infile,idtype)
intc,e1=umfile.get_const(infile,fixhdr,idtype,rdtype,bytes,field='intc')
realc,e2=umfile.get_const(infile,fixhdr,idtype,rdtype,bytes,field='realc')
ihead,ec=umfile.get_const(infile,fixhdr,idtype,rdtype,bytes,field='lookupi')
rhead,ed=umfile.get_const(infile,fixhdr,idtype,rdtype,bytes,field='lookupr')

nsh=ihead.shape
print nsh
np=nsh[0]

dstart=ihead[0,28]*8
offset=ihead[0,28]*8
dpack=ihead[0,29]

datao=py.zeros([np,dpack],dtype=rdtype)
for i in range(np):
  offset=ihead[i,28]*8
  dum=py.memmap(infile,dtype=rdtype,offset=offset,shape=dpack,mode='r')
  dum2=copy(dum)
  cum=dum2[0:1776].reshape(37,48)
  dfix1=fill_in(cum,sw=1)
  dfix=fill_in(dfix1,sw=1)
  dum2[0:1776]=dfix.reshape(37*48)
  datao[i,:]=dum2


fixhdr_new=py.copy(fixhdr)
### Write file
f=open('/DUM/DUM/Data/acacia-11/FAMOUS/ancil/qrclim.sst-FILL3','wb')
print f.tell()
f.write(fixhdr_new)
f.write(intc)
f.write(realc)
print f.tell(), (fixhdr_new[149]-1)*8
for i in range(np):
  f.write(ihead[i,:])
  f.write(rhead[i,:])
print f.tell(), (fixhdr_new[159]-1)*8, 'here and f159'
f.seek((ihead[0,28])*8,0)
print f.tell(), (ihead[0,28])*8, 'here again and ihead0,28'
for i in range(np):
  f.write(datao[i,:])




f.close()






#makesst1v3.py
#Thu Dec 22 09:57:50 GMT 2016

• xmvlf re-run submitted with corrected SST. Run so far indicates diff(T_1.5m) of ~ 2.75K. This equated to a Fettweis/SMB anomaly of 409 Gt.
• xmvlg extended: another 300 years submitted (previous runs suggest a 100-year snow spinup is not enough).
• Annual 1.5m temperatures Dec-to-Dec calculated (alternative to above).

New Runs:

• xmvlh: perturbed run with alternative (maybe warmer) SSTs from Jonathan (400 years).
• xmvli: perturbed run with alternative (maybe warmer) SSTs from Jonathan + perturbed CO2(400 years).

mix_ratio=(44/28.8) * (ppm_val/1000000)

Sat Dec 24 17:39:37 GMT 2016

Tue Jan 3 15:28:22 GMT 2017

Problems

• xmvlf: 2058-2182.
• xmvlg: 2058-2337.
• xmvlh: 1958-2212 (actually showing as still running at 12x100%. KILLED).
• xmvli: 1958-2211.

Attempted fix

• xmvlf: Restarted at 2182 for a 175 year run (complete).
• xmvlg: Restarted at 2337 for a 20 year run (complete).
• xmvlh: Restarted at 2212 for a 145 year run (complete).
• xmvli: Restarted at 2211 for a 146 year run (complete).

Wed Jan 4 10:10:34 GMT 2017

GrIS Surface Mass Balance (SMB) from model, compared with Fettweis value

Comparison of SST winter/summer forcing

TODO: Check with ECMWF derived SST climatology.

Thu Jan 5 12:19:06 GMT 2017

xlmvlf

SST Data/acacia-11/FAMOUS/ancil/qrclim.sst-FILL3
SIC Data/acacia-11/FAMOUS/ancil/qrclim.ice-FILL

xmvlg

SST jonathan/ancil/sst_FAMOUS_HadGEM2-ES_rcp45_2280-2299.anc
SIC jonathan/ancil/sic_FAMOUS_HadGEM2-ES_rcp45_2280-2299.anc

xmvlh/xmvli

SST jonathan/sst_FAMOUS_IPSL-CM5A-LR_rcp45_2180-2200.anc
SIC jonathan/sic_FAMOUS_IPSL-CM5A-LR_rcp45_2180-2200.anc

Fri Jan 6 15:34:01 GMT 2017

xmvlj

SST jonathan/ancil/sst_FAMOUS_HadGEM2-ES_piControl_2280-2299.anc
SIC jonathan/ancil/sic_FAMOUS_HadGEM2-ES_piControl_2280-2299.anc

xmvll

SST jonathan/ancil/sst_FAMOUS_IPSL-CM5A-LR_piControl_2181-2200.anc
SIC jonathan/ancil/sic_FAMOUS_IPSL-CM5A-LR_piControl_2181-2200.anc

Mon Jan 9 15:15:46 GMT 2017

• xmvlj: 200 year run (complete).
• xmvll: 200 year run (complete).

Tue Jan 10 12:17:31 GMT 2017

Wed Jan 11 15:53:13 GMT 2017

Thu Jan 12 10:48:12 GMT 2017

• SP Solid precip (snow): GrIS sum = 468 Gt
• SMB: GrIS sum = 275 Gt.
• AB Difference (ablation/melt): GrIS sum = 193 Gt. These are also the mapped values.

xmvlg (PERT) vs. xmvlj (CTL)

• dSP = 453 - 396 = 57 Gt (increase in precip)
• dSIMB = 140 - 263 = -123 Gt (decrease in accumulation)
• dAB = 313 - 133 = 180 Gt (increase in ablation/melt)

xmvi (PERT) vs. xlmvl (CTL)

• dSP = 418 - 326 = 92 Gt (increase in precip)
• dSIMB = 146 - 234 = -88 Gt (decrease in accumulation)
• dAB = 272 - 92 = 180 Gt (increase in ablation/melt)

mon Jan 16 12:00:00 GMT 2017

Working on a calculation of albedo (α) using grain size.
From Kuipers(2011), and using a simplified form of equation (4) (defined in equation (5)).

α = 1.48-(1.27048 *  r**0.07)

Tue Jan 17 15:08:43 GMT 2017

1|    1 |    3 |  335 |NET DOWN SURFACE SW FLUX ON TILES

Short runs, for all configurations; started at 2157, 30-years in length (acacia-12).

• xmvlf: 2157, 30-years completed
• xmvlg: 2157, 30-years completed
• xmvlh: 2157, 30-years completed
• xmvli: 2157, 30-years completed
• xmvlj: 2157, 30-years completed
• xmvll: 2157, 30-years completed

New Simulations

SST jonathan/ancil/sst_FAMOUS_HadGEM2-ES_abrupt4xCO2_1989-2008.anc
SIC jonathan/ancil/sic_FAMOUS_HadGEM2-ES_abrupt4xCO2_1989-2008.anc

SST jonathan/ancil/sst_FAMOUS_MIROC5_historical_1980-1999.anc
SIC jonathan/ancil/sic_FAMOUS_MIROC5_historical_1980-1999.anc

SST jonathan/ancil/sst_FAMOUS_MIROC5_rcp45_2081-2100.anc
SIC jonathan/ancil/sic_FAMOUS_MIROC5_rcp45_2081-2100.anc

• xmvlm: 1957, 200-years submitted UM2
• xmvln: 1957, 200-years submitted UM2
• xmvlo: 1957, 200-years submitted UM2

Wed Jan 18 17:27:37 GMT 2017

Thu Jan 19 11:55:14 GMT 2017

Wed Jan 25 11:17:51 GMT 2017

             IF (gbm_rho_snow(l) .GT. 600.0)
    &             fsnow(l) = 1.0 - ((gbm_rho_snow(l)
    &                                       - 600.0) / 200.0) 
             fsnow(l) = MIN(fsnow(l), 1.0)
             fsnow(l) = MAX(fsnow(l), 0.0)

               ALB_TYPE(L,N,BAND) = FSNOW(L)*ALB_SNOW(L,N,BAND) +
     &                          (1. - FSNOW(L))*ALB_TYPE(L,N,BAND)               

!!! Thus, at gbm_rho_snow=800, ALB_TYPE ignores any ALB_SNOW contribution
!!! Reverts to bare ice, with a starting albedo of 0.78, which is adjusted
!!! wrt to the tile temperature
        TM=273.15
        aicemax(1) = 0.78         
        daice(1) = -0.075  

        alb_type(l,n,1) = aicemax(1) 

        IF (tstar_tile(l,ntiles-nelev+k) .GT. (tm - 1.0)) THEN
          dt = tstar_tile(l,ntiles-nelev+k) - (tm - 1.0) 
          alb_type(l,n,1) = aicemax(1) + (dt * daice(1))

Lowest ice-elevation area > 4C  1.76376007122 %
Lowest ice-elevation area 3-4C  11.5015728019 %
Lowest ice-elevation area 2-3C  32.1883520289 %
Lowest ice-elevation area 1-2C  54.546315098 %
Percentage check... 100.0

MAR Ice Albedo

albedo = albedo_min + (albedo_max - albedo_min) * exp(-M *  K)
albedo_min,albedo_max - variable. In this paper set to 0.4,0.55 (elsewhere 0.3,0.55)
K - scaling factor (200 kg/m^2)
M - time-dependent  accumulated  excess  surface  meltwater  before runoff (kg/m^2).

FAMOUS Refreeze

Fri Jan 27 14:38:02 GMT 2017

Tue Jan 31 09:44:07 GMT 2017

Mon Feb 6 16:00:30 GMT 2017

Thu Feb 16 10:27:47 GMT 2017

stash1235 - TOTAL DOWNWARD SURFACE SW FLUX
stash3335 - NET DOWN SURFACE SW FLUX ON TILES   

albedo = 1.0 - (stash3335/stash1235)

From the regridded FAMOUS data

• xmvlp: copy of xmvlf, but with adjusted bare-ice characteristics.

tilealb.f

aicemax(1) = 0.55   # Previously  0.78
aicemax(2) = 0.55   # Previously 0.36 
daice(1) = -0.35    # Previously -0.075
daice(2) = -0.35    # Previously -0.075

# Minimum albedo allowed 0.2

Meeting 16-Feb-2017: notes

New Runs

• xmvln: MIROC5 CTL forcing for MAR comparison. 30-year run from 2057 started on um2. Complete

Fri Feb 17 17:27:59 GMT 2017

• xmvln: MIROC5 CTL forcing for MAR comparison. 30-year run from 2157 started on um2.
• xmvlo: MIROC5 RCP45 forcing for MAR comparison. 30-year run from 2157 started on um2.
• xmvlq: MIROC5 CTL forcing for MAR comparison with new bare-ice albedo. 30-year run from 2157 started on um2.
• xmvlr: MIROC5 RCP45 forcing for MAR comparison with new bare-ice albedo. 30-year run from 2157 started on um2.

Mon Feb 20 17:07:54 GMT 2017

Wed Feb 22 18:05:11 GMT 2017

• xmvls: MIROC5 CTL forcing for MAR comparison with new bare-ice albedo. Also max albedo of fresh snow in IR band changed from 0.7 to 0.98. 30-year run from 2157 started on um2.

Thu Feb 23 10:35:28 GMT 2017

Fri Feb 24 17:00:55 GMT 2017

Fri Mar 3 19:35:49 GMT 2017

Mon Mar 6 17:30:03 GMT 2017

Tue Mar 7 21:17:46 GMT 2017

Wed Mar 8 16:24:32 GMT 2017

Thu Mar 9 15:21:25 GMT 2017

Mon Mar 20 17:28:42 GMT 2017

Tue Mar 21 14:02:48 GMT 2017

Mon Mar 27 18:00:57 BST 2017

• xmvln: MIROC5 CTL forcing for MAR comparison. 10-year run from 2187 started on um2.
• xmvlo: MIROC5 RCP45 forcing for MAR comparison. 10-year run from 2187 started on um2.
• xmvlq: MIROC5 CTL forcing for MAR comparison with new bare-ice albedo. 10-year run from 2187 started on um2.
• xmvlr: MIROC5 RCP45 forcing for MAR comparison with new bare-ice albedo. 10-year run from 2187 started on um2.

Tue Mar 28 15:09:57 BST 2017

3314  Surface net radiation on tiles
3316  Surface temperture on tiles

Thu Mar 30 15:06:43 BST 2017

Fri Mar 31 10:31:45 BST 2017

Thu Apr 6 10:08:44 BST 2017

MAR related changes

FAMOUS analysis

!+ seg 4/2017
! Calculated version:
!        tgrad=(t(i,BL_LEVELS)-t(i,1))/bldepth
!
! Use fixed temperature lapse rate
! What to use?
!     DALR ~ 0.0098 deg/m  (9.8 deg/km)
! Model:
!     LAPSE=0.0065     !  NEAR SURFACE LAPSE RATE
!     LAPSE_TROP=0.003 !  TROPOPAUSE LAPSE RATE
! Set to latter
        tgrad=0.003
!- seg        

• xmvlt: As xmvlq, but with fixed temperature lapse rate for elevation calc.
• xmvlu: As xmvlr, but with fixed temperature lapse rate for elevation calc.

Tue Apr 18 16:56:36 BST 2017

Wed Apr 19 18:03:43 BST 2017

Thu Apr 20 14:34:23 BST 2017

Mon Apr 24 13:08:22 BST 2017

• xmvlt: from 2187.

Tue Apr 25 10:24:40 BST 2017

• xmvlv: Copy of xmvlt. From 2157 (30 years).

Wed Apr 26 10:51:22 BST 2017

New run (xmvlv) with 6K km-1 lapse on tiles

Fri May 19 12:01:18 BST 2017

• xmvlt: 3K/km tlapse (CTL)
• xmvlu: (RCP45)
• xmvlv: 6k/km tlapse (CTL) RERUN
• xmvlw: (RCP45) RERUN
• xmvlx: 6K/km tlapse;dLWdT = 5.4 (CTL) RERUN
• xmvly: (RCP45) RERUN

• xnkub: 6K/km tlapse;dLWdT = 3.6 (CTL)

Wed May 24 12:14:25 BST 2017

        tgrad=-0.006
! Specific humidity gradient still locally calculated
        qgrad=(q(i,BL_LEVELS)-q(i,1))/bldepth

        dLWdT  = 3.6
        lwgrad = dLWdT * tgrad

Change of SMB between CTL and RCP45

Thu Jun 29 10:05:23 BST 2017

Fri Jun 30 11:57:53 BST 2017

Reference (coupled FAMOUSA/GLIMMER) experiments

xlmpn: base coupled model with my fudge fixes.

xlmpr: xlmpn with Robin's updates to nsnow etc.

xnfua: xlmpn with my changes to the horizonal/vertical scheme implemented GCM→GLIMMER.

---

Thu Jul 6 16:10:32 BST 2017

Topography issues; what is real?

Looking at Figure 63a, with particular attention to the third row down. FAMOUS BG is the initial model orography (year 1), before coupling to the ice sheet. FAMOUS AG represents the orography after coupling. Third column is the difference field. The interactive region (i.e. vals > 0) is (on average) ~ 200m lower in the GLIMMER topography.

I did some offline calculations betwen the FAMOUS/GLIMMER data and this is confirmed in Figure 63b (a rough downscaling of the GLIMMER orography to that of FAMOUS gives similar results; so it's not a model issue).

The final row in Fig 63a shows the difference on height between our FAMOUS and the portable version. Generally, the portable version has lower elevation, though the Himalayas are higher.

I'm looking for some advice on this. One issue is what happens at the boundaries between the GLIMMER sub-domain and the surrounding model space (mountain ranges introduced).

One option, is to make the sub-domain Greenland only and try and match the orography (and the elevation fractions etc).

Figure 63a: Topography comparision

Figure 63b: Topography comparision offline

---

Fri Jul 7 00:28:50 BST 2017

Similar results comparing FAMOUS surface elevation with TBASE downscaled 2-minute data. Above zeros average difference ~200m.

Figure 64a: Topography comparision (FAM vs TBASE)

---

Tue Jul 11 20:00:28 BST 2017

New coupled setup (xnfuc). GLIMMER domain is now over Greenland only; as might be expected, the initial topographic adjustment is only over said region.

Figure 65a: Topography comparision (GLIMMER domain now Greenland only)

---

Thu Aug 10 10:08:09 BST 2017

Offline spinup (5000 years, with xnkub SMB. Figure 66b is from a run that tries to do something clever with the non-ice snow field (and fails).

Figure 66a: Ice sheet thickness: yr1, yr5000 and diff. Plus ice volume (km3)

Figure 66b: Ice sheet thickness: yr1, yr5000 and diff. Plus ice volume (km3)

---

Mon Aug 14 14:28:07 BST 2017

Modelled and observed surface ice velocities; modelled values are from offline pinup (i.e., ice-model with non-evolving SMB). Figure 67a: Ice surface velocity (magnitude) m/yr (model)

Figure 67b: Ice s urface velocity (magnitude) m/yr (MEaSUREs InSAR winter data)

Tue Aug 15 11:12:31 BST 2017

Interestingly, if you define the ice-shit run as using a 'hotstart' it expects to find a 3D temperature field in the input file ... but doesn't complain if it's not there (just creates a random, non-physical temp field).

Figure 68a: Ice temperature (model)

---

Fri Sep 8 10:30:58 BST 2017

• xnfuh: 3D Interpolation masks, with average SMB per elevation.
• sg2uh: 3D Interpolation masks, with halo SMB
• sg3uh: No 3D interpolation mask
• sg4uh: As xnfuh but with average elevation SMB set to zero.

Figure 69a: GLIMMER SMB

---

Fri Oct 6 14:21:31 BST 2017

Job monitor

---

Tue Nov 7 14:56:15 GMT 2017

Updated: Thu Nov 9 11:11:12 GMT 2017

Figure 70: Volume difference with nasty snow

Non-ice snow in out of GLIMMER (on FAMOUS side)

• sgfac: Snow in as per normal (x10 accel).
• sgnac: Zero (non-ice) now in (x10 accel).

Figure 71a: sgfac snow in

Figure 71b: sgfac snow out

Figure 71c: sgnac snow in

Figure 71d: sgnac snow out

The SMB is different, but not dramatically so:

Figure 72: SMB differences

---

Wed Nov 8 17:13:44 GMT 2017

Contemplating. Ice erodes but snow accumlates. Switches to ice when ice is zero. And so the process repeats. Anomalous accumulation point (in FAMOUS) ... biasing the GLIMMER input as smoothed over a large area.

Figure 73: Ice to no ice to ice

---

---

Thu Nov 9 10:28:26 GMT 2017

Figure 74: Non-ice snow smearing

---

Figure 75: Ice breakout from mask

---

Mon Nov 20 13:07:23 GMT 2017

Mass accumulation issues.

---

Thu Nov 23 17:57:04 GMT 2017

A closeup of plot 73.

Description

• Blue line is "potential" non-ice snow; interp from FAMOUS to GLIMMER space.
• Green line is ice-thickness in FAMOUS timesteps; dotted lines are internal accel steps.
• Red line is actual non-ice snow developed in GLIMMER space (accel time steps)

Process

• Ice is being ablated, and near the end of tstep49 ice is turned to snow.
• This amount of snow should be sent back to FAMOUS, and be part of the blue line at tstep50
• At tstep50 the box is free of ice, potential snow is turned into actual snow; red jump to blue.
• This is turned into ice and the snow converted into a negative anomaly to send back to FAMOUS.
• At tstep51 one would expect potential snow to be reduced, but it has increased.
• This is why I thought there was a sign error somewhere (but could find nothing).

Figure 76: Closeup of 73

---

Mon Dec 4 18:02:22 GMT 2017

New CTL run with outmask turned off. Also, outside the mask, the bathymetry is now set to -3000m.

RCP85 run started from a seemingly settled GrIS volume.

Figure 77: Ice sheet evolution

Snow Albedo

Considering the snow albedo, I've started another control run with a more realistic (visible) snow albedo (Roesch[2002]). Offline calculation of albedo ranges below.

After ∼500 years, GrIS is only experiencing a modest increase in volume (see above).

Figure 78: Snow albedo with grain size

---

Tue Dec 5 15:40:17 GMT 2017

Seasonal evolution of grain size on ice and non-ice tile fractions (one year). It's a bit messy, but the takehome is that we're only getting chunky grains on the ice fractions (which sounds correct). On the lowest elevation, non-ice snow grain size gets up to ∼1600 microns ( ∼0.65 albedo).

Figure 79: Snow albedo height/month

---

Thu Dec 14 10:43:29 GMT 2017

Looking at OLR TOA. The coupled FAMOUS/ice model with the original high albedo in the visible spectrum has a reduced OLR TOA. I assume this reflects the increased SW reflected by the big, shiny high ice blob.

Source	Annual average W/m^2
Standard FAMOUS ref xfhcu	237.322
Back of envelope	239.00
NOAA 1degree	237.541
FAM_ice original albedo sgfj4	231.831
FAM_ice albedo3 sgfj9	237.685
FAM_ice albedo4 sgfja	237.714
FAM_ice albedo5 sgfjv	237.835

Figure 80: Planetary OLR TOA

As per Figure 61

Figure 81: SMB components (cyan CTL)

---

Fri Dec 15 11:56:40 GMT 2017

As per Figure 61

Figure 82a: SMB components (magenta CTL: sgfj9)

Figure 82b: SMB components (browdn CTL: sgfjb)

Figure 82c: SMB components (original blue CTL: sgfj4)

---

Figure 83: SMB vs. Albedo parameter)

---

Mon Dec 18 13:02:58 GMT 2017

Figure 84: Anim 1

Figure 84b: Anim 2

---

Wed Dec 20 10:14:42 GMT 2017

Started a new CTL run (brown variant), with non-constrained ice.

So far, there is some spread (as to be expected without a significant percent of required calving), but it's not looking too bad. The new run is in orange in the general development plot, and here is a map of the ice-thickness evolution (so far).

Figure 85: ice thickness, no constraint

---

Tue Feb 13 11:27:33 GMT 2018

Looking at the endstates for some of the longer rcp85/4xCO2 runs.

Figure 86: Ice sheet endstates

Animations for acceleration/bounce.

• JAAAC(fields;acab;thk_diff)
• JAAAH
• JAAAI(fields)

Ice collapse in NE (jaaak)

Figure 87a: SLE jaaak

Figure 87b: Thickness jaaak

placer.

---