ITCZ/Monsoon Model Intercomparison Project
Aiko Voigt, Michela Biasutti and Jack Scheff, August 4, 2015
4 August 2015: Clarified whether fields should be saved as means over the
output period or as snapshots.
14 July 2015: The land setup is revised and has changed with respect to
earlier versions. We also now include plots for the ECHAM6.1 AquaControl
and LandControl simulations for reference.
This document describes the simulation setup and the requested simulations for the modelintercomparison project organized within the 2015 WCRP Grand Challenge workshop on
ITCZ and monsoons. Five simulations in idealized aquaplanet and idealized continent setup
are performed to study the dynamics and model robustness of how carbon dioxide, land and
the orbital configuration affect the ITCZ and the monsoons as well as the interaction between
the two. We note that this MIP is not part of CMIP6, but we are very much interested in
discussions and exchanges with CMIP6 activities related to monsoons and the ITCZ.
1 Model setup
The models are to be coupled to a slab ocean with time-mean zonal-mean ocean heat transport (“q-flux”) given in section 4. The boundary conditions largely follow the CMIP5 aquaplanet protocol except for interactive sea-surface temperatures and the inclusion of a seasonal
cycle in insolation. The following parameters should be specified:
• CO2 : 348 ppmv
• CH4 : 1650 ppbv
• N2 O: 306 ppbv
• No Halocarbons (CFCs)
• Ozone following values for the AquaPlanet Experiment (see http://www.met.
reading.ac.uk/~mike/APE/ape_ozone.html)
• Total solar irradiance: 1365 Wm−2
• No radiative effects of aerosols
• Diurnal cycle in insolation
1
• Calendar: a 360 days calendar (each month has 30 days) is desirable, but if such a
calendar is not available then a 365 days calendar without leap years should be used
(second preference) or a 365 days calendar with leap years (third preference)
• Orbit: zero eccentricity, 23.5◦ obliquity, NH spring equinox on March 21; modelling
groups should check the timing of the NH spring equinox by looking at daily TOA
shortwave data because a correct timing is essential to meaningfully compare different
models (a deviation from March 21 by 1-2 days is considered acceptable)
• Slab ocean depth: 30 m
• Sea-ice formation should be inhibited, and ocean temperatures should be allowed to
cool below the freezing temperature of sea water (see, e.g., Kang et al., 2008; Voigt
et al., 2014)
For the land simulations, a rectangular continent extending from 30◦ S to 30◦ N and 0◦ E to
45◦ E should be introduced. The q-flux should be zeroed out over this region, which requires
a small uniform correction of the q-flux over ocean areas of -0.59 Wm−2 (see Sect. 4). Land
should be represented as a very shallow ocean of depth 0.1 m with increased surface albedo
and with decreased evaporation.
Regarding the land surface albedo: the land albedo should be the ocean albedo plus 0.07.
So if the ocean has an albedo of 0.06, the land should have an albedo of 0.13. For some
models the ocean albedo is not a constant number but depends on the zenith angle and the
ratio of diffuse vs. direct radiation. In this case, the land albedo should be the ocean albedo
calculated for the specific conditions plus 0.07.
Regarding decreased evaporation over land: over land we aim at a reduction of evaporation
compared to ocean. This is achieved by rescaling the moisture transfer coefficient, Cq , which
affects the surface evaporation via
E = Cq |V~1 |(q1 − qs ),
(1)
with subscripts 1 and s denoting values at the lowest model level and at the surface. Over
land, the transfer coefficient Cq should be halved, i.e.,
1
Cq → Cq · ,
2
(2)
which assuming no changes in surface wind speed and boundary-layer humidity will reduce
the evaporation by a factor of 2. Over ocean the transfer coefficient must not be decreased.
For the LandOrbit simulations, the orbital parameters should be set to
• eccentricity: 0.02
• obliquity: 23.5◦
2
• NH spring equinox on March 21
• longitude of perihelion=270◦ , implying that NH solstice occurs during aphelion.
This orbit roughly corresponds to Earth’s comtemporary orbit (Joussaume and Braconnot,
1997). The solar constant and greenhouse gases should be as in the other simulations.
Table 1 – List of requested simulations.
Simulation
years
specifications
AquaControl 15+30 aquaplanet control simulation (includes 15 years of spin up)
Aqua4xCO2 40
as AquaControl but with quadrupled CO2 ,
to be restarted from end of AquaControl (Dec 30 of year 45)
LandControl 40
as AquaControl but with idealized continent,
to be restarted from end of AquaControl (Dec 30 of year 45)
Land4xCO2 40
as LandControl but with quadrupled CO2 ,
to be restarted from end of LandControl (Dec 30 of year 40)
LandOrbit
40
as LandControl but with changed orbital parameters,
to be restarted from end of LandControl (Dec 30 of year 40)
2 Requested simulations
Table 1 lists the requested simulations. To quantify the transient response and the adjusted
radiative forcing, all of the simulations except AquaControl should be restarted from another
simulation as described in the table.
3 Requested Output
We request output following the standard output of CMIP5 (see http://cmip-pcmdi.llnl.
gov/cmip5/docs/standard_output.pdf). To facilitate the analysis of the simulations, the
data should be “cmorized”. By cmorizing we here mean that the variables are named according to the CMIP5 names and have the same units as in CMIP5, and that 3d-data is
interpolated to the 17 CMIP5 pressure levels (1000, 925, 850, 700, 600, 500, 400, 300, 250,
200, 150, 100, 70, 50, 30, 20, 10 hPa)
• Monthly-mean output: all fields of the CMOR Table Amon except fields related to
carbon mass flux and mole fractions of ozone etc., saved for all years and simulations
except the 15 years of initial spinup for AquaControl
• Daily output: same fields as for monthly-mean output, saved for the last 10 years of
each simulation
3
• 3-hourly output: same fields as for monthly-mean output, saved for the last 3 years of
each simulation
As for whether fields should be saved as means or snapshots, we adapt the CMIP5 conventions. For the monthly and daily output streams, all fields should be means over the output
period. For the 3-hourly output stream, T (including surface and atmosphere), u, v, ω, q, z
(geopotential height), should be snapshots whereas all other fields should be means.
4 Specification of the ocean q-flux
The slab ocean allows interactive sea-surface temperatures without the large computational
burden from coupling to a dynamic three-dimensional ocean model. The q-flux is derived
from present-day observations, with details given below. It is provided as a netcdf file that
includes the q-flux for the aquaplanet simulations and a corrected q-flux for the simulations
with land. The q-flux can also be implemented in the models by using the polynomial
representation that is given in section 4.1. Fig. 1 shows the q-flux and the corresponding
ocean heat transport, and compares the two with the present-day values. The q-flux is
constant in time. For the aquaplanet simulations, the q-flux is zonally-symmetric. For the
land simulations, the q-flux is zonally-symmetric over ocean regions and zero over land.
4.1 Derivation of the q-flux
For the q-flux we make use of present-day observations of the annual-mean TOA radiation
budget and reanalysis estimates of the atmospheric energy transport (AT M ). From the
observed time-mean estimate of the total ocean heat transport,
qobs (λ, ϕ) = T OA(λ, ϕ) − AT M (λ, ϕ),
(3)
we calculate the zonal-mean time-mean q-flux
1 Z 2π
q dλ.
q obs (ϕ) =
2π 0 obs
(4)
For the zonal average, land points are included with their q-flux being set to zero. I.e., the
zonal average is not weighted by the sea-land mask. Then, we fit q obs by a hemisphericallydependent polynomial of degree four to smooth out smaller-scale wiggles in the mid- and
high-latitudes. This is done to avoid that these wiggles aggravate model differences in the jet
position, and to avoid an imprint on these observed wiggles on the extratropical atmospheric
circulation in the idealized simulations pursued here. Specifically, the q-flux is given by
q(ϕ) = p0 + p1 · ϕ + p2 · ϕ2 + p3 · ϕ3 + p4 · ϕ4 .
4
(5)
qflux at ocean grid cells (Wm-2)
30
20
10
0
10
20
30
40
50
60
observed q-flux
idealized q-flux on aquaplanet
idealized q-flux with continent
60S
30S
Eq
30N
60N
60S
30S
Eq
latitude
30N
60N
ocean heat transport (PW)
2.0
1.5
1.0
0.5
0.0
0.5
1.0
1.5
Figure 1 – Q-flux and ocean heat transport from observations (blue) and the fit used in the
MIP (green). The q-flux for simulations with land is given in red; note that in the case of land
the q-flux shown is not the zonal-mean q-flux but the q-flux over ocean points only.
5
For the aquaplanet simulations, the polynomial coefficients are in the Northern hemisphere
(ϕ > 0)
H
= −50.1685
pN
0
H
pN
= 4.9755
1
H
pN
= −1.4162 · 10−1
2
H
= 1.6743 · 10−3
pN
3
H
pN
= −6.8650 · 10−6 ,
4
where ϕ has units of deg latitude. In the Southern hemisphere (ϕ < 0),
= −56.0193
pSH
0
pSH
= −6.4824
1
pSH
= −2.3494 · 10−1
2
pSH
= −3.4685 · 10−3
3
pSH
= −1.7732 · 10−5
4
As described above, for the land simulation the q-flux is set to zero over land. This requires
a small uniform correction of -0.59 Wm−2 to guarantee that the global-mean q-flux is zero.
The uniform correction is applied to all ocean points (but not to the land points) by setting
H
H
pN
→ pN
− 0.59 and pSH
→ pSH
− 0.59.
0
0
0
0
6
5 ECHAM6.1 reference plots
This section provides plots for the ECHAM6.1 AquaControl and LandControl simulations
as a reference for other modelling groups. This does not mean that other models must show
similar or the exact same results, but the plots might provide guidance to check whether
the setup was implemented correctly. We particularly encourage modelling groups to check
the top-of-atmosphere solar insolation, the land surface albedo diagnosed from the surface
shortwave fluxes, and the q-flux over ocean, which can be diagnosed as the sum of shortwave
and longwave radiative fluxes and the latent and sensible heat fluxes.
5.1 TOA shortwave downward insolation
TOA incoming solar (Wm-2)
12
550.0
10
400
TOA incoming solar (Wm-2)
412.5
8
month
450
481.2
343.8
275.0
6
206.2
4
137.5
68.8
2
50
0
latitude
0.0
50
350
300
250
200
150
100
50
0
latitude
50
100
Figure 2 – Top-of-atmosphere downward shortwave irradiance.
5.2 AquaControl
ECHAM6.1 AquaControl (itczmip001)
20
2.0
diagnosed from sfc balance
protocol
ocean heat transport (PW)
10
qflux
0
10
20
30
40
4
1.0
0.5
0.0
0.5
1.0
50
0
latitude
50
100
2.0
100
2
0
2
4
1.5
50
60
100
6
1.5
atm. heat transport (PW)
30
50
0
latitude
50
100
6
100
50
0
latitude
50
Figure 3 – ECHAM6.1 AquaControl: ocean q-flux (right) and meridional energy transport
of the ocean (middle) and the atmosphere (right).
7
100
ECHAM6.1 AquaControl (itczmip001)
12
Tropical Precipitation (mm/day)
26.00
22.75
10
13.00
6
9.75
4
6.50
2
3.25
12
15
10
5
0
5
latitude
10
Surface temperature (K)
10
15
20
0.00
308.00
305
301.71
300
Surface temperature (K)
month
16.25
20
295.43
8
month
Precipitation (mm/day)
19.50
8
289.14
6
282.86
4
276.57
2
50
0
latitude
50
10
9
8
7
6
5
4
3
2
1
100
global mean: 4.05 mm/day
50
0
latitude
50
100
0
latitude
50
100
global mean: 295.4 K
295
290
285
270.29
280
264.00
275
100
50
Figure 4 – ECHAM6.1 AquaControl: seasonal cycle of tropical precipitation (top left) and
surface temperature (bottom left). Right: time-mean zonal-mean precipitation and surface
temperature; global-mean values are given on top of the plots.
8
5.3 LandControl
ECHAM6.1 LandControl (itczmip003)
diagnosed from sfc balance
protocol
10
20
30
40
50
60
100
6
1.5
0
ocean heat transport (PW)
qflux over ocean points (W/m2)
10
2.0
4
1.0
atm. heat transport (PW)
20
0.5
0.0
0.5
1.0
0
latitude
50
100
2.0
100
0
2
4
1.5
50
2
50
0
latitude
50
100
6
100
50
0
latitude
50
Figure 5 – ECHAM6.1 LandControl: ocean q-flux over ocean points (right), and meridional
energy transport of the ocean (middle) and the atmosphere (right).
9
100
ECHAM6.1 LandControl (itczmip003)
Surface albedo (%)
14
12
50
latitude
10
0
8
50
6
150
100
50
0
longitude
50
100
150
4
Figure 6 – ECHAM6.1 LandControl: time-mean surface albedo calculated by the surface
downward and upward shortwave radiative flux.
10
ECHAM6.1 LandControl land+ocean (itczmip003)
Tropical Precipitation (mm/day)
10
22.75
7
19.50
6
16.25
13.00
6
9.75
4
3
6.50
2
3.25
2
0.00
1
100
12
15
10
5
0
5
latitude
10
Surface temperature (K)
10
15
20
310.00
305
303.43
300
296.86
8
290.29
6
283.71
4
277.14
2
50
0
latitude
50
global mean: 3.94 mm/day
5
4
20
month
8
Surface temperature (K)
month
8
26.00
Precipitation (mm/day)
12
50
0
latitude
50
100
0
latitude
50
100
global mean: 294.8 K
295
290
285
270.57
280
264.00
275
100
50
Figure 7 – ECHAM6.1 LandControl: seasonal cycle of tropical precipitation (left) and surface
temperature (bottom left). Right: time-mean zonal-mean precipitation and surface temperature; global-mean values are given on top of the plots.
11
ECHAM6.1 LandControl ocean only (itczmip003)
Tropical ocean precipitation (mm/day)
10
22.75
8
16.25
13.00
6
9.75
4
6.50
2
3.25
20
12
15
10
5
0
5
latitude
10
Ocean surface temperature (K)
10
15
20
0.00
4
277.14
270.57
50
3
2
300
283.71
0
latitude
4
303.43
6
50
5
305
290.29
2
6
310.00
296.86
8
7
1
100
Ocean surface temperature (K)
month
9
19.50
8
month
26.00
Ocean precipitation (mm/day)
12
264.00
50
0
latitude
50
100
50
0
latitude
50
100
295
290
285
280
275
100
Figure 8 – ECHAM6.1 LandControl: seasonal cycle of tropical precipitation (left) and surface
temperature (bottom left) only over ocean points. Right: time-mean zonal-mean precipitation
and surface temperature only taking into account ocean points.
12
ECHAM6.1 LandControl land only (itczmip003)
12
Tropical land precipitation (mm/day)
26.00
19.50
8
16.25
13.00
6
9.75
4
6.50
2
3.25
20
12
15
10
5
0
5
latitude
10
Land surface temperature (K)
15
20
0.00
310.00
303.43
10
296.86
8
290.29
6
283.71
4
277.14
270.57
2
50
0
latitude
50
6
5
4
3
2
1
Land surface temperature (K)
month
Land precipitation (mm/day)
22.75
10
month
7
264.00
306
305
304
303
302
301
300
299
298
297
30
20
10
0
latitude
10
20
30
30
20
10
0
latitude
10
20
30
Figure 9 – ECHAM6.1 LandControl: seasonal cycle of tropical precipitation (left) and surface
temperature (bottom left) only over land points. Right: time-mean zonal-mean precipitation
and surface temperature only taking into account land points.
13
ECHAM6.1: LandControl (itczmip003)
90
Surface temperature (K)
60
30
0
-30
-60
-90
90
-120
0
120
Latent heat flux (W/m2)
60
30
0
-30
-60
-90
-120
0
120
308.00
303.88
299.75
295.62
291.50
287.38
283.25
279.12
275.00
180
160
140
120
100
80
60
40
20
0
90
Precipitation (mm/day)
60
30
0
-30
-60
-90
90
-120
0
120
Sensible heat flux (W/m2)
60
30
0
-30
-60
-90
-120
0
120
Figure 10 – ECHAM6.1 LandControl: time-mean surface temperature, precipitation and
surface latent and sensible heat fluxes.
14
10.00
8.89
7.78
6.67
5.56
4.44
3.33
2.22
1.11
0.00
40.00
35.56
31.11
26.67
22.22
17.78
13.33
8.89
4.44
0.00
References
Joussaume, S. and P. Braconnot, 1997: Sensitivity of paleoclimate simulation results to
season definitions. Journal of Geophysical Research: Atmospheres, 102 (D2), 1943–1956,
doi:10.1029/96JD01989.
Kang, S. M., I. M. Held, D. M. W. Frierson, and M. Zhao, 2008: The Response of the ITCZ
to Extratropical Thermal Forcing: Idealized Slab-Ocean Experiments with a GCM. J.
Climate, 21, 3521–3532, doi:10.1175/2007JCLI2146.1.
Voigt, A., B. Stevens, J. Bader, and T. Mauritsen, 2014: Compensation of hemispheric
albedo asymmetries by shifts of the ITCZ and tropical clouds. J. Climate, 27, 1029–1045,
doi:10.1175/JCLI-D-13-00205.1.
15