A revised version of
Contents:
One of the main goals of the MIT Joint Program on the Science and Policy of Global Change is to evaluate the role of these uncertainties and their interactions in climate change predictions (Jacoby and Prinn, 1994). Only very limited studies of this kind can be performed with AOGCMs due to their large requirements for computational resources. Moreover, the only uncertainty that can be addressed in simulations with a given AOGCM is uncertainty in the forcing (Cubasch, et al., 1992; IPCC, 1996). In most cases different upwelling diffusion-energy balance (UD/EB) models have been used for studying uncertainty in climate change (IPCC, 1990, 1992 and 1996; Murphy, 1995; Wigley and Raper, 1993; Jonas, et al., 1996).
A modified version of the 2D (zonally averaged) statistical-dynamical atmospheric model (Yao and Stone, 1987; Stone and Yao, 1987 and 1990) developed on the basis of the GISS GCM (Hansen, et al., 1983) has been chosen for use in climate simulations in the integrated framework of the MIT Joint Program. Unlike energy balance models, the 2D model includes parameterizations of all the main physical processes and is, therefore, capable of reproducing many of the nonlinear interactions taking place in GCMs. At the same time it is about 20 times faster than the GISS GCM with the same latitudinal and vertical resolutions. At the present time the 2D model is coupled with a simple diffusive ocean model. In climate change simulations discussed below, the 2D climate model is driven by changing greenhouse gas concentrations. Studies of climate change caused by different emission scenarios (Prinn, et al., 1996) have also been carried out with a version of the MIT 2D climate model which includes fully interactive atmospheric chemistry and transport of chemical species (Wang, et al., 1995), and calculates oceanic carbon uptake.
Brief descriptions of both the atmospheric and oceanic models are given in Section 2, with some results of a present-day climate simulation presented in Section 3. For studying uncertainties in climate change, model versions with different sensitivities were obtained by changing the cloud feedback in a way described by Hansen, et al. (1993). In Section 4 responses of different versions of the MIT 2D model to the doubling of the CO2 concentration are compared with the results obtained in simulations with different AGCMs. This comparison shows that the 2D model's behavior is similar to that of AGCMs. It is also shown that the 2D model can match the transient responses of different AOGCMs to a gradual increase in atmospheric CO2 by using different diffusion coefficients. Uncertainties in projected climate change, namely, in surface warming and sea level rise, associated with uncertainties in the climate sensitivity and rate of heat uptake by the deep ocean are studied in the Section 5. Finally, some conclusions are given in Section 6.
A number of modifications have been made to the model at MIT to make it more suitable for climate change studies (Sokolov and Stone, 1995; Prinn, et al., 1996; Xiao, et al., 1997). The first one was to include in the 2D model a real land-ocean distribution. The modified MIT 2D model, as well as the GISS GCM, allows up to four different kinds of surface in the same grid cell, namely, open ocean, ocean-ice, land, and land-ice. The surface characteristics (e.g., temperature, soil moisture, albedo) as well as turbulent and radiative fluxes are calculated separately for each kind of surface while the atmosphere above is assumed to be well mixed horizontally. The area weighted averages of fluxes from different kinds of surfaces are used to calculate the change of temperature, humidity, and wind speed in the atmosphere.
In the GISS GCM turbulent fluxes of temperature, moisture and momentum are calculated under the assumption that the atmospheric surface layer is in equilibrium, which leads to a complicated algorithm including nested iterations. That algorithm, while used successfully in both the GISS GCM and 2D model without land, produces computational problems when land is included. Because of this, the equilibrium assumption has been replaced by the assumption that the layer between the surface and the model's first level is well mixed. The surface wind speed calculation also has been simplified compared to the procedure used in the GISS GCM. The absolute value of the vector Vs is assumed to be equal to:
[Eq1]
where V1 is wind at the model's first level and overbar[V'2s ] is the zonal mean surface wind variance. Calculation of the surface wind variance is similar to that of temperature variance described in Yao and Stone (1987). The cross isobar angle, alpha, is calculated as a function of absolute value of [Vs] and a bulk Richardson number Ris, according to formulas derived by G. Russell (personal comm., 1993):
[Eq2] alpha = [ ] , if
[Eq3] alpha = [ ] , if
Here alphaois an empirical coefficient that relates cross isobar angle
to the surface wind which is taken equal to 0.3.
Two different types of clouds are taken into account in the model: convective
clouds, associated with moist convection, and large-scale or supersaturation
clouds, formed due to large-scale condensation. The amount of convective clouds
in a given layer is proportional to the mass flux due to moist convection
through the lower boundary of this layer. The amount of supersaturated clouds
is expressed as a height-dependent function of the critical value of relative
humidity for cloud formation and the critical value of relative humidity for
condensation (hcon). Due to course model resolution, a value of
hcon = 90% has been chosen as the criterion for condensation.
Convection occurs in just the unstable fraction of a latitude belt, based on a
parameterization of subgridscale variations of temperature and humidity (Yao
and Stone, 1987).
Some changes have been made to the GISS 2D model's radiation calculations. A
dependence of snow albedo on surface temperature has been included; snow
density and heat conductivity are now calculated as functions of snow mass
instead of being fixed, and the temperature distribution in the ground layer is
assumed to be linear rather than quadratic as in the GISS GCM.
In the GISS 2D model a 16th-order filter is applied to air
temperature and specific humidity. We found that this has a significant impact
on the present-day climate simulations and the model's climate sensitivity. For
example, the zonal wind simulated by the 2D model without filter is much closer
to the observed, especially for Northern Hemisphere winter. On the other hand,
specific humidity in the equatorial region turned out to be unrealistically
low. The latter was improved by incorporating into the model a second order
horizontal diffusion for specific humidity. This diffusion is, however, small
compared to the model's parameterised diffusion due to large-scale eddies
outside the tropics. Results of simulations with both the MIT 2D model and
chemistry-transport model (Wang, et al., 1995), as well as with the
original GISS 2D model (Stone and Yao, 1990), have shown that the original 2D
model's parameterisation of eddy transports tends to underestimate transports
of heat and tracers. Consequently the value of the scaling constant A in
Equation 1 in Stone and Yao (1990) was increased from 0.6 to 0.9.
In simulations of the present-day climate and equilibrium climate change the
2D model is coupled with a mixed-layer ocean model. In order to simulate the
current climate, the equation for the mixed-layer temperature includes a term
representing the effect of horizontal heat transport in the ocean and heat
exchange between the mixed layer and deep ocean. The heat balance equation for
the mixed layer is expressed in terms of the specific heat capacity and density
of salt water, sea-ice mass, mixed layer depth, latent heat of freezing, heat
balance on the ocean surface, heat flux through the lower surface of sea-ice,
and fractions of grid cell covered by open ocean and sea-ice. The horizontal
heat flux can be calculated from this equation using the results of a climate
simulation with prescribed climatological sea surface temperature and sea-ice
distributions. This ocean treatment is essentially similar to the one used by
Meleshko, et al. (1991), except that in the MIT model the mixed layer
depth is a prescribed function of season and latitude (Hansen et al., 1988).
The algorithm used for calculation of the thermal energy of the mixed layer
with variable depth is described in Russell, et al. (1985). The above
mentioned change in the eddy flux parameterization led to a better agreement of
the implied ocean heat transport with observations.
In simulations of transient climate change the heat uptake by the deep ocean
has been parameterized by diffusive mixing of the perturbations of the
temperature of the mixed layer into deeper layers (Hansen, et al.,
1988). The zonally averaged values of diffusion coefficients calculated from
measurements of tritium mixing have been chosen as "standard" ones (Table
2.1).
The global average value of the diffusion coefficients, denoted as
Kv, equals 2.5 cm2/s for these "standard" values.
However, Hansen, et al. (1984) found that the equivalent value of
diffusion coefficient that gives similar results when used in a 1D model is
only 1 cm2/s. As will be shown in Section 4, two times the
"standard" diffusion is required to match the behavior of the UD/EB model used
in IPCC, 1996, which uses a diffusion coefficient equal to 1 cm2/s,
but also includes upwelling with a rate which decreases as global temperature
increases.
Neither the 2D model nor the GISS GCM reproduces the seasonal cloud change
(see Sokolov and Stone, 1995) in the tropics associated with the shift of the
Intertropical Convergence Zone. Nevertheless, the overall pattern of seasonal
changes in clouds (see Sokolov and Stone, 1995) and cloud radiative forcing
(Figure 3.5) simulated by the 2D model, is quite similar to the observed.
As mentioned above, the eddy flux parameterization has been slightly changed,
so as to increase eddy transports. As a result, total atmospheric energy
transport (Figure 3.6) is in better agreement with observation (e.g.,
Trenberth and Solomon, 1994) than it was in a previous version of the MIT 2D
model (Sokolov and Stone, 1995). The same is true for the implied ocean heat
transport, which is, however, still higher than observed (e.g.,
Trenberth and Solomon, 1994), especially, in the Northern Hemisphere tropics,
because of the model's underestimation of the Hadley cell heat transport in the
Northern Hemisphere.
There is one essential problem with the simulation of sea surface temperature
(SST) and sea ice distribution in a 2D model. Because of longitudinal
variations of sea surface temperature and sea ice, zonal mean SST may be above
0 degrees C even when part of the ocean surface at the same latitude is covered
by ice. However, in the formulation of the mixed layer ocean model, SST is kept
at or below 0 degrees C until all ice melts, and no sea ice forms if SST is
above the freezing point for salt water, that is -1.56 degrees C. As a result,
the above mentioned feature of the SST and sea ice distribution cannot be
simulated by a 2D model. Because of this, data used in the simulations with
prescribed SST and sea ice have been adjusted. Namely, if, in any given
latitude belt, less than 10% of the ocean surface is covered by ice and zonal
mean SST is above 0 degrees C ice is removed. If ice covers more than 10% of
the ocean, then SST is set to 0 degrees C if the mass of ice is decreasing, and
to -1.56 degrees C otherwise. The adjusted data are shown in Figures 3.7 and
3.8 instead of direct observations. As one can see, despite the use of a
"Q-flux" in the mixed layer model there are some differences in the prescribed
and simulated values of SST and sea ice cover and depth. Some reasons for these
differences are described in Sokolov and Stone (1995). Similar problems with
predicting sea ice distribution in simulations with the GISS GCM with 8 degree
x 10 degree resolution are discussed by Hansen, et al. (1984).
Variations of the globally averaged annual mean surface air temperature in two
simulations of the present-day climate with the 2D model are shown in Figure
3.9. The standard deviations of surface temperature in the simulation with the
2D model coupled with just the mixed layer ocean model is 0.10 degrees C;
however, it decreases to 0.06 degrees C if mixing of mixed layer temperature
perturbations into the deep ocean is taken into account. Thus, the unforced
interannual variability produced by the MIT 2D model is rather close to that of
the GISS AGCM but somewhat less than in simulations with coupled AOGCMs. The
corresponding number for the GISS AGCM coupled with mixed layer model and
diffusive deep ocean is 0.05 degrees C (Hansen, et al., 1997); and for
the coupled GFDL and MPI AOGCMs (Santer, et al., 1995) 0.10 degrees C
and 0.12 degrees C, respectivly.
As a whole, a comparison of the model's results with the observational data
shows that it reproduces reasonably well the major features of the present-day
climate state. Of course, there are some essential 3D features of the
atmospheric circulation that cannot be simulated by a 2D model. However the
depiction of the zonally averaged circulation by the 2D model is not very
different from that by 3D GCMs. Since the model is to be used for climate
change prediction, it is noteworthy that the seasonal climate variations are
also reproduced quite well. Use of a 2D model allows as to perform a
significantly larger number of climate simulations than would be possible with
an AOGCM. A 100 year simulations takes about 12 hours on DEC Alpha Station 250
with the model described above and three times more with couple
climate-chemistry model.
As mentioned in the introduction, the model's versions with different
sensitivities were obtained by inserting an additional cloud feedback, in the
way proposed by Hansen, et al. (1993). Namely, calculated cloud amount
is multiplied by the factor (1 + k DeltaTs), where DeltaTs is the
increase of the globally averaged surface air temperature with respect to its
value in the present-day climate simulation. Equilibrium responses of the MIT
2D model to the doubling of the atmospheric CO2 for several values
of k are shown in Table 4.1. The natural sensitivity, DeltaTeq, of the
MIT 2D model, that is without an additional cloud feedback (k = 0), is 3.0
degrees C, about 1 degree C less than for the version described in Sokolov and
Stone (1995). This decrease in the model's sensitivity is mainly due to changes
in the simulated water vapor distribution and somewhat weaker sea ice-albedo
feedback especially in the Northern Hemisphere. The natural cloud feedback of
the 2D model is very weak, i.e., in a simulation with fixed clouds
DeltaTeq equals 2.9 degrees C. Thus the ratio of DeltaTeq in the
simulation with calculated cloud to that in the simulation with fixed cloud is
1.03 compared to 1.1 for the previous version of the 2D model (Sokolov and
Stone, 1995), 1.25 for the GFDL (Wetherald and Manabe, 1988) and 1.75 for GISS
(Hansen, et al., 1984) GCMs. The magnitude of cloud feedback for the
GISS GCM was evaluated by means of calculations with a 1D model using results
of simulations with the GISS GCM (Hansen, et al., 1984).
The equilibrium surface air temperature increase due to a doubling of the
CO2 concentration predicted by different GCMs ranges from 1.9 to 5.4
degrees C. A significant part of this difference is related to the differences
in cloud feedbacks (Cess, et al., 1990; Senior and Mitchell, 1993;
Washington and Meehl, 1993) produced by the different GCMs. That, in turn, is
caused mainly by different treatments of cloud optical properties. The feedback
associated with changes in the optical properties of clouds in the GCM
experiments is, of course, rather different from that associated with the
changes in cloud amount, used in our simulations. However, it is interesting to
note, that the relationship between changes in the globally averaged annual
mean net cloud forcing (DeltaNCF) and surface air temperature (DeltaTeq) in
the simulations with different versions of the 2D model is qualitatively
similar to that in the simulations with different versions of the UKMO GCM
(Senior and Mitchell, 1993). In the simulations described by Senior and
Mitchell DeltaNCF equals -1.04, 0.21, 0.73, and 2.05 (W/m2) for
DeltaTeq equal to 1.9, 2.8, 3.3 and 5.4 degrees C, respectively.
Different versions of the MIT 2D model reproduce well the relationship between
surface warming and increase in precipitation obtained in simulations with
AGCMs (Figure 4.1). In Table 4.2 changes in the components of the globally
averaged annual mean surface heat budget obtained in the simulation with the
version of the MIT 2D model with DeltaTeq= 4 degrees C are compared with
the results of AGCMs with similar sensitivities. The 2D model produces a
relatively large increase in solar radiation absorbed by the surface which is
mainly compensated by a larger decrease in latent heat flux than in the AGCMs.
Changes in longwave radiation and sensible heat flux lie in the range produced
by the AGCMs. A comparison of the results of the 2D model with those from the
GISS AGCM shows that the larger increase in absorbed solar radiation is caused
mainly by a substantial decrease in the cloud amount (see Table 4.1) and sea
ice cover
Here S and F are the short and longwave radiation at the surface respectively;
LE and H are the turbulent fluxes of latent and sensible heat. Results for
AGCMs are from Washington and Meehl (1993).
At the same time the dependence of changes in the latent heat flux and
radiation components of the surface energy budget on surface warming in
simulations with different versions of the 2D model is very similar to that in
simulations with different versions of the UKMO AGCM (Figures 4.2 and 4.3). In
particular, although the 2D model's simulated change in solar radiation
absorbed at the surface is quite different from that simulated by the NCAR,
GFDL and GISS AGCMs (Table 4.2), it is consistent with the changes simulated by
the UKMO AGCM (Figure 4.3). In contrast, the changes in sensible heat flux
simulated by the 2D model are quite different from those simulated by the UKMO
AGCM (Figure 4.2), but, as can be seen form Table 4.2, there is significant
discrepancy in the changes of sensible heat flux even among GCMs with close
sensitivities.
As discussed by Boer (1993), the CCC AGCM produces a decrease in the net solar
radiation at the surface due to a negative feedback associated with an increase
in cloud albedo. The versions of the MIT 2D model with sensitivities less than
2.5 degrees C and the version of the UKMO AGCM with calculated radiative
properties of cloud show a similar change in solar radiation at the surface.
The increase in evaporation/precipitation produced in the simulations with both
the 2D model and the UKMO AGCM are, however, relatively larger than in the
simulation with the CCC AGCM.
The latitudinal distributions of the CO2 induced changes in surface
air temperature, precipitation and cloud cover obtained in simulations with
three versions of the MIT 2D model are shown in Figures 4.4 and 4.5 for DJF and
JJA respectively. One of these versions is that without additional cloud
feedback (k = 0). The other two are the versions that have sensitivities close
to the high and low ends of the sensitivity range proposed by the IPCC (1990),
namely 4.6 and 1.6 degrees C. Results of the doubled CO2 simulation
with the GISS GCM (Hansen, et al., 1984) are also shown for comparison.
All versions of the 2D model show an amplification of the surface warming in
high latitudes, especially during winter. The GISS GCM produces an increase in
surface temperature close to that of the 2D model with sensitivity 4.6 degrees
C in the equatorial region and middle latitudes of the summer hemisphere and
close to that of the 2D model with sensitivity 3.0 degrees C in high latitudes
of the winter hemisphere. The sensitivity of the GISS GCM is 4.2 degrees C. At
the same time, the latitudinal structure of surface air temperature change
produced by the 2D model closely resembles the results of the GFDL and UKMO
GCMs (Wetherald and Manabe, 1988; Senior and Mitchell, 1993; and Wilson and
Mitchell, 1987).
Differences in the latitudinal gradient of surface warming between the 2D
model and the GISS GCM can be explained by comparing changes in precipitable
water and sea ice cover (Figures 4.6 and 4.7). The increase in precipitable
water obtained in the simulation with the GISS GCM is larger in low latitudes
than that produced by any version of the 2D model, and this causes a larger
warming in the equatorial region. At the same time, the 2D model shows a
significantly larger decrease in the sea ice cover in the winter hemisphere
and, as a result, surface warming amplification, which is additionally
increased by the decrease in cloud cover (Figures 4.4c, 4.5c and 4.10) in this
region. The differences in the sea ice changes between the 2D model and the
GISS GCM are, at least in part, due to the fact that the thickness of sea ice
produced by the 2D model in the present-day climate simulation is less than
that simulated by the GISS GCM (see Figures 3.7 and 3.8). As was shown by Rind,
et al. (1995), both the global average and latitudinal profile of
surface warming obtained in simulations with the GISS GCM depend significantly
on the sea ice thickness used in the control simulation (which is not well
constrained by observations); and on how the sea ice mass decrease caused by
mixed layer warming is distributed between sea ice thinning and horizontal
contraction. In the MIT 2D model, when mixed layer temperature reaches 0
degrees C additional heat is spent to reduce sea ice depth, until it reaches a
minimal value, and is used to reduce horizontal extent after that, while in the
simulation with the GISS GCM shown here it is spent equally on vertical and
horizontal reduction of sea ice. There is, however, no difference between these
two approaches after the sea ice depth reaches its minimal value.
The imposition of a minimal sea ice thickness leads to a discontinuity in the
rate of sea ice cover decrease in response to global warming in a given grid
cell. This, together with the low horizontal resolution of the 2D model,
results in a discontinuity of the model sensitivity with respect to changes in
the parameter k (see above given formula for cloud calculation). For 0.0875 <= k <=
0.0879 DeltaTeq = 1.6 degrees C, but DeltaTeq drops to 1.35 degrees
C for k = 0.088. Such a sharp change is caused by the fact that in the last
case about 30% of the ocean surface at 60 degree S is covered by sea ice with
minimal thickness, while sea ice is completely melted in the others. There is
no such effect associated with sea ice changes in the high latitudes of the
Northern Hemisphere, apparently, due to the smaller ocean area.
Height-latitude cross sections of the change in zonal mean temperature
(Figures 4.8 and 4.9) show all the main features found in most GCM simulations
with doubled CO2, such as stratospheric cooling, maximum warming in
the upper troposphere in the tropics and the above mentioned strong surface
warming in high latitudes of the winter hemisphere. The magnitude of the
decrease in stratospheric temperature is practically the same in all
simulations regardless of sensitivity, similar to the results of the
simulations with different versions of the UKMO AGCM (Senior and Mitchell,
1993). At the same time, the upper-tropospheric warming, in contrast to the
results of the UKMO simulations (Senior and Mitchell, 1993 and Wilson and
Mitchell, 1987), penetrates into the southern hemisphere in both DJF and JJA,
more strongly during the northern hemisphere winter. Height-latitude cross
sections of changes in annual mean zonally averaged cloud (Figure 4.10) bear an
overall resemblance to the results produced by different GCMs (Hansen, et
al., 1984; Wetherald and Manabe, 1988; Senior and Mitchell, 1993). However,
the increase of high cloud at low latitudes associated with the upward shift of
the tropopause is almost absent in the simulation with the standard version of
the 2D model (k = 0).
In general, the results presented above show that responses of different
versions of the MIT 2D model to the doubling of the CO2
concentration, in terms of both global average and zonal mean, are similar to
those obtained in simulations with different GCMs. Since the climate model
outputs are used in simulations with the Terrestrial Ecosystem Model (Prinn,
et al., 1996; Xiao, et al., 1997), it is important to note that
inserting the additional cloud feedback described above, while allowing us to
change model sensitivity, does not lead to any physically unrealistic changes
in climate. On the contrary, the changes in other climate variables, such as
precipitation, evaporation and so on, are generally consistent with the results
produced by different GCMs. It is worth noting that the ecosystem impacts of
the climate change due to a CO2 doubling, as simulated by the
standard version of the MIT 2D model, are quite similar to those produced by
the climate changes simulated by different AGCMs (Xiao, et al., 1997).
As shown in Figure 4.11, the transient responses of the 2D model with the
"standard" deep ocean diffusion coefficients doubled is very similar to those
obtained in the simulations with the GFDL AOGCM with different rates of
CO2 increase (IPCC, 1996)[1]. Ten
times "standard" values of the diffusion coefficients are required to match the
delay in warming produced by the MPI AOGCM (Cubasch, et al., 1992) (see
Figure 4.12).[2] Data for the MPI model have
been corrected by taking into account error in the rate of surface warming
associated with a "cold start" (Hasselmann, et al., 1993; IPCC, 1996).
At the same time, no heat diffusion into the deep ocean is required to
reproduce the fast warming produced by the NCAR AOGCM (IPCC, 1996). The UD/EB
model used in IPCC (1996) was tuned to reproduce the globally averaged results
of the GFDL AOGCM. This implies, as noted in Section 2, that it has a rate of
heat uptake close to that for the 2D model with doubled diffusion
coefficients.
The only significant difference between results of the 2D model and the GFDL
AOGCM occurs in the simulation with 0.25% per year increase in CO2,
and only after some 120-150 years of integration. Aside from that, the 2D
climate model reproduces quite well the globally averaged surface warming
predicted by different AOGCMs for a variety of forcing scenarios, for periods
of at least 100 years.
At the same time, there is one noticeable difference in the zonal pattern of
temperature increase obtained in the transient simulations with the MIT 2D
model and that produced by most of the AOGCMs. That is, there is no strong
interhemispheric asymmetry in the transient warming simulated by the 2D model.
The change of zonally averaged surface air and sea surface temperature for the
decade of doubling of CO2 concentration obtained in the simulation
with 1% per year increase in CO2 with the version of the 2D model
matching the GFDL AOGCM is shown in Figure 4.13, together with the equilibrium
change obtained in the simulation with the same version couple with a mixed
layer ocean model. The temporal evolution of the surface air temperature and
sea surface temperature in the same transient simulation are shown in Figures
4.14 and 4.15, respectively. As can be seen from Figures 4.14 and 4.15, there
is some delay in the surface warming in high latitudes of the southern
hemisphere, much smaller, however, than in the similar simulation with the GFDL
AOGCM (Manabe, et al., 1991). The time dependent response of the sea
surface temperature is closer to that in the simulation with the GFDL GCM,
although the 2D model does not show cooling near Antarctica. The differences in
the temporal changes of sea ice thickness (Figure 4.16) from that shown by
Manabe, et al. (1991), are consistent with the differences in the
surface warming. In particularly, in the simulation with the 2D model sea ice
depth steadily decreases in both hemispheres as CO2 increases,
whereas it increases in the Southern Hemisphere in the GFDL simulation.
The deep ocean temperature change for the decade of CO2 doubling in
the simulation with the 2D model (Figure 4.17), bears a general resemblance
with the results of the simulations with the GFDL AOGCM (Manabe, et al.,
1991) and with the UKMO AOGCM (Murphy and Mitchell, 1995). The 2D model
produces zones of heat penetration into the deep ocean at about 60 degrees N
and about 50 degrees S. Consistent with the significant surface warming in high
latitudes of the Southern Hemisphere, the zone of very deep penetration of heat
south of 60 degrees S is missing in the simulation with the MIT 2D model.
However, some recent studies show that current ocean GCMs may produce excessive
vertical mixing in this region and that, as a result, the corresponding
retardation of warming may be exaggerated (IPCC, 1996).
Another characteristic describing changes in the deep ocean temperature is sea
level rise due to thermal expansion. The thermal expansion has been calculated
from the deep ocean temperature increase using the method described in Gregory
(1993). Levitus' (1992) data have been used for the unperturbed state of the
deep ocean. In spite of our model's simplified representation of the deep
ocean, it reproduces the thermal expansion of the deep ocean as simulated by
the GFDL AOGCM quite well (Figure 4.18), except for the already noted
differences in the last stage of the simulation with 0.25% per year increase in
CO2. There is less agreement between results of the simulations with
the MIT 2D model and the MPI AOGCM (Figure 4.19), especially for scenario D
case. A similar problem in trying to reproduce the results of the MPI AOGCM
with an UD/EB model was reported by Raper and Cubasch (see IPCC, 1996, Section
6.3.1).
Estimates of the impact of uncertainty in any given parameter depend strongly
on both the range for this particular parameter and "best guesses" for the
others. The range for the rate of heat uptake by the deep ocean used in the
simulations below has been based in part on values needed to match the
transient warming produced by different AOGCMs. The "standard" values of the
diffusion coefficients, with Kv = 2.5 cm2/s, obtained, as
mentioned above, from observations of tritium mixing into the deep ocean have
been chosen as a "best guess." The heat uptake by the deep ocean in 100 years
produced by the version of the MIT 2D model using these coefficients is about
15% less than that when using the coefficients matching the GFDL AOGCM. The
coefficients with Kv = 12.5 cm2/s, that is half as large
as those matching the MPI AOGCM, have been used as a high end of the range.
Since the zero rate of heat penetration into the deep ocean required to match
the behavior of the NCAR AOGCM, does not seem to be realistic, the coefficients
with Kv = 0.5 cm2/s have been used as a lower limit. This
range, while somewhat narrower than that given by AOGCMs, is still wider than
that used by Wigley and Raper (1993), who arbitrarily chose coefficients twice
as large and half as large as their standard value for upper and lower bounds
or the uncertainty range. Since their UD/EB model has rate of heat uptake close
to that of the MIT 2D model with Kv = 5.0 cm2/s, the
corresponding range for the MIT 2D model would be from 2.5 to 10
cm2/s. For climate sensitivity a range close to that suggested by
the IPCC, namely 1.6 degrees C to 4.5 degrees C, has been used in this study.
The simulations discussed below have been performed with a 1% per year increase
in the CO2 concentration, while all other forcings were held
constant.
As can be seen from Figure 5.1, if the rate of heat uptake by the deep ocean
is close to that matching the behavior of the NCAR model, the increase of the
surface temperature will be significantly higher than the highest estimate of
possible warming given by the IPCC. Actually, in the simulation with DeltaTeq
= 4.5 and Kv = 0 the surface temperature increase for years
91-100 of the integration, DeltaT91-100, is 5.7 degrees C, compared to
4.6 degrees C for heat diffusion with Kv = 0.5 cm2/s. One
might argue that the combination of high sensitivity with low rate of heat
uptake has quite a low probability. However, there is a noticeable difference
in the simulations with Kv = 2.5 cm2/s and Kv
= 12.5 cm2/s. For a climate sensitivity of 2.5 degrees C an
increase in diffusion coefficient from 0.5 to 12.5 cm2/s leads to a
decrease in DeltaT91-100 from 3.0 degrees C to 2.4 degrees C (not shown).
This shows that for the upper part of the sensitivity range the impact of
uncertainty in the rate of the heat uptake by the deep ocean on surface warming
is comparable in magnitude with that of some other uncertainties, for example,
uncertainty in radiative forcing associated with aerosols. In the case of low
climate sensitivity the impact of the deep ocean on warming is much smaller
(Figure 5.1).
Sea level rise due to thermal expansion (Figure 5.2), in contrast with the
surface warming, is very sensitive to the rate of heat penetration into the
deep ocean especially for low climate sensitivity. An increase in diffusion
coefficient from 0.5 to 12.5 cm2/s leads to a doubling of sea level
rise by the end of the simulation with DeltaTeq = 1.6 degrees C, while
causing about a 40% change for DeltaTeq = 4.5 degrees C. It is worth
noting, that, while in the case of high climate sensitivity an increase in sea
level due to thermal expansion may be somewhat offset (Wigley and Raper, 1993;
IPCC, 1996) by the decrease in land ice melting (or vice versa), this
would not be the case for low climate sensitivity.
Our results show that there is a wide disagreement between coupled AOGCMs
simulations on the rate of heat uptake by the ocean. The corresponding
uncertainty in the surface warming is comparable in magnitude with the
uncertainties in other parameters. The impact of oceanic heat uptake on the sea
level rise is more complicated and strongly depends on chosen values of model
parameters. As a whole, the impact of the uncertainty in oceanic heat uptake is
significant enough to be taken into consideration in determining overall
uncertainty in climate change.
Acknowledgments. We thank Ron Stouffer, Catherine Senior and Reinhard
Voss for providing the results of the simulations with the GFDL, UKMO and MPI
GCMs, respectively. We also thank Gary Russell for providing us with the
procedure for surface wind calculation.
Boer G.J., 1993, Climate change and the regulation of the surface moisture and
energy budget, Climate Dynamics, 8, 225-239
Cess R.D., et al., 1990, Intercomparison and interpretation of climate
feedback processes in 19 atmospheric general circulation models, J. Geophys.
Res., 95, 16,601-16,615.
Cess R.D., et al., 1993, Uncertainties in carbon dioxide radiative
forcing in atmospheric general circulation models, Science, 262,
1251-1255.
Cubasch U., et al., 1992, Time-dependent greenhouse warming computations
with a coupled ocean-atmosphere model, Climate Dynamics, 8,
55-69.
Gregory, J.M., 1993, Sea level changes under increasing atmospheric
CO2 in a transient coupled ocean-atmosphere GCM experiment, J.
Climate 6, 2247-2262.
Gleckler, P.J., et al., 1995, Cloud-radiative effects on implied oceanic
energy transports as simulated by atmospheric general circulation models,
Geophys. Res. Lett, 22, 791-794.
Hansen J., et al., 1983, Efficient three dimensional global models for
climate studies: Models I and II, Mon. Wea. Rev., 111, 609-662.
Hansen J., et al., 1984, Climate sensitivity: Analysis of feedback
mechanisms, In: "Climate Processes and Climate Sensitivity," Geophys.
Monogr. Ser., 29, J.E. Hansen and T. Takahashi (eds.), pp.
130-163, AGU.
Hansen J., et al., 1988, Global climate change as forecast by Goddard
Institute for Space Studies three-dimensional model, J. Geoph. Res.,
93, D8, 9341-9364.
Hansen, J., et al., 1993, How sensitive is world's climate?, National
Geographic Research and Exploration, 9, 142-158.
Hansen, J., M. Sato and R. Ruedy, 1997, Radiative forcing and climate response,
J. Geoph. Res., in press.
Hansen, J., et al., 1997, Forcings and chaos in interannual to decadal
climate change, in preparation.
Hasselmann K., R. Sausen, E. Maier-Reimer and R. Voss, 1993, On the cold start
problem in transient simulations with coupled atmosphere-ocean models,
Climate Dynamics, 9, 53-61.
IPCC, 1990, Climate Change: The IPCC Scientific Assessment, Cambridge
University Press. Cambridge, UK.
IPCC, 1992, Climate Change 1992: The Supplementary Report to the IPCC
Scientific Assessment, Cambridge University Press, Cambridge, UK.
IPCC, 1996, Climate Change 1995: The Science of Climate Change,
Cambridge University Press, Cambridge, UK, 572 p.
Jacoby, H.D. and R.G. Prinn, 1994, Uncertainty in Climate Change Policy
Analysis, MIT Joint Program on the Science and Policy of Global
Change, Report 1, MIT.
Jonas, M., et al., 1996, Grid point surface air temperature calculations
with a fast turnaround: Combining the results of IMAGE and a GCM, Climatic
Change, 34, 469-512.
Leemans, R. and W.P. Cramer, 1990, The IIASA Database for Mean Monthly
Values of Temperature, Precipitation and Cloudiness on a Global Terrestrial
Grid, WP-90-41, International Institute for Applied System Analysis,
Luxemburg, Austria.
Levitus, S., 1982, Climatological Atlas of the World Ocean, NOAA
Professional Paper 13, Washington, D.C.
Manabe, S., et al., 1991, Transient responses of a coupled
ocean-atmosphere model to gradual changes of atmospheric CO2. I:
Annual mean response, J. of Climate, 4, 785-818.
Meleshko, V.P., et al., 1991, Atmospheric general circulation/Mixed
layer ocean model for climate studies and long-range weather prediction,
Meteorology and Hydrology, 5, 1-9.
Murphy, J.M., 1995, Transient response of the Hadley Centre coupled
ocean-atmosphere model to increasing carbon dioxide. III: Analysis of
global-mean responses using simple models. J. of Climate, 8,
496-514.
Murphy, J.M. and J.F.B. Mitchell, 1995, Transient response of the Hadley Centre
coupled ocean-atmosphere model to increasing carbon dioxide. II: Spatial and
temporal structure of response, J. of Climate, 8, 57-80.
Oberhuber, J.M., 1988, An Atlas Based on the "COADS" Data Set, Max
-Planck-Institute for Meteorology, Report 15, MPI.
Peixoto, J.P. and A.H. Oort, 1992, Physics of Climate, AIP, New York,
520 pp.
Prinn, R., et al., 1996, Integrated global system model for climate
policy analysis. I: Model framework and sensitivity study, MIT Joint Program
on the Science and Policy of Global Change, Report 7, MIT.
Rind D., et al., 1995, The role of sea ice in 2 x CO2
climate model sensitivity. I: The total influence of sea ice thickness and
extent, J. of Climate, 8, 449-463.
Russell, G.L., J.R. Miller and L.-C. Tsang, 1985, Seasonal ocean heat transport
computed from an atmospheric model, Dyn. Atmos. Oceans, 9,
253-271.
Santer, B.D., et al., 1995, Towards the detection and attribution of an
anthropogenic effect on climate, PCMDI, Report 21, LLNL.
Senior, C. A. and J. F. B. Mitchell, 1993: Carbon Dioxide and Climate: The
Impact of Cloud Parameterization, J. of Climate, 6, 393-418.
Sokolov, A.P. and P.H. Stone, 1995, Description and validation of the MIT
version of the GISS 2D model, MIT Joint Program on the Science and Policy of
Global Change, Report 2, MIT.
Sokolov, A.P. and P.H. Stone, 1996, Global warming projections: Sensitivity to
deep ocean mixing, MIT Joint Program on the Science and Policy of Global
Change, Report 11, MIT.
Stone, P.H. and M.-S. Yao, 1987, Development of a two-dimensional zonally
averaged statistical-dynamical model. II: The role of eddy momentum fluxes in
the general circulation and their parameterization, J. Atmos. Sci.,
44, 3769-3536.
Stone, P.H. and M.-S. Yao, 1990, Development of a two-dimensional zonally
averaged statistical-dynamical model. III: The parameterization of the eddy
fluxes of heat and moisture. J. of Climate, 3, 726-740.
Trenberth, K.E. and A. Solomon, 1994, The global heat balance: Heat transports
in the atmosphere and ocean, Clim. Dyn., 10, 107-134.
Walsh, J. and C. Johnson, 1979, An analysis of Arctic sea ice fluctuations,
J. Phys. Oceanogr., 9, 580-591.
Wang, C., et al., 1995, A coupled atmospheric chemistry and climate
model for chemically and radiatively important trace species, WMO-IGAC,
WMO/TD , 720, 182-184.
Washington, W.M. and G.A. Meehl, 1993, Greenhouse sensitivity experiments with
penetrative cumulus convection and tropical cirrus albedo effects, Climate
Dynamics, 8, 211-223.
Wetherald, R.T. and S. Manabe, 1988, Cloud feedback processes in a general
circulation model, J. Atmos. Sci., 45, 1397-1415.
Wigley, T.M.L. and S.C.B. Raper, 1993, Future changes in global mean
temperature and sea level, In: "Climate and Sea Level Change:
Observations, Projections and Implications," R.A. Warrick, E.M. Barrow and
T.M.L. Wigley (eds.), Cambridge University Press, Cambridge, UK, p. 111-133.
Wilson, C.A. and J.F.B. Mitchell, 1987, A doubled CO2 climate
sensitivity experiment with a global climate model including a simple ocean,
J. Geophys. Res., 92, D11, 13,315-13,343.
Xiao, X., et al., 1997, Linking a global terrestrial biogeochemical
model with a 2-dimensional climate model: Implication for the global carbon
budget, Tellus, 49B, 18-37.
Yao, M.-S. and P.H. Stone, 1987, Development of a two-dimensional zonally
averaged statistical-dynamical model. I: The parameterization of moist
convection and its role in the general circulation, J. Atmos. Sci.,
44, 65-82.
Figure 3.2 Same as Fig. 3.1 but for JJA.
Figure 3.3 Zonally averaged evaporation (top) and precipitation
(bottom) simulated by the MIT 2D model, GISS GCM and observations for DJF.
Figure 3.4 Same as Fig. 3.3 but for JJA.
Figure 3.5 Zonally averaged cloud radiative forcing simulated by the
MIT 2D model (unlabeled solid line) and observations for DJF (top) and JJA
(bottom).
Figure 3.6 Annual mean northward atmospheric (top) and oceanic (bottom)
meridional energy transport from the MIT 2D model.
Figure 3.7 Zonally averaged sea surface temperature (top), ocean ice
cover (middle) and ice depth (bottom) simulated by the MIT 2D model (unlabeled
solid line), GISS GCM and observations for DJF. Data for sea surface
temperature and sea ice extent are from Alexander and Mobley (1974) and Walsh
and Johnson (1979); sea ice depth is calculated from sea ice extent as
described by Hansen et al., 1984.
Figure 3.8 Same as Fig. 3.7 but for JJA.
Figure 3.9 Globally averaged annual mean surface air temperature
variation in the simulation with the 2D model coupled with mixed layer ocean
model (top) and that in the simulation with diffusive mixing of mixed layer
temperature perturbations into the deep ocean model (bottom).
Figure 4.1 Percentage change in globally and annually averaged
precipitation as a function of global mean warming as produced by different
GCMs (triangles) from IPCC 1990 and different versions of the MIT 2D model
(circles).
Figure 4.2 Changes in in globally and annually averaged surface latent
(top) and sensible (bottom) heat fluxes as produced by different versions of
the UKMO AGCM (triangles) and different versions of the MIT 2D model
(circles).
Figure 4.3 The same as Fig. 4.2 but for net shortwave and longwave
radiation at the surface.
Figure 4.4 Change in surface air temperature (top), precipitation
(middle) and cloud cover (bottom) due to CO2 doubling for DJF in
simulations with versions of the MIT 2D model with sensitivities 1.6, 3.0 and
4.6 degrees C and the GISS AGCM.
Figure 4.5 Same as Fig. 4.4 but for JJA.
Figure 4.6 Change in precipitable water (top) and sea ice cover
(bottom) due to CO2 doubling for DJF in simulations with versions of
the MIT 2D model with sensitivities 1.6, 3.0 and 4.6 degrees C and the GISS
AGCM.
Figure 4.7 Same as Fig. 4.6 but for JJA.
Figure 4.8 Height-latitude cross section of changes in temperature for
DJF due to CO2 doubling in simulations with versions of the MIT 2D
model with sensitivities 4.6 (top), 3.0 (middle) and 1.6 degrees C (bottom).
Contour interval 1 degrees C; negative contours dashed.
Figure 4.9 Same as Fig. 4.8 but for JJA.
Figure 4.10 Height-latitude cross section of changes in annual mean
cloud cover due to CO2 doubling in simulations with versions of the
MIT 2D model with sensitivities 4.6 (top), 3.0 (middle) and 1.6 degrees C
(bottom). Contour interval 1%; negative contours dashed.
Figure 4.11 Global mean surface temperature change caused by a 4%, 1%,
0.5% and 0.25% per year increase in CO2 in the simulations with the
MIT 2D model with DeltaTeq =3.7 degrees C and K v=5.0
cm2/s (solid curves) and the GFDL AOGCM (dashed curves).
Figure 4.12 Global mean surface temperature change caused by increases
in CO2 in simulations with the MPI (scenarios A and D from Cubasch
et al., 1992) and NCAR (1% per year) AOGCMs (dashed curves) and by the
matching versions of the MIT 2D model with the same scenarios for increasing
CO2 (solid curves).
Figure 4.13 Changes in annual mean zonal mean surface air temperature
(top) and sea surface temperature (bottom) in the simulations with the MIT 2D
model with sensitivity 3.7 degrees C. Equilibrium response to CO2
doubling (solid curves) and transient change in the decade of CO2
doubling (dashed curves).
Figure 4.14 Temporal variation of zonally averaged annual mean surface
air temperature (difference from control climate) produced by the version of
the MIT 2D model with DeltaTeq = 3.7 degrees C and Kv = 5.0
cm2/s.
Figure 4.15 Same as Fig. 4.14 but for sea surface temperature.
Figure 4.16 Same as Fig. 4.14 but for sea ice thickness.
Figure 4.17 Depth-latitude cross section of deep ocean temperature
changes in the decade of CO2 doubling in the simulation with the
version of the MIT 2D model with DeltaTeq = 3.7 degrees C and Kv
= 5.0 cm2/s. Contour interval solid curves every 0.5 degrees C;
dashed curve 0.25 degrees C.
Figure 4.18 Global mean sea level rise due to thermal expansion caused
by a 4%, 1%, 0.5% and 0.25% per year increase in CO2 in the
simulations with the MIT 2D model with DeltaTeq = 3.7 degrees C and
Kv = 5.0 cm2/s (solid curves) and the GFDL AOGCM (dashed
curves).
Figure 4.19 Global mean sea level rise due to thermal expansion caused
by increases in CO2 in simulations with the MPI AOGCM (scenarios A
and D from Cubasch et al., 1992, dashed curves) and by the matching
version of the MIT 2D model with the same scenarios for increasing
CO2 (solid curves).
Figure 5.1 Global mean surface temperature change caused by a 1% per
year increase in CO2 in simulations with DeltaTeq = 4.5
degrees C (solid curves, with Kv = 0.5, 2.5 and 12.5 cm2/s,
as indicated) and DeltaTeq = 1.6 degrees C (dashed curves with Kv
= 0.5, 2.5 and 12.5 cm2/s, as indicated) together with upper
and low bounds (shown by straight lines) for the IPCC's projections for the
same scenario (IPCC, 1990, Fig. 6.8).
Figure 5.2 Sea level rise due to thermal expansion caused by a 1% per
year increase in CO2 in the simulations with DeltaTeq = 4.5
degrees C (solid curves, with Kv = 0.5, 2.5 and 12.5 cm2/s,
as indicated) and DeltaTeq = 1.6 degrees C (dashed curves with Kv
= 0.5, 2.5 and 12.5 cm2/s, as indicated).
[1] In previous simulations (see Sokolov and Stone, 1996) we concluded that our
"standard" diffusion coefficients give a good match to the GFDL AOGCM. However,
the earlier simulations were carried out with a version of the 2D model with
sensitivity 3.5 degrees C (the value for the GFDL GCM given by Murphy and
Mitchell, 1995). In the simulations described here we use the value 3.7
degrees C for the GFDL model sensitivity, as given in IPCC, 1996, and by
Stouffer (personal communication).
[2] It should be noted that radiative forcing
produced by the MPI AGCM in response to the CO2 doubling is about
0.5 W/m2 less than that produced by the GISS AGCM (Cess, et
al., 1993) and the MIT 2D model.
| Contents |
Table 2.1. Coefficients of heat diffusion into the deep ocean
(cm2/s).
3. Present-Day Climate Simulation | Contents |
In developing the MIT 2D climate model a significant number of
present-day climate simulations were performed and different versions of the
above mentioned parameterization schemes were tested (Sokolov and Stone, 1995).
Here we show the results averaged for twenty years of a run carried out with
the 2D model coupled with the mixed layer ocean model as described above.
Figures 3.1 and 3.2 show zonal wind and meridional streamfunction averaged for
December, January, February (DJF) and June, July, August (JJA) for the 2D
model, both of which are simulations similar to those observed (e.g.,
see Peixoto and Oort, 1992), except for the streamfunction being too weak in
Northern Hemisphere winter. Precipitation and evaporation obtained in
simulations with the 2D model and the GISS (Model II) GCM (Hansen, et
al., 1983) are shown in Figures 3.3 and 3.4 together with observations
(Leemans and Cramer, 1990; Oberhuber, 1988). Both the 2D and GISS models have
difficulty matching the observed precipitation particularly, in DJF (Figure
3.3). The underestimation of precipitation in the tropics in DJF by the 2D
model is consistent with the above mentioned deficiency of the simulated mean
meridional circulation. As one would expect, the agreement between the results
of the 2D model simulation and observed precipitation in the equatorial region
is much better in JJA. It should be mentioned that there are in any case
significant disagreements among observational data sets for precipitation. The
pattern of evaporation is reasonably well reproduced by the 2D model.4. The 2D Model Response to Instantaneous and Gradual Increases in the Atmospheric CO2 Concentration
4a. Equilibrium response to a doubling of CO2 concentration | Contents |
When using a 2D model to study uncertainty in climate change, it
is desirable to have a model capable not only of simulating the present-day
climate but also of reproducing the climate change pattern obtained in
simulations with different GCMs. In this section responses of different
versions of the MIT 2D model to an instantaneous doubling of CO2
concentration in the atmosphere are compared with the results of similar
simulations with different GCMs. The atmospheric model coupled with the mixed
layer ocean model was used in these simulations. The seasonally changing
horizontal heat transport by the ocean was specified from the results of the
climate simulation with climatological sea surface temperature and sea-ice
distribution and was held fixed, thereby neglecting the impact of possible
changes in ocean circulation. This assumption has been always considered to be
a significant weakness of equilibrium climate change simulations with mixed
layer ocean models. However, as discussed in IPCC (1996), the results of a
simulation with the GFDL AOGCM suggest that the oceanic heat transport for an
equilibrium doubled CO2 climate is similar to that for the
present-day climate (see IPCC, 1996, Section 6.2.4). In any case the GCM
simulations we compare with were carried out under the same assumption.
Table 4.1 Change in globally averaged annual mean surface air
temperature, precipitation, cloud cover and cloud forcing due to a doubling of
the atmospheric CO2 .
Table 4.2 Change in globally and annually averaged terms of the surface
energy budget due to a doubling of the CO2 concentration.
4b. Transient response to a gradual CO2 increase | Contents |
The transient behavior of different AOGCMs can be matched by
choosing appropriate values for the model's sensitivity and the rate of heat
diffusion into the deep ocean. The change in the latter was obtained by
multiplying the "standard" diffusion coefficients (Table 2.1) by the same
factor at all latitudes, thereby preserving the latitudinal structure of heat
uptake by the deep ocean. The time dependent globally averaged surface warming
produced by different versions of the 2D model are compared with the results of
simulations with the GFDL, MPI and NCAR AOGCMs in Table 4.3 and Figures 4.11
and 4.12. Values for sensitivities of the GCMs are taken from IPCC, 1996.
Table 4.3 Responses of different AOGCMs and matching versions of the MIT
2D model to a gradual increase of CO2 concentration.
5. Uncertainty in the Rate of Heat and Carbon Uptake by the Deep Ocean and its Impact on the Transient Model Response to the Increase of Atmospheric Greenhouse Gas Concentrations | Contents |
Uncertainty in the rate of heat uptake by the deep ocean has not been
included in the projections of climate change made by the Intergovernmental
Panel on Climate Change (IPCC, 1990 and 1996). For example, the IPCC (1996)
projections of a warming of 1 degrees C to 3.5 degrees C by 2100 only take into
account uncertainties in greenhouse gas emission scenarios and in climate
sensitivity, while the rate of heat uptake was chosen to reproduce results of
the GFDL AOGCM. The same is true for IPCC projections of possible sea level
rise. Among the reasons for not taking uncertainty in the heat uptake into
account is, apparently, the conclusion of Wigley and Raper (1993) that it is
relatively unimportant. However, the results for different coupled AOGCMs given
above, together with the absence of any direct measurements of heat uptake by
the ocean, indicate that there is more uncertainty than Wigley and Raper
assumed. Thus, as an example of the use of the MIT 2D model, in this section we
present results of simulations with the MIT 2D model concerning sensitivity of
surface warming and sea level change to the rate of ocean heat uptake. While
being qualitatively similar to those presented by Wigley and Raper (1993), the
results lead us to a different conclusion.6. Conclusions | Contents |
The results presented from the simulations with the MIT 2D climate
model show that it, while having some limitations compared to GCMs, reasonably
reproduces the main features of the present-day climate, including seasonal
variability. Both globally averaged values and zonal distributions of
equilibrium changes in the different climate variables, such as temperature,
precipitation, evaporation, and radiation balance at the surface, as produced
by different versions of the 2D model in response to a doubling of the
atmospheric CO2, are similar to those obtained in simulations with
different GCMs. Use of an artificial cloud feedback for changing the 2D model
sensitivity does not produce any physically inconsistent results. The 2D model
coupled with a diffusive ocean model, in spite of the simplicity of the latter,
reasonably reproduces the transient behavior of different AOGCMs, at least for
climate simulations on time scales of 100-150 years. These results together
with the relatively moderate computer resource requirements make the 2D model a
very useful tool for studying uncertainty in climate change both independently
and as a part of the integrated framework of the MIT Joint Program on the
Science and Policy of Global Change.7. References | Contents |
Alexander, R.C. and R.L. Mobley, 1974, Monthly Average Sea Surface
Temperatures an Ice-pack Limits on 1 degree global grid, Rep. 4-1310-ARPA, Rand
Corp., Santa Monica, CA, 30 pp.
Figure 3.1 Height-latitude cross section of zonally averaged zonal wind
(top) and meridional streamfunction (bottom) for DJF from the MIT 2D model.
Endnotes:
[*] Center for Global Change Science,
Massachusetts Institute of Technology, 77 Massachusetts Ave. Room 54-1312,
Cambridge, MA 02139, USA.