S S S Eq N N N

300 400 500

700 850 1000

300 400 500

700 850 1000

240

^ ___220--(

I "200" ]) \

V —200 J

V--210----

----—220-—

230—JX

—240-CX

—250—00-0"-—

260—OSO—---

--270—-

-280-X\O\

30S EQ 30N 60N 90N 90S 60S 30S EQ 30N Figure 2 As in Fig. 1, but for the NCEP climatology.

60N 90N

polar stratosphere. In addition to their general performance, specific aspects of model performance have been investigated in AMIP. For example, a comparison of precipitation and evaporation over the continents was made by Lau et al. (1996). We present our results in the same format in Fig. 3. Our result is close to the observed value. This result also confirms that the performance of our model is comparable with that of other models.

Figure 3 Mean rainfall and surface air temperature averaged over land based on the AMIP results. The CCSR result is plotted on Fig. 1 by Lau et al. (1996). Abbreviations: BMR, Bureau of Meteorology; CCC, Canadian Climate Center; CCSR, Center for Climate System Research; CNR, Centre National de Recherches Météorologiques; COL, COLA; CSI, Commonwealth Scientific and Industrial Research Organization; CSU, Colorado State University; DER, Dynamical Extended Range Forecasting at GFDL; DNM, Department of Numerical Mathematics of the Russian Academy of Sciences; ECM, ECMWF; GFD, GFDL; GIS, GISS; GSF, GSFC; IAP, Institute of Atmospheric Physics; J MA, Japan Meteorological Agency; LMD, Laboratoire de Météorologie Dynamique; MGO, Main Geophysical Observatory; MPI, Max-Plank-Institut; MRI, Meteorological Research Institute; NCA, NCAR; NRL, Naval Research Laboratory; OBS, Observation; SUN, State University of New York at Albany; UCL, UCLA; UGA, U.K. Universités' modeling program, UIU, University of Illinois; UKM, U.K. Meteorological Office; YON, Yonsei University, Korea.

Figure 3 Mean rainfall and surface air temperature averaged over land based on the AMIP results. The CCSR result is plotted on Fig. 1 by Lau et al. (1996). Abbreviations: BMR, Bureau of Meteorology; CCC, Canadian Climate Center; CCSR, Center for Climate System Research; CNR, Centre National de Recherches Météorologiques; COL, COLA; CSI, Commonwealth Scientific and Industrial Research Organization; CSU, Colorado State University; DER, Dynamical Extended Range Forecasting at GFDL; DNM, Department of Numerical Mathematics of the Russian Academy of Sciences; ECM, ECMWF; GFD, GFDL; GIS, GISS; GSF, GSFC; IAP, Institute of Atmospheric Physics; J MA, Japan Meteorological Agency; LMD, Laboratoire de Météorologie Dynamique; MGO, Main Geophysical Observatory; MPI, Max-Plank-Institut; MRI, Meteorological Research Institute; NCA, NCAR; NRL, Naval Research Laboratory; OBS, Observation; SUN, State University of New York at Albany; UCL, UCLA; UGA, U.K. Universités' modeling program, UIU, University of Illinois; UKM, U.K. Meteorological Office; YON, Yonsei University, Korea.

D. Transient Experiments to Explore the Effects of Increasing C02

The AGCM and the OGCM have been coupled to each other, along with a thermodynamic sea-ice model and a simple river runoff model. The AGCM used in the coupled model is the T21L20 version, and the OGCM

grid uses the same Gaussian grid as the AGCM (approximately 2.8° X 2.8°). The OGCM has 17 levels.

Using this coupled model, transient experiments on COz doubling were conducted, including a control run with COz fixed at present values. We used a 1% per year increase of C02 in order to participate in CMIP. Such participation in international comparison projects helps us to evaluate our model's performance, because the results of other research centers can be obtained and it is easy to compare our model results with them. In the transient experiment, we decided to use flux correction. This ensures that the mean SST distribution is similar to the observed one for the control run. In addition to the means, note that variabilities are well simulated. For example, the coupled model produces an ENSO-like fluctuation over the tropical Pacific Ocean and a decadal fluctuation over the northern Pacific Ocean, although their amplitudes are less than observed.

In Fig. 4, the globally averaged surface temperatures are shown for a run with a sudden increase of C02 and a run with the transient increase of COz, along with the control run. There is about a 2 K increase in the transient case, which is similar to results obtained with other models. The spatial pattern of the surface temperature increase is also similar to those

Figure 4 The time sequence of the annual average of global surface temperature based on the transient C02 experiments by the CCSR climate model. The thick full line is for the control case, and the dotted line is for the transient CO, case (1% increase). The thin full line is for the sudden CO, doubling case.

Figure 4 The time sequence of the annual average of global surface temperature based on the transient C02 experiments by the CCSR climate model. The thick full line is for the control case, and the dotted line is for the transient CO, case (1% increase). The thin full line is for the sudden CO, doubling case.

obtained with other models: There is more warming over the continents in the Northern Hemisphere, and less warming over the oceans and in the Southern Hemisphere.

In Fig. 5, the difference of the precipitation fields between the doubled-COz case and the control case is shown. Positive differences occur in the equatorial central Pacific and the Asian monsoon region. Note that the difference over the tropical Pacific basin is similar to the ENSO pattern. To investigate the reason for this change, we calculated the difference of the surface heat budget over the tropical Pacific Ocean. Based on these results, the following scenario can be presented: (1) When C02 increases, the sea surface temperature (SST) increases over the tropical Pacific basin. (2) However, in the western Pacific region, the reflection of solar radiation by clouds increases and so the increase of SST in this region is not as large as in the eastern Pacific. (3) As a result, the east-west SST gradient tends to decrease, that is, an ENSO-like situation is realized. (4) The center of convection tends to shift eastward and becomes located in the central Pacific. The circulation pattern also changes. For example, the southeaster-lies over southern China are enhanced. This feature has been noted in observations (Zhang et al., 1996). These responses to an increase of C02 are similar to the results of others (Meehl and Washington, 1996). Further work is needed to determine whether or not this response is model dependent.

E. Simulation of the QBO

An example of scenario 2 given in the preceding paragraph is simulation of the QBO (quasi-biennial oscillation). In the past, climate models have been criticized because they could not simulate the QBO, even though it has been theoretically explained (Lindzen and Holton, 1968). However, the physics contained in our climate model should include that of the simple theory, so we could not believe that our model could not simulate the QBO as long as it is dynamically maintained by the Kelvin and the mixed-Rossby waves. The AGCM has been expanded to include the stratosphere in order to investigate the QBO. A simulation of a QBO-like phenomenon by using three-dimensional GCM was first conducted by Takahashi and Boville (1992), who noted that the amplitude of the mixed-Rossby wave is larger than observed. Several researchers pointed out that the contributions of gravity waves are essential to simulate the QBO, so we introduced high vertical and horizontal resolution and weak diffusion in our simulation. As a result of these efforts, we were able to simulate the QBO (Takahashi, 1996; see Fig. 6).

CCSR annual precip, C02x2-xl rrrri/y e ar

0 0

Post a comment