|
MODELLING WITH
|
|
Monsoon Variation
The interannual variation of the south Asian and Indian summer monsoon is analysed based on one 21 year and three 10 year simulations of the present day climate using the U.K. Universities' Global Atmospheric Modelling Programme (UGAMP) GCM with different land surface parametrization schemes and sea surface temperature (SST) variations. Generally, a negative relationship is found between south Eurasian winter/spring snow mass (cover) and the amount of summer monsoon rainfall over India in all simulations. However, the significance of this relationship is dependent on the land surface parametrization scheme. The simulations using the no-flux boundary condition at the bottom of the 3 layer soil model give a strong negative correlation. Fig. 26a shows the normalized snow mass deviation in February to May (FMAM) over south Eurasia and the corresponding normalized summer from June to September (JJAS) rainfall over the India, north India, and south India for the simulation with the no-flux boundary condition and the A&M climatological SSTs. The results indicate an inverse relationship between the snow mass in FMAM and Indian summer monsoon rainfall. The correlation coefficients between snow mass and precipitation over India and north India in summer and early summer are well above the 95% confidence level. Excess than normal snow mass over south Eurasia in winter and spring is associated with less than normal summer monsoon precipitation over India, especially over north India, and vice versa. Fig. 26b gives spatial distribution of the correlation coefficients between snow mass in FMAM over south Eurasia and the precipitation JJAS and indicates that there is a negative correlation in most regions over India except for the southern tip of the peninsula. The maximum negative correlation coefficients are found over north India, between 20N and the foothills of the Himalayas. This spatial distribution is in good agreement with observational studies Composite analyses suggest that weak winter/spring snowfall over south Eurasia is associated with a strong Indian summer monsoon, characterized by strong southwesterlies over the Arabian Sea in the lower troposphere in JJA season and heavy precipitation in early summer over north India and the foothills of the Himalayas. In contrast, heavy winter/spring snowfall delays the onset of the Indian summer monsoon through the feedback of snowmelt, soil moisture and evaporation processes, and is associated with weak summer precipitation over the two regions. Sensitivity studies have confirmed that the snow-Indian monsoon relationship identified in the simulation at T42 is also robust in the simulation at T31. However, such negative relationship does not exist in the simulation at T21 horizontal resolution, indicating the importance of horizontal resolution in maintaining the snow-monsoon relationship in the UGAMP model.
Dong, B.-W. and Valdes, P. J., 1997: Asian and Indian summer monsoon rainfall and Eurasian winter/spring snow mass. Draft available from authors.
Initial results with the Extended Extended UGCM (EEUCGM)
In order to avoid upper boundary effects when examining the mesosphere in the EUGCM (see UGAMP technical report No. 41), the EUGCM has been extended upward with the addition of 5 more model levels to form the EEUGCM (which has a top at approximately 128 km). The EEUGCM, reassuringly, runs with exactly the same model parameters as used in the EUGCM, apart from the CO2 transmittance table which has been recalculated for the extra model levels.
It is intended to use the EEUGCM to examine the tides, the two-day wave and other planetary waves in the mesosphere, and their interaction with gravity wave drag. This will build on the work reported in Norton and Thuburn (1996a,b). This article presents some preliminary results from the EEUGCM.
Fig. 27 shows the zonal mean winds, averaged over 1-7 July, from two runs of the EEUGCM. Fig. 27a is from a run where the gravity wave drag parametrization has been made to break strongly. This deposits all the vertical momentum flux of the parametrized gravity waves over a relatively small vertical height range and hence produces large drag on the mean flow, which in the northern hemisphere near 80 km is over 200 m/s/day. In contrast, Fig. 27b is from a run where the gravity wave drag parametrization breaks weakly and the maximum gravity wave drag is only 80 m/s/day. The most noticeable differences between the two runs in the zonal mean wind, is the stronger upper mesosphere/thermosphere jets (above 80 km) in Fig. 27a and the corresponding steeper shear zones. There is also some variability in the stratosphere but this is probably not significant. Examination of the zonal mean temperatures (not shown) shows that the summer mesopause temperatures are 120 K for the strong gravity wave drag run (which is probably too cold by about 10 K), and 160 K for the weak gravity wave drag run (which is too warm). This strongly confirms the fact the gravity wave drag is responsible for the very cold summer mesopause.
The mesospheric planetary waves are very different between these two model runs. Figs. 28a,c shows northern hemisphere PV (summer hemisphere) on the 3000 K isentrope (near 64 km) for 1 and 31 July respectively from the strong gravity wave drag run. Figs. 28b,d shows the corresponding plots from the weak gravity wave drag run. Evident in Fig. 28a is a strong wave-3 pattern; this is the two-day wave. However Fig. 28b shows that the two-day wave is much weaker in the weak gravity wave drag run. This was expected since we believe the two-day wave is excited by baroclinic instability, and the vertical shear on the top of the summer easterly jets is much stronger in the strong gravity wave drag run. Figs. 28c,d show the two-day wave in late July from the two runs. In the weak gravity wave drag run, Fig. 28d, the two-day wave has almost completely disappeared, while in the strong gravity wave drag run, the two-day wave has become almost chaotic, Fig. 28c shows a transition from wave-3 to wave-4 occurring.
The migrating diurnal tide in the two runs are also very different. Figs. 29a,b shows the temperature amplitude of the tides for the period 1-7 July from the strong and weak gravity wave drag runs respectively. The amplitude of the tides in the strong gravity wave drag run is considerably smaller than in the weak gravity wave drag run. There are several possible explanations for this. It could be the change in zonal mean state between the two runs has affected the propagation of the tides; alternatively it could be the influence of the gravity wave drag directly on the tides; a third possibility is that there has been a nonlinear interaction with the large amplitude two-day wave in the strong gravity wave drag run which has reduced the amplitude of the tides. The investigation of these possibilities is the subject of future work.
References
Norton, W. A., and Thuburn, J., 1996: The two- day wave in a middle atmosphere GCM. Geophys. Res. Lett., 23, 2113-2116.
Norton, W. A., and Thuburn, J., 1996: The Mesosphere in the Extended UGAMP GCM. To appear in: Gravity wave processes and their parameterization in global climate models. Ed. K. Hamilton, Springer-Verlag.