FUTURE PROJECTIONS OF PRECIPITATION
CHARACTERISTICS IN ASIA
AKIO KITOH, MASAHIRO HOSAKA, YUKIMASA ADACHI, KENJI KAMIGUCHI
Meteorological Research Institute
Tsukuba, Ibaraki 305-0052, Japan
It is anticipated that global warming will result in large changes in Asian climate,
including changes in the monsoon intensity and duration in South Asia, Southeast Asia
and East Asia, and even in ENSO-monsoon relationship. Using the daily precipitation
output of global climate model simulations under the SRES A2 and B2 scenarios,
projected changes in precipitation characteristics around mid-21st century and end-ofthe-century are analyzed. It is found that both the frequency and intensity increase in
about 40% of the globe, while both the frequency and intensity decrease in about 20% of
the globe. Over the rest of the globe (about one third), the precipitation frequency
decreases but its intensity increases, suggesting a shift of precipitation distribution
toward more intense events by global warming. East Asia is a region where summertime
wet-day frequency decreased but precipitation intensity increased. This is related to
increased atmospheric moisture content due to global warming and an El Nino-like mean
SST change and associated circulation changes, i.e. a strengthened subtropical high.
MODEL AND EXPERIMENT
We conducted the history simulation during 1850-2000 with observed atmospheric
greenhouse gas concentrations and aerosols (history run) and then future projections up
to 2100 under the SRES A2 and B2 scenarios (IPCC [1]). The model used is the MRI
coupled ocean-atmosphere GCM (MRI-CGCM2, Yukimoto et al. [2]). The atmospheric
part of the model has a horizontal resolution of T42 (about 280km) and 30 vertical levels.
The oceanic part adopted variable resolution 0.5°-2.0° for latitudinal direction with finer
grid in the tropics and fixed 2.5° for longitudinal direction. The model performance in
simulating Asian summer monsoon is found in Rajendran et al. [3]. Three ensemble runs
were performed for each scenario from different initial conditions. Daily precipitation
output was analyzed.
RESULTS
In this report, we compare 3 ensemble averages of the mid-21st century 20 yr (2041–
2060) simulations and the end of the 21tst century (2081–2100) under the SRES-A2 and
B2 scenarios with 3 ensemble averages of the present (1981-2000) simulations.
Figure 1 shows the time series of the annual and global mean surface air temperature
and precipitation of the history run and the A2 and B2 scenario runs. The annual mean
temperature indicates a strong warming trend particularly towards late 21st century in the
1
2
SRES-A2 scenario to attain a warming of 2.19oC at 2100 compared to the present. In the
SRES-B2 scenario the warming is less (1.40 oC). At the mid-21st century, the difference
in the warming is small between A2 (0.84 oC) and B2 (0.72 oC). Changes in the annual
mean precipitation show an increase of 4.54% in SRES-A2 scenario, while 2.48% in the
B2 scenario. Again the difference between A2 and B2 scenarios is negligible until mid21st century (1.41% and 1.57%, respectively).
Figure 1. Time series of annual and global mean (a) surface air temperature and
(b) precipitation of the history and the SRES-A2 and B2 scenario runs
Figure 2 shows the change in annual mean surface air temperature at 2041–2060 and
2081–2100 for both A2 and B2 scenarios. Strong warming is evident in high latitudes at
the end of the 21st century in SRES-A2 scenario. However its spatial pattern is similar to
each other, with the higher latitudes being warmer than the lower latitudes, and the land
surface warmer than over the oceans. It is noted that over the tropical Pacific the eastern
part warms up more than the western part, suggesting a permanent El Nino-like SST
change by global warming.
Figure 3 shows the change in annual mean total precipitation at 2041–2060 and
2081–2100 for both A2 and B2 scenarios. In the tropics, an eastward displacement of
major precipitation center is noted associated with El Nino-like background mean SST
changes (not shown). In Asia annual mean precipitation increased in both scenarios and
both time times. Increased rainfall in the South Asian summer monsoon is noted. In
higher latitudes total precipitation increased, while there is a decrease in precipitation
over the subtropical oceans and the Mediterranean region.
3
Figure 2. Annual mean surface air temperature difference (F–P) at (a) 2041–2060 for A2,
(b) 2041–2060 for B2, (c) 2081–2100 for A2, and (d) 2081–2100 for B2.
Figure 3. As in Fig.2 except for total precipitation.
4
Figure 4. As in Fig. 2 except for rainy-day frequency.
Figure 5. As in Fig. 2 except for precipitation intensity in rainy day.
5
Figure 6. Changes in three characteristics in precipitation based on total precipitation
(T), wet-day frequency (F) and precipitation intensity of wet-day (I).
Classification of types 1-8 are + + +, ++ –, + – +, + – –, – + +, – + –, – – + and
– – – for T F I, respectively.
Figure 4 shows the changes in annual rainy-day frequency. Here we defined the day
as rainy when daily precipitation exceeding 1.0 mm/d. An overall spatial pattern is
similar to that in total precipitation, but relatively large changes in frequency can be
found in the middle and high latitudes. Also the areal extent of negative changes in wetday frequency is larger than that in total precipitation
Figure 5 shows the changes in the intensity of precipitation in rainy days. An overall
pattern is again similar to Figs. 3 and 4 in the tropics and high latitudes. However, in the
mid-latitudes, there are regions with different signs between intensity and frequency. For
example, over the Mediterranean region, the frequency change is negative while the
intensity change is positive.
If areal extent of positive or negative changes is compared, total precipitation has
increased over about 65% of the globe at 2081-2100 in the SRES-A2 scenario (Fig. 3c).
It is 42% of the globe where wet-day frequency has increased, while it is 79% where the
intensity of precipitation in rainy days has increased. These numbers differ only a few
percent from decades to decades and between A2 and B2 scenarios.
6
In order to investigate the relationship among the three characteristics we classified
each grid point into 8 categories. Figure 6 shows the changes in three characteristics in
precipitation based on total precipitation, wet-day frequency and wet-day precipitation
intensity. Table 1 shows the ratio of area each type. It is shown that both the frequency
and intensity increased in about 40% of the globe (type 1), while both the frequency and
intensity decreased in about 20% of the globe (type 8). In between, which occupies
around one third of the globe, the precipitation frequency decreased but its intensity
increased (type 3 or 7). Existence of types 3 and 7 suggests a shift toward more intense
precipitation events by global warming. Moreover, the ratio of types 3 and 7 becomes
larger from the SRES-B2 scenario to the SRES-A2 scenario, and also from the period of
2041-2060 to the period 2081-2100, corresponding the magnitude of the warming.
Table 1. Relative area ratio (% of the globe) of 8 classifications for the middle (20412060) and the end (2081-2100) of the 21st century for the SRES-A2 and B2
scenario runs. Types 2, 4, 5 and 6 are grouped.
Type 1
Type 3
Type 7
Type 8
Others
Total
+
+
–
–
Freq
+
–
–
–
Intens
+
+
+
–
A2/41-60 A2/81-00
41.46
39.48
21.80
25.66
14.83
16.04
18.72
16.50
3.19
2.32
B2/41-60
45.95
20.15
12.64
18.59
2.67
B2/81-00
41.47
25.06
14.62
16.88
1.97
The Mediterranean and the Middle East is the type 7 region where both total
precipitation and number of rainy days decreased but precipitation intensity itself
increased. It is interesting to note that the observed rainfall over the Mediterranean
showed an increase of extreme rainfall in spite of decrease in total precipitation (e.g.,
Alpert et al. [4]).
The East Asia and the Southeast Asia are the type 3 regions and may be regarded as
a more flood vulnerable region in the future. Figure 7 compares the seasonal changes of
the total precipitation, wet-day frequency and precipitation intensity averaged for South
China (100oE–120oE, 26oN–33oN) between the present and the period 2081-2100 for the
SRES-A2 scenario run. The SRES-A2 scenario shows increase total precipitation in early
rainy season of March–May, while wet-day frequency decreased in mid-summer season
of May–August. However the precipitation intensity has increased almost all the year
round.
Figure 8 shows the June-August (JJA) total moisture flux and its convergence in the
model control climate and its changes in 2081-2100 of the SRES-A2 scenario. There is a
moisture flux convergence over the wide area in Asia. Increased atmospheric moisture
content in the warmer air in the scenario run leads to larger moisture flux and increased
7
precipitation (e.g., Kitoh et al. [5]). It is also noted that the Pacific subtropical
anticyclone becomes stronger and more southerly moisture flux converges over South
China. It is plausible that wet-day frequency decreases due to a westward shift and an
intensification of the subtropical anticyclone while larger moisture flux results in more
intense precipitation. Moisture flux convergence is also larger over Japan, suggesting
more intense Baiu activity in this region. These circulation anomalies are often observed
at times of El Nino events, and thus the simulated El Nino-like mean climate response in
the tropical Pacific (Fig. 2) may be related to these changes in Asia.
Figure 7. Seasonal changes of (a) total precipitation, (b) wet-day frequency and (c)
precipitation intensity for South China (100 oE–120oE, 26oN–33oN). Solid line
denotes present and dashed line denotes 2081-2100 for the SRES-A2 scenario
run. Box is plotted at abscissa where the difference between future and present
is statistically significant based on 3 ensemble members where black (dark)
color denotes future is larger (smaller) than present.
Figure 8. (a) Total moisture flux and its divergence in JJA in the control run. Shaded
area denotes moisture flux convergence, (b) Changes in total moisture flux
and its divergence in JJA for the period 2081-2100 of the SRES-A2 scenario
run.
8
SUMMARY
The daily precipitation data produced by the MRI-CGCM ensemble runs under the
SRES–A and SRES-B2 scenarios are analyzed. In the tropics, eastward displacement of
precipitation was noted associated with El Nino-like mean SST changes. From the
analysis of precipitation characteristics, it is found that over one third of the globe,
number of rainy days decreased but precipitation intensity has increased. Areal extent of
such a region will increase towards the end of the 21st century by global warming.
ACKNOWLEDGMENTS
This research was partly funded by "Study of the Prediction of Regional Climate Changes
over Japan due to Global Warming" of JMA and by "Water Resource and Its Variability
in Asia in the 21st Century" of "Special Coordination Funds for Promoting Science and
Technology" of MEXT.
REFERENCES
[1] IPCC, "Climate Change 2001: The Scientific Basis. Contribution of Working Group
I to the Third Assessment Report of the Intergovernmental Panel on Climate
Change", Houghton JT et al. (eds.), Cambridge University Press, (2001).
[2] Yukimoto S., and Coauthors, "The new Meteorological Research Institute coupled
GCM (MRI-CGCM2) - Model climate and variability", Pap. Meteor. Geophys., Vol.
51, (2001), pp 47-88.
[3] Rajendran K., Kitoh A. and Yukimoto S., "South and east Asian summer monsoon
climate and variation in the MRI coupled model (MRI-CGCM2)", J. Climate, Vol.
17, (2004), pp 763-782.
[4] Alpert P., and Coauthors, "The paradoxical increase of Mediterranean extreme daily
rainfall in spite of decrease in total values", Geophys. Res. Lett., Vol. 29, No. 11,
(2002), 10.1029/2001GL013554.
[5] Kitoh A., Yukimoto S., Noda A. and Motoi T., "Simulated changes in the Asian
summer monsoon at times of increased atmospheric CO2", J. Meteor. Soc. Japan,
Vol. 75, No. 6, (1997), pp 1019-1031.