Docstoc
EXCLUSIVE OFFER FOR DOCSTOC USERS
Try the all-new QuickBooks Online for FREE.  No credit card required.

Impacts of Global Warming on Hea

Document Sample
Impacts of Global Warming on Hea Powered By Docstoc
					IMPACTS OF GLOBAL WARMING ON HEAVY PRECIPITATION FREQUENCY AND FLOOD RISKS H. Higashi1, K. Dairaku2, T. Matsuura2
2

National Institute for Environmental Studies, Japan; National Research Institute for Earth Science and Disaster Prevention, Japan higashi@nies.go.jp

1

INTRODUCTION Many governments and organizations focus on global warming because these may be related to the increase in the extraordinary weather frequency in recent years [1]. In the last 30 years, tropical cyclones have generally developed over the warmer waters of the western tropical Pacific, and sea temperatures increase towards the central and eastern Pacific caused by global warming, bringing with them more cyclone threats. Scientists believe that climate change could make cyclone seasons increasingly unpredictable. Local people, used to current cyclone patterns, may in future be less able to prepare [2]. Climate-related disasters have been remarkably increased in the Asian region [2]. Advances in the understanding of the meteorology and developments of monitoring and forecasting system have enhanced early warning systems, contributing immensely to the reduction of fatalities due to typhoons, cyclones and floods. However, the frequency of the extreme events which cause the water-related disasters is increasing in last decade. Therefore, the number of people who die from hydrometeorological hazards is still considerable. A key parameter in basin water managements is characteristics of local precipitation, and an accurate estimate of the climate-change impacts on the heavy precipitation is needed to prevent flood disasters in the future. Climate changes due to global warming impacts have been investigated mainly using global circulation and regional climate models [1, 3]. Some of their outputs are useful for mean local climate: the annual, seasonal precipitation and typhoon track in monsoon Asia [4-6]. However, the climate models cannot capture the short-term and basin-scale precipitation accurately. This study presents statistical changes in the heavy precipitation caused by global warming and their influences on flood risks in a watershed. We analyzed time series of precipitation in the future predicted by 12 general circulation models (GCMs), which were provided by Intergovernmental Panel of Climate Change Data Distribution Centre (IPCC-DDC), and changes in 200-year quantile in Tokyo (139.76E, 35.69N) were discussed. Those influences on runoff discharge and flood risks in the Tama River basin located near Tokyo were evaluated using the numerical simulations. STUDY AREA AND STATISTICAL DATA The Tama River basin is an urbanized watershed (Fig. 1). The area is 1240 km2, the length of mainstream is 138km. The basic and the estimated high water discharge are 8700 and 6500 m3/s, respectively, which are determined using the 2-day precipitation with a 200-year return period. Changes in 200-year quantile in Tokyo caused by global warming were investigated using the results of 12 GCMs. The 2-day precipitation data in the periods, 1981-2000(2000), 2046-2065 (2050), 2081-2100 (2100), 2181-2200 (2200), 2281-2300(2300) under the condition of SRES A1B and B1 were used. Table 1 shows the list of the GCMs and their data set. The detail description can be found in IPCC (2001) [1].

Rainfall gauge Runoff gauge

Ishih

ara

Tokyo

N 10km Flood area

Fig. 1: Outline of the Tama River basin

The calculated results of the precipitation in 2000 were compared with the observed data in order to investigate model bias. Table 1 also shows the observed and the calculated results of the average annual precipitation, the maximum and the 40th largest 2-day precipitation for the 20 years. Although it is regrettable that for some technical reasons, there is no model whose outputs agree well with the observed data.

Table 1: List of GCMs and their resolution
Resolution lon.× lat.
(degree)

Model

IPCC ID

Precipitation (1981-2000) 2-day Annual 2-day Average Maximum 40th
(mm/year) (mm/2-day) (mm/2-day)

Quantile (1981-2000) 100-year 200-year
(mm/2-day) (mm/2-day)

Observed(Tokyo) Model emsemble cccma_cgcm3_1 cnrm_cm3 csiro_mk3_0 gfdl_cm2_0 giss_aom giss_model_e_r iap_fgoals1_0_g ipsl_cm4 miroc3_2_hires miub_echo_g ncar_pcm1 CGCM3.1(T47) CNRM-CM3 CSIRO-Mk3.0 GFDL-CM2.0 GISS-AOM GISS-ER FGOALS-g1.0 IPSL-CM4 MIROC3.2(hires) ECHO-G PCM 3.8 2.8 1.9 2.5 4.0 5.0 2.8 3.8 1.1 2.8 3.8 2.8 3.7 2.8 1.9 2.0 3.0 4.0 2.8 2.5 1.1 2.8 3.7 2.8

1479 1733 1834 2227 1600 1671 1795 1065 2077 2106 1823 1863 1138 1601

294 151 170 214 124 175 68 66 144 363 143 108 110 123

86 66 71 86 61 66 49 41 77 104 70 63 46 57

376 170 216 266 139 201 89 72 162 344 147 128 130 145

415 184 236 290 149 219 94 76 173 376 157 136 141 157

miroc3_2_medres MIROC3.2(medres)

Heavy precipitation was defined as the precipitation in the partial duration series (PDS) composed of the 40 large 2-day precipitations for the 20 years. Figure 2 shows frequency distribution for the precipitation in PDS. Here, the precipitation in PDS is set to dimensionless using the maximum and the threshold precipitation in each model or observed data. The model ensemble average of the dimensionless precipitation frequency in 2000 agrees with the observed one, and these can be approximate to the exponential distribution well. Moreover, it is clarified that the frequency distribution does not change in 2050, 2100, 2200 and 2300.

5.0

Probability distribution function

4.0

20c3m Observed(Tokyo) Calculated (ensemble average) SRES (ensemble average) A1B B1 2050 2100 2200 2300

3.0 2.0

1.0 0.0 0.0

0.2 0.4 0.6 0.8 1.0 Dimensionless precipitation : (x-x0)/(xmax-x0)

Fig. 2: Frequency distribution of the precipitation in PDS

CHANGES IN HEAVY RAINFALL FREQUENCY CAUSED BY GLOBAL WARMING Frequency Analysis The frequency analysis based on PDS[7] was carried out. The PDS composed of the 40 large 2-day precipitations in each 20 years were made. The annual frequency of the precipitation exceeding thresholds follows Poisson distribution:

P

m
m!

e 

3  m  1, 2 , ,

(1).

Here, P: probability, : average of the annual frequency (= 2.0 time/year), m: annual frequency of the exceedance events. The probability density function for the amount of the 2-day precipitation in PDS (g) follows the exponential distribution:

g

1  x  x0  exp    a a  

(2).

Here, x0: threshold, a: average precipitation in PDS. The return period can be evaluated by Eq. 3 calculated from Eqs. 1 and 2.

f 

Here, f: probability density function, c: constant (= x0 + a ln ). In this method, the quantile can be calculated if even  is decided, since a and x0are easily known. Changes in 200-year Quantile The threshold (x0) and the average precipitation (a) in PDS of 2000, 2050, 2100, 2200, and 2300 under the condition of SRES A1B and B1 were calculated. Figures 3 and 4 show changes in the threshold (x0) and the average precipitation (a) in PDS caused by global warming, respectively. Here, the values of x0 and a in 2000 are set to 1 in each model. The ensemble average ratio of x0 and that of a to the present one are 1.08-1.21 and 1.09-1.20 in the A1B scenario, 1.05-1.10 and 1.03-1.07 in the B1 scenario. In Figs. 3(a) and 4(a), almost all model outputs in the A1B scenario indicate that the values of x0 and a in the future are larger than the ones in the present. There are some models whose outputs represent the slight decrease trend in the x0 and a in the B1 scenario. Figures 5 shows changes in the predicted 200-year quantile caused by global

1  xc  x  c  exp   exp    a a    a 

(3)

warming. The ensemble average ratio of the 200-year quantile to the present one is 1.07-1.20 in the A1B scenario, indicating that the heavy precipitation frequency increases due to the global warming. However, the ratio remains stable at 1.0 in the B1 scenario, one of the causes is that greenhouse gas concentration in the B1 scenario is smaller than that in the A1B scenario.
cccma_cgcm3_1 cnrm_cm3 csiro_mk3_0 gfdl_cm2_0 giss_aom giss_model_e_r iap_fgoals1_0_g ipsl_cm4 miroc3_2_hires miroc3_2_medres miub_echo_g ncar_pcm1
Ratio

1.8 1.6 1.4
Ratio

1.8 1.6 1.4 1.2 1.0

cccma_cgcm3_1 cnrm_cm3 csiro_mk3_0 gfdl_cm2_0 giss_aom giss_model_e_r

iap_fgoals1_0_g ipsl_cm4 miroc3_2_hires miroc3_2_medres miub_echo_g ncar_pcm1

1.2 1.0 0.8 0.6 2000 Ensemble average Ensemble average + standard deviation Range 2100 Year 2200 2300

0.8 0.6 2000

Ensemble average Ensemble average + standard deviation Range 2100 Year 2200 2300

(a) A1B

(b) B1

Fig. 3: Changes in threshold caused by global warming
cccma_cgcm3_1 cnrm_cm3 csiro_mk3_0 gfdl_cm2_0 giss_aom giss_model_e_r iap_fgoals1_0_g ipsl_cm4 miroc3_2_hires miroc3_2_medres miub_echo_g ncar_pcm1
Ratio

1.8 1.6 1.4

1.8 1.6 1.4 1.2 1.0

cccma_cgcm3_1 cnrm_cm3 csiro_mk3_0 gfdl_cm2_0 giss_aom giss_model_e_r

iap_fgoals1_0_g ipsl_cm4 miroc3_2_hires miroc3_2_medres miub_echo_g ncar_pcm1

Ratio

1.2 1.0 0.8 0.6 2000 Ensemble average Ensemble average + standard deviation Range 2100 Year 2200 2300

0.8 0.6 2000

Ensemble average Ensemble average + standard deviation Range 2100 Year 2200 2300

(a) A1B

(b) B1

Fig. 4: Changes in average precipitation in PDS caused by global warming
cccma_cgcm3_1 cnrm_cm3 csiro_mk3_0 gfdl_cm2_0 giss_aom giss_model_e_r iap_fgoals1_0_g ipsl_cm4 miroc3_2_hires miroc3_2_medres miub_echo_g ncar_pcm1

1.8 1.6 1.4

1.8 1.6 1.4

cccma_cgcm3_1 cnrm_cm3 csiro_mk3_0 gfdl_cm2_0 giss_aom giss_model_e_r

iap_fgoals1_0_g ipsl_cm4 miroc3_2_hires miroc3_2_medres miub_echo_g ncar_pcm1

Ratio

Ratio
Ensemble average Ensemble average + standard deviation Range 2100 Year 2200 2300

1.2 1.0 0.8 0.6 2000

1.2 1.0 0.8 0.6 2000 Ensemble average Ensemble average + standard deviation 2100 Year 2200 Range 2300

(a) A1B

(b) B1

Fig. 5: Changes in 200-year quantile caused by global warming

IMPACTS OF GLOBAL WARMING ON FLOOD RISKS Numerical Method To evaluate changes in the estimated high water discharge in Tama River basin in the A1B scenario, rainfall runoff analysis were carried out under the present geophysical condition. In this study, the kinematic runoff model and the unit hydrograph method were adopted in order to calculate the direct discharge and the base flow at Ishihara (see Fig. 1), respectively. In kinematic runoff model [8], topography, land cover conditions, channel networks and storage facilities are all considered. After dividing the basin into sub-basins, each sub-basin is modeled by two slopes and a channel. The flow on a slope and channel flow are approximated by a kinematic wave.

h q   re t x 1 1 5 q  i 2h 3 N

(4) (5)

Here, h: depth of overland flow, i: average slope of sub-basin, q: flow rate per unit width, N: equivalent roughness coefficient, x: distance, t: time, and re: effective rainfall. The effective rainfall (re) was estimated using the cumulated-retained curve in this study. The governing equations of channel flow are expressed as

A Q  q t x 2 A 1 Q  I 2  k1 Az  3 n

(6) (7)

Here, A: discharge area, Q: flow rate, I: slope of channel, n: roughness of Manning's formula, k1 and z: model constants determined by the shape of the channel. These governing equations were solved by using a characteristics curve. The unit hydrograph method [9] is determined by the following matrix method. The relationship between rainfall, base flow and unit hydrograph is given in matrix notation by

RU = Q

(10).

Here, R: rainfall matrix, U: unit hydrograph, and Q: base flow. The flood risk was evaluated using a numerical simulation if the precipitation with a 200-year return period occurs. The downstream area at Ishihara was defined as inundation flow analysis area (see Fig. 1). The flow of the Tama River was analyzed one-dimensionally by applying St. Venant equations, whereas inundation of the flood plain was analyzed two-dimensionally. Flows in the river and the flood plain were combined using a discharge formula for a weir [10]. St. Venant equations are adopted in order to describe one-dimensional flow gradually varied channel.

A Q  q t x 2 v   vv  H gn v v  g  0 t x x R4/3

(11) (12)

Here, v: mean velocity, R: hydraulic radius, and H: water surface elevation. Q at Ishihara was set to the results calculated by the runoff model (Eqs. 4-10) as the upstream boundary-condition; H at Tokyo Bay was set to +3.85m (T.P.) as the downstream boundary-condition. The two dimensional equations for the flood flow are as follow:

h M N   q t x x M   uM    vM  H gn2u u 2  v 2    gh  0 t x y x h1/ 3

(13) (14) (15)

N   uN    vN  H gn2v u 2  v 2    gh  0 t x y y h1/ 3

Here, M, N: x, y-direction flow flux, u and v: x, y-direction mean flow velocity. A leap frog finite difference scheme was applied in order to solve Eqs. 11-15. Changes in Estimated High Water Discharge and Flood Risks The 200-year quantiles in the periods, 2000, 2050, 2100, 2200, and 2300 were set to 457, 523, 519, 491, 548 mm/2-day, considering the ensemble average in Fig. 5(a). Each hyetograph was defined as the following; the observed hourly-precipitation in from 10:00 on August 30 to 10:00 on September 1 in 1949, which is one of the largest 2-day precipitations, was multiplied by a constant so that 2-day precipitation become equal to the 200-year quantile in each period.

Rainfall (mm/h)

0 10 20 30 40 50 8000 6000 4000 2000 0

457mm/2-day(2000)

Flood volume (106 m3)

Runoff (m3/sec)

Ishihara 2000 2050 2100 2200 2300

40 30 20 10 0 0

2000 2050 2100 2200 2300

48 24 Time (hour)

Fig. 6: Changes in hydrograph and flood volume in the A1B scenario

Ishihara

N

Ishihara
N

5km

5km

Flood depth 0 - 1m 1 - 2m 2 - 3m 3 - 4m 4m -

Flood depth 0 - 1m 1 - 2m 2 - 3m 3 - 4m 4m -

(a) 2000 (b) 2300 Fig. 7: Distribution of flood depth

Figure 6 shows the calculated results of the hydrograph and the flood volume in the A1B scenario. The ratios of the estimated high water discharge to the present one are 1.10-1.26, and those of the flood volume are 1.46-2.31. As compared with the increase of precipitation, the flood volume is increasing dramatically. Figure 7 shows the distribution of the maximum flood depth. We conclude that the risk of flood in the basin is projected to be much higher than the present one due to global warming. CONCLUSION We describe impacts of global warming on heavy precipitation characteristics and flood risk using the 2-day precipitation of the 12 GCMs. The frequency analysis based on PDS was carried out. The model ensemble average 200-year quantiles in Tokyo during 2050-2300 under the climate condition of IPCC SRES-A1B scenario were 1.07-1.20 times as large as the one under the present climate condition. Those influences on runoff discharge and flood risk in the Tama river basin were investigated using the numerical simulations. The estimated high water discharge rose by 10-26%, and the flood volume increased by 46-131% if the precipitation with a 200-year return period occurred. The risk of flood in the basin is projected to be much higher than the present one due to global warming. 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, J.T., Ding, Y., Griggs, D.J., Noguer, M., van der Linden, P.J., Dai, X., Maskell, K. and Johnson, C.A. eds., Cambridge University Press, Cambridge, United Kingdom and New York, NY, U.S.A., (2001). [2 ] International Federation of Red Cross and Red Crecent Societies, “World Disaster Report 2002”, (2003). [3 ] The International Ad Hoc Detection and Attribution Group, “Review Article: Detecting and Attributing External Influences on the Climate System: A Review of Recent Advances”, J. Climate, vol. 18, (2005), pp. 1291-1314. [4 ] Ropelewski, C.F., and Halpert, M.S., “Global and regional scale precipitation patterns associated with the El Nino/Southern Oscillation”, Mon. Wea. Rev., vol. 115, (1987), pp. 1606-1626. [5 ] Matsuura, T., Yumoto, M., Iizuka, S., and Kawamura, R., “Typhoon and ENSO simulation using a high-resolution coupled GCM”, Geophys. Res. Letters, vol. 26, (1999), pp.1755-1758. [6 ] Yumoto, M., and Matsuura, T., “Interdecadal variability of tropical cyclone activity in the western North Pacific”, J. Meteor. Soc. Japan, vol. 81, (2003), pp.

1069-1086. [7 ] Maindment, D.J., “Handbook of Hydrology”, McGraw-Hill, (1993). [8 ] Iwagaki, Y., “Fundamental Studies on the Runoff Analysis by Characteristics”, Bulletin of the Disaster Prevention Research Institute, Kyoto University, vol. 5, No. 10, (1955), pp. 1-25. [9 ] Viessman. W., Knapp, J.W., Lewis, G.l., and Harbaugh, T.E., “Introduction to Hydrology”, Harper & Row, Publishers, (1977). [10] Inoue, K., Toda, K., and Maeda, O., “Inundation model in the region of river network system and its application to Mekong delta”, Annual Journal of Hydraulic Engineering, JSCE, vol. 44, (2000), pp. 485-490.