x y w» w œ» w w. w» j w» (dynamical downscaling) m w» (statistical downscaling) ù. w»»z x(rcms, Regional Climate Models) ù LAMs(Limied Area Models), G

Transcription

1 ª Œª Œ 29ƒ 5B Á œ pp. 441 ~ 452 ª x w GCM» w Downscaling Technique of Monthly GCM Using Daily Precipitation Generator Á»Á½x Kyoung, Min SooÁLee, Jung KiÁKim Hung Soo Abstract This paper describes the evaluation technique for climate change effect on daily precipitation frequency using daily precipitation generator that can use outputs of the climate model offered by IPCC DDC. Seoul station of KMA was selected as a study site. This study developed daily precipitation generation model based on two-state markov chain model which have transition probability, scale parameter, and shape parameter of Gamma-2 distribution. Each parameters were estimated from regression analysis between mentioned parameters and monthly total precipitation. Then the regression equations were applied for computing 4 parameters equal to monthly total precipitation downscaled by K-NN to generate daily precipitation considering climate change. A2 scenario of the BCM2 model was projected based on 20c3m(20th Century climate) scenario and difference of daily rainfall frequency was added to the observed rainfall frequency. Gumbel distribution function was used as a probability density function and parameters were estimated using probability weighted moments method for frequency analysis. As a result, there is a small decrease in 2020s and rainfall frequencies of 2050s, 2080s are little bit increased. Keywords : climate change, downscaling, K-NN, daily precipitation generator, frequency analysis IPCC DDC mw œ»z x k w x wš, w»z yƒ e wsƒ»» w wš w. w x 2 state g x», j v w y Gamma-2 s ³ x z w mw. z»z x l K-NN w w»z yƒ š g.»z x BCM2 x w, 20c3m ù» ù w A2 ù w d w. w w sx Gumbel s w, w y ƒ p w. 2020s, 2050s, 2080s ƒw. w :»z y,», K-NN, x, w 1. ùy w»z y y ù ƒ w ƒ ƒ j w,, k, w,,, w eš. p ù y ƒ w vw l j» w ƒ»z y w wš, w ewš»z y w ƒ y w w.»z y ƒ e w sƒw» w w ƒ w w.»z yƒ e wsƒ w GCM ù RCM»z x š. ù w GCM, RCM œ, wù j»ƒ 300~500 km w.»z z Á Á w w z» lœw ( gigatg@inha.ac.kr) z Á w w z» lœw ( jungki5425@hanmail.net) z Á w w z» lœw Áœw ( sookim@inha.ac.kr) 29ƒ 5B œ 441

2 x y w» w œ» w w. w» j w» (dynamical downscaling) m w» (statistical downscaling) ù. w»»z x(rcms, Regional Climate Models) ù LAMs(Limied Area Models), GCMs(General Circulation Models) w y šw»z ù š w. w», x ù w w ƒ wù w.»z y RCMs w»z yƒ w e w w ƒ Flower et al.(2005a), Ekstrom et al.(2005), Frei et al.(2006) w. ù w RCMs, 30 ƒ control run , projection run w w w transient RCMs š (Flower et al. 2007a). m w» t z x, weather typing,»»(weather generator). z x, Giorgi and Hewitson(2001) w w (Transition function) w., (Predictors) d (Predictand) yw w, w y w w w. ƒ»»» t ù w Hanssen-Bauer and Førland(1998), Hellström et al.(2001) w, œ (ANN, artificial neural network)» w ù (CCA, canonical correlation analysis) w w. Weather typing t p w» x w»» s w w» Goodess and Palutikof (1998), Conway et al.(1996), Flower et al.(2000, 2005b), Bardossy et al.(2002, 2005)., ƒ w,»» w y de w w, x w (EOF, empirical othogonal function)ù r,» w.»» w» ƒ ƒ p w w ƒ y y š m w» wù, w»» j y y w w j ƒ w w»k,, w ù.»z yƒ š» j» w x w»z yƒ e w sƒw» w wš w. w IPCC (Intergovernmental Panel on Climate Change) DDC(Data Distribution Centre) œw»z x Fig. 1 Spatial and temporal downscaling scheme BCM2 x (2009) w KNN w w z, w x w w ww. w œ» w BCM2 x xy 2.1 BCM2 x BCCR(Bjerknes Centre for Climate Research) BCM2(Bergen Climate Model Version) x» x ARPEGE w x MICOM ww»/w w x o 90, o sww x IPCC DDC mw GRIB, netcdfxk œ. IPCC»z y x œw» w DDC wš, IPCC 4 š wì AR4 ù» x sww 2, 3 š ù PCMDI (Program for Climate Model Diagnosis and Intercomparison) l œwš. AR4 ù A1, A2, B1 ù SRES(Special Report on Emissions Scenarios) commit, PIcntrl, 20c3m, 1%2, 1%4 ù ƒ sw ( ù w html#pictl). 20c3m» ù w A2 ù š w 2010~2039(2020s), 2040~2069 (2050s), 2070~2099(2080s) y dw. 20c3m, A2 ù sww AR4 ù» w 2, BCM2 x w Fig. 2 Experiment period of BCM2 442 ª Œª Œ

3 Fig. 3 Anormaly in o C from mean of from 20c3m and A2 Table 1. IPCC DDC AR4 data Symbol variable Unit huss surface specific humidity kg/kg orog surface altitude m pr convective precipitation flux kg/m 2 /sec psl sea level pressure Pa rsds surface downwelling shortwave radiation W/m 2 sftlf land area fraction % tas surface monthly average Tmean K tasmax surface monthly average Tmax K tasmin surface monthly average Tmin K uas vas zonal surface wind speed (eastward wind speed) meridional surface wind speed (northward wind speed) m/s m/s Fig. 4 Coordination of NCEP grid around the Korea y (20c3m and A2) 3. AR4 ù t 1» œw. 2.2 NCEP w» w (NOAA ESRL-PSD, Physical Sciences Division) ú»z l ww»»z, ú, Á» w w wš.»,, «y»z ú w, ú j w ÿ,»z x d wš. w»»z y ql wš wš,»z y w w w v w»z» š. w wùƒ d l š w. PSD yr mw NCEP reanalysis data sww 29 d l œwš.» w NCEP/NCAR Reanalysis NOAA National Center for Environment Prediction l œ PSD wš.»z x d w»z x» d. ù»z x, w w w d d»ƒ ƒ w. œ NCEP/NCAR Reanalysis ƒ v w., NCEP/NCAR Reanalysis d w»z x» w» z x d ƒ ƒ w. w NCEP/NCAR Reanalysis,»z x w» w» de œw. x ¾ w» š, w» w»z x mw» NCEP/NCAR Reanalysis mw». 4 w NCEP Reanalysis w. 2.3»» 77 d 1961 l x ¾» œwš ( d d w), v w w., 1 d wš 3 d œwš.» mw œ» 9 ( s³»,»,», 29ƒ 5B œ 443

4 , s³t,,,, ú ). w» x mw» d 1961 d w x š, p ew š.» d d m p t 3 w. 3. K-NN w œ» 3.1 K-NN w» (2009) w K-NN w BCM2 x w., x w e,, t, ƒ w e š ƒ w,,, t, ƒ ƒ ƒ w š ƒ w w. K-NN w» w» w, d NCEP œw» wš w» w w w. w NCEP BCM2 x 5 w. w p wš. ù»z x BCM2, 20c3m, A2 ù ƒ (Precipitable water) sƒ Fig. 5 Atmospheric variable of NCEP and 20c3m and A2 in BCM2 444 ª Œª Œ

5 Seoul Station Monthly total prec.(mm) Table 2. Correlation coefficient of NCEP and Seoul station data Tas ( o C) Tmax ( o C) Tmin ( o C) NCEP huss (kg/kg) uas (m/s) vas (m/s) Prec. water (kg/m 2 /sec) õncep data (monthly average Tmean, monthly average Tmax, monthly average Tmin, specific humidity, eastward wind speed, northward wind speed, precipitable water) Fig. 6 Calibration Result Fig. 7 Validation results. w» mw» w NCEP s³,, š,, uas, vas w (t 2)., NCEP» ü ƒ. ù» w» w NCEP»z x p š w w w w. w,, s³ w, ƒ û 11 ƒ w ƒ w ƒ w w. w ƒ w sw j, A2 ù 20c3m, ƒ w w ƒ w p xw w y w.,, uas, vas K-NN w» w» w 20c3m A2 ù w w BCM2 x l w. 3.2 K-NN w» K-NN w BCM2 x œw A2 ù 20c3m ù ƒƒ w. w ƒƒ». : 1979~2008 NCEP : 1979~ c3m ù : 1950~1999 A2 ù : 2000~2099 K-NN w Nearest Neighbor 9 6, 7. K-NN, Nearest Neighbor š w s³w w». w ù. ù ƒ ƒ š w y w. w r w» w 29ƒ 5B œ 445

6 Fig. 8 Concept of quantile mapping(½, 2008) Quantile Mapping» y w. 3.3 Quantile mapping»z y x l d r. w r w» w ƒ Panofsy and Brire(1963) w Quantile mapping. Quantile mapping w (Wood et al., 2004; Hashino et al., 2007),»z y w Durman et al.(2001), Palmer et al.(2004), Fowler et al.(2007b) GCM w. ü ½ (2008), (2009) r w» w. w Quantile mapping 8. Quantile mapping w K-NN w 20c3m(1961~1999) ù» w d de»z x w w r w z, w BCM2 x 20c3m ù A2 ù l r CASE w w. Fig. 9 Box plot(obs. vs Case1) CASE 1(1980s) : 1970~1999(20c3m, Reference period) CASE 2(2020s) : 2010~2039(A2, Projection period) CASE 3(2050s) : 2040~2069(A2, Projection period) CASE 4(2080s) : 2070~2099(A2, Projection period) CASE m p s (t 3. 9~12). Fig. 10 Box plot(case1 vs Case2) BCM2 l ù ƒ de m e w ƒ š y w,, 2020s, 2050s 150 mm ƒwš Table 3. Statistical characteristic of observed and quantile mapped data (mm) Observation CASE1 CASE2 CASE3 CASE4 Annual precipitation 1, , , , ,701.1 Monthly precipitation Variance 25,135 26,878 27,711 30,673 28,925 Maximum Minimum Skewness ª Œª Œ

7 Fig. 11 Box plot(case1 vs Case3) Fig. 12 Box plot(case1 vs Case4) 2080s 300 mm ƒw y w. w 7 wš 8 ƒ ƒw y w. 4. x»» x d m w p w w j. w»» j y y w w j ƒ w w»k,, w ù. w Richardson(1981) Richardson-type š,,, y y w š.»z y w Wilks (1992). ù w»» ¼»»» j w ƒ š» w w w» w š Markov chain(mason, 2004; Dubrovsky et al. 2004) w» w. NSRPM(Neyman Scott rectangular pulses model) Watts et al.(2004) w» w j ww x Kilsby et al.(2007),» Markov chain x w ù w x ù. x»z y w ƒ y w š x Racsko et al.(1991) l»» x LAR-WG w. Semenov et al.(1998) Richardson-type»» LAR-WGƒ s³ ù w y w. ù ¾ x s³ ù y xw w ƒ š.» w x w» d p» xw š w. x d w x w z, BCM2 x l w» z yƒ e w sƒw wš w. 4.1 x» w» w w w w j» w x v w. w x» x 2-state markov chain. j 2 w w ù. w y». M M M P ij = Pr X k + 1 = j X k = i ( ) M» P ij w w y ùkü 2 2xk p xk, w w y w» w y ùkü. Markov chain x j, 2 Gamma-2 s w. Gamma-2 s w g. w Gamma-2 s w «xw ½ (2009) markov chain x w» w. f x ( ) 1 = x β 1 α β Γ( β) e ( x α)», x, α, β 0 j, α ³ (Scale Parameter) x (Shape parameter) ùkü. w v w α, β, NR (Transition probability of Rain after No rain), RR(Transition probability of Rain after Rain) mw, w d l z mw ³ w. ù z w α, β, NR, RR w x w, ƒ w ƒ w. w w w» w ƒƒ w w» yw (1) (2) 29ƒ 5B œ 447

8 Fig. 13 Regression analysis between monthly total precipitation vs parameters ƒƒ w w 4 w w z w w w. ƒ ƒ z 13. mw x 13 z k»z x l w w w z, w w w j. 4.2 x x w w w z, w» w w d w w. w x, j. w s³ œwš j 500 w z, 500 w w s³ ü w. s³ d s³ w. 14 x d ù ƒ. w ƒ ù» d ƒ ƒ š w Fig. 15 Monthly total precipitation of Obs. and Model Fig. 14 Averaged monthly total precipitation of Obs. and Model. s³ d w 15 mw y ƒ w. w z w ƒƒ y w ƒ x w y w w d x y kœ (State space) w d ƒ ƒ š w x x w y w ( 16). 448 ª Œª Œ

9 Fig. 16 State space between transition probabilities of Obs. and Model x y w w ww mw» w d w w ww x w w ww ù 2z w. w w s x Gumbel sx w š y ƒ p w w. Part I : 1961~1990( 30 ) Part II : 1978~2007( 30 ) Fig. 18 Frequency analysis results(part II) Table 5. Modeling error(part II) Fig. 17 Frequency analysis results(part I) Table 4. Modeling error(part I) yr Obs1 Model1 (%) yr Obs2 Model2 (%) » 5%. w d w w w w w w j ƒ.» x w ƒ w ƒw.»z y w w sƒw» w» z x l d w 29ƒ 5B œ 449

10 Perturbation» wš. mw x»z y w» w x w» (Part I) x» (Part II)» (Part I) d w x» (Part II) w w d w x w v w. t 6 w w. x»z y w w m ww 2-3% yw, ƒ ƒw w» z yƒ ü w e w sƒw» w w x j ƒ ƒw. 4.3»z y š w w»z yƒ e w sƒw» w x w K-NN w BCM2 x l w w ƒƒ Case 500z w. w l e 500set w w ww z, s³ w. ƒƒ Case (Case2, Case3, Case4)» ù (Case1) w w x d w 2020s, 2050s, 2080s w. ƒ (2020s) ƒ (2050s) Fig. 21 Comparison of climate scenarios (CASE1,3) Fig. 22 Comparison between present and climate change (case3) ƒ (2080s) Fig. 19 Comparison of climate scenarios (CASE1,2) Fig. 23 Comparison of climate scenarios (CASE1,4) Fig. 20 Comparison between present and climate change (case2) Fig. 24 Comparison between present and climate change (case4) 450 ª Œª Œ

11 Freq. Model (past) (1) Table 6. Applicability of daily precipitation generator to climate change study Model (present) (2) Model P (3)=(2)-(1) Observation (past)(4) Estimated present (5)=(4)+(3) Observation (present) (6) Error(%) ((6)-(5))/(6) Table 7. Increasing and Decreasing percentage of rainfall frequency by climate change 2020s 2050s 2080s s ƒ ƒw x w w w. 2050s ƒw, % ƒw y w. 2080s 2050s w ƒw y w. 5.»z yƒ e w sƒw» w»z x BCM2 x» w» w w. w d NCEP BCM2 x K-NN w w z, x w y š w. y ƒ p w w z, GUMBEL s w w ww»z yƒ e w sƒw. mw BCM2 A2 ù 20c3m ù ƒ wš K-NN x y w š»w.»z x w», w»z x» ÿ ƒ ùkú». wz BCM2 x w x š w»z x y yw» w ƒ v w ƒw. m w m» sƒ»z» y w w 2005»» (05-» -D03-01) w. š x,, ½x, ½ (2009)»z yƒ w» e w sƒ; AR4 SRES A2 ù», wm wz, wm wz, 29«2By, pp «xw, ½ (2009) Markov Chain Model w m w Downscaling», w wz, w wz, 42«3y, pp ½, ½,, ½x (2008)»z yƒ w I- D-F e w sƒ. w wz, w wz, 41«4y, pp Bárdossy, A., Bogardi, I., and Matyasovszky, I. (2005) Fuzzy rulebased downscaling of precipitation. Theoretical and Applied Climatology, Vol. 82, No. 1-2, pp Bárdossy, A., Stehlík, J., and Caspary, H.J. (2002) Automated objective classification of daily circulation patterns for precipitation and temperature downscaling based on optimized fuzzy rules. Climate Research, Vol. 23, No. 1, pp Conway, D., Wilby, R.L., and Jones, P.D. (1996) Precipitation and air flow indices over the british isles. Climate Research, Vol. 7, No. 2, pp ƒ 5B œ 451

12 Dubrovsky, M., Svoboda, M.D., Tranka, M., Hayes, M.J., Wilhite, D.A., Zalud, Z., and Hlavinka, P. (2009) Application of relative drought indice in assessing climate-change impacts on drought conditions in Czechia. Theoretical and Applied Climatology, Vol. 96, No. 1-2, pp Durman, C.F., Gregory, J.M., Hassell, D.C., Jones, R.G., and Murphy, J.M. (2001) A comparison of extreme European daily precipitation simulated by a global and a regional climate model for present and future climates. Quarterly Journal of the Royal Meteorological Society, Royal Meteorological Society, Vol. 127, No. 573, pp Ekström, M., Fowler, H.J., Kilsby, C.G., and Jones, P.D. (2005) New estimation of future change in extreme rainfall across the UK using regional climate model integrations. 2. Future estimates and use in impact studies. Journal of hydrology, Vol. 300, No. 1-4, pp Fowler, H.J., Ekström, M., Kilsby, C.G., and Jones, P.D. (2005a) New estimation of future change in extreme rainfall across the UK using regional climate model integrations. 1. Assessment of control integrations. Journal of hydrology, Vol. 300, No. 1-4, pp Fowler, H.J. and Kilsby, C.G, O Connell, P.E. (2000) A stochastic rainfall model for the assessment of regional water resource systems under changed climatic conditions. Hydrology and Earth System Sciences, Vol. 4, No. 2, pp Fowler, H.J., Kilsby, C.G, O Connell, P.E., and Burton, A. (2005b) A weathertype conditioned multi-site stochastic rainfall model for generation of scenarios of climatic variability and change. Journal of Hydrology, Vol. 308, No. 1-4, pp Fowler, H.J., Blenkinsop, S., and Tebaldi, C. (2007a) Linking climate change modelling to impacts studies : recent advances in downscaling techniques for hydrological modeling, International journal of climatology, Vol. 27, No. 12, pp Fowler, H.J., Kilsby, C.G., and Stunell, J. (2007b) Modeling the impacts of projected future climate change on water resources in north-west England. Hydrologic & Earth System Sciences, EGU, Vol. 11, No. 3, pp Frei, C., Schöll, R., Fukutome, S., Schmidli, J., and Vidale P.L. (2006) Future change of precipitation extremes in Europe: an intercomparison of scenarios from regional climate models. Journal of Geophysical Research-Atmospheres, Vol Giorgi, F. and Hewitson, B.C. (2001) Regional climate informationevaluation and projections. In Climate Change 2001, The Scientific Basis. C, Houghton JT, Ding Y, Griggs DJ, Noguer M, vander Linden PJ, Dia X, Maskell K, Johnson CA (eds). Cambridge University Press: Cambridge. Goodess, C.M. and Palutikof, J. (1998) Development of daily rainfall scenarios for southeast Spain using a circulation-type approach to downscaling. International Journal of Climatology, Vol. 18, No. 10, pp Hanssen-Bauer, I. and Førland, E.J. (1998) Long-term trends in precipitation and tempuer ure in the Norwegian Arcong-: can they be explained by changes in atmospherng-circulation patterns?. Climate Research, Vol. 10, No. pp Hashino, T., Bradley, A.A., and Schwartz, S.S. (2007) Evaluation of bias-correction methods for ensemble streamflow volume forecasts. Hydrology and Earth System Science, EGU, Vol. 11, pp Hellström C., Chen D., Achberger C., and Räisänen J. (2001) Comparison of climate change scenarios for Sweden based on statistical and dynamical downscaling of monthly precipitation. Climate Research Vol. 19, No. 1, pp Kilsby, C.G., Jones, P.D., Burton, A., Ford, A.C., Fowler, H.J., Harpham, C., James, P., Smith, A., and Wilby, R.L. (2007) A daily weather generator for use in climate change studies. Environmental Modelling and Software, Vol. 22, No. 12, pp Mason, S.J. (2004) Simulating climate over western north america using stochastic weather generators. Climatic Change, Vol. 62, No. 1-3, pp Palmer, R., Wiley, M., and Kameenui, A. (2004) Will Climate Change Impact Water Supply and Demand In the Puget Sound?, Department of Civil and Environmental Engineering University of Washington, Seattle WA. Panofsy, H.A. and Brire, G.W. (1963) Some application of Statistics to Meteorology, Pennsylvania State University, University Park, Pennsylvania, pp Racsko, P., Szeidl, L., and Semenov, M. (1991) A serial approach to local stochastic weather models. Ecological Modelling, Vol. 57, No. 1-2, pp Richardson, C.W. (1981) Stochastic simulation of daily precipitation, temperature and solar radiation. Water Resources Research, Vol. 17, No. 1, pp Semenov, M.A., Brooks R.J., Barrow, E.M., and Richardson, C.W. (1998) Comparison of the WGEN and LARS-WG stochastic weather generators for diverse climates. Climate Research, Vol. 10, No. 2, pp Watts, M., Goodess, C.M., and Jones, P.D. (2004) The CRU Daily Weather Generator, BETWIXT Technical Briefing Note 1, Version 2, February Wilks, D.S. (1992) Adapting stochastic weather generation algorithms for climate change studies. Climatic Change, Vol. 22, No. 1, pp Wood, A.W., Leung, L.R., Sridhar, V., and Lettenmaier, D.P. (2004) Hydrologic implications of dynamical and statistical approaches to downscaling climate model outputs. Climatic Change, Vol. 62, No. 1-3, pp ( : / : / : ) 452 ª Œª Œ