Characteristics of Extreme Events in Korea: Observations and Projections Won-Tae Kwon Hee-Jeong Baek, Hyo-Shin Lee and Yu-Kyung Hyun National Institute.

Characteristics of Extreme Events in Korea: Observations and Projections

Won-Tae Kwon Hee-Jeong Baek, Hyo-Shin Lee and Yu-Kyung Hyun

National Institute of Meteorological Research

Korea Meteorological Administration

11IMSC, 12-16 July 2010, U of Edinburgh

11IMSC, 12-16 July 2010, Edinburgh2

Motivations and Objectives Observed changes (Previous studies)

Temperature: +1.7℃ (including urbanization effect) Extremely cold day (T<-10℃): 30 days (1950s) recently 10

days Changes in seasonal length [onset and duration]

Precipitation: +19% Amount ↑, Wet days ↓, intensity ↑, large temporal variability Heavy rain days (80 mm/day )↑

1.6 day/year → 2.3 day/year (1954-2003, 14 stations)

Focused on the changes of mean values

This study Trend analysis using climate indices Changes in extreme values Future Changes in extreme values

11IMSC, 12-16 July 2010, Edinburgh3

Data and Methodology

Observa-tion

Six StationsDaily T & P1912-2009

Projection

SRES A1BMM5 (~27 km)Daily T & P1971-2100

Trend in cli-mate indices (27)(CCI/WCRP-ETCCDMI)-16 for temp-11 for precip

Extreme Anal-ysis- GEV- Tmax, Tmin,

Precip

Extreme Anal-ysis- GEV- Tmax, Tmin,

Precip

11IMSC, 12-16 July 2010, Edinburgh4

Example of climate indices: annual precipitation

11IMSC, 12-16 July 2010, Edinburgh5

Climate Indices: temperature

GSL sd25 tr25 id0 fd0 CooldaysCoolnights Coldspell DTR

-4

-3

-2

-1

0

1

2

-Growing season length-Summer days-Tropical nights -Ice days

-Frost days-Cool days-Cool nights-Cold spell duration-Diurnal temperature range

℃/decade

UNIT: day/decade

11IMSC, 12-16 July 2010, Edinburgh6

Climate indices: wet days

wet

day

(Pr>

0mm

)

light

rai

n

(Pr<

1mm

)1m

m

10m

m

20m

m

30m

m

50m

m

80m

m

150m

m

-1.2

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4 Unit: days/decade

Negative trend in total wet days but positive trends in heavy rain daysTotal precipitation change is mostly due to the increase of very wet day precipitation.

Tren

d Very wet day (95th): 41.4mm/day Extremely wet day

(99th): ~91.1mm/day

11IMSC, 12-16 July 2010, Edinburgh7

Results: trend/decade

Term Trend/decade

Summer days +0.830***

Ice days -0.998***

Tropical nights +0.671***

Frost days -3.546***Growing season length +3.528***

Max Tmax (Tmin) +0.093*** (+0.201***)

Min Tmax (Tmin) +0.127*** (+0.281***)

Cool(Warm) nights -0.758*** (+0.176)

Cool (Warm) days -0.583*** (-0.074)

Warm (Cold) spell duration indicator -0.295 (-1.845***)

Diurnal temperature range -0.115***

***: significant at 99% level, **: significant at 95% level, *: significant at 90% level

Term Trend/decade

Max 1-day (5-day) precipitation amount

+2.321*:Aug(+4.994**): Aug

Simple daily intensity index +0.228***

Number of 10mm, 20mm, 30mm precipitation days

+0.320*+0.268*+0.246**

Consecutive dry (wet) days

-0032(-0.011)

Extremely (Very) wet day precipitation

+9.004**(+17.339)***

Annual total wet-day precipitation 20.478**

Summary of Climate Change in Korea

8

Temperature +

Summer day +

T_min +T_max +

Ice day -

Ice day - Frost day -

Winter season - Summer +

Rain Rate +

Lightning +

Relative Humidity -

Visibility +

Cloudiness -

Early Spring Blossom

Less Heating Energy

Late Autumn Leaves

More Cooling Energy

Tropical night +

Sea Surface Temp +

Sea Level +

★ significant level is different at different locations highly significant (α = 0.01) significant (α = 0.05) trend but not significant (α > 0.05)

Precipitation +★

Rain Day - ★

Fog - ★

Heavy Rain +★

Wind speed -★

11IMSC, 12-16 July 2010, Edinburgh

GSL +DTR -

Wet day precip +

11IMSC, 12-16 July 2010, Edinburgh9

GEV Analysis: Methods Data windows

Sliding window (SW) fixed sample size (30 years)

Cumulative window (CW)

increase in sample size

Parameter estimations Maximum Likelihood method (ML) L-Moment method (LM)

11IMSC, 12-16 July 2010, Edinburgh10

Annual maximum daily precipitationGangneung Seoul Incheon

Daegu Busan Mokpo

191019201930194019501960197019801990200020100

50

100

150

200

250

300

Average of annual maxima precip-itation

11IMSC, 12-16 July 2010, Edinburgh11

Annual daily maximum precipitation

Red: MLBlack: LM

Location Scale Shape Return value: 100yr

30-year sliding window

Cumulative window

11IMSC, 12-16 July 2010, Edinburgh12

Annual maximum of Tmax

Red: MLBlack: LM

Location Scale Shape Return value: 100yr

30-year sliding window

Cumulative window

11IMSC, 12-16 July 2010, Edinburgh13

Cumulative window

30-year sliding window

Annual minimum of Tmin

Red: MLBlack: LM

Location Scale Shape Return value: 100yr

Regional Climate Change Projection

14

Emission Sce-nario

(population, energy, socio-economics)

Anthropogenic Radiative Forc-

ing

(carbon cycle, bio-geochemical cycle)

Regional Cli-mate Scenario

(Regional Cli-mate Model)

Impact As-sessment and

Adaptation

High-resolution Global Scenario

(Time-slice Exp)

Global Climate Scenario

(Global Climate Model)

· Based on A1B Scenario· Model: MM5· Spatial Resolution~27km· Period: 1971~2100· Data size: 5TB (52 vari-

ables)

11IMSC, 12-16 July 2010, Edinburgh15

20yr-return value 20 yr-return value for precipitation (OBS)

present (1979-2007) future (2077-2099) difference

present (1979-2007) future (2077-2099) difference Min Temperature

present (1979-2007) future (2077-2099) difference Max Temperature

11IMSC, 12-16 July 2010, Edinburgh16

Summary Observation

Trends in climate indices show major characteristics of regional climate change.

Recent increase of extreme temperature/precipitation events have been recognized well in GEV analysis. Precipitation and most of min/max Temperature

increasing of mean and variance in extreme higher probability of occurrence of extreme events

Annual maximum of Tmax Decrease of mean and return values but increase in variances

L-moment method show stable performance for parameter estimation

Future Projection (SRES A1B) At the end of 21st century: -20-50% of precipitation and 2.5-4.5

degrees of temperature increase over the Korean Peninsula Precipitation: extreme events are increasing but with large regional variation Temperature: the increasing trend in daily minimum temperature is highest

Plan for further analysis when CORDEX data become available.

Thank you!