The 2nd National Workshop on Economics of Climate Change and

Low Carbon Growth Strategies in Northeast Asia

Grand Ambassador Hotel, Seoul, Korea, 15 October 2010

Projection of Hydrologic
Impacts of Climate Change on
the Korean Watersheds
- Preliminary assessment -

Huicheul JUNG
(hchjung@kei.re.kr)
Korea Adaptation Center for Climate Change,
Korea Environment Institute

1
 Impacts of climate change on
global freshwater resources (IPCC AR4, WG-II, SPM, 2007)
Warming temperature:
 Snowmelt runoff will be changed in timing and the peak volume
 Water supplies from glaciers and snow cover are projected to decline
 Water quality and aquatic ecosystem could be deteriorated due to increasing
water temperature
Intensifying hydrological cycle:
 10-40% increase of available water resources (AWR) at high latitude and in
some wet tropic regions; 10-30% decrease over dry regions at mid-latitude and
in the dry tropics
 Drought-affected areas will likely increase in extent
 Heavy rainfall events (≈flood risk) are likely to increase in frequency and
magnitude
Rising sea level:
 Coastal mega cities are vulnerable and water supplies from ground water will be
restricted due to sea water intrusion near coastal region

2
 Impacts of climate change on
Korean water resources (June 2010 by MLTM)
Increases in water scarcity (KMA RCM A2 scenario):
 14% increase of mean annual precipitation and +3.64 ˚C increase of mean
annual temperature (2061-2090 compared with1971-2000, 12.5 ˚C/1,230 mm)
 Changes in seasonal water balance (runoff peak: from JJA to JAS, agricultural
water demand peak: JJA)
 Water deficit due to increasing evapotranspiration (ET) : 33 x 109 ton / yr in
2060s (over the storage of Soyangang Dam, 29 x 109 ton )
 Runoff decrease in Nackdong (-2.4%), Geum (-13.3%), Youngsan (-10.8%) in 2060s

Increases in extreme events:

 Increase in frequency of heavy rainfall
 2.7 times (100 mm/d)
 current 100yr return period rainfall → 58 yr
 Increase in frequency of drought years
 3.4 times, 57% decrease of runoff

3
 Objectives

Assessing potential hydrological impacts of climate change

on Korean watersheds
 To identify the spatial distributions of hydrological hot-

spots (future runoff, flood and drought events)

 To evaluate the future impacts estimated by MLTM

 To evaluate ecosystem contribution to future runoff changes

 About 70% of the Korean peninsula is occupied by forests; About 74% of

annual precipitation (= 1,276 x 108 ton / yr, 1971 to 2000) in forested land
 Forests play the key roles in controlling the quantity and quality of
downstream watersheds
 Future effects of changes in climate, CO2 conc., and tree species on annual
water balance Bilt-up, 2.2
Deciduous
Conifer, 34.4 Mixed, 21.6 Cropland, 20.3
, 17.5
Water, 3.1

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 4
 Outline of impact assessment
Define and develop geographic and ecological scope of Korean
ecosystem and watersheds

Development and evaluation of Stochastic weather

Atmospheric CO2
climate change scenarios (5km generation with monthly
concentration
resolution) climate

Threats to habitat and Threats to hydrological

Land management biodiversity function Model calibration
and conservation and
strategy parameterization
Future Land and
A GIS-based Hydrological
Vegetation Cover
Model
Database

Changes in biodiversity and Changes in hydrological

conservation status Define Ecological and responses in watersheds
• Changes in potential vegetation Hydrological Hot-
Hot-spots • Changes in water availability 5
• Extents, continuity and fragmentation for implementing • Changes in flood events and intensity
of remaining forest cover adaptation strategies • Changes in drought events and
• Changes in land use allocation intensity
5
 Defining geographic scope: Delineating
Watersheds and Hydro-Net
The 3 arc second (about 90m) Digital Elevation Models (DEMs) from the
NASA shuttle radar topographic mission (SRTM, http://srtm.csi.cgiar.org/)
Assigned by the Pfafstetter watershed coding system (Pfafstetter, 1989) to
identify the upstream and downstream watersheds (ROK-820, DPK-737).
Validated with 174 river gauging stations (r2 = 0.9994)
1600 unit basins from DEM Pfafstetter encoding
30000

25000

GIS Area (km2)

20000

15000

10000
R2 = 0.9994, (n=174)
5000

0
0 10000 20000 30000
Reported Area ( km2)
6
 Hydrological model

A deterministic, process-oriented, distributed parameter hydrologic model

based on 1D-SVAT BROOK90 model + Muskingum-Cunge routing model
Day-night simulation generates daily water balance and streamflow in each
unit basin
 Shuttleworth-Wallace (SW) model and Javis equation is used to
consider the effect of the rising CO2 concentration on potential ET (PE)
10 1400
Pan (Jeonju) Pan conifer forest
9 deciduous forest mixed forest
PE (Coniferous forest) 1200

.
Simulated PE and Pan evaporation (mm/d) .

10 day average (Pan) cropland grassland

8

20-year mean annual PE (mm/year)

10 day average (SW) 1000
7

6 800

5
600
4

3 400

2
200
1

0
0
89/1 89/5 89/9 90/1 90/5 90/9 91/1
Jeonju (SJ) Chuncheon (SY) 7
 Baseline climate scenario (1971-2000):

5km spatial and 1day temporal resolution during 1971-2100

 Spatial interpolation using 111 GTS-weather stations (ROK-84, DPK-
27) and IDSW method for temperature and spline method for
precipitation, wind, humidity and solar radiation.
 Temporal interpolation using observed stochastic characteristic and
WXGEN model
Annual mean Annual mean
temperature (oC) precipitation (mm/year)
Generated daily precipitation at Mt. Junbong flux tower

8
 Indicator for hydrological impacts

Future impacts: relative changes in frequency and magnitude of

runoff, flood and drought risk flow
 Flood risk flow is defined as the daily Q5 flow of the BASE period (1981-
2000);
 Drought risk flow is the monthly Q95 flow of the BASE period; and
 Average water availability is the monthly Q50 flow (median) of the
BASE period

 Q5 flow (high flow) is the flow

exceeded 5% of the time (or the 95th
percentile of a probability density
function), and
Q95 flow (low flow) is the flow
exceeded 95% of the time

9
 Model validation (1): Daily inputs

Model efficiency and water balance error in the selected eight

forested watersheds using daily time step
 Annual mean bias error: -7~10%
 Monthly flow r2 : 0.799~0.955
 Daily flow efficiency : 0.513~0.865
800 800 600 600
A.Soyanggangdam B.Hwacheondam C.Chungjudam D.Namgangdam

Simulated monthly runoff (mm)

Simulated monthly runoff (mm)

Simulated monthly runoff (mm)

Simulated monthly runoff (mm)

calibration calibration 500 calibration 500 calibration
600 validation 600 validation validation validation
400 400

400 400 300 300

200 200
200 200
2 2
r -cal. = 0.953 r -cal. = 0.905 100 r2-cal. = 0.852 100 r2-cal. = 0.962
r2-val. = 0.955 r2-val. = 0.807 r2-val. = 0.935 2
r -val. = 0.912
0 0 0 0
0 200 400 600 800 0 200 400 600 800 0 100 200 300 400 500 600 0 100 200 300 400 500 600
Observed monthly runoff (mm) Observed monthly runoff (mm) Observed monthly runoff (mm) Observed monthly runoff (mm)

500 500 600 500

E.Hapcheondam F.Andongdam G.Deacheongdam H.Seomjingangdam
Simulated monthly runoff (mm)

Simulated monthly runoff (mm)

Simulated monthly runoff (mm)

Simulated monthly runoff (mm)

calibration calibration 500 calibration calibration
400 400 400
validation validation validation validation
400
300 300 300
300
200 200 200
200

100 2
r -cal. = 0.953 100 2 100
r -cal. = 0.964 100 r2-cal. = 0.965 2
r -cal. = 0.919
2 2
r -val. = 0.799 r -val. = 0.949 r2-val. = 0.928
0 0 0 0
0 100 200 300 400 500 0 100 200 300 400 500 0 100 200 300 400 500 600 0 100 200 300 400 500
Observed monthly runoff (mm) Observed monthly runoff (mm) Observed monthly runoff (mm) Observed monthly runoff (mm)
10
 Model validation (2): Generated daily inputs

Monthly flow and model efficiency using generated daily climate

 Monthly mean bias error: -8~0.5 % (19 stations)
 Monthly flow r2 : 0.9989
 Daily flow efficiency : 0.593~0.871
10
Long-term average monthly flow
Simulated discharge (km /month)

1
Monthly mean flow
3

Bias = -0.03 km3/month

2 Monthly Q5 flow
R = 0.9989
Bias = -2.15 km3/month
0.1 R2 = 0.9328

Monthly mean flow

0.01
Monthly Q95 flow
Monthly Q95 flow Monthly Q5 flow
3
Bias = 0.14 km /month
2 Regression line
R = 0.9514
95% confi, line
0.001
0.001 0.01 0.1 1 10

Observed discharge (km3/month) 11

 Model validation (3): Annual runoff estimation
Area (km3)
Precipitation, P (mm) (km3)
Runoff, RO (mm) RO / P Jun.-Sep. fraction
(1000km2) Jun.-Sep. Oct.-May Annual Jun.-Sep. Oct.-May Annual (-) P RO
ROK 99.4 85.7 44.6 130.3 46.9 23.3 70.2 0.539 65.8% 66.8%
DPK 122.5 75.9 31.8 107.8 38.3 17.9 56.3 0.522 70.4% 68.1%
1600 0.8
Annual mean 30-year mean annual precipitation (P) and runoff (RO)
runoff 1400
0.7
(mm/year)
1200

0.6
1000

(mm/year)

RO/P
800 0.5

600
0.4

400
Reported RO('69-98)
0.3 Simulated RO('71-00)
200 Reported P('69-98)
GIS-based P('71-00)
0 0.2 Simulated runoff rate
n g m in n n n g in n st st st )
Ha kdon Geu omj ngsa cheo cheo yeon ngj ngsa coa coa coa land
e o
S Ye ong gyo ang o o t t h i n
Na D ye Eas es t
W S o u io n a
l(
se Sap M H
An N a t

Comparing with the National Report, Water Vision 2020

(2000)
30yr(1971-2000) mean annual runoff difference: -4.3%
12
 Future climate change scenarios
2040s (2031-2050) 2090s (2081-2100)
o o
ΔT ( C) ΔP (%) ΔT ( C) ΔP (%)
Jun.-Sep. Oct.-May Annual Jun.-Sep. Oct.-May Annual Jun.-Sep. Oct.-May Annual Jun.-Sep. Oct.-May Annual
ROK 1.6 2.6 2.3 42.1 4.0 24.4 2.0 3.1 2.7 16.9 -5.1 6.7
DPK 1.5 3.0 2.5 37.0 20.4 30.9 2.3 3.5 3.1 9.4 6.7 8.4

Japan MRI-RCM SRES A2 and MIROC high res. A1B GCM

14% increase of mean annual precipitation and +3.64 ˚C increase of mean annual
temperature (2061-2090 compared with1971-2000, 12.5 ˚C/1,230 mm)
60
ROK (2081-2100) DPK (2081-2100)
ROK (2031-2050) DPK (2031-2050)
a24

MIROC-high res.
MRI-RCM a24 a11 A1-CSIRO(TAR)
Precipitation change (%)

40 a1b A1B-MICRO(AR4)
a1b a21 A2-CCCMA(TAR)
a22 A2-CSIRO(TAR)
a1b r40 a23 A2-ECHAM(TAR)
a24 A2-HADCM(TAR)
a25 A2-NCARPCM(TAR)
r40
b24 a26 A2-CCCMA(AR4)
20 b24 a27 A2-CSIRO(AR4)
a28 A2-ECHAM(AR4)
b25 a29 A2-GFDL(AR4)
b23 b1 B1-CSIRO(TAR)
a23
r90 b21 B2-CCCMA(TAR)
r90b23
b21 a29 a29
b22 B2-CSIRO(TAR)
a25 GCMs from TAR and AR4
a21
b25 b21
a27
a28b22
a26
b23 B2-ECHAM(TAR)
0 a28b22
b1 a26 a23 b24 B2-HADCM(TAR)
a27 a11 b1 a11 b25 B2-NCARP CM(TAR)
a22 a25 a21 r40 A2-MRI(RCM)2040s
a22 r90 A2-MRI(RCM)2090s

1 2 3 4 5 1 2 3 4 5
o
Temperatuer increase ( C)
13
 Changes in runoff

Base scenario (1971-2000):

 Annual mean precipitation: 1058 mm (ROK: 1280 mm, DPK:
877 mm)
 Annual runoff: 570 mm (54% of Precipitation) (ROK: 706 mm
(55%), DPK: 459 mm (52%))
 Wet season (June-September) runoff: 68% of annual runoff
Changes in annual runoff (2081-2100):
 MRI-RCM A2: ROK +9.4%, DPK: +8.4%
 MIROC A1B: ROK +32.1.%, DPK: +41%
Changes in wet season (June-September) runoff:
 MRI-RCM A2: ROK 23.7%, DPK: 17.7%
 MIROC A1B: ROK 35.2%, DPK: 44.9%

14
 Changes in Mean Annual Runoff by 2040s /2090s
40 40%
37.5%
35.1% A) 20-year mean (2031-2050)
Oct.-May.
30 30%

Change in runoff (km /year) .

Jun.-Sep.

Percent change in runoff .

%change
20 20%

3
12.7%
10 11.0% 10.5% 10%
6.5% 5.8%
0 0%
-3.2%
-7.7%
-10 -10%
-15.9% -16.5% -16.6%
-20 -20%
ROK DPK ROK DPK ROK DPK ROK DPK ROK DPK ROK DPK

MRIRCM-A2 Micro-A1B csiro-A1 csiro-A2 csiro-B1 csiro-B2

46.6%
Change in runoff (km3/year)

Oct.-May.

% Change in annual runoff .

35.3% 2081-2100 40%
40
Jun.-Sep.
30%
30 %change
20%
20 9.4% 8.4%
6.8% 4.3% 10%
10 -2.5% -1.9%
0%
0 -12.0%
-17.9%-16.9% -10%
-10 -20%
-28.2%
-20 -30%
ROK DPK ROK DPK ROK DPK ROK DPK ROK DPK ROK DPK

MRIRCM- MIROC-A1B CSIRO_A1 CSIRO-A2 CSIRO-B1 CSIRO-B2 15

 Climate Change Impacts on Water Availability: Changes in Mean Annual and
Seasonal Runoff by MRIRCM A2 scenario
Provinces in S.Korea Provinces in N.Korea
400 A.ROK-Gyeonggi Absolute change Percent change 400 B.DPK-Pyeongan

300 2090s (mm/year) 2090s (%) 300

Runoff (mm)
Runoff (mm)

200 200

Present Present
2040s 2040s
100 100
2090s 2090s

0 0
J F M A M J J A S O N D J F M A M J J A S O N D
400 C.ROK-Gangwon 400 D.DPK-Gangwon

300 300
Runoff (mm)

Runoff (mm)
200 200
Present Present
2040s 2040s
100 2090s 100 2090s

0 0
J F M A M J J A S O N D J F M A M J J A S O N D
400 E. ROK-Chungcheong 400 F.DPK-Hwanghae

300 300

Runoff (mm)
Runoff (mm)

200 200
Present Present
2040s 2040s
100 2090s 100 2090s

0 0
J F M A M J J A S O N D J F M A M J J A S O N D
400 G.ROK-Jeolla 400 H.DPK-Hamgyeong

300 DJF MAM JJA SON 300

Runoff (mm)

Runoff (mm)
200 200
Present Present
2040s 2040s
100 2090s 100
2090s

0 0
J F M A M J J A S O N D J F M A M J J A S O N D

400 I.ROK-Gyeongsang 400 J.DPK-Yanggange/Jagang

300 300
Runoff (mm)

Runoff (mm)
200 200
Present Present
2040s 2040s
100 2090s 100 2090s

0 0
J F M A M J J A S O N D J F M A M J J A S O N D

16
 Climate Change Impacts on Water Availability: Changes in Mean Annual and
Seasonal Runoff by MIROC A1B scenario
Provinces in S.Korea Provinces in N.Korea
300 A.ROK-Gyeonggi Absolute change Percent change 400 B.DPK-Pyeongan

250
2090s (mm/year) 2090s (%) 300
200

Runoff (mm)
Runoff (mm)

150 200

Present Present
100
2040s 2040s
100
2090s 2090s
50

0 0
J F M A M J J A S O N D J F M A M J J A S O N D
250 C.ROK-Gangwon 400 D.DPK-Gangwon

200
300
Runoff (mm)

Runoff (mm)
150
200
100 Present Present
2040s 2040s
2090s 100 2090s
50

0 0
J F M A M J J A S O N D J F M A M J J A S O N D
400 E. ROK-Chungcheong 400 F.DPK-Hwanghae

300 300

Runoff (mm)
Runoff (mm)

200 200
Present Present
2040s 2040s
100 2090s 100 2090s

0 0
J F M A M J J A S O N D J F M A M J J A S O N D
400 G.ROK-Jeolla 400 H.DPK-Hamgyeong

300 DJF MAM JJA SON 300

Runoff (mm)

Runoff (mm)
200 200
Present Present
2040s 2040s
100 2090s 100
2090s

0 0
J F M A M J J A S O N D J F M A M J J A S O N D

400 I.ROK-Gyeongsang 400 J.DPK-Yanggange/Jagang

300 300
Runoff (mm)

Runoff (mm)
200 200
Present Present
2040s 2040s
100 2090s 100 2090s

0 0
J F M A M J J A S O N D J F M A M J J A S O N D

17
 Climate change impacts on extreme flows (1):
Flood risk flow
Changes in flood risk flow (2081-2100):
 MRI-RCM A2: ROK +11.4~44.1%, DPK +7.7%~38.1%
 MIROC A1B: ROK +24.8~40.3%, DPK +37.1%~58.8%

MRIRCM-A2 MIROC-A1B

Increase in Increase in Increase in Increase in

Flood events (2090s) Flood severity (2090s) Flood events (2090s) Flood severity (2090s)

18
 Climate change impacts on extreme flows (2):
Drought risk flow
Changes in drought risk flow (2081-2100):
 MRI-RCM A2: ROK -5.4~-14.7%, DPK: -12.3%,
 MIROC A1B: None
MRIRCM-A2 MIROC-A1B

Increase in Increase in Increase in Increase in

Drought events (2090s) Drought severity (2090s) Drought events (2090s) Drought severity (2090s)

19
 Climate change impacts on extreme flows (3):
Potential risk areas
Extreme event risk area affected by floods and drought,
simultaneously (MRI-RCM): ROK about 53%, DPK 24% of
inland area
2040s (2031-2050) 2090s (2081-2100)
MRIRCM-A2 (2040s) MIROC-A1B (2040s) MRIRCM-A2 (2090s)

20
 Ecosystem Contribution to Runoff Change (1):
Experiment setting
Runoff changes in the select eight forested watersheds (23% of
the whole nation) by 2081-2100
Three novel approaches to evaluate the ecosystem contribution
to future runoff increase
 Case C: Climate change only (MRIRCM-A2 scenario)
 Case CP: C + Physiological forcing, P(Atmospheric CO2 conc,. =700
ppmv.)
 Case CPV: CP + Vegetation distribution change, V (35% increases in the
deciduous forest fraction)
Variables SY HW CJ NG HC AD DC SJ Mean
ΔT(˚C) 3.0 2.8 2.9 2.8 2.8 2.9 2.7 2.6 2.8
ΔP(%) 11.4% 14.6% 16.3% 7.1% 6.7% 22.1% 13.7% 13.2% 13.1%
ΔWind(%) 0.2% -0.1% 0.1% -0.3% -0.1% 0.5% -0.7% -1.6% -0.3%
ΔHumidity(%) 18.7% 17.7% 18.0% 17.6% 17.7% 17.8% 16.8% 16.7% 17.6%
ΔRadiation(%) 1.0% 0.5% 1.8% -0.3% 0.1% 0.7% 0.8% 0.4% 0.6%

21
 Ecosystem Contribution to Runoff Change (2):
Results
Climate change alone: 12% (2~17%) increase in runoff
 CP change: 16% (5~32%) increase in runoff
 CPV change: 17% (8~28%) increase in runoff
Increased CO2 effect on runoff: 3.3% (1.1~7.7%)
Increased CO2 and deciduous forests effects: 5.0% (2.5~10.8%)

(mm/yr)
40 500
2
C y = 0.979x - 70.661 R = 0.919
C P V 2
runoff

35 CP y = 1.016x - 52.781 R = 0.970

.

400 2
change in runoff

.
CPV y = 1.039x - 44.591 R = 0.976

(mm/year)
30

runoff
300
in annual

25

Runoff change
20 200
change in annual
Percent

15 100
% change

10
0
5
-100
0
-100 0 100 200 300 400 500 600
SY HW CJ NG HC AD DC SJ Mean 22
Precipitation
Precipitationchange (mm/yr)
change (mm/year)
 Summary
Runoff tends to be increased in western coastal region of the peninsula
and upper interior region of the Han River Basin around the Gwangwon
province.
Extreme flows tend to be increased in maritime parts of the peninsula.
Floods may be increased over the whole peninsula, especially western
coastal region and North Korea, because of increases of the heavy rainfall
in summer season.
Low flow also tends to be increased however the western coastal region
of the peninsula and the middle parts of the Han River basin are showing
reduced low flow by 2090s in MRIRCM scenario.
Changes in forest ecosystem has an additional contribution to future
runoff increases. Increases in atmospheric CO2 concentration and
deciduous forest species fraction produce about 5% of more runoff ranging
from 2.5% to 10.8% in the forested watersheds.

23
 Concluding remarks
Serious changes in extreme events are concerned and water
scarce region in the future is basically corresponding to exiting
hot-spots near the coastal region
In order to provide useful information for impact and
adaptation assessment studies,
 The uncertainties in regional climate projections need to

fully characterized and reduced (probabilistic assessment)

 The role of ecosystem (and current socio-economic

activities) should be considered in the modeling framework

(e.g., DAM building, unification in the near future,
integration of human water security and river biodiversity
etc.)

24