Assessing the Regional Effects of Climate and Land Use Change on the - - PowerPoint PPT Presentation

SLIDE 1

Assessing the Regional Effects of Climate and Land Use Change on the Korean Watersheds

Hui-cheul JUNG Kyoto university The 12th AIM International Workshop

20-Feburary, 2007

SLIDE 2

Contents

• 1. Objectives
• 2. Overview of research flow
• 3. Mapping and Modeling Land Use and Land Cover Change

3.1 Detecting and projecting land cover change using RS dataset 3.2 Spatial allocation model 3.3 Projecting future land use change and conservation priority

• 4. Modeling the Effects of Climate and Land Use Change

4.1 River network system of Korean Peninsula 4.2 Hydrological modeling system 4.3 Projecting potential effects of climate change on the Korean watersheds

• 5. Conclusion

SLIDE 3

• 1. Objectives

Climate Change and Land Use/Cover Change (LUCC) caused by anthropogenic pressures and

natural processes are increasingly recognized as the major drivers of global environmental change which may be a synthesis of results of these two changes.

Since they have a large effect on ecosystem processes, biogeochemical cycles, biodiversity and

even more human activities, many scientists have gained great efforts to understanding the causes and effects of land use and climate changes.

In addition, Land use change of Korean Peninsula, especially urban sprawl and deforestation,

also have been accelerated by the rapid increasing of urban population during the last 30 years.

To understand the effects on the future environment caused by climate and LULC changes. It is

needed to develop diagnostic model of land use and ecosystem changes, and to analyze regionally comparable results of climate and land use change.

The objective of this research is assessing potential effects of climate and land use change on the ecosystem and hydrologic processes in Korean watersheds by developing impact assessment model and regional dataset (land cover, river-network system etc) This presentation mainly focus on the development of impact assessment models- 1) land use allocation, 2) hydrologic model including their validation

SLIDE 4

• 2. Overview of research flow

Figure 1. Framework for research flow and Relationship of Climate and Land Use change on the Ecosystems and Hydrologic processes in the Korean watersheds

3 4 2 1

SLIDE 5

• 3. Mapping and Modeling LULC Change

Remote Sensing Analysis

• Data: Landsat 5 TM and 7 ETM+ (Spatial resolution: 30m)
• Classification classes: Built-up, Agriculture, Forest, Grass, Barren, Wetland,

Water (7 classes)

• Spatial domain and Target date: Korean Peninsula (ROK,DPK), End of 1980s

(1985-1989, rep. 1987) and 1990s (1995-1998, 1997)

• Classification method: Hybrid methods (unsupervised: competitive training +

supervised: maximum-likelihood with prior probability)

• Classification error criteria: Overall accuracy 80%
• Validation: 1) Statistical method using Kappa and overall accuracy, 2) Area

comparison with statistical year book (forest, agriculture)

Statistical Analysis for detecting the changing pattern

• Descriptive variables: distance to road, existing city, river and conservation

area, pop.density, topographic (slope, height etc), Soil (textures, pH, AWC etc), Climate variables (T, P, PET etc)

• Proximity analysis of urban expansion
• Markov transition matrix by using multi-logit regression for future change

estimation

SLIDE 6

3.1 Land cover classification results and validation

(a) Forest - National areas

20 40 60 80 100

Classified area (1000 sq. km)

20 40 60 80 100 Land cover 1987, ROK Land cover 1997, ROK Land cover 1997, DPRK Land use 2000, ROK

(b) Agriculture - National areas

20 40 60 80 100 20 40 60 80 100 Land cover 1987, ROK Land cover 1997, ROK Land cover 1997, DPRK Land use 2000, ROK

(f) Agriculture- County areas

Reported land use area (103 km2)

0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5 Land cover 1997, ROK Land use 2000 - Rice paddy Land use 2000 - Dry field

(e) Forest - County areas

Reported land use area (103 km2)

0.0 0.5 1.0 1.5 2.0

Classified area (1000 sq. km)

0.0 0.5 1.0 1.5 2.0 Land cover 1997, ROK Land use 2000, ROK

Figure 2 Current status of land cover distribution on the Korean Peninsula (left) and Comparison of the major land use classes in terms of remotely sensed land cover and census data set (Right)

SLIDE 7

3.1 Major change among the classes

Green-belt zone (c) Metropolitan area (1980s) National highway (1990s)

Figure III Net change and Relative conversion of land cover classes (Left) and Urban sprawl in the Metropolitan area surrounding Greenbelt zone

SLIDE 8

3-1. Relationship of Land Cover Fraction and Urban population density

(a) Built-up area

Population density (persons/km2)

5000 10000 15000 20000

Fraction (%)

10 20 30 40 50 60 70 DPRK ROK Urb%=0.0056 * Pden + 1.88 (r2 = 0.6888), DPRK Urb%=0.0032 * Pden + 4.0628 (r2 = 0.8706), ROK

(b) Urban forest

Population density (persons/km2)

5000 10000 15000 20000 20 40 60 80 100

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 2627 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83

Chungcheong Gangwon Gyeongsang Jeollajeju Metropolitan DPRK

Figure 4 Land cover distribution of major cities relating with urban population density: Built-up (Left) and Urban forest (Right)

SLIDE 9

3-1. Proximity analysis to urbanized area

(a) ROK

• km. to nearest 1980s development

5 10 15 20 25 Cumulative percent 20 40 60 80 100 cumulative new development cumulative total land

(b) ROK

• km. to nearest national highway

2 4 6 8 10 Percent of total land by distance 2 4 6 8 10 20 40 60 80 100

% urban 1990s % urban 1980s cumulative development 1990s cumulative total land

DPRK

• km. to nearest 1980s development

5 10 15 20 25 20 40 60 80 100 cumulative new development cumulative total land DPRK

• km. to nearest national highway

2 4 6 8 10 2 4 6 8 10 cumulative percent 20 40 60 80 100

% urban 1990s % urban 1980s cumulative development 1990s cumulative total land

(c) ROK

• km. to nearest ocean

20 40 60 80 100 Percent of total land by distance 2 4 6 8 10 20 40 60 80 100

% urban 1990s % urban 1980s cumulative development 1990s cumulative total land

DPRK

• km. to nearest ocean

20 40 60 80 100 2 4 6 8 10 cumulative percent 20 40 60 80 100

% urban 1990s % urban 1980s cumulative development 1990s cumulative total land

(d) ROK - Greenbelt zones

• km. to nearest greenbelt

10 20 30 40 50 60 Percent of total land by distance 10 20 30 40 20 40 60 80 100

% urban 1990s % urban 1980s cumulative development 1990s cumulative total land

ROK - National Parks

• km. to nearest national parks

10 20 30 40 50 60 2 4 6 8 10 cumulative percent 20 40 60 80 100

% urban 1990s % urban 1980s cumulative development 1990s cumulative total land

(a) Distance to existing built-up area (b) Distance to nearest highway (circa. ’80s road) (c) Distance to nearest sea (d) Distance to greenbelt and national park Figure 5 Relationship with urbanized area and proximity (km) to (a) existing urban (b) road (c) Ocean and (d-1) Greenbelt zone (d-2) National parks (Conservation area). (Left) ROK, (Right) DPK excepting (d)

SLIDE 10

3-1. Future land cover change estimation using Markov transition matrix

Future population at provincial level: SRES A2 scenario Transition probability, (Pij ) is estimated by multi-logit regression:

Mlogit-Forest (1997)

y = 0.9723x + 3.9618 R

2 = 0.939

20 40 60 80 100 120 20 40 60 80 100 120 Fraction of 1997 land cover Fraction of simulated land cover

Mlogit-Builtup (1997) y = 0.9572x + 0.4341 R

2 = 0.9011

20 40 60 80 100 120 20 40 60 80 100 120 Fraction of 1997 land cover Fraction of simulated land cover Mlogit-Agriculture (1997) y = 0.914x + 0.6391 R

2 = 0.8922

20 40 60 80 100 120 20 40 60 80 100 120 Fraction of 1997 land cover Fraction of simulated land cover

1 1 1 2

ln( / ) ( , ,..., ), 1

i j m

p p f X X X i j = ≠ ≠

Figure 6 Comparison of observed and simulated land cover fraction of Forest (left), Built-up (middle) and Agri. (right) at 5 by 5 km grid

SLIDE 11

3-1. Estimated future land cover change

DPK 0% 20% 40% 60% 80% 100% 1987 2000 2010 2020 2030 2040 2050 Forest Developed Agriculture Other ROK 0% 20% 40% 60% 80% 100% 1987 2000 2010 2020 2030 2040 2050 Forest Developed Agriculture Other

Nation Classes 1987 2000 2010 2020 2030 2040 2050 Forest 91.733 89.530 88.025 86.925 86.042 85.277 84.574 Agriculture 21.328 22.567 22.977 22.920 22.584 22.077 21.462 Developed 1.405 2.453 3.639 4.920 6.279 7.713 9.204 Other 6.022 5.939 5.848 5.724 5.583 5.423 5.249 Forest 66.411 68.340 68.871 68.784 68.387 67.829 67.183 Agriculture 23.533 20.548 18.748 17.468 16.450 15.572 14.774 Developed 2.112 3.849 5.482 7.093 8.698 10.298 11.889 Other 5.007 4.326 3.963 3.718 3.528 3.364 3.216 DPK ROK

Table 1. Land cover change simulation results by Markov transition matrix, start form 1987 and SRES A2 (provincial level)

SLIDE 12

3-2. Land Use Change Model

Figure 7 Framework of land use change modeling and Ecosystem effects analysis by using land use suitability and management option

• 1. LU allocation probability, Mcj
• Modified form of constrained

logistic model

• Double constrains: 1) national

demand constrain 2) grid cell fraction constrain

• Water fraction is fixed
• 2. Suitability adjusting factor, β:
• Optimization of parameters
• Minimize RMSE(Fo,Ag.Bu) sum
• 1. Land use suitability, Scj:
• Regression of descriptive

variables for Built-up, Agri., Forest, Others and Water (5 types)

• 5 by 5 km grid

2.Conservation weight, Ecj:

• Using Land Environment Mapping

(LEI) criteria (5 grades)

• for the negative change, changing

amount is adjusted using weight

SLIDE 13

3-2. Spatial allocation method

( )

c cj c j

M L D ∑ ⋅ =

1

j cj

M ∑ =

• Step 1: Estimated land use fraction at cell, c and land use type, j

exp( )

cj j c j cj

M a b S β = ⋅ ⋅ ⋅

• Step 2: Adjusting Mcj considering conservation weight Ecj

at cell, c and land use type, j

• Step 3: Calculate coefficients, aj and bc by double constrains

( )

' 1 1 t t cj cj cj cj cj

M x M x E

− −

= + − ⋅

Step 4: Global optimization of βj

( )

( )

( )

2

minimize ( ) / / for

cj cj cj j c x

M N j M − ∀ >

∑ ∑

SLIDE 14

Relative effect of β coefficient 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0

Land use suitability, Scj Expected land use fraction, X

β=0.1 0.5 1 2 5

Figure 8 Relative effect of βcoefficient on Expected land use fraction, Xi

3-2 Suitability adjust coefficient, βj

Minimize Residual error

SLIDE 15

3-2. Conservation weight, Ecj

Spatial importance and criteria for the environmental land management:

• Legal system-ecosystem(14 categories) , water quality(11) and Land Use (28)
• Natural properties-Biodiversity Naturality (3), Richness, Rarity, Weakness (2), Stability

(2), Connectivity (1)

National Land Environmental Index (LEI)

• Absolute conservation area (grade 1) - A very valuable land for environmental

conservation Weight 0.0 for negative land use change

• Priority conservation area (grade 2) Weight. =0.5,
• Others(3,4,5) Weight = 1.0

Spatial averaged conservation weight

/

cj g g g g g

E N W N = ∑ ⋅ ∑

SLIDE 16

3-2. Conservation weight, Ecj

(a) Legal criteria (b) Environmental criteria (c) Final results

Figure 9. Spatial analysis results of assessing Land Environment Index. Table 2. Relative change of land cover classes during 1987-1997 by the LEI region

LC classes grade 1 grade 2 grade 3 grade 4 grade 5 Buit-up 0.12 0.13 0.36 0.21 1.00 Agriculture

• 0.24
• 0.57
• 0.67
• 0.19
• 1.00

Forest 0.30 0.68 0.08

• 0.32
• 1.00

Others*

• 0.15

0.22 0.59 0.50 1.00 LEI index

SLIDE 17

3-2. Optimization of LU model

Figure 10. Spatial difference(%) of LU modeling by time varying types with legal criteria except water quality(E0)

• Table9. optimization results of land use change model with natural protection

Observed Simulated Observed Simulated β RMSE(%) β RMSE(%) β RMSE(%) β RMSE(%) 1 Forest 66304.7 66304.6 86676.4 86676.3 24.92 12.41 21.68 9.96 2 Agriculture 21600.1 21600.1 23982.8 23982.8 19.62 10.32 16.26 9.23 3 Built-up 3429.8 3429.8 2121.1 2121.1 2.29 2.16 3.23 3.23 4 Others 5937.9 5937.9 7792.3 7792.3 9.64 9.64 7.62 7.62 5 Water 7427.5 7427.5 4190.1 4190.1 0.00 0.00 0.00 0.00 1 Forest 66304.7 66304.7 86676.4 86676.3 24.84 12.26 21.27 8.89 2 Agriculture 21600.1 21600.1 23982.8 23982.8 19.60 10.25 16.09 8.61 3 Built-up 3429.8 3429.8 2121.1 2121.1 4.67 3.56 7.21 5.65 4 Others 5937.9 5937.9 7792.3 7792.3 9.64 9.64 7.47 7.47 5 Water 7427.5 7427.5 4190.1 4190.1 0.00 0.00 0.00 0.00 ID LU types Type Model optimization and spatial error of land use fraction Land use demand ('97) constrains 0.250 0.250 Time varing ('87 to '97) Self- correlation ('97 to '97) 0.250 3.015 2.638 3.045 2.770 ROK DPK 0.250 ROK-before DPK-before ROK-after DPK-after

Forest Agriculture Built-up

SLIDE 18

3-3. Projecting future land use change and conservation priority

Figure 11. Projection of potential land use change 2050 (%change of Built-up) (a) E0 (b) E1 (c) Markov

SLIDE 19

3-3. Projecting future land use change and conservation priority

Figure 12. Projection of potential land use change 2050 (%change of Forest) (a) E0 (b) E1 (c) Markov

SLIDE 20

3-3. Land use change in the major BS

Figure 13. Land use change of the major Korean watersheds (E0 and E1 management)

• 20.0
• 15.0
• 10.0
• 5.0

0.0 5.0 10.0 15.0 20.0

Han Nakdong Geum Seomjin Yeongsan Seomjin_south Han_east Nakdong_east Geum_west Nakdong_south Yeongsan_west Cheju Han_west Sapgyocheon Anseongcheon Mangyeong Yeongsan_south Dongjin Hyeongsan Hoeya Tehwa Tamjin Baengnyeong Ulnung Amrok Tuman Taedong Chongchon Tuman_east Yesong Chaeryong Taeryong Chaeryong_west Namdecheon_east Yesong_west Amrok_west Seongcheon Daedong_east Dancheon_Namdaecheon Yonghung Eorangcheon Bukcheong_Namdaecheo Taedong_west Han_east_N Puktaecheon Kilju_Namdaecheon Deokji Anbyon_Namdaecheon Suseongcheon

% Area change Forest-E1 Forest-E0 Agri-E1 Agri-E0 Builtup-E1 Builtup-E0

SLIDE 21

• 4. Modeling the Effects of Climate and Land Use Change

Development of Korean River Network

• Spatial domain: Korean

Peninsula (ROK,DPK)

• Data: DEM-SRTN 3sec DEM

(100m), River-ROK; Digital topo-map layer, DPK; Digitizing 1:250000 paper map, Basin boundary-ROK: national unit basin map

• Delineation method: ArcHydro

tool same method with GDBD (BS criteria: 250km2)

• Validation: 1) Reported basin

area comparison – Yellow colored circles in Fig 14.

• Figure 14 Korean Major Basin

and unit basin

SLIDE 22

4-1 River Network Systems of Korean Peninsula

(n=174) 20 40 60 80 100 120 140 < -50

• 50~-30
• 30~-20
• 20~-10
• 10~-5
• 5~5

5~10 10~20 20~30 30~50 > 50 |(Obs-Cal)/Obs*100| (%) Freguency R2 = 0.9994, (n=174)

5000 10000 15000 20000 25000 30000 5000 10000 15000 20000 25000 30000 Reported Area (sq. km) GIS Area (sq. km)

Figure 15. Basin area comparison: (Left) Delineated area (sq. km) (Right) Frequeny of % difference of BS area (N=174)

SLIDE 23

4-2. Hydrological modeling system

Figure 16. Flow diagram of the semi-distributed hydrologic modeling system

SLIDE 24

4-2. Study region and station info. for validation

Table 10. Location and topographic information of discharge monitoring stations for model validation

Code Guage name Latitude of centroid (deg) Dainage area (km2) M ean elevation (m) M ean slope (deg) M ean aspect (deg)

• 1. Republic of Korea (ROK)

015100 Soyanggangdam 38.0327 2694.3 657.7 17.60 181.60 015700 Hwacheondam 38.4824 4086.1 580.3 15.51 183.71 017500 Chungjudam 37.2814 6659.3 616.1 16.88 181.08 021500 Namgangdam 35.4172 2287.5 432.0 14.32 175.43 023100 Hapcheondam 35.7095 929.0 508.4 14.08 179.34 028320 Imhadam 36.5497 1366.9 400.5 13.86 183.24 029100 Andongdam 36.8809 1593.3 563.5 15.21 177.18 036000 Daecheongdam 36.1258 4189.8 366.3 13.38 182.29 052270 Juamdam 34.9428 1028.6 273.3 12.34 181.66 057700 Seomjingangdam 35.6357 763.7 363.9 12.67 182.50

• 2. Democratic Peoples Republic of Korea (DPRK)

031255 M irim 39.39773 12396.8 465.1 15.22 182.30 031258 Samdung 38.99291 2801.9 476.6 17.87 183.40 031259 Dokchon 39.94235 3346.0 818.0 18.16 183.76 031265 Kumchang 41.08397 18278.1 1319.8 14.67 180.83 031496 Jonchon 40.53503 2146.1 982.7 18.47 185.38 031497 Songchon 39.33842 1731.3 432.9 15.78 184.47 Topographic

SLIDE 25

4-2. Model optimization and validation

Sta tio n = 036000, (O p tim iza tio n : 1996-1998, V a lid a tio n : 1999-2001) N SE = 0.688, B I A S% =

• 9.1%

fo r v a lid a tio n 1 2 3 4 5 6 1 3 6 6 7 3 1 1 9 6 1 4 6 1 1 8 2 6 2 1 9 1 D ays Discharge(cms) Q

• bs

Q sim

No optimized monthly mean discharge (1996-2001) 100 200 300 400 500 600 700 800 100 200 300 400 500 600 700 800 Observed discharge(cms) Simulated discharge(cms)

C alib ratio n V alid atio n IN F E X P C L IN C S Q S G R G S C C R G L M X Q D E P T H C alib ratio n V alid atio n C alib ratio n V alid atio n 1 5 1 S

• y

an g g an g d am 1 9 9 2

• 1

9 9 9 2

• 2

1 1 .9 9 9 7 .1 4 6 .0 3 5 8 7 .1 4 7 5 4 .0 2 5 6 4 1 5 .8 8 5 .8 7 7

• 1

4 .6 8 4

• 5

.9 8 5 1 5 7 H w ach eo n d am 1 9 9 2

• 1

9 9 9 2

• 2

1 1 .9 9 7 6 .1 1 9 9 6 .0 2 4 4 7 .0 7 6 5 9 .0 1 6 8 7 1 1 .7 4 3 .8 2 9 6 .0 1 3 6 .4 8 8 1 7 5 C h u n g ju d am 1 9 9 2

• 1

9 9 9 2

• 2

1 1 .9 9 8 7 .1 1 .0 1 4 6 4 .0 9 1 4 .0 2 2 7 6 1 5 .6 8 5 .8 1

• 2

.8 9 7

• 7

.5 4 2 2 3 1 H ap ch eo n d am 1 9 9 2

• 1

9 9 9 2

• 2

1 1 .8 4 1 6 .8 1 4 1 8 .0 4 6 9 3 .2 8 9 1 4 .0 1 1 1 5 .8 6 6 .8 3 6 .1 9 5

• 1

.3 4 5 2 9 1 A n d

• n

g d am 1 9 9 2

• 1

9 9 9 2

• 2

1 1 .9 9 5 7 .1 1 7 3 .0 2 9 7 6 .1 4 1 4 9 .0 2 7 3 2 1 5 .7 1 6 .8 5 3

• 7

.1 2 1

• 2

.2 9 5 3 6 D aech eo n g d am 1 9 9 2

• 1

9 9 9 2

• 2

1 1 .9 9 8 6 .2 1 5 4 .0 2 5 5 9 .0 7 9 7 7 .0 1 2 1 7 1 5 .7 4 1 .7 4 7 1 .1 3 3 2 .0 1 5 2 2 7 Ju am d am 1 9 9 6

• 1

9 9 9 2

• 2

1 1 .1 2 6 .1 4 3 2 .2 9 9 8 8 .2 9 8 5 .0 1 1 1 5 .8 3 6 .5 7

.6 5 1 8 .7 7 9 5 7 7 S eo m jin g an g d am 1 9 9 2

• 1

9 9 9 2

• 2

1 1 .8 6 7 .4 7 8 9 7 .1 3 8 4 7 .1 2 6 1 1 .0 1 1 4 1 1 5 .9 4 .8 6 5 .1 6 4 .4 9 7 S tatio n N

• .

S tatio n P erio d

• f reco

rd N ash

• S

u tcliffe efficien cy B ais (in p ercen t) O p tim izatio n p aram eters

Table 11. Parameters of fitted model, with efficiency criterion over calibration and validation pe riods for 15 Korean basins Figure 17. Location and topographic information of discharge monitoring stations

SLIDE 26

4-2. Land surface parameterization

Figure 18. Land surface parameterization and estimated hydrologic component (Station = 036000, 1993/01/01-1994/12/31)

100 200 300 400 1 31 61 91 121 151 181 211 241 271 301 331 361 391 421 451 481 511 541 571 601 631 661 691 721

Soil water (mm)

S W AT (depth = 1 m ) S W AT(depth = 5 cm ) 5 1 1 5 2 2 5 3 1 Discharge (cms) Q

• bs

Q sim .0 2 .0 4 .0 6 .0 8 .0 1 AET (mm)

0.7*PA N A E Tsim 10 day m

• ving

averag e of 0.7*PA N E T 10 day m

• ving

averag e of sim ulated A E T

SLIDE 27

4-3. Projecting potential effects of climate change on the Korean watersheds: Examples for evapotranspiration

Figure 19. Spatial distribution of current climate(1971-2000, daily KMA and ECMWF-Reanalyzes results, (a) annual mean temperature, (b) precipitation) and (c) simulated vegetation water stress ratio (Actual Trans. / Potential Trans.) (a) (b) (c)

SLIDE 28

4-3. Future Evaporation Change

Figure 20. Change of annul evaporation rate under the same land cover condition (mm/day); (a) current (b) MRI-RCM 2050s (c) Change rate (a) (b) (c)

SLIDE 29

4-3 Future Transpiration Change

Figure 21. Change of annul transpiration rate under the same land cover condition (mm/day); (a) current (b) MRI-RCM 2050s (c) Change rate (a) (b) (c)

SLIDE 30

Thank You