Kwan Tun LEEKwan Tun LEE

Professor of River and Harbor EngineeringNational Taiwan Ocean University, Keelung, Taiwan

Adjunct Senior Research FellowTaiwan Typhoon and Flood Research InstituteNational Applied Research Laboratories, Taipei, Taiwan 24 November 201524 November 2015

2015 APEC Typhoon Symposium2015 APEC Typhoon SymposiumUniversity of the Philippines, Diliman, Philippines

Runoff estimation for typhoon Runoff estimation for typhoon storms in ungauged watershedsstorms in ungauged watersheds

2

Objectives of this study Objectives of this study

More frequent and severe typhoon storms were noticed in recent years. The impact resulted from climate variability has increased the extent of flood disaster.

To develop a hydrological model for watershed runoff estimation in gauged or ungauged watersheds is an important task for hydrologists.

An integrated window-based hydrological system would be convenient and efficient for hydrologists to perform engineering design and flood forecasting.

3

Conventional rainfall-runoff modeling Conventional rainfall-runoff modeling Linear system model / black-box model /conceptual models

unit hydrograph (UH; Sherman, 1932)instantaneous unit hydrograph (IUH)time-area method (Clark, 1945)linear reservoir method (Nash, 1957)tank model (Sugawara, 1974)

Practical application problemsThe models can not be applied to flow ungauged watersheds, and simulation is poor for watersheds with highly nonlinear & time-variant characteristics.

Grid-based hydraulic models kinematic wave model (Singh, 1976)

diffusion wave model (Akan & Yen, 1981)dynamic wave model (Fread, 1978)SHE model (Abbott et al., 1986)

Practical application problemsIt is time consumed for performing a grid-based runoff routing models.

4

All purposes rainfall-runoff modelingAll purposes rainfall-runoff modeling

The model should be derived only based on watershed geomorphologic characteristics. So, it can be applied to any location with/without flow recorded data.

It should be a nonlinear & time-variant model to account for the hydrodynamic phenomena of the watershed.

The model should be performed in an efficient way to perform real-time flood forecasting.

5

Geomorphologic Instantaneous Unit Hydrograph Geomorphologic Instantaneous Unit Hydrograph

(GIUH)(GIUH)

xxxxxxOA kjiioii

P P PPwP

xxxxw TTTTT

jioi

unit rainfall → N independent raindrops

xo1 x1

x2 x3

xo1 x1

x3 xo2 x2

x3

xo3 x3

xo1→ x1→ x2→ x3

xo2→ x2→ x3

Q

t

xxxxxxOA kjiioii

P P PPwP

watershed geomorphologic IUH (Rodriguez-Iturbe & Valdes, 1979)

where f(t) is the runoff travel time distribution

the probability for a raindrop adopting a specified path

total runoff travel time along the path

u t f t f t f t f t P wx x x xW w

oi i jw

6

GIUH model considering subsurface flowGIUH model considering subsurface flow

effective rainfall

groundwater table

subsurface-flow region

subsurface-flow region

surface-flow region

surface-flow region

effective rainfall

groundwater table saturated zone

unsaturated zonesurface flowsubsurface flow

V-shaped model for

ith-order subwatershed

Rainfall infiltrates as subsurface flow on the upper portions of the watershed.

Surface flow can only occur on the lower portions of the watershed during storms.

(Lee and Chang, 2005)

7

Flow path probability

surface flow

subsurface flow

Runoff travel time

surface flow

subsurface flow

Surface & Subsurface IUH

GIUH model considering subsurface flowGIUH model considering subsurface flow

surface-flow region

2

2

2

1 1

1

1

1

11

3

xxxxw TTTTT

jiisubsub

xxxxw TTTTT

jiios

swxWw

xxxs wPtftftftftus

ss

jiio

)(**)(*)(*)(

subwxWw

xxxsub wPtftftftftusub

ubssub

jiisub

)(**)(*)(*)(

( ) ( ) ( )s subu t u t u t

surface flow paths

xo1 →x1 →x2 →x3

xo1 →x1 →x3

xo2 →x2 →x3

xo3 →x3

subsurface flow paths

xsub1 →x1 →x2 →x3

xsub1 →x1 →x3

xsub2 →x2 →x3

xsub3 →x3

i i o ii

s PCA OA x xP w R P P i jx xP

kx xP

1i i sub ii

sub PCA OA x xP w R P P ji xxP xxk

P

8

Runoff travel time estimationsRunoff travel time estimations

0

i

subii

sub

xsub

L

SKT

swxWw

xxxs wPtftftftftus

ss

jiio

)(**)(*)(*)(

subwxWw

xxxsub wPtftftftftusub

ubssub

jiisub

)(**)(*)(*)(

( ) ( ) ( )s subu t u t u t

Overland-flow travel time

Subsurface-flow travel time

Channel-flow travel time

1

1 2 -1e

i

oi

i

mo

x mo

o LT

S i

n

1/

1 2

2

2i i

i i i

i i

m

sub cmi ex co co

sube ci

cB ni L LT h h

i L B S

Kinematic-wave-based geomorphologic IUH (KW-GIUH)(Lee and Yen, 1997; Lee and Chang, 2005)

(Lee and Yen, 1997)

(Henderson and Wooding, 1964)

(Henderson and Wooding, 1964)

9

Comparison of surface- and subsurface-flow IUHsComparison of surface- and subsurface-flow IUHs The rising limb of the IUH is dominated by surface-flow mechanism, and the

recession limb is dominated by subsurface-flow mechanism.

0 6 1 2 1 8 2 4

T im e (h r)

0

0 .1

0 .2

0 .3

IUH

(hr

-1)

H e n g -C h i

ie = 1 0 m m /h rK 0 = 0 .0 2 5 m /sR P C A = 0 .5n o = 0 .6 , n c = 0 .0 5

su rfa ce -f lo w IU H

su b su rfac e -flo w IU H

to ta l IU H

The duration of the subsurface-flow IUH is longer than that of the

surface flow IUH.

10

Runoff simulation in Heng-Chi watershedRunoff simulation in Heng-Chi watershed

0 1 2 2 4 3 6 4 8 6 0 7 2

T im e (h r)

0

1 0 0

2 0 0

3 0 0

Q (

m3/s

)

3 0

2 0

1 0

0

i e (m

m/h

r)

H en g -C h i Ju ly 1 9 9 6

reco rd ed su b su rface flo w d irec t ru n o ff

Rainstorm in July 1996

11

Time varying & nonlinearity of the KW-GIUHTime varying & nonlinearity of the KW-GIUH

0 1 2 2 4 3 6

T im e (h r)

0 .0

0 .1

0 .2

0 .3

0 .4

IUH

(hr

-1)

S an -H sian o = 2 .0n c= 0 .0 5ie (m m /h r)

11 05 01 0 0

1

1

1 2i

oi

i

mo

mx

o e

o

i

n LT

S

1

1 2

2

2i i

i i i

i i

mo cmi c

x co coo c i

e

e

B n L LT h h

L S B

i

i

Flow travel time estimation

the IUH is a function of the rainfall intensity a set of IUHs instead of only one IUH for a specified watershed

12

Input of KW-GIUH modelInput of KW-GIUH model

Watershed geomorphologic factors Loi ith-order overland-flow length Soi ith-order overland-flow slope Lci ith-order channel-flow length Sc ith-order channel-flow slope Ai ith-order subwatershed contributing area B channel width at the watershed outlet no roughness coefficient for overland flow nc roughness coefficient for channel flow Ko hydraulic conductivity

Watershed hydrological characteristics rainfall intensity

ei

surface-flow region

2

2

2

1 11

1

1

11

3

13

Digital elevation model (DEM)Digital elevation model (DEM)

30m

Flow direction is determined according to the elevation in the eight adjacent cells

• Flow direction determination

• Depressionless

• Flow accumulation value calculation

• Channel network extraction

• Subwatershed delineation

• Geomorphologic factors determination

In 2014, NASA released globally 30m resolution topographic data generated from the Shuttle Radar Topography Mission (SRTM).

14

Geomorphologic factors calculationGeomorphologic factors calculation

A H

L S Ai

Sci

Soi

Lci

watershed areamean elevationmean slopemainstream lengthmainstream slopestream orderith-order subwatershed areaith-order channel slopeith-order overland slopeith-order channel slopewetness index

S

digital elevation dataset

watershed boundary and stream network

watershed geomorphologic factors

)tan

ln(

a

15

Overland roughness determination – SPOT imageOverland roughness determination – SPOT image

Remote sensing images can be used to classify the land cover distribution of a watershed for overland roughness coefficients determination.

SPOT satelliteNear-infraredRedGreen

WaterBuilding and roadBared SoilGrassForest

16

Merits of the KW-GIUH model

0 5 10 15T im e ( hr )

0.0

0.3

0.6

0.9

IUH

( h

r-1 )

10010

1

H eng-C hi

i (m m /hr)e

n =0.3n =0.03

o

c

0 5 10 15Tim e ( hr )

0.0

0.2

0.4

0.6

IUH

( h

r

)-1 0 .1

0.3

0.5

n =0.03c

i =10 m m /hre

n o

H eng-C hi

1 10 100R ecurrence interval T E (years)

0

500

1000

1500

Q (

m3/s

)

Heng-Chi(1975-1992)

Type Ith is study

Influence of rainfall intensity variation can be considered. Influence of environmental changes can be simulated. Flow frequency analysis in ungauged areas can be performed. The model can be performed in an efficient way.

17

Flow frequency analysis in ungauged watershedsFlow frequency analysis in ungauged watersheds

Lee, K. T. (1998). “Generating design hydrographs by DEM assisted geomorphic runoff simulation: a case study,” J. Am. Water Resour. Assoc., 34(2), 375-384.

18

Flow analysis in landslide-dammed-lake watershedFlow analysis in landslide-dammed-lake watershed

• Watershed geomorphologic factors and characteristic curve of the landslide-dammed lake are calculated using DEM• Land cover condition is obtained from remote sensing image analysis• Flow analyses are performed by using KW-GIUH model and TOPMODEL

y3

y1

y2 y1

y2

y3

1 2 3

Lee, K. T., Lin, Y.-T. (2006). “Flow analysis of dammed-up-lake watersheds: a case

study,” Journal of the American Water Resources Association, 42(6), 1615-1628.

19

Runoff analysis in USARunoff analysis in USA

Salt watershed (A=876 km2), Kaskaskia watershed (A=2692 km2), Illinois, US

Yen, B. C. and Lee, K. T. (1997). “Unit hydrograph derivation for ungauged watersheds by stream order laws,” J. Hydrologic Engrg., ASCE, 2(1), 1-9.

20

Sadovy watershed, Komarovka River Basin, Russia (Area=395 km2)

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

T im e (h r )

0

1 0 0

2 0 0

3 0 0

Q (

m3 /

s)

2 0

1 0

0

i e (

mm

)

1 6 A u g . 1 9 6 8

R eco rd edK W -G IU H

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

T im e (h r )

0

4 0

8 0

1 2 0

1 6 0

Q (

m3 /

s)

86420

i e (

mm

)

6 A u g . 1 9 7 1

R ec o rd e dK W -G IU H

Lee, K. T., Chen, N.-C., Gartsman, B. I. (2009). Impact of stream network structure on the transition break of peak flows, Journal of Hydrology, 367, 283-292.

Runoff analysis in RussiaRunoff analysis in Russia

21

Faria catchment, Palestine (Area=334 km2)

Shadeed et al. (2007). “GIS-based KW-GIUH hydrological model of semiarid catchments: the case of Faria catchment, Palestine. Arabian Journal for Science and Engineering.

Runoff analysis in PalestineRunoff analysis in Palestine

22

Gagas watershed, Uttarakhand, India (Area=506 km2)

Kumar, A. and Kumar, D. (2008). Predicting Direct Runoff from Hilly Watershed Using Geomorphology and Stream-Order-Law Ratios: Case Study, Journal of Hydrologic Engineering, ASCE, 13(7), 570-576.

Runoff analysis in IndiaRunoff analysis in India

23

Yasu River basin, Japan (Area=387 km2)

Chiang, S., Tachikawa, Y., and Takara, K. (2007). Hydrological model performance comparison through uncertainty recognition and quantification, Hydrological Processes, 21(9), 1179-1195.

input uncertainty = 1 mm/hr

Runoff analysis in JapanRunoff analysis in Japan

24

Yi-Jin River watershed, Sichuan, China (Area=1700 km2)

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

T im e ( h r )

0

3 0 0

6 0 0

9 0 0

1 2 0 0

1 5 0 0

Q (

m3 /s

)

2 01 61 2

840

i e (

mm

/hr)

A u g . 1 9 8 6

R e c o rd ed

S im u la te d

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

T im e ( h r )

0

3 0 0

6 0 0

9 0 0

1 2 0 0

1 5 0 0

Q (

m3 /s

)

1 086420

i e (

mm

/hr)

S e p . 1 9 8 6

R e c o rd ed

S im u la te d

3.0

0.06o

c

n

n

3.0

0.06o

c

n

n

Cao, S.-Y., Lee, K. T., Ho, J.-Y., Liu, X. N., Huang, E., Yang K.-J. (2010). “Analysis of runoff in ungauged mountain watersheds in Sichuan, China using kinematic-wave- based GIUH model,” Journal of Mountain Science, 7, 157-166.

Runoff analysis in Mainland ChinaRunoff analysis in Mainland China

25

Runoff analysis in IranRunoff analysis in Iran

Estimating Runoff and Sediment Yield Using KW-GIUH and MUSLE Models: Runoff and Sediment yield modeling Saleh ArekhiAssistant professor, Agriculture College, ILam University, Ilam, Iran,

Amazon.com Price:$67.00

26

Geomorphologic & hydrological information systemGeomorphologic & hydrological information system

Design discharge estimatingGeomorphologic factors calculatingLand cover analysisFlow frequency analysisAverage rainfall calculatingRainfall frequency analysisDesign hyetograph developmentRainfall-runoff simulating

27

Integrated channel-flow-routing systemIntegrated channel-flow-routing system

Subwatershed boundary delineatingSubwatershed rainfall calculatingSubwatershed inflow calculatingHEC-RAS channel-flow routingMuskingum-Cunge channel-flow routing

28

Geo. & Hydro. Analy. Channel Routing Sys. Inundation Simu.

Windows-based integrated analysis systemWindows-based integrated analysis system

flow frequency analysis

Geomorphologic factors

land cover analysis

Regional rainfall

Rainfall frequency analysis

Designhyetograph

rainfall-runoff simulation

designdischarge

Non-inertia wave model

FLO-2DHEC-RAS

Muskingum-Cunge

Subwatershed boundary delineating

Subwatershed rainfall calculating

Subwatershed inflow calculating

Subwatershed rainfall calculating

Subwatershed inflow calculating

29

ConclusionsConclusions

KW-GIUH model can be applied to gauged and ungauged watersheds for runoff simulation only based on watershed geomorphologic information.

Instead of using grid-based routing models, the IUH concept operating provides an efficient way for rainfall-runoff simulation, and the runoff nonlinearity can also be considered in the KW-GIUH model.

The integrated windows-based platform can provide both geomorphologic and hydrological information for hydrologists to perform engineering design work and real-time flood forecasting at any desired point in the watershed.

Thank YouThank You

Prof. Kwan Tun LEEGeographic Information System Research CenterNational Taiwan Ocean University, Keelung, TaiwanE-mail: ktlee@ntou.edu.tw

31

Kinematic-wave-based runoff travel time

C

A

B

A B

C

hcoi

m

meo

oox

iS

LnT

i

i

oi

1

121

Travel time for ith-order overland-flow

Travel time for ith-order channel-flow

TB

i Lh

i n L L

Shx

i

e ocom e c o c

c i

m

coii

i

i i

i

i

2

21

B

h

i n N A AP

N B Sco

e c i i OA

i i c

m

i

i

i

1 2

1

Lee, K. T. and Yen, B. C. (1997). Geomorphology and kinematic-wave based hydrograph derivation. J. Hydraulic Engrg., ASCE, 123(1), 73-80.

32

Kinematic-wave-based geomorphologic IUH

( ) exp( ) exp( ) exp( ) ( )o ii

i

x x xo i

w Ww

t t t

T T Tu t a b b P w

If the runoff travel time distribution can be assumed to follow an exponential distribution, then the IUH can be expressed analytically as

1

1 2

i

i

i

mii OA

co

ci i

e c N A APh

N

i

B S

n

1

1 2

2

2i i

i i

i i

i

mo cmi

co coo c i

c

e

xeB L L

h hi n

iT

L S B

The flow travel time in different order of overland areas and channels can be estimated by

Kinematic-wave-based geomorphologic IUH model(KW-GIUH; Lee and Yen, 1997)

33

Runoff simulation in Heng-Chi watershed

Storm event in Oct. 2000 ( no = 0.8, nc= 0.05 )

0 1 2 2 4 3 6 4 8 6 0 7 2

Tim e (h r)

0

1 0 0

2 0 0

3 0 0

4 0 0

Q (

m3/s

)

3 0

2 0

1 0

0

i e (m

m/h

r)

H en g -C h i O c to b e r 2 0 0 0

reco rd ed

K W -G IU H

34

KW-GIUH for soil erosion simulations

Lee, K. T., Yang, C.-C. (2010). Estimation of sediment yield during storms based on soil and watershed geomorphology characteristics. Journal of Hydrology, 382, 145–153.

• sediment travel time in overland areas and channels were derived only based on soil and watershed geomorphologic information, • A geomorphologic instantaneous unit sedimentgraph (GIUS) was obtained for soil erosion simulation during storms.

0 6 12 18 24 30 36 42 48

T im e (h r)

0

5 0

1 0 0

1 5 0

2 0 0

2 5 0

Qs (

kg/s

)

3 0

2 0

1 0

0

i e (m

m/h

r)

S T A 0 1 1 5 F e b . 2 0 0 1

R eco rd edS im u la ted

0 6 12 18 24 30 36 42 48

T im e (h r)

0

1 0 0

2 0 0

3 0 0

4 0 0

Qs (

kg/s

)

6 0

4 0

2 0

0

i e (m

m/h

r)

S T A 0 1 2 8 N o v . 2 0 0 1

R ec o rd edS im u la te d

Goodwin Creek watershed, Mississippi, USA

35

Overland roughness coefficientOverland roughness coefficient (SCS, 1986) (SCS, 1986)

36

Flow attenuation in ungauged reservoir watershedFlow attenuation in ungauged reservoir watershed

Lee, K. T., Chang, C.-H., Yang, M.-S., and Yu, W.-S. (2001). “Reservoir attenuation of floods from ungauged basins,” Hydrological Sciences Journal, 46(3), 349-362.

37

KW-GIUH & HEC-RAS for overbank-flow simulationKW-GIUH & HEC-RAS for overbank-flow simulation

Lee, K. T., Ho, Y.-H., Chyan, Y.-J. (2006). “Bridge blockage and overbank flow simulations using HEC-RAS in the Keelung River during the 2001 Nari typhoon.” J. Hydraulic Engineering, ASCE, 132(3), 319-323.

• Using KW-GIUH model to estimate runoff hydrographs from subwatersheds and lateral flow areas• Using HEC-RAS model to simulate flood wave transport in open channel