43 122010 12

pp. 10111027

DOI: 10.3741/JKWRA.2010.43.12.1011

--

Assessment of Rainfall-Sediment Yield-Runoff Prediction Uncertainty Using a

Multi-objective Optimization Method

* / ** / *** / ****

Lee, Gi Ha / Yu, Wan Sik / Jung, Kwan Sue / Cho, Bok Hwan
..............................................................................................................................................................................................
Abstract
In hydrologic modeling, prediction uncertainty generally stems from various uncertainty sources
associated with model structure, data, and parameters, etc. This study aims to assess the parameter uncertainty effect on hydrologic prediction results. For this objective, a distributed rainfall-sediment yield-runoff
model, which consists of rainfall-runoff module for simulation of surface and subsurface flows and
sediment yield module based on unit
stream power theory, was applied to the mesoscale mountainous area
2
(Cheoncheon catchment; 289.9 km ). For parameter uncertainty evaluation, the model was calibrated by a
multi-objective optimization algorithm (MOSCEM) with two different objective functions (RMSE and
HMLE) and Pareto optimal solutions of each case were then estimated. In Case I, the rainfall-runoff module
was calibrated to investigate the effect of parameter uncertainty on hydrograph reproduction whereas in
Case II, sediment yield module was calibrated to show the propagation of parameter uncertainty into
sedigraph estimation. Additionally, in Case III, all parameters of both modules were simultaneously
calibrated in order to take account of prediction uncertainty in rainfall-sediment yield-runoff modeling.
The results showed that hydrograph prediction uncertainty of Case I was observed over the low-flow
periods while the sedigraph of high-flow periods was sensitive to uncertainty of the sediment yield module
parameters in Case II. In Case III, prediction uncertainty ranges of both hydrograph and sedigraph were
larger than the other cases. Furthermore, prediction uncertainty in terms of spatial distribution of erosion
and deposition drastically varied with the applied model parameters for all cases.
Keywords : distributed rainfall-sediment yield-runoff model, parameter uncertainty, prediction uncertainty,
multi-objective optimization, MOSCEM, pareto optimal solution
..............................................................................................................................................................................................

, ,
. - - --
, MOSCEM
*

post-doc

(e-mail: leegiha@gmail.com)

Corresponding Author, Research Associate., Construction and Disaster Research Center, Chungnam National Univ., Daejeon 305-764, Korea

**

(e-mail: yuwansik@gmail.com)

Graduate student. Dept. of Civil Eng., Chungnam National Univ., Daejeon 305-764, Korea

***

(e-mail: ksjung@cnu.ac.kr)

Professor, Dept. of Civil Eng., Chungnam National Univ., Daejeon 305-764, Korea

****

(e-mail: james7799@rocketmail.com)

Engineer, Div. of Water Resources, Korean Engineering Consultants Corp., Seoul 143-715, Korea

43 12 2010 12

1011

 - , - (Case I II),
-- (Case III) Pareto ,
. (Case I),
, , (Case II)
. --
Case I II ,
.
: -- , , , , MOSCEM, Pareto

.............................................................................................................................................................................................
1.

( )

. (lumped conceptual model)


(integrated)
,

(Sorooshian and Gupta, 1995).
(physics-based distributed model) , ,
,

. Todini (1988), Wheater et al. (1993),
Vieux(2004)
, ,
, Beven(1989,
2003) Madsen(2003) (measurement
scale) (model grid scale) (incommensurability)

.

, (feasible
parameter range)
GA(genetic algorithm;
Wang, 1991) SCE(shuffled complex evolution; Duan
et al., 1992)
(
, 2002; , 2004).
(single-objective optimization method)

1012

, Beven and Binley(1992) Gupta

et al. (1998, 2003)
,


.
Yapo et al. (1998) Vrugt et al. (2003a, 2003b)

Pareto
MOCOM(multi-objective complex
evolution) MOSCEM(multi-objective shuffled complex
evolution metropolis) - SACSMA BATS
. ,
Boyle et al. (2000)
, ,


, (2007) (2007)
NSGA-II Tank

. Tang et al. (2005)
(NSGA-II, MOSCEM, SPEA2)
Pareto
SPEA2 .


, Madsen(2003)
,
MIKE SHE
,


.

, ,


(Ambroise et al., 1995;
Mroczkowski et al., 1997; Franks et al., 2006). ,
- -
,
SCE

,

.

,
(uncertainty propagation)
.
- -
--
MOSCEM - , ,
--
Pareto
, .
2. --

(Tachikawa et al., 2004)
(unit stream power; Yang, 1972)
(Sayama, 2003)
raster --
.
- DEM
, ,


,

- (kinematic
wave) .
(micropore, )
,
(macropore, )
,

43 12 2010 12

.

Eq. (1) , Eq.
(2) .

(1)

(2)
, , , , ,
(m3/s), (m), ,
(m/s),
(m/s), (m),
(m),
(m-1/3s), ( ), 5/3.
- ,
( )
( )
, ( )

.
Yang(1972)
,
. Eq. (3)
, Eq. (4) (net erosion)
. Fig. 1

.

(3)

(4)
, (kg/m3),
(m), (m3/s), (kg/
m2/hr), ,
,
Eqs. (5) and (6) (Morgan et al, 1998).

(5)

(6)

, (kg/J),
1013

(J/m2), ( ,
/ .

( ) Yang(1972)
, ( )

( ) Eq. (7)
.

(7)

Yang(1973) .

log

(8)

, (ppm), (unit
stream power), ,
, m/s),
(m/s),2 (
(m/s ), (mm).
ArcGIS tool
, ,
250250m

3. MOSCEM

log log

log log log

DEM ,
Vieux(2004)
(WAMIS)
Landsat
.
4( , , , ),
5( , , , , ) 9
(Table 1)
Pareto ,
,
.

3.1 Pareto



.



.
Eq. (9)
.

Fig. 1 Flowchart of the Distributed Rainfall-runoff-sediment Yield Model

1014

Table 1. Model Parameters and Feasible Parameter Ranges for Uncertainty Assessment
Parameter

KE

Description
The depth of the unsaturated soil layer (mm)
The depth of the saturated soil layer (mm)
The hydraulic conductivity of the saturated soil layer (m/s)
The non-linear exponent constant for the unsaturated soil layer
The median grain size (mm)
The soil detachability (kg/J)
The detachment or deposition efficiency
The total kinetic energy of the net rainfall (J/m2)
The critical unit stream power (m/s)

(a)

Range of value
50300
1700
0.0010.1
210
110
0.00080.006
0.3351.0
130
0.0020.100

(b)

(c)

Fig. 2 Schematic Diagrams of the MOSCEM Algorithm (Vrugt et al., 2003b)

(9)
, ,
,
. ,

.

,

. Fig. 2(a)
,
Pareto
(Gupta et al., 1998).
Figs. 2(b) and (c) MOSCEM
,
Pareto .
Fig. 2(b) A B
min

, Fig. 2(c)
A B Pareto
. Figs. 2(b) and (c)
, Fig. 2(c)
Pareto rank
. , MOSCEM


, Pareto .
Fig. 2(c) A B A B
()
(),
. Pareto
(group)

(compromise solution or
balanced optimum) .

43 12 2010 12

1015

3.2 MOSCEM
MOSCEM Pareto
SCE
, SCEM(shuffled complex evolution metropolis;
Vrugt et al. 2003a)
, Zitzler and Thiele(1999)
,
SCEM

SCEM (Vrugt et al., 2003b). Fig.
3 MOSCEM
, .
1. ,
( ) .
2. ( )
.
3. ,

.
4.
.
5.

. ,
,
.
6. , SEM(sequence evolution metropolis)


.
SEM
,
.
7.

.
8.
5
.
CMATLAB versionMOSCEMhttp://www.sahra.
arizona.edu
Vrugt et al. (2003b), Bos and Vreng
(2006) . MOSCEM

Pareto ,

Fig. 3. Flowchart of the MOSCEM Algorithm (Vrugt et al., 2003b)

1016

 .
4.

4.1
-
-
. 289.8km2,
EL. 549.13m 25.5km
.

,
- , -
.
(2002, 2003,
2007) - -

. , ,
, 2
Thiessen
. Fig.
4(a)
, Fig. 4(b)
Thiessen .
4.2 MOSCEM Pareto

(a)

MOSCEM
- (Case I), - (Case II)
,
-- (Case III)
2002
Pareto .
Case I -

RMSE(Root Mean Square Error) HMLE
(Heteroscedastic Maximum Likelihood Estimator;
Sorooshian and Dracup, 1980) -
( , , , ) Pareto
. Case II
- Case I
- ( , , ,
, ) RMSE HMLE
Pareto . Case III,
-- ,

RMSE RMSE,
Pareto
.
RMSE HMLE Eqs. (10) and (11) .

(10)

(b)

Fig. 4. (a) Thiessen Polygons of the Yongdam Dam Basin and (b) Hill-shade Topography of the Study
Site: Cheoncheon Catchment Marked by a Bold Solid Line in Fig. 4(a).
43 12 2010 12

1017

, ,
, .
HMLE HMLE
Duan(1991) HMLE .

exp

(11)

, , ln , =0.3.
MOSCEM (algorithmic parameters) ,
5, (random sample) Yapo et al.
(1998)
500 Case I Case II 3000
(iteration) , Case III
Pareto (convergency)
7000 Pareto
.

4.2.1 Case I: -
Fig. 5 - RMSE HMLE
Pareto Pareto
, , , (normalized parameter value) . Fig. 5(a)
Fig. 5(b)

, Pareto
(Fig. 5(b)
) .
Pareto RMSE 37.2147.35,
HMLE 275.37398.20 ,
RMSE
HMLE 40.61 323.71 .
Pareto
RMSE HMLE
Table 2 .

Table 2. Parameter Uncertainty Range and the Compromise Solution in Case I

Parameter

Uncertainty Range
54.4 286.7 (mm)
1 74.3 (mm)
0.004 0.050 (m/s)
2.016 9.970

(a)

Optimal Parameter
RMSE
HMLE
54.4
253.9
56.2
4.2
0.0065
0.047
2.050
9.936

Compromise
Solution
154.3
2.2
0.025
4.394

(b)

Fig. 5. Parameter Uncertainty Assessment: (a) Pareto Solution and (b) Normalized Optimal Parameter
Sets of Case I
1018

4.2.2 Case II: -

-
Case I
-
RMSE
HMLE , , , ,
, Pareto Pareto
,
Fig. 6 .
Pareto RMSE 334.4367.96,
HMLE 52,417.769,113.6
, RMSE HMLE
337.9 59,187.3 -
. Fig. 6(b)

, Franchini
(1996)

(a)

.


. Pareto
, - Pareto
,

. Pareto
RMSE HMLE

Table 3 .

4.2.3 Case III: --

- - ,
Case I Case II Pareto

(b)

Fig. 6. Parameter Uncertainty Assessment: (a) Pareto Solution and (b) Normalized Optimal Parameter
Sets of Case II
Table 3. Parameter Uncertainty Range and the Compromise Solution in Case II
Parameter

43 12 2010 12

Uncertainty Range
9.789.99 (mm)
0.00080.002 (kg/J)
0.3350.401
1.0042.598 (J/m2)
0.09760.1 (m/s)

Optimal Parameter
RMSE
HMLE
9.980
9.987
0.0012
0.001
0.350
0.348
1.775
1.016
0.1
0.0998

Compromise
Solution
9.989
0.001
0.339
1.817
0.0998

1019

. Case III - RMSE , Fig. 7

7000 , Pareto
Case I II

.
(Case I: 4, Case II: 5, Case III: 9)
Pareto
, Case III
Case I II RMSE HMLE
, .

Case II

(a)

4.3


Case Pareto

(ensemble simulation and prediction)
. Case I (
) Fig. 5(b)

(Fig. 8(a)).

(b)

Fig. 7. Parameter Uncertainty Assessment: (a) Pareto Solution and (b) Normalized Optimal Parameter
Sets of Case III
Table 4. Parameter Uncertainty Range and the Compromise Solution in Case III
Parameter

1020

Uncertainty Range
15.48350.6 (mm)
1.58120.99 (mm)
0.00110.0988 (m/s)
2.0459.703
4.6599.998 (mm)
0.00080.0038 (kg/J)
0.35640.9972
10.01523.626 (J/m2)
0.0560.0995 (m/s)

Optimal Parameter
RMSE
HMLE
39.70
158.30
88.64
42.15
0.0046
0.0073
6.653
4.749
9.15
9.97
0.0009
0.0008
0.6127
0.8622
14.282
10.573
0.0914
0.0954

Compromise
Solution
133.09
57.32
0.0123
8.264
9.43
0.0008
0.7588
16.037
0.0969



(equifinality; Beven and Binley, 1992; Savenije, 2001;
Beven, 2006),
.
,
.

2003 2007

Figs. 8(b) and 8(c) . Pareto
,

,

.
Case II
(Fig. 6(b))
Case I
, Fig. 8(d)
,
. ,
Table 3
,

(a) Rusa (2002)

(d) Rusa (2002)

(b) Maemi (2003)

(e) Maemi (2003)

(c) Nari (2007)

(f) Nari (2007)

Fig. 8. Hydrological Uncertainty Ranges Associated with the Pareto Solution Sets: (a)(c) Ensemble
Hydrographs of Case I and (d)(f) Ensemble Sedigraphs of Case II
43 12 2010 12

1021

Figs. 8(e) and 8(f)

. -

-

,
-
.
Case III
, Figs. 9(a)9(c)
3 Case I
,

Case I
.
, 2
Case I
.
, Figs. 9(d)9(f)

, Case II
.
2
,

(a) Rusa (2002)

(d) Rusa (2002)

(b) Maemi (2003)

(e) Maemi (2003)

(c) Nari (2007)

(f) Nari (2007)

Fig. 9. Hydrological Uncertainty Ranges Associated with the Pareto Solution Sets: (a)(c) Ensemble
Hydrographs of Case III and (d)(f) Ensemble Sedigraphs of Case III
1022

 (Fig. 9(f) ).
,



,


.

4.4

Case II III

.

,
(+) (-)
.
Figs. 1012 ,
Case II III Table
5 .
Case II

-1.42+5.34cm ,
0.055cm

. -2.19
+5.57cm ,
0.11cm . ,


. Fig. 10

,
0.23cm ,
0.96cm .
Case II
Figs. 11 and 12 . Figs.
1012
,
Thiessen

Thiessen .
Case III

-8.24+12.75cm
Case II 4 0.24cm .

-59.11+98.87cm
, 1.65cm
. Case III

Fig. 10. Spatial Uncertainty of Erosion and Deposition During the Rusa Event
43 12 2010 12

1023

Fig. 11. Spatial Uncertainty of Erosion and Deposition During the Maemi Event

Fig. 12. Spatial Uncertainty of Erosion and Deposition During the Nari Event
98.1cm,
50.87cm .
Case III
Figs. 11 and 12 .
Case III
Case II
1024

, Case II

Thiessen
.

(drainage area)

Table 5. Summary of Spatial Uncertainty Statistics

Event
Case II
Rusa Case
III
Case II
Maemi Case
III
Case II
Nari Case
III

Maximum Values of Erosion and Deposition

Uncertainty
Min. parameter set
Max. parameter set
Erosion Deposition Average Erosion Deposition Average Erosion Deposition
(cm)
(cm)
(cm) (cm)
(cm)
(cm) (cm)
(cm)
5.34
-1.42
0.055
5.57
-2.19
0.110 0.23
0.96
12.75 -8.24
0.239 98.87 -59.11 1.649 98.1
50.87
1.69
-0.65
0.055
1.81
-1.40
0.063 0.12
0.75
4.59
-6.50
0.138 14.47 -42.61 0.846 9.88
36.11
0.80
-0.91
0.060
0.91
-1.91
0.123 0.15
1.01
2.39
-8.91
0.151 40.87 -61.36 0.978 40.6
52.45

(integrated catchment response)

.


,

.
5.

MOSCEM
-- -
-
, --
Pareto
.

Pareto

,
.
.
1) MOSCEM Pareto
, Case I

Pareto
. Case II
Case I
,

Pareto Case I
. Case III
43 12 2010 12

Pareto
Case I II
.
2) Case I
,

,
.
,



. Case II , Case
I

,
. ,


. Case III

Case I II
,
Case I II .
3)


, Case II III

. ,
,
Thiessen
1025

 .
,
-
,

.

(08F01)

.

, , , (2002).
- .
, , 35, 5,
pp. 541-552.
, , , (2007).
Tank (II):
. ,
, 40, 9, pp. 687-696.
, , , (2007).
Tank (I):
. ,
, 40, 9, pp. 677-685.
, , , (2004).
SIMHYD TANK
. ,
, 37, 2, pp. 121-131.
(2002).
.
(2003).
.
(2007).
.
Ambroise, B., Perrin, J.L., and Reutenauer, D. (1995).
Multicriterion validation of semi-distributed conceptual model of the water cycle in Fecht catchment
(Vosgesmassif, France). Water Resources Research,
Vol. 31, No. 6, pp. 1467-1481.
Beven, K. (1989). Changing ideas in hydrology-The
1026

case of physically-based models. Journal ofHydrology,

Vol. 105, pp. 157-172.
Beven, K., and Binley, A.M. (1992). The future of distributed models: model calibration and uncertainty
prediction. Hydrological Processes, Vol. 6, No. 3,
pp. 279-298.
Beven, K. (2003). On environmental models of everywhere on the GRID. Hydrological Processes, Vol.
17, No. 1, pp. 171-174.
Beven, K. (2006), A manifesto for the equifinality
thesis. Journal of Hydrology, Vol. 320, No. 1-2, pp.
18-36.
Bos, A., and Vreng, A.D. (2006). Parameter optimiza-

tion of the HYMOD model using SCEM-UA and

MOSCEM-UA, Modelling Geo-Ecological Systems

Computational Bio- and Physical Geography Faculty

of Science, University of Amsterdam.
Boyle, D.P., Gupta, H.V., and Sorooshian, S. (2000).
Toward improved calibration of hydrologic models:
Combining the strengths of manual and automatic
methods. Water Resources Research, Vol, 36, No.
12, pp. 3663-3674.
Duan, Q. (1991). A global optimization strategy for

efficient and effective calibration ofhydrologic models,

Ph.D. dissertation, University of Arizona, Tucson.

Duan, Q., Sorooshian, S. and Gupta, V.K. (1992). Effective and efficient global optimization for conceptual
rainfall-runoff models. Water Resources Research,
Vol. 284, pp. 1015-1031.
Franchini, M. (1996). Use of a genetic algorithm
combined with a local search method for automatic
calibration of conceptual rainfall-runoff models.

Hydrological Sciences-Journal-des Sciences Hydrologiques, Vol. 41, No. 1, pp. 21-39.

Franks, S.W., Uhlenbrook, S., and Etchevers, P. (2006).

Hydrological simulation, in Hydrology 2020: An
integrating Science to Meet World Water Challenges, edited by Oki, T., Valeo, C., and Heal, K., pp.
105-122, IAHS Press, Wallingford.
Gupta, H.V., Sorooshian, S., Hogue, T.S., and Boyle,
D.P. (2003). Advances in automatic calibration of
watershed models. in Advances in calibration of
watershed models, edited by Duan, Q., Sorooshian,
S., Gupta, H.V., Rosseau, A. and Turcotte, R., pp.
29-47, AGU, Washington, D.C.

Gupta, H.V., Sorooshian, S., and Yapo, P.O. (1998).

Toward improved calibration of hydrologic models:
Multiple and noncommensurable measures of information. Water Resources Research, Vol. l34, No.
4, pp. 751-763.
Madsen, H. (2003). Parameter estimation in distributed
hydrological catchment modelling using automatic
calibration with multiple objectives. Advances in
Water Resources, Vol. 26, pp. 205-216.
Morgan, R.P.C., Quinton, J.N., Smith, R.E., Govers, G.,
Poesen, J.W.A., Chisci, G., and Torri, D. (1998). The
European soil erosion model(EUROSEM): a dynamic
approach for predicting sediment transport from
fields and small catchments. Earth Surface Process
and Landforms, Vol. 23, pp. 527-544.
Mroczkowski, M., Raper, G.P., and Kuczera, G. (1997).
The quest for more powerful validation of conceptual
catchment models. Water Resources Research, Vol.
33, No. 10, pp. 2325-2336.
Savenije H.H.G. (2001), Equifinality, a blessing in
disguise? Hydrological Processes, Vol. 15, pp. 28352838.
Sayama. (2003). Evaluation of reliability and complexity of rainfall-sediment-runoff models. Master's
Thesis, Kyoto University, pp. 5-10.
Sorooshian, S., and Dracup, J.A. (1980), Stochastic
parameter estimation procedures for hydrologic rainfall-runoff models: Correlated and heteroscedastic
error cases. Water Resources Research, Vol. 16,
No. 2, pp. 430-442.
Sorooshian, S., and Gupta, V.K. (1995). Model calibration, in computer models of watershed hydrology.
edited by Singh, V.P., pp. 23-68, Water Resources
Publications, Highland Ranch.
Todini, E. (1988). Rainfall-runoff modeling-past, present
and future. Journal of Hydrology, Vol. 100, pp. 341352.
Tachikawa, Y., Nagatani, G., and Takara, K. (2004).
Develpment of stage-discharge relationship equation
incorporating saturated-unsaturated flow mechanism.
Annual Journal of Hydraulic Engineering, JSCE,
Vol. 48, pp. 7-12. (Japanese with English abstract)
Tang, Y., Reed, P., and Wagener, T. (2005). How

43 12 2010 12

effective and efficient are multiobjective evolutionary

algorithms at hydrologic model calibration. Hydrology
and Earth System Sciences Discussions, Vol. 2, pp.
2465-2520.
Vieux, B.E. (2004). Distributed hydrologic modeling
using GIS, Kluwer Academic Publishers.
Vrugt, J.A., Gupta, H.V., Bouten, W., and Sorooshian,
S. (2003a). A shuffled complex evolution metropolis
algorithm for optimization and uncertainty assessment
of hydrologic model parameters. Water Resources
Research, Vol. 39, No. 8, 1201, doi:10.1029/2002WR
001642.
Vrugt, J.A., Gupta, H.V., Bastidas, L.A., Bouten, W.,
and Sorooshian. S. (2003b). Effective and efficient
algorithm for multi objective optimization of hydrologic models. Water Resources Research, Vol. 39,
No. 8, 1214, doi:10.1029/2002WR001746.
Wheater H.S., Jakeman, A.J., and Beven, K. (1993).
Progress and directions in rainfall-runoff modeling.
in Modeling change in environmental systems, edited
by Jakeman, A.J., Bech, M.B., and McAleer, M.J., pp.
101-132, John Wiley and Sons, Chichester.
Wang, Q.J. (1991). The genetic algorithm and its
application to calibrating conceptual rainfall-runoff
models. Water Resources Research, Vol. 27, No. 9,
pp. 2467-2471.
Yapo, P.O., Gupta, H.V., and Sorooshian, S. (1998).
Multi-objective global optimization for hydrologic
models. Journal of Hydrology, Vol. 204, pp. 83-97.
Yang, C.T. (1972). Unit stream power and sediment
transport. Journal ofthe Hydraulics Division, ASCE,
Vol. 98, No. HY10, pp. 1805-1826.
Yang, C.T. (1973). Incipient motion and sediment
transport. Journal ofthe Hydraulics Division, ASCE,
Vol. 99, No. HY10, pp. 1679-1704.
Zitzler, E., and Thiele, L. (1999). Multi-objective
evolutionary algorithm: A comparative case study
and the strength pareto approach. IEEE Transactions on Evolutionary Computation, Vol. 3, No. 4, pp.
257-271.
: 10-084
: 2010.10.04
: 2010.11.11
: 2010.11.11

1027