Journal of Hydrology 394 (2010) 370–383

Contents lists available at ScienceDirect

Journal of Hydrology
journal homepage: www.elsevier.com/locate/jhydrol

An evaluation of impacts of DEM resolution and parameter correlation

on TOPMODEL modeling uncertainty
Kairong Lin, Qiang Zhang ⇑, Xiaohong Chen
Department of Water Resources and Environment, Sun Yat-sen University, Guangzhou 510275, China
Key Laboratory of Water Cycle and Water Security in Southern China of Guangdong High Education Institute, Sun Yat-sen University, Guangzhou 510275, China

a r t i c l e i n f o s u m m a r y

Article history: Hydrological modeling uncertainties are the results of many factors such as input error, calibration accu-
Received 25 March 2010 racy, parameter uncertainty, model structure, and so on. Wherein, input errors and parameter uncertain-
Received in revised form 2 September 2010 ties are the two of the major factors influencing the uncertainties of hydrological modeling. TOPMODEL is
Accepted 17 September 2010
a rainfall–runoff model that bases its distributed predictions on analysis of watershed topography, which
This manuscript was handled by Geoff
is widely used in hydrological modeling practices. In this study, the effects of DEM resolution and param-
Syme, Editor-in-Chief, with the assistance of eter correlation on TOPMODEL modeling uncertainties are evaluated by using GLUE technique. The
Soroosh Sorooshian, Associate Editor uncertainty evaluation is performed by modeling the rainfall–runoff processes of three tributaries in
the Hanjiang River, one of the major tributaries of the Yangtze River, China. The results show no evident
Keywords: effects of the DEM resolution on the uncertainty intervals of the TOPMODEL simulation. This can be
Modeling uncertainty attributed to the fact that the modeling uncertainty is due solely to changes of DEM resolution by fixing
GLUE the parameter values to avoid the artifacts resulted from interactions between ln(a/tan(B)) and the
DEM resolution parameters. In addition, the copula functions are used to produce more behavioral parameter sets for
Parameter correlation the same sample time intervals when the model parameters are in good correlation, and which can ben-
TOPMODEL efit thorough evaluation of effects of parameter correlation on the hydrological modeling uncertainty.
With the same number of the behavioral parameter sets, after putting the parameter correlation under
consideration, the simulated runoff series by the TOPMODEL with the behavioral parameter sets can
fit reasonably better the observed runoff series. Thus, the uncertainty due to parameter correlation of
the TOPMODEL modeling can be considerably removed. This study is of great theoretical and practical
merits in sound understanding of the modeling behaviors of the TOPMODEL under the influences of
inputs and parameter correlation.
Ó 2010 Elsevier B.V. All rights reserved.

1. Introduction
Hydrological models have been widely used to investigate
many practical and pressing issues that arise during planning, de-
sign, operation, and management of water resources systems
(Benke et al., 2007; Lin et al., 2007). The crucial step for hydrolog-
ical modeling is to identify values of model parameters and this
procedure is also referred to as calibration procedure (Sorooshian
and Gupta, 1995). Estimation and specification of parameters are
the two important procedures for calibration of hydrological
model. However, the parameter redundancy and correlations be-
tween parameters result in universal equifinality in modeling
behaviors of the hydrological models (Beven and Binley, 1992).
Accordingly the generalized likelihood uncertainty estimation
(GLUE) method proposed by Beven and Binley (1992) is devoted
to the investigation of the hydrological modeling uncertainty by
⇑ Corresponding author. Tel./fax: +86 20 84113730.
E-mail address: zhangq68@mail.sysu.edu.cn (Q. Zhang).
generating the prediction limits for the modeled streamflow series
and a set of behavioral parameter sets (Freer et al., 1996; Blazkova
and Beven, 2002; McMichael et al., 2006; Montanari, 2005, 2007;
Yang et al., 2007, 2008; Xiong and O’Connor, 2008; Jin et al., 2010).
The hydrological system is complicated, being affected by the
climate changes such as atmospheric circulation, precipitation,
air temperature, the underlying surface properties such as the geo-
logical conditions, vegetation and soil conditions, and also human
activities such as water reservoirs and land use changes (Zhang
et al., 2009, 2010). Generally, a hydrological model consists of a
large number of mathematical equations describing changing
properties of hydrological processes, e.g. streamflow series, and
estimating the streamflow variations of the future. Additional in-
puts represent the spatial mosaic of climate, soil type, topography
and land use (Benke et al., 2007). Topography was taken as an
important factor in the evaluation of the hydrological responses
of the upland and forested watersheds to precipitation changes
(e.g. Beven and Wood, 1983; Wolock and Price, 1994) due to the
effects of gravity on the movement of water in a watershed. Many

0022-1694/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.jhydrol.2010.09.012
 K. Lin et al. / Journal of Hydrology 394 (2010) 370–383
researches found that topography can influence many aspects of
the hydrological system (Wolock and Price, 1994; Zhang and
Montgomery, 1994; Wolock and McCabe, 1995; Sørensen and
Seibert, 2007). TOPMODEL (Beven and Kirkby, 1979) is a rainfall–
runoff model that bases its distributed predictions on an analysis
of watershed topography in a semi-distributed way. Thus, the spa-
tial distribution of topographical properties extracted by digital
elevation model (DEM) should be firstly identified for the sake of
hydrological modeling with TOPMODEL. Previous studies have
shown that DEM resolution has the potential to influence the
spatial patterns of the topographic index ln(a/tan B) and thus the
TOPMODEL simulation results. Quinn et al. (1995) indicated that
different DEM resolutions can cause different spatial patterns of
the ln(a/tan B). Zhang and Montgomery (1994) showed that the
mean of the ln(a/tan B) distribution increased as data resolution
became coarser. Wolock and Price (1994) showed that model pre-
dictions of the depth to the water table, the ratio of overland flow
to total flow, peak flow, and variance and skew of predicted
streamflow were affected by both the DEM map scale and data res-
olution. Further analyses showed that the effects of DEM map scale
and data resolution on model predictions should be attributed to
the sensitivity of the predictions to the mean of the ln (a/tan B) dis-
tribution, which was affected by both DEM map scale and data res-
olution. Thus, DEM resolution has the potential to affect the
TOPMODEL simulations. In this case, one objective of this study
is to clarify whether or not hydrological modeling uncertainty
could be affected by the DEM resolutions.
Prediction uncertainty is the result of various factors such as
input error, calibration accuracy, parameter correlation, model
structure, and so on. Beven and Freer (2001a) attempted to ad-
dress the effects of some factors on the modeling uncertainty,
such as model nonlinearity, covariation of parameter values and
errors in model structure, input data or observed variables, using
the GLUE procedure. Blasone and Vrugt (2008) and Yang et al.
(2008) found that parameter correlation can result in the hydro-
logical modeling uncertainty. Recent years have witnessed an
explosion of methods devoted to derive meaningful uncertainty
bounds for hydrological model predictions. Methods aiming to
represent model parameter, state and prediction uncertainty in-
clude classical Bayesian (Kuczera and Parent, 1998; Thiemann
et al., 2001; Vrugt et al., 2003), pseudo-Bayesian (Beven and
Binley, 1992; Freer et al., 1996), set-theoretic (Keesman, 1990),
multiple criteria (Gupta et al., 1998; Madsen, 2003), sequential
data assimilation (Vrugt et al., 2005; Moradkhani et al., 2005),
and multi-model averaging methods (Ajami et al., 2007; Vrugt
and Robinson, 2007). Generally, it is assumed that the parameters
of the model were independent mutually. In fact, complexity and
correlation within the parameter space are the two important fac-
tors having the potential to cause hydrological modeling uncer-
tainty. With this in mind, another objective of this study is to
address the effects of parameter correlation on the hydrological
modeling uncertainty.
With the help of GLUE technique, this study attempts to discuss
the influences of parameter correlation and DEM resolution on the
hydrological modeling uncertainty by taking TOPMODEL as the
case model. Undoubtedly, this study will be of theoretical and
practical merits in obtaining deep insight into the causes behind
the hydrological modeling uncertainty, one of the crucial but tough
problems in the hydrological modeling practices. This paper is or-
ganized as the follows: Section 2 briefly describes the basic equa-
tion and parameters of TOPMODEL; In Section 3, we introduce
three catchments and related hydrological data analyzed in this
study; Section 4 is devoted to analyze the effects of grid DEM on
uncertainty of TOPMODEL; and Section 5 discusses the effect of
parameter correlation on uncertainty of TOPMODEL. Finally, the
last section contains the major conclusions.
371
2. Methodology
2.1. Basic equations and parameters of TOPMODEL
Since Beven and Kirkby firstly proposed TOPMODEL in 1979,
many changes, some minor and some substantial, have been made
over the past 20 years (e.g. Beven and Wood, 1983; Beven et al.,
1995; Duan and Miller, 1997; Scanlon et al., 2000; Beven and Freer,
2001b; Walter et al., 2002; Xiong and Guo, 2004). The 1995 version
(Beven et al., 1995) is subjected to the most frequent application
and assessment (Beven, 1997) and is recognized widely. Therefore,
the 1995 version of TOPMODEL is referred to as the original version
in this study.
The structure of the original version of TOPMODEL is shown in
Fig. 1, which shows that the total runoff is generally the sum of two
major flow components: saturated excess overland flow from var-
iable contributing areas and subsurface flow from the saturated
zone. The infiltration excess overland flow component can also
be included based on the properties of soil and rainfall of the river
basin (Beven et al., 1995). Basic equations of this version are listed
as the follows.
The actual evaporation rate, E, is calculated by:
 
Srz
Ea ¼ Ep 1  ð1Þ
SRmax
where Srz and SRmax are the root zone storage deficit and maximum
allowable root storage deficit, respectively; Ep is the reference or po-
tential evaporation rate.
The precipitation that falls over the root zone in saturation state
forms the unsaturated store, and is transferred (unsaturated flow
or recharger) to the saturated store at a rate proportional to the
depth of the unsaturated store (Suz) and inversely proportional to
both the local saturated deficit (SD) and the recharger delay
parameter (Td). It can be expressed as:
Suz
qv ¼ ð2Þ
SD  T d
where SD is the local saturated deficit due to gravity drainage and
depends on the depth of the local water table (z).
Another fundamental equation shows the relation of local
transmissivity T(Zi) to the groundwater table depth Zi, that is,
 
Z i
TðZ i Þ ¼ T 0 exp ð3Þ
Szm
where T0 is the transmissivity of the soil in the saturated state and
Szm is the maximum moisture deficit.
The original TOPMODEL has four parameters, i.e. the maximum
allowable root storage deficit (SRmax), the transmissivity of the soil
in saturated state (T0), the maximum moisture max deficit (Szm),
and the recharger delay parameter (Td). All of these parameters
must be optimized. Ranges of parameters in TOPMODEL for Monte
Carlo simulations are listed in Table 1. Ranges for T0 are shown in
log in relation to the graph scales (Beven and Freer, 2001a).
2.2. The indices selected for uncertainty evaluation
Uncertainty interval at each time step is the major result by the
GLUE technique in terms of evaluations of the hydrological model-
ing uncertainty. In this study, three indices, i.e. containing ratio
(CR), interval width (IW), and the Nash–Sutcliffe efficiency index
(R2), are adopted aiming to evaluate the uncertainty interval. The
definitions of these three indices are introduced as the follows:
Containing ratio (CR) is devoted to estimate the capability of the
uncertainty intervals to capture the observed values, which is de-
fined as the ratio of the number of the observations falling within
372 K. Lin et al. / Journal of Hydrology 394 (2010) 370–383

Fig. 1. Structure of the TOPMODEL.

Table 1
Ranges of parameters used in TOPMODEL model.

Parameter Unit Physical meaning Minimum value Maximum value Sampling strategy
Szm m Maximum moisture deficit 0.01 1 Uniform
T0 m2 h1 Transmissivity of the soil at saturated 2.33a 3.4a Uniformb
Td h Time parameter 1 20 Uniform
SRmax m Maximum allowable root storage deficit 0.01 0.3 Uniform
a
Ranges for T0 are shown in log in relation to the graph scales.
b
The sampling method for ln(T0).

their respective uncertainty intervals to the total number of obser- Copula functions have good performance in description of correla-
vations (Beven and Binley, 1992; Montanari, 2005; Xiong and tion structure of variables. In addition, one of the advantages of
O’Connor, 2008). The CR is formulated as: Copula functions is that the marginal distributions are beyond nor-
Pn mal and extreme value distributions. Gumbel–Hougaard Copula is
i¼1 J½Q obs ðiÞ another kind of Copula function having been taken as the right
CR ¼ ð4Þ
n technique in the description of the joint probability behaviors of
where the hydrological variables (Salvadori and De Michlele, 2004; Zhang
 and Singh, 2006). This study uses Gumbel–Hougaard Copula to de-
1; Q low ðiÞ < Q obs ðiÞ < Q up ðiÞ scribe the correlative structure of the model parameters, and can
J½Q obs ðiÞ ¼ ð5Þ
0; otherwise be expressed as
Qobs(i), Qlow(i) and Qup(i) in Eqs. (4) and (5) denote the observed run- Cðu; v Þ ¼ expf½ð ln uÞh þ ð ln v Þh 1=h g; h1 ð8Þ
off, the lower and the upper uncertainty intervals at time i respec-
tively; and n is the length of the observed series. where u and v represent two random variables U and V, and C is the
Interval width (IW) is defined as the average width of the uncer- copula function, a kind of multivariate distribution function, and its
tainty intervals (Choi and Beven, 2007; Blasone and Vrugt, 2008), one-dimensional margins are uniform on the interval (0, 1), and h is
which can be calculated by the parameter of Copula. Larger h value implies stronger correlation,
Pn and vice versa. h = 1 denotes independency of variables.
i¼1 ½Q up ðiÞ  Q low ðiÞ
IW ¼ ð6Þ The parameter of Copula, h, is related to the Kendall’s rank cor-
n relation coefficient s by 1/(1  s). Once an estimate s ^ of s is ob-
The Nash–Sutcliffe efficiency index (R2) (Nash and Sutcliffe, 1970) is tained, an estimate ^ h of h can be calculated. In the CGLUE
used to judge whether or not the median values MQ0.5 and the technique, correlative parameter sets can be generated directly
uncertainty intervals are effective crisp simulations of the observa- by the Copula function.
tions, which is defined as: The theory behind the link between copula and bivariate distri-
Pn butions was addressed by Sklar’s theorem (Sklar, 1959): Let H be a
2
i¼1 ½Q obs ðiÞ  MQ 0:5 ðiÞ joint distribution function with margins F and G. Then there exists
R2 ¼ 1:0  Pn 2
ð7Þ
i¼1 ½Q obs ðiÞ  Q obs 
a copula C for x, y,

where Q obs is the mean value of the observed runoff series. Hðx; yÞ ¼ CðFðxÞ; GðyÞÞ ð9Þ

2.3. Copula function
In this study, a method based on the GLUE technique (CGLUE) is
adopted using the Copula function to describe the correlations
amongst parameters (e.g. Lin et al., 2009). Copula technique, after
it was proposed in 1959, has been a kind of important method ded-
icated to analyze the multivariate joint distribution (Nelson, 1999).
By Eqs. (8) and (9), the bivariate joint distribution H(x, y) of X and Y
can be written as
Hðx; yÞ ¼ expf½ð lnðFðxÞÞÞh þ ð lnðGðyÞÞh 1=h g ð10Þ
The interesting point is that the properties of H(x, y) can be dis-
cussed in terms of the structure of C. In fact, it is precisely the cop-
ula that captures many of the features of a joint distribution, and
 K. Lin et al. / Journal of Hydrology 394 (2010) 370–383
GLUE
Select model and determine
the range of each parameter
Select likelihood function, L(θ ) , and set a
threshold of likelihood function, choose a
desired % of observations to be contained
in the uncertainty bounds
Monte Carlo Sampling with independent parameters
Repeatedly run the model, obtain the
behavioral parameter sets and the
posterior distribution of parameters
Generate the uncertainty bounds
according to the likelihood weights
Based on the behavioral parameter sets, analyze the
parameters correlation, and calculate the correlative
parameters, the Kendall’s τ , and the parameter of Copula, θ
Generate the independent model parameter sets with
Monte Carlo method, and generate another correlative
parameter sets directly by Copula function
Fig. 2. Calculating flowchart of the GLUE and CGLUE techniques.
measures of association and dependence properties between ran-
dom variables.
2.4. The CGLUE technique
The main difference between the GLUE technique and CGLUE
technique is that the CGLUE technique uses copula function to de-
scribe the parameter correlation. In the CGLUE technique, the ori-
ginal behavioral parameter sets is obtained by the GLUE technique
firstly base on these behavioral parameter sets, the parameters cor-
relation are analyzed, and the correlative parameters, the Kendall’s
s, and the parameter of Copula, h, are calculated. Then, the inde-
pendent model parameter sets are generated with Monte Carlo
method, and correlative parameter sets are generated directly by
the Copula function. The calculating procedures of the GLUE and
CGLUE technique are illustrated in Fig. 2.
3. Study regions and data
3.1. River basins
In this study, three river basins, i.e. Mumahe, Xunhe, and Laogu-
anhe, located in the Hanjiang River basin (Fig. 3) are accepted as
case studies. Hanjiang River is one of the largest tributaries of
the Yangtze River basin. The importance of this river basin lies in
the fact that the Danjiangkou reservoir in the Hanjiang River is
the headwater of the middle route of the South-to-North Water
Transfer Project of China (SNWTP), bearing the heavy responsibil-
373
ity for the water supply for the north China (Fig. 3). The basin is
dominated by the subtropical monsoon climate and hence the
water resource is subject to large variability. Flood occurs quite of-
ten in the Hanjiang River basin and effective water resource man-
agement is a must in this river basin. Daily rainfall and runoff data
of these three river basins covering 1978 and 1987 are analyzed in
this study. The geographical properties and data in these three riv-
er basins are introduced briefly as the follows:
Mumahe River basin is 1224 km2 in drainage area. The topo-
graphical properties of this river are characterized by low hills
and basins with moderate slope. Mumahe River basin is the largest
basin in the south bank of the Hanjiang River.
Xunhe River basin is 6448 km2 in drainage area. The topography
is dominated by mountains with the altitude of the mountain peak
between 2000 m and 3000 m. The vegetation coverage of the Xun-
he River basin is good and the soil erosion is not serious.
Laoguanhe River basin is 3219 km2 in drainage area. The upper
Laoguanhe River is dominated by hill landforms, and the lower
Laoguanhe River is the Danjiangkou Reservoir. The slope of the ter-
rain of this river basin is steep. Due to the frequent rainstorms and
steep terrains, soil erosion is serious.
3.2. DEM data
In this study, DEM data are extracted from topographic maps of
the Chinese State Bureau of Surveying and Mapping (with map
scale of 1:50, 000, about 25 m per grid). Then the 50 m, 100 m,
200 m, 400 m, and 800 m grid DEMs are resampled by using ArcGIS
software package.
Table 2 shows the topographic characteristics calculated from
different DEM resolutions. It can be seen from Table 2 that decreas-
ing DEM resolution tends to decrease the average slopes, surface
area/2D area, and volume/2D area, and which should be attributed
to the loss of detailed topographic characteristics as a result of
decreasing DEM resolutions. Surface roughness and elevation vari-
ations are used to denote the relief degree of land surfaces. Coarser
DEM resolution can cause larger surface roughness and elevation
variations. Average and maximum flow concentration path lengths
are the average and maximum flow concentration path lengths of
all grids in the catchment, being related to the size, shape and land-
forms of the river basin considered. The flow directions on each
grid are changing with different DEM resolutions, having the im-
pacts on the flow concentration paths on grids (Table 2). These re-
sults and previous studies clarified that the extracted topographic
characteristics hinge on the DEM resolutions.
4. Results
4.1. Effects of DEM resolution on uncertainty of TOPMODEL
4.1.1. ln (a/tan B) distribution
TOPMODEL derives a topographic index by ln (a/tan B), which is
used to identify the hydrological similarities of points in a catch-
ment (Beven et al., 1995). The ln (a/tan B) distribution is an indis-
pensable input for TOPMODEL, which can be extracted from DEM.
Previous studies have shown that the effects of DEM resolution on
model predictions were due to the sensitivity of the predictions to
the mean of the ln (a/tan B) distribution (Wolock and Price, 1994).
Thus, it is necessary to understand the effects of DEM resolution on
ln (a/tan B) distribution in the study regions before investigation of
the effects of DEM on uncertainty of TOPMODEL simulations. The
ln (a/tan B) distributions are calculated by using multiple flow
direction algorithm using Eq. (11) based on different DEM resolu-
tions in three river basin considered in this study.
lnða= tan BÞ ¼ ln½A=ðC  tan BÞ ð11Þ
374 K. Lin et al. / Journal of Hydrology 394 (2010) 370–383

Fig. 3. Locations of the study region, the Hanjiang River basin and three major tributaries considered in this study, i.e. Mumahe, Xunhe and Laoguanhe Rivers. SNWTP denotes
South-to-North Water Transfer Project of China.

Table 2
Topographic characteristics of different DEM resolutions.

Catchment DEM Topographic characteristics DEM resolution

25 m 50 m 100 m 200 m 400 m 800 m
Mumahe Average slope (%) 44.90 40.50 33.61 26.27 19.68 14.51
Surface area/2D area (km2/km2) 1.1337 1.1204 1.0984 1.0716 1.0466 1.0290
Volume/2D area (km3/km2) 0.5936 0.5923 0.5901 0.5854 0.5778 0.5688
Surface roughness 2.39 6.35 13.86 25.15 42.03 67.60
Elevation variation 0.010 0.018 0.030 0.046 0.069 0.103
Average flow concentration path length (km) 47.4 46.7 46.15 43.01 39.56 37.78
Maximum flow concentration path length (km) 85.21 84.42 83.77 78.51 80.19 69.41
Xunhe Average slope (%) 64.72 59.96 51.34 40.18 29.04 19.69
Surface area/2D area (km2/km2) 1.2246 1.1954 1.1528 1.1013 1.0562 1.0267
Volume/2D area (km3/km2) 0.9759 0.9756 0.9747 0.9718 0.9640 0.9345
Surface roughness 3.02 8.23 19.62 39.16 66.86 102.93
Elevation variation 0.013 0.023 0.041 0.066 0.098 0.135
Average flow concentration path length (km) 104.76 103.93 101.13 93.82 84.43 82.88
Maximum flow concentration path length (km) 184.32 182.98 176.71 163.02 145.22 143.12
Laoguanhe Average slope (%) 53.20 47.56 38.63 29.09 20.99 14.42
Surface area/2D area (km2/km2) 1.1674 1.1489 1.1168 1.0781 1.0457 1.0245
Volume/2D area (km3/km2) 0.7246 0.7238 0.7226 0.7196 0.7154 0.7056
Surface roughness 2.92 7.89 17.32 30.81 48.93 74.89
Elevation variation 0.013 0.023 0.039 0.060 0.089 0.126
Average flow concentration path length (km) 75.89 75 73.07 69.04 51.59 45.73
Maximum flow concentration path length (km) 181.19 179.1 171.3 156.71 123.44 112.35

where ln is the Napierian logarithm, a is the upslope area per unit
contour length, and tan(B) is the slope gradient. The DEMs are used
to calculate a and tan(B) for each grid cell in the watershed. To do so
required the calculation of the total area draining into each cell (A),
as well as the contour length (C) and slope gradient (tan B) along
which drainage from the cell occurs (Wolock and McCabe, 1995).
The results are illustrated in Fig. 4 and Table 3. It can be ob-
served from Fig. 4 and Table 3 that all statistics of the ln (a/tan B)
distribution are heavily affected by DEM resolutions. Coarser DEM
resolution tends to cause greater mean ln(a/tan(B)) values within
three catchments considered in this study, being represented by
increased areas covered by higher ln(a/tan(B)) values. Coarser
DEM resolution causes larger mean variance of the ln(a/tan B) with
smaller variation coefficients. Thus, it is evident that coarser grid
resolution of the DEM tends to increase absolute dispersion degree,
but tends to decrease relative dispersion degree. The coarser the
DEM resolution is, the smaller the skewness coefficient of the
ln(a/tan(B)), implying that the asymmetry of the ln(a/tan(B)) distri-
bution tends to be smaller.
4.2.1. Effects on uncertainty of TOPMODEL
An implicit assumption in using topography data in TOPMODEL
is that the water table configuration is related to the land surface
topography. Previous study (e.g. Wolock and Price, 1994) has
shown that it should not be concluded that coarse resolution is
inappropriate source of topographic information for TOPMODEL.
In the current study, we investigate how the DEM resolution af-
fects the efficiency of uncertainty interval in TOPMODEL simula-
tion. Three indices, i.e. containing ratio (CR), interval width (IW),
and the Nash–Sutcliffe efficiency of the median MQ0.5, are selected
to evaluate the efficiency of uncertainty. For the datasets consid-
ered in this paper, the GLUE methodology is applied for the likeli-
hood function selected as the global Nash–Sutcliffe efficiency R2
 K. Lin et al. / Journal of Hydrology 394 (2010) 370–383 375

0.25 Mumahe 25m 0.3 Xunhe 25m 0.3 Laoguanhe

0.2 50m 50m 25m
100m 0.2 100m 0.2 50m

Ac/A

Ac/A
Ac/A
0.15 200m 100m
200m
0.1 400m 400m 200m
800m 0.1 0.1 400m
0.05 800m
800m
0 0 0
0 5 10 15 20 0 5 10 15 20 25 0 5 10 15 20 25
ln (a/tanB) ln (a/tanB) ln (a/tanB)

Fig. 4. Relations between topographical index distribution and DEM resolution.

Table 3
Statistical characteristics of topographical index corresponding to different DEM resolutions.

Catchment Statistic 25 m 50 m 100 m 200 m 400 m 800 m

Mumahe Mean 5.06 5.74 6.53 7.05 7.78 8.38
Mean variance 1.33 1.51 1.59 2.08 2.3 2.36
Cv 0.26 0.26 0.24 0.30 0.30 0.28
Cs 1.57 1.63 1.54 0.29 0.28 0.32
Xunhe Mean 4.97 5.67 6.36 7.06 7.93 8.81
Mean variance 1.48 1.71 1.73 1.83 1.9 2.01
Cv 0.30 0.30 0.27 0.26 0.24 0.23
Cs 1.67 1.84 1.76 0.76 0.48 0.15
Laoguanhe Mean 5.06 5.69 6.45 7.05 7.79 8.58
Mean variance 1.53 1.68 1.74 1.93 2.00 2.23
Cv 0.30 0.30 0.27 0.27 0.26 0.26
Cs 1.63 1.62 1.55 0.23 0.01 0.15

Table 4
Assessing indices of uncertainty for different DEM resolutions in the GLUE technique.

Catchment Indices 25 m 50 m 100 m 200 m 400 m 800 m

Mumahe CR 0.636 0.634 0.635 0.638 0.639 0.637
IW (m3/s) 30.487 29.684 30.431 32.069 32.758 31.857
R2 (MQ0.5) (%) 79.560 79.440 79.460 79.160 79.050 79.020
Xunhe CR 0.652 0.651 0.652 0.658 0.657 0.652
IW (m3/s) 58.836 58.062 58.361 62.704 62.159 60.767
R2 (MQ0.5) (%) 76.820 76.820 76.880 76.630 76.460 76.740
Laoguanhe CR 0.646 0.645 0.645 0.652 0.654 0.648
IW (m3/s) 33.477 33.158 33.802 35.808 36.112 34.848
R2 (MQ0.5) (%) 76.790 76.850 76.870 76.650 76.480 76.660

(Nash and Sutcliffe, 1970). The analysis uses 50,000 parameter sets
generated by the uniform Monte Carlo sampling schemes. The
uncertainty intervals at a given confidence level of 90% are ob-
tained with the GLUE technique by setting the threshold value,
R2, as 70%. Table 4 displays the results of GLUE-based evaluation
of the uncertainties of the TOPMODEL with respect to different
DEM resolutions. R2 (MQ0.5) in Table 4 represents the Nash–
Sutcliffe efficiency of the median MQ0.5 produced from the
GLUE-based analysis by fitting the observed runoff series. The vari-
ations of the containing ratio (CR), and interval width (IW) with
different the DEM resolutions are demonstrated in Fig. 5.
It can be seen from Table 4 and Fig. 5 that changes of the con-
taining ratio (CR), and interval width (IW) are not significant when
the DEM resolutions are changing, showing no evident influences
of DEM resolutions on the uncertainty intervals of TOPMODEL sim-
ulations. It should be noted here that DEM resolution can affect the
topographic characteristics and the ln(a/tan(B)) distribution, how-
ever the effects of changing DEM resolutions on the ln(a/tan(B))
distribution can be compensated by the proper adjustments of
the parameters with sampling techniques. Availability of numer-
ous parameter sets can greatly reduce the impacts of DEM resolu-
tions on the modeling uncertainty, and the similar modeling
results by hydrological models with different sets of parameters
is also called as equifinality. Figs. 6–8 illustrate comparisons
amongst the posterior distributions of parameters with respect to
different DEM resolutions in the Mumahe, Xunhe, and Laoguanhe
River basins respectively. The numbers 1–6 showing the DEM res-
olutions in Figs. 6–8 denote 25 m, 50 m, 100 m, 200 m, 400 m, and
800 m respectively. Figs. 6–8 show that the posterior distributions
of parameters are affected by DEM resolutions, but this effect is rel-
atively minor. The major influence of DEM resolution on the pos-
terior parameter distributions is the difference of proration of
parameter values. The modeling uncertainty should be due solely
to DEM resolution changes by fixing the parameter values to avoid
the artifacts resulted from interactions between ln(a/tan(B)) and
the parameters. Besides, it should be noted here that Table 4 indi-
cates larger IW of the Xunhe basin compared to the other two ba-
sins, which should be due to larger drainage area of the Xunhe
basin when compared to other two river basins considered in this
study. Correspondingly, the discharge of the Xunhe River basin is
larger than that of the other two basins. The same phenomenon
can also be identified in Tables 5–7.
5. Effects of parameter correlation on uncertainty of TOPMODEL
5.1. Parameter correlation in TOPMODEL
Complexity and correlations of the parameter space are the
two important factors that can trigger the parameter uncertainty.
376 K. Lin et al. / Journal of Hydrology 394 (2010) 370–383

0.66 70

0.65 60
Mumahe

IW (m3/s)
0.64 Xunhe
50

CR
0.63 Laoguanhe
Mumahe 40
0.62 Xunhe
Laoguanhe 30
0.61
0.60 20
0 200 400 600 800 0 200 400 600 800
DEM Resolution (m) DEM Resolution (m)

Fig. 5. Relations between index of uncertainty and DEM resolution.

(a) Szm posterior distribution (b) ln (T ) posterior distribution

0
0.12
0.08
Parameter distribution

Parameter distribution
0.1
0.06
0.08

0.06
0.04
0.04
0.02
0.02

0 0
0 −3
0.2
−2
0.4
−1
Pa 0.6 Pa
ram ram 0
ete 0.8 ete
r ra 1 r ra 1
ng ng
e 1.2 e 2
56 56
34 ion
1.4 12
re solut 3 12
34
lution
DEM reso
DEM

(c) Td posterior distribution (d) SRmax posterior distribution

0.12 0.12
Parameter distribution

Parameter distribution

0.1 0.1

0.08 0.08

0.06 0.06

0.04 0.04

0.02 0.02

0 0
0 0
5 0.05
0.1
Pa
ram10 Pa
ram0.15
ete 15 ete 0.2
r ra r ra
ng ng0.25
e 20 e
56 0.3 56
25 12
34
eso lution 34
1 2 esolutio
n
EM r
0.35
D DE Mr

Fig. 6. Posterior distribution of parameters against DEM resolution in Mumahe River basin.

The following procedures are taken in this study to identify the formed on the parameters. Figs. 9–11 illustrate the correlation
parameter correlations: (1) the GLUE technique is applied to relations amongst the parameters from posterior distributions ob-
achieve the posterior distributions of the parameters by setting a tained by the GLUE technique in the three river basins respectively
threshold efficiency as 70% (Nash–Sutcliffe efficiency criterion); considered in this study. Not significant correlations can be found
(2) a total of 100,000 parameter sets are generated by the uniform from Figs. 9–11. Panels showing correlations relations of {Szm, T0}
Monte Carlo sampling schemes; and (3) correlation analysis is per- are the exceptions, which show evident correlations with correla-
 K. Lin et al. / Journal of Hydrology 394 (2010) 370–383 377

(a) Szm posterior distribution (b) ln (T0) posterior distribution

0.1 0.1

Parameter distribution

Parameter distribution
0.08 0.08

0.06 0.06

0.04 0.04

0.02 0.02

0 0
0 −3
0.2 −2
0.4
−1
Pa 0.6 Pa
ram ram 0
ete 0.8 ete
r ra r ra 1
ng 1 ng
e e
1.2 2
56 56
34 ion 34
1.4 12 solut 3 12 lution
M re reso
DE DEM

(c) Td posterior distribution (d) SRmax posterior distribution

0.1 0.1
Parameter distribution

Parameter distribution
0.08 0.08

0.06 0.06

0.04 0.04

0.02 0.02

0 0
0 0
5 0.05
0.1
Pa
ram10 Pa 0.15
ete 15 ram
r ra ete 0.2
ng r ra 0.25
e 20 ng
56
e 0.3
34 56
12 tion 34
esolu lution
25 0.35 12
r eso
DEM D EM r

Fig. 7. Posterior distribution of parameters against DEM resolution in Xunhe River basin.

tion coefficients of 0.66, 0.62 and 0.62, and the Kendall correlation same number of behavioral parameter sets, i.e. for the sake of com-
coefficients, s, of 0.69, 0.67 and 0.67, and the corresponding param- parison, the number of behavioral parameter sets of CGLUE tech-
eter of copula function h, of 3.23, 3.03, 3.03 in Mumahe, Xunhe, and nique is set to be the same as that of the GLUE technique. Firstly,
Laoguanhe River basins respectively. The evident correlation be- comparison is carried out under the same threshold criterion. Nine
tween Szm and T0 can be well explained according to the hypothe- scenarios are made by setting the sampling numbers as 10,000,
ses behind TOPMODEL. Generally, three hypotheses were 50,000, and 100,000, and threshold values as 50%, 60%, and 70%.
introduced in TOPMODEL, and the third hypothesis assumes that The number of behavioral parameter sets of these nine scenarios
transmissivity is the negative exponent function of the water depth is listed in Table 5. The variations of the assessing indices corre-
of the saturated groundwater, see Eq. (3) for more details. sponding to different DEM resolutions with parameter correlation
in the first scheme are displayed in Table 6, where 50,000 are ac-
5.2. Uncertainty of TOPMODEL simulation cepted as the sampling numbers, 90% as confidence level, and
70% as threshold value.
In this study, the GLUE and the CGLUE techniques are used to The GLUE and CGLUE techniques behave in evidently different
perform the uncertainty analysis. The objective is to investigate ways in terms of analysis of the parameter correlation: (1) evident
the effects of parameter correlation on modeling uncertainty of differences can be found for GLUE and CGLUE techniques with
TOPMODEL by comparing the results from these two methods. respect to the uncertainty estimation of the model parameters.
Two schemes are done to investigate the effects of parameter cor- Table 5 shows that the number of behavioral parameter sets
relation on uncertainty of TOPMODEL simulation in this study: (1) decreases when the threshold increases, and increases when the
comparison is carried out between the GLUE and CGLUE tech- sampling times increase, and for any given scenario, more behav-
niques with the same threshold value of R2, 70%; (2) comparison ioral parameter sets are obtained in CGLUE technique than in GLUE
is conducted between the GLUE and CGLUE techniques with the technique; (2) the containing ratios, CR, and the interval width IW
378 K. Lin et al. / Journal of Hydrology 394 (2010) 370–383

(a) Szm posterior distribution (b) ln (T0) posterior distribution

0.08
0.1
Parameter distribution

Parameter distribution
0.06 0.08

0.06
0.04
0.04
0.02
0.02

0 0
0 −3
0.2 −2
0.4
−1
Pa 0.6 Pa
ram ram 0
ete 0.8 ete
r ra r ra 1
ng 1 ng
e e
1.2 2
56 56
1.4 34
1 2 esolutio
n 3 12
34
eso lution
DE Mr DEM r

(c) T posterior distribution (d) SRmax posterior distribution

d
0.08 0.08
Parameter distribution

Parameter distribution

0.06 0.06

0.04 0.04

0.02 0.02

0 0
0 0
5 0.05
0.1
Pa 10
ram Pa 0.15
ete ram 0.2
r ra15 ete
ng r ra 0.25
e 20 ng
56 e 0.3
34 n 56
25 1 2 resolutio 0.35 12
34
lution
DE M eso
DEM r

Fig. 8. Posterior distribution of parameters against DEM resolution in Laoguanhe River basin.

Table 5
List of the total number of the behavioral parameter sets.

Catchment Threshold (%) 10,000 50,000 100,000

GLUE method CGLUE method GLUE method CGLUE method GLUE method CGLUE method
Mumahe 50 1911 2997 9361 15,048 18,821 30,156
60 1102 1728 5304 8487 10,604 16,729
70 571 904 2841 4421 5639 8775
Xunhe 50 1420 2438 6956 12,241 13,987 24,532
60 831 1405 3997 6904 8066 13,764
70 430 735 2141 3596 4325 7219
Laoguanhe 50 1619 2601 7930 13,061 15,946 26,174
60 947 1499 4556 7366 9196 14,685
70 490 784 2440 3837 4930 7702

obtained by CGLUE technique are larger when compared to those one million parameters are required due to the large amount of
by GLUE technique (Table 6 and Fig. 12). Besides, more sampling parameter sets involved in the model structures. In this case,
times are necessary to evaluate the uncertainty of the parameters Monte Carlo procedure usually consumes massive computing re-
in a more comprehensive way. In practice, ten thousand or even sources. If the samples are not adequate, the sampling procedure
 K. Lin et al. / Journal of Hydrology 394 (2010) 370–383 379

Table 6
Assessing indices of uncertainty in the CGLUE technique considering parameter correlation in the first scheme.

Catchment Indices 25 m 50 m 100 m 200 m 400 m 800 m

Mumahe CR 0.668 0.667 0.670 0.671 0.673 0.672
IW (m3/s) 34.738 34.434 35.159 36.320 36.421 36.086
R2 (MQ0.5) (%) 79.740 79.910 79.140 79.170 79.130 79.030
Xunhe CR 0.689 0.688 0.691 0.695 0.694 0.691
IW (m3/s) 67.334 67.643 67.721 71.328 69.420 69.136
R2 (MQ0.5) (%) 76.810 76.880 77.000 76.700 76.780 76.710
Laoguanhe CR 0.681 0.680 0.682 0.687 0.690 0.686
IW (m3/s) 38.245 38.564 39.156 40.661 40.258 39.577
R2 (MQ0.5) (%) 76.780 76.830 76.680 76.760 76.550 76.640

Note: In the first scheme, comparison is carried out between the GLUE and CGLUE techniques with the same threshold value of R2, 70%.

Table 7
Assessing indices of uncertainty in the CGLUE technique considering parameter correlation in the second scheme.

Catchment Indices 25 m 50 m 100 m 200 m 400 m 800 m

Mumahe CR 0.630 0.628 0.629 0.632 0.634 0.631
IW (m3/s) 27.340 26.845 27.620 27.802 27.930 28.343
R2 (MQ0.5) (%) 84.130 84.040 84.180 84.310 83.980 83.750
Xunhe CR 0.644 0.643 0.644 0.650 0.649 0.644
IW (m3/s) 52.586 52.335 52.795 54.171 52.812 53.881
R2 (MQ0.5) (%) 81.350 81.150 81.400 81.120 81.110 80.980
Laoguanhe CR 0.639 0.637 0.638 0.644 0.647 0.641
IW (m3/s) 29.954 29.921 30.612 30.971 30.718 30.934
R2 (MQ0.5) (%) 81.370 81.190 81.470 81.090 81.130 81.020

Note: In the second scheme, the number of behavioral parameter sets of the CGLUE technique is set to be the same as that of the GLUE technique.

3 20 0.3

2 0.25
15
0.2
1
SRmax
ln (To)

Td

10 0.15
0
0.1
5
−1
0.05

−2 0 0
0 0.5 1 0 0.5 1 0 0.5 1
Szm Szm Szm

20 0.3 0.3

0.25 0.25
15
0.2 0.2
SRmax

SRmax
Td

10 0.15 0.15

0.1 0.1
5
0.05 0.05

0 0 0
−2 0 2 −2 0 2 0 10 20
ln (To) ln (To) Td

Fig. 9. Correlations amongst the parameters from posterior distribution in Mumahe River basin.

which assumes independence of the parameters may not include
all possible behavioral parameter combinations, and which is ex-
pected to underestimate the uncertainty. The results showed that
the CGLUE technique performs well in terms of estimation of the
parameter correlation when the model parameters are in good cor-
relations, and this method can help to obtain more behavioral
parameter sets by the same sampling times, and thus the uncer-
tainty of parameters could be estimated thoroughly.
In addition, comparison is conducted with the same number of
behavioral parameter sets to study the effects of parameter corre-
lation on uncertainty of TOPMODEL simulation. The variations of
the assessing indices corresponding to different DEM resolutions
380 K. Lin et al. / Journal of Hydrology 394 (2010) 370–383

3 20 0.3
2 0.25
15
1 0.2

SRmax
ln (To)

Td
0 10 0.15

−1 0.1
5
0.05
−2
0 0
0 0.5 1 0 0.5 1 0 0.5 1
Szm Szm Szm

20 0.3 0.3

0.25 0.25
15
0.2 0.2
SRmax

SRmax
Td

10 0.15 0.15

0.1 0.1
5
0.05 0.05

0 0 0
−2 0 2 −2 0 2 0 10 20
ln (To) ln (To) Td

Fig. 10. Correlations amongst the parameters from posterior distribution in Xunhe River basin.

3 20 0.3
2 0.25
15
1 0.2
SRmax
ln (To)

Td

0 10 0.15

−1 0.1
5
−2 0.05

−3 0 0
0 0.5 1 0 0.5 1 0 0.5 1
Szm Szm Szm

20 0.3 0.3

0.25 0.25
15
0.2 0.2
SRmax

SRmax
Td

10 0.15 0.15

0.1 0.1
5
0.05 0.05

0 0 0
−2 0 2 −2 0 2 0 10 20
ln (To) ln (To) Td

Fig. 11. Correlations amongst the parameters from posterior distribution in Laoguanhe River basin.

with parameter correlation considered in the first scheme are dis-
played in Table 7. Fig. 13 shows comparison of the assessing indi-
ces between independent and correlative parameters in the second
scheme. The runoff uncertainty intervals generated by the GLUE
technique for the Xunhe catchment during the time period of July
3th, 1983–December 10th, 1983 when threshold of likelihood
equals to 70%, are demonstrated in Fig. 14, and the corresponding
runoff uncertainty intervals by the CGLUE technique with the same
number of behavioral parameter sets obtained by the GLUE tech-
nique are illustrated in Fig. 15.
It can be observed from Table 7 and Figs. 13–15 that the Nash–
Sutcliffe efficiency of the median value, MQ0.5, R2 (MQ0.5) increases
accompanying the enhancing parameter correlations. Increase of
the Nash–Sutcliffe efficiency of the median value, MQ0.5, indicates
 K. Lin et al. / Journal of Hydrology 394 (2010) 370–383 381

Fig. 12. Comparison of the assessing indices between the GLUE and the CGLUE techniques with the same threshold value of R2, 70% (the first scheme).

Fig. 13. Comparison of the assessing indices between the GLUE and the CGLUE techniques with the same number of behavioral parameter sets, i.e. for the sake of comparison,
the number of behavioral parameter sets of CGLUE technique is set to be the same as that of the GLUE technique (the second scheme).

4000
3500 observed data
discharge (m3/s)

3000 upper bounds of GLUE

2500 lower bounds of GLUE
2000
1500
1000
500
0
1983-7-3 1983-8-12 1983-9-21 1983-10-31 1983-12-10
time

Fig. 14. The runoff uncertainty intervals generated by the GLUE technique for the Xunhe catchment during the time period of July 3th, 1983–December 10th, 1983 with the
threshold of likelihood of 70%.

4000
3500 observed data
discharge (m3/s)

3000 upper bounds of CGLUE

2500 lower bounds of CGLUE
2000
1500
1000
500
0
1983-7-3 1983-8-12 1983-9-21 1983-10-31 1983-12-10
time

Fig. 15. The runoff uncertainty intervals generated by the CGLUE technique for the Xunhe catchment during the time period of July 3th, 1983–December 10th, 1983 with the
same number of behavioral parameter sets obtained by the GLUE method.

that, when considering the parameter correlation, the simulated 6. Conclusion

runoff series by the TOPMODEL with the behavioral parameter sets
will fit better the observed runoff series. It can also be found from
Table 7 and Fig. 13 that the containing ratio, CR, decreases a little
but significant decrease can be found in the interval width, IW,
implying that the CGLUE technique can greatly reduce the uncer-
tainty as a result of parameter correlation to some degree.
Modeling uncertainties are possibly due to a variety of sources,
such as input error, calibration accuracy, parameter uncertainty,
and model structure, and so on. Holistic evaluation of impacts of
these various sources on modeling uncertainty is important for sci-
entific decision making and will help direct resources towards
382 K. Lin et al. / Journal of Hydrology 394 (2010) 370–383
model structural improvements and uncertainty reduction
(Blasone and Vrugt, 2008). Beven and Freer (2001a,b) addressed
the effects of some factors on modeling uncertainty, such as model
nonlinearity, covariation of parameters and drawbacks in model
structure, input error or observed variables, within the GLUE pro-
cedure. Considering the importance of modeling uncertainty and
related underlying causes, based on the GLUE technique, the cur-
rent study thoroughly investigated the effects of DEM resolution
and parameter correlations on uncertainties of TOPMODEL simula-
tions and some interesting, novel and important conclusions have
been obtained:
(1) DEM resolution has the potential to affect the topographic
characteristics and the ln(a/tan(B)) distribution. However,
the results of this study indicates little influence of DEM res-
olutions on the modeling uncertainty of the TOPMODEL,
which is represented by no evident differences of the uncer-
tainty intervals corresponding to different DEM resolutions.
It should be attributed to the fact that the modeling uncer-
tainty should be due solely to DEM resolution changes by
fixing the parameter values to avoid the artifacts resulted
from interactions between ln(a/tan(B)) and the parameters.
(2) The Copula function is used to describe the correlation struc-
tures of the model parameters with the CGLUE technique.
Comparison of the results by the GLUE and CGLUE technique
shows that, under the same threshold, more behavioral
parameter sets are required when the model parameters
are in good correlation relations, and in so doing the uncer-
tainty of parameters could be evaluated in a thorough way.
(3) The Nash–Sutcliffe efficiency of the median value, MQ0.5, and
R2 (MQ0.5) increases when the parameter correlation is put
under consideration, implying that, when considering the
parameter correlation, the simulated runoff series by the
TOPMODEL with the behavioral parameter sets can fit better
to the observed runoff series, and correspondingly, the
abstracted median value, MQ0.5, can be improved for better
forecasts of the runoff. Besides, the results of this current
study also confirm that the CGLUE technique considering
the parameter correlation can greatly reduce the modeling
uncertainty of the TOPMODEL as a result of parameter corre-
lation to some degree.
The hydrological model structure becomes increasingly compli-
cated to satisfy the requirements of the gradually complicated
environment problems we are facing. However, complicated
hydrological model structure is easy to produce more uncertainties
within the hydrological modeling results. In this study, impacts of
DEM resolution and parameter correlations of the TOPMODEL on
the uncertainties of the modeling results are thoroughly investi-
gated by using CGLUE technique. The results of this study will be
of theoretical and practical merits in scientific understanding of
uncertainties of the modeling behaviors of the TOPMODEL, and will
also be helpful for further improvement of the TOPMODEL in terms
of hydrological modeling.
Acknowledgements
This study was financially supported by the National Natural
Science Foundation of China (Grant Nos.: 50809078; 41071020;
50839005), the Open Foundation of the State Key Laboratory of
Water Resources, Hydropower Engineering Science (Wuhan
University) (Grant No.: 2008B043), the Scientific Research Founda-
tion for Yong Scholars of Sun Yat-sen University (Grant No.:
3161395), the Natural Science Foundation of Guangdong province
(Grant Nos.: 4203384; 2009-37000-4203384), the Program for
Outstanding Young Teachers of the Sun Yat-sen University (Grant
No.: 1132381) and by a grant from the Research Grants Council
of the Hong Kong Special Administrative Region, China (Project
No. CUHK405308). The last but not the least, cordial gratitude
should be extended to the editor, Prof. Dr. Geoff Syme, and the
two anonymous reviewers for their pertinent and professional
comments which greatly help to improve the quality of this
manuscript.
References
Ajami, N.K., Duan, Q., Sorooshian, S., 2007. An integrated hydrologic bayesian multi-
model combination framework: confronting input, parameter and model
structural uncertainty. Water Resources Research 43, W01403. doi:10.1029/
2005WR004745.
Benke, K.K., Lowell, K.E., Hamilton, A.J., 2007. Parameter uncertainty, sensitivity
analysis and prediction error in a water-balance hydrological model.
Mathematical and Computer Modeling 47 (11–12), 1134–1149.
Beven, K.J. (Ed.), 1997. Distributed Modeling in Hydrology: Applications of
TOPMODEL. Wiley, Chichester.
Beven, K.J., Binley, A., 1992. The future of distributed models, model calibration and
uncertainty prediction. Hydrological Processes 6, 279–298.
Beven, K.J., Freer, J., 2001a. Equifinality, data assimilation, and uncertainty
estimation in mechanistic modeling of complex environmental systems using
the GLUE methodology. Journal of Hydrology 249, 11–29.
Beven, K.J., Freer, J., 2001b. A dynamic TOPMODEL. Hydrology Processes 15, 1993–
2011.
Beven, K.J., Kirkby, M.J., 1979. A physically-based variable contributing area model
of basin hydrology. Hydrology Science Bulletin 24 (1), 43–69.
Beven, K., Wood, E.F., 1983. Catchment geomorphology and the dynamics of runoff
contributing areas. Journal of Hydrology 65, 139–158.
Beven, K.J., Lamb, R., Quinn, P.F., Romanowicz, R., Freer, J., 1995. TOPMODEL. In:
Singh, V.P. (Ed.), Computer Models of Watershed Hydrology. Water Resources
Publications, Highlands Ranch, CO, pp. 627–668.
Blasone, R.S., Vrugt, J.A., 2008. Generalized likelihood uncertainty estimation (GLUE)
using adaptive Markov Chain Mote Carlo sampling. Advances in Water
Resources 31, 630–648.
Blazkova, S., Beven, K.J., 2002. Flood frequency estimation by continuous simulation
for a catchment treated as ungauged (with uncertainty). Water Resources
Research 38 (8), 1139. doi:10.1029/2001WR000500.
Choi, H.T., Beven, K., 2007. Multi-period and multi-criteria model conditioning to
reduce prediction uncertainty in an application of TOPMODEL within the GLUE
framework. Journal of Hydrology 332, 316–336.
Duan, J., Miller, N.L., 1997. A generalized power function for the subsurface
transmissivity analysis of the TOPMODEL. Journal of Hydrology 175, 293–338.
Freer, J., Beven, K.J., Ambroise, B., 1996. Bayesian estimation of uncertainty in runoff
prediction and the value of data: an application of the GLUE approach. Water
Resources Research 32 (7), 2161–2173.
Gupta, H.V., Sorooshian, S., Yapo, P.O., 1998. Toward improved calibration of
hydrologic models: multiple and noncommensurable measures of information.
Water Resources Research 34, 751–763.
Jin, X., Xu, C., Zhang, Q., Singh, V.P., 2010. Parameter and modeling uncertainty
simulated by GLUE and a formal Bayesian method for a conceptual hydrological
model. Journal of Hydrology 383, 147–155.
Keesman, K.J., 1990. Set theoretic parameter estimation using random scanning and
principal component analysis. Mathematics and Computers in Simulation 32,
535–543.
Kuczera, G., Parent, E., 1998. Monte Carlo assessment of parameter uncertainty in
conceptual catchment models: the Metropolis algorithm. Journal of Hydrology
211, 69–85.
Lin, K., Guo, S., Zhang, W., Liu, P., 2007. A new baseflow separation method based on
analytical solutions of the Horton infiltration capacity curve. Hydrological
Processes 21 (13), 1719–1736.
Lin, K., Chen, X., Zhang, Q., Chen, Z., 2009. A modified generalized likelihood
uncertainty estimation method by using Copula function, IAHS Publication 335,
51–56.
Madsen, H., 2003. Parameter estimation in distributed hydrological catchment
modeling using automatic calibration with multiple objectives. Advance in
Water Resource 26, 205–216.
McMichael, C.E., Hope, A.S., Loaiciga, H.A., 2006. Distributed hydrological modeling
in California semi-arid shrublands: MIKE SHE model calibration and uncertainty
estimation. Journal of Hydrology 317 (3–4), 307–324.
Montanari, A., 2005. Large sample behaviors of the generalized likelihood
uncertainty estimation (GLUE) in assessing the uncertainty of rainfall–runoff
simulations. Water Resources Research 41, W08406. doi:10.1029/
2004WR003826.
Montanari, A., 2007. What do we mean by ‘uncertainty’? the need for a consistent
wording about uncertainty assessment in hydrology. Hydrological Processes 21,
841–845.
Moradkhani, H., Hsu, K.-L., Gupta, H., Sorooshian, S., 2005. Uncertainty assessment
of hydrologic model states and parameters: sequential data assimilation using
the particle filter. Water Resources Research 41 (5), 1–17.
Nash, J.E., Sutcliffe, J.V., 1970. River flow forecasting through the conceptual models,
1: a discussion of principles. Journal of Hydrology 10 (3), 282–290.
 K. Lin et al. / Journal of Hydrology 394 (2010) 370–383
Nelson, R.B. (Ed.), 1999. An introduction to Copulas. Springer, New York.
Quinn, P., Beven, K.J., Lamb, R., 1995. The ln(a/tan b) index: how to calculated it and
how to use it within the TOPMODEL framework. Hydrological Processes 9, 161–
182.
Salvadori, G., De Michlele, C., 2004. Frequency analysis via copulas, theoretical
aspects and applications to hydrological events. Water Resources Research 40
(12), 1–17.
Scanlon, T.M., Raffensperger, J.P., Hornberger, G.M., Clapp, R.B., 2000. Shallow
subsurface storm flow in a forested headwater catchment: observations and
modeling using a modified TOPMODEL. Water Resources Research 36 (9), 2575–
2586.
Sklar, K., 1959. Fonctions de repartition à n dimensions et leura marges. Paris,
Publications de l’Institut de de l’Université, vol. 8, pp. 229–231.
Sørensen, R., Seibert, J., 2007. Effects of DEM resolution on the calculation of
topographical indices: TWI and its components. Journal of Hydrology 347, 79–89.
Sorooshian, S., Gupta, V.K., 1995. Model calibration. In: Singh, V.P. (Ed.), Computer
Models of Watershed Hydrology. Water Resources Publication, Highlands
Ranch, CO.
Thiemann, M., Trosset, M., Gupta, H., Sorooshian, S., 2001. Bayesian recursive
parameter estimation for hydrological models. Water Resources Research 7
(10), 21–35.
Vrugt, J.A., Robinson, B.A., 2007. Treatment of uncertainty using ensemble
methods: comparison of sequential data assimilation and Bayesian model
averaging. Water Resources Research 43, W01411. doi:10.1029/
2005WR004838.
Vrugt, J.A., Gupta, H.V., Bouten, W., Sorooshian, S., 2003. A Shuffled complex
evolution metropolis algorithm for optimization and uncertainty assessment of
hydrologic model parameters. Water Resources Research 39 (8), 1201.
doi:10.1029/2002WR001642.
Vrugt, J.A., Diks, C.G.H., Gupta, H.V., Bouten, W., Verstraten, J.M., 2005. Improved
treatment of uncertainty in hydrologic modeling: combining the strengths of
global optimization and data assimilation. Water Resources Research 41 (1),
1–17.
383
Walter, M.T., Steenhuis, T.S., Mehta, V.K., Thongs, D., Zion, M., Schneiderman, E.,
2002. Refined conceptualization of TOPMODEL for shallow subsurface flows.
Hydrology Processes 16, 2041–2046.
Wolock, D.M., McCabe Jr., G.J., 1995. Comparison of single and multiple flow
direction algorithms for computing topographic parameters in TOPMODEL.
Water Resources Research 31 (5), 1315–1324.
Wolock, D.M., Price, C.V., 1994. Effects of digital elevation model map scale and data
resolution on a topography-based watershed model. Water Resources Research
30 (11), 3041–3052.
Xiong, L., Guo, S., 2004. Effects of the catchment runoff coefficient on the
performance of TOPMODEL in rainfall–runoff modeling. Hydrological
Processes 18, 1823–1836.
Xiong, L., O’Connor, K.M., 2008. An empirical method to improve the prediction
limits of the GLUE methodology in rainfall–runoff modeling. Journal of
Hydrology 349, 115–124.
Yang, J., Reichert, P., Abbaspour, K.C., 2007. Bayesian uncertainty analysis in
distributed hydrologic modeling: a case study in the Thur River basin
(Switzerland). Water Resources Research 43, W10401. doi:10.1029/
2006WR005497.
Yang, J., Reichert, P., Abbaspour, K.C., Xia, J., Yang, H., 2008. Comparing uncertainty
analysis techniques for a SWAT application to the Chaohe Basin in China.
Journal of Hydrology 358, 1–23.
Zhang, W., Montgomery, D.R., 1994. Digital elevation model grid size, landscape
representation, and hydrologic simulations. Water Resources Research 30,
1019–1028.
Zhang, L., Singh, V.P., 2006. Bivariate flood frequency analysis using the copula
method. Journal of Hydrologic Engineering 11, 150–164.
Zhang, Q., Xu, C.-Y., Singh, V.P., Yang, T., 2009. Multi-scale variability of sediment
load and streamflow of the Lower Yangtze River basin: possible causes and
implications. Journal of Hydrology 368, 96–104.
Zhang, Q., Xu, C.-Y., Zhang, Z., Chen, Y.D., 2010. Changes of atmospheric water vapor
budget in the Pearl River basin and possible implications for hydrological cycle.
Theoretical and Applied Climatology. doi:10.1007/s00704-010-0257-z.