Automated Water Extraction Index: A new technique for surface water
mapping using Landsat imagery
Gudina L. Feyisa a,
⁎, Henrik Meilby a
, Rasmus Fensholt b
, Simon R. Proud b
a
Department of Food and Resource Economics, University of Copenhagen, Rolighedsvej 23, DK-1958 Frederiksberg C, Denmark
b
Department of Geosciences and Natural Resource Management, University of Copenhagen, Oester Voldgade 10, DK-1350 Copenhagen K, Denmark
a b s t r a c ta r t i c l e i n f o
Article history:
Received 16 April 2013
Received in revised form 3 August 2013
Accepted 21 August 2013
Available online 17 September 2013
Keywords:
Classification accuracy
Threshold stability
Subpixel
Mixed pixel
Classifying surface cover types and analyzing changes are among the most common applications of remote
sensing. One of the most basic classification tasks is to distinguish water bodies from dry land surfaces. Landsat
imagery is among the most widely used sources of data in remote sensing of water resources; and although
several techniques of surface water extraction using Landsat data are described in the literature, their application
is constrained by low accuracy in various situations. Besides, with the use of techniques such as single band
thresholding and two-band indices, identifying an appropriate threshold yielding the highest possible accuracy
is a challenging and time consuming task, as threshold values vary with location and time of image acquisition.
The purpose of this study was therefore to devise an index that consistently improves water extraction accuracy
in the presence of various sorts of environmental noise and at the same time offers a stable threshold value. Thus
we introduced a new Automated Water Extraction Index (AWEI) improving classification accuracy in areas that
include shadow and dark surfaces that other classification methods often fail to classify correctly. We tested the
accuracy and robustness of the new method using Landsat 5 TM images of several water bodies in Denmark,
Switzerland, Ethiopia, South Africa and New Zealand. Kappa coefficient, omission and commission errors were
calculated to evaluate accuracies. The performance of the classifier was compared with that of the Modified Normalized
Difference Water Index (MNDWI) and Maximum Likelihood (ML) classifiers. In four out of five test sites,
classification accuracy of AWEI was significantly higher than that of MNDWI and ML (P-value b 0.01). AWEI improved
accuracy by lessening commission and omission errors by 50% compared to those resulting from MNDWI
and about 25% compared to ML classifiers. Besides, the new method was shown to have a fairly stable optimal
threshold value. Therefore, AWEI can be used for extracting water with high accuracy, especially in mountainous
areas where deep shadow caused by the terrain is an important source of classification error.
© 2013 Elsevier Inc. All rights reserved.
1. Introduction
Environmental changes and their impacts on natural systems and
human societies are topics of research in a wide range of scientific fields.
Surface water is among the most vital earth resources undergoing
changes in time and space as a consequence of land use/cover (LULC)
changes, climate change and other forms of environmental changes in
many parts of the world. The ecological, social, health and economic
effects of surface water changes have been the subject of academic
study for many years (Alderman, Turner, & Tong, 2012; Bond, Lake, &
Arthington, 2008; Charron et al., 2004; Kondo et al., 2002; Lake, 2003;
Li, Wu, Dai, & Xu, 2012); Sun, Sun, Chen, and Gong (2012). Changes in
surface water may result in disasters such as flooding, outbreaks of
waterborne disease and water shortage in dry tropical areas, which
may involve loss of lives. Timely monitoring and delivery of data on
the dynamics of surface water are, therefore, essential for policy and
decision making processes (Giardino, Bresciani, Villa, & Martinelli,
2010; Morss, Wilhelmi, Downton, & Gruntfest, 2005).
Remote sensing has become an important source of information in
analyzing and delivering data on changes in different earth resources,
and surface water in particular. Examples of studies applying remote
sensing and GIS techniques for various applications in relation to water
resources include flood hazard/damage assessment and management
(Dewan, Islam, Kumamoto, & Nishigaki, 2007; Ji, Zhang, & Wylie, 2009;
Proud, Fensholt, Rasmussen, & Sandholt, 2011), change in surface
water resources (Gardelle, Hiernaux, Kergoat, & Grippa, 2009; Haas,
Bartholomé, & Combal, 2009; Prigent et al., 2012), water quality assessment
and monitoring (Guttler, Niculescu, & Gohin, 2013; He et al.,
2012; Novoa et al., 2012), and water-related disease epidemiology
(Charoenpanyanet & Chen, 2008; Dambach et al., 2012; Lacaux, Tourre,
Vignolles, Ndione, & Lafaye, 2007).
Satellite sensors of varying spatial, temporal and spectral resolution
have been used to extract and analyze information regarding surface
water. Landsat satellites are among the most widely used optical
Remote Sensing of Environment 140 (2014) 23–35
⁎ Corresponding author. Tel.: +45 91414185; fax: +45 353 31508.
E-mail addresses: fgudina@gmail.com, fgudina@ifro.ku.dk (G.L. Feyisa),
heme@ifro.ku.dk (H. Meilby), rf@geo.ku.dk (R. Fensholt), srp@geo.ku.dk (S.R. Proud).
0034-4257/$ – see front matter © 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.rse.2013.08.029
Contents lists available at ScienceDirect
Remote Sensing of Environment
journal homepage: www.elsevier.com/locate/rse
sensors in surface water and other environmental research. The use of
these remotely sensed data commonly starts with classification of
land use/cover types. Common water classification methods for optical
imagery could be categorized into four basic types (Ji et al., 2009):
(a) thematic classification (Lira, 2006), (b) linear unmixing (Sethre,
Rundquist, & Todhunter, 2005), (c) single-band thresholding (Jain,
Singh, Jain, & Lohani, 2005) and (d) two-band spectral water indices
(Jain, Saraf, Goswami, & Ahmad, 2006; McFeeters, 1996; Rogers &
Kearney, 2004; Xu, 2006). Combinations of various methods are also
proposed to improve water extraction accuracies. Examples are, Jiang,
Qi, Su, Zhang, and Wu (2012), Sheng, Shah, and Smith (2008), Sun
et al. (2012) and Verpoorter, Kutser, and Tranvik (2012). Single band
thresholding and two-band indices are commonly used water extraction
methods because of ease of use and the fact that these methods
are computationally less time-consuming than alternative approaches
(Ryu, Won, & Min, 2002).
McFeeters (1996) introduced the Normalized Difference Water
Index (NDWI) to delineate open water features using the green (band
2) and near-infrared (band 4) of Landsat TM. Rogers and Kearney
(2004) used another NDWI for water extraction where they applied
bands 3 and 5 of Landsat TM. McFeeters (1996) proposed a threshold
of 0 for extracting surface water using the raw digital number of
Landsat, where all positive NDWI values would be classified as water
and negative values as nonwater. However, Xu (2006) found that the
NDWI cannot efficiently suppress the signal from built-up surfaces
and using an NDWI threshold of 0 does not accurately enable discriminating
built-up surfaces from water pixels. Xu (2006) therefore proposed
another index, called Modified Normalized Difference Water
Index (MNDWI), where McFeeters (1996) NDWI was modified by
replacing band 4 by band 5 of Landsat 5 TM. The MNDWI of Xu (2006)
is one of the most widely used water indices for various applications,
including surface water mapping, land use/cover change analyses
and ecological research (Davranche, Lefebvre, & Poulin, 2010; Duan
& Bastiaanssen, 2013; Hui, Xu, Huang, Yu, & Gong, 2008; Poulin,
Davranche, & Lefebvre, 2010).
Even though a number of water extraction techniques are described
in the literature, the choice between them is constrained by accuracy
problems. Environmental monitoring and change detection techniques
such as post-classification comparison are likely to be less reliable
when classifiers of low accuracy are used (Congalton & Green, 2009;
Mucher, Steinnocher, Kressler, & Heunks, 2000). For instance, in a
study focusing on water dynamics monitoring, Ji et al. (2009) faced
two major problems in appropriately using water indices: first, the results
obtained using different indices were inconsistent and unreliable;
second, the threshold values applied to distinguish water from nonwater
were unstable, varying with scene and locations. These authors
compared four different water indices using simulated datasets of
four satellite sensors: Landsat ETM+, Système Pour l'Observation de
la Terre (SPOT), the Advanced Space-borne Thermal Emission and
Reflection radiometer (ASTER), and the Moderate Resolution Imaging
Spectroradiometer (MODIS), aiming to identify the best method for
delineating water features. Among the four alternatives, they found
that the MNDWI performed best in delineating water, and featured
the most stable threshold.
Water classification accuracy problems may be especially pronounced
in areas where the background land cover includes low albedo
surfaces such as asphalt roads in urban areas, and shadows from
mountains, buildings and clouds. The presence of shadows may cause
misclassification due to the similarity in reflectance patterns, and this
may lessen the accuracy of surface water mapping and change analysis
(Frey, Huggel, Paul, & Haeberli, 2010; Verpoorter et al., 2012; Xu, 2006).
In environments where nonwater dark surfaces are found, simple classification
methods such as two-band water indices and single-band
thresholding may not sufficiently and accurately distinguish between
water pixels and nonwater dark surfaces, particularly shadows
(Verpoorter et al., 2012). In a study of land cover dynamics using
Landsat TM data, we noted accuracy problems due to failure of existing
water extraction methods in accurately distinguishing water from
shadows and low albedo urban surfaces. Particularly, no existing
water index was able to automatically separate water and shadowed
surfaces. In this paper, therefore, we introduce a multiple-band index
called Automated Water Extraction Index (AWEI), with the objectives
to: (a) improve accuracy of surface water mapping by automatically
suppressing classification noise from shadow and other nonwater
dark surfaces, and (b) test the robustness of the new method under
different environmental conditions and evaluate its relative accuracy
in comparison with existing classification techniques.
2. Study areas and data sources
2.1. Test sites
The accuracy and robustness of the Automated Water Extraction
Index (AWEI) were tested considering several lakes and other water
bodies in different environmental conditions ranging from humid temperate
through sub-tropical to tropical dry regions. The test water bodies
were obtained from five different countries: Denmark, Switzerland,
Ethiopia, South Africa and New Zealand. The water bodies that include
small freshwater reservoirs, large lakes, harbors and the sea differ
with regard to depth, turbidity, chemical composition and surface appearance.
A summary of the basic characteristics of the test sites is
shown in Table 1.
The test sites were deliberately selected so that the sub-scenes
consist of complex surface features, such as hill shade, built-up areas
and other dark surfaces as background to the water bodies. The test
sites in Switzerland, Ethiopia and South Africa are characterized by the
presence of built-up surfaces and shadows of mountains. The site in
Denmark also consists primarily of urban background but with no
major shadow problems since the terrain is predominantly flat and
tall buildings in the urban area are rare. The test site in New Zealand
consists of mountain slopes with deep shadows, but no major urban
surfaces are included.
In addition to the five test sites for which detailed accuracy analyses
and comparisons were carried out, further validation of the robustness
of the new index was undertaken considering shadow-dominated
water bodies in Norway, rivers with urban surfaces and shadows from
tall buildings in Shanghai, China, and several crater lakes with built-up
background surfaces in Bishoftu, Ethiopia. However, these additional
test sites were not analyzed in detail and classification output from
these sites is not included in the Results section; instead, the classification
maps are included in Appendix A for visual inspection of classification
accuracy.
2.2. Landsat images
Landsat 5 TM images were acquired from USGS GLOVIS portal
(United States Geological Survey (USGS), 2012). All Landsat images
used are of product type L1T and with a scene quality score of 9,
which means perfect scenes with no errors detected. The images were
also georeferenced with precision better than 0.4 pixels (NASA, 2012).
The sub-scenes were all free of clouds. Descriptions of the Landsat
images are presented in Table 2.
2.3. Reference data
Reference data used in accuracy assessment are described in Table 2.
For the test site in Denmark, colored Digital Orthophoto Quadrangles
(DOQs) from year 2010 were used as reference. These aerial photos
have a spatial resolution of 12.5 cm and location accuracy better than
0.5 m (COWI, 2010). For the four other test sites, high spatial resolution
images provided by Google Earth™ were used for reference. The acquisition
dates of the reference data and the Landsat 5 TM images were
24 G.L. Feyisa et al. / Remote Sensing of Environment 140 (2014) 23–35
closely matched to minimize bias in the surface water boundaries that
could arise because of large differences in time. The dates of acquisition
of the Landsat images and reference data are shown in Table 2.
The “true” boundaries of all the test water bodies were digitized
manually on-screen from the reference data. In the analysis, the manually
digitized water map was used to assess the accuracy of the different
water extraction methods when applied to the Landsat images.
3. Methods
3.1. Image preprocessing
Landsat 5 TM images acquired in the form of raw digital number
were calibrated to surface reflectance values. Atmospheric correction
was applied to all images using the Fast Line-of-Sight Atmospheric Analysis
of Spectral Hypercubes (FLAASH) module in ENVI v. 4.8 (Exelis
Visual Information Solutions, 2010). Aerosol Optical Depth (AOD) and
total Column Water Vapor data to be used in the FLAASH atmospheric
correction module were retrieved from MODIS Terra atmospheric products
(Jimenez-Munoz, Sobrino, Mattar, & Franch, 2010) for each of the
five test sites. The MODIS Terra products that had been acquired on
the same date as the Landsat images were used to retrieve the calibration
constants. Initial visibility was estimated using the aerosol optical
thickness obtained from MODIS data using Eq. (1), where, VIS = initial
visibility, Z = mixing layer height and AOD = Aerosol Optical Depth.
The mixing layer height values for the test sites were not available.
Therefore, a daytime mixing layer height of 1.5 km was used for all
test sites (typical values range from 1 to 2 km (Butcher, Charlson,
Orians, & Wolfe, 1992)). Water vapor calibration constants for each of
the Landsat images are summarized in Table 3. The overpass time of
Landsat TM and MODIS Terra at each test site was closely matched
(less than 2 h difference).
VIS ¼
3:912
AOD
à Z: ð1Þ
Image-to-image co-registration between the reference data and
Landsat images was undertaken for the test sites in Denmark and
Ethiopia. Manual co-registration was performed with a Root Mean
Square Error (RMSE) of less than 0.4 pixels. At least 25 control points
were used for co-registration of each image with the reference data.
The co-registration between Landsat and Google Earth™ images at the
rest of the test sites was already highly accurate and hence there
was no need for manual co-registration. Since the L1T Landsat TM and
ETM+ products are geometrically corrected (NASA, 2012), no such
corrections were applied in the pre-processing.
3.2. Pure-pixel selection
An independent set of “pure” pixel reflectance values of nine major
land cover types was sampled from the six reflective bands of a Landsat
5 TM image of Addis Ababa, acquired on Dec. 9, 2010. The land cover
types are: water, vegetation (forest and non-forest), bright soil, dark
soil, brown soil, bright built, asphalt, other dark built and shadow.
Spectral data from these pure pixels were used to examine reflectance
patterns and identify land cover types that affect water extraction accuracy,
aiming to design a method that accurately discriminates between
Table 1
Characteristics of the study sites.
The source of climate information is: (http://www.climatedata.eu/).
Country and name of water bodies Center point coordinate
(UTM)
Area
(ha)
General characteristics
of water bodies
Mean alt. (m) Topography Climate
Denmark
Several artificial lakes, a harbor and
the sea (Øresund and Køge Bugt)
6,172,085 m N, 12°34′57.42″E 2085 Shallow clear artificial
lakes, clear seas
9 Predominantly flat Temperate
Switzerland
Lake Lauerz 5,209,030 m N, 469,608 m E 289 Clear lake 1100 Mountainous Temperate
Ågeri lake 5,218,774 m N, 471,530 m E 719 Clear lake
Sihl lake 5,218,191 m N, 484,028 m E 1034 Clear lake
Wägitaler lake 5,214,616 m N, 494,092 m E 402 Clear lake
Klöntaler lake 5,207,839 m N, 498,040 m E 309 Clear lake
Ethiopia
Gefersa 1,002,432 m N, 459,709 m E 144 Clear reservoir 2377 Mountainous Tropical dry
Dire 1,011,794 m N, 493,000 m E 106 Turbid reservoirs
Legedadi 1,002,374 m N, 497,446 m E 423 Turbid reservoirs
South Africa
Berg river 6,244,161 m S, 320,605 m E 426 Clear reservoir 600 Rugged hilly Subtropical semi-arid
Wemmershoek 6,255,473 m S, 323,355 m E 195 Clear reservoir
Brandvlei 6,265,857 m S, 354,413 m E 3097 Clear reservoir
New Zealand
Lake Te Anau 5,004,239 m S, 723,800 m E 6495 Large clear lake 800 Rugged hilly Humid temperate
Table 2
Description of Landsat TM scenes and corresponding reference data.
Test site Landsat scene Reference data and sources
Acquisition date Path Row
Water bodies in Copenhagen, Denmark June 29, 2010 195 21 Colored Digital Orthophoto quadrangles acquired between May and July 2010, ©COWI
Lakes in Switzerland Sept 30, 2009 195 27 Google Earth™ image acquired on Jul 1 2009
Reservoirs in Addis Ababa, Ethiopia Dec 9, 2010 168 54 Google Earth™ image acquired on Oct 13 and Dec 20, 2010, ©Digital globe, CNES/SPOT Image
Reservoirs in South Africa Mar 29, 2010 175 83 Google Earth™ image acquired on Jan 4, Jan 7 and Jan 18, 2010, ©GeoEye and AfriGIS
A lake in New Zealand Feb 3, 2010 76 91 Google Earth™ image acquired on Apr 09, 2010 and Feb 16, 2011, ©GeoEye
25G.L. Feyisa et al. / Remote Sensing of Environment 140 (2014) 23–35
such surfaces and water. The pure pixel data were intended neither for
classification nor accuracy assessment and therefore, only the image of
Addis Ababa and its surroundings was used for pure pixel extraction.
The reason for choosing Addis Ababa for pure pixel extraction was
that this area includes all the major challenging features influencing
water extraction accuracy: shadow, dark built-up surfaces and other
low albedo surfaces such as black soil.
The methods used to extract pure pixels of the selected land cover
types include spectral feature space scatter plot from Minimum Noise
Fraction Transform (MNFT) images, Pixel Purity Index (PPI), manual
digitization from images accessed through Google Earth™, groundbased
land cover assessment and the familiarity of the first author
with the local area. Pure pixel samples for water were taken from the
middle of lakes to avoid mixed edge pixels. Similarly, high forest
with closed canopy from Menagesha national forest was applied for
sampling vegetation pure pixels. Pure pixels of built-up land cover
were sampled from homogenous surfaces such as airport runways and
large warehouse roofs in Addis Ababa city. Since land cover types in
Addis Ababa are highly heterogeneous, detection of pure built pixels
was assisted by PPI and spectral feature space in ENVI v.4.8 (Exelis
Visual Information Solutions, 2010). Shadow pixels were extracted by
thresholding hill-shade images derived from elevation data using the
ASTER Digital Elevation Model (DEM) in mountainous parts of the test
site. Homogenous agricultural fields with exposed black, brown and
bright soils were also sampled from the outskirts of the city.
For each land cover type, 312 pure pixels were extracted from the six
reflective bands of the Landsat 5 TM images. Average reflectance values
of the pure pixels are shown in Fig. 1. Separability of the spectral
signatures of the selected nine major land cover types was tested by
Jeffries–Matusita's pairwise separability measure (Richards, 1993) in
ENVI v. 4.8. All pairs of land cover types were found to be separable
with values ranging from 1.89 to 2.0.
3.3. Formulation of the Automated Water Extraction Index (AWEI)
Five spectral bands of Landsat 5 TM were used in developing the
new index (AWEI) to increase the contrast between water and other
dark surfaces. The primary aim of the formulation of AWEI was to
maximize separability of water and nonwater pixels through band
differencing, addition and applying different coefficients. Accordingly,
two separate equations are proposed to effectively suppress nonwater
pixels and extract surface water with improved accuracy (Eqs. (2) and
(3)). The coefficients used in Eqs. (2) and (3) and the arithmetic combinations
of the chosen spectral bands were determined based on critical
examination of the reflectance properties of various land cover types.
The coefficients of these equations are empirical results determined
based on reflectance patterns observed across the dataset of pure pixels
of various land cover types. An iterative process was applied to identify
parameters that maximize the separability of water and nonwater surfaces
characterized by low reflectance. In the final index, the coefficients
were rounded for ease of use. Particular emphasis was given to the enhancement
of the separability of water and dark surfaces such as shadow
and built-up structures that are often difficult to distinguish due to
similarities in reflectance patterns. In addition to enhancing separability
of water and nonwater pixels, the choice of the coefficients also aimed
to stabilize the threshold needed to distinguish water from nonwater
pixels by forcing nonwater pixels below 0 and water pixels above 0,
implying that 0 could be used as a reasonable starting threshold for
classifying land cover into binary classes of water and nonwater under
a wide range of environmental conditions.
AWEInsh ¼ 4 Â ρband2−ρband5ð Þ− 0:25 Â ρband4 þ 2:75 Â ρband7ð Þ ð2Þ
AWEIsh ¼ ρband1 þ 2:5 Â ρband2−1:5 Â ρband4 þ ρband5ð Þ−0:25 Â ρband7 ð3Þ
where ρ is the reflectance value of spectral bands of Landsat 5 TM:
band 1 (blue), band 2 (green), band 4 (NIR), band 5 (SWIR) and band
7 (SWIR).
AWEInsh is an index formulated to effectively eliminate nonwater
pixels, including dark built surfaces in areas with urban background
and AWEIsh is primarily formulated for further improvement of accuracy
by removing shadow pixels that AWEInsh may not effectively eliminate.
The subscript “nsh” in Eq. (2) is included to specify that the index is
suited for situations where shadows are not a major problem. The subscript
“sh” in Eq. (3) indicates that the equation is intended to effectively
eliminate shadow pixels and improve water extraction accuracy in
areas with shadow and/or other dark surfaces. But in areas with highly
reflective surfaces such as ice, snow and reflective roofs in urban areas,
Eq. (3) may misclassify such surfaces as water.
In Eq. (2), quadrupling the difference between the band 2 and band 5
results in large positive values for water pixels and negative values for
most nonwater pixels. To help in discriminating water from other
surfaces that have similar spectral patterns, band 4 and band 7 are
subtracted from the result and different weights are assigned to these
bands to force nonwater pixels to have even larger negative values;
this subtraction will not to any greater extent lead to negative values
for water pixels because water has very low reflectance in the spectral
ranges of band 4 and band 7. The equation results in large negative
values for pixels covered by vegetation, soil, bright built and other
surfaces that have large reflectance for band 4 or 7. The equation is
also intended to enhance separability between water, dark surfaces
and other nonwater surfaces. In many cases, water absorbs almost all
of the incoming radiation in bands 4, 5 and 7 and achieves relatively
highest reflectance between bands 1 and 2 of Landsat 5 TM (Lillesand,
Kiefer, & Chipman, 2004). Shadowed surfaces also have low reflectance
in all spectral bands, but the magnitude of reflectance varies due to variation
in surface characteristics and the depth of shadow. Hence, Eq. (2)
alone may not completely eliminate all types of shadows and other low
albedo surfaces. As shown in Fig. 1, for instance, subtracting band 5 from
band 2 could yield positive values for both water and shadows. Quadrupling
the difference and subtracting bands 4 and 7 may result in some
shadow pixels obtaining similar values as that of water due to the
similarity in reflectance patterns within these bands, hence making it
difficult to exclude shadow pixels from water class.
Due to these limitations of Eq. (2), Eq. (3) was formulated to achieve
enhanced separability of water and shadows and/or dark surfaces. It can
Table 3
Calibration values used in atmospheric correction using FLAASH.
Test Site Date of Landsat/MODIS
Terra overpass
Average water vapor column
from MODIS (g/cm2
)
Water vapor multiplier Average aerosol optical
thick-from MODIS
Visibility (km)
Denmark Jun 29, 2010 3.00 1.03 0.33 17.8
Switzerland Sept 30, 2009 1.43 0.49 0.08 73.4
Ethiopia Dec 9, 2010 2.30 0.56 0.10 58.7
South Africa Mar 29, 2010 1.90 0.46 0.033 177.8
New Zealand Feb 3, 2010 2.67 0.91 0.12 48.9
26 G.L. Feyisa et al. / Remote Sensing of Environment 140 (2014) 23–35
be noted from Fig. 1 that the largest difference between the reflectance
of water and shadow is found in bands 1 and 2. Therefore, adding these
two bands, while at the same time multiplying band 2 by the specified
coefficient, enhances the separability between water and shadow
pixels, yielding relatively large positive values for water pixels compared
to shadow pixels. Subtracting bands 4, 5 and 7 forces nonwater
pixels in the negative direction, and the net effect of this subtraction
on water pixels is minimal compared to nonwater surfaces including
shadows, which are forced considerably below zero. Band 3 was not
used in Eqs. (2) and (3) because during the preliminary tests, including
this band did not improve separability and accuracy. From the arithmetic
formulation of Eq. (3), it may be noted that the addition of the short
wave bands (bands 1 and 2) may result in large positive values for
high albedo surfaces such as ice, cloud, and highly reflective building
roofs. Eq. (3) may therefore not be able to distinguish these highalbedo
surfaces from water.
Therefore, the intended use of the two AWEI equations is as follows:
1) in situations where shadows are major sources of accuracy loss but
surfaces such as snow, ice and high albedo built surfaces are not present,
AWEIsh alone is proposed to automatically enhance the separability of
pixels of water from nonwater (more importantly from shadow pixels)
so that application of a threshold close to 0 is suitable for the extraction
of surface water; 2) in areas where shadows are not a major problem,
AWEInsh alone is proposed; 3) in conditions where both high albedo
surfaces and shadow/dark surfaces are found, we propose using
Eqs. (2) and (3) sequentially in a classification tree; 4) in areas with
no shadowed areas, no dark urban backgrounds and no high-albedo
surfaces, either of the two can be used alone.
3.4. Classification, threshold optimization and per-pixel accuracy assessment
At the test sites in Denmark and Ethiopia, urban background
dominates the sub-scenes. Therefore, both equations of AWEI (AWEInsh
and AWEIsh) were applied sequentially: first, AWEInsh was applied to the
image; next, AWEIsh was used to eliminate misclassified pixels with
shadows and other dark surfaces. At the test sites in Switzerland, New
Zealand and South Africa, only AWEIsh was applied because urban
surfaces are rare in these sites.
To compare accuracy of the proposed water extraction technique
with other methods, we made preliminary tests of various water indices
including the Water Index (WI) of Ouma and Tateishi (2006), the
Normalized Difference Water Index (NDWI) of McFeeters (1996) and
other indices that Ji et al. (2009) used in their studies. Based on this preliminary
evaluation, it appeared that all indices, except the MNDWI,
performed poorly at our test sites. We therefore only considered
MNDWI for comparison with the new index proposed in this paper. A supervised
maximum likelihood (ML) classifier was also included in our
comparison as this classifier is one of the most widely used methods in
land cover classification. For the ML classifier, water and nonwater training
data were produced for each test site. The minimum size of reference
datasets for training was determined using the multinomial conservative
sample size equation described in Congalton and Green (2009). The
Blue band (b1)
Reflectance
0.00
0.05
0.10
0.15
0.20
0.25
0.30
Green band (b2)
NIR (b4)
Reflectance
0.0
0.1
0.2
0.3
0.4 SWIR (b5)
Red band (b3)
SWIR (b7)
Eq. 2, AWEInsh
Forest
Water
Shadow
Non-for. green
Asphalt
Dark built
Bright built
Bright soil
Black soil
Brown soil
Indexvalue
-1.5
-1.0
-0.5
0.0
0.5
1.0
Eq. 3, AWEIsh
Surface category
Forest
Water
Shadow
Non-for. green
Asphalt
Dark built
Bright built
Bright soil
Black soil
Brown soil
MNDWI
Forest
Water
Shadow
Non-for. green
Asphalt
Dark built
Bright built
Bright soil
Black soil
Brown soil
Fig. 1. Reflectance distributions of pure pixels of major land cover types. Each box plots shows the location of the 10th, 25th, 50th, 75th, and 90th percentiles using horizontal lines (boxes
and whiskers) and the circles are 5th and 95th percentiles.
27G.L. Feyisa et al. / Remote Sensing of Environment 140 (2014) 23–35
reference data were generated by digitizing multiple polygons on the
true-color composites of Landsat bands and evenly distributing the samples
across all parts of the sub-scenes. It was easy to generate large reference
data units since the classes considered are only water and nonwater
and it is relatively easy to visually distinguish between water and
nonwater surfaces from high spatial resolution images retrieved through
Google Earth™. These images were used to differentiate nonwater dark
areas from water surfaces. No separate validation data were necessary
for accuracy assessment of the ML classifier since the classification result
was compared against the true map of water.
Since the AWEI equations are formulated to enhance separability of
water and nonwater pixels by applying coefficients that force nonwater
pixels below 0 and water pixels above 0, a threshold of 0 can be used as
a default starting point. But due to variation in scene brightness and
contrast with time and space, the default threshold may not always
result in the highest possible water extraction accuracy that can be
achieved by application of the index. In order to determine the optimal
threshold, multiple thresholds were considered, and for each threshold
value corresponding commission errors (over-estimation) and omission
errors (under-estimation) were calculated and the percentage
errors were plotted against threshold values. The intersection point of
commission and omission error graphs was then considered as the
optimal threshold since it approximates the minimum possible sum of
the two error types. We evaluated the stability of optimal thresholds
of the new method and of MNDWI by examining the variation of the
optimal threshold values for the two indices across the five test sites.
Classification accuracy of the three methods, i.e. AWEI, MNDWI and
ML, was assessed by calculating kappa coefficients and error matrices.
The accuracy comparison between AWEI and MNDWI was made at
their optimal thresholds. McNemar's statistical test was applied to
examine whether the new water extraction method significantly
improves accuracy compared to MNDWI and ML in the test sites.
McNemar's continuity corrected chi-square statistic was computed as
shown in Eq. (4) (De Leeuw et al., 2006):
X
2
¼
f 12−f 21j j−1ð Þ
2
f 12 þ f 21
ð4Þ
where, f12 and f21 denote the frequencies of cases that are correctly classified
by one classification method but wrongly classified by the other.
3.5. Sub-pixel accuracy assessment
The sensitivity of different classifiers to various mixtures of water
and nonwater was evaluated using sub-pixel commission–omission
errors and by plotting graphs showing the cumulative percentage of
edge pixels classified as water against the proportion of each individual
pixel covered by water for mixed edge pixels at test sites in Denmark,
Switzerland and Ethiopia. Detailed sub-pixel accuracy analysis and
comparisons were undertaken using the three reservoirs in Ethiopia
(Gefersa, Dire and Legedadi). The total number of mixed edge pixels
in the three reservoirs was 1819 (164 ha). In the sub-pixel accuracy assessment,
commission and omission errors brought about by edge
pixels were quantified by the use of an overlay analysis in ArcGIS. Any
pixels that included water and nonwater surfaces were considered to
be mixed edge pixels (Fig. 2). If a mixed edge pixel was classified as
water, the fraction of it that fell outside the “true” boundary was considered
to be sub-pixel commission error. Similarly, in cases where mixed
pixels are classified as nonwater, the fraction of these pixels that fell
inside the “true” water body was considered to be an omission error
at the sub-pixel level. Mixed pixels consisting predominantly of water
(N50% water) should ideally be classified as water and vice versa. In
the sub-pixel accuracy assessment, influences of misregistration artifacts
and manual digitization of true water boundaries were assumed
to be insignificant.
4. Results
4.1. Water extraction maps
The outputs of water extraction using the three classifiers at the five
test sites are presented in Fig. 3. Visual inspection of Fig. 3 indicated that
AWEI resulted in better accuracy of surface water mapping compared to
MNDWI and ML. Particularly at test sites in Switzerland, South Africa
and New Zealand, the new index (AWEI) was consistently better in suppressing
shadow and other nonwater surfaces. In most cases, MNDWI
Fig. 2. Edge pixels around Gefersa reservoir (Ethiopia) showing mixed pixels with different proportions of water (shown on high spatial resolution image accessed through Google Earth™).
28 G.L. Feyisa et al. / Remote Sensing of Environment 140 (2014) 23–35
and especially ML produced noisy results. However, at test sites in
Denmark and Ethiopia, visual inspection of Fig. 3 indicated smallest
difference among the three classification methods.
Visual inspection of classification outputs at the three additional test
sites shown in Appendix A (Figs. A1–A3) also indicates that AWEI is
effective in extracting surface water in the presence of shadow and
urban surfaces. At Bishoftu lakes in Ethiopia, where no major shadow
surfaces were present, both AWEIsh and MNDWI resulted in (visually)
similar classification outputs. By contrast, at the test sites in Norway
and Shanghai where dark shadows were abundant, visual inspection
clearly shows that AWEIsh suppressed shadowed surfaces more effectively
than MNDWI (shown in Appendix A).
4.2. Classification accuracy and edge pixel effects
The results of mapping accuracy at each of the five main test sites are
summarized in Table 4. At all test sites the accuracy achieved by AWEI
was higher than that of the MNDWI and ML classifiers. Averaged over
LMIWDNMIEWATest site
Denmark
Switzerland
Ethiopia
LMIWDNMIEWASite
S. Africa
New
Zealand
Fig. 3. Comparison of water extraction results using three classifiers at the five test sites.
29G.L. Feyisa et al. / Remote Sensing of Environment 140 (2014) 23–35
the five test sites the total omission and commission error of AWEI was
only about 50% of that of the MNDWI and 25% of that of the ML classifiers
(Fig. 4). Details of accuracy assessment including users' and producers'
accuracy are shown in Appendix A (Table A1). Since visual
inspection from Fig. 3 indicates small variation in accuracy at test sites
in Denmark and Ethiopia, McNemar's chi-square test of significance of
accuracy difference at the test sites in Denmark, Switzerland and
Ethiopia are included in Table 5. At these three test sites, significant accuracy
improvement was achieved by AWEI (P-value b 0.01) compared
to ML. At the test site in Denmark, accuracy difference between AWEI
and MNDWI was insignificant (Table 5). ML performed worst at test
site in South Africa (kappa coefficient 0.62) and at this test site, the
highest accuracy was achieved by AWEI, with a kappa coefficient of
0.98 (Table A1 in Appendix A).
The sub-pixel accuracy analysis is presented in Fig. 5. The comparison
shows the ability of the three classifiers in correctly classifying
edge pixels with various mixtures of water and nonwater components.
The vertical line in Fig. 5 indicates the 50% water–nonwater mixture and
the figure shows that among the edge pixels that AWEI classified as
water, only 13% were predominantly nonwater. Conversely, 87% of
mixed edge pixels that were classified as water were correctly classified
by AWEI. Using MNDWI, 81% of the mixed edge pixels were correctly
classified. Further analysis of mixed edge pixels at test sites in Addis
Ababa showed that sub-pixel commission error of AWEI corresponded
to an overestimation of 16.6 ha (total area of predominantly nonwater
edge pixels classified as water), and omission error corresponded to
4.3 ha (total area of predominantly water edge pixels classified as
nonwater). For comparison, edge pixel commission and omission errors
of MNDWI corresponded to overestimation and underestimation of
18.3 ha and 4 ha, respectively. At this site edge pixel omission and commission
error of ML corresponded to 49.3 ha and 0.6 ha, respectively.
Based on the sum of total overestimation and underestimation of edge
pixels, AWEI performed slightly better than MNDWI, and ML achieved
the lowest accuracy in classifying mixed edge pixels.
4.3. Optimal threshold and its variability
A comparison of the stability of the optimum thresholds of AWEI and
MNDWI is shown in Fig. 6. It clearly appears that the optimal threshold
of MNDWI at different test sites exhibited large variation compared
to AWEI. The optimal threshold of MNDWI ranged from 0.005 in
Denmark to 0.6 in South Africa, whereas for AWEI the optimal threshold
only varied from −0.15 (AWEInsh in Denmark) to 0.045 (AWEIsh in
South Africa), and in the three other sites the optimal threshold of
AWEIsh was 0 (Fig. 6).
5. Discussion and perspectives
The new water extraction index introduced in this paper contributes
to the efforts being made to improve the accuracy of surface water
mapping and change analysis for various environmental studies and
applications. This method uses a simple and systematic technique of
enhancing class separability without a need for additional data to
remove shadow and dark surface noises, which are often major causes
of misclassification in surface water mapping. Using a simple classification
tree approach, the AWEI was shown to extract surface water with
high accuracy, particularly in mountainous areas where hills cast
shadows on background surfaces and in urban areas with complex
land cover. AWEI is not only a simple technique but was also shown
to be robust under various environmental conditions and for different
types of water bodies.
Totalerror(%omissionandcommission)
0
10
20
30
40
50
60
South Africa
Switzerland
Ethiopia
Denmark
New Zealand
South Africa
Switzerland
Ethiopia
Denmark
New Zealand
South Africa
Switzerland
Ethiopia
Denmark
New Zealand
Classifier
AWEI MNDWI ML
Fig. 4. Total classification error (combined commission and omission error). The box plots
show the variability of classification errors among test sites. Each box plots shows the
location of the 10th, 25th, 50th, 75th, and 90th percentiles using horizontal lines
(boxes and whiskers) and the circles are 5th and 95th percentiles.
Table 5
Summary of McNemar's continuity corrected χ2 test for differences in classification
accuracy.
Test sites Classifier χ2
P-value
MNDWI ML MNDWI ML
Denmark AWEI 0.8 114 0.30 0.00
Switzerland AWEI 408.0 619 0.00 0.00
Ethiopia AWEI 19.0 30 0.00 0.00
AWEI
MNDWI
Max. Like.
Percentage of water in edge pixels
0 20 40 60 80 100
Cum.%ofedgepixelsclassifiedaswater
0
20
40
60
80
100
A A
A
A
A
A
A
A
A
A
A
M
M
M
M
M
M
M
M
M
M
M
L
L
L
L
L
L
L
L
L
L
L
A
M
L
Fig. 5. Cumulative frequency of mixed edge pixels classified as water (average of test sites
in Denmark, Switzerland and Ethiopia).
Table 4
Summary of classification accuracy of the three classifiers by test site.
Classifier Denmark Switzerland Ethiopia S. Africa New Zealand
Kappa coeff. Kappa coeff. Kappa coeff. Kappa coeff. Kappa coeff.
AWEI 0.93 0.95 0.97 0.98 0.98
MNDWI 0.92 0.89 0.95 0.94 0.90
ML 0.89 0.81 0.93 0.62 0.97
30 G.L. Feyisa et al. / Remote Sensing of Environment 140 (2014) 23–35
In many water indices, the lack of stability of the threshold is a problem
(Ji et al., 2009), making it difficult to decide which value should be
used in classification trees. The lack of a reasonably stable threshold
may make the classification more time-consuming and lead to a subjective
choice of threshold which may also affect accuracy. In addition to
accuracy improvement, our new index was also shown to have a relatively
stable optimal threshold which makes the use of the method
even simpler. It should be noted that in our study, images from all test
sites were atmospherically corrected applying the FLAASH module in
ENVI v.4.8. In classifying images that are calibrated to TOA reflectance,
but with no atmospheric corrections, the optimal thresholds may differ
slightly from what is observed in this study.
A number of authors contributed to previous research on the improvement
of surface water mapping accuracy using remotely sensed
data, including those that have emphasized the improvement of general
land cover classification accuracies (Aguirre-Gutierrez, Seijmonsbergen,
& Duivenvoorden, 2012; Rozenstein & Karnieli, 2011) and surface water
mapping in particular (Ji et al., 2009; Sun et al., 2012; Verpoorter et al.,
2012; Xu, 2006). In a recent work of Verpoorter et al. (2012), a six-step
water extraction method called GeoCover Water Bodies Extraction
Method (GWEM) was introduced. In GWEM, the authors proposed a
combination of various classification techniques for improvement of
accuracy. The same authors identified shadows of cloud and mountains
as major sources of accuracy problems and proposed a method where
elevation data were used to detect shadows and water surfaces that
overlap with shadow were removed from the classification dataset.
Our new method which automatically suppresses shadow pixels without
the need for other data input or separate shadow detection procedures,
may ease surface water mapping, particularly in situations
where mapping, monitoring and change detection of surface water
resources across multiple scenes or over regional and global scales are
required.
Despite a number of surface water mapping and accuracy improvement
methods reported in the literature, limited research has been undertaken
on accuracy assessment at sub-pixel level. This is particularly
important when satellite images such as Landsat are used. Because of
the limited spatial resolution of reflective bands of Landsat TM, edge
pixels cover relatively large areas which likely consist of a mixture of
water and nonwater components. In the use of Landsat TM data for
environmental studies where monitoring and detecting changes in
waterline are of interest, the accuracy of classifying mixed edge pixels
may become an important issue.
As mentioned in the Results section, when applying the ML classifier
to the reservoirs around Addis Ababa, a substantial number of edge
pixels that predominantly consist of water were classified as nonwater,
thus obviously leading to underestimation of surface water extents. The
implication of this could be that even if the water boundary increases by
certain distance, thus changing the proportion of water in mixed edge
pixels, say from 40% to 60%, the ML method could still classify the pixels
as nonwater since it seems that the classifier is sensitive to nonwater
components of the mixture. The ability of different classifiers to classify
such mixed pixels correctly into water and nonwater classes may
vary depending on the spectral bands and algorithms used. The
nonwater components of the mixture could be composed of many
combinations of different land cover types. So, the reflectance values
of mixed pixels can vary considerably, even for pixels where the proportion
of water is similar (Ji et al., 2009). The relative improvement
in sub-pixel accuracy achieved by AWEI may make it suitable for
consistent and reliable estimation of surface water dynamics using
Landsat data.
Though the new water extraction index was tested under wide
range of environmental conditions and water body types, several variables
that were not considered at our test sites are likely to affect the accuracy
of water extraction methods. Seasonal and daily variation in the
angle of the sun, atmospheric composition, and changes in biophysical
and chemical properties of water bodies, such as changes in phytoplankton
(Zhang et al., 2010) may influence the reflectance patterns of
water bodies. The use of different atmospheric correction methods
may also influence thresholds and accuracies. Therefore one may need
to consider the importance and type of atmospheric correction applied
in the image preprocessing stage in evaluating accuracies of different
water extraction methods. AWEI was tested using Landsat TM data
only and its use may therefore need to be evaluated on data from
other sensors.
In our test cases, we did not consider the influence of seasonal variation
in appearances of water bodies. Therefore, the robustness of the
new method also needs to be tested in different seasons. In addition
to the five test sites examined in details, we included three additional
sites in Appendix A (Figs. A1–A3). Nevertheless, more sites may need
to be included for a thorough evaluation of the performance of the
index.
6. Conclusion
The main purpose of this study was to devise a method that improves
water extraction accuracy by increasing spectral separability between
water and nonwater surfaces, particularly in areas with shadows
and urban backgrounds that are often major causes of low classification
accuracy. Using Landsat 5 TM data, we introduced a new automated
water extraction method (AWEI) and compared its per-pixel and subpixel
accuracy and threshold stability with that of the MNDWI and
ML classifiers. AWEI significantly improved accuracy in areas where
shadow and other dark surfaces were the main sources of classification
errors.
A sub-pixel analysis of errors at the edges of water bodies revealed
that the AWEI classifier was relatively more accurate in classifying
edge pixels compared to the MNDWI and ML classification methods. Besides,
the optimal threshold of AWEI was shown to be less variable with
images of different locations and times compared to that of MNDWI.
Therefore, AWEI is proposed as an alternative and improved water
index, especially in extracting water information from areas where
noisy results are expected because of the presence of shadows and
built-up surfaces. This new method would also be suitable for surface
water change detection studies since it classifies edge pixels with high
accuracy and with a stable threshold.
AWEI MNDWI
Denmark
Switzerland
Ethiopia
South Africa
New Zealand
Denmark
Switzerland
Ethiopia
South Africa
New Zealand
Indexvalue
-1.0
-0.5
0.0
0.5
1.0
x
x x x x x
x
x
x
x
nsh
sh sh
sh
sh
Fig. 6. Threshold variability and distribution of index values for AWEI and MNDWI. Dashed
lines show mean optimal threshold of the five test sites, and symbol “x” shows optimal
threshold for each site.
31G.L. Feyisa et al. / Remote Sensing of Environment 140 (2014) 23–35
Appendix A
AWEIsh at 0 threshold
MNDWI at 0.0 threshold
Landsat true color composite
AWEIsh at 0.17 threshold
Google 3D viewshowing water and mountain shadow
MNDWI at 0.7 threshold
Fig. A1. Water extraction images applying AWEIsh (top row) and MNDWI (middle row) at test site in Norway using Landsat 5 TM acquired on Aug 13, 2011 (2382 by 2382 pixels, top-left
corner coordinate in UTM: 6,989,655 N, 371,295 E).
32 G.L. Feyisa et al. / Remote Sensing of Environment 140 (2014) 23–35
AWEIsh at 0.0 threhold
X
X
X
X
X
X
X
X
X
X
AWEIsh at 0.1 threhold MNDWI at 0.15 thresholdMNDWI at 0.0 threshold
Fig. A3. Water extraction images applying AWEIsh (top row) and MNDWI (middle row) at test site in Bishftu Ethiopia using Landsat ETM+ acquired on Nov 27, 2002 (400 by 400 pixels,
top-left corner coordinate UTM: 3,461,925 N, 350,415 E). Location of actual water bodies is shown by “X” mark on true color composite of the Landsat ETM+ image (bottom image).
AWEIsh at 0.15 threshold
MNDWI at 0.2 threshold
Landsat true color composite Google 3D view showing building shadow and river
MNDWI at 0.5 t hresholdMNDWI at 0.4 threshold
AWEIsh at 0.25 thresholdAWEIsh at 0.2 threshold
Fig. A2. Water extraction images applying AWEIsh (top row) and MNDWI (middle row) at test site in China Shanghai using Landsat ETM+ acquired on Nov 27, 2002 (400 by 400 pixels,
top-left corner coordinate UTM: 3,461,925 N, 350,415 E).
33G.L. Feyisa et al. / Remote Sensing of Environment 140 (2014) 23–35
Table A1
Summary of accuracy assessments at the five main test sites showing various accuracy measures.
Test site Classification method Threshold Land cover class User accu. Produc accu. Kappa Comm. error % Omi. error % Total error %
Denmark a
AWEInsh 0.000 Water 97.08 91.43 0.93 2.92 8.57 11.49
Nonwater 98.57 99.54 1.43 0.46 1.89
−0.2 Water 96.35 92.89 0.94 3.65 7.11 10.77
Nonwater 98.81 99.41 1.19 0.59 1.79
−0.15 Water 98.30 92.58 0.95 1.70 7.42 9.12
Nonwater 98.75 99.72 1.25 0.28 1.53
MNDWI 0.00 Water 95.13 91.04 0.92 4.87 8.96 13.82
Nonwater 98.50 99.22 1.50 0.78 2.28
0.05 Water 97.10 89.89 0.92 2.90 10.11 13.02
Nonwater 98.32 99.55 1.68 0.45 2.14
0.1 Water 98.05 88.65 0.92 1.95 11.35 13.31
Nonwater 98.12 99.70 1.88 0.30 2.18
MaxLike Water 96.85 84.61 0.89 3.15 15.39 18.53
Nonwater 97.46 99.54 2.54 0.46 3.00
Switzerland AWEIsh −0.050 Water 52.89 96.77 0.66 47.1 3.2 50.3
Nonwater 99.84 95.81 0.2 4.2 4.4
0.100 Water 99.34 76.61 0.86 0.7 23.4 24.1
Nonwater 98.88 99.98 1.1 0.0 1.1
0.000 Water 99.01 90.96 0.95 1.0 9.0 10.0
Nonwater 99.56 99.96 0.4 0.0 0.5
MNDWI −0.005 Water 99.34 76.61 0.94 0.7 23.4 24.1
Nonwater 99.84 95.81 0.2 4.2 4.4
0.100 Water 73.92 95.64 0.82 26.1 4.4 30.4
Nonwater 99.79 98.36 0.2 1.6 1.9
0.300 Water 87.99 91.92 0.89 12.0 8.1 20.1
Nonwater 99.61 99.39 0.4 0.6 1.0
0.500 Water 96.76 78.95 0.86 3.2 21.1 24.3
Nonwater 98.99 99.87 1.0 0.1 1.1
MaxLike – Water 74.46 92.08 0.81 25.5 7.9 33.5
Nonwater 99.61 98.47 0.4 1.5 1.9
Ethiopia AWEIsh −0.050 Water 95.47 98.85 0.95 4.53 1.15 5.68
Nonwater 100.00 99.97 0.00 0.03 0.04
0.000 Water 95.47 98.85 0.97 4.53 1.15 5.68
Nonwater 100.00 99.98 0.00 0.02 0.02
0.100 Water 96.60 91.48 0.94 3.40 8.52 11.92
Nonwater 99.97 99.99 0.03 0.01 0.05
MNDWI 0.000 Water 92.57 96.66 0.95 7.43 3.34 10.77
Nonwater 99.99 99.97 0.01 0.03 0.04
0.100 Water 95.01 95.39 0.95 4.99 4.61 9.60
Nonwater 99.98 99.98 0.02 0.02 0.04
0.150 Water 97.53 92.66 0.95 2.47 7.34 9.81
Nonwater 99.97 99.99 0.03 0.01 0.04
MaxLike – Water 99.71 86.96 0.93 0.29 13.04 13.33
Nonwater 99.95 100.00 0.05 0.00 0.05
S. Africa AWEIsh 0.020 Water 83.23 98.86 0.90 16.77 1.14 17.91
Nonwater 99.97 99.50 0.03 0.50 0.53
0.045 Water 98.32 98.30 0.98 1.68 1.70 3.38
Nonwater 99.96 99.96 0.04 0.04 0.09
0.060 Water 98.43 97.51 0.98 1.57 2.49 4.06
Nonwater 99.94 99.96 0.06 0.04 0.10
MNDWI 0.300 Water 70.54 97.67 0.81 29.46 2.33 31.79
Nonwater 99.94 98.97 0.06 1.03 1.09
0.450 Water 89.07 96.37 0.92 10.93 3.63 14.56
Nonwater 99.91 99.70 0.09 0.30 0.39
0.600 Water 94.46 93.58 0.94 5.54 6.42 11.96
Nonwater 99.84 99.86 0.16 0.14 0.30
MaxLik – Water 46.74 97.12 0.62 53.26 2.88 56.14
Nonwater 99.93 97.21 0.07 2.79 2.87
N. Zealand AWEIsh −0.100 Water 98.74 99.87 0.96 1.26 0.13 1.39
Nonwater 99.29 93.65 0.71 6.35 7.06
0.000 Water 99.82 99.56 0.98 0.18 0.44 0.61
Nonwater 97.85 99.13 2.15 0.87 3.02
0.100 Water 99.90 99.45 0.98 0.10 0.55 0.65
Nonwater 97.33 99.51 2.67 0.49 3.17
MNDWI 0.000 Water 96.93 99.79 0.89 3.07 0.21 3.28
Nonwater 98.79 84.23 1.21 15.77 16.98
0.150 Water 97.39 99.44 0.90 2.61 0.56 3.16
Nonwater 96.91 86.72 3.09 13.28 16.37
0.200 Water 98.14 98.50 0.90 1.86 1.50 3.35
Nonwater 92.39 90.71 7.61 9.29 16.89
MaxLike – Water 99.84 99.18 0.97 0.16 0.82 0.98
Nonwater 96.05 99.22 3.95 0.78 4.73
a
At this test site, shadow is not a major source of classification noise but built-up surfaces are predominant land cover type. Therefore, the use of AWEInsh resulted in high accuracy of
water extraction (a combined use of both AWEIsh and AWEInsh did not improve accuracy).
34 G.L. Feyisa et al. / Remote Sensing of Environment 140 (2014) 23–35
References
Aguirre-Gutierrez, J., Seijmonsbergen, A.C., & Duivenvoorden, J. F. (2012). Optimizing land
cover classification accuracy for change detection, a combined pixel-based and
object-based approach in a mountainous area in Mexico. Applied Geography, 34,
29–37.
Alderman, K., Turner, L. R., & Tong, S. L. (2012). Floods and human health: A systematic
review. Environment International, 47, 37–47.
Bond, N. R., Lake, P.S., & Arthington, A. H. (2008). The impacts of drought on freshwater
ecosystems: An Australian perspective. Hydrobiologia, 600, 3–16.
Butcher, S. S., Charlson, R. J., Orians, G. H., & Wolfe, G. V. (Eds.). (1992). Global
biogeochemical cycles. London: Academic Press Limited.
Charoenpanyanet, A., & Chen, X. (2008). Satellite-based modeling of Anopheles mosquito
densities on heterogeneous land cover in Western Thailand. The International Archives
of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 37, 159–164.
Charron, D. F., Thomas, M. K., Waltner-Toews, D., Aramini, J. J., Edge, T., Kent, R. A., et al.
(2004). Vulnerability of waterborne diseases to climate change in Canada: A review.
Journal of Toxicology and Environmental Health-Part a-Current Issues, 67, 1667–1677.
Congalton, R. G., & Green, K. (2009). Assessing the accuracy of remotely sensed data:
Principles and practices (2nd ed.)Boca Raton: Lewis Publishers.
COWI (2010). DDOland2010: Danish National Digital Orthophoto, version 2010. http://
www.geodata-info.dk/Portal/ShowMetadata.aspx?id=d6a57343–c930–4d28–94ce-
85b5ca2847dc
Dambach, P., Machault, V., Lacaux, J. -P., Vignolles, C., Sié, A., & Sauerborn, R. (2012). Utilization
of combined remote sensing techniques to detect environmental variables
influencing malaria vector densities in rural West Africa. International Journal of
Health Geographics, 11.
Davranche, A., Lefebvre, G., & Poulin, B. (2010). Wetland monitoring using classification
trees and SPOT-5 seasonal time series. Remote Sensing of Environment, 114, 552–562.
De Leeuw, J., Jia, H., Yang, L., Liu, X., Schmidt, K., & Skidmore, A. K. (2006). Comparing accuracy
assessments to infer superiority of image classification methods. International
Journal of Remote Sensing, 27, 223–232.
Dewan, A.M., Islam, M. M., Kumamoto, T., & Nishigaki, M. (2007). Evaluating flood hazard
for land-use planning in Greater Dhaka of Bangladesh using remote sensing and GIS
techniques. Water Resources Management, 21, 1601–1612.
Duan, Z., & Bastiaanssen, W. G. M. (2013). Estimating water volume variations in lakes
and reservoirs from four operational satellite altimetry databases and satellite imagery
data. Remote Sensing of Environment, 134, 403–416.
Exelis Visual Information Solutions (2010). ENVI v. 4.8. Boulder, Colorado: EXELIS.
Frey, H., Huggel, C., Paul, F., & Haeberli, W. (2010). Automated detection of glacier lakes
based on remote sensing in view of assessing associated hazard potentials. Grazer
Schriften der Geographie und Raumforschung, 45, 261–272.
Gardelle, J., Hiernaux, P., Kergoat, L., & Grippa, M. (2009). Less rain, more water in ponds:
A remote sensing study of the dynamics of surface waters from 1950 to present
in pastoral Sahel (Gourma region, Mali). Hydrology and Earth System Sciences
Discussions, 6, 5047–5083.
Giardino, C., Bresciani, M., Villa, P., & Martinelli, A. (2010). Application of remote sensing
in water resource management: The case study of Lake Trasimeno, Italy. Water
Resources Management, 24, 3885–3899.
Guttler, F. N., Niculescu, S., & Gohin, F. (2013). Turbidity retrieval and monitoring of
Danube Delta waters using multi-sensor optical remote sensing data: An integrated
view from the delta plain lakes to the western-northwestern Black Sea coastal
zone. Remote Sensing of Environment, 132, 86–101.
Haas, E. M., Bartholomé, E., & Combal, B. (2009). Time series analysis of optical remote
sensing data for the mapping of temporary surface water bodies in sub-Saharan
western Africa. Journal of Hydrology, 370, 52–63.
He, B., Oki, K., Wang, Y., Oki, T., Yamashiki, Y., Takara, K., et al. (2012). Analysis of stream
water quality and estimation of nutrient load with the aid of Quick Bird remote sensing
imagery. Hydrological Sciences JournalJournal Des Sciences Hydrologiques, 57,
850–860.
Hui, F., Xu, B., Huang, H., Yu, Q., & Gong, P. (2008). Modelling spatial–temporal change of
Poyang Lake using multitemporal Landsat imagery. International Journal of Remote
Sensing, 29, 5767–5784.
Jain, S. K., Saraf, A. K., Goswami, A., & Ahmad, T. (2006). Flood inundation mapping using
NOAA AVHRR data. Water Resources Management, 20, 949–959.
Jain, S. K., Singh, R. D., Jain, M. K., & Lohani, A. K. (2005). Delineation of flood-prone areas
using remote sensing technique. Water Resources Management, 19, 337–347.
Ji, L., Zhang, L., & Wylie, B. (2009). Analysis of dynamic thresholds for the normalized difference
water index. Photogrammetric Engineering and Remote Sensing, 75, 1307–1317.
Jiang, Z., Qi, J., Su, S., Zhang, Z., & Wu, J. (2012). Water body delineation using index composition
and HIS transformation. International Journal of Remote Sensing, 33, 3402–3421.
Jimenez-Munoz, J. C., Sobrino, J. A., Mattar, C., & Franch, B. (2010). Atmospheric correction
of optical imagery from MODIS and reanalysis atmospheric products. Remote Sensing
of Environment, 114, 2195–2210.
Kondo, H., Seo, N., Yasuda, T., Hasizume, M., Koido, Y., Ninomiya, N., et al. (2002).
Post-flood-infectious diseases in Mozambique. Prehospital and Disaster Medicine, 17,
126–133.
Lacaux, J. P., Tourre, Y. M., Vignolles, C., Ndione, J. A., & Lafaye, M. (2007). Classification of
ponds from high-spatial resolution remote sensing: Application to Rift Valley Fever
epidemics in Senegal. Remote Sensing of Environment, 106, 66–74.
Lake, P.S. (2003). Ecological effects of perturbation by drought in flowing waters.
Freshwater Biology, 48, 1161–1172.
Li, K. Z., Wu, S. H., Dai, E. F., & Xu, Z. C. (2012). Flood loss analysis and quantitative risk
assessment in China. Natural Hazards, 63, 737–760.
Lillesand, T. M., Kiefer, R. W., & Chipman, J. W. (2004). Remote sensing and image interpretation
(5th ed.): John Wiley & Sons, Inc.
Lira, J. (2006). Segmentation and morphology of open water bodies from multispectral
images. International Journal of Remote Sensing, 27, 4015–4038.
McFeeters, S. K. (1996). The use of Normalized Difference Water Index (NDWI) in the
delineation of open water features. International Journal of Remote Sensing, 17,
1425–1432.
Morss, R. E., Wilhelmi, O. V., Downton, M. W., & Gruntfest, E. (2005). Flood risk, uncertainty,
and scientific information for decision making — Lessons from an interdisciplinary
project. Bulletin of the American Meteorological Society, 86 (1593-+).
Mucher, C. A., Steinnocher, K. T., Kressler, F. P., & Heunks, C. (2000). Land cover characterization
and change detection for environmental monitoring of pan-Europe.
International Journal of Remote Sensing, 21, 1159–1181.
NASA (2012). Landsat 7 science data users handbook. Online.
Novoa, S., Chust, G., Sagarminaga, Y., Revilla, M., Borja, A., & Franco, J. (2012). Water quality
assessment using satellite-derived chlorophyll-a within the European directives,
in the southeastern Bay of Biscay. Marine Pollution Bulletin, 64, 739–750.
Ouma, Y. O., & Tateishi, R. (2006). A water index for rapid mapping of shoreline changes
of five East African Rift Valley lakes: An empirical analysis using Landsat TM and
ETM+ data. International Journal of Remote Sensing, 27, 3153–3181.
Poulin, B., Davranche, A., & Lefebvre, G. (2010). Ecological assessment of Phragmites
australis wetlands using multi-season SPOT-5 scenes. Remote Sensing of Environment,
114, 1602–1609.
Prigent, C., Papa, F., Aires, F., Jimenez, C., Rossow, W. B., & Matthews, E. (2012). Changes in
land surface water dynamics since the 1990s and relation to population pressure.
Geophysical Research Letters, 39, L08403.
Proud, S. R., Fensholt, R., Rasmussen, L. Y., & Sandholt, I. (2011). Rapid response flood detection
using the MSG geostationary satellite. International Journal of Applied Earth
Observation and Geoinformation, 13, 536–544.
Richards, J. A. (1993). Remote sensing digital image analysis: An introduction (2nd ed.)
Berlin: Springer-Verlag.
Rogers, A. S., & Kearney, M. S. (2004). Reducing signature variability in unmixing coastal
marsh Thematic Mapper scenes using spectral indices. International Journal of Remote
Sensing, 25, 2317–2335.
Rozenstein, O., & Karnieli, A. (2011). Comparison of methods for land-use classification
incorporating remote sensing and GIS inputs. Applied Geography, 31, 533–544.
Ryu, J. H., Won, J. S., & Min, K. D. (2002). Waterline extraction from Landsat TM data in a
tidal flat — A case study in Gomso Bay, Korea. Remote Sensing of Environment, 83,
442–456.
Sethre, P. R., Rundquist, B. C., & Todhunter, P. E. (2005). Remote detection of prairie pothole
ponds in the Devils Lake Basin, North Dakota. GIScience and Remote Sensing, 42,
277–296.
Sheng, Y. W., Shah, C. A., & Smith, L. C. (2008). Automated image registration for hydrologic
change detection in the lake-rich Arctic. IEEE Geoscience and Remote Sensing
Letters, 5, 414–418.
Sun, F., Sun, W., Chen, J., & Gong, P. (2012). Comparison and improvement of methods for
identifying waterbodies in remotely sensed imagery. International Journal of Remote
Sensing, 33, 6854–6875.
United States Geological Survey (USGS) (2012). Landsat data archive. Global Visualization
Viewer (GLOVIS).
Verpoorter, C., Kutser, T., & Tranvik, L. (2012). Automated mapping of water bodies using
Landsat multispectral data. Limnology and Oceanography-Methods, 10, 1037–1050.
Xu, H. (2006). Modification of normalised difference water index (NDWI) to enhance
open water features in remotely sensed imagery. International Journal of Remote
Sensing, 27, 3025–3033.
Zhang, Y. L., Feng, L. Q., Li, J. S., Luo, L. C., Yin, Y., Liu, M. L., et al. (2010). Seasonal–spatial
variation and remote sensing of phytoplankton absorption in Lake Taihu, a large eutrophic
and shallow lake in China. Journal of Plankton Research, 32, 1023–1037.
35G.L. Feyisa et al. / Remote Sensing of Environment 140 (2014) 23–35