The Egyptian Journal of Remote Sensing and Space Sciences (2015) 18, 99106

H O S T E D BY

National Authority for Remote Sensing and Space Sciences

The Egyptian Journal of Remote Sensing and Space

Sciences
www.elsevier.com/locate/ejrs
www.sciencedirect.com

RESEARCH PAPER

Flood monitoring and damage assessment using

water indices: A case study of Pakistan ood-2012
Akhtar Ali Memon

a,*

, Sher Muhammad b, Said Rahman a, Mateeul Haq

Pakistan Space & Upper Atmosphere Research Commission (SUPARCO), Pakistan

Key Laboratory of Tibetan Environmental Changes and Land Surface Processes, Institute of Tibetan Plateau Research,
Chinese Academy of Sciences, China
b

Received 7 August 2014; revised 4 February 2015; accepted 22 March 2015

Available online 16 April 2015

KEYWORDS
Water Index;
Flood-2012;
Pakistan oods;
Damage assessment;
MODIS;
Landsat ETM+

Abstract This paper uses Normalized Difference Water Index (NDWI) of McFeeters (1996),
Water Index (WI) introduced by Rogers and Kearney (2004), referred to as Red and Short
Wave Infra-Red (RSWIR) and WI suggested as the best by Ji et al. (2009), referred to as Green
and Short Wave Infra-Red (GSWIR) for delineating and mapping of surface water using
MODIS (Terra) near real time images during 2012 oods in Pakistan. The results from above
indices have been compared with Landsat ETM+ classied images aiming to assess the accuracy
of the indices. Accuracy assessment has been performed using spatial statistical techniques and
found NDWI, RSWIR and GSWIR with kappa coefcient (j) of 46.66%, 70.80% and 60.61%
respectively. It has been observed using statistical analysis and visual interpretation (expert knowledge gained by past experience) that the NDWI and GSWIR have tendencies to underestimate and
overestimate respectively the inundated area. Keeping in view the above facts, RSWIR has proved
to be the best of the three indices. In addition, assessment of the damages has been carried out
considering accumulated ood extent obtained from RSWIR. The information derived proved to
be essential and valuable for disaster management plan and rehabilitation.
2015 National Authority for Remote Sensing and Space Sciences. Production and hosting by Elsevier
B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/
by-nc-nd/4.0/).

1. Introduction
Pakistan is a ood-prone country with a history of widespread
and repeated ooding that causes loss of lives, substantial
damage to property, infrastructure, loss of agricultural crops
and land (Sardar et al., 2008). It has two dominant types of
oods i.e. riverine and ash. The riverine ood, mainly
* Corresponding author.
Peer review under responsibility of National Authority for Remote
Sensing and Space Sciences.

originating from glacier melt as well as extreme weather events,

is predictable which in result allows sufcient time to react
whereas the converse is true for ash oods. Flash oods,
originating from extreme weather events, have relatively less
duration but intensity and impacts are severe. These oods
normally occur during the South Asian monsoon period
between July and September.
The country has been continuously experiencing major
oods since 2001. The major ood in 2010 inundated an area
of 70,238 km2 and affected around 884,715 homes (Haq,
2010). In 2011, another major ood caused an inundation of

http://dx.doi.org/10.1016/j.ejrs.2015.03.003
1110-9823 2015 National Authority for Remote Sensing and Space Sciences. Production and hosting by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

100
21,200 km2, affected 5.88 million people, 1500 km of road network, 382 km of railway tracks, 498 km2 of forests and
16,440 km2 of agricultural land (Haq et al., 2012). In this year,
moderate scale rains have been observed, starting from August
and intensied in September, resulting in huge economic
losses, destruction of ecological sources, food shortage and
starvation of thousands of people. During the rst half of
September, widespread rains were experienced throughout
the country, which severely affected the areas of Punjab,
Sindh and Balochistan provinces. During 0611 September,
Jacobabad, Larkana, Sukkur, Thatta and Tharparkar in
Sindh received heavy rains resulting in ood. In addition, rains
in Balochistan not only brought disaster in Balochistan but
also the districts of Sindh received devastating ash oods
due to hill torrents in the Suleman range. In Punjab, Rahim
Yar Khan, Sahiwal, Okara, Jehlum, Mandi Bahaudin, DG
Khan and Rajanpur experienced extensive and continuous
downpour from 08 to 12 September. The available daily rainfall data of 8 districts for the months of August and September
are given in Table 1(a) and (b).
Estimation of Land surface water (area), which includes
rivers, lakes, wetlands and ood inundated areas, plays an
important role in the evaluation of the water cycle. With global
climate change and the growing inuence of human activities,
river changes (Xu, 2004; Xia and Zhang, 2008), shrinkage of
wetlands, glacier melting (Chen et al., 2007), ooding, and
other spatial and temporal distribution changes in surface
water resources are increasingly highlighted. A variety of studies have been focused on impact assessment to ood disaster
on different groups of people and communities (e.g.,
Schmuck-Widmann, 1996; Rashid, 2000; Rasid and Haider,
2003) and found the poor people to be hardest hit (IPPC,
2001).
Modern techniques of remote sensing provide tremendous
potential for monitoring and managing dynamic changes in
large surface water bodies; extracting hydrological parameters,
and modeling the water balance (El Bastawesy, 2015;
El-Gamily et al., 2010). McFeeters (1996) proposed the
Normalized Difference Water Index (NDWI) using Near
Infrared (NIR) and green channels of Landsat that can delineate and enhance open water features. Xu (2006) modied the
NDWI proposed by McFeeters (1996) by substituting the middle infrared band for NIR band. The NDWI dened as (re
(0.86 lm)  re (1.24 lm))/(re (0.86 lm) + re (1.24 lm)),
was introduced by Gao (1996) for extracting vegetation liquid
water. Rogers and Kearney (2004) introduced NDWI using
red and Short Wave Infra-Red (SWIR) channels of Landsat
TM. Xiao et al. (2005) dened a Land Surface Water Index
(LSWI) using NIR and SWIR bands of MODIS. Ouma and
Tateishi (2006) proposed a new Water Index (WI) for shoreline
edge extraction using a logical combination of Tasseled Cap
Wetness (TCW) index and NDWI using Landsat Thematic
Mapper (TM) and Enhanced Thematic Mapper Plus
(ETM+) data. Wu et al. (2008) introduced a new combinated
method for extraction of water body information using
ETM+. The method is based on combining the spectral
relationship between different bands with the TCW index
and removing other objects from the mask image using maximum likelihood classication methods. Li and Zhou (2009)
presented a method for Mountain area Enhanced Water-body
Identication (MEWI) through the Normalized Difference
Vegetation Index (NDVI) and red band of MODIS. Ji et al.

A.A. Memon et al.

(2009) used spectral data to simulate the satellite sensors data
and calculated the simulated NDWI in various forms and concluded that the calculated NDWI from (Green  SWIR)/
(Green + SWIR), where the SWIR is a region from 1.2 to
1.8 lm, has the most stable threshold. Ling et al. (2010) used
the Combined Index of NDVI and NIR for Water Body
Identication (CIWI) aiming to extract coastal wetlands water
information using Landsat TM images. Lu et al. (2011) proposed an integrated water body mapping method through
combination of difference between NDVI and NDWI
(NDVINDWI) with slope and NIR band using HJ-1A/B
satellite images.
In this paper, we have selected three water indices
proposed by McFeeters (1996), Rogers and Kearney (2004)
and Ji et al. (2009). The methods selected in this study are
due to their generality in delineating and enhancing surface
water features. The high temporal resolution of MODIS
images used in this study played a major role in monitoring
the 2012 oods in Pakistan. The paper recommends the best
of the three mentioned indices for mapping ood inundated
areas and to assess the damage caused by the 2012 oods in
the country.
2. Study area
The study area includes 20 adjacent districts of the
Baluchistan, Khyber Pakhtunkhwa, Punjab and Sindh
provinces of Pakistan ranging from 2650 13.900 to 33130 2.2100
in latitude and 6760 52.6800 to 7200 15.8600 in longitude
(Fig. 1). On average the mountainous area of the study area
receives 200960 mm rain annually and rest of the area, which
is mostly plain, receives 180460 mm rain annually.
3. Methodology
Five images of MODIS onboard the TERRA satellite acquired
in 0.5 km resolution through NASAs website http://
rapidre.sci.gsfc.nasa.gov/realtime/, have been used as a
main input to map ood inundated areas. Images have been
geo-referenced using the header le allied with the data. The
study area has been extracted from the images using District
administrative boundaries. Cloud shadow was frequently
identied as water, by most water indices. So careful removal
of cloud shadow using visual interpretation has been
carried out. Cloud masking has been performed using NIR
(band 2 of MODIS) by clipping the image that has reectance
above 0.7 as the spectral reectance of cloud is greater than 0.7
in NIR region.
Flood mapping has been carried out using NDWI, RSWIR
and GSWIR as shown in Eqs. 13 respectively by setting the
threshold from 0 to 1. Two images of Landsat ETM+, having
path 152, row 041, of September 17, 2012 and October 03,
2012, have been acquired for accuracy assessment of the results
obtained through mentioned indices. A standard supervised
maximum likelihood classication of Landsat ETM+ images
has been performed. Validation of the classication has been
performed by visual interpretation of Landsat ETM+ images
using false color 543 composite. Contingency matrix and its
allied statistics (i.e. overall accuracy, producers accuracy,
users accuracy and kappa coefcient (j)), reported at
Congalton (1991), have been performed for accuracy

(a) Daily rainfall (mm) data for the month of August 2012 (b) Daily rainfall (mm) data for the month of September 2012.

Districts

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

(a)
Dera Ismail Khan
Bhakkar
Dera Ghazi Khan
Rahim Yar Khan
Dadu
Jacobabad
Larkana
Sukkur

0
0
0
0
0
0
0
0

0
0
0
0
0
0
0
0

0
0
0
0
0
0
0
0

0
0
0
0
0
0
0
0

49
21
0
0
0
0
0
0

0
0
0
0
0
0
0
0

0
0
0
14
0
0
0
0

0
0
0
0
0
0
0
0

0
0
0
0
0
0
0
0

0
0
4
0
0
0
0
0

16
22
0
0
0
0
0
0

0
0
0
0
0
0
0
0

0
0
0
0
0
0
0
0

0
0
0
0
0
0
0
0

0
0
0
0
0
0
0
0

1
0
1
0
0
0
0
0

0
0
0
5
0
0
0
0

0
0
0
0
0
0
0
0

0
0
0
0
0
0
0
0

0
0
0
0
0
0
0
0

0
0
0
0
0
0
0
0

0
0
0
0
0
0
0
0

0
16
0
0
0
0
0
0

0
0
2
0
0
0
0
0

0
0
0
0
0
0
0
0

5
2
0
0
0
0
0
0

63
0
0
0
0
0
0
0

0
0
0
0
0
0
0
0

0
5
0
0
0
0
0
2

0
0
0
0
0
0
0
0

0
0
0
0
0
0
0
0

(b)
Districts
Dera Ismail Khan
Bhakkar
Dera Ghazi Khan
Rahim Yar Khan
Dadu
Jacobabad
Larkana
Sukkur

1
0
0
0
0
0
0
0
0

2
0
0
0
0
0
0
0
0

3
0
0
0
0
0
0
0
0

4
49
0
1
0
0
0
0
0

9
13
15
35
102
2
9
47
4

10
8
20
76
32
0

58
164

11
0
0
4
96
5
143
58
36

12
0
0
0
0
0
0
0
0

13
0
0
0
0
0
0
0
0

14
0
0
0
0
0
0
0
1

15
0
0
0
0
0
0
0
0

16
0
7
0
0
0
0
0
0

17
0
54
0
0
0
0
0
0

18
34
0
0
0
0
0
0
0

19
0
0
0
0
0
0
0
0

20
0
0
0
0
0
0
0
0

21
0
0
0
0
0
0
0
0

22
0
0
0
0
0
0
0
0

23
0
0
0
0
0
0
0
0

24
0
0
0
0
0
0
0
0

25
0
0
0
0
0
0
0
0

26
0
0
0
0
0
0
0
0

27
0
0
1
0
0
0
0
0

28
0
0
0
0
0
0
0
0

29
0
0
0
0
0
0
0
0

30
0
0
0
0
0
0
0
0

5
17
56
8
0
0
0
0
0

6
0
0
2
9
0
0
1
0

7
0
3
0
0
82
24
42
1

8
0
0
5
6
0
0
10
1

Flood monitoring and damage assessment using water indices

Table 1

data not available.

101

102

A.A. Memon et al.

Figure 1

Figure 2

Location of study area.

Inundated areas obtained through mentioned indices on different dates.

Table 2 Accuracy assessment of NDWI using error matrix.

Values show the area in km2.
NDWI results

MODIS
Water
Others
Total

Landsat ETM+

Table 3 Accuracy assessment of RSWIR using error matrix.

Values show the area in km2.
RSWIR results

Total

Water

Others

390
185
575

80
331
411

470
516
986

assessment. After analyzing the best Water Index resulting

ood water class, obtained from the corresponding index,
has been converted into a shapele and used as an input to
assess damage for various land use/land cover categories.

MODIS
Water
Others
Total

NDWI

Landsat ETM+

Total

Water

Others

505
65
570

75
341
416

580
406
986

GREEN  NIR
GREEN NIR

RED  SWIR
RED SWIR

RSWIR

Flood monitoring and damage assessment using water indices

Table 4 Accuracy assessment of GSWIR using error matrix.
Values show the area in km2.
GSWIR results

MODIS
Water
Others
Total

Landsat ETM+

GSWIR

GREEN  SWIR
GREEN SWIR

Total

Water

Others

510
70
580

115
291
406

Figure 3

Figure 4

103

4. Results and discussions

625
361
986

In order to evaluate the performance of the mentioned indices

we have compared them with classied images of Landsat
ETM+ through statistical techniques mentioned before.

Results of statistical analysis of error matrix.

Extent of cumulative damage assessment on September 17, 2013.

104

A.A. Memon et al.

Also visual interpretation, using expert knowledge, of MODIS

images with 721 composite has been performed for the purpose. In order to preserve normalization for better comparison
we have set the threshold from 0 to 1 while applying water
indices to MODIS images. It has been observed that varying
the threshold to the best values could impact the efciency
of the individual method, also reported by McFeeters (1996),
Xu (2006), Ouma and Tateishi (2006) and Wu et al. (2008),
but gives relatively same accuracy. This might be due to poor
atmospheric correction of images. Note that the Earth
Observing System Data and Information System (EOSDIS)
provides MODIS images that are atmospherically corrected
but due to diurnal variations in atmospheric parameters it
would not be possible to purely normalize the images by
atmospheric correction.
Fig. 2 shows inundated areas obtained on different dates
through mentioned water indices. It reveals that the variations
in inundated areas follow similar trends for RSWIR and
GSWIR while the converse is true for NDWI. This suggests
that the NDWI maps roughly the inundated area. It also
shows that NDWI underestimates the inundated area. The validation shown as error matrices (Tables 24), is obtained
through MODIS and Landsat ETM+ images of 17
September 2012 and 03 October 2012. The rasters obtained
from the mentioned indices applied on MODIS images and
classication on Landsat ETM+ were converted to vector
les. The vector les were then intersected to get the results
as shown in Tables 24. Rows in these tables show the inundated area derived from MODIS while the columns represent
the inundated area derived from Landsat ETM+ images.
The row total shows water and other land use/land cover area
in square kilometers derived from MODIS, similarly for

Table 5

Landsat ETM+ via column. Each cell value shows the agreement/disagreement of MODIS with Landsat ETM+ results.
Tables 24 and Fig. 3 show underestimation of 18.26% for
NDWI and overestimation of 1.75% and 7.76% for RSWIR
and GSWIR respectively, while j (it is a measure of overall
agreement of a matrix) for NDWI, RSWIR and GSWIR is
measured to be 46.66%, 70.80% and 60.61% respectively.
Overall accuracy, producers accuracy (it is a measure of
how well a certain area is classied), users accuracy (it is a
measure of how much the results, obtained through producers
accuracy, are reliable) and kappa coefcient of NDWI,
RSWIR and GSWIR for water and other land use/cover
(Fig. 3) clearly depict in detail that RSWIR is relatively more
accurate in delineating water bodies.
The vegetation has relatively high reectance in NIR
region, so the NDWI could not take water under vegetation
into account and ultimately underestimate the inundated area.
Most of the study area is vegetated, that is why NDWI has signicantly underestimated the inundated area. Vegetation has
reectance of around 17%, 10% and 20% for green, red and
SWIR regions respectively of electromagnetic spectrum. The
difference in reectance between green and SWIR is very small
that could vanish in chaotic atmospheric conditions during
oods and make GSWIR to take vegetation into account (as
mentioned above) which results in overestimation in the
inundated area. This could not be the case for RSWIR because
red and SWIR have large differences regarding reectance.
4.1. Damage assessment
One of the most important applications of water indices in
Pakistan is the mapping and monitoring of ood events.

Detailed statistics of ood affected land use/land cover categories based on MODIS data.

Province

District

Accumulated
inundated area
(km2)

Villages
aected
(Count)

Aected
railway
(km)

Aected
roads (km)

Aected
agriculture
(km2)

Forest
aected
(km2)

BALOCHISTAN
BALOCHISTAN
BALOCHISTAN
BALOCHISTAN
BALOCHISTAN
KHYBER
PAKHTUNKHWA
PUNJAB
PUNJAB

BOLAN
JAFARABAD
JHAL MAGSI
LORALAI
NASIRABAD
D I KHAN

736.24
1144.69
581.29
53.97
725.46
328.65

39.00
214.00
24.00
2.00
35.00
27.00

2.26
11.07
0.00
0.00
7.58
0.00

64.07
332.14
39.69
0.00
61.60
19.20

0.00
295.64
62.09
0.00
330.26
5.27

0.00
0.00
0.00
0.00
29.42
0.00

BHAKKAR
DERA GHAZI
KHAN
Layyah
MUZAFFARGARH
RAHIM YAR
KHAN
RAJANPUR
DADU
GHOTKI
JACOBABAD
KASHMORE
LARKANA
QAMBAR
SHADAD KOT
SHIKARPUR
SUKKUR

27.34
527.11

7.00
78.00

0.00
0.00

0.00
29.51

20.07
18.88

0.00
35.04

201.93
821.33
500.24

57.00
197.00
108.00

0.00
0.00
0.00

0.00
11.91
10.41

20.15
104.46
165.68

0.00
134.24
38.83

1038.30
413.98
494.56
1557.67
1185.42
250.32
1390.14

213.00
102.00
148.00
573.00
291.00
75.00
332.00

37.44
2.01
4.22
12.63
4.26
3.03
0.00

160.14
95.31
74.07
377.43
90.24
26.74
172.12

200.44
115.03
160.09
294.87
350.99
43.69
304.16

93.02
9.04
123.92
0.00
192.94
70.08
0.00

682.92
495.80

184.00
107.00

11.50
13.24

119.34
37.89

220.45
234.05

46.82
141.16

PUNJAB
PUNJAB
PUNJAB
PUNJAB
SINDH
SINDH
SINDH
SINDH
SINDH
SINDH
SINDH
SINDH

Flood monitoring and damage assessment using water indices

During ood events, policy makers and disaster management
authorities require rapid identication and extent of ooding
at a reasonable scale aiming at minimizing damage and
vulnerability. The accumulated ood extent (Fig. 4) of
13,157 km2 has been observed on September 17, 2012 with
severe rainfall recorded from 06 to 11 September. It has been
observed that inundation in the low lying areas in upper
Sindh is greater than other provinces despite the low amount
of rainfall measurements. This is due to ash oods in
the catchment areas of Baluchistan province that ows
into Sindh province causing inundation in Jacobabad,
Shahdadkot, Kashmore and Ghotki districts. Damages
assessed include agriculture, villages, roads, forest and railway
in order of severity from high to moderate. All the provinces
were affected by oods but assessed damages and losses in
Sindh were relatively severe followed by Baluchistan, Punjab
and Khyber Pukhtoonkhwa. Table 5 shows the accumulated
inundated areas with damage assessment of land cover/land
use in various districts.
5. Conclusion
Due to high temporal resolution of MODIS it is used widely to
map ood events. But its low spatial resolution questions on
accuracy of delineating ood extent. In this regard, one of
the aims of the study is to check the accuracy of MODIS data
with respect to LANDSAT in mapping ood events in the
country without taking the difference in spatial resolution into
account. It is estimated that the MODIS data could help to
map the inundated area with an accuracy of 7381% depending on the mentioned indices used. Further, the comparative
analysis of NDWI, RSWIR and GSWIR using spatial statistical techniques and visual interpretation suggests RSWIR
to be more efcient for delineating water bodies during oods.
Visual interpretation using past experience provides qualitative
results and statistical analysis presents quantitatively. The
ood was caused by heavy monsoon rains mostly in the
Eastern part of Baluchistan and owed toward upper Sindh
province during early September 2012. It inundated 22 districts
of Pakistan with maximum cumulative extent of 13,157 km2.
The study also provides detailed information about the
affected land use/land cover in the country as mentioned in
Table 5.
Acknowledgements
A part of the work has been carried out for the National
Disaster Management Authority (NDMA), Pakistan and
Sindh Irrigation Department, Pakistan. We acknowledge the
use of Rapid Response imagery from the Land Atmosphere
Near-real time capability for EOS (LANCE) system operated
by the NASA/GSFC/Earth Science Data and Information system (ESDIS) with funding provided by NASA/HQ. We are
grateful to Mr. Rahmatullah Jilani of SUPARCO for encouraging and giving us moral support to carry out the study.
We thank Dr. Muhammad Shaq from Earth Sciences
Directorate of SUPARCO for giving useful suggestions to
improve the manuscript.

105
References
Chen, X.Q., Cui, P., Li, Y., Yang, Z., Qi, Y.Q., 2007. Changes in
glacial lakes and glaciers of post-1986 in the Poiqu River basin,
Nyalam, Xizang (Tibet). Geomorphology 88 (3), 298311.
Congalton, R.G., 1991. A review of assessing the accuracy of
classications of remotely sensed data. Remote Sens. Environ. 37
(1), 3546.
El Bastawesy, M., Gabr, S., Mohamed, Ihab, 2015. Assessment of
hydrological changes in the Nile River due to the construction of
Renaissance Dam in Ethiopia. Egypt. J. Remote Sens. Space Sci.
18 (1), 6575.
El-Gamily, I.H., Selim, G., Hermas, E.A., 2010. Wireless mobile
eld-based GIS science and technology for crisis management
process: a case study of a re event, Cairo, Egypt. Egypt. J. Remote
Sens. Space Sci. 13 (1), 2129.
Gao, B.C., 1996. NDWInormalized difference water index for
remote sensing of vegetation liquid water from space. Remote Sens.
Environ. 58 (3), 257266.
Haq, M., Said, R., Rahmatullah, J., Mohammad, S., 2010. Pakistan
ood 2010 monitoring using MODIS data. In: The 9th
International Workshop of the CAS-TWAS-W Forum (2010
CTWF) on Climate and Environmental Changes and Challenges
for Developing Countries, November 1720, 2010, Beijing, China.
Haq, M., Akhtar, M., Muhammad, S., Paras, S., Rahmatullah, J.,
2012. Techniques of remote sensing and GIS for ood monitoring
and damage assessment: a case study of Sindh province, Pakistan.
Egypt. J. Remote Sens. Space Sci. 15, 135141.
Houghton, J.T., 2000. The IPCC Report 2001. In: The Solar Cycle
and Terrestrial Climate, Solar and Space weather, vol. 463, pp. 255.
Ji, L., Zhang, L., Wylie, B., 2009. Analysis of dynamic thresholds for
the normalized difference water index. Photogramm. Eng. Remote
Sens. 75 (11), 13071317.
Li, B., Zhou, X.Z., 2009. Research on extraction method of river
water-body in mountain area of Western Sichuan based on MODIS
data. Geogr. Geoinf. Sci. 2, 017.
Ling, C., Zhang, F.Q., Lin, H., 2010. Research on extracting coastal
wetlands water information using composition ciwi water index.
Resour. Environ. Yangtze Basin 2, 007.
Lu, S., Wu, B., Yan, N., Wang, H., 2011. Water body mapping
method with HJ-1A/B satellite imagery. Int. J. Appl. Earth Obs.
Geoinf. 13 (3), 428434.
McFeeters, S.K., 1996. The use of the normalized difference water
index (NDWI) in the delineation of open water features. Int. J.
Remote Sens. 17 (7), 14251432.
Ouma, Y.O., Tateishi, R., 2006. A water index for rapid mapping of
shoreline changes of ve East African Rift Valley lakes: an
empirical analysis using Landsat TM and ETM+ data. Int. J.
Remote Sens. 27 (15), 31533181.
Rashid, S.F., 2000. The urban poor in Dhaka city: their struggles and
coping strategies during the oods of 1998. Disasters 24 (3), 240
253.
Rasid, H., Haider, W., 2003. Floodplain residents preferences for
water level management options in ood control projects in
Bangladesh. Nat. Hazards 28 (1), 101129.
Rogers, A.S., Kearney, M.S., 2004. Reducing signature variability in
unmixing coastal marsh thematic mapper scenes using spectral
indices. Int. J. Remote Sens. 25 (12), 23172335.
Sardar, M.S., Tahir, M.A., Zafar, M.I., 2008. Poverty in riverine areas:
vulnerabilities, social gaps and ood damages. Pak. J. Life Soc. Sci
6 (1), 2531.
Schmuck-Widmann, H., 1996. Living with the Floods: Survival
Strategies of Char-Dwellers in Bangladesh. LN-Vertrieb.
Wu, W.Y., Shen, X.H., Zou, L.J., Lu, S.L., Zhang, G.F., 2008. An
integrated method for water body detection and delineation using
Landsat ETM+ data. Bull. Sci. Technol. 24, 252259.

106
Xia, Jun, Zhang, Yongyong, 2008. Water security in North China and
countermeasure to climate change and human activity. Phys.
Chem. Earth Parts A/B/C 33 (5), 359363.
Xiao, X., Boles, S., Liu, J., Zhuang, D., Frolking, S., Li, C., Moore III,
B., 2005. Mapping paddy rice agriculture in Southern China using
multi-temporal MODIS images. Remote Sens. Environ. 95 (4),
480492.

A.A. Memon et al.

Xu, J., 2004. A study of anthropogenic seasonal rivers in China.
Catena 55 (1), 1732.
Xu, H., 2006. Modication of normalised difference water index
(NDWI) to enhance open water features in remotely sensed
imagery. Int. J. Remote Sens. 27 (14), 30253033.