Académique Documents
Professionnel Documents
Culture Documents
H O S T E D BY
RESEARCH PAPER
a,*
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.
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) 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
Table 1
101
102
Figure 1
Figure 2
MODIS
Water
Others
Total
Landsat ETM+
Total
Water
Others
390
185
575
80
331
411
470
516
986
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
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
104
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
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.