PresentationPDF Available



Abstract and Figures

The analysis of Landsat images acquired with the application of Remote Sensing (RS) and Geographic Information System (GIS) provides an inexpensive technique in estimating and mapping spatio-temporal shifts in the river. Concerning the case, this study modeled the temporal changes of the Bagmati River within 25 years (1991–2016) using the multi-temporal Landsat images. We adopted the Normalized Difference Water Index (NDWI for the unsupervised extraction of the water feature and monitoring the changes. A model was developed in Arc-GIS by discerning the river, and the difference was determined for 25 years. The result indicated a major temporal shift in the river channel with a decreasing trend from 1991 to 2016. Over 25 years, the river loss almost one-third of its original water-flow channel with a severe sweep in the south-western portion of the study area. With this precise information, a field-based study can be undertaken either to analyse the damage caused by the river in those particular portions or to assess the factors affecting the river shift. Hence, we strongly recommend employing the cost-effective methods, RS and GIS, for detecting, analysing and monitoring the shifts and changes in the rivers and lakes over a while.
Content may be subject to copyright.
A. Subedi 1, P. Poudel 2, T. D. Acharya3, 4
1Institute of Forestry, Pokhara Campus, Tribhuvan University, Pokhara 33700, Nepal
2 Himalayan College of Agricultural Sciences & Technology Purbanchal University, Kathmandu 44600, Nepal
3 Dept. of Civil Engineering, Kangwon National University, Chuncheon 24341, Korea
4 School of Geomatics and Urban Spatial Information, Beijing University of Civil Engineering and
Architecture, Beijing 102616, China
Presentation Outline
Results and Discussion
Nepal, a country rich in its water resource, encompasses more than 6000
rivers originated from the lap of majestic Himalayas.
Major source to fulfil the basic needs of people,
Important role in other major aspects like drinking water, agriculture,
electricity production, transportation etc.
Regulation of natural water cycle, climatic regulation as well as other
ecological cycle.
Decrease in quality due to factors like climate change, natural disasters,
environmental hazards, undesirable anthropogenic influences, etc.
Variation in river shape and size as well as water quantity, however, are slow and
takes a long time.
Relying on information acquired through the use of remote sensing and geographic
information system, this dynamism could easily be identified, at a very low cost,
Remote sensing satellite continuously provide very reliable, accurate and cost-
effective information
Multi-temporal images acquired through the satellite like Landsat are widely used in
detecting the changes (Acharya et al., 2019a).
In this study, we analyzed the image-data from satellite image Landsat-series
images and geographical information system (GIS) to find the temporal shift in
the Bagmati River.
A plethora of methods have been developed for extracting water areas from
optical remote sensing imagery.
Between supervised and unsupervised methods of image classification and
the analysis, we used the NDWI method to detect and monitor the temporal
shift of the Bagmati River for a period of 25 years using the Landsat series data
of 1991,2002 and 2016.
Test site
A geographical area between
26.73850Nto 27.18240N latitude
and 85.27550Eto 85.53590E
longitude was selected as the
study site.
The area incorporates the
Bagmati River which separates
the two major districts: Rautahat
and Sarlahi. Fig. 1 Map of study area showing district boundary
Data acquired by Landsat series
available from the USGS GLOVIS
portal were used.
Three images from 1991 to
2016, a 25 years gap data, were
used for the study of a temporal
shift of the Bagmati River within
the study area.
Fig. 2 Map of study area showing Natural colour
composite image of the Landsat imagery for 1991,
2002 and 2016 from left to right
The accessed image file was subjected for eliminating radiometric as well as a
geometric aberration if any.
The area of interest (Bagmati River) was delineated in the software.
The images, pre-georeferenced to the Universal Transverse Mercator zone 45
north, were exported into the GeoTIFF format for further analysis.
After the pre-processing step, all three images of the corresponding year were used
in the Arc Map for calculating the NDWI.
A model was developed in ArcGIS for the unsupervised change detection, recognized
as the prioritized option for achieving the Spatio-temporal monitoring of the river
dynamics, thus, the McFeeters’s formula for NDWI, employing the green band and
NIR band, was used for the calculation as follows:
Where, ρBand2 = Reflectance value of the green band
ρBand4 = Reflectance value of the NIR band
NDWI =𝑝𝐵𝑎𝑛𝑑2−𝑝𝐵𝑎𝑛𝑑4
A comparative study
of the river, through
the unsupervised
indicated a major
shift in the Bagmati
River channel with a
decreasing trend
from 1991 to 2016.
Fig. 3 Detected surface water in Bagmati River for the years
1991, 2002 and 2016
The overall surface is coverage of the river was found to be 32.80 sq. km, which
was the highest among the compared with the other two years.
In 2002,on another hand, the area of the surface water of the Bagmati River
shrank to become 28.34 which dramatically reduced to 11.99 in
2016 along with drastic fluctuations in the water-current flow channel.
In the first decade, i.e. from 1991 to 2002, Bagmati River is confirmed to have
shrunk 0.45 year; whereas in another 15 years of period (2002-2016),
the river is shrinking with the higher rate: 1.09 sq. km/year.
The river is, thus, having constant displacement and shrinkage in the bank and
the water availability.
The river channel is erratically fluctuating from the base year
to present.
The result, overall, connotes the loss of about one-third of the
original water-flow channel by the river, (with a severe sweep
in the south-western portion of the study area).
There must be several factors (natural to anthropogenic)
which have heightened such displacement and shifts in the
Additionally, such shifting might have resulted in dire
ramifications in the particular area where the river is seen
shifting at a greater extent (might have marred the natural
integrity and affected different living creatures).
Fig. 4 A map showing the changes
during the 25-year period,
Concerning the performance of the unsupervised classification, the NDWI
performed well in the detection, delineation, and shift identification of the
surface water.
This index could be used in the similar approaches of detecting and monitoring
the changes in the rivers, lakes, and other water sources.
The Bagmati River is shrinking at the greatest rate”, is what our result showed.
These sweeping - and shrinking - trends of the Bagmati River during just 25-
year period is the indication of the vulnerability of the river.
The river may ultimately, if the situation persists and the river goes losing the
coverage in the same rate, will be bereft of water.
Thus, appropriate policies, laws, planning and other research and analysis must
be carried out for the constant monitoring of the river and preserving our
greatest natural assets. The risk zone delineation and identification for the
Such results can be useful in monitoring the river system as well as planning for
future actions, most importantly in countries like Nepal where field-based-
sophisticated research and analysis are barely prone to happen.
Constant River monitoring, risk zone mapping, disaster risk area identification
and prediction, etc. could be an effective way to minimize the uncertain risk to
living creatures and to keep the water conserved.
The series of Landsat images are, therefore, totally useful in discerning the
shifts in the river, lakes, and other water bodies.
Hence, we strongly recommend employing the cost-effective methods: RS and
GIS, for detecting, analyzing and monitoring the shift and changes in the rivers
and lakes over a while.
Thank You!
ResearchGate has not been able to resolve any citations for this publication.
ResearchGate has not been able to resolve any references for this publication.