Geomatics, Natural Hazards and Risk

ISSN: 1947-5705 (Print) 1947-5713 (Online) Journal homepage: https://www.tandfonline.com/loi/tgnh20

Changes in turbidity along Ganga River using

Sentinel-2 satellite data during lockdown
associated with COVID-19

Vaibhav Garg, Shiv Prasad Aggarwal & Prakash Chauhan

To cite this article: Vaibhav Garg, Shiv Prasad Aggarwal & Prakash Chauhan (2020)
Changes in turbidity along Ganga River using Sentinel-2 satellite data during lockdown
associated with COVID-19, Geomatics, Natural Hazards and Risk, 11:1, 1175-1195, DOI:
10.1080/19475705.2020.1782482

To link to this article: https://doi.org/10.1080/19475705.2020.1782482

© 2020 The Author(s). Published by Informa

UK Limited, trading as Taylor & Francis
Group

Published online: 27 Jun 2020.

Submit your article to this journal

Article views: 17

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=tgnh20
GEOMATICS, NATURAL HAZARDS AND RISK
2020, VOL. 11, NO. 01, 1175–1195
https://doi.org/10.1080/19475705.2020.1782482

Changes in turbidity along Ganga River using Sentinel-2

satellite data during lockdown associated with COVID-19
Vaibhav Garga, Shiv Prasad Aggarwala and Prakash Chauhanb
a
Water Resources Department, Indian Institute of Remote Sensing, Dehradun, Uttarakhand, India;
b
Indian Institute of Remote Sensing, Dehradun, Uttarakhand, India

ABSTRACT ARTICLE HISTORY

India had announced the longest ever lockdown from 24 March Received 3 May 2020
2020 to 14 April 2020 amid COVID-19 pandemic. It was reported Accepted 2 June 2020
that the water quality of the Ganga River has improved as com-
KEYWORDS
pared to regular during this country-wide lockdown. In the pre-
Water quality; turbidity;
sent study, an attempt has been made to study the change in remote sensing; reflectance;
water quality of the river in terms of turbidity purely through Ganga River; lockdown;
remote sensing data, in the absence of ground observations, NDTI; COVID-19
especially during this time period. The change in spectral reflect-
ance of water along the river in the visible region has been ana-
lyzed using the Sentinel-2 multispectral remote sensing data at
Haridwar, Kanpur, Prayagraj, and Varanasi stretches of the river. In
the present study, it was found that the red and NIR bands are
most sensitive, and can be used to estimate the turbidity. Further,
the temporal variation in turbidity was also analyzed through nor-
malized difference turbidity index at each location. It was
observed that the turbidity in the river has reduced drastically at
each stretch of the river. The study elicited that the remote sens-
ing approach can be used to make qualitative estimates on tur-
bidity, even in the absence of field observations.

1. Introduction
Mapping of water quality of inland waters using remote sensing is being carried out
since 1970s, with the launch of Landsat series of satellites (Klemas et al. 1971;
Kritikos et al. 1974; Johnson 1975; Ritchie et al. 1976). Every feature on the surface
of the earth behaves differently on interacting with electromagnetic radiation (EMR).
Based on their spectral response these features are identified in a satellite image. The
slight change in their composition, changes the spectral properties of the feature.
Similarly, many factors affect the spectral response of the water such as time of the
year, sun-elevation angle, the concentration of atmospheric constituents, roughness of
the water, suspended matter, turbidity, depth of water, and submerged or emergent
vegetation (Moore 1980). On simultaneously analyzing the change in composition

CONTACT Vaibhav Garg [email protected]

ß 2020 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial License (http://
creativecommons.org/licenses/by-nc/4.0/), which permits unrestricted non-commercial use, distribution, and reproduction in any
medium, provided the original work is properly cited.
1176 V. GARG ET AL.

and changes in the reflectance properties of the water, the abundance or concentra-
tion of a particular component can qualitatively or sometimes quantitatively be esti-
mated using remote sensing data. It is to be noted that the water quality parameters
that affect its optical properties are generally being studied through remote sensing.
Therefore, the studies are mainly constrained to the estimation of turbidity and
chlorophyll, colored dissolved organic matter (CDOM) concentration (Lim and Choi
2015; Gholizadeh et al. 2016; Trinh et al. 2017; Chander et al. 2019; Luis et al. 2019).
Among all the water quality parameters suspended sediments are the most com-
mon problem in inland waters such as rivers, lakes, and estuaries (Ritchie et al.
1974) . These suspended particles attenuate the light required for aquatic life (Ritchie
et al. 1974; Doxaran et al. 2002; Garg et al. 2017). It is also considered as an indicator
of eutrophication (G€ uttler et al. 2013; Sebastia-Frasquet et al. 2019). Traditionally, the
concentration of these sediments is assessed optically or through gravimetric methods
or laboratory analysis (Pavelsky and Smith 2009). Nowadays, it is measured visually
through Secchi disk depth or directly using the light turbidimeters in the field
(Pavelsky and Smith 2009; Quang et al. 2017). The field measurements are considered
to be more accurate, but, they are both labor-intensive and time-consuming (Pavelsky
and Smith 2009; Quang et al. 2017). However, these measurements are limited to
point information, and the concentration of total suspended particles varies both tem-
porally and spatially (Gholizadeh et al. 2016; Garg et al. 2017). It is reported in the
literature that the three variables turbidity, suspended particulate matter and Secchi
disk depth are closely related (Sebastia-Frasquet et al. 2019).
Turbidity is an important optical property of water, where suspended sediments
scatter the light rather than transmit it along the water column (Sebastia-Frasquet
et al. 2019). The turbidity increases with an increase in the concentration of sus-
pended solids or sediments in water (Ritchie et al. 1976; Garg et al. 2017). Turbidity
enhances the opacity of water, which hampers the aquatic life (G€ uttler et al. 2013;
Quang et al. 2017; Sebastia-Frasquet et al. 2019). It is also mentioned in the literature
that there can be temporal variation in turbidity concentration due to fluctuation in
weather, climate pattern and human activities along the banks (Luis et al. 2019). This
property of water is being mapped for long using remote sensing data as it provides
synoptic coverage of earth at regular temporal time domain. It has been done by ana-
lyzing how suspended sediment concentration change alters the optical properties of
the water column. It has been reported in the literature that the reflectance in visible
region specifically red region increases with increase in sediments in the water or tur-
bidity (Ritchie et al. 1976; Moore 1980; Doxaran et al. 2002; Pavelsky and Smith
2009; Gholizadeh et al. 2016; Garg et al. 2017). It has also been noticed that the
reflectance peak shifts from green to red region of the spectrum (Gholizadeh et al.
2016). It has been mentioned that the concentration of suspended sediments or tur-
bidity shows good relation with the visible region of the spectrum whereas red and
NIR regions are more sensitive towards turbidity (IOCCG 2000; Toming et al. 2016;
Caballero et al. 2019; Sebastia-Frasquet et al. 2019).
Novel Corona Virus (COVID-19) has emerged as a deadliest disease across the
globe (Chauhan and Singh 2020; Paital et al. 2020), which has affected around 210
countries. Considering, its nature and spread, World Health Organization declared it
 GEOMATICS, NATURAL HAZARDS AND RISK 1177

as a pandemic (Saadat et al. 2020). To safeguard their countrymen from the pan-
demic, each country declared lockdown time to time in phased manner (Chauhan
and Singh 2020). India had also called the longest ever lockdown in the entire coun-
try in view of the pandemic. Studies reported that the environmental condition over
major cities has improved in terms of reduction in pollution either water or air across
globe due to the lockdown (Braga et al. 2020; Chauhan and Singh 2020;
Collivignarelli et al. 2020; Dantas et al. 2020; Lal et al. 2020; Li et al. 2020;
Muhammad et al. 2020; Otmani et al. 2020). In India also, it has been reported that
the air quality in the major cities and the water quality of the rivers/water body
improved with the lockdown period (Chauhan and Singh 2020; CPCB 2020; Paital
et al. 2020; Saadat et al. 2020; Yunus et al. 2020). It was also stated that the water of
the rivers such as Ganga and Yumuna passing through major cities is much clear and
their turbidity has reduced. In the present study, an attempt has been made to verify
it using remote sensing data for different stretches of the Ganga River. The temporal
changes in reflectance in visible region bands of the Sentinel-2 data and normalized
difference turbidity index (NDTI) have been analyzed for pre and post-nationwide
lockdown dates. Further, as no field data were available for this time period, the
results are presented in qualitative terms.

2. Study area and data used

Ganga River is the most sacred river and has mythological importance in India. It
originates in Himalaya as Bhagirathi and Alakhnanda rivers, after the confluence of
these tributaries at Deoprayag, the river is then called as Ganga River. Its basin area
falls in four adjacent countries India, Nepal, Tibet (China), and Bangladesh. The
Ganga River basin in India covers nearly one-third of its total geographical area. The
river runs a total length of 2525 km before draining into the Bay of Bengal (Jain et al.
2007). From mountainous Himalaya, the river flows mostly in the south direction
and enters in plains near Haridwar, one of the most pilgrimage towns in the country.
Then it turns in south-east direction and flows through the industrial city of Kanpur.
The most important tributary of the river, Yamuna, joins it at Prayagraj (formerly
known as Allahabad). The confluence of Yamuna and Ganga at Prayagraj is called
Sangam (means confluence only). From here the river flows in east-ward direction
through eternal cities of Varanasi (also known as Banaras or Kashi) and Patna. Then
it enters the state of West Bengal, gets split into two Bhagirathi and Padma. The first
branch directly drains into the Bay of Bengal and the other enters Bangladesh and
then ends in the Bay of Bengal. The river feeds billions of people along its course. As
mentioned earlier, the river is most worshipped and holy in nature for Indians, mil-
lions of people take a holy dip in it almost every day. The government of India keeps
on trying to maintain its water quality through various national-level plans and pro-
grams (NMCG 2020). Still, the water quality in many places is not fit for dir-
ect drinking.
In the present study, the stretches of the river at Haridwar, Kanpur, Prayagraj, and
Varanasi have been studied for change in turbidity during the Phase-I (25 March–14
April 2020) of the lockdown through remote sensing data. The location of these
1178 V. GARG ET AL.

cities/stretches is provided in Figure 1. Haridwar lies in Uttarakhand and the other

03 are major cities of Uttar Pradesh State of India. Haridwar is the city where the
Ganga River enters the plains. It has mythological importance in India, therefore, a
very large number of pilgrims visit this place every day throughout the year and take
a holy dip in Ganga water. The place people take holy bath ‘Har-ki-Pauri’, is the
diverted water from Bhimgoda Barrage in its right bank canal (see Figure 1). Further,
the upstream region of the river, from its origin, has several places for pilgrims,
therefore, many developmental activities going on in the vicinity to cater to the huge
population of the visitors. Due to these types of activities along the river upstream,
the water in the river remains turbid, most of the time. Kanpur is a city famous for
its leather and textile industries. It is generally said that the water quality of the river
deteriorates after the city. It can be seen from Figure 1 that there is Ganga Barrage
on the river just before it enters Kanpur city. Therefore, the turbidity remains low to
moderate along the river across the city. The city at the confluence of Yamuna and
Ganga called Prayagraj is again a pilgrimage city. Large devotees take a holy bath at
the Sangam, every day. Last year, the city conducted the major bath festival (bath in
Ganga River) ‘Kumbh’. Then the next important city is Varanasi , another sacred city.
It is also regarded as the spiritual capital of India. The city is known for its silk fabric
industry along with perfumes, ivory, sculpture, etc.
To map the turbidity in these stretches of Ganga River, the Sentinel-2A/B dataset
has been used. The Sentinel-2A and Sentinel-2B satellites were launched on 23 June
2015 and 07 March 2017, respectively. Its Multi-spectral Instrument (MSI) provides
data in 13 spectral bands ranging from visible and near-infrared to short wave infra-
red (443–2190 nm) regions, with a swath width of 290 km and a spatial resolution of
10 m (four visible and near-infrared bands), 20 m (six red edge and shortwave infra-
red bands) and 60 m (three atmospheric correction bands). The details of each band
and wavelength regions are given in Table 1.
The details of stretch wise satellite data used are given in Table 2. The cloud-free data
of respective dates were downloaded from the Sentinel Scientific Data Hub (https://
scihub.copernicus.eu/). In the present study, the Level-1C, that is, radiometrically and
geometrically corrected top of atmosphere reflectance product of Sentinel-2A/B were
used to maintain the uniformity in terms of atmospheric correction.
The intermittent rainfall, during the study period, has also changed the turbidity
pattern at each of the stretches under consideration. Therefore, the daily rainfall (in
mm) data available for Haridwar, Kanpur, Prayagraj, and Varanasi stations at India -
Water Resources Information System (India-WRIS) for the period from 15 March–15
April 2020 was analyzed.

3. Methodology
Due to the COVID-19 outbreak, India has declared country-wide lockdown in two
phases from 25 March 2020–14 April 2020, and 15 April 2020– 03 May 2020 (MHA
2020). Because of this, all the industrial activities other than essential were closed,
and people asked to confine themselves in their houses. The lockdown resulted in
minimum disturbance to the nature, especially, the Ganga River (Yunus et al. 2020).
 GEOMATICS, NATURAL HAZARDS AND RISK 1179

Figure 1. Location of Ganga River stretches under consideration on Google Earth Image.
1180 V. GARG ET AL.

Table 1. Sentinel-2 Multispectral Imager bands and

their details.
Central
Band (wavelength region) wavelength (nm) Resolution (m)
Band – 1 (Coastal aerosol) 443 60
Band – 2 (Blue) 490 10
Band – 3 (Green) 560 10
Band – 4 (Red) 665 10
Band – 5 (Vegetation red edge) 705 20
Band – 6 (Vegetation red edge) 740 20
Band – 7 (Vegetation red edge) 783 20
Band – 8 (NIR) 842 10
Band – 8A (Vegetation red edge) 865 20
Band – 9 (Water Vapor) 945 60
Band – 10 (SWIR - Cirrus) 1375 60
Band – 11 (SWIR) 1610 20
Band – 12 (SWIR) 2190 20

There was less industrial waste effluent in the water, minimum anthropogenic activ-
ities along its banks due to restricted pilgrim visits, and other activities along its
course (CPCB 2020). It was reported that the water quality, in terms of clarity or tur-
bidity, of the river has improved at many places along its course during this short
time period (CPCB 2020; Yunus et al. 2020). In the present study, an effort has been
made to study it using remote sensing data, as there was no scientific investigation
supporting it. For the analysis, the major cities or cities associated with major pilgrim
activities namely Haridwar, Kanpur, Prayagraj, and Varanasi were identified as study
stretches of the river. The overall methodology adopted to map the changes in turbid-
ity is provided in Figure 2.

3.1. Identification of water pixels

To study the water quality of the river, in terms of change in turbidity, the optical
remote sensing data from Sentinel-2 data (as given in Table 2) of pre- and post-lock-
down were analyzed for change in reflectance and hence change in turbidity. The
MSI sensor provides data from visible to SWIR wavelength region in reasonably high
spatial resolution. Initially, the water pixels on each date were estimated using the
normalized difference water index (NDWI) given by McFeeters (1996) as provided in
Eq. (1).

RG  RNIR
NDWI ¼ (1)
RG þ RNIR

Where RG and RNIR are reflectance in green and NIR bands, respectively. The reason
for considering green and NIR wavelength regions was that water reflects the highest
in the visible region and absorbs all energy in the NIR region of EMR. The value of
NDWI varies from 1 to þ1. The value of more than 0 is theoretically being consid-
ered as water in the NDWI image. However, it is recommended that one should fol-
low a trial and error approach and decide the threshold for water pixels. After
identifying the water pixels, the visible to VNIR bands were masked only for water
 GEOMATICS, NATURAL HAZARDS AND RISK 1181

Table 2. The list of data used in the present study.

City Sensor Date
Haridwar Sentinel-2B 20 March 2020
Sentinel-2A 25 March 2020
Sentinel-2B 30 March 2020
Sentinel-2A 04 April 2020
Sentinel-2A 14 April 2020
Kanpur Sentinel-2A 19 March 2020
Sentinel-2A 29 March 2020
Sentinel-2B 03 April 2020
Sentinel-2B 13 April 2020
Prayagraj Sentinel-2A 09 March 2020
Sentinel-2B 24 March 2020
Sentinel-2A 29 March 2020
Sentinel-2B 03 April 2020
Sentinel-2B 13 April 2020
Varanasi Sentinel-2A 16 March 2020
Sentinel-2A 26 March 2020
Sentinel-2B 31 March 2020
Sentinel-2A 05 April 2020
Sentinel-2B 10 April 2020

Figure 2. A broad framework of the methodology used.

pixels on the respective dates. It has been reported in the literature that the red and
NIR region of EMR is most sensitive to turbidity. Therefore, initially, the change in
reflectance in each visible to NIR wavelength bands was studied.
1182 V. GARG ET AL.

3.2. Change in reflectance in each visible to NIR band of sentinel-2

It is to be noted that the suspended sediment concentration and the turbidity are fre-
quently being treated as similar or same in the remote sensing technology field
(Ritchie et al. 2003). Therefore, in the present study, the turbidity term has been used
throughout, as a proxy for suspended sediment concentration. It has been reported in
the literature and proved that, due to the change in turbidity of the water, the
changes in spectral reflectance in visible regions of the spectrum are significant
(Ritchie et al. 1976; Liedeke et al. 1995; Brezonik et al. 2005). Literature suggests that
even a single band, if chosen appropriately, can provide a robust estimate of turbidity
(Pavelsky and Smith 2009; Nechad et al. 2010; Gholizadeh et al. 2016). It was sug-
gested that a single red wavelength region can be used to estimate the water turbidity
(Miller and McKee 2004; Hellweger et al. 2007; Shi and Wang 2009). However, it was
reported that the NIR region is also equally sensitive to turbidity and less influenced
by bottom reflectance in shallow waters (Caballero et al. 2019; Sebastia-Frasquet et al.
2019). Studies also mentioned the use of red and NIR together for better turbidity
assessment (IOCCG 2000; Doxaran et al. 2002; Toming et al. 2016). These single
band approaches were used where the reflectance increases with increment in turbid-
ity. There are a large number of studies, where all the bands of the entire visible
wavelength region are analyzed either individually or in combination for turbidity
estimation (Sebastia-Frasquet et al. 2019). In the present study also, the bands
through blue to NIR were analyzed for change in their spectral response due to a
change in turbidity concentrations in view of the longest ever lockdown in India. The
simple approach of density slicing has been adopted over each visible region band of
each day image. The reflectance on each water pixel of each day was sliced or classi-
fied into a number of groups. It was considered that as the sediment concentration
increases, the reflectance in visible region increases or vice-versa. Therefore, the pixels
with high reflectance in each band are regarded as high turbid regions, whereas low
reflectance as low turbid. In this way, all the bands of Sentinel-2A/B are classified as
very low, low, moderate, high, very high turbid. However, these results couldn’t be
validated as field data was not available due to lockdown in the country.

3.3. Qualitative estimate of turbidity through NDTI

It has also been reported in the literature that relying on a single band or algorithm
sometimes results in overestimation or underestimation of suspended sediment con-
centration (Kuhn et al. 2019; Pahlevan et al. 2019). It was suggested, either use com-
bination of sensitive bands or different algorithms to reach the final conclusion.
Therefore, an attempt has been made to qualitatively estimate the temporal turbidity
in each stretch using the NDTI developed by Lacaux et al. (2007).

RR  RG
NDTI ¼ (2)
RR þ RG

Where RR is the reflectance in the red band. Generally, the reflectance of pure water
is more in green than the red wavelength region. However, it has been reported that
 GEOMATICS, NATURAL HAZARDS AND RISK 1183

the red region reflectance increases with an increase in turbidity. Therefore, the red
and green bands were used to enhance the image for turbidity. Initially, the water
pixels were identified as mentioned above, then Eq. (2) was applied using these two
bands to map NDTI. The higher value of turbidity yields a high value of NDTI and
vice versa.

4. Results and discussion

The water quality, in terms of turbidity, has been analyzed to verify the change in the
quality of Ganga River water due to the longest ever country-wide lockdown. In the
present study, the Sentinel-2 multispectral data of the first phase of lockdown, that is,
from 25 March–14 April 2020, was analyzed. Ganga River is considered as most
sacred river and people have associated sentiments or rituals with it in India. Further,
a million people reside at the bank of the river along its course. However, in many
places, the quality of water in the river is not fit for drinking. The change in turbidity
at the most popular and crowded stretches of the river, that is, Haridwar, Kanpur,
Prayagraj, and Varanasi has been analyzed purely based on remote sensing data
adopting the methodology discussed in the previous section. The results are discussed
in the subsequent section.

4.1. Change in reflectance in each visible to NIR band of sentinel-2

It has been reported in the media that the water of the Ganga River has improved a
lot in terms of its clarity during the lockdown. In this study, it has been investigated
simply through the remote sensing data and techniques. The data of Sentinel-2 satel-
lites (both A & B) as mentioned in Table 2 were analyzed. Initially, the change in
reflectance in each visible to NIR band was accessed, as the minute change in turbid-
ity, changes the reflectance in these wavelength regions. In the case of river stretch at
Haridwar (Figure 3a), the depth of water is mostly shallow downstream Bhimgoda
Barrage, therefore, there is not much change in the water reflectance in the blue band
with a central wavelength of 490 nm. However, there was a decrease in reflectance in
the water upstream of the barrage from 25 March 25–04 April 2020; which may be
due to a reduction in sediments causing less scattering of blue light. Similar changes,
reduction in blue band reflectance, were found near the ‘Har ki Pauri’ region, which
might be attributed to less disturbance by pilgrimage activities near this place. The
same kind of results was found in the green band (560 nm), that is, reduction of
reflectance of water due to reduction in human activities in and around Ganga River
near Haridwar leading to a reduction in turbidity. Major changes in reflectance were
noticed in red and NIR regions of the spectrum, as these bands have relatively less
interference from the bottom and return backscattered energy from suspended par-
ticles mainly.
The decrease and increase in reflectance of water due to turbidity have also been
affected by intermittent rainfall during the period of analysis, in this regard, the rain-
fall data of this time period was also analyzed as shown in Figure 4a. It can be seen
that after 15 March 2020, there was no heavy rainfall in the region till 20 March
1184 V. GARG ET AL.

Figure 3. Change reflectance in Visible to NIR bands of Sentinel – 2 due change in turbidity con-
centration on respective dates; (a) Haridwar, (b) Kanpur, (c) Prayagraj, (d) Varanasi.

2020, therefore, the water was relatively clear on 20 March 2020. However, there was
rainfall from 21 to 24 March 2020 of around 2.5 mm in Haridwar, leading to more
turbidity in the water of the Ganga River. That is why, all the bands were showing
high reflectance due to more sediments in the river on 25 March 2020. On the other
hand, there was not much precipitation after 25 March 2020, therefore, the
 GEOMATICS, NATURAL HAZARDS AND RISK 1185

Figure 3. Continued.
1186 V. GARG ET AL.

Figure 3. Continued.

reflectance of water reduced throughout in the visible region spectrum bands. Based
on visual interpretation, it can be said that water is less turbid. Again, the pattern of
reflectance or turbidity has changed on 14 April 2020, which might be attributed to
rainfall occurred during 10–11 April 2020. It is inferred that the water quality of the
river water has improved, however, it is affected by the contribution from intermit-
tent rainfall occurred during the lockdown period at Haridwar.
 GEOMATICS, NATURAL HAZARDS AND RISK 1187

Figure 3. Continued.

Kanpur is again an important city, known for the textile and leather industry.
Being on the bank of the river, the effluent from these industries reach Ganga River
and pollute its water. A similar approach of monitoring change in reflectance in each
visible to NIR band, as mentioned in the previous section was used. The results of
the analysis are provided in Figure 3b. In the case of the Ganga River stretch at
Kanpur, it can be observed that the water quality of the water has improved continu-
ously till 08 April 2020. It has to be noted that the water in this particular stretch is
1188 V. GARG ET AL.

deeper than the depth at Haridwar stretch considered in the study. Therefore, there
was less interference from bottom surface reflectance. However, the rainfall on 08
April 2020 (Figure 4b), has altered the turbidity on 14 April 2020. There was very lit-
tle change in reflectance in the blue band, whereas, it has slightly reduced in the
green band. That might be attributed to reduced turbidity during the period of ana-
lysis. Again, it can be noticed that the red and NIR regions are more sensitive
towards turbidity.
The river stretch at Sangam, Prayagraj, showed major changes in turbidity. At this
location, Ganga is shallow as compared to the Yamuna and brings more sediments along
with it. Further, heavy pilgrimage activity at this location keeps water turbid. As Kanpur
and Prayagraj are not very far away from each other, the rainfall pattern is almost simi-
lar. There was no rainfall from 24 March–08 April 2020, in the region. Therefore, the
reduction in reflectance or in other terms ‘turbidity’ was very well analyzed. Yamuna
River, as it is deep at this location, some changes, that is, reduction in reflectance, even
in the blue region were observed (Figure 3c). This might have occurred due to less sus-
pended sediments yielding less scattering or reflectance. Further, there was less effect of
bottom reflectance in the analysis. Even after the confluence, where the river is deeper, it
showed a reduction in reflectance in this particular band. Similar results were found in
the green band. Again, with the help of red and NIR most sensitive bands, one can easily
make out the concentration of turbidity qualitatively. It was noticed that the turbidity at
Sangam was continuously low as compared to the condition before lockdown. However,
the rainfall of 08 April 2020 (Figure 4c), has changed the turbidity condition.
The analysis at river stretch near Varanasi, again showed continuous lowering of tur-
bidity, as reflectance throughout the visible and NIR region has reduced till 08 April 2020
(Figure 3d). It must be noted that the depth of water at this stretch is also deep as com-
pared to Haridwar. Figure 3d depicted that there was a significant reduction in reflectance
of the blue band, after the start of the lockdown. However, pixel-wise better spatial vari-
ation of turbidity can be studied in longer wavelengths, that is, red and NIR. Again the
rainfall on 08 April 2020 (Figure 4d), in the city has altered the turbidity variation.
The study elicited that the red and NIR regions of the spectrum are more sensitive
towards suspended sediments or the turbidity, and are in the line of findings reported
in the literature. Moreover, these bands have less interference from the bottom. The
results of Haridwar upstream Bhimgoda Barrage, Kanpur, the Yamuna at Prayagraj,
and Varanasi support this statement. However, for shallow water (2–5 m depth), it is
suggested to use the NIR band for inferring about turbidity concentration. Further,
the most appropriate band for turbidity mapping using Sentinel-2 is ‘red edge’ with
705 nm wavelength (Liu et al. 2017; Caballero et al. 2019; Sebastia-Frasquet et al.
2019). Further, the turbidity concentration is affected by the precipitation in the
region, therefore, it is recommended to analyze precipitation data before reaching a
conclusion on change in turbidity.

4.2. Analysis of spatio-temporal change in NDTI

There may be some extent of radiometric inconsistency in temporal remote sensing
data, even with the same sensor, therefore, sometimes relying on single-band
 GEOMATICS, NATURAL HAZARDS AND RISK 1189

Figure 4. Rainfall at each location during 15 March–15 April 2020 (Source: India - Water Resources
Information System, India-WRIS: http://indiawris.gov.in/wris/#/ … accessed on 23 April 2020).

information for turbidity estimation is not always advisable (Kuhn et al. 2019;
Pahlevan et al. 2019). It is suggested to blend approaches that use different algorithms
at different reflectance ranges to map water quality parameters using remote sensing
data such as turbidity (Hu et al. 2012). Tassan (1997) suggested the use of green and
red band reflectance to estimate variable suspended sediment concentration in the
stratified water column. Therefore, later the results were validated through mapping
of NDTI, which is a normalized ratio of red and green band reflectance as given in
Eq. (2), of each date at each location as shown in Figure 5. At Haridwar, it is perceiv-
able that the turbidity was low almost every day after the lockdown started. However,
intermittent rainfall has changed the turbidity pattern in-between. It was observed
that the flow has also increased due to rainfall in the region. The results of Kanpur
are much better than Haridwar, as rainfall was less in this stretch. At Kanpur, it was
noticed that the turbidity has reduced continuously during the lockdown period. The
Yamuna River at Sangam, Prayagraj was showing uniform low turbidity, however, the
impact of lockdown are evident on Ganga River water. Even after the intermittent
rainfall, the water in the Ganga River stretch at Sangam is less turbid then ever
before. At this location also the flow has increased after the rainfall. It should be
noted that the Ganga River water is shallow here than the Yamuna River. The results
of NDTI are much better than the single-band approach in shallow water stretches
due to the normalization of band reflectance.
Varanasi is another stretch where the water is comparatively deeper. The NDTI
analysis at this location showed a consistent reduction in turbidity with the lockdown
time period. The analysis depicted that the NDTI value is consistently low in the
deep water during this period. It signifies that the turbidity has reduced during the
1190 V. GARG ET AL.

Figure 5. The temporal NDTI maps of Ganga River water at each location along with their FCC; (a)
Haridwar, (b) Kanpur, (c) Prayagraj, (d) Varanasi.

period of analysis. Based on the finding, it can be deduced that in the absence of
ground observed data, remote sensing approach can be used for preliminary estimates
on water quality. The present study focused on turbidity concentration, however, the
other optical properties of water can also be analyzed.
 GEOMATICS, NATURAL HAZARDS AND RISK 1191

Figure 5. Continued.
1192 V. GARG ET AL.

5. Conclusions
Turbidity is an important optical property of water that reduces the energy required for
aquatic growth. It has been reported that the water quality of Ganga River has improved
in terms of clarity due to recent Nationwide lockdown from 25 March–14 April 2020
(Phase-I) in India due to COVID-19. In the present study, the spatial and temporal
change in turbidity have been analyzed along the Ganga River during the said period
purely through remote sensing. The most attractive pilgrimage stretches of Ganga River
i.e., Haridwar, Kanpur, Prayagraj, and Varanasi were selected in this study. Initially, the
change in turbidity has been analyzed in terms of change in reflectance in the visible and
NIR region of Sentinel-2A/B data. It has been reported in the literature that with the
increase in turbidity the reflectance in visible region increases and vice versa. Similar
results were found in the present study, the temporal change study showed that reflect-
ance in each visible to NIR region has reduced, which might be attributed to a reduction
in turbidity in the water of the river. The blue and green bands could not map the spatial
heterogeneity in the turbidity variation due to interference from the bottom. However, a
slight reduction in reflectance or turbidity, even in these bands, could be seen in the
deep water. It was noticed that red and NIR bands are more sensitive towards turbidity
estimation. The deep water absorbs almost all the energy incident upon it, however, due
to the presence of suspended matters some energy reflects in these regions of the spec-
trum. These bands are very useful to quantitatively estimate the turbidity in the absence
of field observed data for optically deep water. However, it was also observed that the
small amount of precipitation can bring large sediments in the river and change in tur-
bidity immediately. Further, the results were verified through the NDTI band ratio tech-
nique. The analysis also confirmed that the turbidity has reduced during the lockdown
period. The main reason for the reduction in turbidity in the river, might be less effluent
generation and discharge into the river. Further, the activities of pilgrimage along the
river were below the minimum during this period of lockdown. It was realized that the
spatial water quality information can be generated using the remote sensing approach,
which is much better than field-based point information. It is well proven that remote
sensing can provide preliminary qualitative estimates of turbidity, however, it is always
recommended to verify the results through observed turbidity values. Therefore, the
retrieved turbidity through remote sensing approach required verification, which may
be conducted once the lockdown will be over. Further, the accuracy of the remote sens-
ing approach for water quality studies may be improved with very high spatial, spectral,
and temporal resolution datasets, especially for river water.

Disclosure statement
The authors declare no conflicts of interest.

References
Braga F, Scarpa GM, Brando VE, Manfe G, Zaggia L. 2020. COVID-19 lockdown measures
reveal human impact on water transparency in the Venice Lagoon. Sci Total Environ. 736:
139612.
 GEOMATICS, NATURAL HAZARDS AND RISK 1193

Brezonik P, Menken KD, Bauer M. 2005. Landsat-based remote sensing of lake water quality
characteristics, including chlorophyll and colored dissolved organic matter (CDOM). Lake
Reservoir Manage. 21(4):373–382.
Caballero I, Stumpf RP, Meredith A. 2019. Preliminary assessment of turbidity and chlorophyll
impact on bathymetry derived from Sentinel-2A and Sentinel-3A satellites in South Florida.
Remote Sens. 11(6):645.
Chander S, Gujrati A, Hakeem KA, Garg V, Issac AM, Dhote PR, Kumar V, Sahay A. 2019.
Water quality assessment of River Ganga and Chilika lagoon using AVIRIS-NG hyperspec-
tral data. Curr Sci. 116(7):1172–1181.
Chauhan A, Singh RP. 2020. Decline in PM2.5 concentrations over major cities around the
world associated with COVID-19. Environ Res. 187:109634.
Collivignarelli MC, Abba A, Bertanza G, Pedrazzani R, Ricciardi P, Carnevale Miino M. 2020.
Lockdown for COVID-2019 in Milan: What are the effects on air quality?. Sci Total
Environ. 732:139280.
CPCB (Central Pollution Control Board) 2020. Impact of lockdown on water quality of river
Ganga. CPCB, Ministry of Environment, Forest and Climate Change, Govt. of India, New
Delhi; [accessed 2020 May 12]. https://cpcb.nic.in/openpdffile.php?id=TGF0ZXN0RmlsZS8y
OTNfMTU4Nzk3ODU3MV9tZWRpYXBob3RvMTY3MDYucGRm.
Dantas G, Siciliano B, França BB, da Silva CM, Arbilla G. 2020. The impact of COVID-19 par-
tial lockdown on the air quality of the city of Rio de Janeiro, Brazil. Sci Total Environ. 729:
139085.
Doxaran D, Froidefond J-M, Lavender S, Castaing P. 2002. Spectral signature of highly turbid
waters Application with SPOT data to quantify suspended particulate matter concentrations.
Remote Sens Environ. 81(1):149–161.
Garg V, Senthil Kumar A, Aggarwal SP, Kumar V, Dhote PR, Thakur PK, Nikam BR,
Sambare RS, Siddiqui A, Muduli PR, et al. 2017. Spectral similarity approach for mapping
turbidity of an inland waterbody. J Hydrol. 550:527–537.
Gholizadeh MH, Melesse AM, Reddi L. 2016. A comprehensive review on water quality
parameters estimation using remote sensing techniques. Sensors. 16(8):1298.
G€uttler FN, 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 Sens Environ. 132:
86–101.
Hellweger FL, Miller W, Oshodi KS. 2007. Mapping turbidity in the Charles River, Boston
using a high-resolution satellite. Environ Monit Assess. 132(1–3):311–320.
Hu C, Lee Z, Franz B. 2012. Chlorophyll a algorithms for oligotrophic oceans: a novel
approach based on three-band reflectance difference. J Geophys Res. 117(C1):C01011.
IOCCG. 2000. Remote Sensing of Ocean Colour in Coastal, and Other Optically-Complex,
Waters; Reports of the International Ocean-Colour Coordinating Group, No. 3.
Sathyendranath S, editor. Dartmouth, NS, Canada: IOCCG.
Jain SK, Agarwal PK, Singh VP. 2007. Ganga Basin. Chapter 8 in hydrology and water resour-
ces of India, water science and technology library. Netherlands: Springer, p. 333–418.
Johnson RW. 1975. Quantitative sediment mapping from remotely sensed multispectral data.
In: Shahrokhi F, editor. Remote sensing of earth resources, Vol. IV. Tullahoma (TN): The
University of Tennessee Space Institute; p. 565–576.
Klemas V, Borchardt JF, Treasure WM. 1971. Suspended sediment observations from ERTS-1.
Remote Sens Environ. 2:205–221.
Kritikos H, Yorinks L, Smith H. 1974. Suspended solids analyses using ERT-A data. Remote
Sens Environ. 3(1):69–78.
Kuhn C, de Matos Valerio A, Ward N, Loken L, Sawakuchi HO, Kampel M, Richey J, Stadler
P, Crawford J, Striegl R, et al. 2019. Performance of Landsat-8 and Sentinel-2 surface reflect-
ance products for river remote sensing retrievals of chlorophyll-a and turbidity. Remote
Sens Environ. 224:104–118.
1194 V. GARG ET AL.

Lacaux JP, Tourre YM, Vignolles C, Ndione JA, Lafaye M. 2007. Classification of ponds from
high-spatial resolution remote sensing: application to Rift Valley Fever epidemics in Senegal.
Remote Sens Environ. 106(1):66–74.
Lal P, Kumar A, Kumar S, Kumari S, Saikia P, Dayanandan A, Adhikari D, Khan ML. 2020.
The dark cloud with a silver lining: assessing the impact of the SARS COVID-19 pandemic
on the global environment. Sci Total Environ. 732:139297.
Li L, Li Q, Huang L, Wang Q, Zhu A, Xu J, Liu Z, Li H, Shi L, Li R, et al. 2020. Air quality
changes during the COVID-19 lockdown over the Yangtze River Delta Region: an insight
into the impact of human activity pattern changes on air pollution variation. Sci Total
Environ. 732:139282.
Liedeke J, Roberts A, Luternauer J. 1995. Practical remote sensing of suspended sediment con-
centration. Photogramm Eng Remote Sens. 61(2):167–175.
Lim J, Choi M. 2015. Assessment of water quality based on Landsat 8 operational land imager
associated with human activities in Korea. Environ Monit Assess. 187 (6):384..
Liu H, Li Q, Shi T, Hu S, Wu G, Zhou Q. 2017. Application of Sentinel 2 MSI images to
retrieve suspended particulate matter concentrations in Poyang Lake. Remote Sens. 9(7):761.
Luis KMA, Rheuban JE, Kavanaugh MT, Glover DM, Wei J, Lee Z, Doney SC. 2019.
Capturing coastal water clarity variability with Landsat 8. Mar Pollut Bull. 145:96–104.
McFeeters SK. 1996. The use of the Normalized Difference Water Index (NDWI) in the delin-
eation of open water features. Int J Remote Sens. 17(7):1425–1432.
MHA (Ministry of Home Affairs) 2020. Order: Extension of Lockdown. Order No. 40-3/2020-
DM-I(A), MHA, Govt. of India. [accessed 2020 May 12]. https://www.mha.gov.in/sites/
default/files/MHA%20DO%20letter%20dt.14.4.2020%20to%20Chief%20Secretaries%20and%
20Administrators%20for%20strict%20implementation%20of%20Lockdown%20Order%20during%
20extended%20period.pdf.
Miller RL, McKee BA. 2004. Using MODIS Terra 250 m imagery to map concentrations of
total suspended matter in coastal waters. Remote Sens. Environ. 93(1-2):259–266.
Moore GK. 1980. Satellite remote sensing of water turbidity. Hydrol Sci J. 25(4):407–421.
Muhammad S, Long X, Salman M. 2020. COVID-19 pandemic and environmental pollution: a
blessing in disguise? Sci Total Environ. 728:138820.
Nechad B, Ruddick K, Park Y. 2010. Calibration and validation of a generic multisensor algo-
rithm for mapping of total suspended matter in turbid waters. Remote Sens Environ.
114(4):854–866.
NMCG (National Mission for Clean Ganga) 2020. Environmental and Social Management
Framework. NMCG, Ministry of Jal Shakti, Department of Water Resources, River
Development & Ganga Rejuvenation, Govt. of India, New Delhi; [accessed 2020 May 12]
https://nmcg.nic.in/writereaddata/fileupload/7_Environmental%20and%20Social%20Management%
20Framework.pdf..
Otmani A, Benchrif A, Tahri M, Bounakhla M, Chakir EM, El Bouch M, Krombi M. 2020.
Impact of COVID-19 lockdown on PM10, SO2 and NO2 concentrations in Sale City
(Morocco). Sci Total Environ. 735:139541.
Pahlevan N, Chittimalli SK, Balasubramanian SV, Vellucci V. 2019. Sentinel-2/Landsat-8 prod-
uct consistency and implications for monitoring aquatic systems. Remote Sens Environ. 220:
19–29.
Paital B, Das K, Parida SK. 2020. Inter nation social lockdown versus medical care against
COVID-19, a mild environmental insight with special reference to India. Sci Total Environ.
728:138914.
Pavelsky TM, Smith LC. 2009. Remote sensing of suspended sediment concentration, flow vel-
ocity, and lake recharge in the Peace-Athabasca Delta, Canada. Water Resour Res. 45(11):
W11417.
Quang NH, Sasaki J, Higa H, Huan NH. 2017. Spatiotemporal variation of turbidity based on
Landsat 8 OLI in Cam Ranh Bay and Thuy Trieu Lagoon, Vietnam. Water. 9(8):570.
Ritchie JC, McHenry JR, Schiebe FR, Wilson RB. 1974. The relationship of reflected solar radi-
ation and the concentration of sediment in surface water of reservoirs. In: Shahrokhi F,
 GEOMATICS, NATURAL HAZARDS AND RISK 1195

editor. Remote sensing of earth resources, Vol. III. Tullahoma (TN): The University of
Tennessee Space Institute; p. 52–72.
Ritchie J, Schiebe FR, McHenry JR. 1976. Remote sensing of suspended sediments in surface
waters. Photogramm Eng Remote Sens. 42(12):1539–1545.
Ritchie JC, Zimba PV, Everitt JH. 2003. Remote sensing techniques to assess water quality.
Photogramm Eng Remote Sens. 69(6):695–704.
Saadat S, Rawtani D, Hussain CM. 2020. Environmental perspective of COVID-19. Sci Total
Environ. 728:138870.
Sebastia-Frasquet M-T, Aguilar-Maldonado JA, Santamarıa-Del-Angel  E, Estornell J. 2019.
Sentinel 2 analysis of turbidity patterns in a coastal lagoon. Remote Sens. 11(24):2926.
Shi W, Wang M. 2009. Satellite observations of flood-driven Mississippi River plume in the
spring of 2008. Geophys Res Lett. 36(7):L07607.
Tassan S. 1997. A numerical model for the detection of sediment concentration in stratified
river plumes using Thematic Mapper data. Int J Remote Sens. 18(12):2699–2705.
Toming K, Kutser T, Laas A, Sepp M, Paavel B, N~ oges T. 2016. First experiences in mapping
lake water quality parameters with Sentinel-2 MSI imagery. Remote Sens. 8(8):640.
Trinh RC, Fichot CG, Gierach MM, Holt B, Malakar NK, Hulley G, Smith J. 2017. Application
of Landsat 8 for monitoring impacts of wastewater discharge on coastal water quality. Front
Mar Sci. 4:329.. 00329.
Yunus AP, Masago Y, Hijioka Y. 2020. COVID-19 and surface water quality: improved lake
water quality during lockdown. Sci Total Environ. 731:139012.