TANAYA DAS

Saktipur; Agarpara; Kolkata 700109; West Bengal;

Contact No: 9073150378; E-Mail: tanaya16.das@gmail.com

Academic Qualification
B.Sc with Honours in Physics from University of Calcutta (2009).
Masters in Computer Application from West Bengal University of Technology (2012).
M.Phil in Oceanography and Coastal Management from School of Oceanographic Studies, Jadavpur University
(2014).
Ph.D from School of Environmental Studies, Jadavpur University (2021).

 Research Area:

 Estuarine Hydrodynamics and Ecological Modeling (Development of hydrological and ecological model,
Parameterization, Data analysis, Simulation of hydrological and ecological data);
 Coastal Hazard Mapping (Identification of hazardous areas, Vulnerability analysis, Development of
sustainable solution);
 Climate Research (Monsoon, Indian Ocean);

 Thesis

2014

Uncertainty Estimation in Projection of Indian Summer Monsoon Rainfall using Nonlinear Relationship
between Indiaces for Equatorial Indian Ocean Oscillation.River System. Submitted for the Degree of M.Phil in
Oceanography and Coastal Zone Management. Under guidance of Prof. Goutam Kumar Sen.

2021

Tidal Hydrodynamics of Sundarbans Estuarine System. School of Environmental Studies Faculty Council of
Interdisciplinary Studies, Law and Management. Jadavpur University.Kolkata, India (for the Degree of Doctor
of Philosophy (Science)). Name of the Supervisor: Goutam Kumar Sen. Professor (Retd.). School of
Oceanographic Studies. Jadavpur University.

 Publications

2019
 Chakraborti, S., Das, T., Sen, G. K., & Mukherjee, J. (2019, November). Monsoonal variability of
phytoplankton in a tide dominated river of Sundarban. In Biotechnology and Biological Sciences:
Proceedings of the 3rd International Conference of Biotechnology and Biological Sciences
(BIOSPECTRUM 2019), August 8-10, 2019, Kolkata, India (p. 219). CRC Press.

2018
 Das, T., Chakraborti, S., Mukherjee, J., & Sen, G. K. (2018). Mathematical modelling for phytoplankton
distribution in Sundarbans Estuarine System, India. Ecological Modelling, 368, 111-120.

2015
 Goutam, K. S., Tanaya, D., Anwesha, S., Sharanya, C., & Meenakshi, C. (2015). Tide and mixing
characteristics in Sundarbans estuarine river system. Hydrology: Current Research, 6(2).

 Conferences

2020

 Das Tanaya, Sen Goutam Kumar. Dynamical control on estuarine hydrodynamics of Sundarbans, India:
Classification based on vulnerability indices. 1st Online International Conference on Recent Advances in
Computational and Experimental Mechanics. September 4-6, 2020, IIT Kharagpur.
2017

 Das Tanaya, Chakraborti Sharanya, Mukherjee Joydeep Sen Goutam Kumar. A note on dynamic
initialization of numerical model for plankton dynamics in Sundarbans Estuarine Systems, India. 7th
International Conference on Theoretical, Applied, Computational and Experimental Mechanics.(ICTACEM
2017). Organized by Department of Aerospace Engineering, Indian Institute of Technology, Kharagpur,
December 28-30.

2015

 Das Tanaya, Sen Goutam Kumar. Uncertainty Analysis of Indian Summer Monsoon Rainfall (ISMR) in
Stochastic Framework. National Conference on Emerging Trends in Physics of Fluids and Solids
(NCETPFS-2015). Organized by Department of Mathematics, Jadavpur University, February 24-25.

2014

 Das Tanaya, Sen Goutam Kumar. A Note on Uncertainty Estimation in the Projection of Indian Summer
Monsoon Rainfall (ISMR). 6th International Conference on Theoretical, Applied, Computational and
Experimental Mechanics.(ICTACEM 2014). Organized by Department of Aerospace Engineering, Indian
Institute of Technology, Kharagpur, December 29-31.

 Das Tanaya, Sen Goutam Kumar. Relationship between Index for the Equatorial Indian Ocean Oscillation
(EQUINOO) and Indian Summer Monsoon Rainfall (ISMR). 3rd International Symposium on Complex
Dynamical Systems and Applications. Organized by Physics and Applied Mathematical Unit and
Agricultural and Ecological Research Unit, Indian Statistical Institute, March 10-12.

 Das Tanaya, Sen Goutam Kumar. Comparison between Indices for the Equatorial Indian Ocean Oscillation
(EQUINOO) and Indian Summer Monsoon Rainfall (ISMR). National Conference on Emerging Trends in
Physics of Fluids and Solids (NCETPFS-2014). Organized by Department of Mathematics, Jadavpur
University, March 06-07.

 Workshops

2015
 International Training Course on 'Indian Ocean Dynamics: From the large-scale circulation to small-scale
eddies and fronts’ by ITCOocean, INCOIS, Hyderabad, November 16-27.

2014
 Near Field Communication, Electronics and Computational Intelligence (NFCECI-2014). Organized by
Department of Electronics and Communication Engineering, College of Engineering and Management,
Kolaghat in technical collaboration with The Institute of Engineering and Technology (IET), April 11-13.

2013
 Short term Course Study in Sustainable Development under RIO+20b India Program on Sustainable
Development by IARC Education and Outreach Division.

 Work Experiences

2015-2017

 Name of the Project: Study of Tidal Dynamics and Interaction with Surge in Sundarban Estuary

Designation: Project Fellow, School of Oceanographic Studies, Jadavpur University

Project Overview: A national level project funded by INCOIS, Ministry of Earth Science in HOOFS under
supervision of Prof. Goutam Kumar Sen (PI), Dr. Joydeep Mukherjee(Co-PI) , Jadavpur University and Dr.
Meenakshi Chtterjee (Co-PI), Basanti Devi College, Kolkata. The main objective of the project was to simulate
hydrological parameters in estuarine rivers to develop a full-fledged hydrological estuarine model for
forecasting.
Key Responsibilities:
(i) Analysis of hydrological data (water level, current, salinity etc.) of estuarine rivers of Sundarbans and
identification of major physical processes.
(ii) Investigation of the interaction of hydrological flow regime with other physical and climatic factors during
extreme weather conditions.
(iii) Validation of outcomes to the real data by conducting field surveys.
(iv) Developing a hydrological simulation model in the estuary to simulate hydrological parameters at different
locations of the estuary.
(v) Drafting technical notes, letters, conference proceedings, project reports, research papers, reviews and other
related documents as per the requirement of the project.
(vi) Preparation of work plan, budget and draw up questionnaires for assessment of stakeholder’s viewpoint.

2014

 Name of the Project: Sundarbans Estuarine Programme

Designation: Field Scientist, Sundarbans Estuarine Programme 2014
Project Overview: A national level project funded by INCOIS, Ministry of Earth Science in HOOFS under
supervision of Dr. Meenakshi Chtterjee (PI), Basanti Devi College, Kolkata and Prof. Goutam Kumar Sen (Co-
PI), Dr. Joydeep Mukherjee(Co-PI) , Jadavpur University. The main objective of the project was to collect
continuous data for hydrological and ecological parameters of the estuary and analysis to understand the
characteristics of the flow regime.

Key Responsibilities:
(i) Organization workshop and training classes to build scientific temperament among the stakeholders to
involve them in a month long data collection process to be carried out in seven principal estuaries of Sundarban.
(ii) Organization and implementation of data collection process in continuous tidal cycle in the monsoon period
(July, 2014) and day-to-day coordination to smooth the entire process in a planned way.
(iii) Analysis of time series data for continuous monitoring and minimization of error during collection period.

I declare that the entries in this document are true to the best of my knowledge and belief.

Date: 8th September, 2021

Place: Kolkata
(TANAYA DAS)
 Tidal Hydrodynamics of Sundarbans, India

An immense archipelago of islands, known as “the Sundarban”, which means “the beautiful

forest”, lies on the easternmost coast of India, between the Bay of Bengal and the plains of

Bengal. Some of the islands are vast and some are no larger than sandbars; some have existence

trough recorded history while others have just reshaped by the tides. There are no borders to

divide fresh water from salt, river from sea, even land from water. At the beginning of the last

century a visionary Scotsman, Sir Daniel Mackinnon Hamilton founded a settlement where

peoples of all races, classes and religions could live together along with the man-eating tigers

and the crocodiles, who rule the land and waters of Sundarbans. Indeed, the idea was utopian!

The Saptamukhi, Thakuran, Matla, Bidya, Gomdi, Gosaba, Gona, Harinbhanga and Raimangal-

all the seven principle estuaries of Sundarbans are north-south flowing rivers, lying the east of

River Hoogly. Sundarbans, India (21.25o – 22.5o N and 88.25o – 89.5o E), is a funnel-shaped,

wide mouthed complex estuarine system, interconnected through numerous east-west flowing

rivers and tidal creeks. Vulnerability to the climate extremes are the major stress and threat to

the living beings and environment of this tidal country. The mangrove forest acts as natural

protected barrier for the densely populated urbanization to its north reducing the direct impact

of the surge water from the sea accelerated by cyclonic storms. The vulnerability was well

exposed when on 25 May, 2009, surges of 2-3 m due to Cyclone Aila swept through the river

channels causing severe destructions of river protective constructions and households on the

either banks of rivers. The entire biosphere reserve of Sundarbans was inundated under 2-6 m

water level for several days. The loss of lives of man and animal were massive. The agricultural

fields on the river banks lost productivity due to saline water intrusion. The region needs special

care towards the environment management. It is futile without the knowledge of tides and

circulations in estuarine waters in this tidal country.

Since the surge water takes time to propagate through the long estuarine channels to reach the

densely populated northern sector, the prediction of surge can help in prior decision making

strategies to prevent inevitable destruction of life and property from huge water mass. Any

progress in understanding the tidal water propagation along the estuarine rivers cannot be

visualized without nearly perfect reconstruction of water level variations. These requirements

emphasize to study tidal hydrodynamics, first.

Study Area

River Saptamukhi is the westernmost among the seven principal estuaries of Sundarbans. It lies

on the east of River Hoogly and carries major portion of the inflow from the bay in form of large

flood volume through two eastern (Saptamukhi East Gullay) and western (Saptamukhi West

Gullay) branches during spring, neap and mean phases.

Jagaddal, a small estuarine river is the area of very interest to study the tidal activities. River

Jagaddal flows parallel to the stronger branch of the River Saptamukhi, the Saptamukhi East

Gullay. Debouching into bay (mouth), the river is connected with the adjoining estuary of River

Thakuran through Dhanchi Khaal (Canal). In the northern limit (head), Jagaddal is connected to

River Bar Chara, River Kal Chara and Pakhi Nala.

It is on the Jagaddal that the three observation stations are situated (south to north): Indrapur

(21.651o N, 88.413o E; Distance: 8.15 km), Dhanchi (21.704o N, 88.433o E; Distance: 13.75 km)

and Ramganga (21.796o N, 88.37o E; Distance: 35 km). The hydrological conditions of the

station Indrapur, about 8 km north of the mouth and closest to the Bay of Bengal, are suitable to

serve as the mouth of the estuary. The station Dhanchi is at the junction of river Jagaddal and

connected to river Thakuran through Dhanchi Cannel. The station Ramganga is at the junction of
three different linkages of river Bar Chara, river Kal Chara (connected to Hoogly River to the

west) and Pakhi Nala (connected to Thakuran) [Chatterjee et al. 2013].

The comprehensive state of the knowledge about water circulation in this flow regime is given

by Chatterjee et al. 2013. Complex river bathymetry, varying river width, meandering, effect of

channel geometry, transition between flow-dominated to ebb-dominated nature in presence of

freshwater source, nonlinearities from frictional effect have made the propagation of semidiurnal

tide along the channel more ambiguous, where duration asymmetries are found throughout the

channel. One of the striking features of this asymmetry is the „wall of water‟ during tidal stand

due to significant modification of tidal asymmetry. This period of longer persistence of the high

water level around high tide indicates the more likely occurrence of storm surge in coincident

with high tide may amplify water level during their passage through the river channels leading to

enormous chance of destruction. This river plays a crucial role as the spawning zone and nursery

of a wide variety of fishes, shrimps and crustaceans. The frontal zones i.e., the boundary zones

between stratification and mixing, found along the river, are the area of upwelling. These small

and frequent pockets of flow regime located near the mangrove forest covered river banks along

with marshland (Dhanchi Reseve Forest and Rakkaskhali) are the most productive zones. In

short, River Jagaddal is a flow regime of considerable environmental and economic importance.

Objectives

The present study aims at a novel approach to understand the tide and the estuarine processes by

examining the influence of tidal motion to the estuarine waters through following queries:

I. To develop a tidal model of estuary to carry out :

a) Simulation of tidal propagation along the estuarine channel

b) Computation of tidal current

c) To analyze the spatial and temporal characteristics of tidal elevation and current

in estuarine conditions such as: strong variable feature of bathymetry and

 nonlinear interactions, river width, inflow of fresh water.

d) To simulate tidal motion in extreme sea level condition.

II. Using tidal characteristics nature of stratification and mixing in tidal rivers are

investigated to understand the nature of inward salinity transport and distribution in the

estuarine river system.

III. To develop a numerical model to understand the distribution of nutrient and

phytoplankton in tidal river.

Methodology

The Tide Model

It is known that the tide at the mouth of the Sundarbans estuaries is predominantly semi-diurnal

[Majumdar, 1942]. The semi-diurnal tides propagate along the deep central portion of the

funnel-shaped Bay of Bengal and reach first the head of the „Swatch of No Ground‟ (21.083◦N,

89.283◦E), a deep submarine trench in the Bay almost in line with the central part of the delta. It

then travels along the delta face, bifurcating into an easterly and westerly branch. This

bifurcation and the irregular seaward extension of the delta contribute to the arrival of the tide on

different time at different points at the same latitude along the seaface [Majumdar, 1942]. There

are some other interesting features in tidal variations in Sundarbans [Chatterjee et al. 2013]: the

tidal range increased from mouth to head in all estuaries, but the degree and rates of

amplification over the various estuarine stretches were not uniform. The rate of change of water

level is low around high and low tide. Slow changes in the water levels were evident in the

water-level data, suggesting the existence of a stand. The tidal stand modified significantly the

tidal duration asymmetry in the Sundarbans estuary. The magnitude asymmetry of the tidal

currents is a direct consequence of the duration asymmetry. Since the observational was

conducted during the dry season, when the freshwater inflow is possible only from the Hoogly in

the west and into the Raimangal in the east, therefore, all the seven estuaries exhibited a flow-

dominant duration asymmetry. However, the flow dominance gave way to ebb dominance during
some tidal cycles at a few stations. Ebb dominant tidal asymmetries are caused by the presence

of upstream freshwater [Chugh, 1961] and the presence and effects of tidal flats [Speer and

Aubrey, 1985]. Cross-channel 9east-west) confluences and bottom roughness (contributed by

mangrove swamp [Mazda et al. 1995]) significantly modifies the tidal velocity and amplitude in

Sundarbans estuary [Goutam et al. 2015].

Very little else is known about tidal dynamics of Sundarbans. Physical processes are important

for hydrology of mangrove estuary, like Sundarbans [Wolanski et al. 1990]. We use depth-

averaged hydrodynamic model for shallow estuary to solve the tidal motion numerically for tidal

water level and current [Li et al. 1998]. The calculation involves the following steps:-

First, the observed horizontal velocity components are analyzed by harmonic decomposition so

that the amplitude and phase of the velocity at the major tidal frequency are obtained.

Second, the phase of the longitudinal velocity is then fitted to a quadratic function of the water

depth to obtain the phase of the pressure gradient.

Third, the drag coefficient, as a function of the transverse position, is then calculated from the

phase relationship of the longitudinal momentum equation.

In order to correctly recover the drag coefficient or validate this method, the following tests will

be designed [Li et al. 2004]:

– First step involves solving an analytic model with a given spatial distribution of

bottom drag coefficient in a rectangular domain.

– Second step uses the velocity field obtained from the analytic model as

hypothetical observational data and applies the phase-matching method to recover

the drag coefficient.

Nonlinear and Asymmetric Features

Tidal dynamics in estuary vary due to different complex, often nonlinear and non-stationary

forcing. The nonlinear response of tide to estuarine regime can be non-stationary and asymmetric
in nature, and often interact with each other. The asymmetric nature of tidal cycle in estuaries of

Sundarbans indicates underlying nonlinear and nonstationary processes, those are poorly

understood. The Hilbert Huang transform (HHT, composed of empirical mode decomposition

and Hilbert transformation) can be applied for processing non-stationary, non-linear signals

[Huang et al. 1998]. Empirical mode decomposition method decomposes the data set into several

intrinsic mode functions (IMFs). Through the Hilbert transform, IMFs could obtain amplitude

and instantaneous frequency with time. Those IMFs with amplitude and frequency can be used to

represent those factors that affect the tidal dynamics in estuary. The physical interpretation

represented by IMFs has been inferred in conjunction with the hydrological data available in the

estuary.

In this study we use that observational data to investigate the determining factors that affect the

tidal propagation and salinity variation. The results show that the frequency modes of the EMD

can capture the semidiurnal and other (non M2) variability and diurnal tides. The presence of

higher harmonics suggests the strong interactions between the incoming tide and the channel

geometry and nonlinear friction. This short report is meant to gain a new insight into tidal

analysis that can help studies of changes in tidal characteristics and the relation of these changes

to extreme sea level conditions.

Salinity Structure and Circulations in Estuary

Salinity profile in estuaries of Sundarbans plays a key role in maintaining delicate balance in

marine ecosystem. The interconnected estuaries are connected to River Hoogly to the west.

Sufficient freshwater flow through the south-west region from upstream is necessary to maintain

the marine ecosystem in Sundarbans. Siltation, mixing with ocean water, possible global

warming led evaporation, lack of freshwater and tidal water supply due to anthropogenic

activities has done considerable changes in variations in salinity pattern [Banerjee K (2013)].

The tide-model generated water level and current are used to investigate variations in both the
 salinity and estuarine current structure, through characteristics of horizontal salinity gradient,

diffusivity, mixing and stratification [Geyer, 2005].

The Plankton-Nutrient Distribution Model and Uncertainty Estimation

Sundarbans Estuarine System is a unique laboratory of ecological biodiversity that interacts with the
estuarine hydrology. Under the changing climate scenario, it is very crucial to develop a predictive
hydrological model coupled with ecological dynamics to monitor the nature of the planktonic group as well
as the flow regime to build a sound environmental management system. The main difficulty arises when
describing plankton dynamics under wide range of estuarine environmental conditions due to inadequate
knowledge to interpret the mechanism to which the environmental nature of a given situation force the
model condition towards the observation and the extent to which the details of the model implementations
do.
1. A simple 2 – state variable modeling approach is taken for the simulation of

phytoplankton and nutrient distribution coupled with mixing process in the estuary

[Frank, 2002; T.Das et al. 2018].

2. A sensitivity analysis is performed to determine the dominant parameters in the estuarine

ecological system.

3. The stability analysis test is conducted to identify key parameters to maintain the

ecological stability.

4. The influence of salinity structure and estuarine circulations on the distribution of

phytoplankton and nutrients in estuarine water has been investigated.

5. While parameterizing model variables and formulate model structure, incomplete

understanding of the effect of tidal hydrodynamics on plankton distribution, response of

planktons to the hydrological forcing and the natural variability of plankton dynamics

contribute to uncertainties. Our aim is to estimate model uncertainties using two criteria

[Giorgi et al., 2002]: the performance of the model in reproducing distribution of state

variables in model domain (“model performance” criteria) and the convergence of

 simulated projections over certain time period (“model convergence” criteria). We

attempt in quantification of uncertainties while modeling the distribution of

phytoplankton and nutrients along the estuarine river.

Salient Outcomes

1. Propagation of tide and tidal current are simulated along the R.Jagaddal. Tidal range

increased from mouth to head. The model simulations follow the observation pattern.

2. Most important factors: Channel geometry and frictional dissipation.

3. Nonlinear effect due to bottom drag coefficient is important driving factor in tidal

elevation variation in this mangrove estuary.

4. Along R. Jagaddal, tidal range increased from mouth to head, but decreased between

Indrapur and Dhanchi. These characteristics have been captured by the model.

5. Cross-channel flow (west-east) is a dominant feature in this region [Goutam et al. 2015].

6. Variation in bottom drag coefficient, tidal current, inflow through two intervening

channels near Dhanchi, one connected with Thakuran and other connected with

Saptamukhi, contribute to tidal amplitude and current variation.

7. The drag coefficient is larger in the shallow water than in the deep water. It indicates an

effect of increased nonlinear friction. This leads to the generation of higher harmonics.

The superposition of higher harmonics leads to tidal asymmetry.

8. River Jagaddal (as observed from, gradient Richardson number variation) shows

stratification characteristics from Indrapur to Dhanchi. But beyond Dhanchi upto the head

(further north) it exhibits a zone of strong mixing.

9. This zone of strong mixing is indicative of the feature of considerable fresh water flow

from River Hoogly.

10. This study on phytoplankton distribution is an initial attempt to model the plankton-

nutrient dynamics (distribution, growth rate, mortality rate) in the wide range of estuarine

condition (nutrient distribution, mixing level etc.).

11. Using sensitivity analysis we determine which parameters having the most influence,

interaction with other variables and the possible mechanism underlying the ecological

processes in estuarine environment.

12. The rate of change of mixed layer provides the physically driven rate to change the

concentration of phytoplankton and nutrient in the water column.

13. The state variables (phytoplankton and nutrient concentration) maintain the stable

behavior during tidal cycle [T.Das et al. 2018].

14. The model reproduces different aspects of spatial variations of state variables

(phytoplankton population and nutrient concentration) within specified time scale. The

model simulations converge over the time within specified hydrological and ecological

forcing scenario.

15. The greater model convergence implies higher reliability of capturing interactions of

processes with much better sensitivity.

The study provides a detailed methodology to understand the processes behind the nature of
estuarine waters of Sundarbans and the influence of the tides and associated physical processes
on the plankton dynamics. The study sets a strong background needed for the solid
environmental management plan to keep the nature and lives of estuaries sustainable and
flourished.

References

1 Amitav Ghosh, The Hungry Tide (HarperCollins Publishers India, Fifth impression,

2013).

2 Banerjee K (2013) Decadal Change in the Surface Water Salinity Profile of Indian

Sundarbans: A Potential Indicator of Climate Change.J Marine Sci Res Development

S11: 002. Doi: 10.4172/2155-9910.S11-002.

3 Chatterjee Meenakshi, Shankar D, Sen G K, Sanyal P, Sundar D, Michael G S, Chatterjee

Abhisek, Amol P, Mukherjee Debabrata, Suprit K, Mukherjee A, Vijith V, Chatterjee

Siddhartha, Basu Anwesha, Das Madhumita, Chakraborti Saranya, Kalla Aravind, Misra

Surja Kanta, Mukhopadhyay Soumya , Mandal Gopal and Sarkar Kankan, Tidal

variations in Sundarbans Estuarine System, India , J. Earth Syst. Sci. 122, No. 4, August

2013, pp. 899–933.

4 Chugh R S 1961 Tides in Hooghly River; Hydrol. Sci. J. 6(2) 10–26. Speer P E and

Aubrey D G 1985 A study of non-linear tidal propagation in shallow inlet/estuarine

systems, part II – Theory; Estuarine Coast. Shelf Sci. 21207–224.

5 Eric Wolanski, Yoshihiro Mazda, Brian King, Stephen Gay, Dynamics, flushing and

trapping in Hinchinbrook channel, a giant mangrove swamp, Australia, Estuarine, Coastal

 and Shelf Science, volume 31, Issue 5, 1990, Pages 555-579, ISSN 0272-7714,

https://doi.org/10.1016/0272-7714(90)90013-H.

6 Franks Peter J.S. NPZ models of plankton dynamics: their construction, coupling to

physics, and application, Journal of Oceanography, Vol. 58, pp.379 to 387, 2002.

7 India Meteorological Department 2009 Severe Cyclonic Storm, AILA: A Preliminary

Report; Regional Specialized Meteorological Centre–Tropical Cyclones, New Delhi, 26p.

8 Geyer W.R. 2005. Estuarine salinity structure and circulation.

9 Giorgi Filippo, Mearns Linda O. 2002. Calculation of average, uncertainty range, and

reliability of regional climate changes from AOGCM simulations via the “Reliability

Ensemble Averaging” (REA) method. American Meteorological Society. 1141-1158.

10 Goutam, K.S., Tanaya, D., Anwesha, S., Sharanya, C., Meenakshi, C., 2015. Tide

andMixing Characteristics in Sundarbans Estuarine River System, Hydrol. Curr. Res. 6,

204.

11 Huang Norden E., Shen Zheng, Long Steven R., Wu Manli C., Shih Hsing H., Zheng

Quanan, Yen Nai-Chyuan, Tung Chi Chao and Liu Henry H., The empirical mode

decomposition and the Hilbert spectrum for nonlinear and non-stationary time series

analysis, Proc. R. Soc. Lon. A (1998) 454, 903-995.

12 Wikipedia 2012 Cyclone Aila; http://en.wikipedia.org/ wiki/Cyclone Aila.

13 Li, C., A. Valle-Levinson, K.-C. Wong, and K.M.M. Lwiza (1998), Separating baroclinic

flow from tidally induced flow in estuaries, J. Geophys. Res., 103, 10,405-10,417.

14 Li, C., Valle-Levinson Arnoldo, Atkinson Larry P., Wong Kuo Chuin and Lwiza

Kamazima M.M., J. Geophys. Res., Vol. 109, C03034, 2004.

15 Majumdar S C 1942 Rivers of the Bengal Delta; Calcutta University Readership

Lectures, Calcutta University,124p.

16 Mazda, Y., Kanazawa, N. and Wolanski, E., 1995. Tidal asymmetry in mangrove creeks.

Hydrobiologia, 295(1): 51-58.

17 Speer, P.E., and D.G. Aubrey, A study of nonlinear tidal propagation in shallow inlet/

estuarine systems, II, Theory, Estuarine, Coastal Shelf Sci., 21, 207-222,1985.

18 T.Das, S.Chakraborti, J.Mukherjee, G. Sen. Mathematical Modelling for Phytoplankton

Distribution in Sundarbans Estuarine System, India. Ecological Modelling 368 (2018)

111–120.
 Tidal Hydrodynamics and Vulnerability of Sundarbans,
India

Tanaya Das

email: tanaya16.das@gmail.com.

Abstract. Sundarbans Estuarine System, India, is an interesting laboratory for

studying the wide range of macro-tidal conditions and the impact on tidal dy-
namics. Energy is dissipated in tidal current by friction on the bottom. The en-
ergy is dissipated by the viscous drag layers of water moving parallel to the bot-
tom of the river. In macro-tidal mangrove estuaries the effect of bottom friction
on dissipation of energy in tidal currents is modulated by the mechanism of
drag force in mangrove swamps. The dissipation of energy in the tides has re-
cently formed the subject of critical research to address the question of adapta-
bility of tidally-dominated estuaries to increasing trends of extreme sea-level
events (like cyclone accompanied by storm surge ) associated with global envi-
ronmental changes. In more recent times, since cyclones began to be named in
2006, the Sundarbans, the largest monsoonal, macro-tidal, mangrove delta-front
estuarine system of India, has seen Sidr (November, 2007), Rashmi (October,
2008), Aila (May, 2009), Bulbul (2019) and the latest, Amphan on 20 May
2020. In seaward region of this macro-tidal estuary, impact from changes in sea
level on energy dissipation levels is significant. Every time a cyclone has hit the
Sundarbans, the need to deliver climate risk resilient sustainable infrastructures
is strongly felt. The method of study and findings are applicable to assess the
vulnerability of tidal estuaries due to anthropogenic and environmental hazards
and to formulate sound ecological management plan for coastal protection
against storms.

Keywords: Vulnerability, Tidal energetics, Tidal dissipation energy, Drag

force, Estuary, Sundarbans.

1 Introduction

During the super cyclone Amphan, the sea-facing islands (Patharpratima, G-Plot)
along with mangrove-swamp system in Dhanchi Reserve Forest, on the Jagaddal Riv-
er Estuary (Figure 1), mainly the western sector of Sundarbans were the worst hit.
This shallow estuary, enriched by mangrove-swamp system in Dhanchi Reserve For-
est, potential fishing zone and marine resource trading ports in Patharpratima serves
as one of the most potential marine resource zones of Sundarbans. The hydrodynamic
conditions of Jagaddal River Estuary make it extremely risk-prone zones in extreme
sea-level conditions ( Figure.2).
2

The Jagaddal, flowing parallel to the Saptamukhi East Gulley, is a shallow estuary
(depth around ). Jagaddal, debouching into the bay, has links with the adjoining
estuary, the Thakuran in the east through the Dhanchi Khaal, Pakhi Nala, and Shibua
Gaang. In the northward direction (head) river Jagaddal has linked with river Bar-
chara, river Kalchara and Pakhi Nala. River Bar Chara and river Kal Chara are influ-
enced by fresh water discharge of Hoogly river whereas Pakhi Nala is connected with
water flow of river Thakuran (Chatterjee et al. 2013; Goutam, K.S. et al.2015).
Cross-channel flow (west-east) and stronger tidal current, ebb-flow transition in
tidal flats are the controlling factor on drag force leading to nonlinear and asymmet-
rical pattern in tidal regime of Jagaddal. The height of inundation happens to be lower
and over shorter duration in the forested area of Dhanchi, which also drain out water
quickly. But, the southern part (G-Plot) towards Indrapur (closest to the Bay of Ben-
gal, about 8 km north) there is no forest and completely unguarded from surge waters.
The height of inundation happens to be lower and over shorter duration in the forested
area, which also drain out water quickly. In the upstrem (Patharpratima - Ramganga),
where adjoining river confluences contribute to strong tidal current and supress drag
force, destructions due to surge impact on tidal dissipation have been clearly visible
during Cyclone Amphan.
In this context, identification of vulnerable risk-prone zones, the areas potentially
experiencing severe destruction due to tidal activities in extreme conditions, consti-
tutes a significant step in terms of risk mapping is a fundamental basis for developing
sound environmental management plan and decision-making strategies. This study
addresses the question of how tidal variation in estuary response to vulnerabilities
during such extreme sea level conditions. The aim was to determine the vulnerability
conditions based on energetics and controlling factors of tidal hydrodynamics in the
macro-tidal mangrove estuary.

In order to answer these queries the plan of the study is as follows: Tidal propagation
dynamics and method of vulnerability assessment are described in Section 2. Section
3 contains analysis results to gain insights into estuarine dynamics, factors controlling
vulnerability and conditions of vulnerability. Finally, in Section 4, we present the
main conclusions those can be drawn from the study.

2 Methodology

We use depth-averaged hydrodynamic model to simulate the tidal wave propagation

along the Jagaddal River Estuary. The tidal water level and current obtained as model
solutions are used to calculate dissipation of energy in tidal currents. We apply classi-
fication tree methods to identify the vulnerability classes based on calculated tidal
energetics. We apply Random Forest and support vector machines (SVMs) for more
precision and to understand the importance of the dynamic estuarine processes.
 3

2.1 The depth-averaged hydrodynamic model

The model geometry is chosen to be symmetric about an axis along the channel, with
constant lateral depth. The axis lies along the boundary and points toward the head
of the estuary. The axis is taken to lie along the open boundary at A single
frequency semi-diurnal tide is imposed at the mouth of the estuary. The tidal ampli-
tude and phase of sea level variation are assumed to be uniform across the estuary and
specified. The depth-averaged, shallow water momentum and continuity equations are
used (Li and Valle-Levinson, 1999):

( )

( ) ( )

where u, v, , h, x, y, t, are longitudinal velocity, lateral velocity, eleva-

tion, water depth, longitudinal coordinate, lateral coordinate, time, friction coefficient,
and the gravitational constant, respectively. , the friction coefficient is defined by
(Proudman, 1953; Parker, 1984):

in which and are the bottom drag coefficient and the velocity component of the
tidal flow. In general, is a function of position and is to be determined by observa-
tions. Here, we consider the semi-diurnal tidal component only (Huntley et al. 1994).
The friction coefficient , a function of along and cross-channel spatial variation, is
dependent on the current amplitude and the bottom drag coefficient. Spatial distribu-
tion of for every segment of the channel and also the interpolated longitudinal
velocity along the channel from observed current data (see, Figure. 2)are used to
provide initial values for and then β to solve equation (1) numerically.
For propagation of tidal wave through estuarine channel we solve the depth-
averaged model equations for tidal current and amplitude following the steps:-
(i) The first step solves a numerical model with a given spatial distribution of fric-
tion coefficient in a channel with depth and width variation along and cross channel to
compute tidal amplitudes and velocity components.
(ii) In second step we use the velocity component computed in the previous step
and apply the phase-matching method (Li et al. 2004) to recover the drag coefficient.
(iii) Then by using the drag coefficient estimated from the phase-matching method
(Li et al. 2004) as a function of along-channel position in the numerical model, we
4

have estimated the tidal current and tidal amplitude and compare with the observed
amplitude.
(iv) Step (i) to (iii) are repeated if the computed tidal amplitude are not close to the
observed data.

2.2 Tidal energetics

If the friction coefficient β is known, tidal energy dissipation by bed friction occurs
at the rate 𝜌β 3 per square meter (𝜌 is the density of water). At any estuarine channel
of depth, , the summation of all such upstream dissipation, DISS, is approximated
by (Taylor, 1920):

𝜌 ( )
where is the phase difference between tidal water elevation and current ampli-
tudes. is the linearized friction coefficient and , tidal fre-
quency.
In estuaries of Sundarbans, the tidal ebb and flood mass exchanges can be directly
represented by the mean tidal current amplitude, , which conveniently corresponds
to the (predominant) semi-diurnal lunar tidal constituent, M2. Net tidal energy dissipa-
tion is proportional to , and net upstream dissipation can be directly calculated
from the numerical solutions of depth-averaged hydrodynamic model described in
Section 3.1.

2.3 Statistical methods

We applied classification tree methods, an approach based on machine learning, pro-
vides effective ways to generate models that are very easy to explore numerically and
graphically complex relationships in the data (De‟ath and Fabricius 2000; Selkimäki et al.
2011). Among the algorithms that generate decision trees, we use classification and
regression trees (CART), a recursive partitioning method popular for its non-
parametric approach and interpretability (Breiman et al. 1984; Quinlan et al. 1986). We
followed the following steps of this analysis: a) divide data using splitting criteria, b)
forming a final tree, c) pruning the final tree to reduce its size and increase its classifi-
cation abilities avoiding over fitting. We use Gini index as a measure of impurity to
define the splitting criteria. To achieve more precision in predictives, balancing error
in imbalanced data and to estimate the importance of variables used in the classifica-
tion, we use Random Forest and support vector machines (SVMs). These ensemble
learner methods generate many classifiers and aggregate their results. Random Forest
method creates multiple classifications and regression (CART) trees, each trained on
a bootstrap sample of the original training data and searches across a randomly select-
ed subset of input variables to determine the split. Each tree in random forest will cast
a vote for certain input, then the output of the classifier is determined by majority
voting of the trees. All the decision trees were calculated with the R package rpart
(Therneau et al. 2015). We evaluate and compare the performance of the recursive parti-
 5

tioning method, random forest and support vector machines on hydrological data sets
(tidal energetics, tidal water level, current, friction coefficient, bottom drag coeffi-
cient, water depth and along channel distance) of Sundarbans Estuary for different
vulnerability categories.

3 Results

The mechanism by means which energy is dissipated by tidal current in Jagaddal

River Estuary in Sundarbans is encapsulated here by the following parameters: up-
stream distance from Bay of Bengal, tidal water elevation and current, water depth,
friction coefficient and bottom drag coefficient. We simulate the depth-averaged hy-
drodynamic model equations (See, Section 3.1) for the model domain (i.e., Jagaddal
River Estuary) using observed data and obtain the model solutions for the hydrologi-
cal processes. The tidal ebb and flood mass exchanges are directly represented by the
mean tidal current amplitude, U, which conveniently corresponds to the (predomi-
nant) semi-diurnal lunar tidal constituent, M2 in Jagaddal River Estuary. Net tidal
energy dissipation is proportional to U3, and net upstream dissipation, DISS are di-
rectly calculated using the depth-averaged hydrodynamic model solutions described
in Section 3.2.
The net upstream dissipation, DISS takes five peaks (Figure. 4) along the estuarine
channel of River Jagaddal. The first peak occurs in upstream distance of 10 km (near
Indrapur). Second peak takes place around 13 km (near Dhanchi Canal). The location
of third peak is nearby 15 km (at the confluence of River Jagadaal and east-west
channels on adjacent side. The fourth peak occurs towards Ramganga (around 17 km
upstream) at the confluence of another east-west channel. Finally in Ramganga
(around 21 km upstream), at the confluence of Rivers Bar Chara, Kal Chara and Pakhi
Nala, the last peak in net upstream dissipation, DISS is observed.
The variables used in the classification trees were mainly explaining estuarine hydro-
dynamics of tidal regime of Sundarbans. The six most important variables defining
vulnerability criterion based on tidal energetics are: upstream distance from Bay of
Bengal, tidal water level and current, water depth, bottom drag coefficient, and fric-
tion coefficient. From the tree in Figure 5, we find that the primary control on the net
upstream dissipation, DISS is bottom drag coefficient ('st') followed by the upstream
distance ('d').
The six terminal nodes classify the net upstream dissipation, DISS in six classes with
varying hydrological conditions (Figure. 6). If we compare performance of the classifi-
cation algorithm, random forest and support vector machines ( Table. 1), we find that
high to severe categories of vulnerability (A, AA, AAA) are the most challenging
aspect.
6

Table 1. Accuracy of class detection. Gini index provides initial classification of

vulnerability based on the detection accuracy of the net upstream dissipation, DISS.
To achieve more precision in predictives, random forest and support vector machines
have been used.
A AA AAA B BB BBB
Gini Index 0.9068 0.83333 NA 0.9500 0.9815 0.9783
Random 0.64286 0.64815 NA 0.9500 0.5 0.8478
Forest
Support 0.7143 0.64815 NA 0.9500 0.9630 0.8199
Vector
Machines
(SVMs)

The classification tree and supervised learning algorithms rank ( Figure. 7), most of the
regions along the estuarine channel of River Jagaddal from low to high vulnerability
zones depending on tidal energetics. Bottom drag coefficient acts as the most influ-
encing parameter in determining distribution of tidal energetics.
If we interpret the outcome of classification trees in the light of estuarine hydrody-
namics, the results can be best understood. The net upstream dissipation, DISS peaks
share the occurrence of water depth range 3.6-4.5 m along the river ( Figure. 8). Tidal
current in the range of 0.7 to 1 m/s contributes the most impact on upstream dissipa-
tion ( Figure. 9). Tidal water elevation around 3.5 to 5.5 m has major influence on tidal
energy dissipation ( Figure. 10). It is clear that not only high tide but also the ebb-tide
exchange phase is susceptible to high energy dissipation. The role of bottom drag
coefficient is the most interesting in the tidal phenomena along this mangrove estuary
( Figure. 11). In Jagaddal River Estuary, mangrove-swamp system in river-bed acts as
water reservoir as well as enhances the bed friction. As a result tidal water elevation
and current are modified significantly. The river regime adjacent to the Dhanchi Re-
serve Forest bottom drag force owing to mangrove vegetation lowers the tidal eleva-
tion. The accumulated water mass, thus, is dissipated towards Indrapur (southward)
and Ramganga (northward) and causes increased accumulation of net tidal energy in
these regions causes massive destruction, as observed in case of super-cyclone Am-
phan, this year.

4 Conclusions

Estuaries in Sundarbans with remarkable geometry of interconnected-east west chan-

nels, enhanced bed stress and bottom friction due to mangrove-swamp system, shal-
low water regime and asymmetry in ebb-tide transitions make the tidal energy dissi-
pation a complex process. Tidal energetics of estuaries of Sundarbans, apart from
being poorly understood, has a significant influence on the life and marine environ-
ment of this tidal country. The study area, Jagaddal River Estuary, an offshoot of
Saptamukhi river in Sundarbans, has witnessed wrath of tidal destruction during super
cyclone Amphan. In this study we attempt to identify estuarine conditions influencing
 7

the dynamic phenomena of tidal energy dissipation. To get a complete picture of vul-
nerability of a coastal region, we have to take into account salinity intrusion and eco-
logical health of coastal water to pinpoint the risk zones (Goutam, K.S. et al. (2015);
T.Das et al. (2018)). Finally a composite coastal vulnerability index formulated by
using tidal energetics, inundation factor, salinity intrusion and ecological health will
provide a comprehensive tool for decision making.
The beauty of the methodology used in the work is that we have used minimum
controlling factors and incorporated them in simple hydrodynamic model to generated
simulations. There is a way out to link the hydrodynamic model to the ecological
model. This framework can be developed into a strong decision-making tool to identi-
fy the vulnerable coastal zones and prioritize the resilience plan.

References
1. Chatterjee Meenakshi, Shankar D, Sen G K, Sanyal P, Sundar D, Michael G S, Chatterjee
Abhisek, Amol P, Mukherjee Debabrata, Suprit K, Mukherjee A, Vijith V, Chatterjee Sid-
dhartha, Basu Anwesha, Das Madhumita, Chakraborti Saranya, Kalla Aravind, Misra Surja
Kanta, Mukhopadhyay Soumya , Mandal Gopal and Sarkar Kankan, Tidal variations in
Sundarbans Estuarine System, India , J. Earth Syst. Sci. 122, No. 4, August 2013, pp. 899–
933.
2. De‟ath G., Fabricius K.E. (2000). Classification and regression trees: a powerful yet sim-
ple technique for ecological data analysis. Ecology 81(11): 3178–3192.
https://doi.org/10.1890/0012-9658(2000)081[3178:CARTAP]2.0.CO;2.
3. D. Prandle, A. Lane / Estuarine, Coastal and Shelf Science 160 (2015) 60e68. Sensitivity
of estuaries to sea level rise: Vulnerability indices.

4. Gönnert G, Dube SK, Murty, Siefert T (hrsg) (2001) Global storm surges. Theory, obser-
vations and applications. Boyens& Co, Heide. Die Küste 63, 623 p. ISBN 3-8042-1054-6.

5. Goutam, K.S., Tanaya, D., Anwesha, S., Sharanya, C., Meenakshi, C., 2015. Tide and
Mixing Characteristics in Sundarbans Estuarine River System, Hydrol. Curr. Res. 6, 204.
6. Huntley, D. A., R. J. Nicholls, C. Liu, and K. R. Dyer (1994), Measurements of the semi-
diurnal drag coefficient over sand waves, Cont. Shelf Res., 14, 437–456.
7. Li, C., A. Valle-Levinson (1999), A two-dimensional analytic tidal model for a narrow es-
tuary of arbitrary lateral depth variation: The intratidal motion, J. Geophys. Res.,104,
23,525-23,543.
8. Li, C., Valle-Levinson Arnoldo, Atkinson Larry P., Wong Kuo Chuin and Lwiza Kamazi-
ma M.M., J. Geophys. Res., Vol. 109, C03034, 2004.
9. Parker, B.B. (1984), frictional effects on the tidal dynamics of shallow estuary, Ph.D dis-
sertation, 291 pp., Johns Hopkins Univ., Baltimore Md.
10. Proudman, J., Dynamical Oceanography, 409 pp., New York, Methuen, 1953.
11. Selkimäki M., González-Olabarria J.R., Pukkala T. (2011). Site and stand characteristics
related to surface erosion occurrence in forests of Catalonia (Spain). European Journal of
Forest Research 131(3): 727–738. https://doi.org/10.1007/s10342-011-0545-x.
8

12. Socio Economic and Caste Census 2011.

https://secc.gov.in/statewiseHousingDwellingWallTypeReport?reportType=Housing%20D
welling%20Wall%20Type.
13. Taylor, G.I., 1920. Tidal friction in the Irish Sea. Philos. Trans. R. Soc. Lond. A220,
239e264. http://dx.doi.org/10.1098/rsta.1920.0001.
14. Therneau T.M., Atkinson E.J., Mayo Foundation. (2015). An introduction to recursive par-
titioning using the RPART routines. Technical report. Mayo Foundation.
15. Ulman, D., and R.E.Wilson (1998), Model parameter estimation from data assimilation
modeling: Temporal and spatial variability of the bottom drag coefficient, J. Geophys.
Res., 103, 5531-5549.
16. Wikipedia 2020 Cyclone Amphan; https://en.wikipedia.org/wiki/Cyclone_Amphan.
17. T.Das , Chakraborti Saranya, Mukherjee Joydeep, Sen Goutam Kumar, Mathematical
modelling for phytoplankton distribution in Sundarbans Estuarine System, India, Ecologi-
cal Modelling 368 (2018) 111–120.

Figure Captions

Figure 1. Jagaddal River Estuary, the study area. Three observation stations (from
south to the north) on the river Jagaddal are: Indrapur (21.651oN, 88.413oE), Dhanchi
(21.704oN, 88.433oE) and Ramganga (21.796oN, 88.37oE). Abbreviatons are as fol-
lows: R=River.

Figure 2. Width (top) and tidal water depth profile (bottom) for River Jagaddal. As
we move from the head (i.e.,Ramganga; 21 km. approx..) the width of the river in-
creased towards mouth (i.e., Indrapur; 8 km. approx.). From Indrapur towards Dhan-
chi, at the junction of Jagaddal and Dhanchi canal (around 14 km), the depth has de-
 9

creased suddenly. Here the width of the channel has increased due to linking channel
with two rivers (Saptamukhi, in the west) and Thakuran, in the east connected through
Dhanchi canal. The semidiurnal tidal amplitude generally increased from mouth to
head due to “funneling effect”. But we get minimum tidal water level at Dhanchi.
Frictional dissipation contributed plays vital role for such anomaly. Tidal swamp and
mud flats in the region of Dhanchi Reserve Forest increases the frictional effects
which in turn decreases tidal amplitude amplification. Near Ramganga, width and
depth both increased. Here the effect of river discharge from the confluences of rivers
in Jagaddal increases the tidal water level.

Figure 3. Observed water level variations (in between 4:00 a.m. to 3:00 a.m) in River
Jagaddal. The water level at Indrapur is considered representative of the water level at
mouth of all estuaries. Tidal range increased from mouth (Indrapur) to head (Ram-
ganga), but it decreased at Dhanchi.
10

Figure 4. Distribution of tidal energetics along Jagaddal River Estuary

 11

Figure 5. Classification tree model predicting net upstream dissipation, DISS occur-
rence. Range of DISS in along Jagaddal River Estuary and percentage of occurrence
are encapsulated in number of nodes based on controlling parameters. Here, the vari-
ables upstream distance and bottom drag coefficient are represented as „d‟ and „st‟,
respectively, for convenience.

Figure 6. Class probabilities associated with classes (A, AA, AAA, B, BB, BBB) are
represented at each node of the tree. Classes A, AA, B, BB and BBB represent classes
for increased probabilities of net upstream dissipation, DISS and are assumed to be
vulnerable. Here, the net upstream dissipation, DISS and bottom drag coefficient are
represented as „bn‟ and „st‟, respectively, for convenience.
12

Figure 7. Detection prevalence rate in percentage. Gini index (top-left), random for-
est (top-right) and support vector machines predicts extremely low (BBB) to medium
vulnerability classes the most prevalent. Classes of higher vulnerability categories (A
and AA) are detected next. The most severe category (AAA) remains undetected.

Figure 8. Variation of DISS with water depth along Jagaddal River Estuary
 13

Figure 9. Variation of DISS with tidal current along Jagaddal River Estuary

Figure 10. Variation of DISS with tidal water elevation along Jagaddal River Estuary
14

Figure 11. Variation of DISS with bottom drag coefficient along Jagaddal River Es-
tuary
The entire stretch of Jagaddal River Estuary is categorized based on tidal energetics and associated
Hydrological parameters into low risk (green), medium risk (orange) and high risk (red) zones.
This exercise can be extended in case of salinity intrusion and ecological health.
The composite index computed by taken into account of these hydrodynamical and ecological indices
can give a complete picture of vulnerability of any coastal area.

No. of Rank Distance from Current depth Drag force Water DISS
Zones sea (Km) level (KW)
1 0.559% 8-10.9 0.96074 3.9140 0.00037023 3.9357 19.206-
50.381
2 0.391% 10.9-11 0.98729 3.9697 0.00067947 3.8147 18.480-
19.203
3 1591.1% 11-11.033 0.90968 3.9554 0.00031841 3.9981 14.039-
87.757

4 134.5% 11.033-12.7 0.84317 3.6443 0.0056911 3.0538 0.76583-

17.942

5 233.3% 12.7-12.9 0.84793 3.9078 0.00033961 3.8374 13.518-

76.611

6 68.2% 12.9-14.7 0.84918 4.0699 0.0018739 4.3921 5.2225-

22.737

7 62.8% 14.7-15.1 0.81110 3.9698 0.00021987 4.1717 22.786-

80.074

8 14.9% 15.1-16 0.77417 3.8358 0.00062271 3.7979 12.671-

29.667
9 37.3% 16-16.4 0.74161 3.7400 0.00021560 3.5275 18.003-
58.302

10 66.5% 16.4-17 0.72512 3.7210 0.0013227 3.4773 3.5962-

17.953

11 0.41% 17-19 0.71928 3.8593 0.0041898 3.9021 1.9697-

3.5937

12 42.2% 19-20.1 0.72469 4.0711 0.0025140 4.5104 2.7076-

13.008
13 39.8% 20.1-20.9 0.69907 4.0904 0.00033986 4.4863 13.026-
54.559

14 511.6% 20.9-22.49 0.62681 3.9989 0.0058715 4.0440 0.30542-

25.154
 : Current R
gy e

Hydrology Goutam et al., Hydrol Current Res 2015, 6:2

olo

se
Hydr

arch
http://dx.doi.org/10.4172/2157-7587.1000204

ISSN: 2157-7587
Current Research

Research Article Open Access

Tide and Mixing Characteristics in Sundarbans Estuarine River System

Goutam KS1*, Tanaya D1, Anwesha S2, Sharanya C1 and Meenakshi C3
1
School of Oceanographic Studies, Jadavpur University, Kolkata 700032, West Bengal, India
2
Department of Mathematics, Jadavpur University, Kolkata 700032, West Bengal, India
3
Basanti Devi College, Kolkata 700029, West Bengal, India

Abstract
Sundarbans Estuarine System (SES, 21.25°-22.5° N and 88.25°-89.5° E), comprising the southernmost part of
the Indian portion of the Ganga-Bramhaputra delta bordering the Bay of Bengal, is India’s largest monsoonal, macro-
tidal delta front system. Sundarbans Estuarine Programme (SEP), the first comprehensive observational programme
to study tidal as well as salinity features was conducted during 18-21 March, 2011 (Equinoctial spring phase). The
main objective of this program was to monitor tides and salinity characteristics within the SES. Out of 30 observation
stations, spread over more than 3,600 sq km covering seven inner estuaries, we have chosen river Jagaddal, which
is connected with Saptamukhi East Gulley (SEG) in the West and river Thakuran in the East, due to the fact that the
station Indrapur situated on this river at location very close to Bay of Bengal represents the condition at the mouth
of all seven estuaries.
Tidal elevation, salinity, bathymetry and vertical profile of salinity using CTD were measured during the
observation period. Observed current data collected from different sources have been used for comparison with
computed tidal current. Finally, the estuarine current, bottom drag coefficient and gradient Richardson number
have also been computed. Computed values of these parameters have analyzed for interpreting variations for tidal,
current and mixing feature prevailing in the estuary.

Keywords: Tidal current; Estuarine circulation; Salinity; Mixing;
Sundarban; Estuary; Bottom drag coefficient
Introduction
Sundarbans Estuarine System (SES, the region lying between
21.25°-22.5° N and 88.25°-89.5° E) is India’s largest monsoonal, macro-
tidal delta-front estuarine system in the eastern coastal state of West
Bengal. River Hoogly, the first deltaic offshoot of the river Ganga forms
the western boundary of the SES. River Raimangal, a tributary of river
Ichhamati, an easterly distributary of the Ganga forms the eastern
boundary of the SES. Dense natural mangrove forests of Sundarbans
in the southern fringes of the Bay of Bengal forms the southern limit.
An imaginary line named Dampier-Hodges line based on 1829-1832
survey is the northern boundary of SES. North-south flowing rivers:
the Saptamukhi, Thakuran, Matla, Bidya, Gomdi, Gosaba, Gona,
Harinbhanga and Raimangal form the principal estuaries of SES
on the east of the Hoogly as a complex network of numerous west-
east interlinked channels, canals and creeks. In addition to the tidal
effects at their mouths, during floods these inner estuaries accumulate
considerable amount of Gangetic fresh water in their upstream along
with summer monsoon rainfall which are the major driving factors in
retraining the estuarine character of the SES. The rivers, water bodies,
intertidal mudflats, creeks, saline swamps, sandy shoals in SES are
remarkable not only for the astounding biodiversity including a wide
variety of benthic and pelagic fauna (such as fish and crustaceans),
spawning zone of reptiles (such as crocodiles), fisheries but also for
their important role as a natural filter to pollutants released from
the human settlements and industrial zone in northern reaches
and major pathways for nutrient recycling. Moreover impenetrable
mangrove forests acts like a protective barrier to the densely populated
metropolitan city of Kolkata to its north by absorbing direct impact of
cyclonic storms and surges from the Bay of Bengal to a great extent [1].
bends. Over this distance the characteristics of the SES has shaped in
these different environments (variable depth, width and stretches) by
the interplay of a variety of physical processes. This investigation is
a part of a series of intensive on-going study of the tidal elevations,
salinity and temperature at 30 locations situated on the principal inner
estuaries of the SES, conducted by Sundarbans Estuarine Programme
(SEP) during 18-21 March 2011(Pre-monsoon period), 2013 and one
month phase of study in August, 2015 (Monsoon period).
In the first observation programme (during 18-21 March 2011, Pre-
monsoon period, equinoctial spring phase) in SES it was reported that
estuaries have shown a great diversity of tidal forcing characteristics
with different depth, width, size and shape. In general tidal range has
increased from mouth to head (northwards) in all estuaries due to
convergent channel geometry (funnel-shaped) and frictional effects
with the variable degree and rates of amplification over the various
estuarine stretches as a result of flood-dominant and ebb-dominant
tidal asymmetries [1]. In such estuarine system tidal currents, salinity
variation and river discharge predominates in estuarine flow pattern
more than meteorological forcing (wind stress). So, the study of
tidal flows and salinity structure are essential to assess a broad range
of estuarine phenomena including dispersion of salinity, nutrients,
pollutants, sediments and their transportation into and out of the
estuary and to know how quickly the estuary responds to different
*Corresponding author: Goutam KS, School of Oceanographic Studies,
Jadavpur University, Kolkata 700032, West Bengal, India, Tel: 919831282057;
E-mail: gksju@yahoo.co.in
Received April 21, 2015; Accepted May 06, 2015; Published May 08, 2015
Citation: Goutam KS, Tanaya D, Anwesha S, Sharanya C, Meenakshi C (2015)
Tide and Mixing Characteristics in Sundarbans Estuarine River System. Hydrol
Current Res 6: 204. doi:10.4172/2157-7587.1000204

All the principal estuaries in the SES are funnel-shaped, widths Copyright: © 2015 Goutam KS, et al. This is an open-access article distributed
under the terms of the Creative Commons Attribution License, which permits
converge rapidly northwards, having very wide mouths at confluences unrestricted use, distribution, and reproduction in any medium, provided the
with the Bay of Bengal and following meandering courses with sharp original author and source are credited.

Hydrol Current Res

ISSN: 2157-7587 HYCR, an open access journal Volume 6 • Issue 2 • 1000204
Citation: Goutam KS, Tanaya D, Anwesha S, Sharanya C, Meenakshi C (2015) Tide and Mixing Characteristics in Sundarbans Estuarine River
System. Hydrol Current Res 6: 204. doi:10.4172/2157-7587.1000204
forcing such as tidal mixing and river discharge which affect the
ecological health and the water quality [1,2-8].
In this paper we are intended to study the flow pattern in the
narrow tidal channel of River Jagaddal, a part of SES. Our attempt is to
quantify the indicators such as tidal amplitude, tidal current, residual
current (estuarine circulation), bottom drag coefficient and level of
stratification which are essential to characterize the flow pattern in
different channel geometry along a single stretch of a river.
Study Area and Relevant Measurements
River Jagaddal is flowing parallel to the Saptamukhi East Gullay
(SEG) in the west, the stronger branch of the river Saptamukhi
debouncing into the Bay of Bengal having links with the adjoining
estuary of River Thakuran in the east through Dhanchi Khaal (cannel).
In the northern limit (head) river Jagaddal has linked with river
Barchara, river Kalchara and Pakhi Nala (Figure 1). River Barchara and
Figure 1: Study Area: Jagaddal River Estuary. Triangles show the observation stations along the river; Abbreviation R. is used to refer “River”.
Hydrol Current Res
ISSN: 2157-7587 HYCR, an open access journal
Page 2 of 7
river Kalchara are influenced by fresh water discharge of Hoogly river
whereas Pakhi Nala is connected with water flow of river Thakuran.
In this study the data required for tidal elevation, salinity are
collected from the field observation of four stations on Jagaddal:
Indrapur (mouth), Dhanchi and Ramganga/Pakhirala (head).Observed
current data collected from different sources have been used. These
stations are chosen for their significant importance along the channel.
The station Indrapur (S1) on river Jagaddal is closest (about 8 Km north
of Bay of Bengal) to the Bay of Bengal, representing the conditions at
the mouth. The station Dhanchi (S2) is at the junction of river Jagaddal
and river Thakuran through Dhanchi Cannel. The station Ramganga
(S6) is at the junction of three different linkages of river Barchara, river
Kalchara (connected to Hoogly River) and Pakhi Nala (connected to
Thakuran). The station Pakhirala (S7) is in the bank of Pakhi Nala,
serving as the connecting junction of river Jagaddal to river Thakuran.
Both Ramganga and Pakhirala serve as the head of this channel.
Volume 6 • Issue 2 • 1000204
Citation: Goutam KS, Tanaya D, Anwesha S, Sharanya C, Meenakshi C (2015) Tide and Mixing Characteristics in Sundarbans Estuarine River
System. Hydrol Current Res 6: 204. doi:10.4172/2157-7587.1000204

Page 3 of 7

Methodology equations [Li and Levinson] [9] are numerically solved to obtain the
tidal amplitude and tidal current:-
The methodology follows five steps:-
∂u ∂u ∂u ∂ς u β
(i) The first step solves a numerical model with a given spatial +u + v − fv =
−g − β + 2 uς
distribution of friction coefficient in a channel with variable along ∂t ∂x ∂y ∂x h h
(-x) and cross (-y) channel depth variation and along channel width
variation to compute tidal amplitudes and velocity components. ∂v ∂v ∂v ∂ς v β
+ u + v + fu =
−g − β + 2 vς (2)
∂t ∂x ∂y ∂y h h
(ii) The second step uses the velocity field computed in the previous
step and applies the phase-matching method to recover the drag
coefficient. ∂ς ∂ (h + ς )u ∂ (h + ς )v
+ + 0
=
(iii) Then by using the drag coefficient estimated from the ∂t ∂x ∂y
phase-matching method as a function of along-channel position in Where u:- longitudinal velocity, v: lateral velocity, ς : elevation,
the numerical model, we have estimated the tidal current and tidal h: water depth, x: longitudinal coordinate, y: lateral coordinate, t:
amplitude and compare with the observed amplitude. Step (i) to time, β : friction coefficient, f: coriolis parameter, g: gravitational
(iii) are repeated if the computed tidal amplitude are not close to the acceleration.
observed data.
The friction coefficient β is defined by Proudman and Parker
(iv) Using the values of bottom drag coefficient at different depth [12,13]:-
the magnitude of residual current (estuarine circulation) is calculated.
8CDU O
β= (3)
(v) At last, the level of stratification along the channel is 3π
parameterized as the gradient Richardson number using the values of where CD and U0 are bottom drag coefficient and magnitude of
tidal current and residual current or estuarine circulation. longitudinal velocity respectively. The friction coefficient â , a function
of x and y, is dependent on the velocity amplitude and the bottom drag
First step: Computing depth functions
coefficient. Spatial distribution of CD for every segment of the channel
We have assumed bathymetric functions for computational and also the interpolated longitudinal velocity U0 along the channel
convenience, by fitting bathymetry data along cross-channel directions. from observed current data are used to provide initial values for CD and
Several such bathymetry profiles have been used for different segment then β to solve equation (2) numerically.
of the river to examine the effect of complicated bathymetry. The depth
function is assumed to vary in both along-(x) and cross-channel (y) Recovering bottom drag coefficient
directions as needed having the following form: In depth-averaged hydrodynamic model, the intensity of the
overall tidal energy dissipation is related to mainly bottom drag
  D 
2

 y−  
2 
h= b0 + b1 * exp  −  (1)
 b2  significant spatial variation in bottom drag coefficient with an effect on
 
 
where D is the transverse coordinate, the estuary width, and b0, b1,
b2 are the parameters that adjust the depth profile and vary with x. The
width D is allowed to vary from 0 (at mouth) to maximum value for
each of the transect starting from mouth to head of the channel.
Second step: Solving numerical model
We have used the depth-averaged hydrodynamic model for
shallow tidal estuaries (maximum water depth is on the order of a few
tens of meters) to compute tidal current and amplitude as described
by Li et al. [9,10] to study the variation of intratidal flow field in James
river estuary. This method is a modification of the adjoint variational
method, the full-scale assimilative numerical models to compute
tidal current and amplitude with variable bottom drag coefficient in
Chesapeake Bay by Ulman and Wilson et al., Spitz and Klinck et al.
[6,11].
We have assumed the estuarine geometry as a curved channel with
variable width and an arbitrary lateral depth variation. By following
approach similar to Li et al. [7], due to anisotropy of width and depth
variation, we have to opt for numerical approach. The x axis is taken to
lie along the river boundary and points toward the head of the estuary.
The y axis lies along the open boundary at x=0.
The depth-averaged, shallow water momentum and continuity
coefficient for the major tidal frequency component [9,10]. Again we
have assumed that the bathymetry effect alone will be able to produce
tidal current and amplitude in a tidal channel. Hence we have chosen
arbitrary along and across channel depth variations with the emphasis
on the effect of channel-shoal configuration and bottom slopes.
Using the computed results for tidal current and amplitude from
numerical simulation of depth-averaged, shallow water momentum
and continuity equations (2), spatial variation of CD in tidal channels
with significant lateral variation of depth is estimated by applying the
phase-matching method [10]. The calculation involves the following
steps:
(a) The amplitude and phase of the computed tidal current ( ϕU )
are obtained by harmonic decomposition of horizontal velocity.
(b) The phase of the tidal current is then fitted to a quadratic
function of the water depth to obtain the phase of the pressure gradient
ϕ Ax .
(c) From the phase relationship of the longitudinal momentum
equation, the drag coefficient is calculated as a function of the transverse
position using the following formula:
3πσ h
CD = (4)
8U o tan ϕ
( Ax−ϕU )
Where σ is the angular frequency of the semi-diurnal tide (M2 tide
for this study), h is the undisturbed water depth, U0 is the amplitude

Hydrol Current Res

ISSN: 2157-7587 HYCR, an open access journal Volume 6 • Issue 2 • 1000204
Citation: Goutam KS, Tanaya D, Anwesha S, Sharanya C, Meenakshi C (2015) Tide and Mixing Characteristics in Sundarbans Estuarine River
System. Hydrol Current Res 6: 204. doi:10.4172/2157-7587.1000204

Page 4 of 7

of the tidal current, ϕU is the phase of the computed tidal current and
ϕ AX is the phase of the pressure gradient.
Estimation of residual or estuarine circulation
Pritchard [14] first introduced the concept of “residual” or
estuarine circulation driven by horizontal salinity gradient, a key
dynamical variable that makes estuaries different from any other
marine environment and plays a key role in maintaining salinity
stratification in estuary which inhibits vertical mixing and leads to
one way of classifying estuaries based on strength of stratification:
well mixed, partially mixed, highly stratified and salt edge. He pointed
out that if the vertically varying horizontal currents are measured
through the course of the tidal cycle, and then averaged, the estuarine
circulation would be revealed, though the tidal currents are typically
much stronger the estuarine circulation.
Using tidal current U T we have calculated the magnitude of
estuarine circulation, U e expressed as (Geyer et al.,) [15]:
∂s
β s h2 g
U e = a0 ∂x (5)
CDU T
Where a0is a dimensionless constant related to turbulent
momentum flux, β s is the coefficient of saline contraction, g is the
acceleration of gravity, h is the water depth, UT is the tidal velocity.
Estimation of level of mixing
Numerous studies have undertaken to analyze importance
of salinity gradient, stratification and tidal currents in estuarine
circulation. Both river discharge and mixing are crucial for the different
tidal flushing process and tide-induced transport of materials in flood
and ebb conditions. During the flood tide when the more saline sea
water is coming into estuary with a faster advancement at the surface,
the heavier salt water on the surface promotes vertical mixing to
reach a more stable state. During ebb tide when the fresher less saline
estuarine water is going out of the estuary with a faster advancement at
the surface on the heavier more saline water at bottom, a more stable
water column is formed suppressing vertical mixing. The sensitive
dependence of stratification on tide also has been revealed in a number
of studies of the spring-neap variation of tidal cycle with stratification
with maximal tidal mixing and minimum stratification during spring
tides [7,15].
Geyer et al. [15] obtained the top-to-bottom salinity difference, ∆s
∂s
is estimated from , the horizontal salinity gradient obtained from
∂x
observed salinity data in Hudson estuary,
∂s
hU e
∆s =a1 x
∂ (6)
CDU T
Where a1 is a constant that depends on the shape of salinity and
∂s
velocity profile, is the horizontal salinity gradient and CD is the
∂x
bottom drag coefficient.
To quantify the importance of stratification on mixing, the
gradient Richardson number RiT , is estimated which is defined as the
ratio between vertical salinity difference and tidal velocity,
β S g ∆sh
RiT (7)
UT 2
On the basis of the estimated value of gradient Richardson number
we have differentiated the level of mixing along the channel.
Results and Discussions
During observation period of 18-21 March 2011 (Pre-monsoon
period) an increased trend in tidal range from mouth (Indrapur) to head
(Ramganga, Pakhirala) was observed along Jagaddal. But interestingly,
a decrease in tidal range was observed between Indrapur and Dhanchi
(Figure 2: near the junction of Jagaddal and Dhanchi Canal (Chatterjee
et al., [1]. Some striking features regarding salinity variation was also
observed during observation where the salinity decreases from mouth
to head upto Ramganga, but it increases from Ramganga to Pakhirala
(Figure 3).
From Indrapur towards Dhanchi, at the junction of Jagaddal and
Dhanchi canal, the depth has decreased suddenly (Figure 4). In this
shallow region, the stronger tidal current has observed. Here the width
of the channel has increased due to linking channel with two rivers

Indrapur
5 Dhanchi
Ramganga
Pakhirala
4
Water
Level
in m. 3

0
0 10 20 30 40 50 60 70 80
Time in Hours
Figure 2: Observed water level variation in four observation stations.

Hydrol Current Res

ISSN: 2157-7587 HYCR, an open access journal Volume 6 • Issue 2 • 1000204
Citation: Goutam KS, Tanaya D, Anwesha S, Sharanya C, Meenakshi C (2015) Tide and Mixing Characteristics in Sundarbans Estuarine River
System. Hydrol Current Res 6: 204. doi:10.4172/2157-7587.1000204

Page 5 of 7

28.6
Indrapur
28.4 Dhanchi
Ramganga
28.2
Pakhirala

28
Salinity
in 27.8
psu
27.6

27.4

27.2

27
26.8
0 10 20 30 40 50 60 70
Time in Hours
Figure 3: Observed salinity variation in four observation stations.

6.9

6.8

6.7

6.6
Depth (m)

6.5

6.4

6.3

6.2
5 10 15 20 25 30

Along Channel Distance in Km

Figure 4: Depth distribution along the channel.
Drag Coefficient

Dhanchi

Indrapur (Mouth)

Along channel distance in km

Figure 5: Drag coefficient variation along the channel.

Hydrol Current Res

ISSN: 2157-7587 HYCR, an open access journal Volume 6 • Issue 2 • 1000204
Citation: Goutam KS, Tanaya D, Anwesha S, Sharanya C, Meenakshi C (2015) Tide and Mixing Characteristics in Sundarbans Estuarine River
System. Hydrol Current Res 6: 204. doi:10.4172/2157-7587.1000204

Page 6 of 7

(Saptamukhi, in the west) and Thakuran, in the east connected through
Dhanchi canal. In addition, the bottom drag coefficient (Figure 5)
has increased suddenly after this junction. With the combined effect
of increased CD and river width, the water level (Figure 2) and tidal
current (Figure 6) has dropped after this point.
Near the west bank of Jagaddal, the cross channel current is stronger
and generates stronger residual circulation (Figure 7). The estuarine
current has started to increase from Dhanchi to Indrapur. Vertical
salinity gradient has increased due to landward advection of denser and
more saline seawater at bottom and seaward advection of fresh riverine
water on the top layer. As a result, an increase in gradient Richardson
number (>0.25, Figure 8) from Dhanchi to Indrapur (towards the sea)
has found which indicates that the stratification has already started.
Along the channel from Dhanchi to Ramganga, there are three
adjoining rivers: Kalchara, Barchara and Pakhi Nala which draw large
amount of less saline water from Hoogly. Depth has increasing trend
here and tidal current tends to take a sharp rise. From CTD cast it is
observed that vertical gradient of salinity over this stretch of the river
is negligible. The gradient Richardson number at this region initiates
strong mixing by making stratification negligible (the channel in this
area acts as a well-mixed estuary) with an increased bottom drag
coefficient along this portion of the river. The increased river inflow
(through linked rivers or channels) increases the tidal current and
amplitude suppressing the role of drag coefficient (opposite to the
observations in Dhanchi). Similar type of phenomenon was reported
by Li et al. in James river estuary where increased river discharge has
limited tidal mixing, lowering the bottom drag coefficient. For this
reason, decrease in tidal current and amplitude due to large value of
bottom drag coefficient have not observed here in contradiction with
Dhanchi. A small peak in estuarine circulation is found due to this
limited effect of drag coefficient.
Conclusion
The main intension of this study is to interpret the cause about the

0.78

0.76

0.74
Tidal Current in m/sec

0.72

0.7

0.68

0.66
4 6 8 10 12 14 16 18 20 22

Along Channel Distance in (km)

Figure 6: Tidal current variation along the channel.

0.1

0.08
Estuarine circulation in m/sec

0.06

0.04

0.02

-0.02
5 10 15 20 25 30
Along Channel Distance in km
Figure 7: Estuarine current variation along the channel.

Hydrol Current Res

ISSN: 2157-7587 HYCR, an open access journal Volume 6 • Issue 2 • 1000204
Citation: Goutam KS, Tanaya D, Anwesha S, Sharanya C, Meenakshi C (2015) Tide and Mixing Characteristics in Sundarbans Estuarine River
System. Hydrol Current Res 6: 204. doi:10.4172/2157-7587.1000204

Page 7 of 7

0.16

0.14

0.12
Gradient Richardson Number

0.1

0.08

0.06

0.04

0.02

0
5 10 15 20 25 30
Along Channel Distance in Km.

Figure 8: Gradient Richardson number variation along the channel.

variation of water elevation at several locations on Jagaddal. Cross-
channel flow (west-east) is a dominant feature in this region (though
rigorous measurement on this aspect is not available except some cross
channel current measurement near river junction). Variational features
of bottom drag coefficient and estuarine current, due to inflow through
two intervening channels near Dhanchi, one connected with Thakuran
and other connected with Saptamukhi, contribute to tidal amplitude
and current variation. River Jagaddal (as observed from, gradient
Richardson number variation) also shows stratification characteristics
from Indrapur to Dhanchi. But beyond Dhanchi upto the head (further
north) it exhibits a zone of strong mixing. This zone of strong mixing
3. Jay DA, Smith JD (1990) Residual circulation in shallow estuaries. 2. Weakly
stratified and partially mixed, narrow estuaries. J Geophys Res 95: 733-748.
4. Valle-Levinson A, Li C, Royer TC, Atkinson LP (1998) Continental Shelf
Research 18: 1157-1177.
5. Mac Cready P (1999) Estuarine adjustment to changes in river flow and tidal
mixing. J Phys Oceanogr 29: 708-726.
6. Spitz YH, Klinck JM (1998) Estimate of bottom and surface stress during a
spring-neap tide cycle by dynamical assimilation of tide gauge observations in
the Chesapeake Bay. J Geophys Res 103: 12761-12782.
7. Li C (2006) Modeling of Bathymetry Locked Residual Eddies in Well mixed
Tidal Channels with Arbitrary Depth Variations, J Phys Oceanogr 36: 1974-
1993.

is indicative of the feature of considerable fresh water flow from river
Hoogly. All the salient findings corroborated in above mentioned
statements have been observed with considerable elaboration in our
discussion part.
Acknowledgements
This work has been supported by the Sundarbans Estuarine Programme
(SEP) funded by Indian national Centre for Ocean Information Services (INCOIS),
Ministry of Earth Sciences, Government of India. The authors are grateful to Dr.
Meenakdhi Chatterjee, Project Coordinator of Sundarbans Estuarine Programme
(SEP) for appreciable assistance and for providing data and other necessary
information.
8. Li C, Swenson E, Weeks E, White JR (2009) Asymmetric tidal straining across
an inlet: Lateral invertsion and variability over a tidal cycle, Estuarine, Coastal
and Shelf Science 85: 651-660.
9. Li Chunyan, Valle-Levinson A (1999) A two-dimensional analytic tidal model
for a narrow estuary of arbitrary lateral depth variation: The intratidal motion. J
Geophys Res 104: 23525-23543.
10. Li C, Valle-LA, Atkinson, LP, Wong K, Lwiza K (2004) Estimation of drag
coefficient in James River estuary using tidal velocity data from a vessel-towed
ADCP, J Geophys Res 109: C3.
11. Ullman D, Wilson RE (1998) Model parameter estimation from data assimilation
modeling: Temporal and spatial variability of the bottom drag coefficient. J
Geophys Res 103: 5531-5549.

References
1. Chatterjee M, Shankar D, Sen GK, Sanyal P, Sundar D (2013) Tidal variation
in the Sundarbans Estuarine System, India. J Earth Syst Sci 122(4): 899-933.
2. Jay DA, Smith JD (1990) Residual circulation in shallow estuaries: 1. Highly
stratified, narrow estuaries J Geophys Res 95: 711-731.
12. Proudman J (1953) Dynamical Oceanography, Methuen, New York, p. 409.
13. Parker BB (1984) Frictional effects on the tidal dynamics of shallow estuary,
(Ph.D. dissertation) Johns Hopkins Univ., Baltimore, pp. 291.
14. Pritchard DW (1956) The dynamic structure of a coastal plain estuary. J Mar
Res 15: 33-42.
15. Geyer WR (2001) Estuarine salinity structure and stratification.

Hydrol Current Res

ISSN: 2157-7587 HYCR, an open access journal Volume 6 • Issue 2 • 1000204
 Ecological Modelling 368 (2018) 111–120

Contents lists available at ScienceDirect

Ecological Modelling
journal homepage: www.elsevier.com/locate/ecolmodel

Mathematical modelling for phytoplankton distribution in

Sundarbans Estuarine System, India
Tanaya Das a,b,∗ , Saranya Chakraborti a,b , Joydeep Mukherjee b , Goutam Kumar Sen a
a
School of Oceanographic Studies, Jadavpur University, Kolkata 700032, West Bengal, India
b
School of Environmental Studies, Jadavpur University, Kolkata 700032, West Bengal, India

a r t i c l e i n f o a b s t r a c t

Article history: Mathematical modeling is an important tool to study planktonic ecosystem dynamics and various mecha-
Received 2 August 2017 nisms involved in its interaction with hydrological regime. The difficulty arises when describing plankton
Received in revised form dynamics under wide range of estuarine environmental conditions due to inadequate knowledge to inter-
17 November 2017
pret the mechanism to which the environmental nature of a given situation force the model condition
Accepted 18 November 2017
towards the observation and the extent to which the details of the model implementations do. We present
a model of the phytoplankton dynamics and nitrogen cycling in the estuarine environment of Sundarbans.
Keywords:
We have tested the sensitivity of each model parameter in turn by running the model to a steady-state
Phytoplankton
Nutrient
within a certain range of value satisfying the trend of phytoplankton and nutrient distribution. In this
Estuary way, it is possible to determine which parameters had the most influence on which variables and the
Ecosystem model possible mechanism underlying the ecological processes in estuarine environment. The model exhibits
Sundarbans stable behavior for the state variables over the tidal cycle and follows the trend of phytoplankton and
nutrient distribution along the river channel. This exercise will obviously set a background to have some
knowledge of the tuning of the ecological model to the parameter values.
© 2017 Elsevier B.V. All rights reserved.

1. Introduction Richards, 2001; Oschlies and Garcon, 1988). Changes in temper-

Estuarine waters are one of the fluvial-dynamic ecological sys-
tem which undergo through a continuous changes due to the
physical processes of coastal water (e.g. diffusion, stratification)
and as well as due to the inherent biological processes (e.g. nutrient
uptake, physiological changes, growth rate) which are instrumental
in producing spatial variation of planktonic distribution (Bonachela
et al., 2011). Hence models of planktonic processes have gaining
much interest of researchers in order to understand the functions
of different part of estuarine ecosystem. Such works are needed
to assess the health of the marine life and environment condi-
tions of the estuaries in changing climate scenario and to locate
the potential marine economic zones.
Several researchers have attempted to explain the growth
of phytoplankton in different environmental conditions. Oceanic
plankton distribution is related to the physical features like eddies
and fronts (Biktashev and Brindley, 2004; Abraham, 1998; Flier
and Davis, 1993; McGillicuddy and Robinson, 1997; Martin and
∗ Corresponding author at: School of Environmental Studies, Jadavpur University,
Kolkata 700032, West Bengal, India.
E-mail address: tanaya16.das@gmail.com (T. Das).
ature and nutrient availability induced by seasonal changes such
as, variation in thermocline depth and strength are considered to
be responsible to the rapid and massive phytoplankton bloom in
rivers, lakes and oceans. The effect on inflows on phytoplankton
population is an interesting problem to study how the magnitude
and the timings of inflows influence the physical features and nutri-
ent availability by altering hydrology that affect phytoplankton
growth and productivity. Moreover, seasonal forcing also acts as the
controlling factor on phytoplankton population that lives in natu-
ral systems in a periodically varying environment and population
interactions. Since phytoplankton have the capacity to acclimate
to changes in environmental conditions by altering their chemi-
cal composition in response to environmental variability, physical
aspects of nutrient uptake and the change in plankton dynamics in
physical limitations in response to limited nutrient condition have
been the focus of intense study in changing hydrological regime.
Temperature, availability of light in addition to nutrients regulates
the phytoplankton photo-physiology (Steele, 1978; Bainbridge,
1957; Venrick, 1990; Mukhopadhyay and Bhattacharya, 2006; Jang
and Allen, 2015; Bonachela et al., 2011).
To date a large number of model formulations have made to
express the nutrient and plankton dynamics in a wide variety of
marine environment i.e., ocean, lake, estuary etc. Among them

https://doi.org/10.1016/j.ecolmodel.2017.11.020
0304-3800/© 2017 Elsevier B.V. All rights reserved.
112 T. Das et al. / Ecological Modelling 368 (2018) 111–120
Michaelis-Menten Kinetics was one of the remarkable pioneer
attempts to describe the influence of ambient nutrient concen-
tration (external and internal) on nitrogen uptake rate by algal
cells (Caperon, 1967; Dugdale, 1967; Droop, 1973; Burmaster,
1979). A number of models tended to concentrate on interaction
among nutrients, phytoplankton and zooplankton and the flows
of materials among these components of ecosystem. Effects of
light limitation, temperature, zooplankton grazing, role of bacte-
ria, protozoans and dissolved organic matter and toxic elements
on planktonic system are extensively explored (Pace et al., 1984;
Fasham, 1985; Moloney et al., 1986; Williams and Druffel, 1988).
But these models have not put much light on the variability of
plankton dynamics in seasonal cycle in the ambient of explicit phys-
ical processes as addressed by Evans and Parslow (1985) using a
simple model of plankton and nutrient dynamics in a seasonally
varying mixed layer. Here, the role of physical mixing on seasonal
recurrence of spring bloom formation is well explained. A simple
model of planktonic ecosystem was developed by Fasham et al.
(1990) (hence called FDM model) to describe the ecosystem of
Sargasso Sea, near Bermuda in which temporal changes in the ther-
mocline depth plays a major role in ecosystem. Later Haney and
Jackson (1996) modified this model with changes in algal growth
rate formulations to combine the effects light and nutrient limita-
tion.
Sundarbans Estuarine System (21.25◦ –22.5◦ N and 88.25◦ –89.5◦
E), situated in the eastern coastal state of West Bengal, is India’s
largest monsoonal, macro-tidal delta-front estuarine system. It
comprises the southernmost part of the Indian portion of the
Ganga-Brahmaputra delta bordering the Bay of Bengal. This estu-
ary has gained considerable impetus as an interesting study area
on phytoplankton dynamics. Twice-daily mixing of the estuarine
waters by the tidal phenomenon (Chatterjee et al., 2013) causes
changes in water level, salinity and nutrient and such wide range
of environmental conditions which have an important impact on
the characteristics of phytoplankton distribution in rivers of Sun-
darbans. River discharge, stratification of water column, grazing
pressure by zooplankton, nutrients and light availability are the
main controlling factors in growth of phytoplankton. A number
of comprehensive studies have conducted to diagnose the current
state of ecosystem in Hoogly-Matla estuarine system for exploring
the possible impact of future changes in environmental factors on
the ecosystem. The impact of mangroves on nutrients and plankton
dynamics, temperature dependent growth rate of phytoplankton in
saline environment and implication of body sizes plankton for total
system dynamics in optimization of exergy are critically analyzed
(Roy et al., 2016; Mandal et al., 2012; Ray et al., 2001; Joyita et al.,
2015)
On these developments we have taken a simple but realistic
phytoplankton-nutrient distribution model (Fasham et al., 1990;
Haney and Jackson, 1996) as a tool to determine the controlling
forcing parameters of plankton and nutrient dynamics in rivers of
Sundarbans estuary. The organization of the paper is as follows.
Section 2 gives a brief overview on physical and environmental con-
ditions of the study area. Section 3 provides the description of the
base model along with the parameters and significance. Parameter
sensitivity analysis and stability test are included in this section.
Section 4 analyzes the behavior of plankton and nutrient dynamics
with the help of sensitivity analysis, stability conditions and model
output. Section 5 deals with the salient features and significance of
the work in the concluding part.
2. Study area
The principal estuaries of the Sundarbans Estuarine System
(SES), India, are: the Saptamukhi, Thakuran, Matla, Bidya, Gomdi,
Gosaba, Gona, Harinbhanga and Raimangal. All these north-south
flowing estuaries are lying east of the Hoogly River. A wide range of
tidal variations in different environmental and estuarine conditions
of these estuaries are reported by Chatterjee et al. (2013). According
to this report and an extensive hydrological survey of the southern
stretch of the river Saptamukhi during 1961–1963 (Jha et al., 1999)
have noted that Saptamukhi East Gulley (SEG), the stronger branch
of Saptamukhi, carries a major portion of the inflow from the bay.
Tide induced daily mixing of the estuarine waters cause changes
in hydrological and physico-chemical characteristics of the estuary
which makes the eastern part of river Saptamukhi an ideal estuary
to study the dynamic interaction of estuarine hydrology with the
ecosystem of the estuary.
In this paper we have chosen the River Jagaddal, an easterly
branch of the SEG, flowing parallel to the SEG straight to the Bay of
Bengal as our study area to model the phytoplankton and nutri-
ent distribution along the river Jagaddal. Goutam et al. (2015)
discussed about the nature of nonlinearity (role of bottom drag
coefficient), mixing characteristics, variation of water level, tidal
current and estuarine current along the river Jagaddal based on the
observations in equinoctial spring phase (18–21 March) of 2011, on
the three stations on this river: Indrapur, Dhanchi and Ramganga
and numerical simulation of hydrological parameters (water level,
tidal current, estuarine current, bottom drag coefficient, gradient
Richardson number etc.). Since an initial overview on tidal hydro-
dynamics of flow regime is available in this river, so it is a straight
approach to undertake a comprehensive study on variation of eco-
logical parameters (phytoplankton, nutrients etc.) in the influence
of tidal action on this river.
A number of studies have conducted in different seasons to
understand the variation of physico-chemical parameters (salinity,
nutrients, phytoplankton) in the eastern part of river Saptamukhi,
as well as on the principal estuaries of Sundarbans (Chaudhuri
et al., 2012; Mitra et al., 2014; Mitra et al., 2009; Raha et al., 2012).
To model the variation of phytoplankton and nutrient concentra-
tion along the river Jaggaddal, we have considered these available
references to get an idea about the nature of phytoplankton and
nutrient variation on this river in monsoon and identify the con-
trolling parameters to understand the interaction of hydrological
and ecological processes.
3. Methodology
3.1. Model structure and equations
To model the phytoplankton and nutrient dynamics along the
estuarine channel of river Jagaddal (Fig. 1), coupling of physics and
biology is essential for simulation of phytoplankton and nutrient
distribution. Evans and Parslow (1985) explored a 1D (vertical)
model in which the plankton were affected by shoaling and deepen-
ing of the mixed layer where the effect of shallowing or deepening
of mixed layer on the volumetric concentration of a state-variable
in mixed layer remains unchanged. In this study we have taken this
model to induce the hydrodynamics of estuarine river in a simpli-
fied form in phytoplankton and nutrient dynamics. For a mixed
layer of depth M, the rate of change of mixed layer depth is
dM
h+ (t) = (1)
dt
Now the mixing process rate is fast compared to the growth rate
of the organisms (Robinson et al., 1979). Following these features
we make the basic assumption that in the shallow estuarine chan-
nel of Jagaddal river, the vertical water column can be considered
biologically homogeneous. With mixing rate m, the rate of change
of mixed layer provides the physically driven rate to change the
 T. Das et al. / Ecological Modelling 368 (2018) 111–120 113

Fig. 1. The southwestern part of Sundarbans Estuarine System. From west to east rivers are: Hoogly River, Saptamukhi River and River Thakuran. All these rivers are drained
into the Bay of Bengal to the south. Bordered in red is the study area (outlined in green box), i.e., River Jagaddal, an eastern branch of River Saptamukhi. Major observation
stations, from south to north, along the River Jagaddal are denoted by the triangles in black: (i) Indrapur (21.651◦ N, 88.413◦ E; 8 km north from Bay of Bengal, serves as
the mouth of Jagaddal River Estuary) (ii) Dhanchi (21.704◦ N, 88.433◦ E; at the junction of River Jagaddal and Dhanchi Canal) and (iii) Ramganga (21.796◦ N, 88.37◦ E). The
R. Jagaddal is interconnected with a number of west-east flowing channels. The junctions of the river with these channels in the study area are denoted by circles in blue
and the name of the connecting rivers are presented in the parentheses: (i) J1 (R. Barchara) (ii) J2 (R. Namkhana, that is connected to River Hoogly to the west) (iii) J3 (River
Saptamukhi west) (iv) J4 (Anonymous River channel) and J5 (Dhanchi Canal, that is connected to River Thakuran to the east). The entire study area is of approximately 21 km
in length (Bay of Bengal to the north). The map is modified form Google Map. The places not relevant to the study are not denoted and pointed out to the Abbreviation R. is
used to refer River and E is used to refer East. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

concentration of phytoplankton and nutrient in the water column The equation for phytoplankton is described by Haney and
by the term Jackson (1996),
 
  dP m + h+ (t)
m + h+ (t) = (1 − 1 ) ␴P − G1 − 1 P − P (2)
. (1A) dt M
M
where ␴ is the average daily phytoplankton specific growth rate,
This set of equations (Eqs. (1) and (1A) allow nutrients to be G1 represents the rate of phytoplankton loss to grazing, ␮1 is the
injected into the mixed layer during entrainment (deepening), phytoplankton specific mortality rate, ␥1 is the fraction of phyto-
while the phytoplankton concentration becomes diffused through- plankton growth excreted as DON, m is a mixing velocity, M and
out the mixed layer M. During stratification, the phytoplankton h+ (t) has been defined above.
concentration below the new mixed layer M is lost. The nitrate equation is described as Haney and Jackson (1996),
We have taken a simple 2-state variable model following the  
dNn m + h+ (t)
compartmental ecosystem model developed by Fasham et al. = −n P + (N0 − Nn ) (3)
(1990). The state variables are (with their mathematical symbols dt M
and units in parentheses, Haney and Jackson, 1996): where N0 is the subsurface nitrate concentration which is assumed
to be constant, ␴n is the phytoplankton specific growth rate for
nitrate uptake and other parameters have been defined above. The
i Phytoplankton Concentration (P, ␮M-N) and
parameters in model equations together with their unit are listed
ii Nitrate Nitrogen (Nn , ␮M-N)
in Table 1.

During monsoon the phytoplankton growth rate in Sundar-
bans estuary is nitrogen-limited and nitrate is usually considered
to be added to the euphotic zone by phytoplankton nutrient
uptake/growth, hence, we have used it as model currency (in form
of nitrate concentration, Nn ) (Chaudhuri et al., 2012; Dugdale and
Goering, 1967) and the state variables are modeled in terms of their
nitrogen content.
3.2. Choice of parameters and numerical methods
We model the flow in the river Jagaddal (Fig. 1) as being
one-dimensional. Using a single spatial parameter x (m) we have
segmented the distance (approximately 21 km) down the estuary
to the mouth (Indrapur) from its head (Ramganga). Concentrations
of state variables, i.e., phytoplankton and nitrate nitrogen are only
114 T. Das et al. / Ecological Modelling 368 (2018) 111–120

Table 1
Notation. List of Model Parameters. Note that ␮M is equivalent to mmol m−3.

Serial No. Symbol Meaning Units

1 ␥1 Fraction of phytoplankton growth excreted as DON –

2 ␴ Phytoplankton average daily specific growth rate day−1
3 G1 Phytoplankton loss to zooplankton grazing ␮M-N day−1
4 ␮1 Phytoplankton specific mortality rate day−1
5 m Eddy diffusion mixing rate
6 M Mixed – layer depth m
7 ␴n Phytoplankton average daily specific growth rate for nitrate uptake day−1
8 N0 Subsurface nitrate concentration ␮M-N
9 P Initial value for vertically averaged concentrations of phytoplankton in the mixed layer respectively ␮M-N
10 Nn Initial value for vertically averaged concentrations of nitrate in the mixed layer respectively ␮M-N

allowed to vary along the length of the river. They are treated as Set of all variables except Xi is indicated by X∼i . Due to the
homogeneous across the river cross-section (Dobbins, 1964). Model interaction effect of Xi and Xj ,
calibration is done by adjusting selected model parameter values
derived from studies on the Sundarbans estuary or from other d

coastal marine water body with similar estuarine conditions. The
model equations are integrated using a fourth-order Runge – Kutta
algorithm implemented by C++ programming language. The model
is simulated for 2 h, 6 h and 12 h. The observational data (described
in Section 2) of the state variables along the river Jagaddal are used
for the validation of the model run (Fig. 4). The sensitivity analysis
is performed using MatLab.
3.3. Parameter sensitivity analysis
The values for biological parameters cannot strictly be regarded
as constants. In an attempt to match the modeled distribution of
phytoplankton and nitrate to the observations obtained in river
Jagaddal, we have chosen parameter values from the references
with similar environmental and hydrological conditions. To inves-
tigate sensitivity of the model response to the model parameters
and interactions among the parameters, a variance – based sensi-
tivity analysis method developed by Sobol (1993) has been used.
In this method, within a probabilistic framework, the variance of
the output of the model is decomposed into fractions which can
be attributed to the input parameters. To understand the method
further, let Y = f(X), is a function, where X is a vector of ‘d’ uncer-
tain model inputs {X1 , X2 , X3 ,. . ..Xd } and Y is a chosen univariate
model output. To know the contribution to the output variance of
the main effect Xi , a direct variance – based measure of sensitivity
Si , called the “First – order sensitivity index” or “Main effect index”
Sobol (1993) is stated as follows,
Vi
Si =
Var (Y )
 STi ≥ 1
i=1
The sum of STi will only be equal to 1, when the model is purely
additive.
The variance-based sensitivity analysis method of Sobol (1993)
is carried out on phytoplankton and nutrient equation (Eqs. (2) and
(3) respectively) to determine the first order sensitivity index (Si )
and total effect index (STi ). The first order sensitivity index (Si ) indi-
cates the main effect output variance of input parameter, Xi , that
could be reduced if it could be fixed to a specific value (Eq. (4)).
The STi indicates the sum that includes first order index (Si ), second
order index (Eq. (5)), third order index and so on. It is generally
acceptable to determine only Si and STi . (Fox et al., 2010). To know
whether each parameter interacts as a pair or more with the other
parameters, (STi − Si ), is used to estimate interaction effect index
of second order and/or higher order higher order effects. Finally
this variance-based method is used to compare the contribution of
parameters based on the sensitivity analysis results.
3.4. Stability of the model
Stability analysis of a subsystem of an ecological model con-
structed by a number of differential equations provides useful
information about the significance or contribution of each model
component to the stability of the whole system. In this study we
have used a straightforward approach to perform the stability anal-
ysis of model equations (Eqs. (2) and (3)) performing Routh –
Hurwitz stability test.
Let a dynamical system of n variables X1 , X2 , X3 ,. . ..Xn is
(4)
described by n linear differential equations

Therefore the first order sensitivity index, Si , measures the effect dXi
of varying Xi alone, but averaged over variations over other input = fi (X1 , . . .., Xn ; C1 , . . .., Cn ) ,
dt
parameters. Higher – order interactions Sij and Sijk and so on can
be formed by dividing other terms in the variance decomposition where C1 , . . .., Cn are parameters, then an equilibrium point is a
by Var (Y), vector X such that all of the derivatives are 0. The system is stable,
if for all such X, the system will return to X after a small perturbation

d

d of the components of X.
Si + Sij + . . .S12....d = 1 (5) The characteristic polynomial |I − A| can be represented as
i=1 i<j
n + a1 n−1 + . . . + an−1  + an ,
To measure the contribution to the output variance of Xi , includ-
ing all variance caused by its interactions, of any order, with any Where the coefficients ai are real constants, i = 1, 2, . . ., n,
other input variables, an another index called “Total – effect index” The system is stable if and only if all the real parts of the
or “Total – order index” (Homma and Saltelli, 1996) is stated as characteristic polynomial (i.e., the eigenvalues) are negative. The
follows, straightforward approach of testing the stability of system is to find
the eigenvalues. Routh – Hurwitz Theorem provides an alternative
EX ∼i (Var Xi (Y |X ∼i )) Var X ∼i (EX i (Y |X ∼i )) way to find the necessary and sufficient conditions that the real
STi = =1 − (6)
Var (Y ) Var (Y ) parts of the roots of a polynomial are negative.
 T. Das et al. / Ecological Modelling 368 (2018) 111–120 115

Table 2 relations, and this was followed by parameters m, ␴n, Nn, M and P
Model Parameter Value for Standard Run. Refer to Table 1 for parameter definition.
in decreasing order (Table 4).
Serial No. Symbol Model Parameter Value References The total effect sensitivity index in Fig. 2 shows that the STi has
1 ␥1 0.05 Haney and Jackson (1996) a consistent trend as the Si , in case of ␮1, m and G1, that is, the
2 ␴ 2 day−1 Calibrated ranking of parameters by Si values agreed well with those from STi
3 G1 0.021 ␮M-N day−1 Calibrated values (Table 3). This clearly indicates that phytoplankton specific
4 ␮1 0.54 day−1 Haney and Jackson (1996) mortality rate, ␮1 and phytoplankton loss to zooplankton grazing,
5 m 0.2 m day−1 Fasham et al. (1990)
G1 play a dominant role in interaction of ecological processes with
6 M 4.6 m Goutam et al. (2015)
7 ␴n 2.3 day−1 Calibrated estuarine hydrodynamics through mixing. In case Fig. 3, STi shows
8 N0 1.2 ␮M-N Calibrated that the has a consistent trend as the Si , in case of ␴n, Nn, M and P
9 P 0.0549 ␮M-N Observed Data (Fig. 4) (Table 4).
10 Nn 0.134 ␮M-N Observed Data (Fig. 4)
To identify the contribution of parameters and the interaction
with the other parameters (STi – Si ) is used to estimate the inter-
The n Hurwitz matrices can be defined using the coefficients ai action effect index of second order and/or higher order effects. The
of the characteristic polynomial: highest interaction effect was showed by STP – S P followed by
parameters  1 , m, G1 ,  and 1 in decreasing order (Table 3) in
a1 1 Fig. 2. In Fig. 3, the highest interaction effect was showed by S␴n –
H1 = (a1 ), H2 = ( ) and
a3 a2 S␴n followed by parameters m, N0 , M, P and Nn in decreasing order
⎛ ⎞
a 1 0 0 ··· 0 (Table 4).
1
⎜ ⎟
Hn = ⎝ a3 a2 a1 1 0 ⎠ 4.2. Discussions on the stability of the system
0 0 0 0 ··· an
In this study we have used the system of linear differential equa-
Where aj = 0 if j > n. All of the roots of the polynomial are negative or tions with two state variables, Phytoplankton Concentration, P and
have negative real part iff the determinants of all Hurwitz matrices Nitrate Nitrogen, Nn . Phytoplankton and nitrate distribution are
are positive: modeled using Eqs. (2) and (3) respectively.
det Hj > 0, j = 1, 2, . . ., n. The characteristic equation of the system (2) and (3) may be
represented as

|I − A| = 0 with
4. Results
⎡ m + h+ (t)
⎤
The model is run to study the spatial variation of phytoplank- (1 − 1 )  − 1 − 0
ton abundance and nutrient (nitrate) distribution in monsoon along A=⎣
⎢ M ⎥
⎦ (7)
the river Jagaddal up to Ramganga away from Indrapur (i.e., form m + h+ (t)
the mouth of Bay of Bengal). In this section, first we will discuss −n −
M
the result of sensitivity analysis described in Section 3.3 and signif- or
icance of the major contributing parameters. Second, the conditions    2
for the stability of the system will be discussed. Finally the nature 2 (m + h+ (t)) m + h+ (t)
2 +  1 − (1 − 1 )  + +
of phytoplankton and nutrient distribution obtained from model M M
run is explained.

1 (m + h+ (t)) m + h+ (t)
4.1. Discussions on sensitivity analysis + −  (1 − 1 ) =0 (8)
M M

Local and global sensitivity analyses were carried out on phyto- When n = 2, the Routh-Hurwitz criteria (described in Section 3.4)
plankton and nutrient model Eqs. (2) and (3) respectively based on simplify to det H1 = a1 > 0 and
the input parameters shown in Table 2. Both analyses were studied  
for each model equation at 105 s simulation time. Local sensitiv- a1 1
det H2 = det = a1 a2 > 0
ity analysis was carried out to determine normalized sensitivity 0 a2
coefficients by varying input parameters at prescribed percentage,
while global sensitivity analysis was carried out to determine the or a1 > 0 and a2 > 0.
fraction of model output variance, that is, first order index (Si ) and The Routh array of coefficients of Eq. (8) becomes:
total effect index (STi ), for each parameter, based on uncertainty in m + h+ (t)
2
1 (m + h+ (t)) m + h+ (t)
input value range. Input parameters were varied ±5% (Zheng and 2 1 ( ) + − (1 − 1 )
M M M
Bennett, 2002) for Eqs. (2) and (3), and the results of total effect
index, main effect index and interaction index for model parame- 2(m + h+ (t))
1 1 − (1 − 1 ) + 0
ters in phytoplankton and nutrient equations (Eqs. (2) and (3)) are M
2
shown in Figs. 2 and 3 respectively. m + h+ (t) 1 (m + h+ (t)) m + h+ (t)
Quantitative contribution for each parameter to model output 0 ( ) + − (1 − 1 ) 0
M M M
variance for model Eqs. (2) and (3) is given by first order index
For the stability of the system the all the elements of the first
(Si ) and total effect index (STi ). In Fig. 2, among the model param-
column must be greater than zero. According to the values used for
eters of phytoplankton Eq. (2), parameter ␮1 appeared to have
model parameters listed in Table 2, this condition satisfies. Hence
the most sensitivity in positive relations, and this was followed by
the system is in stable condition.
parameters ␥1 , ␴, P, m and G1 in decreasing order. In negative rela-
If the system is in stable condition, then two conditions arise:
tions, parameters m and G1 are in decreasing order of importance
(Table 3). In case of Fig. 3, among the parameters of nutrient Eq. 2 (m + h+ (t))
1 − (1 − 1 )  + >0 (9)
(4), parameter N0 appeared to have the most sensitivity in positive M
116 T. Das et al. / Ecological Modelling 368 (2018) 111–120

Fig. 2. Sensitivity Indices versus Model Parameters of Phytoplankton Equation (Eq. (2)). The parameters are arranged sequentially as: (1) ␥1 is the fraction of phytoplankton
growth excreted as DON, (2) ␴ is the average daily phytoplankton specific growth rate, (3) G1 represents the rate of phytoplankton loss to grazing, (4) ␮1 is the phytoplankton
specific mortality rate, (5) m is a mixing velocity (6) P is the initial value (Here we have taken the value from the observation data starting from Ramganga, the head of the
estuary) for vertically averaged concentrations of phytoplankton in the mixed layer. The contribution of the parameter M, the depth of vertical water column of the estuary,
is too insignificant to consider. So we have excluded it from the table.

Fig. 3. Sensitivity Indices versus Model Parameters of Nutrient Equation (Eq. (3)). The parameters are arranged sequentially as: (1) ␴n is the phytoplankton specific growth
rate for nitrate uptake, (2) m is a mixing velocity, (3) M is the depth of vertical water column of the estuary, (4) N0 is the subsurface nitrate concentration, (5) P is the initial
value (Here we have taken the value from the observation data starting from Ramganga, the head of the estuary) for vertically averaged concentrations of phytoplankton
in the mixed layer and (6) Nn is the initial value (Here we have taken the value from the observation data starting from Ramganga, the head of the estuary) for vertically
averaged concentrations of nitrate.
 T. Das et al. / Ecological Modelling 368 (2018) 111–120 117

Table 3
Ranking of Model Parameters Based on Sensitivity Analysis for Phytoplankton Equation (Eq. (2)). The parameters are arranged sequentially as: (1) ␥1 is the fraction of
phytoplankton growth excreted as DON, (2) ␴ is the average daily phytoplankton specific growth rate, (3) G1 represents the rate of phytoplankton loss to grazing, (4) ␮1 is
the phytoplankton specific mortality rate, (5) m is a mixing velocity (6) P is the initial value (Here we have taken the value from the observation data starting from Ramganga,
the head of the estuary) for vertically averaged concentrations of phytoplankton in the mixed layer. The contribution of the parameter M, the depth of vertical water column
of the estuary, is too insignificant to consider. So we have excluded it from the table.

Serial No. Symbol Main Sensitivity Rank Based on Main Total Sensitivity Rank Based on Interaction
Index (Si ) Sensitivity Index Index (STi ) Effect Index (STi – Si )

1 ␥1 0.243403 2 0.25495 2
2 ␴ 0.239142 3 0.24744 5
3 G1 −0.010754 6 8.3078e-011 4
4 ␮1 0.250212 1 0.25642 6
5 m −0.010745 5 1.4294e-007 3
6 P 0.236186 4 0.25615 1

Table 4
Ranking of Model Parameters Based on Sensitivity Analysis for Nutrient Equation (Eq. (3)). The parameters are arranged sequentially as: (1) ␴n is the phytoplankton specific
growth rate for nitrate uptake, (2) m is a mixing velocity, (3) M is the depth of vertical water column of the estuary, (4) N0 is the subsurface nitrate concentration, (5) P
is the initial value (Here we have taken the value from the observation data starting from Ramganga, the head of the estuary) for vertically averaged concentrations of
phytoplankton in the mixed layer and (6) Nn is the initial value (Here we have taken the value from the observation data starting from Ramganga, the head of the estuary)
for vertically averaged concentrations of nitrate.

Serial No. Symbol Main Sensitivity Rank Based on Total Sensitivity Rank Based on
Index (Si ) Main Sensitivity Index Index (STi ) Interaction Effect Index (STi – Si )

1 ␴n 0.2669231 3 0.28362 1
2 m 0.2783879 2 0.28752 2
3 M 0.0324117 5 0.034126 4
4 N0 0.2807541 1 0.28664 3
5 P 0.0026569 6 0.000030409 5
6 Nn 0.1266412 4 0.12354 6

and around the densely forest covered area of Dhanchi. As we cross

 2 Dhanchi and move towards Indrapur, three is a sharp decreasing
m + h+ (t) 1 (m + h+ (t)) m + h+ (t) trend of phytoplankton concentration (Fig. 4).
+ −  (1 − 1 ) > 0 (10)
M M M Nutrient distribution (here, nitrate) shows a wide range of
interesting variation along river Jagaddal. Initially, in Ramganga,
Eq. (9) yields
concentration of nutrient in the river is increasing. Then, there is a
2 (m + h+ (t)) sudden drop in nutrient concentration level. Between Ramganga
1 > (1 − 1 )  − (11)
M and Dhanchi, nutrient concentration shows an increasing trend
Eq. (10) yields again. Near Dhanchi, there is another drop in nutrient concentra-
tion over a specific region on Jagaddal. After this point, there is a
m + h+ (t) sharp rise in nutrient concentration around Dhanchi and decreasing
>0 (12)
M trend in nutrient concentration towards Indrapur (Fig. 4).
and Our goal is to know how much the formulation of model scenario
m + h+ (t) is close to interpreat observed spatial variations of state variables.
+ 1 −  (1 − 1 ) > 0, or Now it is quite interesting to investigate significance of the model
M
parameters (Tables 1 and 2) in context of this estuarine river. In
m + h+ (t) 1 monsoon the phytoplankton average daily specific growth rate (␴n )
< + (13) has maximum of 2.3 day−1 , actually matches the maximum value
M (1 − 1 ) (1 − 1 )
of phytoplankton specific mortality rate (␮1 ) of 0.054 day−1 . The
Eqs. (11) and (13) provides the conditions for the two important growth rate and mortality rate of phytoplankton are within the
parameters phytoplankton specific mortality rate and phytoplank- range as reported by Haney and Jackson (1996). But the growth
ton average daily specific growth rate respectively. Eq. (12) clearly rate is low to the value 2.62 day−1 mortality rate is much higher
indicates that the rate of change of mixed layer together with to the value 0.021 day−1 as prescribed by Ray et al. (2001) and
mixing rate m provides (Eqs. (1) and (1A)) the physically driven Mandal et al. (2012) respectively. During monsoon low growth rate
rate to change the concentration of phytoplankton and nutrient of phytoplankton observed due to much lower penetration of light
in the water column plays the major role as the physical forcing turbidity light in rivers of Sundarban that affects the growth rate.
mechanism to maintain the stable condition of phytoplankton and In this shallow river, changing the subsurface nitrate condition,
nutrient in estuarine environment (Eqs. (1) and (1A)). N0, shows no such significance variation N0 = 1.2 ␮M is the opti-
mal level which tends to match the surface nitrate concentration
4.3. Features of phytoplankton and nutrient dynamics obtained as observed in field data. In case of G1, i.e., Phytoplankton loss to
from model run zooplankton grazing, we have to depend on the reference data from
similar environmental conditions in Sundarbans and other tropical
The observed trend phytoplankton distribution in river Jagaddal estuaries and found that the value 0.021 ␮M-N day−1 has justi-
shows an increasing trend from the first observation station of Ram- fied our model conditions (Mandal et al., 2012; Haney and Jackson,
ganga. A sharp peak of phytoplankton concentration is observed 1996). Similar approach has been taken to check the growth cor-
around Ramganga. Between Ramganga and Dhanchi, there is a rected for algal leakage using (1- ␥1 ), where the value of ␥1 is
decreasing trend of phytoplankton concentration. Another sharp set as 0.05 (Haney and Jackson, 1996). Now, the nutrient-plankton
peak of phytoplankton concentration in the river is observed
118 T. Das et al. / Ecological Modelling 368 (2018) 111–120

Fig. 4. Observed distribution of Phytoplankton and Nutrients along the River Jagaddal during Monsoon. Straight line represents the distribution of phytoplankton concen-
tration and the dotted line segment represents the nitrate concentration distribution. The flow of the river considered being one-dimensional, to describe the distance from
the mouth of the estuary to upwards direction, the entire stretch of river Jagaddal from Indrapur (from extreme left along the x- axis) to Ramganga (21 Km. approximately)
is divided using a single spatial parameter x(n) and represented along the x – axis where ‘n’ is the number of one dimensional grids.

Fig. 5. Evolution of Spatial Distribution of Phytoplankton Concentration during Monsoon along River Jagaddal. Units and parameter values are described in Table 2. The flow
of the river considered being one-dimensional, to describe the distance from the mouth of the estuary to upwards direction, the entire stretch of river Jagaddal from Indrapur
(from extreme left along the x- axis) to Ramganga (21 Km. approximately) is divided using a single spatial parameter x(n) and represented along the x – axis where ‘n’ is the
number of one dimensional grids. The solid black line represents the observed pattern and the dotted black, green and red line segment represent the simulated pattern of
phytoplankton concentration during the model run for 2 h, 6 h and 12 h respectively. (For interpretation of the references to color in this figure legend, the reader is referred
to the web version of this article.)

reaction to the diffusion process in shallow estuaries evolves in
small scale interactions, which need special care to reconstruct or
to understand the sensitivity of the organisms and organic particle
to such processes. In case of rivers in Sundarban, the residual cir-
culation plays a key role in maintaining mixing characteristics in
the mixed layer (Goutam et al., 2015). In this scenario, the model
has not yet modified up to that level so that it can interpret the
role of eddy diffusivity in such stochastic processes. Hence, we do
not observe much variation in Eddy Diffusivity rate is set to 0.2 m
(Haney and Jackson, 1996) to satisfy our model conditions.
Phytoplankton and nutrient in the model run with variation of
time along the river shows an overall stable behavior (Figs. 5 and 6).
Distribution of nutrient and phytoplankton concentration on river
Jagaddal can be explained as the interaction of estuarine hydrolog-
ical processes and dynamics of ecological system of Sundarbans
estuary. As expected, nutrient rich water regime around Ram-
ganga and Dhanchi shows dense population of phytoplankton. The
enriched source of nutrient in Ramganga is the additional contribu-
tion of river channel Bar Chara and Kal Chara which are connected
to the fresh water of Hoogly River through the western channels
(Fig. 1). As we depart from Ramganga, there is a confluence of
another river channel coming to the river Jagaddal. This junction
is characterized by turbulence and mixing phenomenon as well as
variable salinity regime (Goutam et al., 2015), which is the reason
behind the sudden drop of nutrient level. This drop is increasing
with time as observed in model run (4.). Since there is no signif-
icant change in phytoplankton concentration in this location, we
can expect a continuous supply of phytoplankton by the river flow
in this location (Fig. 5).
The river regime between Ramganga and Dhanchi gathers much
interest for the ambiguous nature of phytoplankton and nutrient
distribution. As we discussed earlier, though the river water is
enough enriched with nutrient, there is a sharp decline in phy-
toplankton population. To understand this ubiquity, we have to
consider the hydrological processes underlying the flow regime. If
we consider the depth profile of river Jagaddal, then we could find
that river in this region is quite shallow (Goutam et al., 2015). Tidal
current in this region is very strong. River Jagaddal is connected
to the much saline water of river Thakuran (in the east) through
Dhanchi Canal near Dhanchi forest. This canal and adjacent shal-
low tidal creeks draw enough water volume at this location which
is dispersed by the strong nonlinear tidal phenomena (tidal asym-
metry and bottom drag coefficient). As a result this region shows
low tidal water range and strong mixing in water column (Goutam
et al., 2015). All of these contribute to the resulted phytoplank-
 T. Das et al. / Ecological Modelling 368 (2018) 111–120 119

Fig. 6. Evolution of Spatial Distribution of Nutrient Concentration (Nitrate) during Monsoon along River Jagaddal. Units and parameter values are described in Table 2. The
flow of the river considered being one-dimensional, to describe the distance from the mouth of the estuary to upwards direction, the entire stretch of river Jagaddal from
Indrapur (from extreme left along the x- axis) to Ramganga (21 Km. approximately) is divided using a single spatial parameter x(n) and represented along the x – axis where
‘n’ is the number of one dimensional grids. The solid black line represents the observed pattern and the dotted black, green and red line segments represent the simulated
pattern of nitrate concentration during the model run for 2 h, 6 h and 12 h respectively. (For interpretation of the references to color in this figure legend, the reader is referred
to the web version of this article.)

ton volume decrease with a mismatch to the increasing trend of
nutrient concentration (Figs. 5 and 6).
As we proceed to the river regime adjacent to the densely cov-
ered mangrove forest of Dhanchi, we found the river water is
very nutrient-rich. In model run, we have observed coincidence
of phytoplankton and nutrient concentration peaks in this region
(Figs. 5 and 6). This region is characterized by very high density
of phytoplankton concentration. The adjacent riverside of forest
area is covered with salt marsh, which acts as the water reservoirs
(Friedrichs et al., 1988; Chatterjee et al., 2013). The stability of water
column in this region is another contributor to the higher growth
rate of phytoplankton in this region. The riverine area around the
forest covered region of Dhanchi, thus serve as potential fishing
zone and treasurer of ecological resources. The strong influence
of tidal hydrodynamics on the river regime in this area shows that
importance of the tidal water inflow through river run off in preser-
vation of the ecological health of this fishing zone.
After crossing Dhanchi, as we move towards Indrapur, there is
a sharp decline in nutrient and phytoplankton concentration. The
station Indrapur, as we discussed earlier, is only on 8 Km. distance
from Bay of Bengal and for this reason the estuarine waters in this
experience direct influence of Bay. Small scale eddies; continuous
mixing of riverine fresh water with saline tidal inflow and turbu-
lence in water column are the regular characteristics of the flow
regime in this area. As a result, nutrient concentration in the estu-
arine water is not enough to provide a favorable condition for the
rapid growth of phytoplankton.
The distribution of nutrient concentration is not uniform along
the river channel. It is expected due to the variable contribution
of ecology of a mangrove estuarine river condition (e.g. presence
of mangrove forest). The role of physical processes (e.g. effect of
channel confluence, mixing etc.) of estuarine hydrodynamics are
prominent in the variation of spatial distribution of state variables
in this river. Here we have discussed the possible mechanism that
plays behind the observed model run to match the general trend of
nutrient and phytoplankton distribution to the observed field data.
5. Conclusion
This study is an initial attempt to model the phytoplankton
variation (distribution, growth rate, mortality rate) in the wide
range of estuarine condition (nutrient distribution, mixing level
etc.). This exercise will obviously set a background to have some
knowledge of the tuning of the ecological model to the parameter
values. The procedure used in this paper is to test the sensitiv-
ity of each model parameter in turn by running the model to a
steady-state within a certain range of value satisfying the trend of
phytoplankton and nutrient distribution along the river Jagaddal. In
this way, it is possible to determine which parameters had the most
influence on which variables and the possible mechanism underly-
ing the ecological processes in estuarine environment. The model
simulations were run until a steady-state is achieved. The state
variables (phytoplankton and nutrient concentration) maintain the
stable behavior during tidal cycle. This result clearly indicates that
the conservation principles of state. This conclusion will further
indulge the researchers to develop a full fledge hydrodynamic
model containing physical features of estuarine circulation as a
forcing mechanism to the biological processes inherent in the pop-
ulation dynamics of ecosystem in Sundarbans Estuarine Systems.
Acknowledgments
This work was conducted as part of the research project entitled
“Study of tidal Dynamics and Interaction with Surge in Sundarban
Estuary” at School of Oceanographic Studies, Jadavpur Univer-
sity funded by INCOIS, Ministry of Earth Sciences, Government
of India, under HOOFS programme. The observation work and
analysis of collected data (salinity, phytoplankton and nutrient con-
centration) in laboratory was performed by Saranya Chakraborti.
Author gratefully acknowledges the valuable inputs of Dr. Joydeep
Mukherjee, Associate Professor, School of Environmental Studies,
Jadavpur University and Co-Principal Investigator of the project
during the manuscript preparation and to extend his cooperation
in performing analysis in laboratory. Finally, the author would like
to sincerely thank Prof. Goutam Kumar Sen, Professor, School of
Oceanographic Studies, Jadavpur University and Principal Investi-
gator of the project for his continuous guidance and immense effort
to supervise this study.
References
Abraham, E.R., 1998. The generation of plankton patchiness by turbulent stirring.
Nature 391, 577–580.
Bainbridge, R., 1957. Size, shape and density of marine plankton concentrations.
Biol. Rev. 32, 91–115.
Biktashev, V.N., Brindley, J., 2004. Phytolankton blooms and fish recruitment rate:
effects of spatial distribution. Bull. Math. Biol. 66, 233–259.
Bonachela, Juan A., Raghib, Michael, Levin, Simon A., 2011. PNAS 108 (51),
20633–20638.
Burmaster, D.E., 1979. The continuous culture of phytoplankton: mathematical
equivalence among three unsteady state models. Am. Nat. 113, 123–134.
Caperon, J., 1967. Population growth in micro-organisms limited by food supply.
Ecology 48, 715–722.
120 T. Das et al. / Ecological Modelling 368 (2018) 111–120
Chatterjee, M., Shankar, D., Sen, G.K., Sanyal, P., Sundar, D., 2013. Tidal variation in
the Sundarbans Estuarine System, IndiaTidal variation in the Sundarbans
Estuarine System, India. J. Earth Syst. Sci. 122 (4), 899–933.
Chaudhuri, et al., 2012. Physicochemical and biological factors controlling water
column metabolism in Sundarbans estuary, India. Aquat. Biosyst. 8, 26.
Dobbins, W.E., 1964. BOD and oxygen relationships in streams. J. Sanit. Eng. Div.
Proc. ASCE 90, 53–78.
Droop, M.R., 1973. Some thoughts on nutrient limitation in algae. J. Phycol. 9,
264–272.
Dugdale, R.C., Goering, J.J., 1967. Uptake of new and regenerated forms nitrogen in
primari production. Limnol. Oceanogr. 12, 196–206.
Dugdale, R.C., 1967. Nutrient limitation in the sea: Dynamics, identification, and
significanceNutrient limitation in the sea: dynamics, identification, and
significance. Limnol. Oceanogr. 12, 685–695.
Evans, G.T., Parslow, J.S., 1985. A model ofannual plankton cycles. Biol. Oceanogr. 3,
327–347.
Fasham m, J.R., Ducklow, H.W., McKelvie, S.M., 1990. A nitrogen-based model of
plankton dynamics in the oceanic mixed layer. J. Mar. Res. 34, 591–639.
Fasham, M.J.R., 1985. Flow analysis of materials in the marine euphotic zone. Can.
Bull. Fish. Aquat. Sci. 213, 139–162 (Ecosystem Theory for Biological
Oceanography, R.E. Ulanowicz and T. Platt, eds.).
Fox, G.A., Muñoz-Carpena, R., Sabbagh, G.J., 2010. Influence of flow concentration
on parameter importance and prediction uncertainty of pesticide trapping by
vegetative filter strips. J. Hydrol. 384 (1–2), 164–173.
Friedrichs, Carl, T., Aubrey, David, G., 1988. Non-linear Tidal Distrotion in Shallow
Well-mixed Estuaries: a SynthesisNon-linear tidal distrotion in shallow
well-mixed estuaries: a synthesis. Estuar. Coast. Shelf Sci. 27, 521–545.
Goutam, K.S., Tanaya, D., Anwesha, S., Sharanya, C., Meenakshi, C., 2015. Tide and
Mixing Characteristics in Sundarbans Estuarine River SystemTide and mixing
characteristics in Sundarbans Estuarine River System. Hydrol. Curr. Res. 6, 204.
Haney, Jeffrey D., Jackson, George A., 1996. Modelling phytoplankton growth rates.
J. Plankton Res. 18 (1), 63–85.
Homma, T., Saltelli, A., 1996. Importance measures in global sensitivity analysis of
nonlinear models. Reliab. Eng. Syst. Saf. 52, 1–17.
Jang, Sophia R.-J., Allen, Edward J., 2015. Deterministic and stochastic
nutrient-phytoplankton-zooplankton models with periodic toxin producing
phytoplankton. Appl. Math. Comput. 271, 52–67.
Jha, R., Ghosh, N.C., Chakraborty, B., 1999. Flow computation of river estuaries
using finite element model; Technical Note CS(AR) 21/96-97Flow Computation
of River Estuaries Using Finite Element Model; Technical Note CS(AR)
21/96-97. National Institute of Hydrography, Roorkee (78 p.).
Joyita, M., Debaldev, J., Moitreyee, B., Bhusan, P.G., Santanu, R., 2015. Utilization of
carbon in NPZ model of Hoogly Estuarine System, IndiaUtilization of carbon in
NPZ model of Hoogly Estuarine System, India. J. Earth Sci. Clim. Change 6, 292.
Mandal, Sudipto, Ray, Santanu, Ghosh, Phani Bhusan, 2012. Modelling nutrient
(dissolvd inorganic nitrogen) and planktonic dynamics at Sagar Island of
Hoogly-Matla Estuarine System, West Bengal, IndiaModelling nutrient
(dissolvd inorganic nitrogen) and planktonic dynamics at Sagar Island of
Hoogly-Matla Estuarine System, West Bengal, India. Nat. Resour. Modell. 25
(November (4)).
Martin, A.P., Richards, K.G., 2001. Mechanisms for vertical nutrient transport
within a North Atlantic mesoscale eddy. Deep-Sea Res. II 48, 757–773.
McGillicuddy Jr., D.J., Robinson, A.R., 1997. Eddy-induced nutrient supply and new
production in the Sargasso Sea. Deep-Sea Res. 144, 1427–1450.
Mitra, A., Gangopadhyay, V., Dube, A., Schmidt, A.C.K., Banerjee, K., 2009. Observed
changes in water mass properties in the Indian Sundarbans (northwestern Bay
of Bengal) during 1980–2007. Curr. Sci. 97 (10), 1445–1452.
Mitra, Abhijit, Zaman, Sufia, Sett, Saurav, Raha, Atanu Kumar, Banerjee, Kakoli,
2014. Phytoplankton cell colume and diversity in Indian Sundarbans. Indian J.
Geo-Mar. Sci. 43 (2), 208–215.
Moloney, C.L., Bergh, M.O., Field, J.G., Newell, R.C., 1986. Theeffect of sedimentation
and microbial regeneration in a plankton community: a simulation
investigation. J. Plankton Res. 8, 427–445.
Mukhopadhyay, B., Bhattacharya, R., 2006. Modelling phytoplankton allelopathy in
a nutrient-plankton model with spatial heterogeneity. Ecol. Modell. 198,
163–173.
Oschlies, A., Garcon, V., 1988. Eddy-induced enhancement of primary productivity
in a model of the North Atlantic. Nature 398, 266–268.
Pace, M.L., Glasser, J.E., Pomeroy, L.R., 1984. A simulation analysis of continental
shelf food webs. Mar. Biol. 82, 47–63.
Raha, A., Das, S., Banerjee, K., Mitra, A., 2012. Climatic change impacts on Indian
Sunderbans: a time series analysis. Biodivers. Conserv. 21 (5), 1924–2008.
Ray, Santanu, Berec, Ludek, Straskraba, Milan, Jørgensen, Sven Erik, 2001.
Optimization of exergy and implications of body sizes of phytoplankton and
zooplankton in an aquatic ecosystem model. Ecol. Modell. 140, 219–234.
Robinson, m.K., Bauer, R.A., Schroeder, E., 1979. Atlas of North Atlantic-Indian
ocean monthly mean temperatures and mean salinities of the surface
layerAtlas of North Atlantic-Indian Ocean Monthly Mean Temperatures and
Mean Salinities of the Surface Layer. U.S. Naval Oceanographic Office Reference
Publication, pp. 18.
Roy, Ujjwal, Majee, N.C., Ray, Santanu, 2016. Temperature dependent growth rate
of phytoplankton and salinity induced grazing rate of zooplankton as
determinants of realistic multi-delayed food chain model. Model. Earth Syst.
Environ. 2, 161.
Sobol, I.M., 1993. Sensitivity estimates for nonlinear mathematical models. MMCE
1 (4), 407.
Steele, J.H., 1978. Some comments on plankton patches. In: Steele, J.H. (Ed.), Spatial
Pattern in Plankton Communities. Plenum Press, New York and London, pp.
1–20.
Venrick, E.L., 1990. Mesoscale patterns of chlorophyll-A in the central North
Pacific. Deep-Sea Res. A 37, 101–1031.
Williams, P.M., Druffel, E.R.M., 1988. Dissolved organic matter in ocean: comments
on a controversy. Oceanography 1, 14–17.
Zheng, C., Bennett, G.D., 2002. Applied Contaminant Transport Modeling, 2nd ed.
John Wiley & Sons., New York.