Luận văn Thạc sĩ Kỹ thuật: Assessment of salinity intrusion in the Red River Delta Vietnam

ASSESSMENT OF SALINITY INTRUSION IN THE RED RIVER DELTA VIETNAM

Le Thi Thu Hien

A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering

Examination Committee: Dr Roberto Clemente (Chairman)

Dr Sutat Weesakul (Co-chairman)

Prof Ashim Das Gupta Dr Mukand S Babel

Nationality: Vietnamese

Previous degree: Bachelor of Engineering in

Water Resources Engineering Water Resources University

Hanoi, Vietnam

Fellowship Donor: The Government of Denmark

Asian Institute of Technology School of Civil Engineering

ThailandMay 2005

LIST OF TABLES Vii

LIST OF FIGURES Vili 1 Hoabinh Hydropower Plant

2 Son la Hydropower Plant: On-Going construction

1.2.4 Tidal Regime and Salinity Intrusion 6 1.25 Existing Land Use 6

2.1.3 Salinity Intrusion Study in Vietnam 17

2.2 Salinity Control Requirement for Irrigation and Aquaculture 18

IH THEORETICAL CONSIDERATIONS 19

3.1 Characteristics of Estuary 19

3.1.1 Stratified Estuary 20 3.1.2 Partially Mixed Estuary 20

Chapter Title Page

3.1.3 Well Mixed Estuary jy 20 3.2 Numerical Computation 21

3.2.1 Finite Difference Tidal Hydraulics Equations 22

3.3 Model characteristics 23

Chapter I

1.1 Problem Identification

Intrusion of salt-water in dry season is a well-known phenomenon in the Red-Thaibinh estuaries In the rainy season from June to November the discharge of

freshwater from upstream is high, the saltwater is pushed to the sea and the problem of salinity intrusion is not present But in the dry season from December to May of the next year the discharge of freshwater from upstream is small and the salinity intrusion problem becomes serious In some branches of the Red river system, the distance of

salinity intrusion may be up to 40 km As the increasing of the freshwater intake for irrigation, the salinity intrusion is causes a lot of problems for irrigation, aquaculture and other economic activities due to lack of freshwater The knowledge of the characteristics of salinity intrusion therefore is very necessary for solving the problems and utilizing the river.

Saltwater intrusion in Red river delta has been studied for several years ago Many institutions involve research works for controlling and predicting Red-river

salinity intrusion by various methods with mathematical models VRSAP (Water

Resources Planning Institute, Hanoi Water Resources University), TL1, TL2 (Institute of Mechanics), the hydrological and meteorological models (Hydrological and Meteorological Services), the multivariable relational model (Institute for water resources research) However there is a lack of efficient numerical hydrodynamic

models that consider effect of Hoabinh reservoir as well as calculation and prediction

salinity intrusion in Red river delta Sonla Hydropower Plant is going to built in

upstream of Da River to reduce flood damage and improve irrigation in the Red River

Delta Increasing inflow for irrigation in dry season can cause a change of salinity concentration for planning aquaculture area in Red River Delta’s coastal zone How to supply sufficient freshwater for paddy crops while controlling salinity concentration for

aquaculture area? This is an important issue to assess the entire effect of Sonla

Hydropower Plant to downstream area.

Moreover, global climate changes in some recent years have deep effect to

hydrology condition of Red river delta “Global Warning” could cause sea level rise 0.5 to 1 meter by the current century due to the “Greenhouse Effect’ A rise in sea level enables salt water penetrates upstream and inland, and would threaten human uses of

water particularly during droughts.

To bring a reasonable operation for both electric production and saltwater

prevention is urgent duty of Hoabinh reservoir in the future, finding a numerical model for simulation and prediction salinity intrusion in Red river delta for future is also very important.

1.2 Study Area Introduction

1.2.1 Geographical Condition

Red - Thai Binh River System is the second largest river system in Vietnam,after Mekong River It originates from Nguy Son Mountain in Yunnan province ofChina.

Fig 1.1 Red River System in Vietnam Territory

The whole basin areas occupy 169,020km” of which 86,720km” (representing

51%) are located in Vietnam’s territory as shown in Fig.1.1 and Fig.1.2 It is a population density area with high economic potential The North Delta and Midland

Region cover 14,590km” with a population of 18.56 millions in the year 2000.

Asa large river basin with a complex topography including mountains and hills

(covering 90% of the area), delta and coastal areas, the Red - Thai Binh river basin,

hosting a diverse and more and more developed socio-economy, makes a significant contribution to the national economy.

104° E 105°R 106° E 107

Figure 1.2 River Network in Red River Delta

-3-LOCATION OF STUDY AREA

The Red River Delta is in reality the delta of two river systems: the Red River System and Thai binh River System The Red River System consists of 3 major river

branches namely the Da, Lo and the Thao Rivers The Thaibinh River System is also comprised of 3 river branches, which are the Cau River, Thuong and the Luc Nam River as

shown in Figure 1.4 The two river systems are connected through the Duong and Luoc rivers forming the Red and Thaibinh River Basin.

SCHEMA OF RIVER SYSTEM

Da Thao Lo Cau Thuong

River River River River River Lucnam River

Table 1.1 Catchment’s Area and Distributed Flow of Red River Delta’s Branches Catchment’s | Area Percentage in Distributed Flow

Catchments Area Red River Delta to Red and Thaibinh river

Water resource of Red River is plentiful Annual average volume at Sontay station

is 114km? corresponding with 3643m°/s of discharge Inflow in Thaibinh River is less low

due to upstream rivers of Thaibinh River (Cau, Thuong, Lucnam) have annual inflow very

small Total water volume of Thaibinh river at Phalai is 8.26km” (equal to 7.2% ones of Red river at Sontay station) with annual discharge is 318m’/s.

Apart from inflow from Cau, Thuong and Luc Nam River, one numerous inflow is

passed from Red River at downstream of Phalai through Duong River This flow is nearly

triple are compared with Thaibinh’s (25km* compare with 8.26km*) In addition, Thaibinh

River also gets supplementary volume from Red river through Luoc River with total

volume is 13 km? per year before flowing to the sea.

In the dry season, water level in Red River fall down very low; in somewhere freshwater altitude of river is less than altitude of field’s surface inside the dyke However,

water resources of Red River keep in plentiful state so the lowest monthly average inflow

at Sontay is 691m’/s.

1.2.3 Hydraulic Constructions in Study Area 1 Hoabinh Hydropower Plant

Hoabinh Hydropower Plant was built in 1980 in the northern mountainous province of Hoabinh with assistance from the former Soviet Union.

Some characteristics of Hoabinh reservoirSurface of the reservoir F=200 km2

Hoabinh Hydropower Plant has been completely constructed in 1979 with 8

electricity generation units It has raised the discharge of flow of Da (Black river) and Red

rivers in dry season up to 400-600 m/s The flow regulation also facilitates to put saltwater

into river mouth in dry season.

2 Sonla Hydropower Plant: On-Going Construction

Sonla Hydropower Project to be constructed on the Da River, it is far from Hoabinh Hydropower Plant nearly 250 km towards upstream and about 320 km of Hanoi.

The proposed Sonla Dam would be the largest dam in Vietnam The Sonla

Hydropower Station Project will be the largest of its kind in south East Asia.

Sonla Hydropower together with Hoabinh Hydropower Plant will improve

Vietnam's electricity fuel mix, reduce flood damage and improve irrigation in the Red River Delta Sonla reservoir will hold a total of 25 billion m3 of water Together with the

Hoabinh reservoir, the water volume will total 36 billion mỶ With the Sonla reservoir,

safety discharge to Hoabinh in the dry season is 759 m°/s, raising 115 m°/s if has only

Hoabinh reservoir (Source: Proceedings of the Workshop on Methodologies for EIA of Development Projects, Hanoi, July, 1999).

Electricity of Vietnam (EVN) plans begins construction on Sonla Hydropower

Plant late 2005 First turbine expected operable 2012, the entire of construction expected compliable in 2015.

Major objectives

e Energy production: 14.16 billion KWh/year

e Regulation flood stream: very important for Hoabinh Dam and downstream

areas, including Hanoi (ensuring water level in Hanoi during flood season not to exceed 13 m).

e Water supply: providing to the Red River Delta about 6 billion m3; during dry

season will ensuring a sanitary run-off of 300-600m3/sec

e Creating new opportunities for regional socio-economic development.

Some characteristics of constructionNormal water level: 265 m

Fig 1.5 Location of Sonla and Hoabinh Reservoirs

1.2.4 Tidal Regime and Salinity Intrusion

The mixing of fresh and marine waters also is accelerated by tidal action The tidal regime in this area is irregularly diurnal, but is more regularly diurnal upstream The

maximum tidal range along the coast of the Delta is approximately 4 m The tidal transfer speed in the river mouth approaches 95-150 cm/sec and the tidal influence extend 150-180

km from the river mouths (Source: Nguyen Ngoc Thuy, 1982).

Due to low terrain and improved river mouths so much, seawater and salinity are easy to go Red River Delta in almost of annual In Thaibinh River, low river bottom datum, large estuary and upstream inflow create a good condition for severe saltwater intrusion up

far from the sea to Lucnam, Cau and Thuong River In the Red River, distance of saltwater intrusion was recorded at location which is 10 km far from Hanoi station above and 185 km far from the sea.

Salinities increase from about 0.5 ppt in the rivers to 30.0 ppt Fluctuation widely of salinity depends on the flow in the river and state of the tide Salinity concentration l ppt

can intrude about 30 — 40 km in average in the main branches with complicated characteristic.

1.2.5 Existing Land Use

Almost the entire delta has been reclaimed for agricultural land, aquaculture ponds,

forestry and urban development Approximately 53% of the delta is agricultural land, 6.4% is forestry land and there are only some 3.8% of permanent lakes and ponds for aquaculture as shown in Fig 1.4 and Table 1.2.

1 In general

The principal land use throughout the delta is the annual cultivation of rice, in

addition to the perennial crop as main fruit species Rice occupies around 93 percent of the total

annual crop area as shown in Table 1.3 Corn, sweet potato and cassava followed behind The whole region produces about three million tons of rice per year (an average yield of 2,835

kg/ha in 1995).

32

To facilitate rice production, some 1,080 km of embankments, 34,400 km of canals,

1,310 drains, 217 reservoirs and 1,300 pumping stations have been constructed.

In spite of the low salinity of estuarine water, the production of table salt by traditional measures in estuarine waters has been developed Each year the salt fields of this area have provided North Vietnam a table salt production of 20,000 — 30,000 tons.

Table 1.2 Existing Land Use in 1998 (Unit: 1000 ha)

Total Area Agricultural Land Forestry Land Aquaculture Land

1,266.3 671.8 80.9 48.7

Table 1.3 Agricultural Crop Land (Unit: 1000 ha)

Annual Crop Land

Perennial Crop Land

Total Rice Land

620.9 576.4 10.1

33

Map of Landuse in Red River Delta

(9) Other Unused Area > Rural populated area

ME Unused mountains and hills Š

EM Urban populated area

Fig 1.6 Map of Land Use in Red River Delta

2 In Coastal Zone Area

Almost coastal zone area in Red River Delta no has agricultural land and has

traditionally depended on fishing and salt production Production of catching fish is getting decreased Life of many stakeholders in the area is below poverty line.

Coastal zone has 3 different sorts of water, including fresh water, brackish water

and brine.

Brine surface: set for the exploitation of sea products Some main sea products are

bream, Chinese herring, Khoai fish, grey mullet, Vuoc fish (perch), Van shrimps, Bop shrimps, and pawns At present, the seafood catching activities are natural and being carried out on small-scale A majority of aquatic products are used in processing traditional

lines such as fish sauce, shrimp paste and seafood.

Area of brackish water surface: Being mainly available in the Red, Thaibinh and

Traly river mouths thanks to an abundant source of short-lived creature, algas and aquatic botany as natural food used in process of breeding aquatic products Thaibinh province has about 20,705ha (Tienhai district has 9,949ha and Thaithuy district 10,756ha), of which

34

15,839ha is able to breed brackish water products (Tienhai 7,179ha and Thaithuy 8,660ha), including 10,386ha of tide-water region and 5453ha of low productivity me, transplantation salt-land likely being used for breeding brackish water sea-products At present, about 3,629hais tapped for breeding shrimps, crab, arca, mussel and gracilaia,

Fresh water region: The total area of aquatic products breeding is about 9,256ha, of which 6,020ha has been exploited for breeding Besides, more than 3,000ha of low productivity hollowed.

In brief, the estuaries of Red River Delta offer good conditions for aquaculture as follows:

= Water available for aquaculture development i large, estimated at over 1,000 ha, - Natural food sources are abundant, natural seed stock, particularly sheimp, is diverse in species composition

- High tide level assists in the supply and drainage of water a

natural food and seed from the sea and tothe sanitation of the rearin

so the reception of ponds.

= The mangroves in costal zone help protect aquaculture ponds and contribute to the supply of aquaculture seed (crabs, shrimp and certain species of fish) and feed (molluscs, trash fish, small mangrove erabs ete.)

the early 1980s, the aquaculture farming for export in Red River Delta has been encouraged and promoted by the government Furthermore, a high economic return leads tothe widespread practice of this lucrative activity.

‘There are many districts in coastal zone convert of salt fields, intertidal areas and ‘mangrove forests into aquaculture ponds with highly profitable, at least inthe short term,

Fig L7 Distiets along the Coast Having Aquaculture Production

Basing on different conditions about topography area, tidal regime, salinity concentration and so on, the sort of aquaculture species and pond size in each regional area is diferent.

‘Table 1-4 Aquaculture Productions in Districts along The Coast (Souree data: 2002) 3s

Location Sortof Aree

This study is an attempt to describe the effect of Hoabinh Hydropower Plant and Sonla Hydropower Plant (going-on construction) to salinity intrusion in Red River Delta nd the changes of the flow characteristics of the lower Red River Delta in time and space at present and future condition

In order to archive the above requirements, the mathematical model of MIKEII is used to evaluate the characteristics of the freshwater flow and salinity intrusion based! on the recent observed data, The result will be estimated under the conditions of sea level rise due to the Greenhouse Effect

The objectives ofthis study are as follows:

‘To estimate the longitudinal dispersion coefficients at different braches inthe Red river delta at present

To assess the effect of Hoabinh reservoir t0 salinity intrusion in condition with or ‘without the reservoir.

To assess the future-effect of Sonla reservoir to salinity intrusion ‘To forecast characteristics of flow and salinity intrusion in the Future 1.4 Scope of Study

The scope of this study is to use numerical model MIKEI1 to study the characteristies of salinity intrusion in the estuaries ofthe Red River System,

“The upstream boundaries of study area is stations in Yenbai, Hoabinh, Vuguang, Phalai and downstream ones are stations in the nine estuaries: Day, Ninhco, Balat, Trủy, “Thaibinh, Vanuc, Lachtray, Cam, Dabac.

36

Chapter 11

LITERATURE REVIEW 3⁄1 Theoretical Study of Dispersion Coeff

2.11 Mathematical Formulas

nt and Sali Intrusion

For many years, a number of systematic attempts have been made with more or less success to correlate the intrusion of saline water with tidal characteristics on the basic of actual observations of salinity condition in the estuaries.

TAYLOR (1935) developed the turbulence theory and used the statistical approach, 40 formulate the dispersion coefficients forthe case of two-dimensional motion as follow:

Dh persion coefcens in x andy dtections cst he velo Macaton nan deen Ra, incauo-artlation of elon andy dieons.

‘A requirement is that the velocities be measured according to the Lagrangian standpoint, However, actual data for velocity are normally obtained by measurements taken at fixed points, that is, they are expressed in the Eulerian point of view Therefore, the TAYLOR theorem cannot be applicable to the data available in most cases, A ‘wansformation between the Eulerian and Lagrangian description of velocities was made by HAN and PASQUILL (1957), WADA et al, (1975); they suggested that the dispersion coefficient can be expressed as follows

D, = Bu, 23)

D,=Bv eT, 4) where

wp, Vp: the Eulerian velocities fluctuation in x and y direction

a dimensionless parameter depending upon the scale of turbulence.

‘Ty : the Eulerian time scale,

KETCHUM (1951) has presented an approach to the steady state salinity intrusion problem based on dividing an estuary into segments whose lengths are equal tothe average excursion of a particle of water during the flood tide, Complete mixing is assumed within neat at high ride, and exchange coefficients are based on this assumption, As a result of the complete mixing assumption this method is limited to steady-state studies of estuaries where the well mixed condition is approached Estuaries of this type are characterized by very large rations of tidal prism to freshwater discharge and are a rather limited class as compared to the partially mixed estuary.

ARON and STOMMEL (1951) have proposed a mixing-length theory of tidal ‘mixing as a means of treating the time average (over tidal cycle) salinity distribution in & rectangular estuary The one-dimensional conservation-of-salt equation was employed with a convective term forthe river flow and horizontal eddy diffusivity The latter is assumed to be equal to the product of the maximum tidal velocity at the estuary entrance, the tidal excursion length, and a constant of proportionality By integrating the conservator

Fa

‘equations, 4 family of salinity distribution curves is obtained in terms of the distance along

the estuary divided by the total length of salinity intrusion, The results are primarily useful asa classification of estuaries by means of a “flushing number” oblained by a best fit of field salinity measurements with one of the family of curves.

They used the steady state model to study the problem of salinity intrusion equation

~ø, đề - +Íp 43) es

ae dr ae)

s76 26

‘TAYLOR (1954) established that the longitudinal dispersion ina long straight pipe may be characterized by a one-dimensional dispersion equation, in which the diffusive and convective process occurring throughout the cross section interact to produce a longitudinal dispersion coefficient:

D=100au, en in which:

a: radius ofthe pipe vu: the shear velocity

‘This result was probably the best known as well as the simplest of all equations describing turbulent dispersion,

FURUMOTO and AWAYA (1955) proposed a numerical model to calculate salinity intrusion in tidal estuaries by mean of transforming the independent variable x in the advective-dispersion equation into the storage volume V, They oblained the longitudinal distribution of the dispersion coefficient in the estuary based on the qui steady transformed dispersion coefficient equation with the aid of the observed S-V relationship and fresh water ilove

THOMAS (1958) applied TAYLOR’s concept to flow in an infinitely wide two «dimensional open channel in which the flow is described by a power-law distribution He ‘obtained a complicated functional relationship between dispersion and Reynolds number.

ELDER (1959) duplicated THOMAS's effort, for assuming a logarithmic velocity profile, obtained a remarked simple result

D=5.93hu 28)

in which his the depth of flow.

PRITCHARD (1959) presented a mathematical model representing the variatio of salinity concentration from tidal eycle to tidal eyele:

38

De: time-averaged over a tidal-cyele of dispersion coefficient.

Dy was obtained by integrating the steady state equation corresponding to Eq (2.9)

9, Xí ^ 2.10)

The integration (2.14) with respect tox yields:

-0,5=p/45 ean

By fitting data to Bq (2.15) D', could be obtained.

IPPEN and HARLEMAN (1961) used the steady state model and analyzed the results of salinity intrusion experiment in the tidal plume of the Waterways Experiment Station (WES) to show that

D, the dispersion coefficient a station x at the low tide D) : the dispersion coefficient at x=0 at low tide,

Xx: O atriver mouth

[B the distance seaward from x=0 to the point where S=So at low tide,

he parameter Do is found to be correlated with a stratification parameter Gi, where:

G _rateofenergydissipationperunitmassoffiuid _ (2.12)

J rateofporentialenergygainedperunitmassoffluid

HARLEMAN and ABRAHAM (1966) re-analyzed the WES data and found that the stratification parameter G/J was related to another parameter called “estuary number” ED They formulated the following correlations:

tidal ism, dened as the volume of wate Fp densinetric Froude numer.

a G16

ug: maximum tidal velocity, 1h: depth at the ocean velocity

‘Ap: change of density over the entire length of the estuary, Q¢: fresh water discharge.

39

Ts tidal period,

STIGTER and SIEMON (1967) used the unsteady state diffusion equation tò ‘study the salinity intrusion in a constant width representation of the Rotterdam Waterway ‘The unsteady state diffusion equation

2 (as) S(as)-2 (0.4) aun)

They appisd boundary condiions repeating fom tidal cyle to tidal cycle, thus reiting alo t cgeaing me tuying alaty iatabulon The dspesioncocsint Was assume to be in form:

“The value of ‘condition for salinity.

FISCHER (1966-1968) made an important step in the development of methods for predicting longitudinal dispersion coefficient in natural stream based on Taylor's theory, He presented two ways of predicting a dispersion coefficient for a natural stream: the Method moment and the Routing method His methods required field measurement of channel geometry, concentration and cross-sectional distribution of velocity,

\ any instant of time was determined by using the ocean boundary

‘The method of moment is based on the equation:

the variance ofthe concentration distribution with respect to distance along the

the variance of the concentration distribution with respect to time, measured at ‘fixed point in the stream,

ithe mean velocity ofthe flo.

7: the time of passage of centroid af concentration

Subscripts 1 and 2 reler to the two measuring stations.

Inthe Routing method, a value of Dy is assumed The validity of D may be tested by the beginning with a measured concentration curve at a particular time, applying the theory (o predict a concentration curve at the same later time, at which one Was actually measured The comparison between the observed data and routed results demonstrates the validity of the predict dispersion coefficient

BOICOURT (1969) wsed the approach of Prichard to study the salinity of Upper ‘Chesapeake Bay He obtained the dispersioncoefficient by integrating equation

222) where

KT is a constant dus measured data

‘isthe cross-section area,

‘They derived the longitudinal dispersion coefficient, which was assumes constant during a time period:

the lime period of (2241) and could be computed from

" Aidt

M total mass of salt

‘The value of Q, A, S are measured at station.

THARCHER and HARLEMAN (1972) improved the model used by Stigter and ‘Siemons (1967) They extended the problem to transient boundary condition and proposed 4 formula in which the dispersion coefficient varied with time and space:

LL length of estuary from sea entrance to head of tide So: salinity of sea water

‘Thatcher and Harleman found that the dimensionless param ‘well with the estuary number inthe following form

the deviation of velocity from the ro 4; the traverse mixing coefficient

tional mean velocity LIU (1977) also suggested a similar equation as Equation (2.28)

p,=p2, 229)

a coefficient

Q: the discharge ofthe river

RR: the hydraulic radius

LIU deduced an expression to estimate the coefficient

YVONGVISESSOMJAI, ARBHABHIRAMA and APICHATVULOP (1978) formulated a mathematical model to investigate the effect of upstream fresh water discharge and tidal conditions on the salinity concentration and intrusion length along the (Chao Phaya and the Mae Klong rivers The dispersion coefficient expression suggested by ‘Thatcher and Harleman was used in this mode inthe following form:

KI and K2 are coefficients to be calibrated These coefficients were varied until the model reproduced the observed salinity conditions, and the investigators found that

‘© KI is equal to 600 (mJs)(ppukm) and K2 is equal to 400 (m)/5)(ppUkm)

are appropriate for the Chao Phaya river

‘© KH is equal to 100(mŸ9)/(ppUkm); K2 is equal to 200(m1/9)/(ppUkm) forthe ‘Mae Klong river.

PENPAS (1979) showed that D, boing @funetion ofthe product [28 x]

632) PRANDLE (1981) analyzed the measured data from eight estuaries and shows thai these data could be fitted reasonable well with each of three expressions the dispersion

‘The following are some popular numerical models of salinity intrusion that are ‘mentioned in many references

stuary Model (FWQA) 4) Hydrodynamic

4

FWQA is usually called as ORLOB following the name of Dr Geral T Orlob, ‘This model was to be used in actual eases, Both set of Saint-Vanant equations and

dispersion equation are solve with a consideration of tidal effects The first pplication of FWQA was Sacramento-San Joaquin, California,

b) SALFLOW of Delf Hydraulics

SALFLOW (1987) is production of cooperation between Hydraulics Institute of Netherlands and Mekong Committee It is one of the newest achievements in ‘numerical salinity intrusion model.

‘Test model in Netherlands achieved good results and doing apply in Mek delta

In addition, there are modules of salinity intrusion in some hydrodynamic model in recent year as ISIS (English), MIKEI L (Danish) and HEC-RAS (US) but have not applied in Vietnam.

2.13 Salinity Intrusion Study in Vietnam.

Salinity Intrusion in Mekong Delta Project (Southern of Vietnam) in 1980 under the Mekong Committee assistance promoted the research of salinity intrusion in Vietnam, Within the framework ofthis project, some of saltwater and salinity intrusion models were found by Mekong Committee and Institute of Water Resources Planning and Institute of ‘Mechanics These models are used in eesearch of Mekong delta planning, in estimate effect of anti-salinity-intrusion constructions to enlarge erop area in the dry season as well as

prediction salinity intrusion, These models have important contribution to study of salinity

intrlslon in Vietnam,

On contrary, esearch of salinity intrusion in Red-Thaibinh delta is mentioned less than The following are some previous study

‘VI (1980) by analyzing the data recorded at the stations in the Red river system sates that in dry season the intrusion length of salinity at some branches of the Red river system may be longer than 30 km; also the freshwater discharge and the slope of salinity intrusion was not present due to the large amount of freshwater discharge from upstream,

‘THUY (1985) studied the characteristics of tide inthe Red River estuary He found that the tidal properties vary greatly from the rainy season to dry season and the

‘predominant components of tidal waves are diurnal

‘THUY (1987) applied a numerical model to study the flow in the river system during Mood and dry season He found that in dry season, tidal waves could propagate ‘more than 100 km upstream along main branch of the river system,

PHÚC (1990) used ID numerical model with much success However in the ‘model, te effects of density differences were not considered The data used for calibration ‘were limited and the verification of the model was not possible Moreover, data were used in the model such as datum of all station was not possible to bring to standard altitude; ross sectional areas of river system were not measured atthe same Yeurs Thus the results ‘were very limited,

NGO (1991) based on the recorded data of salinity concentration at stations along, estuaries of the Red River System has draven some primary remarks on the characteristics of salinity intrusion there Details of salinity intrusion in each tributary ofthe iver network ‘were not investigated,

4B

UY (1992) applied a numerical model to determine the dispersion coefficient for the prediction of salinity intrusion in the Mekong estuarine network He found that dispersion coefficient varies in the same manner as those of salinity concentration.

CA (1996) based mainly on two previous publications by Vu (1990) and Vu et al (1991), Using many year recorded data of salinity concentration at stations along the cestuaries, monthly-average salinity concentration at each estuary is computed The salinity intrusion length in each estuary was also estimated Details of salinity concentration distributions along the estuaries were studied using a numerical model of the transport and dispersion of salinity He found that in the dry season, the salinity intrusion length is as long as 20 km in the mai river and mie than 20 km for some tributaries Inthe main river and tributaries with high freshwater discharge, the maximum salinity concentration is ‘observes in January while for the tributaries with low freshwater discharge, the maximum salinity concentration is observed in March,

HUNG's study (1998) of saline intrusion in the Red river delta has been also limsted by data used for calibration of the model and the dispersion coefficients were not accounted for the saline gradient along to estuaries also the verification was not carried out,

AN NIEN, NGUYEN (1999) has summarized studies relating to saline intrusion in Vietnam and has pointed out that at estuaries the salinity is in the range of 22-28ppt The saline inưusion length inthe Red river delta is not so long The distributaries connected to the open sea is at acute angle thus bands affected by salinity are narrow with the width of

‘The above studies are the first studies in some rivers without consideration of whole river system.

23 Salinity Control Requirement for Irrigation and Aquaculture

Control of salinity concentration is primary importance in development of aquaculture in coastal zone as well as water intake to ierigate for crop fields in the dry According to Water Quality Standards (TCVNS943-1995) and Quality Criteria of Water for Aquatic Life (2§TCNI71-2001); (28TCNI91-2004), salinity concentration is required for water intake into paddy fields and aquaculture ponds as followings:

= Gate of weirs under the dykes can be opened to intake for rice seeds fields while

salinity concentration is 1g/l With growing-paddy, maximum salinity concentration is allowed in 4g/1

- The procedure for intensive culture of Tiger shrimp assign that salinity

concentration of shrimp ponds as well as for nursery of shrimp from postlarvae 15 to post-larvae 45 is from 10 to 30 (past per thousand) (the best range: 1Sppt-25pp0)

= River water ean be used for men and livestock with salinity concentration is

‘The characteristics of flow and salinity intrusion in the rivers are governed by three predominant factors, which control the magnitude and direction of current at different {depth and at different distances from the estuary month, The three factors are

4) The effect of tide in generating the tidal current and turbulence;

4i) The effect of upstream discharge of freshwater in producing a net seaward

transport; an

iit) The effect of gravitational forces due to density differences between the ‘upstream freshwater and downstream seawater o the sediment.

The rise and fall of the tide at the mouth and the associated exchange of water ‘masses through the entrance result in the temporary storage of large amounts of sea water in the estuary during high tide and the drainage of this water seaward dung low tide The total volume so exchanged is known as the tidal prism, which, for a given estuary, is variable only with tidal amplitude.

OF significance in relation to this tidal prism is the continual inflow of fresh water from upland sources which results in a volume of fresh water equal to discharge rate totaled over the tidal period While this discharge rate may vary slowly with time, the ratio of fresh-water volume to the tial prism has proved of value in the general classification of ‘The type of an estuary depends on the ratio of volume of seaward freshwater Ilow in a tidal cycle and volume of the tidal prism which govern the mixing of the estuarine ‘water, The volume of seaward freshwater How in a tidal eyele is

‘A: cross-sectional area at the estuarine mouth, Up: seaward freshwater flow velocity

‘a= “4: mean tidal velocity in a half tidal eyele

uu: maximum tidal velocity Tidal period

3d ‘Stratified Estuary

‘When the seaward freshwater flow is large during rainy season with respect to tidal flow, 0,/ P21 the fresh and salt water remain separate The seaward fresh water flows out 40 the sea over the top of the salt water wedge The length of this salt wedge depends on the depth of water, the fresh water discharge, and the density differences between the salt and fresh water There is no reverse flow atthe upstream station and that flows at all depth

45

are in the seaward direction, At the salt wedge, the upper fresh water flow is seaward but the lower sea water is reversed in direction.

3.1.2 Partially Mixed Estuary

‘When the ration of volume of seavard fresh wate flow in a tidal cycle and volume of the tidal prism, Eq, 3.3, is in the range of 0.2 to 05, the estuary type is partially The tidal velocity is strong enough to produce sufficient turbulence to induce mixing horizontally ‘and vertically between fresh and salt water, therefore, there exists no distinet interface of fresh and salt water However there exists gradient of salinity in the vertical direction due to the partial mixing

Well Mixed Estuary

‘When the seaward fresh water flow is small, during the dry season, with respect to tidal low, Q, /P<0.1, the fresh and salt water are well mixed due to strong tidal action.

wer this condition, there exists only a litle gradient of salinity in the vertical direction and salinities decrease progressively from the sea water atthe mouth of the estuary.

In these four estuaries, stratification is strongly season-dependent resulting in partly stratified conditions or well-mixed inthe dry season when the dominant stratifying is tidal Staining, tidal advection, and bed-generated turbulent mixing The rainy season is characterized by stratified conditions when estuarine circulation and advection of ratification by tidal currents and river flow are the main stratifying mechanisms,

In the dry season, almost of estuaries in Red river delta is partially mixing with the ration between Volume of freshwater ina tidal cycle and tidal prism as Follows

‘Table 3.1 Classification of Mixing Type of Main Braches of Red River Delta

No | River Gly | Uotins)| QP | Mixing type

1 [Bata T011 [0a | 079 | Parilly mixed

(Source of data: Center of Estuary and Coastal Engineering)

‘Therefore, applying one-dimension model to evaluate the salinity intrusion in rivers ‘of Red river delta is appropriate

32 Numerical Computation

Salinity intrusion model consists of two portions: the tidal dynamics portion and the salt balance portion The tidal dynamics portion is described by two equations namely.

4 Continuity Equation

46

B: width atthe water surface ofthe siver cross section, including specified storage far each segment (ml.

‘A: c1oss section aa [m"}

1; Manning's roughness coefficient

S salinity concentration at time t's"

us ida velocity [m's"]

Q discharge [m's")

De: dispersion coefficient,

4 lan flow {ms}

Si itera salinity concentration

Both the tidal dynamics and the salt balance equations are solved using implicit Finite difference method The domain, the x1 plane, is discretized into rectangular grids of size Ác by Ar Value of water level (H), discharge (Q), and salinity (S) at all grid points are to be determined

Firstly, the water level and discharge along the river are determined from the tidal dynamics Secondly, using these water level and discharge values the salt balance equation is solved to give the salinity along the river.

‘The general outline of the MIKE 11 is used to compute the tidal hydraulies and salinity intrusion in the Red river deta is as follows:

+ Use the finite difference formulae with Abboth and loneseu 6-point implicit scheme

+ Use the interpolate formulae to determine the intermediate grid points

3.2.1 Finite Difference Tidal Hydraulics Equations Continuity equation — h centered

= Momentum equation ~Q centered

OF +0)

a a a2)

4

Fig 3.1 Staggered Meshes in Space and Time

32.2 Finite Difference Salt Balance Equations

‘Tobe simple, lateral flow is clinsinated An arbitrary six point scheme constructed

& and term of salt balance equation

a os

Fig 36 show how the weighting coefficients are assigned The weighting coefficients bby means of weighting factors and applicable to

gt tm=l

HN tsi)valsp!+s/)+ gi esi]

nan an: G9

Fig 32 Arbitrarily Weighted Six-point Compatational Molecule 33 Model characteristics

On unsteady salinity intrusion ina tidal estuary is solved using ice equation while the tidal dynamics are solved using a system of one-dimensional, ‘unsteady continuity and momentum equations based on Manning's roughness coefficient

Both tidal dynamics and salt balance equations are solved using finite difference ‘methods for implicit schemes The tidal dynamics are calibrated by varying Manning's roughness coefficient until the model produces the water level and discharge relationships ‘observed in prototype.

To develop and calibrate the hydrodynamic and salinity intrusion model, the ‘mathematical model has considered the following basic assumptions

1 The model is one-dimensional, and that is all quantities vary along the longitudinal axis of the estuary.

2 Vertically well-mixed salinity condition exists,

3 The seasonal lateral ouvinflow from groundwater is negligible

dimension salt

MIKE II is ä professional engineering software package for the simulation of lows, water quality and sediment transport in estuaries, rivers, iigation systems, channels, and other water bodies MIKEI is based on an integrated modular structure with a varity ‘of basic modules and add-on modules, each simulating certain phenomena in river systems MIKE 11 includes basic modules forthe following

Rainfall - Runoff Hydrodynamics

Advection-<ispersion and cohesive sediments Water quality

[Non-cohesive sediment transport

‘The modules used for this study consists of hydrodynamic (HD) and adveetion-dispersion (AD) modules The application the MIKE 11 model in this study takes place in

two steps:

‘© Computation of the river flows and water level by MIKE 11-HD

‘© Computation of the river salinity concentration using MIKEII-AD Chapter 1V

DATA COLI ECTION “The collected data involves geometry data of cross

and salinity data at stations in the Red-Thaibin River System in 199:

ections, hydrology, hydraulic 2002 and 2003 4a Geometry

‘Two sets of geometry data in 1993 and in 2000 along main branches of whole Red River Delta are collected The accuracy of both of topographical data is very high and reliable In this study, geometry data of cross-section in 2000 is used,

49

Hydro-topographic measurement was carried out in 2000 for 634 cross-sections ‘They cover river reaches of 1,086 km The distances between cross-sections are about Tkm to 5km,

Fig 4.1 Spatial Distribution of Available Cross-Section in Red-Thaibinh River System 42 Hydrologiealand Salinity Dat

ata of water level, discharge and salinity concentration in the dry season mainly from December to April of the next year at the stations along the main branches of the Red river and Thaibinh River in 1993, 2002, 2003 are collected Table 4.1, 42, 4.3 shows the location of stations, the measured components of each set of data, respectively Station river networks of each year above are displayed in Figure 4.2, 4.3,4.4

Measured data in 1993 for Red river basin are very good for hydrodynamic simulation, However, the measurement of salinity concentration was focus on Bulat estuary which is the main branch of Red river basin, Salinity concentration along the other branches as Day, Ninhco, Traly was not measured or measured at only one station.

Therefore, these data are not enough to calibrate and verify for model

Measured data in 2003 is the newest data, However, the measured campaign was limited in Red river basin from only Hanoi station to the sea Hence the change of "upstream inflow of Red River Delta cannot study

‘Measured data in 2002 in all of Red River delta is the most adequate to requirement of this study The data are both homogenous and long enough in order to be used in ‘numerical computation of salinity intrusion For this reason, hydrological and salinity data from Sth to 20th March, 2002 is used in this study.

‘Measured campaign was carried out in the same time at all rivers from upstream to Gulf of Tonkin Salinity was observed at every hour during high tide and every bihourly during low tide, At every station, salinty is measured at three depths (surface, middle and bottom) Fig 49 illustrates the salinity concentration at three depths at Balat(km0) station

sọ

Salinity concentration tends to increase from surface to bottom of water depth Water levels are measured at

happens when water level is also the maximum.

‘Table 4.1 Collected Data in 2003 for Numerical Model (Time of measurement: Feb/1/2003 to March/27/2003)

ety hour Fig 4.10 presents measured water level and salinity concentration at Balat(km0) station in the Red River The maximum salinity concentra

Fig 42 Locations of Gauging-Stations in Measured Campaign by The Year 1993

Table 4.2 Collected Data in 2002 for Numerical Model

N ‘Measured

No] Rher EP | ot si Coxe | °0EAEMGI Postion | Conponens

[es Tangtie[ [aiue | H | Q S 1 [Da 40103 | 47 Tos Bini Tf toss | ãmp ff

2 thao | 25658 | 3 | Yen B: 2 | is [21a |»

3 la -44500 | 15 Và Quang 3 [tosis [aise |e |= 4 [ Booag —T H0 | 4L Thuong Cat + 8s [aro f= f=

‘ho Tan we | 10606 | 2000 |*[ }+

Tien Hoang | Mới | mm | +] | + s | Day 126300 | 69 | Doe Bo 16 | tosas | 3015 | | +

inh Bình 1s | 10558 | 2016 | *

Phu Ly 18 | 1055 | ano | + | + Phu Le ar} 6e | anos |[*[ + 6 | Nhho | 44000 | 36 | Bian ta m0 | roar | 3 |*| |

‘Trac Phương w | tosis | 2019 | «| 7 |NamDnh | 28800 | 9 | Nam Dink 13 | I8 | 30% | * | =

BaLat (kad) | tosas | amis | |

1 [Tae | 69050 | 33 | Trew Duong s | 10607 a

1í [How 18000 | T9 Quan Khai 2910638 =

Dong Xuyen 2 | 10638 :Ƒ

Cat Pháo han a7 | 10620 +] is 12 | thai inh | 92100 | an | Ch zy ee "nh

Pha Lat 2s | 10617 ae TF [Ga 2900 | 5 _[Ba Na Trang Trang) 34 [106.27 :

14 [Mai 1500 | 2 Tien Tien 3 | 10631 aXenh Khe(Quang Pius) 36 [106.2 =

lan 35600 | 20 trang Trang 35 | 10629 *|rị*

16 | Mi 10 | 317 [tai Va | 2700-13

18 | Lach Tray 42100 | 28 | Kien An 3P | I6 | má [+ |= 30 | Kinh Thay) 37200 | 25 Ben Binh 30] T63 | 2198 | [= =

21 | Han T900 | 5

Ximang Hoang Tach 31 rosa | nas ff

22| DaBae 18800 | 12 | Don Son 33 | tosas | 2100 | +] |

Do Nghị 33 | 10646 | oss [+ | +

Tay 3000 [+

| Rink Mow 3000 | 16 | An Pha 3% ee | ma pep

‘en Pha Binh án | tosai | ms ef

34 | cam 20700 | 13 CaaCem 40 | 10637 | as2 | ©

Cao Kenh 39 | toss | mse | + | + Toa CES

3

‘Table 4.3 Collected Data in 1996 for Numerical Model

(Time of measurement 12 to 11/23/1993 and 11 10 3/17/1996)

Geographical Postion | eared

No) River | Coie] Sion — a] Rais |g s

tions in Measured Campaign by the Figure 4.4 Locations of Gauging

56

43 Based Maps

1 Based map of Red River - Thaibinh River in ArcView of sale 1:250000 with database management system on topography, hydrology, transport system, communes and land use for Red-Thaibinh River are collected to use in application of this study.

Fig 45 Based Maps in ARCVIEW of scale 1:250000

2 Based map of Vietnam in MAPINEO environment including several basic map layers: topographic, geologic, administrative boundaries, basin boundaries, river and stream network, transportation network, hydraulic works systems, ete They are used for determination of gauging-stations are showed in Fig 4.6, 4.7, and 4 8

7

58

Fig 4.8 Map of Salinity Intrusion Observation Stations

Sty Concentration athe depth we: Sora Mle si to,‘ah Son

Fig 49 Salinity Concentration at Three Depth of Water at Balat (km0) Station

9

Salinity Concentration and Water Elcvdlon ta Lak) Staton

Fig 4.10 Salinity Concentration and Water Elevation at Balat (km) Station

40

33.1 Hydrodynamic Module

‘The HD module contains an implicit, finite difference computation of unsteady flow in rivers and estuaries, Within the standard HD module, advanced computational formulations enable flow overa variety of structures to be simulated

The hydrodynamic model takes into account the main rivers and eanals in the area, The connectivity of the river systems and influence of other rivers outside the model area is identified from observed data The river slop and flow direction is computed in the model by considering the cross sections The flood plain depression within the model area is defined as flood cells or storage cells, which have been connected to the main river through the link channel, The runoffs from catchments of the region are directed to the hydrodynamic model either as lateral inflows distributed over a reach or as point inflows ‘The area to be attached at cach point or reach has been determines examining the detailed drainage pattern, established with the help of vopo maps.

332 Advection-Dispersion Module

the AD module is based on a one-dimensional equation of conservation of mass of 4 dissolved or suspended material, ic the dispersion equation The advection-dispersion equation is solved numerically using an implicit finite difference scheme, which hay negligible numerical dispersion An AD simulation is carried out on the basis of resulls from the hydrodynamic model

The movement of salt in the model depends on flow velocities and the amount of ‘mixing, which oceurs with water of differing salinity Water level and discharge ‘boundaries of hydrodynamic model must be specified as either open or closed for salinity

Times series of salinity concentrations needs to be specified in open boundaries, ‘where salt can enter or leave the model In closed boundaries, itis assumed that no net transport of salt occurs, Usvally downstream water level boundaries are open concentration ‘boundaries where as upstream discharge boundaries ae closed,

“The program for salinity intrusion in an estuaty is describe in Fig 34

41

34 Statistical Criterion

For the evaluation of the performance of the mathematical model compared preferably against measurements, some statistical parameters are often used in order to ‘obtain some quantitative measure Tor the deviations

34.1 Efficiency Index or Coefficient of Efficiency

Efficiency index can be used for the measuring the efficiency or accuracy of the

SSE is the sum of square errors (sim of squares of the differences between observed and calculated vale)

SSE =x, -¥)"

‘iis observed (measured) data at tie

¥ is mean value of observed data ữ Sx

‘Yi computed (predicted) data at time ï N numberof datapoints

342 Standard Deviation (s)

Y mean value of computed data ir

343 Root Mean Square Error (RMSE)

Root mean square error is used to compare the performance of two or more ‘models when used for the same data set.

RMSE = 614)

4B

344 Correlation Coefficient (R)

Correlation coelficient isthe ratio if the covariance between X and Y to the product of thei standard deviations The correlation coefficient must have -1R 1 ‘The value R = 1 implies a perfect linear relationship with a positive slope, while & value of R = -1 results from a perfect linear relationship with a negative slope As the covariance gets smaller relative to the variance, so does the correlation coefficient, approaching zero If the variances are very small, then the covariance can be smaH and there will be a strong relationship The coefficient of determination is also defined as R° Both the correlation coefficient, Rand the coefficient of determination are not well suited for model performance evaluation,

A worldwide rise of sea level is among the predicted consequences of the “Greenhouse Effect’ the global warming expected as a result of the accumulation in the Earth's atmosphere of carbon dioxide and other greenhouse gases generated by industrial and agricultural activities It has been suggested that increasing concentrations of these ‘gases will lead to a rise of between 1.5" and 4.5°C in mean atmospheric temperature during this century

‘Such an increase will cause expansion of the volume of near surface ocean water, and partial melting of snowfield, ice sheets and glacier, releasing water to augment the ‘oceans, thereby producing worldwide sea level rise (BATH and TITUS, 1984).

There have been various scenarios for the scale of this sea level rise, Analyses carried out by the Environmental Protection Agency in Washington led Hoffman to predict that global sca level would rise a meter during the next 50 to 150 years, with the most likely scenario of an accelerating sea level rise attaining a meter a century.

‘Sea level rise scenario based on the prediction from the Climatic Research Unit, fersity of East Anglia, that global sea water level will stand to 18 em higher by the year “2030 and the prediction from the Environmental Protection Agency, Washington, that it will rise Im by the year 2090 as shown in Figure 4.2.5 (BIRD, 1989)

AA global sea level rise of one meter would greatly, modify coastal environments, producing erosion and submergence, especially on low-lying sectors Such a rise will tenable the highest tides to reach levels of at last 1m above their present limits, allowing for

a possible increase in tide range as nearshore water deepen

Arising sea level will result in submergence and widening of the mouths of rivers ‘with increased penetration by salt water, which may invade underground aquifers A sea level rise will also enlarge and deepen lagoons, causing erosion of fringing swamp areas,

44