**2.3 Dye tracer experiments**

Hydrodynamics and pollutant discharge dispersion characteristics are determinant factors in river basin planning and management, where different waters uses and aquatic ecosystems protection must be considered.

Net advection and longitudinal dispersion play important roles in determining transport and mixing of substances and pollutants discharged into the aquatic systems. In order to enhance water sources protection, the knowledge of transport processes is of increasing importance concerning the prediction of the pollutant concentration distribution, particularly when resulting from a continuous or accidental spill event caused by industrial and mining activities or road-river accidents.

Generally, there are two approaches to calculate the transport of solutes in water bodies. One is the more classical calculation based on exact river morphological and hydraulic input

A Hydroinformatic Tool for Sustainable Estuarine Management 13

(Caplow et al. 2004) as an alternative method to dye tracer experiments used for advection

MASS MONITORING

 S1 – S2 0.526 Var. 2:37 2:35 14 10 57 **3 rd.** S2 – S3 0.497 Var. 2:41 2:41 51 45 56 (Nov.-90) S3 – S5 0.473 Var. 3:21 3:19 37 35 55 S1 – S3 0.511 Var. 5:18 5:16 34 - - S1 – S5 0.497 Var. 8:38 8:35 35 - - **1 st.** S1 – S2 1.105 Var. 1:14 1:14 52 40 62 (Dec.-89) S2 – S3 0.949 Var. 1:24 1:24 61 70 62 S1 – S3 1.023 Var. 2:38 2:38 58 - - Table 1. Hydraulic and dispersion parameters estimation using tracer dye experiments in a

The dispersion processes in rivers are combined with a specific dynamic characterized by a decrease in maximum dye concentration (Fig. 12). The distribution of the tracer in all directions follows the sluggish injection into the channel. In non-tidal rivers, the lateral and vertical dispersion processes are almost always faster than the continuing longitudinal

TRAVEL TIME (h)

EXPER. DUFLOW EXPER. DUFLOW EXPER. DUFLOW (%)

DISPERSION COEFFICIENT (m<sup>2</sup>

**DECEMBER-89 SAMPLING PROGRAM**

**/s - flood situation)** 

**(Flow=140 m<sup>3</sup>**

s -1) RECOVERED

AVERAGE VELOCITY (ms-1)

and dispersion characterisation.

PROGRAM REACH

non-tidal reach of river Mondego

**0,00 0,10 0,20 0,30 0,40 0,50 0,60**

**8:00**

**8:15**

**8:30**

**8:45**

**R=0,93**

**9:00**

**9:15**

**9:30**

**9:45**

**10:00**

Fig. 12. River Mondego model calibration: correlation between field tracer experiment data

One-dimensional modelling is a reasonably reliable tool to be considered for estimating the distribution of solutes in large rivers. Complex processes, for example in dead zones or downstream from the confluence of two rivers, have to be investigated by direct measurements and should be described by two-dimensional transport models. Calculation of net advection in tidal rivers is fairly straightforward, but longitudinal dispersion is difficult to determine *a priori*, and the application of two or three-dimensional transport

**R=0,98**

**10:15**

**Time (h)**

Model Results Site 1 Site 2 Site 3

**10:30**

**10:45**

**11:00**

**11:15**

**R=0,97**

**11:30**

**11:45**

**12:00**

**12:15**

dispersion process.

**Concentration (**

and model results.

models are often required.

 **g/L)**

data and the other is the calculation based on estimation of transport parameters such as travel time and dispersion coefficients. Since exact morphological data are often unavailable, the parameter estimation technique is more promising.

In both approaches, tracer experiments are needed to provide field data for water quality models calibration and validation procedures. Indeed, model calibration is often a weak step in its development and using experimental tracer techniques, the calibration and validation problems can be solved satisfactorily, improving the needed feasibility of the early warning systems used by many water supply utilities.

Tracer experiments are typically conducted with artificial fluorescent dyes (like rhodamine WT) (Fig. 11), whose concentrations are easily measured with a fluorometre. These tracers should be easily detected, non toxic and non-reactive, as well as, have high diffusivity, low acidity and sorption for a quasi-conservative behaviour.

Fig. 11. Rhodamine spreading after their injection in a river Mondego reach

Based on field experiments data, many investigators have derived semi-empirical equations (Hubbard et al., 1982; Chapra, 1997; Addler et al., 1999) or applied one-dimensional models (Duarte & Boaventura, 2008) to calculate experimental longitudinal dispersion coefficients from concentration time curves at consecutive sampling sites, using the analytical solution of first order decay kinetics (Table 1).

The injected tracer dye mass must be calculated considering the water volume estimated in the river reach or reservoir system and the fluorometre detection limit. Specific problems of the application of tracers to surface water researches include the photosensitivity of dyes, such as fluorescence tracers, and recovery efficiency, which may imply the use of correction techniques for tracer losses. The tracer mass recovered at each site allowed the assessment of the importance of physical and biochemical river processes by quantifying precipitation, sorption, retention and assimilation losses. Usually, total tracer mass losses resulting from all these sinks can reach 40 to 50% of the injected mass (Duarte & Boaventura, 2008; Addler et al., 1999).

In some recent experiments, a gas tracer (SF6) has been shown to be a powerful tool for examining mixing, dispersion, and residence time on large scales in rivers and estuaries

data and the other is the calculation based on estimation of transport parameters such as travel time and dispersion coefficients. Since exact morphological data are often unavailable,

In both approaches, tracer experiments are needed to provide field data for water quality models calibration and validation procedures. Indeed, model calibration is often a weak step in its development and using experimental tracer techniques, the calibration and validation problems can be solved satisfactorily, improving the needed feasibility of the early warning

Tracer experiments are typically conducted with artificial fluorescent dyes (like rhodamine WT) (Fig. 11), whose concentrations are easily measured with a fluorometre. These tracers should be easily detected, non toxic and non-reactive, as well as, have high diffusivity, low

the parameter estimation technique is more promising.

acidity and sorption for a quasi-conservative behaviour.

Fig. 11. Rhodamine spreading after their injection in a river Mondego reach

Based on field experiments data, many investigators have derived semi-empirical equations (Hubbard et al., 1982; Chapra, 1997; Addler et al., 1999) or applied one-dimensional models (Duarte & Boaventura, 2008) to calculate experimental longitudinal dispersion coefficients from concentration time curves at consecutive sampling sites, using the analytical solution

The injected tracer dye mass must be calculated considering the water volume estimated in the river reach or reservoir system and the fluorometre detection limit. Specific problems of the application of tracers to surface water researches include the photosensitivity of dyes, such as fluorescence tracers, and recovery efficiency, which may imply the use of correction techniques for tracer losses. The tracer mass recovered at each site allowed the assessment of the importance of physical and biochemical river processes by quantifying precipitation, sorption, retention and assimilation losses. Usually, total tracer mass losses resulting from all these sinks can reach 40 to 50% of the injected mass (Duarte & Boaventura, 2008; Addler

In some recent experiments, a gas tracer (SF6) has been shown to be a powerful tool for examining mixing, dispersion, and residence time on large scales in rivers and estuaries

systems used by many water supply utilities.

of first order decay kinetics (Table 1).

et al., 1999).


(Caplow et al. 2004) as an alternative method to dye tracer experiments used for advection and dispersion characterisation.

Table 1. Hydraulic and dispersion parameters estimation using tracer dye experiments in a non-tidal reach of river Mondego

The dispersion processes in rivers are combined with a specific dynamic characterized by a decrease in maximum dye concentration (Fig. 12). The distribution of the tracer in all directions follows the sluggish injection into the channel. In non-tidal rivers, the lateral and vertical dispersion processes are almost always faster than the continuing longitudinal dispersion process.

Fig. 12. River Mondego model calibration: correlation between field tracer experiment data and model results.

One-dimensional modelling is a reasonably reliable tool to be considered for estimating the distribution of solutes in large rivers. Complex processes, for example in dead zones or downstream from the confluence of two rivers, have to be investigated by direct measurements and should be described by two-dimensional transport models. Calculation of net advection in tidal rivers is fairly straightforward, but longitudinal dispersion is difficult to determine *a priori*, and the application of two or three-dimensional transport models are often required.

A Hydroinformatic Tool for Sustainable Estuarine Management 15

The RMA4 programme solves depth-integrated equations of the transport and mixing process using the Galerkin Method of weighted residuals. The form of the depth averaged

> *c c c c c Rc h u v D D kc t x yx xy y h*

 

As with the hydrodynamic model RMA2, the transport model RMA4 handles onedimensional segments or two-dimensional quadrilaterals, triangles or curved element edges. Spatial integration of the equations is performed by Gaussian techniques and the temporal variations are handled by nonlinear finite differences consistent with the method

The numerical computation was carried out for all Mondego estuary spatial domains. Several sections were carefully selected and used for calibrating and analysis of the simulation results (Duarte, 2005). The legend includes the designation, section code and

**N6**

**152000**

**156000**

CODE SECTION NAME DISTANCE (km) N0 Estuary mouth 0.0 N1 Recreational harbor 1.3 N2 Figueira Bridge 2.8 N3 *Gramatal* 6.3 N4 *Cinco Irmãos* 7.4 N5 Maria da Mata sluices 10.0 N6 Foja Pumping Station 15.7 N7 River Arunca mouth 20.9 N8 Formoselha Bridge 28.6 N9 Pereira Bridge 31.4 S1 Gala Bridge (Lota) 2.6 S2 Armazéns creek (Negra) 4.4 S3 River Pranto mouth 5.4 S4 *Areeiro novo* 6.7 S5 Alvo sluices 8.7

**160000**

**348000**

**352000**

**356000**

**144000 148000**

( ) <sup>0</sup> *x y*

(1)

transport equation is given by equation (1)

c = concentration of pollutant for a given constituent;

their distance to the mouth of the estuary (Fig. 14).

**136000 140000**

**N0**

**N2 N1**

**S1**

**N3**

**S4 S3 S2**

**N4**

**S5**

**N5**

Fig. 14. The *MONDEST* model finite elements mesh and outline of the control sections

u, v = velocity in x direction and y direction; Dx, Dy, = turbulent mixing (dispersion) coefficient;

k = first order decay of pollutant; σ = source/sink of constituent; R(c) = rainfall/evaporation rate.

Where

t = time;

h =water depth;

described for RMA2.

Ever increasing computational capacities provide the development of powerful and userfriendly mathematical models for the simulation and forecast of quality changes in receiving waters after land runoff, mining and wastewater discharges.

The results of several research works have showed that the linkage of tracer experimental approach with mathematical modelling can constitute a power and useful operational tool to establish better warning systems and to improve management practices for the efficiently protection of water supply sources and, consequently, public health.
