Comparing levels of pollutants in regulated rivers with safe concentrations of pollutants for fishes: a case study

Comparing levels of pollutants in regulated rivers with safe concentrations of pollutants for fishes: a case study

Chemosphere, Vol. 33, No. 1, pp. 81-90, 1996 Coowinht 0 1996 Elsevier Science Ltd PIk 80045-6535(!?6)00154-3 Printed in Great Britain. All rights rese...

597KB Sizes 8 Downloads 32 Views

Chemosphere, Vol. 33, No. 1, pp. 81-90, 1996 Coowinht 0 1996 Elsevier Science Ltd PIk 80045-6535(!?6)00154-3 Printed in Great Britain. All rights resewed 0045-6535/96 SlS.OOtO.CQ

COMPARING WITH

LEVELS

SAFE

OF POLLUTANTS

CONCENTRATIONS

IN REGULATED

OF POLLUTANTS

A CASE

RIVERS

FOR FISHES:

STUDY

Julio A. Camargo

Centro de Ciencias Medioambientales,

C.S.I.C.,

Sermno 115 dpdo., 28006 Madrid, Spain

(Received in Germany 23 February 1996; accepted 10 April 19%)

ABSTRACT

The present study compares levels of fluoride ion (F-) in the regulated Durat6n River receiving an industrial effluent with safe concentrations the responsability

(SCs) of F- for brown trout and rainbow trout to evaluate

of fluoride pollution for the absence of trout populations

downstream

from the

effluent. SCs for each trout species were estimated using the multifactor probit analysis software on acute mortality data. Differential releases from the dam caused short-term

flow fluctuations

at S- 1

(between dam and effluent), S-2, S-3 and S-4 sampling sites (0.1, 2.2 and 7.3 km downstream the effluent). Because of this, fluoride concentrations period downstream

from

exhibited a temporal variation over a one-day

from the industrial effluent; the highest and lowest concentrations

(ppm F-) were

0.11 and 0.10 at S-l,

19.6 and 1.67 at S-2, 7.02

and 0.43

at S-3, and 2.98 and 0.52 at S-4. The

mean F-concentration

at the industrial effluent was 25.3k3.9

ppm. SCs (infinite hours LCO.Ols) of

F- for rainbow trout and brown trout were respectively

5.14 and 7.49 ppm. Comparisons

between

levels of F- in the Durat6n River and SCs for trout species apparently indicate that fluoride pollution was a minor factor in determining effluent.

It is concluded

undertaken

in regulated

the absence of trout populations

downstream

from the industrial

that intensive (24 hours) sampling surveys of pollutant levels should be rivers

with

pollution

sources

(e.g.,

industrial

appropriately the real influence of pollutants on freshwater organisms. Science Ltd 81

effluents)

Copyright

to evaluate

0 1996 Elsevier

82 INTRODUCTION Since the earliest civilizations were established on the Indus, Nile and Tigris-Euphrates

Rivers, man

has tried to control and regulate the flow of rivers for the benefit of agriculture and other human interests, Currently,

dam-building activity throughout the world is resulting in an approximate rate of

550 dams per year. By the year 2000, more than 60% of the total river flow in the world will be regulated (Petts, 1989). On the other hand, the pristine water quality of many rivers is nowadays being dramatically altered because of the global pollution caused by the human species (Meybeck and Helmer,

1989).

radionuclides

Toxic

pollutants,

and surfactants,

such

as ammonia,

profile of river systems (Rand et uf., 1995). Therefore, (e.g., sewage discharges,

cyanide,

dioxins,

PCBs,

pesticides,

may now be found in relatively high levels along the longitudinal when river regulation and pollution sources

industrial effluents) act together, freshwater

daily changes in the concentration

organisms

of toxic pollutants as a direct consequence

can be exposed to of differential water

releases from dams. Nevertheless,

despite its important implication for ecotoxicological

studies, very little attention has

been paid to that fact. In this respect, the present study examines the effect of river regulation on the concentration of fluoride ion (F-) along the recovery gradient of a regulated river (the Durat6n River) receiving an industrial effluent. Likewise, safe concentrations and rainbow

trout (Oncorh~~~chus mykiss

Walbaum)

of F- for brown trout (Salmo truth L.)

are estimated using the multifactor

probit

analysis (MPA) software (US EPA, 1991; Lee et al., 1995) on acute mortality data. After calculation, levels of Fm in the Duraton River and safe concentrationsof evaluate the responsability

F- for trout species are compared to

of fluoride pollution for the absence of trout populations

downstream from

the industrial effluent.

THE STUDY

AREA

The Duraton River is located in northern Spain within the Duero basin. A hydropower completed

in the middle of this river in 1953, and hydroelectric

dam was

power started to be generated by

discharging hypolimnial waters through three turbines. At present, the daily discharge of hypolimnial waters can fluctuate from 0.35 to 10.5 m?ls, with minimum flows at night and during weekends maximum flows during the daytime. Because of the eutrophic condition of the reservoir,

waters cause a significant deficit of dissolved

oxygen downstrem

and

hypolimnial

from the dam (Camargo,

1989;

1991). The industrial effluent entets the river 300 m downstream from the dam, flowing continuously from Monday to Friday. Because of the industrial process, high amounts of F- are discharged into the regulated river. Four downstream

sampling sites were selected along the recovery gradient of Duraton

River. A control station (S-l) was placed between the dam and the industrial effluent. Second (S-2),

83 third (S-3) and fourth (S-4) stations were placed about 0.1, 2.2 and 7.3 km downstream effluent, respectively.

Upstream from Burgomillodo

brown trout and rainbow trout, whereas downstream trout populations are absent (Camargo,

MATERIALS

from the

Dam, the DuraMn River supports populations

of

from the dam (S- 1, S-2, S-3 and S-4 stations)

1989; 199 1).

AND METHODS

PollutantLevels Extensive (four) and intensive (one) sampling surveys were undertaken during the summer of 1987. At all stations, F- concentration

and river depth were estimated in situ. Extensive

surveys involved

punctual (one-time) estimations whereas the intensive survey consisted of estimations over a one-day period. In addition, the concentration of fluoride ion was directly estimated at the industrial effluent. F- concentrations Orion-USA

were assessed

using an Orion-USA

model 94-09 specific ion electrode and an

model 90-02 calomel reference electrode. Water samples were analysed at pH 5.5 after

adding total ionic strength adjustment (CDTA) as complexing

buffer (TISAB-III)

agent for total fluoride analysis.

with cyclohexanediamine

tetra acetic acid

The specific ion electrode was calibrated

according to the analytical method described by Orion Research ( 1983). Spatial differences were determined

in the concentration

through one-way

river depth during the intensive

of fluoride ion along the recovery gradient of Durat6n River

ANOVA. Bivariate relationships sampling

Temporal variations in F- concentration

survey were determined

between through

F- concentration Pearson

and

correlation.

were evaluated by calculating the respective coefficients

of

variation. All statistical analyses were performed in accordance with Sokal and Rohlf (1987). Sufe Concentrutions Safe concentrations

(SCs) and their 95% confidence

using the multifactorprobit mortality

analysis (MPA) software

data from Camargo

methodology

(1989;

1991) and Camargo

utilizes data derived from short-term

24 to infinite hours) causing quanta1 responses

and Tarazona

toxicity bioassays

toxic substances that can exist in an aquatic environment Lee et al. (1995) have recommended

limits for each trout species were estimated (US EPA, 1991; Lee et al., 1995) on acute (1991).

This original

to predict the concentration

(e.g., mortality) at 0.01% population of test species.

minimums

of five different toxicant concentrations

and four

short-term exposure times to obtain estimations with small standard errors. In this investigation, different F- concentrations

of

for a determined exposure time (e.g., from

five

were used, and short-term exposure times were 24, 48, 72, 96, 120, 144,

168 and 192 hours. The MPA methodology

solves the concentration-time-response

iterative reweighed least squares technique (multiple linear regression).

equation simultaneously

via the

The independent variables

are

84

4.5

A

4.0

A

no effluent

3.5

A ??

3.0

0

??

2.53

A 2.0:

:

1.5:

0

LO! 0.5: o.oj

m

A 1

Sl

s-2

A I

A I

s-3

S-4

stations Fig. 1. Concentrations

of fluoride ion (ppm F-) estimatedat

sampling sites (S-l, S-2, S-3 and S-4)

during the four extensive surveys. exposure

time and toxicant concentration.

responding

to each concentration.

heterogeneity

factor (chi-squared

After

The dependent evaluating

variable is the probit of the proportion

several

MPA models

variable divided by degrees of freedom),

with

regard

to the

a parallel-regression-line

model was selected as the best one since its heterogeneity factor was the smallest (see US EPA, 1991;

Lee et al., 1995). This model assumes that the concentration-response

relationship is a continuum in

time, the mode of action of the toxicant being similar as the reciprocal of time varies. In order to improve the calculation of SCs, fluoride concentrations

and exposures

times were loglo transformed

before statistical analyses (US EPA, 1991; Lee et al., 1995). Because there is no probit value for 0% response

(Finney,

approximation

RESULTS

1971), the probit value of 1.281 for 0.01% response

was chosen as the best

to estimate safe concentrations.

AND DISCUSSION

F- concentrations it was expected,

at each sampling site during the four extensive surveys are shown in Figure 1. As S-2, S-3 and S-4 stations exhibited much higher concentrations

station (S- 1) whenever the industrial effluent occurred. It is surprising

than the control

to realize, however,

that in

85 100 90 80 70 G 60 2 c 50 g ,40 g ‘Z ,30

,20

0

4

8

12 16 20 24

80 70 G 60 k 50 E 8 40 g 'e 30 20 10 0 0

4

8

12 16 20 24

0

Fig. 2. Temporal variation over a one-day period of F- concentration and river depth at S- 1, S-2, S-3 and S-4 stations during the intensive sampling survey in the Duratbn River.

86 Table 1. Mean concentrations coefficientsof intensive

(n=17) of fluoride

variation

sampling

was 25.3ti.90

ion (ppm F-) f standard

(%) at each station (S-l,

Coefficients

and

S-2, S-3 and S-4) during the

survey. The mean F- concentration

at the industrial

effluent

ppm (n=14; C.V.=15.4%). S-l

Mean concentrations

deviations,

f SD

s-2

s-3

S-4

0.1 +o.o

6.81t5.7

2.7k2.1

1.3i0.8

0.0

83.8

77.8

61.5

of variation

some surveys S-3 and S-4 had higher concentrations

than S-2 (the station just below the effluent).

Only when the industrial effluent was interrupted for repairing the water treatment plant associated to the industry, all sampling sites exhibited similar F- concentrations.

This fact obviously

indicates that

the industrial effluent was the immediate cause of the high levels of fluoride ion estimated in the waters of Durat6n River. The F- concentrations

at the industrial effluent during the extensive surveys

were 21.4, 24.8 and 22.9 ppm. The surprising occurrence of higher F- concentrations

at S-3 and S-4 as compared with S-2 may be

explained by examining the results obtained during the intensive sampling survey (Figure 2). Owing to the differential experienced

discharge

a temporal

of hypolimnial

waters from Burgomillodo

variation over a one-day

period at sampling

industrial effluent. The highest and lowest concentrations 0.10 at S-l,

Dam, F- concentrations sites downstream

from the

of fluoride ion (ppm F-) were 0.11 and

19.6 and 1.67 at S-2, 7.02 and 0.43 at S-3, and 2.98 and 0.52 at S-4. In this way,

higher F- concentrations

at S-3 and S-4 than at S-2 were possible because of the simultaneous

action

of the industrial effluent and differential water releases from the dam. Nevertheless,

if we take into account all concentrations

sampling survey, (RO.05;

then it may be determined

ANOVA) with increasing

highest mean concentrationof ppm).

This downstream

of fluoride ion estimated during the intensive

that mean F- concentrations

distance from the

industrial

effluent

decreased

significantly

(Table 1). S-2 had the

fluoride ion (6.8 ppm), being followed by S-3 (2.7 ppm) and S-4 (1.3 decrease

also was

apparent

for the coefficients

of variation

in F-

concentration (Table l), despite the fact that the temporal variation in river depth was similar along the study area (Figure 2). Furthermore, significant at S-2 only (r=-O.637; of concentration

Pearson correlation between F- concentration and river depth was P
of alterations in the processes

and dilution with increasing distance from the effluent. Because

variation at S-2, S-3 and S-4 stations were much higher than the coefficient industrial effluent (Table l), it results evident that the major factor responsible variation in F- concentrations

at S-2, S-3 and S-4 stations was river regulation.

coefficients

of

of variation at the for the temporal

87

0

20

40

60

80

120

100

140

160

180

200

exposure time (hours) 120,144,168 and 192 hours LCO.Ols) of fluoride ion

Fig.3 Safe concentrations(24,48,72,96, (ppm F-)and their approximate

95% confidence

limits for rainbow trout (0. mykiss)

and

brown trout (S. mcn~l).

Table 2. Safe concentrations confidence

(infinite hours LCO.Ols) of fluoride ion (ppm F-)and their 95%

limits for rainbow

trout (0. mykiss)

and brown trout (S. fruttu).

addition, application factors (AF=SCinfinite hoursI% hours LO) % hours LC50 values are from Camargo (1991). safe concentration 0. mykiss

5.14

application factor 0.048

(3.10-7.53) S. trutta

7.49 (4.42-lo.%)

In

are also presented.

0.046

88

F- concentrations 24,48,

causing mortality at 0.01% populations

of brown trout and rainbow trout after

72, 96, 120, 144, 168 and 192 hours of exposure to fluoride ion are presented in Figure 3. In

addition,

F- concentrations

mortality

at 0.01%

that can exist in an aquatic environment

populations

of each trout species

confidence limits of safe concentrations higher SCs than 0. mykiss

are presented

for infinite hours

causing

in Table 2. Although

95%

(SCs) overlap (Table 2; Figure 3), S. hutiu always exhibited

as an outcome of differential responses

to fluoride toxicity between these

two trout species. Thus, the estimated SC values suggest that S. trunu might be a more tolerant fish species to fluoride pollution than 0. mykiss during both acute and chronic exposures.

Furthermore,

the estimated SC values of fluoride ion for rainbow trout and brown trout are significantly higher than those for several species of freshwater

invertebrates (see Camargo and La Point, 1995). On the other

hand, application factors (AF=SC ,,,rinlrc ho&96

hours LC50) were very similar between trout species

(Table 2); 0.046 for brown trout and 0.048 for rainbow trout. And it is worth noting that Pimentel and Bulkley (1983), concentrations

after lethal acute toxicity bioassays,

(ppm F-) of 25

arbitrary AF of 0.05. In our case, safe concentrations water hardness of 21.8 ppm CaC03 (Camargo, Comparisons

estimated

maximum

from the industrial effluent.

were lower than the 24 hours LCO.Ols calculated for rainbow trout and brown

(Figure 3; Table 2). Moreover,

Pimentel and Bulkley (1983) have showed

along the study area was about 165 ppm CaCO3 (Camargo, such as low dissolved

oxygen concentrations

for the absence of trout populations

that the tolerance of

and the average value of water hardness 1989; 1991). Consequently,

and short-term

other abiotic

flow fluctuations,

would be

at S-2, S-3 and S-4 stations. Trout species indeed

were already absent just below the dam (Camargo, which

through time,

tend to be lower than the estimated SC values

rainbow trout to fluoride ions increases with water hardness,

cyprinids,

(Table l), rather than punctual

(Figures 1 and 2), as the real level of fluoride affecting trout populations

then it is easy to see that these mean Fm concentrations

responsible

the absence of 0.

estimated at sampling sites during the intensive survey (Figure 2), we can see that

these concentrations

factors,

of Fm for trout

Even in the case of the highest F-

trout (Figure 3). In addition, if we consider mean Fm concentrations concentrations

an

refer to soft water, with an average value of

between levels of F- in the Duratdn River and safe concentrations

and S. truttu downstream

concentrations

trout through

1989; 1991; Camargo and Tarazona, 1991).

species apparently indicate that fluoride pollution was a minor factor in determining mykiss

allowable toxicant

(soft water) and 9.6 (hard water) for rainbow

1989; 1991). In this respect it is noteworthy

in general are more tolerant to low concentrations

of dissolved

salmonids (Alabaster and Lloyd, 1980), were present at all sampling sites (Camargo, There is little doubt that this study has demonstrated

oxygen

that than

1989; 1991).

that, when river regulation and pollution

sources (e.g., industrial effluents) act together, freshwater organisms can be exposed to daily changes in the concentration dams. Thus, regulated

of toxic pollutants

intensive

(24 hours)

rivers with pollution

as a direct consequence

sampling

sources

surveys

of differential water releases from

of pollutant levels must be undertaken

in order to evaluate appropriately

the real influence

in of

89 pollutants on freshwater

organisms.

Moreover,

because alterations in the processes

and dilution can occur with increasing distance from dams and pollution demonstrated),

investigators

should

not expect

to find necessarily

sources

of concentration

(like this study has

the lowest

pollutant

levels

associated to the highest water flows. On the other hand, pollutant levels must be compared with safe concentrations

of pollutants for freshwater organisms

use of the multifactor estimationof

probit analysis

those safe concentrations

whenever data are available. In this respect, the

(MPA) software

on acute mortality data may facilitate the

(SCs). However,

it is imperative to take into consideration

three important points (Camargo and La Point, 1995). First, laboratory investigations

concerning

the

estimation of SCs should be conducted in water quality conditions of highest potential toxicity. This is the case of the relationship between fluoride toxicity and water hardness, with increasing the latter. Second, have been

shown

concentrations

to be consistent

estimated

reproduction).

confirmation

end points

chronic

toxicity

during

acute toxicity

data are based

testing.

After

all, safe

on sublethality

(e.g.,

growth,

And third, SCs estimated via the MPA software should be validated not only by long-

term toxicity bioassays compounds

from

where the former decreases

sublethal effects should be measured instead of mortality if they

but also by field studies. Predicting accurate safe concentrations

for fishes and other aquatic organisms of the estimated safe concentrations

is a priority in ecotoxicology,

of chemical

and therefore the

has to be an essential goal in order to evaluate the

goodness of the method.

ACKNOWLEDGEMENTS I am particularly grateful to F.L. Mayer for a copy of the multifactor probit analysis software and to T.W.

La Point for helpful

corresponds

indications

to use it. The information

on field sampling

surveys

to a portion of my doctoral dissertation.

REFERENCES Alabaster J.S. and Lloyd R. (1980) Water Quality Criteriafor Freshwuter Fish. Butterwoth, Camargo J.A. (1989) Estudio Ecotoxicoldgico

London.

de1 Impact0 Ambiental Genercldo por unu Regulacirin

de Cauoizles y un Verti& de Fltior, sobre ILLSComunidudes de Animales Acudticos de1 Rio Durutdn. Doctoral Dissertation, Camargo J.A.

Universidad

Audnoma,

(1991) Ecotoxicological

Madrid.

analysis of the influence of an industrial

populations in a regulated stream. Aquacult. Fish. Manug. 2 2,509-S Camargo J.A. and Tarazona J.V. (1991) Short-term rainbow trout and brown trout. Chemosphere 2 2,6056

effluent on fish

18.

toxicity of fluoride ion (F-) in soft water to 11.

90 Camargo

J.A.

and La Point T.W.

concentrations

for five species

(1995) Fluoride

of Palearctic

toxicity to aquatic life: a proposal

freshwater

invertebrates.

Arch.

Environ.

of safe Contum.

Toxicol. 29, 159-163. Finney D.J. (1971) ProbitAnulysis, Lee G.,

Ellersieck

M.R.,

3rd Edition. Cambridge University Press, London.

Mayer F.L.

and Krause

G.F.

(1995)

Predicting

chronic lethality of

chemicals to fishes from acute toxicity test data: multifactor probit analysis. Environ. Toxicol. Chem. 14, 345-349. Meybeck M. and Helmer R. (1989) The quality of rivers: from pristine stage to global pollution. Palaeogeogr. Palaeoclim. Palueoecol. 7 5,283-309. Orion

Research

Incorporated, Petts G.E.

(1983)

Instruction

manual: fluoride

electrode,

model

94-09.

Orion

Research

Cambridge, Massachusetts. (1989) Perspectives

RegulatedRiver

for ecological management

of regulated

rivers. In Alternatives

in

Management. Edited by J.A. Gore & G.E. Petts, pp 3-24. CRC Press, Boca Raton,

Florida. Pimentel R. and Bulkley R.V. (1983) Influence of water hardness on fluoride toxicity to rainbow trout. Environ. Toxicol. Chem. 2,381-386. Rand

G.M.,

Fundamentals Washington,

Wells

P.G.

of Aquatic

and McCarty Toxicology.

L.S. Edited

(1995) by G.M.

Introduction Rand,

to aquatic

pp 3-67.

toxicology.

Taylor

DC.

Sokal R.R. and Rohlf F.J. (1987) Introduction to Biostutistics, 2nd Edition. Freeman, New York. U.S.E.P.A.

(1991) Multifactor Prohit Analysis.

Washington,

DC.

Environmantal

Protection Agency,

In

& Francis,

600/X-91-101,