Version classiqueVersion mobile
OpenEdition Books

Trente ans de lutte contre l’onchocercose en Afrique de l’Ouest. Traitements larvicides et protection de l’environnement

 | 
Laurent Yaméogo
, 
Christian Levêque
, 
Jean-Marc Hougard

Environmental Impact Assessment of Settlement and Development in the Upper Léraba Basin

3. Ecotoxicological Impact of Human Activity at the Basin Scale: Assessment of Chemical Loads

Texte intégral

1The statistics on human and livestock populations, crops, fertilizers and pesticides in the Burkina Faso sector of the Basin, which were required for the hazard assessments, were grouped at source according to agro-pastoral zone (Sindou, Loumana and Niangoloko). While the first two of these zones fitted completely into the upper Léraba basin, the third did not. Only about one third of the Niangoloko zone was actually inside the Basin. Therefore, all the figures received for the Niangoloko zone, were divided by three before being entered into the accompanying tables and before being used for the assessments.

Human and Livestock Populations

2The numbers of humans and of livestock in the Burkina Faso sector of the Basin, recorded in 1991 and 1989 respectively, are given in Table 3.1.

Crop Types and Surface Areas Cultivated

3The surface areas (ha) cultivated by the main crops in the Burkina Faso sector of the Basin in 1992/93, are given in Table 3.2. Cotton, the only important cash crop for which pesticides were used, was normally sown at the end of May, and harvested in October.

Use of Fertilizers

4The fertilizers commonly used in the Basin were "NPK" (a product rich in nitrogen, phosphorus and potassium) and urea. The quantities (kg) of those products used in the Burkina Faso sector of the Basin in 1993, are given in Table 3.3.

Pesticides Used and their Ecotoxicological Profiles

5Throughout the Basin, pesticides were only used for the protection of cash crops, of which cotton was the only one grown in significant quantity. In general, pesticides were applied twice during the cotton-growing season, to plants which were attacked by aphids (usually attacking in July or in September), fruit-eating caterpillars (usually attacking in September or October), arachnids and whitefly.

Table 3.1 Numbers of Humans and of Livestock in the Burkina Faso Sector of the Basin, 1991 and 1989, respectively

Table 3.1 Numbers of Humans and of Livestock in the Burkina Faso Sector of the Basin, 1991 and 1989, respectively

Table 3.2 Main Crop Types and Surface Areas Cultivated (ha) in the Burkina Faso Sector of the Basin, 1992/93

Table 3.2 Main Crop Types and Surface Areas Cultivated (ha) in the Burkina Faso Sector of the Basin, 1992/93

6The two most commonly used pesticides for cotton protection were:

7Sherdiphos: containing 30g/litre cypermethrin, 150g/litre dimethoate and 240g/litre triazophos; application rate 30g/ha, 150g/ha and 240g/ha a.i., respectively, 1-2 applications/year.

8Fastac D: containing 7g/litre alpha-cypermethrin and 133g/litre dimethoate; application rate 15 g/ha and 400 g/ha a.i., respectively, 2-3 applications/year.

9In addition, Ciperthion CK (containing alphacypermethrin and isoxathion) was used on a small scale in some localities (2-3 applications/year).

10The quantities (litres) of Sherdiphos and Fastac D dispensed in the Burkina Faso sector of the Basin in 1993, are given in Table 3.4.

11Ecotoxicological profiles of the active ingredients of these products were synthesized, on the basis of physico-chemical properties, application rate, toxicology, terrestrial/aquatic/atmospheric fate, environmental distribution, groundwater ubiquity score (GUS), etc. Summaries are given below.

12Cypermethrin: According to its half-life and GUS values, cypermethrin (a synthetic pyrethroid) appears to be a non-persistent and non-leachable molecule. Its low vapour pressure suggests that it does not volatilize significantly, while its high log Kow value and its low solubility explain its scant affinity for water. Despite its high lipofilicity, it is not bioaccumulative, as the molecule is rapidly metabolized. Cypermethrin exhibits high toxicity to some non-target organisms, such as honey-bees, fish and aquatic crustaceans, but is considered to be only slightly toxic to birds and mammals. Once the molecule has entered the environment, it is reasonable to expect that it will be almost completely absorbed by soil, in which it will degrade in a few days.

13Dimethoate: This organophosphate compound has very low vapour pressure and log Kow values, which cause it to have a high affinity for water. The molecule can therefore be considered as leachable. It is also non-persistent. Its low lipofilicity excludes the possibility of bioaccumulation. While it is only slightly toxic to mammals and birds, it is highly toxic to non-target invertebrates. It is to be expected that, once introduced into the environment, dimethoate will be readily leached into running water by the first rain.

14Triazophos: This organophosphate has a low to medium level of persistence, and can be expected to volatilize moderately from soil. However, in view of its limited lipofilicity, triazophos could leach through the soil profile, thus increasing the probability that it could be transported by runoff from rainfall. The molecule is slightly toxic to mammals and birds, moderately to highly toxic to fish, and highly toxic to crustaceans and bees. Once introduced into the environment, it is expected that triazpophos will most readily enter the water compartment, within which degradation will occur. Because of its limited lipofilicity, small quantities of triazophos may also be found in river sediments.

15From these preliminary assessments, it was concluded that alpha-cypermethrin, dimethoate, and triazophos were non-bioaccumulative and relatively short-lived, with a predicted limited impact on the aquatic environment. However, dimethoate and triazophos could, through leaching and runoff, reach flowing water and have some impact. Accordingly, a more refined evaluation of these compounds should be made with a view to improving this prediction.

Table 3.3 Quantities of Fertilizers (kg) Used in the Burkina Faso Sector of the Basin, 1993

Zone

NPK

Urea

Sindou

14150

10 000

Loumana

5 000

500

Niangoloko

3 666

3000

Totals:

22 816

13 500

Table 3.4 Quantities (litres) Sherdiphos and Fastac D Dispensed in the Burkina Faso Sector of the Basin, 1993

Zone

Sherdiphos

Fastac D

Sindou

490

Loumana

1960

Niangoloko

753

Totals:

2 713

490

Table 3.5 Chemical Properties of the Pesticides Entered into the SoilFug Model

Table 3.5 Chemical Properties of the Pesticides Entered into the SoilFug Model

The SoilFug Model and its Application for Estimating Pesticide Runoff

16For all the pesticide runoff simulations, the SoilFug model of Di Guardo et al. (1994a) was used.

17SoilFug is a model for the prediction of potential surface water contamination derived from pesticide use on agricultural fields. It uses the fugacity approach to a basin scale soil environment and calculates the partition of the Chemical applied to the soil phases and its possible contamination of surface water during the rain events, periods of time starting with a rainfall and ending with the return to the background water level in the adjacent streams. It requires a limited amount of Chemical and environmental data, and it furnishes an average concentration of pesticide in outflowing waters. SoilFug is essentially an unsteady-state but equilibrium event model. This is because it takes into account the disappearance of the Chemical according to different phenomena (degradation, volatilization, runoff), but then calculates the partition among the different phases of the soil according to a Level 1 fugacity calculation (Mackay 1979; Mackay & Paterson 1981) in the rain event periods. It is derived from previously published models (Mackay 1991; Mackay & Stiver 1991; Di Guardo et al. 1993b).

18Briefly, the model considers four different compartments in the soil: soil air, soil water, organic matter and mineral matter. For each of these compartments a capacity (Z) can be calculated and therefore the fugacity can be worked out, once the volume and the Chemical input are known. From the fugacity, Chemical amounts and concentrations in each compartment can be calculated.

Input Data for the SoilFug Model

19Input data for the Soilfug model fell under four headings—Chemicals, soil properties, water input/output balance, and pesticide treatments—each of which is elaborated below.

Chemicals

20Physico-chemical properties of the pesticides, together with their half-life in soil, were the first data to be entered into the SoilFug model. The values actually employed are given in Table 3.5. They were drawn from the recent literature (Worthing & Hance 1991, Howard 1991a and 1991b), with the exception of the half-life figures, which were arbitrarily selected on the basis of the extremes of the ranges of variability encountered under different conditions, and taking into account the high temperatures of the Léraba Basin.

Soil Properties

21The input data required for the soil scenario description of the model were temperature, depth, volume fractions of air and of water in soil at field capacity, organic carbon content, area of the Basin, and the number of simulations, i.e. the number of rain events. The actual values of these parameters, as used for the simulations, are given in Table 3.6. They were selected on the basis of agronomical considerations, e.g. soil depth was derived from data in the BUNASOL Technical Report (No 85, Ouagadougou, 1993), while the soil itself was assumed to be “clay, structured soil” with a field capacity of pF = 2.5.

Table 3.6 Soil Properties Data Entered into the SoilFug Model

Temperature

28°C

Soil depth

0.2 m

Air volume fraction

0.25

Water volume fraction

0.42

Organic carbon fraction

0.01

Basin area

537 600 ha

Rain events

10

Water Input/Output Balance

22The mass balance of the water was calculated on the basis of rain events and outflow. The basic data for the water input calculation were the number of rain events, the interval between rain events, the duration of the rain events and the quantity of rain.

23A rain event was arbitrarily considered to be a period during which consecutive instances of rainfall were separated by no more than 3 days, and each event had to be sufficient to produce a measurable outflow, since the model was required to estimate the quantities of the pesticides in advective water moving out of the basin. In addition, each rain event had to allow for the repartition and the re-equilibration of the Chemicals between the soil phases and the rainwater.

24Outflows were estimated on the basis of the river discharge rates recorded at Léraba Bridge hydrology station, taking into account the estimated amount of rainfall over the whole basin, extrapolated from the Niangoloko rainfall data, which were considered to be reasonably representative of the Basin.

25The actual values used for the simulations, based on the Burkina Faso official records for Niangoloko for the year 1992, are given in Table 3.7.

Pesticide Treatments

26The next parameters necessary for the running of the model were the periods and numbers of pesticide applications, pesticide dosage, days between pesticide applications and rain events, area treated, and the half-life of the pesticides. The actual values of these parameters that were used for the simulation, are given in Table 3.8.

27In assembling these values for the SoilFug model, it was assumed that:

  • cotton was sown at the end of May, that the main growing season was from June to September, and that harvesting was done in October,
  • the first pesticide treatment was around 20 July, and the second treatment in late August or in early September,
  • treatments with Sherdiphos were made throughout the Basin, as this was the product which accounted for more than 80% of total pesticide consumption,
  • there were about 10,000 ha of cotton in the Basin, since there were about 6,000 ha in the Burkina Faso sector, the area of which accounted for about 70% of the total basin.

Table 3.7 Rain Event Data Entered into the SoilFug Model, Niangoloko, 1992

Rain event N°

Days to rain event

Duration of rain event (days)

mm “in” (rainfall)

mm“out” (outflow)

1

3

13

149

75

2

4

5

26

13

3

7

6

71

36

4,

4

3

30

15

5

4

1

39

20

6

5

1

16

8

7

6

1

23

3

8

8

5

56

6

9

4

12

80

8

10

19

2

21

2

Table 3.8 Pesticide Treatment Data Entered into the SoilFug Model

Rain event No

Area treated (ha)

Days before rain event

Dosage (kg/ha)

Half-life (days)

1

10 000

3

0.03

Cypermethrin

20

0.15

Dimethoate

7

0.24

Triazophos

20

2–4

5

10 000

6

0.03

Cypermethrin

20

0.15

Dimethoate

7

0.24

Triazophos

20

6–10

Output of the SoilFug Model

28From the sets of data described above, the model was able to calculate the average concentrations of the different pesticides in water during each rain event, taking into account not only the partitioning phenomena between soil and water, but also the estimated persistence of each molecule (i.e. its half-life). The final output was a series of graphs showing the predicted concentrations of the different pesticides at the basin outlet which, in this case, was represented by the main River Léraba at Léraba Bridge. These graphs are presented in Figures 3.1, 3.2 and 3.3, and are discussed below.

Estimated Impact of the Pesticide Loads

29The results of the SoilFug model calculations demonstrated that certain quantities of the most common cotton protection pesticides (none of which have bio-accumulation potential) could have been present at the Léraba Bridge outflow, during the period July to November. The ecotoxicological significance of the estimated concentrations of the different Chemicals at the Léraba outflow, is discussed below.

3.1 Graph Showing the Concentration of Cypermethrin at the Léraba Bridge Outlet, Predicted by the SoilFug Model

3.1 Graph Showing the Concentration of Cypermethrin at the Léraba Bridge Outlet, Predicted by the SoilFug Model

3.2 Graph Showing the Concentration of Dimethoate at the Léraba Bridge Outlet, Predicted by the SoilFug Model

3.2 Graph Showing the Concentration of Dimethoate at the Léraba Bridge Outlet, Predicted by the SoilFug Model

30Cypermethrin: The concentration of cypermethrin peaked at approximately 0.002 μg/litre immediately after the first treatment, and then rapidly declined to 0.0006 μg/litre. After the second treatment, the concentration peaked at about 0.0026 μg/litre, and subsequently declined to 0.0003 μg/litre.

31According to the literature on cypermethrin, most of the acute toxicity concentrations for the aquatic fauna are on the order of a few μg/litre, i.e., at significantly higher concentrations than those given by the SoilFug model.

32Dimethoate: The concentation of dimethoate peaked at about 0.41 μg/litre immediately after the first treatment, and then declined to 0.015 μg/litre.

3.3 Graph Showing the Concentration of Triazophos at the Léraba Bridge Outlet, Predicted by the SoilFug Model

3.3 Graph Showing the Concentration of Triazophos at the Léraba Bridge Outlet, Predicted by the SoilFug Model

33After the second treatment, the concentration peaked at 0.99 μg/litre, and then rapidly declined to 0.001 μg/litre, due to the short half-life of this molecule.

34Data in the literature on dimethoate indicate that its acute toxicity for the aquatic fauna is in the order of a few mg/litre, i.e. at significantly higher concentrations than those given by the model.

35Triazophos: The concentration of triazophos peaked at approximately 0.39 μg/litre immediately after the first treatment, and then rapidly declined to 0.12 μg/litre. After the second treatment, the concentration peaked at about 0.60 μg/litre, and subsequently declined to 0.07 μg/litre.

36Data in the literature on triazophos indicate that its acute toxicity for fish is on the order of a few mg/litre, and for aquatic crustaceans in the order of a few μg/litre.

37If a risk assessment is made, by examining the three molecules simply on the basis of a comparison between the predicted environmental concentrations (PEC) and acute toxicity data for the aquatic fauna, the results are quite positive and give no immediate cause for concern. Even at the highest peaks of concentration, which lasted only a few days, the distances between the two sets of values were in orders of magnitude. If, however, a more severe criterion is used to evaluate the model predictions, e.g. by applying the water quality objective for aquatic life established by the EEC Scientific Advisory Committee on Toxicology and Ecotoxicology (CSTE 1994) as 0.01 μg/litre for several organophosphate compounds, one can observe that this more stringent level of concentration could have been exceeded for several days after each pesticide treatment.

38It can therefore be concluded that on the basis of the acute toxicity criteria considered in the first of the two types of assessment, there is no immediate cause for concern with regard to the aquatic fauna of the Léraba. However, if the more severe EEC criterion is applied to the assessment, the fact that the criterion was theoretically exceeded for several days has to be accepted as proof of contamination, and should therefore be viewed as an early warning.

Impact of Organic Loads and Nutrients

Overall Conclusions

39Taking into consideration that there were no sewage Systems, no formal Systems of pit latrines and septic tanks, generally dispersed settlement infrastructures (especially in the Burkina Faso sector), and only moderate human and livestock population densities, it is reasonable to assume that by far the greater proportion of organic materials entering the environment were being degraded at the soil compartment level, over a large surface area, and that the nutrient component of organic waste would be taken up directly by vegetation, before it had a chance to be conveyed into the aquatic environment.

40Furthermore, from the limited data available on the use of proprietary fertilizers, it would appear that over the Basin as a whole, those products were dispensed in relatively small quantities.

41For these reasons, it was concluded that there was insufficient justification for making detailed quantitative calculations of the impact of organic matter and nutrients in the Léraba river System.

Point Sources of Organic Loads and Nutrients and their Impact on the Aquatic Environment

42Early on in the implementation of the Pilot Project the authors realized that although there was insufficient justification for making quantitative estimations of organic substances and nutrients, they should not overlook the possibility that there would be some situations where these substances could have localized impact; for instance, where a village was located close to a main river. In such situations, the habits of bathing, washing clothes and culinary utensils and the watering of livestock at the water’s edge, could provide direct point sources of organic substances and nutrients, to such an extent as to modify the quality of the water and consequently, to exert negative effects on aquatic life.

43In March 1994 the opportunity arose to study this assumption in more detail, by making hydrobiological measurements of the biological condition of the Léraba at Léraba Bridge and comparing the results with those obtained from further downstream and from a comparable nearby river. In the first investigation, fish samples were collected from the Léraba at Léraba Bridge, from the Léraba near Léraba Railway Station (a less densely populated area, some 10 km downstream from Léraba Bridge), and from the River Comoé (near Folonso in a sparsely populated section of the Comoé valley, and some 55 km south-east of Léraba Bridge), for the purpose of calculating CPUE values. The second investigation, conducted at these same sampling points, involved the daytime and night-time collection of drifting aquatic invertebrates, and the calculation of the corresponding Drift Indexes (number of individuals collected per cubic meter of filtered water during the daytime and night-time).

44The results of the two investigations are presented in Table 3.9.

Table 3.9 Data on the Biological Condition of Fish and Invertebrates at Three Sampling Sites (Léraba Bridge, Léraba Railway Station and R. Comoé)

Table 3.9 Data on the Biological Condition of Fish and Invertebrates at Three Sampling Sites (Léraba Bridge, Léraba Railway Station and R. Comoé)

1. Data for 1994.
2. Mean data for 1990-1994.
3. Data for 1994.

45It is evident from these data that, for almost every parameter, the R. Comoé was richer in fish and invertebrates than the Léraba, which was to be expected considering that there was very little human activity around the Comoé sampling site.

46However, by far the most informative of the data are those relating to the two Léraba sampling sites. In all respects, the fish and invertebrate populations of the Léraba at the Léraba Railway Station site were richer than those at Léraba Bridge, even though the two sampling sites were separated by only about 10 km of river. The explanation proposed for these differences is that the fish and invertebrate populations of the Léraba at Léraba Bridge were influenced by the very close proximity of Lérabadougou village, whose western perimeter was no more than 300 meters from the sampling site. Evidently, the human and livestock populations of Lérabadougou constituted a point source of organic substances and nutrients which had a very localized impact on the aquatic environment, since the biological condition of the river was so much better only a few kilometers downstream at Léraba Railway Station; a much smaller settlement, twice as far from the river as Lérabadougou, and presumably having little or no adverse effect on the adjacent stretch of the river.

Table des illustrations

Titre Table 3.1 Numbers of Humans and of Livestock in the Burkina Faso Sector of the Basin, 1991 and 1989, respectively
URL http://books.openedition.org/irdeditions/docannexe/image/28722/img-1.jpg
Fichier image/jpeg, 31k
Titre Table 3.2 Main Crop Types and Surface Areas Cultivated (ha) in the Burkina Faso Sector of the Basin, 1992/93
URL http://books.openedition.org/irdeditions/docannexe/image/28722/img-2.jpg
Fichier image/jpeg, 29k
Titre Table 3.5 Chemical Properties of the Pesticides Entered into the SoilFug Model
URL http://books.openedition.org/irdeditions/docannexe/image/28722/img-3.jpg
Fichier image/jpeg, 31k
Titre 3.1 Graph Showing the Concentration of Cypermethrin at the Léraba Bridge Outlet, Predicted by the SoilFug Model
URL http://books.openedition.org/irdeditions/docannexe/image/28722/img-4.jpg
Fichier image/jpeg, 49k
Titre 3.2 Graph Showing the Concentration of Dimethoate at the Léraba Bridge Outlet, Predicted by the SoilFug Model
URL http://books.openedition.org/irdeditions/docannexe/image/28722/img-5.jpg
Fichier image/jpeg, 48k
Titre 3.3 Graph Showing the Concentration of Triazophos at the Léraba Bridge Outlet, Predicted by the SoilFug Model
URL http://books.openedition.org/irdeditions/docannexe/image/28722/img-6.jpg
Fichier image/jpeg, 45k
Titre Table 3.9 Data on the Biological Condition of Fish and Invertebrates at Three Sampling Sites (Léraba Bridge, Léraba Railway Station and R. Comoé)
Légende 1. Data for 1994.2. Mean data for 1990-1994.3. Data for 1994.
URL http://books.openedition.org/irdeditions/docannexe/image/28722/img-7.jpg
Fichier image/jpeg, 37k

© IRD Éditions, 2003

Conditions d’utilisation : http://www.openedition.org/6540