Version classiqueVersion mobile

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

Sélection d’articles du programme OCP / Selected OCP papers

Long-term assessment of insecticides treatments in West Africa: aquatic entomofauna

L. Yaméogo, G. Crosa, J. Samman, K. Nabé, F. Kondé, D. Tholley et D. Calamari

Résumé

For the control of the Onchocerca volvulus vector in West Africa, up to 18,000 km of rivers from 1975 and up to 50,000 km from 1989 had been partly sprayed weekly with insecticides as part of the Onchocerciasis Control Programme (OCP). To evaluate the possible short-term and long-term effects of the application of insecticides on the nontarget fauna, an aquatic monitoring programme was set up during the initial phase of the programme. By analysing the in vertebrate data, which were collected using various sampling strategies from four different countries between 1977 and 1996, this paper evaluates the long-term changes of the invertebrate populations with respect to their taxonomic composition as well as their trophic structures. The discussed results of the applied numerical analysis strategy suggest that neither the taxonomic nor the trophic structures are greatly altered from the range of biological, flow-related variation that normally occurs in the studied river Systems. This allows us to conclude that the biological variation found here is ecologically acceptable. © 2001 Elsevier Science Ltd. All rights reserved.

Texte intégral

1. Introduction

1Onchocerca volvulus is a parasitic worm giving rise to skin reactions and eventually severe ocular lesions followed by blindness (Zimmerman et al., 1992). This human disease had been a major public health problem in many fertile valleys of West African countries in which it was an obstacle to their social and economic development. The parasite is transmitted by the female blackfly of the Simulium damnosum complex (Philippon, 1977), particularly the savannah vectors S. sirbanum and S. damnosum s.s., whose larval instars, requiring a minimum flow of about 50 cm s-1 for survival, breed in the bedrock areas of fast flowing water courses.

2In December 1974 the United Nations Development Programme under the aegis of WHO launched a 20-yr campaign for the control of the vector (Davies et al., 1978) and, being difficult to repress the adult of Simulium, it was decided to treat the larval stages whose distributions are limited to rapids.

3The initial phase of the Onchocerciasis Control Programme (OCP) covers a vast area of 764,000 km2 in which up to 18,000 km of rivers had been partly weekly sprayed with insecticides selected according to the criteria reported in Lévêque (1989) and on account of their efficacy against the larvae of the vector and their low toxicity for the non-target fauna.

4Obviously, such an extensive and prolonged use of insecticides could have important environmental risks, therefore an aquatic monitoring programme as well as lab and on field toxicity tests have been set up from 1974 to evaluate the possible short-and long-term effects of the insecticides on the non-target fauna.

5From 1975 to 1985 temephos (Abate®), chlorophoxim (two organophosphorous compounds) and a biological insecticide, Bacillus thuringiensis var. israelensis [B.t. H-14] (“Teknar”) had been the insecticides used. After these first 10 yr of treatment, the data collected on aquatic insect larvae and on fish populations lead to review papers demonstrating that “... the insecticides employed had little effect on the non-target fauna (Lévêque et al., 1988; Yamèogo et al., 1988). Although the first application of temephos and chlorphoxim had a fairly strong impact on invertebrate communities in the short term, it would seem that these situations disappear fairly quickly after a year or less of successive applications”.

6From 1980, the appearing of certain forest cytotypes of the vector resistant to temephos (Guillet et al., 1980) and to chlorphoxim by 1982 (Kurtak et al., 1982) forced the search of new compounds and the implementation of a new treatment strategy based on the rotational use of different insecticides. The new compounds were searched from groups unlikely to produce cross-resistance and having different modes of action: permethrin (pyrethroid), carbosulfan (carbamate), pyraclofos (organophosphorus compounds) and vectron (pseudopyrethroid). To prevent Aies reinvasion, from 1989 the original programme was expanded to 1,235,000 km2 controlling about 50,000 km of rivers.

7After the above-mentioned review papers that provided a comprehensive evaluation of the first 10-yr monitoring of non-target aquatic fauna, the biological data collected within a wide area till 1998 allow to face new questions arising from the implementation of the OCP previously outlined.

8Besides the main question addressed to the long-term changes of the invertebrate and fish populations with respect to their taxonomic composition as well their trophic structures, a better outlining of it is now possible:

  1. the severity of each specific insecticide used during the programme,
  2. the resistance of the communities to the induced stresses, and
  3. their recovery capacity.

9Analysing the invertebrate data collected in four countries during a period ranging from 1977 to 1996, this paper addresses the above questions.

10With respect to the long-term changes of the invertebrate and fish populations it has to be outlined that the observed patterns by the end of treatments represent an addition of the effect of all the treatments. If such additive effect can bias the evaluation of the biological effects of each specific treatment, however it can be considered a minor problem in our case where the main question focuses on the overall biological effect of the programme.

11All the ecological activities, encompassing the application of biological monitoring protocols, insecticides risk assessments and extensive impact evaluation studies, were recently reviewed in a synthetic paper by Calamari et al. (1998).

2. Materials and methods

2.1. Biological data and sampling methods

12The study addresses the analysis of the invertebrates collected in four rivers located in West Africa during the time extents and within the sampling stations indicated in Table 1. The rivers are savannah type showing high discharges from July to November and a low-water period from January to June, details on hydrological and physicochemical characteristics of the main West African rivers are reported by litis and Lévêque (1982) and by Moniod et al. (1977).

13The biological data were collected from 1977 to 1996 after the larvicides application as well as during suspension and pre-treatment periods; the insecticides applied are detailed in Table 2 with the code adopted in this paper.

14It is difficult to describe or to graphically show the overall treatment strategy but, to fully understand the presented results, the following aspects have to be clarified. The different larvicides were regularly applied during all the treatment period with a rotational strategy to avoid the appearing of resistant forms and the compounds were selected and applied according to the river flow. The biological samples were collected soon after the insecticides application and during the suspension periods, that mainly took place during the low-water season. In the first case the samples, other than to describe the “post-treatment status” of the rivers, mainly represent the effect of the last applied compound; thus, the insecticides labels are used to group the samples for the analyses as well as in the graphs identify such “last” applied compounds.

Table 1. Location of the sampling stations and sampling periods

River

Station

Country

Maximum sampling period

Maraoué

Entomokro

Ivory coast

December 1977 to February 1996

Pru

Asubende

Ghana

January 1980 to April 1995

Niandan

Sassambaya

Guinea

December 1984 to April 1994

Kaba

Outamba

Sierra Leone

March 1989 to April 1994

Table 2. Treatments for which biological samples had been collected

Code

Treatment

no

Pre-treatment

su

Suspension

bt

Bacillus thuringiensis (B.t).

H-14)

ab

Temephos (Abate®)

ph

Phoxim

pe

Permethrin

ch

Chlorophoxim

py

Pyraclofos

15A relative elevate number of pre-treatment observations is available for Niandan (25) and Pru rivers (17), on the contrary only 1 and 8 pre-treatment observations are available, respectively, for Kaba and Maraoué rivers.

16The organisms were sampled by means of Surber net and as day and night drifts. Details of monitoring and sampling methods can be found in Lévêque et al. (1979) and Yamèogo et al. (1991).

2.1.1. Drift samples

17Drift nets, 2 m long, 20 x 20 cm of diameter and 300 μm mesh size, were used to collect the drifting organisms. Considering that the densities of the drifting organisms in natural conditions are low during the daytime and maximum 1-2 h after the sun set, both these two phases had been sampled. Three samples were taken l½ h before sunset (day drift) and six samples 1½ h after sunset (night drift), the sampling time was 10-30 min for daytime drift and 3 min for night drift.

18As the number of drift organisms is related to the volume of water filtered the data were standardised expressing the number of individuais captured per m3 of filtered water (drift index).

2.1.2. Surber samples

19The benthic invertebrates were collected in shallow riffles over rock substrate using a modified 15 x 15 Surber sampler. The sampling stations were chosen because, being the breeding sites of S. damnosum, these areas had been subjected to direct application of the insecticides.

20Five samples were normally taken for each site and the mean number of individuais as well as the 95% confidence limits were calculated (Elliott and Décamps, 1973).

21The three sampling strategies, Surber samples, day and night drifts, had been employed in order to collect biological data showing different information. Night drift, which is supposed to be mainly voluntary, reflects an active period while the number of the day drift organisms is related to their health condition. Schematically, an increase in the number of day drift organisms can be related to external factors stressing the invertebrates, by contrast a high night drift could reflect an increased activity of the benthic organisms.

22As regards the use of the Surber net, this technique allows to sample, in a quantitative way, the organisms living in specific river areas and for this reason the collected organisms reflect the structure of the benthic communities, where this term indícate an assemblage of interacting individuals sharing at the same time the same space. Obviously, this concept of community is less applicable to the drift samples which represent “collections” of organisms turning up from a wider area located upstream the sampling sites.

23All the sampled individuais were classified according to their family levei and their trophic role: predators, shredders, scrapers and filtering or gathering collectors, this second classification method is based on the association between a limited set of feeding adaptations found in freshwater invertebrates and their basic nutritional resource categories (Cummins, 1973). To avoid the presence of rare taxa only the principal systematic units belonging to the Ephemeroptera, Tricoptera and Chironomidae were used for the analysis.

2.2. Numerical analysis methods

24The invertebrate data collected in each river have been analysed independently, distinguishing, for each river, the three techniques adopted: Surber net (named invertebrate communities), day and night drifts (named invertebrate assemblages).

25The numerical analyses strategy was selected in order to assess the long-term invertebrate structure variations with respect to the pre-treatment periods as well as the biological variation occurring during the suspension periods. The attention given to these two situations – treatment and suspension – is justified by the fact that they allow the evaluation of two main attributes of the biological communities, namely, the resistance (the capabilities of contrasting stress factors) and the resilience (the recover capabilities after a stress).

26Since the invertebrates collected were classified with respect to their taxonomic leveis as well as to their functional feeding group (trophic role), the analysis of the biological variation was addressed to both these structural and functional attributes. Whereas the first aspect concern, besides the faunistic interest over loss of global biodiversity, the quality of the biomass available for the upper trophic leveis, the second one is related to the stability of the energetic flows. On account of the invertebrate position in the first levels of the river food webs, changes in this latter aspect can be an alert signal about greater detrimental effects on the ecological characteristics of the whole river System.

27Because the trophic structure is a property of the living organisms emerging at community levei, only the Surber samples were used for the analysis of this biological property.

28On the basis of the results of preliminary data inspection (Crosa et al., 1998), the following analysis techniques were applied:

29• Invertebrate taxonomic diversity. The non-parametric index utilised was the Shannon heterogeneity index: H'=-Σpi In(pi). The deviation of the Shannon index from the pre-treatment situation was statistically tested by means of the Mann-Whitney U-test.

30• Relative abundance of the functional groups. The relative abundance of the invertebrates classified as functional groups was estimated for each treatment.

31• Rank abundance models. This graph approach to the analysis of the biological structures consists in a conventional form of presenting the importance of each taxon as abundance (y-axis) with the different systematic units concerned arranged in rank-order along the x-axis from the commonest to the rarest. The pattern of the line connecting the taxa of each sample allows a visual inspection of the invertebrate structures: Sshaped curves are related to high heterogeneity values, on the contrary high slopes indícate more dominated taxonomic structures.

32The comparison of the curves pattern facilitates the inspection of the changes that can take place in the invertebrate structures during the different sampling periods.

33In the graphs, the species abundance are represented by means of the median values occurring during the pre-treatment, treatment and suspension periods.

34 Multivariate analysis. Due to the linear response of the invertebrate abundance the Principal Components Analysis (PCA) was preferred to the unimodal multivariate analysis approaches (i.e., redundancy analysis). The PCA was applied to the log-transformed abundance of the taxa collected by means of Surber net.

35To avoid the biological variation due to the different environmental conditions of the sampled rivers, outlined by a preliminary factorial analysis applied to an all rivers data matrix, the invertebrates collected in each river were analysed independently.

3. Results

3.1. Invertebrate structures analysis

36The results of the analyses of the invertebrate community structures are shown according to the type of sampling method used to collect the organisms: Surber samples, day and night drifts.

3.2. Surber samples

37The gathering and filtering collectors were the most abundant trophic groups analysed in the sampled comcommunities; the remaining three functional groups show a low contribution to the overall invertebrate abundance and for this reason these guilds are not illustrated in the graphs (Fig. 1). More in detail the pre-treatment samples show a dominance of the gathering collectors in Pru, Niandan and Kaba rivers and a co-dominance of the gathering and filtering collectors in the communities sampled in Maraoué river.

38With reference to the pre-treatment data, changes in the relative abundance of these two collector guilds are noticeable during the insecticides application in all rivers but Pru. These changes, that take place mainly during chlorophoxim, pyraclofos, phoxim and permethrin treatments, are the resuit of the increase of the gathering collectors, related to a decrease of the filtering collectors. A different variation occurs during B.t. treatments in Niandan and Kaba rivers for which the gathering invertebrates decrease and the filtering group increases.

39Regarding the communities sampled during temephos treatments, no appreciable changes occur in the guild structures with respect to the pre-treatment periods. A similar absence of changes is outlined for the invertebrate functional structures identified during the suspension periods; in these occasions the gathering and filtering collectors show percentages similar to the pretreatment periods.

40For the analysis of the diversity of the taxonomic invertebrate communities collected by means of the Surber net during the pre-treatment, treatment and suspension periods, the mean values of the Shannon heterogeneity index are shown in Fig. 2(a).

41The communities sampled in the Niandan river show the most noticeable reductions in the taxonomie diversity during the treatment periods with the greatest changes occurring after permethrin and phoxim treatments. During these periods the Shannon index shows the lowest mean values. The communities sampled during the suspension periods are characterised by lower heterogeneity values compared with the pre-treatment ones.

Fig. 1. Surber samples. Mean percentages of the functional groups sampled during the pre-treatment periods (no), at the end of the different treatment periods (x-axis) and during suspension periods (su); for the treatment code see Table 2. White bars gathering collectors, dark bars filtering collectors, the remaining three functional groups are not illustrated in the graphs because of their very low abundance.

42Maraoué and Kaba rivers show appreciable reductions of the heterogeneity values only for the communities sampled during chlorophoxim and pyraclofos treatments.

43Regarding the comparison of the Shannon index calculated for the biological data sampled during each treatment and those related to the pre-treatment situation, the Mann Whitney U-test reveals statistically lower values (95% confidence range) during temephos, B. t. and suspension periods for Niandan and during chlorophoxim for Maraoué.

44Fig. 2(b) shows the comparison of the taxonomic structures related to each treatment period with respect to the pre-treatment ones by means of rank abundance models.

45The greatest variations in the rank models occur in the Niandan and Kaba rivers for which the total abundance and heterogeneity of the sampled communities were generally reduced during the insecticides applications.

46For these two rivers the ranking gradients with respect to the pre-treatment data can be outlined as folfollows: B.t., temephos, pyraclofos, phoxim and permethrin; the models related to the suspension periods are located in an intermedíate position.

47In Pru river, excluding the model related to the suspension period which shows the highest heterogeneity, no differences are appreciable for the remaining models.

48In Maraoué, only the community structure sampled during chlorophoxim treatment shows a reduction both in heterogeneity and abundance.

3.3. Day drift

49The diversity indices calculated for the day drift invertebrate assemblages show patterns similar to those detected for the Surber sampled communities: no changes appear in Pru river, on the contrary, changes are evident for the invertebrates sampled in the remaining three rivers (Fig. 3(a)).

50With respect to the pre-treatment periods a clear decrease in the heterogeneity can be shown for the assemblages collected during pyraclofos (Kaba; H' approximately from 2 to 0.7), permethrin (Niandan; H' approximately from 1.6 to 0.8) and, in less degree, during temephos and chlorophoxim treatments. In all the sampled rivers the diversity indices calculated for the invertebrates collected during B.t. treatments show values similar to the pre-treatment ones. In the Niandan and Kaba rivers, for which the greatest variation of the heterogeneity indices was measured, the structures of the invertebrate assemblages collected during the suspension periods are characterised by low heterogeneity values. The comparison of the heterogeneity indices between pre-treatment and suspensions periods for the remaining two rivers points out no differences. With reference to the pre-treatment period, the Mann Whitney U-test reveals statistically lower values (P <0.05) during temephos, B.t. and suspension periods for Niandan river and during temephos for Maraoué river.

Fig. 2. Surber samples. (a) Mean values of the Shannon diversity index calculated for the invertebrate communities sampled during the pre-treatment periods (no), at the end of the different treatment periods and during suspension periods (su) (x-axis); for the treatment code see Table 2; (+) denotes a significant (95% confidence range) deviation from the pre-treatment situation according to the Mann Whitney U-test; id – insufficient data to apply the test; vertical bars desígnate 1 S.D. (b) Rank abundance models showing the departure of the invertebrate communities structures sampled at the end of the insecticides application from the pre-treatment condition (bold fines).

51The rank abundance models related to the day drift assemblages sampled during the treatments periods in Kaba, Niandan and Pru rivers (Fig. 3(b)) show a small change in the taxonomic structures from pre-treatment periods mainly due to a general decrease of the taxa abundance; no differences in the structures are evident for the communities sampled in Maraoué river.

3.4. Night drift

52As for the day drift, the mean values of the diversity indices calculated (Fig. 4(a)) for the night drift assemblages sampled in Pru river during the different treatments do not show appreciable variations for that the mean values remain close to 1.6. This value points out invertebrate assemblages characterised by a relatively high heterogeneity.

Fig. 3. Day drift. (a) Mean values of the Shannon diversity index calculated for the invertebrate communities sampled during the pretreatment periods (no), at the end of the different treatment periods and during suspension periods (su) (x-axis); for the treatment code see Table 3. (+) denotes a significant (95% confidence range) deviation from the pre-treatment situation aceording to the Mann Whitney U-test; id – insufficient data to apply the test; vertical bars designate 1 S.D. (b) Rank abundance models showing the departure of the invertebrate communities structures sampled at the end of the insecticides application from the pre-treatment condition (bold fines).

53For the remaining rivers the lowest diversity values of the invertebrate assemblages occur in Niandan river during phoxim and permethrin applications.

54According to the Mann Whitney U-test, significant difference from the pre-treatment heterogeneity, approximately ∆H'= 0.4-0.5, is recognisable for the values related to B. t. and temephos treatments in Maraoué river.

55Regarding the rank abundance models related to the night drift illustrated in Fig. 4(b), a general greater abundance of the systematic units can be outlined with respect to the drift invertebrate assemblages collected during the day. Only the invertebrate structures sampled in Kaba and Niandan rivers present changes according to the different treatments with the wider heterogeneity reductions related to the first river.

56Similar structure, in dominance, pattern and abundance values, is evident for the invertebrate assemblages sampled during the different treatment periods both in Maraoué and Pru rivers.

Fig. 4. Night drift. (a) Mean values of the Shannon diversity index calculated for the invertebrate sampled during the pre-treatment periods (no), at the end of the different treatment periods and during suspension periods (su) (x-axis); for the treatment code see Table 3. (+) denotes a significant (95% confidence range) deviation from the pre-treatment situation according to the Mann Whitney U-test; id – insufficient data to apply the test; vertical bars desígnate 1 S.D. (b) Rank abundance models showing the departure of the invertebrate communities structures sampled at the end of the insecticides application from the pre-treatment condition (bold lines).

3.5. Ordination

57The results of the ordination analysis of the sampled invertebrate communities are illustrated by means of functional graphs which consist in representing the sample co-ordinates along with the sampling time (Figs 5-9).

58For the description of the biological variation the first two components of the analysis have been used; in Table 3 the percentages of variance accounted for each axis are reported.

3.6. Niandan

59The first PCA axis scores show a cyclic pattern with maximum values occurring approximately every 6 1/2 months. This pattern is quite evident during 1988-1995 and can be explained with reference to the periodic variation of the river discharges (Fig. 6). The co-ordinate scores show the maximum values after the rain season then, as the dry season proceed, the values decrease and the following year the cycle repeats itself again.

Fig. 5. Niandan. Functional graphical presentation of the first two axes PCA scores (after 45° axes rotation). Arrows mark the end of the pre-treatment observations; white squares show biological data collected during suspension periods.

Fig. 6. Niandan. Functional graphical presentation of the first PCA axis scores after 45° axes rotation (dotted line) and river flows (lower line). Only the treatments period is shown.

Fig. 7. Maroué, functional graphical presentation of the first two axes PCA scores. Arrows mark the end of the pre-treatment obobservations; white squares show biological data collected during suspension periods.

60This regular variation of the invertebrate communities is not altered by the different treatments and the suspension period scores follow the cyclic pattern.

61The second axis scores show a change of the taxonomic invertebrate structures occurring at the beginning of the treatments (Fig. 5). This difference is mainly due to the reduction of the relative abundance of the Tricorythidae, Leptoceridae and Chironomini. It has to be noted that these changes are mainly related to a decrease of the scores variation that is quite wide during the pre-treatment period. During the treatments no further patterns are evident and the sample scores of the suspension periods still show low values.

3.7. Maraoué

62In Fig. 7, the first and second axes scores plotted along with the sampling time show no evident long-term trends or differences with respect to the pre-treatment scores. The invertebrate taxonomic structures show cyclic variation non-altered by the different insecticides applied during the investigated period.

63The second axis scores, positively correlated to Tricorythidae and Baetidae and negatively to Orthocladiinae, describe a reduction in the abundance of the first two systematic units associated to the increase of the third one as the treatments take place. The decrease of the second axis scores occurring at the beginning of the treatments disappears after 5 yr.

3.8. Kaba

64Although the unique pre-treatment observation does not allow a significant interpretation of the changes of the treated communities (Fig. 8), an increase of the first component scores does occur at the end the treatment period. The second component scores do not resuit structured along with the sampling time.

Fig. 8. Kaba, Functional graphical presentation of the first two axes PCA scores. In both graphs only the first point is a pre-treatment observation; white squares show biological data collected during suspension periods.

Fig. 9. Pru. Functional graphical presentation of the first two axes PCA scores. Arrows mark the end of the pre-treatment obserobservations; white squares show biological data collected during suspension periods.

Table 3. Percentages of variance accounted by each axis of the PCA

Axis 1

Axis 2

Axis 3

Axis 4

Maraoué

38

18

11

9

Pru

40

15

12

10

Niandan

43

22

11

8

Kaba

52

16

11

7

3.9. Pru

65The sample co-ordinates do not show changes as the treatments take place and no long-term variations resuit structured along with the sampling time (Fig. 9). With reference to the first component, it has to be noted that the wide variation of the suspension period scores at the end of the studied period.

4. Discussion

4.1. Comments on the invertebrate communities and assemblage structures

66The results of the analyses allow the following general considerations:

67• Although no absolute reference values exist for the Shannon index (in theory this index can increase to infinite but common values range between 1.5 and 3.5), the heterogeneity values and the rank-abundance models calculated during the pre-treatments periods are compatible with non-altered invertebrate communities.

68• The most informative invertebrate collections were those related to the Surber samples. These allow a detailed examination of the community changes in terms of both taxonomic and functional structures.

69With regard to the drift, the one collected during the night provides slightly better data for the analysis of the invertebrate structures than the day drift.

70• Pru river shows the lowest changes of the invertebrate structures sampled during the different treatments, the greatest changes occur for the invertebrates collected in Niandan and Kaba rivers.

71• Some of the insecticides show different efifects on the invertebrate structures depending on the river in which they are applied. For example, the diversity of the invertebrate assemblages collected in Niandan river was greatly reduced by phoxim, but no effect is recognisable during the same treatment in Pru river.

72A different response also occurs in relation to the type of invertebrate collection analysed: during temephos treatment in Kaba river, a reduction of the heterogeneity indices is detectable only for the day drift assemblages but no or very little changes are shown for the night drift assemblages or for the invertebrate communities sampled by means of Surber net. These differences can be partially explained with the different selective capture of the taxonomic units by the different sampling techniques applied.

73• Irrespective of the above-mentioned differences, the greatest reduction in the heterogeneity and abundance values of the invertebrate assemblages occur during phoxim, permethrin and pyraclofos treatments.

74• During the suspension periods, the invertebrate communities and assemblages do not show structures similar to the ones typical of the pre-treatment periods for that the diversity values and the rank models present more or less evident differences. This suggests that, although the invertebrates show a recover trend, the time for this to be completed strongly depends on the treatments that take place before the suspension periods.

4.2. Comments on the PCA

75The results of the PCA allow the following general considerations.

76For all the investigated rivers, the component scores related to the suspension samples follow the cyclic pattern defined by all the scores that suggest a seasonal, flow-related variation of the invertebrate community structures. This is particularly evident for the rivers showing cyclic variation of their samples’scores as Niandan (first axis in Figs. 5 and 6) or Maraoué (both axes in Fig. 7). In these situations the suspension samples are not altered by the treatments. Only for Niandan river the second axis scores clearly point out a relevant change in the community taxonomie structures occurring as the treatments took place and persisting during the suspension periods.

77No changes are evident for Pru river, while for Kaba river the unique pre-treatment sample does not allow an evaluation of the taxonomie variations that take place during the treatment periods.

5. Conclusions

78For the data analysis we faced different difficulties, among the sources of bias that can introduce additive variation to the biological data due to the natural environmental differences of the treated rivers and to the human factors, the following problems resuit more specific of the sampling protocol adopted:

79• A natural factor that can bias the biological data collected can be identified in the different hydraulic conditions occurring during the treatments. Actually, the operational use of insecticides was set up according to the discharges of the river to be treated, for example, during low discharges (up to 1 m3 s-1) only B.t. was applied and above 450 m3 s-1 only permethrin resulted appropriate. Because it is reasonable to assume that the drift and the benthic organisms react to the changing hydraulic conditions, this correspondent between discharges and treatment makes less comparable the biological data.

80• For some rivers, the number of pre-treatment samples is not sufficient in defining the invertebrate taxonomic and functional structures as well as the natural biological variation occurring before the treatment campaigns. Obviously, this makes more difficult the judgement of the biological changes induced by the treatment and, as the underlying abiotic and biotic processes and functions of running waters are still largely unknown in subtropics rivers, this insufficiency in reference data becomes a more serious problem in the investigated areas.

81• The rivers were differently treated with a rotational use of the insecticides. This cyclic sequence leads to difficulties in interpreting the biological data collected during specific periods because of the additional effects related to “what happend before”. This is clearly the case of the suspension periods during which the invertebrate recovery tendency is related to the severity of the stresses taking place during the past treatments.

82Since it is difficult to compare many graphs of taxa abundance against time, for the functional graphical presentation of the sample co-ordinates an appropriate technique has been revealed that it allows, for each sample, the inspection of the changes of the invertebrate structures with reference to the remaining ones in a single scatterplot, at the same time making it easy to relate each sample to the corresponding period.

83• Biological data often show outlier values that can bias the statistical analyses being, in most cases, related to factors extemal than those in study. Usually a great number of replicates can smooth their effects but in the specific study for operational constrains the number of replicates is limited.

84• The number of invertebrate collection replicates is not enough to show the distribution of the parent distributions sampled and this prevenís from applying parametric test for the comparison of the samples. This is a common problem in the sampling programmes encompassing wide temporal or spatial extents in which the logistic difficulties constrain the field Works.

85This forces the use of descriptive methods, like the rank abundance models, and quantitative ones, like the non-parametric Shannon diversity index and the PCA, to show and measure the biological variation without reference to the statistical descriptors of the data.

86• The effects of each treatment on the biota are river specific, for example, the faunistic changes that occur during pyraclofos and B.t. treatments are greater in the Kaba river than in Maraoué river. The greater induced stress detected in Kaba river is partially explainable because of its pristine environmental condition compared to those of Maraoué for which a number of stressing factors are recognisable.

87This example outlines the importance of considering the different anthropogenic pressures that have been taking place, with different time, in the treated rivers. Obviously, the induced stresses on the non-target fauna are less noticeable in those rivers in which the biological communities are altered because of extemal factors.

88Finally, it has to be outlined that the analysis strategy employed allows itself to face the possible source of bias included in the biological data collected.

89Considering that no unique standard analysis procedures are available for answering the question addressed in this paper or, more in general, in the longterm impact assessment studies, different numerical analyses are necessary to corroborate a comprehensive evaluation of the biological data.

90Regarding the results of the analyses applied to the biological data collected during the 20-yr monitoring programme a first synthesis can be shown for the response leveis of the four sampled rivers. Niandan river presents the most relevant changes in the invertebrate taxonomie and functional structures. Pru river shows the lowest biological variation; Kaba and Maraoué rivers are located in an intermediate levei with the wider biological variation limited, respectively, during pyraclofos and chlorophoxim treatments. For these two rivers it has to be noted that the low amount of pretreatment observations does not allow a significant comparison of the biological variation that takes place during the treatment periods.

91This synthesis is corroborated by the high similarity of the results obtained with all the analyses applied both to the taxonomie as well as to the functional classifications of the invertebrates.

92The most informative data collections for the above conclusions can be identified in the Surber samples.

93Regarding the leveis of the stressing factors induced on the aquatic fauna by the different treatments, the lowest have been revealed for temephos and B.t., the highest for the remaining insecticides.

94A tentative of classifying the applied insecticides on the basis of the stresses induced on the invertebrate communities can be drawn from the results of the taxonomic and functional structure analyses applied to the Surber samples, the gradient, ranking from low to high stress, results: temephos, B.t., chlorophoxim, permethrin, pyraclofos and phoxim.

95From a trophic point of view all the communities are dominated by the gathering collectors and, to a less extent, by the filtering collectors; the abundance of these feeding groups is a direct evidence of the availability of fine particulate organic matter (FPOM) that characterise the food resources within the studied rivers.

96This dominated structure tends to increase with the application of all insecticides but the B.t. During this biological treatment the two guilds show the highest evenness.

97The similar trophic structures shown by the communities sampled during the suspension periods with respect to the pre-treatment ones, as well as the changes occurring after the insecticide treatments, have to be positively considered, because they demonstrate the recovery potential of the guild structures analysed.

98In conclusion the data analyses demonstrate community levei effects of the insecticides applied during the OCP on the invertebrate fauna as well as the maintaining of their flow-related cyclic variation. This second aspect is encouraging and, though we did not discuss population levei data of different species so that no concluding remark is possible with respect to biodiversity, it can be concluded that the taxonomic and functional biological structures examined are not greatly altered from the range of biological variation that would normally occur in these river Systems. In the natural situation these river invertebrate communities would rarely be in equilibrium (constant in taxonomic and trophic composition) because of the natural driving forces, like drought and spate events, which would occur with great frequency and regularity. When these factors, the high priority of the human health protection objective of the programme and the criteria indicated within the mandate of the ecological group: “temporary and seasonal variation in invertebrate populations other then Simulium could be accepted” are taken into consideration, the biological variations previously discussed can be considered ecologically acceptable. Finally, it has to be considered that the programme succeeds in protecting 30 million people from onchocercal disease and it is estimated that none of the 9 million children that has been born within the OCP area since operation began has ever run the risk of contracting onchocercal blindness (Samba, 1994).

Acknowledgements

99Authors are grateful to the OCP Director Y.K. Dadzie for his support and encouragement and to OCP staff for their helpful collaboration during every phase of the work.

Bibliographie

References

Calamari, D., Yamèogo, L., Hougard, J.-M., Lévêque, C., 1998. Environmental assessment of larvicide use in the onchocerciasis control programme. Parasitol. Today 14 (12), 485-489.

Crosa, G., Yamèogo, L., Calamari, D., Hougard, J.M., 1998. Long term quantitative ecological assessment of insecticides treatments in four African rivers: a methodological approach. Chemosphere 37, 2847-2858.

Cummins, K.W., 1973. Trophic relations of aquatic insects. Ann. Rev. Entomology 18, 183-206.

Davies, J.B., Le Berre, R., Walsh, J.F., Cliff, B., 1978. Onchocerciasis and Simulium control in the Volta River Basin. Mosq. News 38, 466-472.

Elliott, J.M., Décamps, H., 1973. Guide pour l’analyse statistique des échantillons d’invertébrés bentiques. Ann. Limnol. 9, 79-120.

Guillet, P.M., Escaffre, M., Ouedraogo, M., Quillevere, D., 1980. Note préliminaire sur une résistance au téméphos dans le complexe Simulium damnosum (S. sanctipauli et S. soubrense) en Côte d’Ivoire. WHO 80.784 Technical Report, Zone du Programme de Lutte contre l’Onchocercose dans la région du bassin de la Volta, Geneva, CH.

Iltis, A., Lévêque, C., 1982. Caractéristiques physico-chimiques des rivières de Côte d’Ivôire. Rev. Hydrobiol. Trop. 15, 115-130.

Kurtak, D., Ouedraogo, M., Ocran, M., Barro, T., Guillet, P., 1982. Premilinary note on the appearance in Ivory Coast of resistance to chlorphoxim in Simulium soubrenselsanctipauli larvae already resistant to temephos (Abate®). WHO VBC/82.850 Technical Report, Geneva, CH.

Lévêque, C., 1989. The use of insecticides in the Onchocerciasis Control Programme and aquatic monitoring in West Africa. In: Bourdeau, P., Haines, J.A., Klein, W., Krishna Murti, C.R. (Eds.), Ecotoxicology and Climate, SCOPE 38, IPCS Joint Symposia 9, Wiley, Chichister, UK, pp. 317-335.

Lévêque, C., Odei, M., Pugh, M., 1979. The Onchocerciasis Control Programme and the monitoring of its effects on the riverine biology of the Volta River Basin. Linn. Soc. Symp. Ser. 5, 133-134.

Lévêque, C., Fairhurst, C.P., Abban, K., Paugy, D., Curtis, M.S., Traore, K., 1988. Onchocerciasis control programme in West Africa: ten years monitoring of fish populations. Chemosphere 17, 421-440.

Moniod, F., Pouyard, B., Sechet, P., 1977. Le Bassin du fleuve Volta, Cah ORSTOM Sér. Hydrologiques No. 5, Institut Français de Recherche Scientifique Pour le Développement en Coopération, Paris Cedex, France.

Philippon, B., 1977. Etude de la transmission d’Onchocerca volvulus (Leuckart, 1893) (Nematoda: Onchocercidae) par Simulium damnosum (Theobald, 1903). Travaux et documents ORSTOM No. 63, Institut Français de Recherche Scientifique Pour le Développement en Coopération, Paris Cedex, France.

Samba, E.M., 1994. The Onchocerciasis Control Programme in West Africa. Public Health in Action series, WHO, Geneva.

Yamèogo, L., Elouard, J.M., Simer, M., 1991. Typologies des entomocènoses benthiques soumises à des èpandages d’insecticides antisimulidiens. Arch. Hydrobiol. 123, 111-127.

Yamèogo, L., Lévêque, C., Traore, K., Fairhurst, C.P., 1988. Dix ans de surveillance de la faune aquatique des rivières d’Afrique de l’Ouest traitèes contre les simules (Diptera: Simuliidae) agents vecteurs de l’Onchocercose humaine. Naturalist 115, 287-298.

Zimmerman, P.A., Dadzie, K.Y., De Sole, G., Remme, J., Alley, E., Unnasch, T.R., 1992. Onchocerca volvulus DNA probe classification correlates with epidemiologic patterns of blindness. J. Infect. Dis. 165, 964-968.

Table des illustrations

Légende Fig. 1. Surber samples. Mean percentages of the functional groups sampled during the pre-treatment periods (no), at the end of the different treatment periods (x-axis) and during suspension periods (su); for the treatment code see Table 2. White bars gathering collectors, dark bars filtering collectors, the remaining three functional groups are not illustrated in the graphs because of their very low abundance.
URL http://books.openedition.org/irdeditions/docannexe/image/28644/img-1.jpg
Fichier image/jpeg, 27k
Légende Fig. 2. Surber samples. (a) Mean values of the Shannon diversity index calculated for the invertebrate communities sampled during the pre-treatment periods (no), at the end of the different treatment periods and during suspension periods (su) (x-axis); for the treatment code see Table 2; (+) denotes a significant (95% confidence range) deviation from the pre-treatment situation according to the Mann Whitney U-test; id – insufficient data to apply the test; vertical bars desígnate 1 S.D. (b) Rank abundance models showing the departure of the invertebrate communities structures sampled at the end of the insecticides application from the pre-treatment condition (bold fines).
URL http://books.openedition.org/irdeditions/docannexe/image/28644/img-2.jpg
Fichier image/jpeg, 41k
Légende Fig. 3. Day drift. (a) Mean values of the Shannon diversity index calculated for the invertebrate communities sampled during the pretreatment periods (no), at the end of the different treatment periods and during suspension periods (su) (x-axis); for the treatment code see Table 3. (+) denotes a significant (95% confidence range) deviation from the pre-treatment situation aceording to the Mann Whitney U-test; id – insufficient data to apply the test; vertical bars designate 1 S.D. (b) Rank abundance models showing the departure of the invertebrate communities structures sampled at the end of the insecticides application from the pre-treatment condition (bold fines).
URL http://books.openedition.org/irdeditions/docannexe/image/28644/img-3.jpg
Fichier image/jpeg, 38k
Légende Fig. 4. Night drift. (a) Mean values of the Shannon diversity index calculated for the invertebrate sampled during the pre-treatment periods (no), at the end of the different treatment periods and during suspension periods (su) (x-axis); for the treatment code see Table 3. (+) denotes a significant (95% confidence range) deviation from the pre-treatment situation according to the Mann Whitney U-test; id – insufficient data to apply the test; vertical bars desígnate 1 S.D. (b) Rank abundance models showing the departure of the invertebrate communities structures sampled at the end of the insecticides application from the pre-treatment condition (bold lines).
URL http://books.openedition.org/irdeditions/docannexe/image/28644/img-4.jpg
Fichier image/jpeg, 43k
Légende Fig. 5. Niandan. Functional graphical presentation of the first two axes PCA scores (after 45° axes rotation). Arrows mark the end of the pre-treatment observations; white squares show biological data collected during suspension periods.
URL http://books.openedition.org/irdeditions/docannexe/image/28644/img-5.jpg
Fichier image/jpeg, 27k
Légende Fig. 6. Niandan. Functional graphical presentation of the first PCA axis scores after 45° axes rotation (dotted line) and river flows (lower line). Only the treatments period is shown.
URL http://books.openedition.org/irdeditions/docannexe/image/28644/img-6.jpg
Fichier image/jpeg, 22k
Légende Fig. 7. Maroué, functional graphical presentation of the first two axes PCA scores. Arrows mark the end of the pre-treatment obobservations; white squares show biological data collected during suspension periods.
URL http://books.openedition.org/irdeditions/docannexe/image/28644/img-7.jpg
Fichier image/jpeg, 30k
Légende Fig. 8. Kaba, Functional graphical presentation of the first two axes PCA scores. In both graphs only the first point is a pre-treatment observation; white squares show biological data collected during suspension periods.
URL http://books.openedition.org/irdeditions/docannexe/image/28644/img-8.jpg
Fichier image/jpeg, 16k
Légende Fig. 9. Pru. Functional graphical presentation of the first two axes PCA scores. Arrows mark the end of the pre-treatment obserobservations; white squares show biological data collected during suspension periods.
URL http://books.openedition.org/irdeditions/docannexe/image/28644/img-9.jpg
Fichier image/jpeg, 22k

Auteurs

Water Research Institute, P.O. Box 38, Achimota, Ghana

National ONCHO Team, clo P.O. Box 529, Freetown, Sierra Leone

Le texte et les autres éléments (illustrations, fichiers annexes importés) sont sous Licence OpenEdition Books, sauf mention contraire.

Cette publication numérique est issue d’un traitement automatique par reconnaissance optique de caractères.
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search