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

Références des articles présentés sur le cédérom / Papers references presented on CD-ROM

Analysis of the effects of rotational larviciding on aquatic fauna of two Guinean rivers: the case of permethrin

G. Crosa, L. Yaméogo, D. Calamari, M.E. Diop, K. Nabé et F. Kondé

Résumé

Within the Onchocerciasis Control Programme about 50,000 km of west African rivers have been regularly sprayed with larvicides to control the vector of dermal filariasis caused by Onchocerca volvulus. Since the beginning of the programme invertebrates and fish data were collected to monitor adverse effects on non-target organisms. The regular series of biological and hydrological data collected in two Guinean rivers were analysed to evaluate the effects of rotational larviciding with particular attention to permethrin, as preliminary acute toxicology tests and semi-field experiments suggest it has stronger effects on non-target fauna in respect to other larvicides. Invertebrates ard fish variations in biomass and species richness are seasonal and flow-related and the results presented here do not support any evidence of specific effects of permethrin application on the biological targets monitored. Larvicide applications influence community structures, putting pressure on some taxonomic groups, causing. for example, the rarefacion of some taxa. In spite of the above results, the scarcity of some invertebrate systematic units does not resuit in a significant reduction of total invertebrate density because of the corresponding increase in other systematic units. In nature the studied aquatic communities would rarely be in equilibrium because of frequent natural stresses, such as drougst and spate events, the biologicai variations discussed are to be considered ecologically acceptable. © 2001 Elsevier Science Ltd. All rights reserved.

Note de l’auteur

Corresponding author.

Texte intégral

1Received 3 March 1999; accepted 10 July 2000

1. Introduction

2In 1974 WHO launched the Onchocerciasis Control Programme for the elimination of an illness of public importance in west Africa, river blindness. The programme was based on the control of the blackfly Simulium damnosum, vector of the filariasis caused by the parasitic worm Onchocerca volvulus (WHO, 1997).

3The strategy of larviciding was designed to interrupt the transmission of the parasite for longer thin the longevity of their adult worms in the human host tes rimated to be about 14 years) by destroying larval stages of the vector through the aerial application of insecticides at breeding sites in the rivers. The developnent oc the aquatic stage from egg to pupae is around one week hence, insecticide was applied weekly. At the peak of larviciding activities about 50.000 km of river: were treated weekly in 11 countries in West Africa The biological insecticide Bacillus thuringiensis stotype H-14 (a biologicai control agent) and temepres. and organophosphorus degradable molecule were, until 1979, the only insecticides applied. Following the development of resistance in vector populations, the strategy was to alternate other insecticides such as chlorphoxim and pyraclophos (organophosphorus compounds), permetrin (pyrethroid), carbosulfan (carbamate) and etofenprox (pseudopyrethroid) (Hougard et al., 1997).

4Because of the extent of the aquatic environment exposed to regular larviciding and the consequent possible effects on non-target organisms, larvicides applications, both in terms of compounds and frequency of spraying, were planned and carefully monitored from the beginning of the programme, taking into account the criteria and advice of an independent ecological group. A comprehensive review of the environmental assessment of larvicide use in the programme has been recently published (Calamari et al., 1998).

5Permethrin was selected in 1984 for potential operational use following the increase in resistance of target populations to applied larvicides. Preliminary assessments of its effects on aquatic fauna were made through acute toxicity on fish (Yaméogo et al., 1991) and by means of semi-field experiments in mini-gutters (Yaméogo et al., 1992). Following the results of these studies, a pilot scale field trial was conducted on the Sassandra river in Cote d’Ivoire to evaluate its impact in operational conditions (Yaméogo et al., 1993).

6The resuit of this study led to the conclusion that the conditions of fish and fisheries before and after the experiment period were significantly unchanged, and thus, operational use of permethrin by the programme would not be expected to have permanent adverse effects on non-target fauna (Lévêque et al., 1998)

7However, this research demonstrated that permethrin has a stronger impact on invertebrates than other insecticides used in the programme (Yaméogo et al., 1992). It has been therefore decided to pay particular attention to a few rivers where the Chemical larvicides permethrin and organophosphates were applied. Organophosphates on several occasions had limited impact on aquatic fauna when used only in combination with the biological insecticide B.t. H-4 (Crosa et al., 1998).

8Detailed, long-term series of biological data were available to make a more specific assessment for permethrin. This paper presents and comments on the results of a comprehensive analysis of the biological data collected during the monitoring programme in two Guinean rivers.

9The objective of this paper is the detailed evaluation of the short-, medium-and long-term biological variations related to larvicide applications with particular attention to permethrin.

10To account for biological variations due to the strong dependence of aquatic organisms on the seasonal fluctuation of river discharges, the results of the data analyses are evaluated and commented on with reference to the hydrological condition of the river.

2. Treatments, biological data, rivers hydrology and analysis strategy

11The analysis regards the invertebrate and the fish sampled in the rivers Niandan (Sansambaya, field station) and Milo (Boussolé, field station), respectively. (Fig. 1). The rivers are located in Guinea and were selected from among other Guinean rivers because of their dctaiied series of biological and hydrological data, as well as their type of treatments.

Fig. 1. Fish (black square) and invertebrate (black star) sampling sites location, respectively, on Milo and Niandan rivers.

2.1. Treatments

12The rivers were treated with B.t. HI4, the organophosphates, themephos and pyraclophos, and the pyrethroid permethrin.

13The length of time, the number of treatments and the available discharge data as well as the number of invertebrate and fish samples for each monitored year are shown in Fig. 2. Invertebrates and fish monitoring started 3 and 2 years, respectively, before the beginning of the treatments.

2.2. Biological data

14Invertebrates were collected in shallow riffles over rock substrate using a modified 25 x 25 cm Surber sampler (Dejoux et al., 1979). For each site five samples were normally replicated and the mean number of individuals as well as 95% confidence limits were calculated.

15All sampled individuals were classified according to their family level 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).

Fig. 2. Schematic representation of the periods of the data analysed. The numbers of the weekly larviciding plus permethrin and the numbers of invertebrates and fish samples arc dctaiied for each year.

16The variables used for the analyses are the total invertebrate individuals collected in each Surber sample (individuals/625 cm2) and the percentage of each trophic guild. Principal component analyses (PCA) were applied to the relative log-transformed invertebrate abundance to detect any change in taxonomie composition or relative abundance of each systematic unit.

17The fish were sampled since 1980 with standard batteries of five mesh size gill nets: 15, 20, 25, 30 and 40 mm. The number and total weight of individuals for each species caught by the different mesh size were calculated, all the results were standardised as catch per unit effort (CPUE) which is the number, or the weight, of fish caught in 100m2 of net per night (Abban et al., 1997).

18The variables used for the analyses are the total catch per unit of effort expressed as number and weight of specimens calculated for each of the five mesh sizes and specific richness. For the principal component analysis the relative abundance of each species was used.

2.3. Hydrological data

19The water regime of the studied rivers is characterised by a well-defined seasonal pattern with high water periods from June-July to December followed by low water periods from January to May–June.

20The hydrographs show a time-related increase of the water volumes flowing yearly through the Milo measure section Figs. 8–10. For the Niandan discharges, two low water periods were recorded during 1992–1993 and no long-term trend is evident (Figs. 3 and 7).

2.4. Data analysis

21The numerical analysis strategy was selected in order to assess community structure variations as well as invertebrate and fish reductions in terms of total number of individuals or species richness.

22For assessing community structure changes, principal component analysis (SAS, 1989) was applied to the logtransformed abundance of the invertebrate and fish taxa collected. The results of this ordination analysis are illustrated by means of functional graphs, which consist in representing the sample co-ordinates along with the sampling time.

23PCA is one of the multivariate techniques that arrange samples along axes (principal component) on the basis of data pertaining to species composition: presence and relative abundance. The aim of this technique is to ordinate the samples so that samples with similar axis scores correspond to sites that are similar in species composition, and samples that are far apart correspond to sites that are dissimilar in species composition

Fig. 3. Niandan hydrograph and invertebrate samples ordination scores (dotted line), percentage of total data variability 21% AXI. 10% AX2. In the upper part of the graphs permethrin applications are marked with black crosses.

Fig. 4. Relationship between the invertebrate samples ordination scores (1st axis) and the invertebrate density; each sample is marked as follows: o = pre-treatment period, x = first year of larvicides application, i = 2nd to last year of larvicides application.

24Due to the strong and seasonal relationship expected between the river discharges and the aquatic community structures, the biological variations measured were related to hydrological conditions in order to assess natural biological changes with respect to larvicide applications.

3. Results

3.1. Invertebrates

3.1.1. Taxonomic classification

25The results of the ordination analysis of Niandan invertebrate communities and the trend in discharge for the period 1988–1996 are illustrated by means of functional graphs in Fig. 3. In the upper part of the graphs, black crosses mark the periods of permethrin.

26The ordination axes account for 31% of total invertebrates variance: 21% first axis, 10% second axis. Considering the high variability of aquatic communities. such percentages can be considered adequate to analyse the major specific sources of variation.

Fig. 5. Detailed representation of discharge and invertebrate density trends during 1993–1995 for Niandan river. Dotted line marks invertebrate densities, the crosses in the upper part of the graph mark permethrin and the arrows mark pyraclofos applications.

27The sample co-ordinates of the first axis (Fig. 3 upper) show no long-term trend, in that the communities sampled during the treatment period have ranges and patterns of variation similar to those occurring during the pre-treatment periods. This pattern results from the cyclic variation of the sample co-ordinates throughout the investigated period with low values at the end of wet seasons and an increasing trend during dry seasons.

28With respect to the second axis, a clear, long-term trend of sample co-ordinates is evident from the graph (Fig. 3 lower). Although maximum values during the pre-treatment period still occurred throughout the research, the variation ranges of sample scores were reduced soon after the beginning of the treatments.

29Biological fluctuation mainly reflects the seasonal variation of invertebrate densities as demonstrated by the significant correlation between the first axis sample co-ordinates and the mean invertebrate density (r = 0.8, P <0.001). This relationship is shown in Fig. 4 where pre-treatment samples, the samples collected during the first year of treatment and the remaining ones are diflerently marked. This graph clearly demonstrates that the sample scores present high, intermediate or low values without any reference to the three different operational phases of the programme (pre-treatment, first year of larviciding, succeeding years of larviciding).

30Excluding the data collected during 1995 (for which a detailed discussion is provided in the following) no significant changes of this seasonal pattern occur during the years of larvicide application.

31To explain invertebrate community changes previously outlined during 1995 by the first co-ordinate scores, and for a better evaluation of the short-term effects of permethrin and pyraclofos applications, invertebrate data collected during 1993–1995 have been analysed in detail. For this, invertebrate densities and discharge values are plotted on the same graph with the indication of permethrin and pyraclofos application as described in Fig. 5.

32The graph indicates low invertebrate density values occurring during the first months of 1995. Because the hydrograph of the previous wet season (1994) is not significantly different from those of other years, the outlined low density values might be partially explained by larvicide applications. A specific effect of permethrin cannot be reasonably assumed because more permethrin application occurred during the 1993 wet season compared to the 1994 wet season and no such reduction of invertebrate densities can be shown for the following dry season.

33On the contrary, the low density values measured during the first months of 1995 can be partially explained considering that, in this period, invertebrate sampling took place shortly after the weekly applications of pyraclofos. For example, during April 1995 Surber samples were collected 17 days after pyraclofos spraying, while during April 1994 the invertebrates were sampled twice, 70 and 79 days after pyraclofos treatments, a sufficient lag time for invertebrate communities to recover.

34Considering that the long-term trend of the invertebrate structure illustrated by the second axis scores in Fig. 4 is not related to river hydrology, a detailed evaluation of taxonomic units showing significant and timecorrelated trends is illustrated in Fig. 6.

Fig. 6. Niandan average invertebrate densities (individuals/625 cm3). Inner box = standard error, whisker = standard deviation.

35Tricorythidae and Leptoceridae decrease in abundance as treatments take place and become increasingly rare from 1990–1991 until the final year of monitoring. On the contrary, the low, pre-treatment densities of both Hydropsychidae and Philopotamidae increase from 1991–1992.

36As a result of this relative shift of taxa abundance, total invertebrate density was not significantly modified within the studied period. The only remarkable invertebrate density reduction occurred during 1995 and was due to the short lag period between treatments and sampling which did not allow time for recovery. It should be noted that this biological alteration did not last long and was completely reversed after the 1995 wet season.

3.1.2. Functional classification

37The two most abundant invertebrate trophic guilds are the gathering and the filtering collectors that, on an average, contain more than 95% of total invertebrates.

38The relative percentage variation of these two guilds during the study period is represented in Fig. 7 superimposed on the river discharge. In the saine figure, permethrin applications are marked with black crosses.

39The percentage values show a cyclic pattern characterised by the dominance of gathering colleetors at the beginning of the dry season shifting to the dominance of the filtering collector group in the months before the beginning of the wet season.

40With reference to pre-treatment years, both trophic guilds present a wider seasonal pattern in their relative abundance for all the monitored years after the beginning of larviciding. This “amplification” of a seasonal and natural biological variation can thus be attributed to larvicide applications. It should be considered, however, that, although the invertebrate communities show a marked reduction in relative abundance of gathering collectors at the end of the dry season (and a corresponding increase of the filtering collectors), after each wet season the trophic compositions of the communities arc similar to those in pre-treatment sampling.

Fig. 7. Niandan flow rates and relative percentages of the two most abundant functional groups of the invertebrate communities (dotted line). In the upper part of the graphs permethrin applications are marked with black crosses.

41In Fig. 7, the periods of permethrin applications do not show corresponding eflects on variation in invertebrate trophic structure commented above.

3.2. Fish

3.2.1. Species structure and fish abundance

42Taxonomie structure changes of fish communities sampled in the Milo river during 1985–1998 are shown in Fig. 8 by means of the first two axes scores plotted along with the sampling time. In the same graphs the river discharges are available from 1988.

43The first two axes account for 25% of the total variance of the sampled communities (15% firs axis, 10% second axis). As previously noted, for invertebrates the percentage of variation can be considered adequate to identify major sources of variation in the fish communities.

44The biological changes described by the first axis scores show two variation patterns with respect to time. The first is related to the increase in sample scores during the dry seasons of 1995–1996 and 1998, the second one is the seasonal, flow-related, variation of the sample scores showing low values during the high flow periods and high values during the dry seasons. No short-term eflects of these patterns are demonstrable after permethrin applications.

45The first axis sample scores are significantly related to the CPUEs (r = 0.63, P> 0.001) showing high scores in concomitance of high catch values recorded during the low water periods lasting from November until January (Fig. 9).

Fig. 8. Milo hydrograph and fish samples ordination scores (dotted line), percentage of total data variability 15% AX1.10% AX2. In the upper part of the graphs permethrin applications are marked with black crosses.

46Visual inspection of long-term variations in CPUEs shows a reduction of the mean yearly catches during the period 1988–1993. Although this reduction can partially be the resuit of the initial phase of the treatments, it can be better explained by the low discharge values of those years. In fact, as the hydrologic regime shows an increase in volume of flowing water (1994–1997), the CPUEs increase in spite of larvicide applications.

3.2.2. Fish species richness

47The number of fish species caught during the monitoring period shows seasonal fluctuation with maximum values at low water periods and a slight increment during the last 5 years (Fig. 10). Both these variations can be related to the number of fish caught, in particular the seasonal fluctuation can partly be explained by the maximum efficiency of the gill nets during the dry season. The long-term trend of species richness corresponds almost exactly to CPUE values whose variation can be ascribed to long-term flow variation. No reduction in fish species occurs during the final years of larvicide application.

4. Conclusion

48Data analysis of fish and invertebrate monitoring in two Guinean rivers treated with B.t. temephos, pyraclophos and permethrin allows for the first time a comprehensive field assessment of the impact of permethrin on aquatic life when used in rotation with other Chemicals already evaluated as having limited or no impact.

49The two biological components investigated for detecting adverse environmental responses after larvicide applications, in particular permethrin, produced a body of experimental evidence on the biological variation occurring in the rivers during the monitoring period. In general, most of this variation is strongly, long-or shortterm related to the hydrological river conditions. Evidence of this finding is provided, for example, by the Milo river, where the long-term trend in fish catches follows the increase in discharges or, more in general, for the invertebrates for which short-term seasonal variation, both in terms of total density and trophic structure, are significantly related to the river discharge.

Fig. 9. Milo hydrograph and trend in fish caiches per 100 m2 of gill nets (dotted line). CPUE in number of individuals. Black crosses mark permethrin applications.

Fig. 10. Milo hydrograph and trend in number of fish species per sample during the monitoring period (dotted line). In the upper part of the graph permethrin applications are marked with black crosses.

50Flow and treatment-related aspects of biological response differ for invertebrates and fish according to the different durations of their life cycles. Invertebrates, in fact, with their short life cycles are better indicators of short-term Chemical shock and seasonal flow changes than fish, for which the seasonal (short-term), variations are mainly related to the high efficiency of the gill nets during low water periods.

51Results on fish communities do not suggest any reduction in CPUEs or species richness that can be attributed to larvicide applications; observed variations are mainly flow-related.

52The invertebrates collected have been classifiai according to their taxonomic levels as well as to their functional feeding group (trophic role). Analysis of biological variation focused on both these structural and functional attributes. Whereas the first aspect concerns, besides the faunistic interest over loss of global biodiversity, the quality of biomass available for the upper trophic levels. the second is related to the stability of energy flows. On account of the invertebrate position in the first levels of the river food webs, changes here can signal greater detrimental effects on the ecological characteristics of the whole river System. In this context most invertebrate variations are seasonal and flowrelated and the results do not support any evidence of specific effects of permethrin application on the biological targets monitored.

53Larvicide applications have influenced the community structures by putting pressure on some taxonomical groups, causing for example the rarefaction of Tricorythidae and Leptoceridae in Niandan river since 1990. In spite of the above results, the scarcity of some systematic invertebrate units does not resuit in a significant reduction of total invertebrate density because of the corresponding increase in other systematic units as Hydropsychidae and Philopotamidae.

54Only occasionally, for example, during 1995 in the Niandan river, can the reduction of total invertebrate density be related to larvicide applications; however, the experimental evidence previously discussed reveals that this adverse biological response is a particular, shortlasting case: invertebrate density completely recovered in 3–8 weeks.

55Invertebrate trophic guilds with an increase in their seasonal variation during the treatment period completely recovered at the end of each high water period.

56Considering the short-time recovery capabilities of invertebrate communities, it can be assumed that sampling sites will be recolonised by taxa shortly after the conclusion of the programme.

57The biological variations discussed above are to be considered ecologically acceptable, according to the rationale for evaluating adverse effects of larvicide applications on aquatic fauna indicated by the ecological group (Calamari, 1998): “temporary and seasonal variation in invertebrate populations other than Simulium could be accepted” and considering that in nature aquatic communities would rarely be in equilibrium because of frequent natural stresses, like drought and spate events.

58In conclusion the acceptability of the discussed results is in agreement with the two fundamental principles for which the monitoring programme was organised (Lévêque et al., 1998):

  1. repeated long-term treatments could change the reproductive cycle of fish, either by affecting their physiology or by acting directly on eggs or offspring. If so, there could be changes in fish recruitment and, on a long-term basis, a decrease in fish abundance. This would apply to the fish community as a whole, or to particular species, which could be more sensitive to insecticides.
  2. Insecticides could affect the food chain leading to a serious reduction in diet.

59Note: A complete set of data is available upon request.

Acknowledgements

60The authors are grateful to OCP Director, Yankum K. Datzie for making this article possible.

Bibliographie

References

Abban, E.K., Yameogo, L., Paugy, D., Traorè, K., Diop, M.E., Samba, E.M., 1997. The fish monitoring programme of the Onchocerciasis Control Programme in west Africa: a model for fish and fisheries preservation in the face of development in African inland fisheries. In: Remane, K. (Ed.), Aquaculture and the Environment FAO/Fishing News Books. Blackwell Scientific Publication, Oxford, pp. 136–149.

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

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

Cummins, K.W., 1973. Trophic relations of aquatic insects. Annu. Rev. Entomol. 18, 183–206.

Dejoux, C., Elouard, J.M., Lévêque, C., Troubat, J.J., 1979. La lutte contre Simulium damnosum en Afrique de l’ouest et la protection du milieu aquatique. C. R. du Congrès de Marseille sur la lutte contre les insectes en milieu tropical, II, pp. 873–883.

Hougard, J.M., Yaméogo, L., Seketeli, A., Boatin, B., Dadzie, K.Y., 1997. Twenty-two years of blackfly control in the onchocerciasis control programme in west Africa. Parassitology Today 13 (11), 425–431.

Lévêque, C., Fairhust, C.P., Abban, E.K., Paugy, D., Curtis, M.S., Traore, K., 1998. Onchocerciasis control programme in west Africa: ten years monitoring of fish populations. Chemosphere 17, 421–440.

SAS Institute Inc, 1989. SAS/STAST® User guide Version 6, fourth ed., vol. 1&2. Cary NC: SAS Institute Inc, p. 1787.

World Health Organization, 1997. Twenty years of onchocerciasis control in west Africa. Review of the Work of the Onchocerciasis Control Programme in west Africa 1974–1984, WHO.

Yaméogo, L., Abban, E.K., Elouard, J.M., Traore, K-, Calamari, D., 1993. Affects of permethrin as Simulium larvicide on non-target aquatic fauna in an African river. Ecotoxicology 2, 157–174.

Yaméogo, L., Elouard, J.M., Simier, M., 1992. Typology of susceptibilities of aquatic insect larvae to different larvicides in a tropical environment. Chemosphere 24, 2009–2020.

Yaméogo, L., Tabsoba, J.M., Calamari, D., 1991. Laboratory toxicity of potential blackfly larvicides on some African fish species in the onchocerciasis control programme area. Ecotoxicol. Environ. Saf. 21, 248–256.

Table des illustrations

Légende Fig. 1. Fish (black square) and invertebrate (black star) sampling sites location, respectively, on Milo and Niandan rivers.
URL http://books.openedition.org/irdeditions/docannexe/image/28794/img-1.jpg
Fichier image/jpeg, 99k
Légende Fig. 2. Schematic representation of the periods of the data analysed. The numbers of the weekly larviciding plus permethrin and the numbers of invertebrates and fish samples arc dctaiied for each year.
URL http://books.openedition.org/irdeditions/docannexe/image/28794/img-2.jpg
Fichier image/jpeg, 71k
Légende Fig. 3. Niandan hydrograph and invertebrate samples ordination scores (dotted line), percentage of total data variability 21% AXI. 10% AX2. In the upper part of the graphs permethrin applications are marked with black crosses.
URL http://books.openedition.org/irdeditions/docannexe/image/28794/img-3.jpg
Fichier image/jpeg, 100k
Légende Fig. 4. Relationship between the invertebrate samples ordination scores (1st axis) and the invertebrate density; each sample is marked as follows: o = pre-treatment period, x = first year of larvicides application, i = 2nd to last year of larvicides application.
URL http://books.openedition.org/irdeditions/docannexe/image/28794/img-4.jpg
Fichier image/jpeg, 28k
Légende Fig. 5. Detailed representation of discharge and invertebrate density trends during 1993–1995 for Niandan river. Dotted line marks invertebrate densities, the crosses in the upper part of the graph mark permethrin and the arrows mark pyraclofos applications.
URL http://books.openedition.org/irdeditions/docannexe/image/28794/img-5.jpg
Fichier image/jpeg, 49k
Légende Fig. 6. Niandan average invertebrate densities (individuals/625 cm3). Inner box = standard error, whisker = standard deviation.
URL http://books.openedition.org/irdeditions/docannexe/image/28794/img-6.jpg
Fichier image/jpeg, 104k
Légende Fig. 7. Niandan flow rates and relative percentages of the two most abundant functional groups of the invertebrate communities (dotted line). In the upper part of the graphs permethrin applications are marked with black crosses.
URL http://books.openedition.org/irdeditions/docannexe/image/28794/img-7.jpg
Fichier image/jpeg, 130k
Légende Fig. 8. Milo hydrograph and fish samples ordination scores (dotted line), percentage of total data variability 15% AX1.10% AX2. In the upper part of the graphs permethrin applications are marked with black crosses.
URL http://books.openedition.org/irdeditions/docannexe/image/28794/img-8.jpg
Fichier image/jpeg, 106k
Légende Fig. 9. Milo hydrograph and trend in fish caiches per 100 m2 of gill nets (dotted line). CPUE in number of individuals. Black crosses mark permethrin applications.
URL http://books.openedition.org/irdeditions/docannexe/image/28794/img-9.jpg
Fichier image/jpeg, 61k
Légende Fig. 10. Milo hydrograph and trend in number of fish species per sample during the monitoring period (dotted line). In the upper part of the graph permethrin applications are marked with black crosses.
URL http://books.openedition.org/irdeditions/docannexe/image/28794/img-10.jpg
Fichier image/jpeg, 70k

Auteurs

© IRD Éditions, 2003

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