macroinvertebrate community responses to long term ...

1 downloads 0 Views 731KB Size Report
Apr 15, 2002 - v-03 fev. -0. 3 ma rc h-0. 3 apr-0. 3 ma y-0. 3 ju ne-0. 3 ju ly-0. 3 aug. -0. 3 se pt-0. 3 oct-03. Time. Ca no nic al Coef fic ie nt (c dt). 5 µg/L. 25 µg/ ...
MACROINVERTEBRATE COMMUNITY RESPONSES TO LONG TERM COPPER EXPOSURE – A LOTIC MESCOSM EXPERIMENT Sandrine Joachim (1), Hélène Roussel (1), Sylvain Lamothe (1), Patrick Baudoin (1), Eric Thybaud (1), Jean-Marc Bonzom (1)(2) (1) National Institut of Industrial and Environmental Risks (INERIS), Laboratory of ecotoxicology, 60550 Verneuil-en-Halatte (2) Current address : Institut of Radioprotection and Nuclear Safety (IRSN), Laboratory de radioecology and ecotoxicology, Cadarache, 13115 Saint Paul Lez Durance

Introduction

In excess, trace elements such as copper can cause severe disturbances in aquatic ecosystems. Anthropic activities such as mining and farming greatly contribute to copper pollution. Macroinvertebrates who have an important position in aquatic and avian food webs are known to be adversely affected by copper. But few long term chronic results are available. To approach ecological realism, a mesocosm experiment was put together in order to assess effects of copper on the structure and functioning of a lotic ecosystem during 18 months. Results of the response of macroinvertebrates are presented below. Water flow 0m

20 m

Figure 1 : View of the mesocosms

Natural and artificial sediment

Material and Methods Experimental Design:: 12 lotic mesocosms of 20 meters long and 1 meter wide. Coarse-grain sediments and a water depth of 30 cm is found upstream. Fine grain sediments and a water depth of 70 cm is found dowstream (Figure 1).

(b)

(a)

Exposure : Flow-through conditions from the 15th April 2002 to the 15th of October 2003 with three copper concentrations (delivered as copper sulphate) 5, 25, et 75 µg/L tri-replicated and three controls.

(c )

Sampling: Invertebrates were sampled each month using different traps (Figure 2 a, b, c).

Figure 2 : Location of traps (a) landing nets (b) tiles, (c ) tubes

Statistical Analysis

The effects of copper on macroinvertebrates was analysed by the Principal Response Curves (PRC) method which is based on Redundancy Analysis, a constrained form of Principal Component Analysis (Van den Brink et Ter Braak, 1999). The PRC results in a diagram showing the sampling weeks in the x-axis and the first Principal Component on the y-axis. The diagram shows the deviations in time of the different treatments compared to the control. The species’ weight, on the right side of the diagram can be interpretrated as the contribution of each species to the response of the community (Figure 3). A high positive weight of a taxon (ex. : Lymnaea) indicates that abundances have decreased in the higher treatments. A high negative weight indicates, on the contrary an increase in abundance. Monte Carlo permutation were used to assess the significance of the PRC for each sampling date. The univariate Williams test was applied to determine the NOECcommunity and to assess the significance of the treatments through time at the species level.

Results

(a) Abundance of Lymnaea adult s

Principal Response Curves for invertebrates

4

+

+

Lymnae a j uveni les

3,5

+ Abunda nc e

2,5

+

7 5 µg/ L

1,5

1

-0,5

Discussion

oc t-03

a ug-03

se pt-0 3

july-0 3

a pr -03

+

+ 10 -oct-0 3

6 -déc -0 2

8 -nov-0 2

11-oct-0 2

13 -sept-0 2

14 -a oût-0 2

19 -juil-0 2

2 6 -avr -0 2

2 1-juin-0 2

+

12 -sept-0 3

+

Time Abundance of Chironomidae

-1,5

10 0 0

+

+ +

Abunda nc e

10 0

Gasteropods of the genus Lymnaea et Physa are the taxas the most affected by copper followed by Crustaceans of the genus Gammarus and Asellus. On the contrary, Chironomidae seem to be « favoured » by high copper concentrations. Indirect effects such as a decrease of predation and/or an increase in food ressources and/or an increase in favourable ecological niches could explain these results. Further analysis of all of the components of the ecosystem is needed to test these assumptions.

june -0 3

f e v-03

Gammarus 2 4 -ma i-0 2

1-mar s-0 2

-1

Time

+ +

+

+

(c)

+ +

+

+

10

+

1 c ontrol 5 µg/ L 2 5 µg/ L 7 5 µg/ L oc t-03

a ug-03

f e v-03

de c -02

0 2 -janv-0 3

nov-02

oc t-02

a ug-02

july-02

se pt-0 2

a pr -02

ma y-0 2

june -0 2

2 9 -ma r c h-0 2

3 1-janv-0 3

+

0 ,1 0 1-ma r c h-0 2

Forty-two percent of the data is explained by time while 33% is explained by treatment. The PRC shows that the macroinvertebrate community, contaminated at 25 and 75 µg/L, deviates from the control. Permutation tests also indicate a significant treatment effect on all sampling dates after the start of the contamination (Figure 3). A NOECcommunity of