Réseaux trophiques et production piscicole en étangs fertilisés (Dordogne, France) Food webs and fish production in fertilized ponds (Dordogne, France)

Résumé

Des séries de données (physicochimie, plancton, poisson) collectées pendant deux saisons consécutives sur deux groupes d'étangs de pisciculture de Dordogne (Sud-ouest de la France) ont été analysées statistiquement (analyses en composantes principales ACP, analyse des correspondances multiples ACM, tris croisés). Des liens significatifs entre les différents constituants de la chaîne trophique ont été identifiés, et les facteurs clés de la production piscicole mis en évidence. En début de saison, une régulation ascendante oriente la composition et l'abondance de la flore algale, suivant les concentrations de N et PO₄ et leurs proportions relatives. L'abondance du zooplancton filtreur est liée à la fois aux ressources trophiques disponibles (cellules algales aisément ingérables) et à la pression de prédation qu'il subit de la part des juvéniles de poissons. La production piscicole de gardon (Rutilus rutilus) et de tanche (Tinca tinca) est plus faible si ces espèces sont associées (compétition alimentaire interspécifique chez les juvéniles) ou produites en présence d'un carnivore. En contre-partie d'une production piscicole plus faible, la stabilité de la structure du réseau trophique est accrue. Les concentrations en P-PO₄, la quantité de grands Crustacés planctoniques et une densité de poissons adaptée aux capacités du milieu constituent les facteurs clés de la production piscicole finale.

Mots-clés:

• Étang piscicole,
• réseau trophique,
• plancton,
• poisson

Abstract

The main parameters (physicochemistry, plankton, fish) of two sets of fish ponds (8 and 9 ponds respectively; area between 0.2 and 1.2 ha; mean depth: 1.5 m) located in south-west of France were monitored during two consecutive seasons (1995, 1996). The benthic macroinvertebrates and the microbial loop were not taken into account for these field studies. No macrophytes were growing into the studied ponds. The multivariate analyses of the two yearly data bases provided informations on the strength of the various links among and between the successive compartments of the food web (physicochemistry of water, phytoplankton, zooplankton and fish) and identified the bottlenecks of the food web to the fish production.

Integrated water samples were collected bimonthly (from May to September) for field measurements (T°, pH, dissolved O₂, Secchi disc transparency, conductivity) and laboratory analyses ([N.NH₄], [N.NO₂], [N.NO₃], [ortho-P], [Ca] using a HACH spectrophotometer). The groups of planktonic algae (Diatoms, Desmidieae, Dinophyceae, Chlorococcales and Volvocales, Euglenophyceae, Cyanobacteriae) were identified and classified in three categories (dominant, present, rare). Rotifers, Cladocerans and Copepods were numbered (N/litre) into two size classes (large or small, with a threshold at 400 µm). One hundred liters of pond water were filtered through a plankton net (mesh size: 80 µm) and the volume of sedimented zooplankters measured as zooplankton biovolume (mL/100 liters). The total numbers of fish stocked and collected are noted (weight and number of fish/species; age classes). Working on private production ponds, we were not allowed to make catches for stomach contents studies.

The data from each season were analysed through PCA, MCA, crosstabs (STATBOX software from GRIMMER); for each species of fish a ratio was computed between the biomass of the collected juveniles and the biomass of the stocked breeders.

In the studied fish ponds like in other eutrophic bodies of water, the densities of the planktonic organisms follow a bimodal evolution, with high numbers at the end of spring (June) and the end of summer (August), separated by a phase of more transparent water. During the first phase, Daphnia and large Cladocerans are predominant with Chlorophyceae, replaced later in the season by species of smaller size and shorter life cycle (Bosmina, Ceriodaphnia, Cyclops), associated with Cyanobacteriae, at a time where predation by fish is more intense as their biomass increase. In spring, different algal communities develop with density depending on the initial concentrations of inorganic N and PO₄ and the PO₄ /N ratio. If the available dissolved N and P are infra-optimal, mixotroph algae (Chrysophyceae, Dinophyceae) and Rotifers predominate. When [N] is not limiting (N/P between 4/1 and 10/1), Chlorophyceae become predominant, but if PO₄ is in relative excess, they are replaced by Cyanobacteria. The abundance of grazing Cladocerans is linked to the capacity of the available algal cells to be ingested. Cyanobacteria can already become predominant in May because of the selective grazing by Cladocerans on more readily ingestible phytoplankton taxa. Links among groups of zooplankton are evidenced: small grazers (Rotifers, Cladocera < 0.4 mm) attain high densities only when there is neither competition from larger Cladocerans nor predation by Copepods.

During the second season as the management of the pond fertilization improved, the links between the zooplankton biovolume and the fish production became more evident through the statistical analysis. High net fish productions are linked to high zooplankton biovolumes (> 2 mL/100 L), a water transparency between 0.30 and 0.70 m and are related to a shift from small size species (< 400 µm) in spring to larger specimen in summer (Cladocerans and Copepods > 400 µm). In spring numerous small zooplankters facilitate the survival of the fish fry, whose growth is later ensured by high densities of larger Crustaceans.

From the fish production data, the ratios between the biomass of collected juveniles and the biomass of stocked breeders were compared under different rearing conditions for the various species. Roach Rutilus rutilus, tench Tinca tinca and gudgeon Gobio gobio have similar production potentials when reared separately (juveniles production in kg is 5 to 7 times the biomass of the breeders). When roach and tench are reared together under the same conditions, the ratio is divided by 2 for each species, but the total net production of fish remains at the same level (500-600 kg/ha). This shows that there is a strong feeding competition on zooplankton between the juveniles of roach and tench. In opposition, there is little feeding competition between roach and gudgeon. When the two benthic species are associated, tench strongly dominates the gudgeon. The addition of a carnivorous species to the association of roach + tench reduces the final total production by 250-300 kg/ha. Between 30 and 40 kg of carnivorous fish (Esox lucius or Stizostedion lucioperca) are produced at the expense of the equivalent of a final biomass of 200 kg/ha of young Cyprinids. The pressure of predation from these two top consumers is 2 to 3 times less intense compared to the predation exerted by the large-mouth bass Micropterus salmoides.

As the basic food for juvenile fish, the zooplankton can be considered here as the key element of the food web, between the bottom-up (production) and top-down (predation) effects. The ortho-phosphate concentration is the limiting factor for the ascending regulation of the food web, and the amount of initially stocked fish regulates the top-down effects. The maximum net final fish production reached about 800 kg/ha with roach alone, or in association with tench. In such cases, Cyanobacteria were dominant in the pond with the presence of Chlorophyceae. The low density of zooplankton and the sub-optimal dissolved oxygen concentrations during the day seem to indicate that at such production level, the food chain is near to collapse.

Keywords:

• Fish pond,
• food web,
• plankton,
• fish