Do highly anthropized hydrological conditions in marshes influence fish communities according to their life‐history strategies?

AbstractAlterations of natural hydrology in aquatic ecosystems are known to strongly impact the community composition of different taxa. Surprisingly, literature on the potential influence of hydrology on fish community composition is still very scarce in agricultural marshes, where canals represent one of the few remaining aquatic habitats. This study is aimed to address this research gap by monitoring fish communities in independent hydrological units differing in hydrology management over a 6 years period. We predicted variable fish responses to the hydrological context according to different life‐history strategies (opportunistic, equilibrium, or periodic species). Periodic and opportunistic species were the most frequently observed. Despite differences in hydrology between canals (but little variation over years), we found that hydrology explained only a very low proportion of variation in the composition of fish communities. In particular, the flooding duration of meadows in early spring did not influence the composition of fish communities, not even the abundance of periodic species expected to rely on such temporary habitats. Instead, fish communities were more influenced by local habitat variables (aquatic vegetation cover, turbidity, tree roots, and refuges under the canal banks). The hydrological management of most hydrological units for agricultural purposes (i.e., severe flood abatement in spring and shallow water depth in canals in summer) was found to be incompatible with conservation goals to promote more diverse fish communities between hydrological units. Therefore, we call for further investigations in similar habitats covering a larger range of hydrological conditions.
Methods
The study was conducted in the Marais poitevin, a ca. 1000-km2 French hypertrophic marshland located along the Atlantic coast. It has been historically divided into different hydrological units in which water management is controlled independently from the surrounding units because of different water regulations fixed by management committees. This leads to different hydrological patterns across units. We selected eleven hydrological units to represent the diversity of hydrology conditions in the study area.

Fish sampling

Each year, fish were sampled between the last week of June and the second week of July. Fish were sampled using the point abundance sampling approach (PAS; Copp and Persat, 1989; Nelva et al., 1979) with an electrofishing apparatus (Heron®, Aigrette®, or EFKO F.E.G. 8000® depending on the year, voltage during fishing operations 255±75V, amperage 7±5A). A total of 30 PAS were performed per canal, spaced out by at least 10 meters to provide independent samples. At each PAS, the anode was immersed about five meters in front of the boat and near the canal bank where fish generally concentrate. All shocked fish within 1 m2 of the impact point of the anode were caught, identified to the species level, measured (fork length, mm), and released back into the water.

Environmental variables
The water depth was calculated for different periods, and measured at two different spatial scales: locally, in the canal where fish were sampled, and across 10 small nearby canals connected to the fish canal in order to better characterize the hydrological conditions of the hydrological unit (average value). We calculated the average water depth for two seasons: winter and spring. Two additional hydrological variables were calculated at the hydrological unit scale to describe the rate of water depth changes at the pivotal period in early spring, namely the coefficients of variation of water depth in March and in April. We also assessed the amount of temporary fish-accessible aquatic habitats as expressed by the flooding duration (days) of at least 5% and at least 20% of the meadows in the whole hydrological unit. Calculations were done for spring and winter so that four flooding duration variables (two quantiles × two seasons) were used for subsequent analyses.

We also considered a series of water physico-chemistry variables, most of them collected as part of a companion monitoring program on the eutrophication of the Marais poitevin. Chlorophyll a content (µg.L-1), oxygen percentage saturation (%), conductivity (µS.cm-1), and ammonium, nitrite, nitrate and orthophosphate concentrations (mg.L-1) were measured in one large canal representative of each hydrological unit, all year round. We also measured basic variables – water temperature (°C), conductivity (µS.cm-1) and turbidity (m) – in the fish canal once a year, when sampling the fish. Lastly, we measured five habitat features: aquatic vegetation cover (%), tree roots and refuges under the banks, width (m) of the canal section where fish were sampled, and the total network length (km/ha) of large permanent canals and that of narrower temporarily dried canals in the hydrological unit.

Please read the article entitled "Do highly anthropized hydrological conditions in marshes influence fish communities according to their life-history strategies?" in River Research and Applications for more details.