Multivariate analysis of salinity effects on mayfly distributions along a complex river continuum.

Abstract

The association between the benthic community structure and the river gradient as described in the river continuum concept is evaluated in the saline stressed Washita River. Physical, chemical and biological variables were monitored at six sites in the river over 14 sampling periods during 1980 and 1981. Physical parameters represented stream discharge, sediment particle sizes, organic content of the sediment and related factors. Physical variables were measures of conductivity, pH and the major salts in the river (calcium, sulfate, sodium and chloride). Biological variables were subdivided into two groups of densities of mayfly genera from dredge samples and Dendy multiple-plate samplers, respectively. The Ephemeroptera represent one of the dominant groups of benthic insects in the river. The genera monitored provide a wide range of functional types (with respect to habitat selection, feeding preference and feeding modes).

Benthic community structure is influenced by physical gradients in the river continuum. But saline effects, although governed by geomorphological processes, do not conform to the "typical" river gradient. This localized saline impact influences the benthic community structure. Therefore, the river continuum concept must be modified if it is to be useful in the generation of ecological models in regions with high land and river salinity.

Two-dimensional ordinations were produced for the physical, chemical, dredge and Dendy groups of variables using the ALSCAL model of three-way nonmetric multidimensional scaling. The ALSCAL model summarizes variation over the 14 sampling periods. Scales were produced which represent river gradient effects and local effects for physical and chemical variables. Rank correlations of the dimensions from all ordinations suggest that both river gradient and local effects influence the structure of the benthic community. River gradient dimensions for physical and chemical variables are highly correlated with a dredge and a Dendy dimension. The remaining dredge dimension is correlated with local saline effects. The other Dendy dimension remains uncorrelated and may represent factors affecting colonization of Dendy samplers.