Choice of biota in stream assessment and monitoring programs in tropical streams: A comparison of diatoms,macroinvertebrates and fish

The objective of this study was to compare the responses of diatoms, macroinvertebrates and fish to agriculture, urbanization and mining in the Manyame River Basin. Water quality sampling and benthic diatom, macroinvertebrate and fish community data were collected in April (end of the rain season) and September (dry season) 2013 at 44 sampling stations spread out across three land-use categories: commercial agricultural, communal agricultural and urban-mining areas. Commercial agricultural areas were relatively pristine as they were characterized by mature deciduous riparian forest strips which acted as riparian buffers thus protecting water resources from nonpoint source pollution. In communal agricultural areas a combination of poor agricultural practices (stream bank cultivation, overgrazing, soil erosions) and high human population densities had negative effects on water quality of streams draining these areas. Streams in urban-mining areas were highly stressed, being impacted primarily by physical habitat degradation and both point and nonpoint sources of pollution. A suite of environmental variables that varied with land-use pattern was assessed to find the combination of variables that best explained patterns of biota community composition. Community metrics i.e. the Trophic Diatom Index (TDI) based on diatoms, the South African Scoring system version 5 (SASS 5) based on macroinvertebrates and the Fish Assemblage Integrity Index (FAII) were used to determine the ecological status of study streams in relation to human-induced stressors. Data were also subjected to multivariate statistical techniques; canonical correspondence analysis (CCA), mantel test and cluster analysis to determine environmental gradients along which the diatom, macroinvertebrate and fish assemblages were distributed as well as to elucidate hypothesized differences in response to stressors among communities per land-use type. Using CCA, we assessed the individual importance of a set of environmental variables on each biotic community structure. ANOVA, showed a significant difference (p < 0.05) in physical and chemical variables among commercial agricultural, communal agricultural and urban-mining sampling stations with no significant differences (p > 0.05) between the 2 sampling periods. Based on CCAs carried out using individual variables, the strengths of relationships between diatoms and macroinvertebrates was generally high for nutrient levels, organic and metal pollution and other variables. However, fish assemblages showed a relatively low association with all water quality variables in the study; this might be explained by the high abundance of omnivores and air breathers which are able to tolerate a variety of environmental conditions. These patterns were also confirmed by the mantel test as well as the other CCAs carried out to investigate the simultaneous effects of environmental variables. These findings indicate that diatoms are more powerful indicators in accessing ecological stream/river quality and have potential for application in routine monitoring programs in tropical streams.