Fish‐based multimetric index for evaluating land use effects on large neotropical reservoirs.

Because of their importance to human well‐being and ecosystem services, it is important to have cost‐effective methods for assessing reservoir ecological condition or health. To do so, we constructed a preliminary reservoir fish assemblage index (RFAI) to assess the effect of predominantly agricultural land use on two oligotrophic neotropical reservoirs that differed in disturbance level. We first determined the disturbance levels around each reservoir through use of an integrated disturbance index (IDI). The IDI was calculated from the combination of land use metrics at local (Local Disturbance Index ‐ LDI) and buffer scales (Buffer Disturbance Index ‐ BDI). Afterwards, we employed the IDI to assess the response of candidate metrics to environmental disturbance in the RFAI construction. The final RFAI included three metrics selected by range, responsiveness, and redundancy tests. Metrics were scored continuously, and the final RFAI was divided in four quality classes: very poor, poor, acceptable and good. The results showed a clear disturbance gradient in both reservoirs with the IDI values varying between 0.07 and 1.33. However, the IDI values were greater around Volta Grande reservoir because of the predominance of agriculture land use (at local and buffer scales). The RFAI scores in Nova Ponte reservoir were significantly higher than in Volta Grande. All the sites scored as very poor or poor were in Volta Grande and all the sites scored as good were in Nova Ponte. In addition, RFAI scores in the reservoir river arms did not differ from those in the reservoir main body, indicating that it was applicable to both habitat types. Finally, our RFAI scores indicated no significant seasonal difference in the two reservoirs; however, the rainy season produced a greater range and more low scores, particularly in Volta Grande. We related this result to the harmful effect of agriculture and the surface runoff of fertilizers, herbicides, and pesticides. Thus, the RFAI was effective in detecting agricultural impacts at the site scale, even in oligotrophic reservoirs. [ABSTRACT FROM AUTHOR]