Fishing and drought effects on fish assemblages of the central Amazon Basin

One method with considerable potential for understanding fishing and environmental impacts on fish assemblages is size spectra (SS) and diversity size spectra (DSS) analysis-a regression of Ln abundance or Shannon diversity against the natural logarithm of body size of fish assemblages. But the usefulness of this method for application to tropical freshwater fish assemblages is uncertain. Here, we assessed the extent to which SS and DSS analyses explained changes in the exploited fish assemblage related to fishing effort and river droughts. To do this, we used correlation analyses on historical datasets of river water level, fishing effort, and fish length measurements for 56 fish species in three of the largest rivers of the Amazon Basin. In calculating the SS and DSS analysis statistics, we found that linear regressions adjusted well to the diversity and Ln(abundance) data plotted against Ln(fork length). In analysing SS and DSS statistics in relation to fishing and environmental impacts, we found that the slope of SS was negatively correlated to drought intensity with a lag of zero years in all rivers. The slope of SS also was negatively correlated to fishing effort with a lag of three years in two rivers and a lag of two years in one year. The slope of DSS was not correlated to drought intensity and fishing effort in any of the rivers. Our results provide support for the use of SS analysis to investigate fishing and environmental effects on exploited fish assemblages in tropical freshwater environments.