Ecology of the fisheries in the river Amazon and its tributaries in the Amazonas state (Brazil) / Miguel Petrere Jr.

Tese (Ph.D.) - University of East Anglia, , 1982

Both general and detailed analyses of the Amazon fishery from the Manaus fishing fleet, are presented based on fisheries catch data. (I) A description of the main characteristics of the fishery wherein the affects of the environment are shown to be strong. (a) Catch data are presented in relation to the year of capture, the rivers visited by the Manaus fishing fleet and within and outside a radius of 500 km from Manaus. The multigear character and the selective strategy of the fishery are described. (b) - From three blackwater regions, catches showed differences in relative composition. (c) - There is a lagged and positive correlation between the coefficients of variation of the monthly average distances travelled by the Manaus fishing fleet and the monthly mean water level of river Negro at Manaus.This is explained by the difficulty that fishermen have in locating the fish in the high water season. There is also a negative correlation between the monthly number of trips and the water level showing that the fishing effort decreases significantly in the high water. This is due to the increased dispersal of the fish throughout the environment. (II). A derivation of empirical models, based in regression analysis for predicting the annual catches in different rivers. (a) - The rivers in the Amazonas State were separated into two groups: rivers of 'normal' floodplain, which inundate on average 9,633 km² of floodplain per 1000 km of river channel, and rivers of 'extensive' floodplain which inundate on average 35,554 km² of floodplain per 1000 km of river channel. (a) - The rivers in the Amazonas State were separated into two groups: rivers of 'normal' floodplain, which inundate on average 9,633 km² of floodplain per 1000 km of river channel, and rivers of 'extensive' floodplain which inundate on average 35,554 km² of floodplain per 1000 km of river channel. (b) - Different variables were considered to be important in determining catch. These are: area of the floodplain (A, km²) inundated annually in each river corresponding to its length (L, km) reached by the Manaus fishing fleet in each year, the annual number of trips (T) per river, and the respective annual number of fishermen (F). Annual number of fishermen was found to be the most appropriate measure of fishing effort. Using results of partial correlation analysis it is concluded that fishermen are selecting the most productive floodplain for the fishing activities. (c) - Equations are presented for prediction of catches from different rivers for a given number of fishermen. (III) Analysis of catch and effort data of two important commercial species, tambaqui (Colossoma macropomum Cuvier, 1818) and tucunaré (Cichla ocellaris, C. temensis) caught in defferent lakes situated within a radius of 500 km from Manaus shows: (a) - The density of the stock of tambaqui vary significantly with affects of year and lake of capture, however the density of the stock of tucunaré varies also with the season of the year (high or low water).(b) - The density of the stocks of both fish increases with the distance of the lake from Manaus. (c) - There is a strong interaction between the lake and season affect upon the density of the stock of tucunaré. This is interpreted as being a consequence of lake morphology and vegetation coverage. (IV) Length frequency distributions of tambaqui landed at Manaus in 1977 and 1978 were analysed. (a) - It is found the total mortality rate (Z) in 1977 was 0.71y-¹ and in 1978 it was 0.74y-¹. (b) - The growth coefficient of the von Bertalanffy equation (K) is 0.227y-¹. (c) - The natural mortality rate (M) is of the order of 0.45y-¹. (d) - A yield per recruit analysis reveals that the stock is underexploited for values of M bigger than 0.30y-¹.