EXPLORATORY DATA ANALYSIS FOR ANOMALY DETECTION IN ILLEGAL FISHING RECORDS IN THE AMAZON BASIN

Resumo

The present study analyzed the records of illegal fishing in the Amazon Basin, based on the seizure records drawn up by the Brazilian Institute for the Environment and Natural Renewable Resources (IBAMA) in Manaus City, Amazon State, from year of 1992 to 2017. A total of 1,875 records (539.36 tons of apprehended fish), Purus (230.31 tons), Japura (179.71 tons), Amazonas (95 tons), 02 ton) and Juruá (69.91 tons), with a total fish seized around 2,765.70 tons. A total of 34 fish species were identified, the main ones were pirarucu (Arapaima gigas) with 66.12%, followed by tambaqui (Colossoma macropomum) with 21.70% and Surubim (Pseudoplatystoma fasciatum) with 2.82%, the other species grouped together 9,36% of the seizures. Were applied the technique of exploratory data analysis associated to the control chart for the detection of outliers in the apprehended fish record time series of the data, in which anomalies were verified in the quantitative of fish seized from the years 1999, 2000, 2001 and 2003, which oscillated at 409.81 tons, 493.37 tons, 530.59 tons and 310.79 tons of apprehended fish, respectively. There was a marked decrease in seizures after the validity of the Defense Law, emphasizing that its due compliance, combined with the intensification of inspections, can make sustainability of inventories feasible. However, it is immature to say that this decrease behaviour occurred only due to the creation of the aforementioned Law, but others causes should also be considered in the current scenario, as external factors that may have influenced the result, such as the reduction of existing fish stocks, difficult access to fishing environments and the lack of effective monitoring and surveillance of fisheries in these areas, as well as the decentralization of inspecione, previously performed exclusively by IBAMA, to the others governamental agencies.