Forecasting population dynamics of the black Amur bream (Megalobrama terminalis) in a large subtropical river using a univariate approach

Abstract

Understanding the stocks and trends of fish species using modern information technology is crucial for the sustainable use and protection of fishery resources. Megalobrama terminalis (Cyprinidae) is endemic to the large subtropical Pearl River (China) and is a commercially important species. Its population has however been suffering from long-term degradation. In this paper, a seasonal autoregressive integrated moving average (ARIMA) model and redundancy analysis (RDA) were proposed to predict larval abundance and its influence, using larva data collected every 2 days from 2006 to 2013. The ARIMA model provided good forecasting performance and estimated that the population trends will follow a relatively stable cycling trend in the near future. The cross-correlation function model further identified that discharge acted as a trigger for population growth; the effect of discharge on the number of larvae will last at least 5 days. However, the predicted breeding period will decrease significantly compared to 2006–2013. Such variability has the potential to cause a decrease in larva numbers that could destabilize the population dynamics of this species. We conclude that the ARIMA model approach is an especially promising tool for effectively predicting the population trend, which could be helpful in formulating realistic policies for effective fishery management.