Modelling and forecasting the fishery for pilchard (Sardina pilchardus) in Greek waters using ARIMA time-series models

Publisher Website
Google Scholar
Abstract
A 17-year record (1964–1980, 204 observations) of monthly catches of pilchard ( Sardina pilchardus ) from Greek waters was analysed using AutoRegressive Integrated Moving Average (ARIMA) techniques (Box and Jenkins, 1976). Two models were found to be suitable for describing the dynamics of the fishery and for forecasting up to 12 months ahead: X t =X t−1 +X t−12 −X t−13 +e t −0.539e t−12 −0.289e t−1 +0.156e t−13 and X 1 =X t−1 +X t−12 −X 1−13 +e t −0.537e t−12 −0.391e t−1 +0.21e t−13 −0.255e t−2 +0.137e t−14 , where X t = landings in month t, and e t = error term in month t. Forecasts were compared with actual data for 1981 that were not used in the estimation of the parameters of either model. Mean error was 14.6% and 12 % for the two models, respectively. This suggests that ARIMA procedures are capable of describing and forecasting the complex dynamics of the Greek pilchard fishery, which have hitherto been regarded as difficult to predict owing to the strong influence of year-to-year changes in oceanographic and biological conditions and socio-economic factors (low commercial value and demand, high discard rate).

This publication has 0 references indexed in Scilit: