Stochastic models for some meteorological outcomes in Niger Delta region of Nigeria

Abstract

In this paper, stochastic models based on autoregressive integrated moving average models of various orders and its seasonalized versions are presented, with a view to identifying the optimal model for some meteorological outcomes in some cities in Niger Delta region of Nigeria, using Normalized Bayesian Information Criterion as model selection criterion. ARIMA (p, d, q) and seasonal ARIMA (p, d, q)(P, D, Q)[c] are formulated for pressure, cloud, air temperature, relative humidity, wind direction, wind speed, rain amount, and sun hours of Calabar, Portharcourt, Uyo, Owerri, Warri, Benin city, Ondo and Akure on daily basis. The choice of autoregressive integrated moving average models of orders p and q is intended to retain any persistence in natural process. Parameters of the models are estimated, causality and invertiblity behaviour of some optimal models are also presented.

Keywords: Meteorological outcomes , stochastic model, Autoreressive model, seasonal effect, SARIMA, BIC