Open AccessMODELING AND FORECASTING HUMIDITY IN BANGLADESH: BOX-JENKINS APPROACH
Author(s) -
Mashfiqul Huq Chowdhury,
Somaresh Kumar Mondal,
Jobaidul Islam
Publication year - 2018
Publication title -
international journal of research - granthaalayah
Language(s) - English
Resource type - Journals
eISSN - 2394-3629
pISSN - 2350-0530
DOI - 10.29121/granthaalayah.v6.i4.2018.1475
Subject(s) - humidity , autoregressive integrated moving average , environmental science , box–jenkins , meteorology , relative humidity , atmospheric sciences , mathematics , climatology , statistics , geography , time series , geology
Humidity (atmospheric moisture) is an important atmospheric component and has significant influence on plant growth and development. The rate of growth and the form that a plant attains is controlled by humidity. The present study is an attempt to analyze the seasonal humidity’s of Bangladesh by employing appropriate statistical techniques. The main objective of this study is to examine humidity over time in Bangladesh and find a suitable model for forecasting. This study utilizes humidity data from Bangladesh Meteorological Department (BMD), recorded at 6 divisional meteorological stations for the period of 1976 to 2015. This study found that annual average humidity of Bangladesh is 78.88%. Initially data set is checked for whether it is stationary or not through Augmented Dickey Fuller test. Data was found non-stationary but it was transformed to stationary after taking first difference. Then seasonal ARIMA model was built using Box and Jenkins approach. After examining of all diagnostic procedures, ARIMA (2,0,1)(2,1,1)12 model has been identified as an appropriate model for forecasting 60 months (2016-2020) seasonal humidity.