World Environment 2016, 6(1): 1-9
DOI: 10.5923/j.env.20160601.01
Sarima Modelling and Forecasting of Monthly Rainfall in
the Brong Ahafo Region of Ghana
Ebenezer Afrifa-Yamoah
1,*
, Bashiru I. I. Saeed
2
, Azumah Karim
3
1
Graduate Studies, Kwame Nkrumah Univ. of Science Technology, Kumasi, Ghana
2
Mathematics and Statistics Department, Kumasi Polytechnic, Ghana
3
Al-Huda Islamic School, Kumasi, Ghana
Abstract Statistical modelling and data analysis is a crucial instrument in understanding rainfall processes. This article
aims at examining the rainfall pattern and fit a suitable model for rainfall prediction in the Brong Ahafo (BA) Region of
Ghana. Data from 1975 to 2009 were collected from the Department of Meteorology and Climatology in the BA region. The
results revealed that the region experience much rainfall in the months of September and October, and least amount of rainfall
in the months of January, December and February. SARIMA (0,0,0)×(1,1,1)
12
, with an AIC score of 8.985894, has been
identified as an appropriate model for predicting monthly average rainfall figures for the Brong Ahafo Region of Ghana and
hope that when adopted by the Ghana Metrological Agency and other relevant governmental organisations like the National
Disaster Management Organisation (NADMO), it will in the long run help in accurate forecasting and education of the
populace on rainfall expectancies.
Keywords Statistical modelling, Forecast, Decision making, Brong Ahafo, Ghana
1. Introduction
Natural resource management is heavily driven by climate.
Far reaching impacts of climate change are being felt in
ecological, resource-degradation and hydrological process in
Ghana. A typical example in Ghana is the constant decrease
in the water levels of the Akosombo Hydropower Generation
Dam, which is the main source of power generation thus the
persistent “dumsor”. The changing climatic conditions have
immensely affected rainfall patterns in Ghana, recording
downward figures since 1993, which, Kwamena & Benneh
(1995) attributed it to the activities of the over 70% citizenry
direct or indirect involvement in rain dependencies
agriculture. The Forestry Commission of Ghana estimated
8.2 million hectares of forest in 1990 reduced to 1.2 million
hectares in 1997 as a result of illegal logging, overgrazing,
bad mining practices, industrialization, environmental
pollution and urbanization. The irony of the situation is that
Ghana is a rain dependent agricultural country.
Amidst this background however, flood and rainstorm
have been identified under hydro-meteorological disasters in
Ghana’s disaster profile for which technical committees have
been tasked to develop various hazard maps and to draw
education and training programme, emergency preparedness
and response plans and advice for the government and
* Corresponding author:
kingmorose@ymail.com (Afrifa-Yamoah, Ebenezer)
Published online at http://journal.sapub.org/env
Copyright © 2016 Scientific & Academic Publishing. All Rights Reserved
major stakeholders (NADMO, 2015). The BA Region has
had its fair share of flood. In 1999, the northern part of BA
Region was among the victims of the “northern floods” in
which over three thousand persons were affected. The flood
re-occurred in 2007 and affected over three hundred
thousand (307,127) persons (NADMO, 2015). A major
shortfall in disaster management in developing countries,
like Ghana, is the limited capacity of responsible agencies to
issue correct and up-to-date happenings of events to the
populace. This has resulted in deaths and loses of valuable
properties that could have been saved. To complement the
efforts of these technical committees, resourced materials
must be made available. There is the need for appropriate
modelling, if possible, of relevant events of life to adequately
inform decision making processes, thus the need for this
study.
Several studies on climatic conditions have been
undertaken to understand and identify the changing trend in
temperature, rainfall, humility and other related conditions
with temperature studies have witnessed consistent
agreement in results among researchers (Chin (1977); Coe
(1982); Cowpertwait (1994); Chandler (1997); Dunn (2004);
Carmona (2005): Beatrice et al (2014); R.C. Noven et al
(2015)). However, rainfall studies have resulted in diverse
results in same spatial area. A typical case in literature is the
study in the Meditarranean, researchers have reported
diverse variability patterns (Maheras (1988), Maheras &
Vafiaris (1990), Benifo, Orellara & Zunta (1994). In Ghana,
climate studies have been conducted and reported in
literature. Several attempts have been made by researchers to