Browsing by Author "Yeboah, Michael"

Now showing 1 - 1 of 1
  • No Thumbnail Available
    Item
    Markov chain modelling of rainfall: a case study of Eastern region of Ghana
    (KNUST, 2019-06) Yeboah, Michael
    It is a fact that Ghana Meteorological Agency (GMA) predicts Weather in Ghana. But empirical research has shown that GMA uses Weather Surveillance Radar (WSR) for weather prediction in the country and for reason of cost of equipment and maintenance the spread of monitoring stations are few. So, in order to keep costs down and make more specific weather predictions in remote areas of the region, a first-order Markov Chain was used to model the monthly rainfall data from 1998-2003. The twelve weather models were developed from Average Probability Vectors and Markov Chain Theory Inverse Technique under the assumption that in the long –term study of the weather, Average Probabilities equal steady-state vectors. The transition matrices showed the short-run behaviour of Markov chains while the equilibrium state probability indicated the long-run behaviour of the chains. The Markov Chain analysis for Eastern region in the short – run showed that the probability of rain there varies day by day whiles the long-term probability values also differ. There was rainfall throughout the year. The developed models revealed bimodal rainfall patterns in the study area. The least rainfall occurred in December and the wettest months are June and October. In the long-term analysis of the weather in the region, typically, there are approximately 64% of dry days and 36% of wet days indicating more dry days than wet days. Further observations made on the long-term probabilities revealed that areas in the region during the months of December, January, February and March tended to be dry. All because the long-term rainfall values were generally low. The developed transition matrices converge and the close conformance between the predicted figures and actual figures validates the use of Markov chains to model rainfall data in Eastern Region of Ghana. The developed Chains have been tested and found to be reliable and besides, predict weather accurately in the region, hence good for application by our National Weather Agency.