A Markov Chain Analysis of Wet and Dry Spell for Agricultural Crop Planning in the Middle Belt Region of Nigeria
Bernard Tarza Tyubee, Michael Terver Iwan

The frequencies of dry and wet spells and probabilities of dry and wet spell weeks were analysed from 1981 – 2010 in the Middle Belt region (MBR) of Nigeria. Daily rainfall (mm) data were obtained from eight synoptic weather stations spread across the region. A wet (dry) spell is defined as a consecutive number or group of days each with at least 1.2mm rainfall (<1.2mm rainfall), and a wet week is defined as a week with an accumulated rainfall amount of at least 8.4mm. Markov Chain probability model was used to analyse the probabilities of occurrences of dry and wet weeks. The result showed that the mean regional onset, cessation and duration of the rainy season were week 15th (9th – 15th April), week 42nd (15th – 21st October) and 29 weeks (203 days) respectively. The observed frequencies generally agreed with the estimated frequencies for wet spells for the entire region compared to dry spells. The initial and conditional probabilities of wet spell week were more than 50% from the week 22nd for all the stations at 8.4mm per week threshold. The study concludes that land preparation for crop planting should commence from week 20th - 22nd whereas crop irrigation farming should start after week 42nd in the Middle Belt region.

Full Text: PDF     DOI: 10.15640/jaes.v8n2a16