Enhanced prediction methods for the onset and cessation dates of the rainy season over the Guinea Savanna, Nigeria.

The need for an effective method of prediction of the onset and cessation dates of the rainy season cannot be overstressed. It becomes more imperative considering that the mainstay of the economy of the tropical Guinea Savanna region and the country at large is agriculture that is still rain-fed. Thus, an effective prediction of the onset dates of the rainy season is the first step to ensure food security. Two efficient methods were developed—the hybrid and the modified Anyadike's methods for the prediction of the onset dates of the rainy season and tested for the Guinea Savanna region of the country. Three stations were chosen: Lokoja, Bida and Jos. The methods proved to perform better than the earlier known and well-used method in the region—the Walter method. While the Walter method had about 12 false starts in Lokoja and 7 in Jos, the modified Anyadike had 5 and 3, respectively. However, the hybrid method had 1 and 2, respectively. The bias method also showed the hybrid method followed by the modified Anyadike's methods outperformed the Walter and Anyadike's methods. This is due to the overestimation associated with the Walter method and the underestimation by the Anyadike's method. Therefore, the two developed methods are recommended for use in the agricultural planning for the Guinea savanna region of Nigeria and any other region with a similar climate. Thus, it is recommended to be applied to other tropical Guinea savanna regions. [ABSTRACT FROM AUTHOR]