Prediction of long dry spells for appropriate cropping system in Gusau Northwestern Nigeria

Abstract

The objective of this study was to predict the probabilities of occurrences of long dry spells and their lengths during the planting period in rainfed farming season for future planning in Gusau and its environs North-Western Nigeria. Markov chain and probability distribution models were used to help predict in advance the longtime dry spells occurrences in the study area. Daily rainfall amount for each year was used to determine the probabilities of wet and dry days at different orders of Markov Chain. Gamma distribution was used with the help of INSTAT plus statistical package to estimate the length of dry spells in May, June and July. The early season dry spells were determined to occur usually between the first and the last decades of May. The Longest dry spells for the month of May were determined to be 26, 25 and 21 days in 2020, 2022 and 2030 respectively. Low frequencies of dry spells are to be anticipated in June with only 10 days in 2026 and July with only 12 days in 2024. The month of May from 2011 to 2020 with mean dry days of measured and predicted data were found to be 14 and 15 days respectively while coefficient of variation (CV) of 0.3, shows a stable dry spell in the coming years in May. The R between the observed and the predicted values were averagely good, mean error (ME) -1.25,-1.00 and 1.63 between the longest monthly observed and predicted dry spell were less than all the observed data. The root mean square error (RMSE) indicated that the month of June has the highest measure of precision 3.18, followed by the month of July4.46 and May 5.50. Since, early season rainfall is uncertain and erratic than the mid-season, early planting of moisture sensitive crop like maize in Gusau without supplementary irrigation would be highly risky.