Analysis of Seasonal Rainfall Variability for Agricultural Water Resource Management in Southern Region, Ethiopia

摘要

The livelihood of over 80% of the people of Southern region of Ethiopia is dependent on agriculture. However, the region is known by its food insecurity problem due to increasing population growth and alarmingly increasing natural resource degradation as well as unpredictability of rainfall. Therefore, it needs to assess the seasonal rainfall variability in selected areas of the region. Instat software version 3.36 was used to analyze and estimate the onset and end of the growing season, and the Length of Growing Period (LGP). Trend analysis for 17 rainfall stations’ data was also made by Mann-Kendall, Spearman test and Ordinary Least Square (OLS) methods. The Pattern of variability (coherent rainfall variability areas) was also identified by using Principal Component Analysis. The result shows that, the annual rainfall in the region varied from slightly over 780 mm in Billate station to more than 2110 mm in Gerese station with mean, SD and CV as 1200 mm, 197 mm and 25% respectively and that of kiremt varied from 157 mm in Konso to 844 mm in Welkite with mean, SD and CV as 506 mm, 202 mm and 39% respectively. Belg seasonal rain varied from 863 mm in Gerese to 246mm in Bue with mean, SD and CV of 409mm, 121mm and 30% respectively. CV of 15%-64% for kiremt, 17%-52% for March April May (MAM) and 12%-46% for annual were observed. For Kiremt season, CV greater than 30% was observed in 18 stations; between 20-30% in 10 stations and below 20% for 5 stations. The mean onset, end and LGP of the main growing season was found to be at Days of year (DOY) 92, DOY 286, and 193 days for Hosaina area; at DOY 117, DOY 290 and 169 days for Welkite area; and at DOY 84, DOY 146 and 62days for Gato area respectively. OLS apparently showed a non significant decreasing linear trend for rainfall amount in 10 stations out of 17 while significant decreasing trend at Sawla; p= 0.04(-9.15 mm/yr) and Chida; p= 0.05(-16.08 mm/yr). However, for the start of MAM, March was unreliable in Hosaina and Welkite while reliable in Gato area. In addition, two homogenous areas of coherent rainfall variability, in terms of both seasons were obtained. This classification could be used for regional water management and rainfall prediction. Therefore, time of planting crops and other soil and water conservation activities should be performed accounting these variability parameters. Key words: Variability, Seasonal rainfall, rainfall trend, DOY (days of the year), Onset, End date, LGP, Homogenous rainfall areas.