The Southern Oscillation and Prediction of “Der” Season Rainfall in Somalia

P. Hutchinson Natural Resources Institute, Overseas Development Administration, Chatham, United Kingdom

Search for other papers by P. Hutchinson in
Current site
Google Scholar
PubMed
Close
Restricted access

Abstract

Somalia survives in semiarid to arid conditions, with annual rainfall totals rarely exceeding 700 mm, which are divided between two seasons. Many areas are arid, with negligible precipitation. Seasonal totals are highly variable. Thus, any seasonal rainfall forecast would be of significant importance to both the agricultural and animal husbandry communities. An investigation was carried out to determine whether there is a relationship between the Southern Oscillation and seasonal rainfall. No relationship exists between the Southern Oscillation and rainfall during the midyear “Gu” season, but it is shown that the year-end “Der” season precipitation is attected by the Southern Oscillation in southern and central areas of Somalia. Three techniques were used: correlation, regression, and simple contingency tables. Correlations between the SOI (Southern Oscillation index) and seasonal rainfall vary from zero up to about −0.8, with higher correlations in the south, both for individual stations and for area-averaged rainfall. Regression provides some predictive capacity, but the “explanation” of the variation in rainfall is not particularly high. The contingency tables revealed that there were very few occasions of both high SOI and high seasonal rainfall, although there was a wide scatter of seasonal rainfall associated with a low SOI.

It is concluded that the SOI would be useful for planners, governments, and agencies as one tool in food/famine early warning but that the relationships are not strong enough for the average farmer to place much reliance on forecasts produced solely using the SOI.

Abstract

Somalia survives in semiarid to arid conditions, with annual rainfall totals rarely exceeding 700 mm, which are divided between two seasons. Many areas are arid, with negligible precipitation. Seasonal totals are highly variable. Thus, any seasonal rainfall forecast would be of significant importance to both the agricultural and animal husbandry communities. An investigation was carried out to determine whether there is a relationship between the Southern Oscillation and seasonal rainfall. No relationship exists between the Southern Oscillation and rainfall during the midyear “Gu” season, but it is shown that the year-end “Der” season precipitation is attected by the Southern Oscillation in southern and central areas of Somalia. Three techniques were used: correlation, regression, and simple contingency tables. Correlations between the SOI (Southern Oscillation index) and seasonal rainfall vary from zero up to about −0.8, with higher correlations in the south, both for individual stations and for area-averaged rainfall. Regression provides some predictive capacity, but the “explanation” of the variation in rainfall is not particularly high. The contingency tables revealed that there were very few occasions of both high SOI and high seasonal rainfall, although there was a wide scatter of seasonal rainfall associated with a low SOI.

It is concluded that the SOI would be useful for planners, governments, and agencies as one tool in food/famine early warning but that the relationships are not strong enough for the average farmer to place much reliance on forecasts produced solely using the SOI.

Save