Vulnerability of crop yields to variations in growing season precipitation in Uganda

Abstract

This study aims at assessing the vulnerability of six key crops (maize, beans, cassava, millet, groundnuts, and sweet potatoes) to variations in growing season precipitation at both the national and regional scale (southwest) in Uganda. To achieve this objective, a vulnerability model that is a function of sensitivity, exposure and adaptive capacity is used. Crop yield data for the period 1998–2017 for both the national scale and regional level analysis were collected from FAOSTAT and the Crop Yield Gap Atlas respectively. Precipitation data for the period 1998–2017 for both the national scale and regional level analysis were collected from the Climate Change Knowledge Portal of the World Bank Group. Adaptive capacity which reflects how the crops and the farmers growing these crops are adapted was measured using two proxies (literacy and poverty rates) and were collected from the Uganda Bureau of Statistics, UBOS (Uganda national household survey 2002/2003, 2002. https://www.ubos.org/wp-content/uploads/publications/03_2018unhs_200203_report.PDF; 2002 Uganda population and household census analytical report: education and literacy, 2006. http://www.ubos.org/onlinefiles/uploads/ubos/pdf%20documents/2002%20CensusEducAnalyticalReport.pdf; Uganda national household survey 2012/2013, 2014. https://www.ubos.org/wp-content/uploads/publications/04_20182012_13_UNHS_Final_Report.pdf; Uganda national household survey 2016/2017, 2018. https://www.ubos.org/wp-content/uploads/publications/03_20182016_UNHS_FINAL_REPORT.pdf) and Daniels (Measuring poverty trends in Uganda with non-monetary indicators, 2011. http://www.fao.org/fileadmin/templates/ess/pages/rural/wye_city_group/2011/documents/session3/Daniels_-_Paper.pdf). The results show that at the national scale, cassava is the most vulnerable crop while maize is the least vulnerable crop while at the regional scale millet is the most vulnerable crop and sweet potatoes are least vulnerable. At both scales, vulnerability has a positive relationship with exposure and sensitivity and an inverse relationship with adaptive capacity. The coefficients of determination at both scales are equally above 50% for most of the indices indicating that the models presented here are generally reliable and can be explained by the linear relationship between the variables. This study underscores the importance of regionally focused studies being that national and more global studies often have results that diverge from regional trends due to internal variations in precipitation, temperature, poverty and literacy rates at different levels. Proposing adaptation options that will be applicable at the regional scale must require studies like this involving results from both regional and national scale analysis.