University of Twente Student Theses


Estimating Consumer Food Accessibility Across Urban-Rural Catchment Areas in Ethiopia

Campomanes V, Florencio (2022) Estimating Consumer Food Accessibility Across Urban-Rural Catchment Areas in Ethiopia.

[img] PDF
Abstract:Food insecurity has long been a problem globally and even more so in low and middle income countries. However, existing indicators to assess and monitor food security are often operationally constrained as these do not provide insights on the underlying causes of food insecurity. One of the under-studied pillars of food security is food access or the ability of people to physically, economically, and socially access food. The food access studies that have incorporated a spatial context have mostly considered physical access but not economic access and have only done so in a limited spatial coverage. Moreover, several social inequalities have been studied but none so far have examined inequalities in food access. This study aimed to quantify food access and food access inequality across the urban-rural continuum of Ethiopia across time. To do this, this study developed a food access index (FAI) from household survey data using principal component analysis (PCA). Variables related to physical and economic access were found to be majority of the significant variables in the first principal component, but economic access variables had stronger contributions. The newly developed index was predicted across space and time for Ethiopia at high spatial detail using geospatial covariates by selecting important geospatial variables using random forest (RF). A generalized additive model (GAM) was trained to predict the FAI using 4 selected variables, namely, (1) road proximity, (2) travel time to markets, (3) cereal prices, and (4) precipitation at coldest quarter, and achieved an R2 of 0.62. The predicted FAI across Ethiopia’s urban-rural continuum showed that peri-urban areas, which make up a considerable proportion of the population, had low food access. Peri-urban areas of intermediate and small cities had the most temporally volatile food access within a year leaving such areas vulnerable to food price spikes. Finally, food access inequality was quantified by calculating the Gini index from the spatial predictions of the FAI and was found to be low for all of Ethiopia, driven by the large area percentage of small cities and catchment areas where food access inequality was low. However, food access inequality was high in large and intermediate cities and their catchment areas. From these findings, this study provided a proof-of-concept of the possibilities for spatiotemporal quantification of food access at high spatial detail. Moreover, a scenario development application1 was also built to simulate possible shocks like price spikes which comes at a time when global food access is threatened by rising prices caused by the war in Ukraine, among others. This study recommends further testing and improving this approach by (1) testing other modelling techniques and (2) implementing at multi-country or global scale to develop more insights into food access across the urban-rural continuum.
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Subject:54 computer science, 74 (human) geography, cartography, town and country planning, demography
Programme:Geoinformation Science and Earth Observation MSc (75014)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page