Analysis of Urban Maize Yield Response to Urban Sprawl in a Changing Climate: Case of Harare Metropolitan City

Abstract

Agriculture in free spaces within urban areas sustains residents as they grow crops to ensure food availability while at the same time selling other produce to fund other needs. Due to the value of urban agricultural production, the main focus of this study was to investigate the implications of urban growth on maize yield in Harare metropolitan city. In order to achieve this, Landsat multispectral and multi-temporal data were used to establish the responses of the estimated maize yield to city growth from 1984 to 2018. Initially, Land Use and Land Cover map for each period were produced using multispectral images and field observation in Support Vector Machine (SVM)-based supervised image classification. The maps were reclassified using a binary scheme of croplands and non-croplands, which was used to quantify cropland area in each period and hence change over time. The maize yield for each period was estimated from data obtained from a study of maize yields obtained at the University of Zimbabwe Farm (UZ) farm for maize grown under variable agricultural practices. Results showed that cropland area was reduced from 11,120 ha in 1984 down to 2631 ha in 2018. The estimated average maize yield decreased from 52,264 tons in 1984 to 12,366 tons in 2018. In addition to showing the value of urban agriculture, the findings of this study are important in informing the government, municipalities, and other stakeholders about how urban growth has the potential to compromise food security and livelihoods, especially for the urban poor.

Keywords

• classification
• land use and land cover
• remote sensing
• support vector machine
• urban agriculture
• yield estimation