Climate change, agricultural productivity and child health outcomes in Ghana: A geospatial analysis.

Abstract

Deaths of children under the age of five are increasingly concentrated in Africa just as the continent is experiencing significant increase in population growth accompanied by a dramatic shift in climate conditions. Agricultural droughts are predicted with high certainty, which can increase the risks of dying for children through bio-behavioural, livelihood and socioeconomic factors at the individual and household levels; thus, placing burden for families as well as for governments to adequately plan and allocate resources. The role of climatic change and related stress on agriculture and food production, and how this affects human health is not systematically understood in Africa, especially in Ghana where climate change and environmental degradation adversely impact agricultural productivity and sustainable livelihoods.

This thesis investigates the long-term changes in climate patterns and crop productivity, and how they relate to child health outcomes, specifically stunting of children under five, across different ecological zones in Ghana. In doing so, a new enviro-social conceptual framework of child mortality is developed by linking existing frameworks to describe the outcomes of morbidity and mortality through the interactions of climate, socioeconomic and demographic factors that control access to resources, agriculture, poverty and diseases.

The initial exploratory analyses describing the climate and validating agricultural variables are performed using Mann Kendall analysis and simple predictive regression modelling. Bayesian geoadditive and logistic spatial regression modelling techniques are employed within multilevel frameworks to explore relationships between the environmental and demographic variables.

Four interlinked analyses are undertaken in this study. First, a descriptive analysis of the evolution of the spatiotemporal distribution of rainfall and surface temperature is created. This analysis establishes the spatial unit of analysis at the district level, at which governmental and non-governmental stakeholders apply policies, strategies and interventions. Climate variables were obtained from the University of Delaware’s Gridded Precipitation and Temperature Monthly Climatology version 4.01. Second, an analysis is attempted to relate the National Aeronautics and Space Administration MODIS/Terra Surface Reflectance 8-Day L3 Global 500 m SIN Grid V005 MOD09A1 Enhanced Vegetation Index (EVI) 2001-2014 with crop yield data from the Ministry of Food and Agriculture Ghana (MOFA) Agricultural Estimates 2009-2012. A phenologically-based classification is developed to predict crop yields. Third, an examination of the drivers of agricultural production is undertaken in order to assess the extent of the influence of climatic and sociodemographic conditions on agricultural production in Ghana. This assessment is critical for locating where food insecurity and poor nutritional outcomes may occur. The fourth analysis attempts to explore the spatial and temporal relationships between climate stress, agricultural production and stunting of children under five in order to understand the drivers of poor child health outcomes in Ghana.

Demographic and land cover variables were obtained from the European Space Agency Climate Change Initiative 300m annual global land cover time series 2001-2014; the Food and Agricultural Organisation of the United Nations Harmonised World Soil Database v1.2; and The International Food Policy Research Institute’s HarvestChoice Project Agro-ecological Zones of Africa; the Ghana Population and Housing Census 2000, 2010; Ghana Demographic and Health Survey 2003, 2008, 2014; WorldPop and Center for International Earth Science Information Network (CIESIN) Ghana 100 m Age structures 2003, 2008, 2014; the Centre for Remote Sensing and Geographic Information Services (CERSGIS) medical facility network dataset; the Centre for Remote Sensing and Geographic Information Services (CERSGIS) and the Ghana Highway Authority road network dataset. This thesis acknowledges the complexity of harmonising the necessary data to perform such analyses, especially in countries, such as Ghana, where data are limited, spatial boundaries are not consistent over time and patterns and processes complex. The data sources used are open source and freely available, with the exception of the Ghana Population and Housing Censuses of 2000 and 2010, with the aim that other researchers are able to reproduce and improve on this thesis.

The results show that the north and southwest of the country are most vulnerable to the risks of stunting and climate stress. In northern and south-western Ghana, the risks of stunting are driven by the climate through poor crop yields. The results suggest that this relationship is influencing a much slower decline in stunting in the north. There is, however, a nationwide pattern of decline in stunting over time.

More information

Identifiers

ORCID for Sabu Padmadas: orcid.org/0000-0002-6538-9374

Catalogue record

Export record