Too Much Rain, Too Little Rain: Impacts of Climate Variability on Child and Adult Health in Sub-Saharan Africa

Raya Muttarak, Wittgenstein Centre for Demography and Global Human Capital (IIASA, VID/OEAW, WU)
Anna Dimitrova, Wittgenstein Centre for Demography and Global Human Capital (IIASA, VID/OEAW, WU)

Chronic seasonal crop and livestock loss due to heat stress and rainfall shortages can pose a serious threat to health, especially in sub-Saharan Africa where subsistence and small-scale farming dominate. Apart from inconclusive evidence, little is known about the differential impacts of climatic shocks on health of population subgroups. This study aims to analyse the impacts of climate variability on health using Demographic and Health Surveys. Health is measured as: 1) stunting and wasting for children aged under 5; and 2) body mass index (BMI) for adults aged 15-59. Measures of climatic shocks include precipitation and temperature anomalies. The analysis was done for Ethiopia (2000, 2005, 2011 and 2016), and we are expanding to other countries in Sub-Saharan Africa. Our preliminary findings show a negative relationship between rainfall and stunting and wasting; and low BMI among adults. The climate impacts vary with population subgroups whereby girls and children whose mother has lower level of education and living in the rural area are more vulnerable to rainfall shocks.

Introduction

Climate change pose serious risksto populations in Sub-Saharan Africa, mainly through undermining food security (Funkand Brown 2009). Millions of people in the region live in rural areas anddepend on rain-fed agriculture for their subsistence. Rising temperatures and irregularrainfalls have increased the frequency of droughts in the region.  At the sametime, heavy precipitation is projected to rise and consequently increase therisk of floods and landslides. The consequences for the well-being ofpopulations exposed to such climatic shocks are manifold – from the loss oflives, water contamination and home damage due to floods, to reducedagricultural production and undermined food security caused by droughts.

The objective of this study is toexplore the impacts of climatic shocks on the nutrition status of children andadults in Sub-Saharan Africa. We will additionally investigate whetherdemographic and socioeconomic characteristics (i.e. wealth, education) andgeographic location determine the extent to which climatic shocks affect thehealth outcome of children and adults in the region.

Data and methods

For the analysis of child malnutrition, weuse DHS data for children aged under 5 years old. Stunting and wasting are usedas indicators of undernutrition. Stunting refers to children with a low heightfor age defined as below -2 standard deviations (SD) of the WHO Child GrowthStandards median. Wasting refers to children with a low weight for heightdefined as below -2 SD of the WHO Child Growth Standards median. Stuntingcaptures the cumulative effects of undernutrition while wasting indicates acuteweight loss. Both variables are measured as binary outcomes, indicating whetheror not a child was stunted/wasted at the time of the interview.

For the analysis of adult malnutrition, weuse data for males and females between the ages of 15 and 59. As a dependentvariable we use the body mass index (BMI), which is calculated as the weight ofthe individual divided by the square of their height.

As our climate measure we use monthlytemperature and precipitation data supplied by the Climatic Research Unit ofthe University of East Anglia for the period 1901-2015 (Universityof East Anglia Climatic Research Unit et al. 2014). We restrict the climatedata to the main agricultural season in each country and calculate seasonalaverages. Seasonal Z-scores are then constructed, which measure yearlydeviations from the long term-average (1950-2015). The gridded temperature andprecipitation data are matched with the geographical location of each clusterof households in the DHS data.

We use linear regression models to quantifythe impact of climate shocks on nutrition status, controlling for individualand household characteristics, time and location fixed effects. 

Results

Here we show preliminary results based on fourrounds of DHS data for Ethiopia (2000, 2005, 2011 and 2016). Table 1 presents aregression estimate of the probability of stunting and wasting for children ageunder 5. We find that higher rainfall during infancy (age 0 to 1) reduces theprobability of a child being stunted by 7.8 percentage point. The second column shows that an increase in precipitation during the latest summerseason leads to decline in the probability of a child suffering from wasting byas much as 15.4 percentage points at 1% significance level.

In Table 2, we include interaction termsbetween the climate measure and various household characteristics in order toidentify what types of households are more vulnerable to climate variabilities.Our findings suggest that children whose mothers’ have lower levels of educationare more likely to be malnourished during drought periods. We also find that girlsand children living in rural areas are more vulnerable to droughts.

With respect to adult health, measured bythe body mass index, we show that above average rainfall in the latest growingseason increases BMI scores for men and does not have a significant effect forwomen. However, when we consider the effect of rain shocks during the latest5-year period women seem to benefit more from higher rainfall than men. Thesepreliminary findings indicate that women are susceptible to long-term droughtsand men to short-term rainfall variations.

As anext step, we will expand our sample to include children and adults from otherSub-Saharan African countries covered by the DHS and will include measures oftemperature variations.

References

Funk,C. C., & Brown, M. E. (2009). Declining global per capita agriculturalproduction and warming oceans threaten food security. Food Security, 1(3),271–289. doi:10.1007/s12571-009-0026-y

University of East AngliaClimatic Research Unit, Harris, I., & Jones, P. D. (2014, September 24).Climatic Research Unit (CRU) Time-Series (TS) Version 3.22 of High ResolutionGridded Data of Month-by-month Variation in Climate (Jan. 1901- Dec. 2013).NCAS British Atmospheric Data Centre.http://dx.doi.org/10.5285/18BE23F8-D252-482D-8AF9-5D6A2D40990C. Accessed 23September 2015

 

Presented in Session 1179: Health, Wellbeing, and Morbidity