ENSO impacts child undernutrition in the global tropics

Climate variability significantly impacts health outcomes and is a major concern for global climate policy and international public health. The El Niño Southern Oscillation (ENSO), a principal component of global climate variability, affects social, economic, and health outcomes. However, its systematic effects on global health remain understudied. ENSO destabilizes agriculture, economic production, and social stability in the global tropics. It's been linked to human health through its effects on infectious diseases and indirectly by decreasing agricultural yields, increasing food insecurity, and conflict likelihood. ENSO's adverse effects on yields are particularly acute in the tropics, where vulnerable, food-insecure children are most numerous. This study estimates ENSO's impacts on human nutrition throughout the global tropics by leveraging over one million child anthropometric records spanning four decades and all developing country regions. The research design estimates the change in nutritional status associated with positive or negative ENSO states compared to neutral conditions. Children's anthropometric measures, particularly weight-for-age z-scores (WAZ), serve as a summary measure of contemporary household food security. The goal is to assess the total influence of ENSO variability through all plausible mechanisms known to affect child nutrition, and to analyze systematic differences in ENSO response across regions with varying precipitation responses to ENSO, across continents, and across decades.

The existing literature highlights the detrimental effects of ENSO on various aspects of human well-being. Studies have documented ENSO's large-scale adverse effects for centuries, and its impact on agricultural productivity, economic production, and social stability within the tropics have been extensively examined. While the links between ENSO and infectious diseases are known, the systematic global impact on human nutrition remains largely unexplored. Previous analyses typically focused on single countries or specific El Niño episodes, lacking the global or regional scope necessary to inform effective national and international public investment strategies. This study builds upon a growing literature using Demographic and Health Surveys (DHS) to document the effects of weather variation on child nutrition, expanding upon previous studies that focused on specific regions or time periods to provide a comprehensive global perspective.

The study uses the NINO3.4 index of equatorial Pacific sea surface temperature to capture ENSO variation. Children's WAZ at the time of survey, calculated using the NCHS/CDC/WHO International Reference Standard, are the primary outcome measure. The analysis identifies all countries with local climates teleconnected to ENSO for which DHS anthropometric data exist, encompassing 1.3 million children aged 0–4 years from 186 household surveys between 1986 and 2018 across 51 countries. Treatment (ENSO state during the survey) is assigned annually by tropical year, using the mean NINO3.4 SST value from May to December. The study accounts for differences in ENSO effects by estimating separate responses in subnational regions with positive versus negative precipitation correlations with warmer ENSO. The empirical distribution of WAZ is compared between El Niño and La Niña years. To isolate the effect of ENSO on child malnutrition, the analysis utilizes variation in ENSO anomalies—deviations from long-run average conditions. The researchers control for potential confounders using fixed effects for spatial location, detrending by major world regions, removing monthly seasonality, and including country-specific controls for mother's age and education. The empirical analysis involves OLS regressions with two-way clustered standard errors at the level of tropical year and subnational administrative unit. The study also explores alternative specifications, observation weighting, ENSO variable definitions, and age categories. Robustness checks include testing different detrending methods, ENSO indicators, and teleconnection definitions. To assess the population-level impact of the 2015 El Niño, the researchers calculated the effect size summed over all children and compared it to the effect sizes of various nutrition interventions.

The study finds a significant negative association between warmer, El Niño-like ENSO conditions and child WAZ across the tropics, except in areas where precipitation increases during El Niño. A 1°C increase in the ENSO index is associated with a 0.03σ decrease in WAZ. This negative association persists across the distribution of ENSO values and is robust across various model specifications and other anthropometric outcomes (weight-for-height and BMI). Warmer ENSO increases the prevalence of underweight by 0.6 percentage points per 1°C. The risk of wasting shows a similar but insignificant effect. These patterns reverse in areas where warmer ENSO leads to wet anomalies. The 2015 El Niño significantly worsened child undernutrition. The magnitude of the effect is substantial, equivalent to the WAZ reduction from moving 46 million children from urban to rural areas, and resulted in an additional 5.9 million underweight children. The negative relationship between child nutrition and warm ENSO state shows little variation across major world regions and decades. The effect is consistent across age groups and robust to alternative model specifications, weighting methods, and ENSO variable definitions. Lagged effects show persistent impacts on child stunting but no persistent effect on weight, height, or BMI, except in the positively correlated rainfall subsample. These findings are robust across a range of alternative specifications and tests.

The study's findings highlight the substantial impact of ENSO on child undernutrition, particularly in the tropics. The consistent negative relationship between warmer ENSO conditions and child nutrition across various regions and decades underscores the vulnerability of these populations. The variation in effect based on precipitation patterns suggests that agricultural production plays a crucial mediating role, though other factors (e.g., conflict) can't be ruled out. The lack of substantial change in the relationship over the decades despite economic growth indicates stringent limits to adaptation. The substantial impact of the 2015 El Niño on child undernutrition emphasizes the need for proactive measures. The predictable nature of ENSO offers the opportunity to develop effective early warning systems and deploy nutrition and humanitarian support proactively. This research emphasizes the necessity of integrating ENSO forecasts into humanitarian planning and budgetary frameworks to anticipate fluctuations in resource availability.

This study demonstrates a robust and consistent negative association between El Niño events and child nutrition in the global tropics, emphasizing the importance of considering predictable climate variability in global development and humanitarian efforts. The findings highlight the need for proactive strategies to mitigate the impacts of ENSO on child nutrition, including the development of early warning systems and incorporation of ENSO forecasts into multi-year humanitarian planning and budgetary frameworks. Future research could explore the specific mechanisms linking ENSO to child nutrition in more detail and assess the effectiveness of different interventions to mitigate the impact of these predictable climatic events on vulnerable populations.

While the study's results are robust, there are limitations. DHS data selectively report migration, making it difficult to account for migration patterns in response to ENSO. The severity of ENSO events might differentially influence the likelihood of children being included in the sample, potentially leading to biases. However, the consistency of results across various specifications suggests these biases are likely small. Future studies should address these limitations to increase the precision and generalizability of findings.