ASSESSING THE IMPACT OF EL NINO AND LA NINA EVENTS ON MAIZE YIELD IN THYOLO DISTRICT

Climate variability driven by the El Niño-Southern Oscillation (ENSO) represents one of the most consequential and recurring threats to agricultural productivity in sub-Saharan Africa. In Malawi, where over 90% of agricultural production is rain-fed and smallholder farmers constitute the backbone of the food system, ENSO-induced fluctuations in rainfall and temperature directly determine the success or failure of the maize harvest the nation's most important staple crop. Thyolo District, located in the southern highlands of Malawi, is particularly exposed to ENSO-driven climate extremes: the 2015-2016 and 2023-2024 El Niño events caused devastating crop losses in the district, the latter prompting a national disaster declaration affecting nearly two million farming households and reducing national maize output by approximately 17%. Despite this well-documented vulnerability, no empirically grounded, district-specific study has quantified the relationship between ENSO phases and maize yield outcomes in Thyolo, leaving agricultural planners and extension services without locally calibrated evidence to guide seasonal decision-making. This study assessed the impact of El Niño and La Niña events on maize yield in Thyolo District, Malawi, over 29 growing seasons (November-April, 1995-2024). The specific objectives were to: verify the quality and distributional properties of the maize yield dataset, characterize historical trends in ENSO phases, seasonal rainfall, and temperature using descriptive and exploratory analysis, examine bivariate linear relationships between the Oceanic Niño Index (ONI), seasonal rainfall, seasonal temperature, and detrended maize yield anomaly using Pearson correlation, conduct a composite analysis comparing mean yield, rainfall, and temperature across El Niño, La Niña, and Neutral ENSO phases, quantify the independent and joint effects of ONI, rainfall, and temperature on detrended maize yield using simple and multiple linear regression. The study utilized secondary data from three institutional sources: maize yield records from HarvestStat Africa, seasonal rainfall and temperature data obtained from the Malawi Department of Climate Change and Meteorological Services (DCCMS) through a formal institutional data access process, and ONI values from the National Oceanic and Atmospheric Administration (NOAA) Climate Prediction Centre. Maize yield data were detrended to isolate climatedriven yield anomalies from long-term technology trends.
ENSO phases were classified using the standard ONI threshold (El Niño: ONI ≥ +0.5; La Niña: ONI ≤ -0.5; Neutral: -0.5 < ONI < +0.5). Statistical analyses including Pearson correlation, composite phase analysis, simple linear regression, and multiple linear regression were performed in R statistical software. Data quality was assessed through boxplot outlier detection and Q-Q normality testing prior to all inferential analyses. Data quality assessment confirmed that the 29-season maize yield dataset was free of outliers and approximated a normal distribution, satisfying parametric analysis assumptions. Descriptive analysis revealed high interannual variability in maize yield (range: 0.27-3.06 mt/ha; mean: 1.82 mt/ha) and seasonal rainfall (691-1,990 mm), with no consistent upward yield trend over the study period indicative of climate-dominated productivity dynamics. The composite analysis demonstrated a coherent directional pattern: El Niño phases were associated with the lowest mean rainfall, highest temperatures, and negative yield anomalies, while La Niña phases corresponded to above-average rainfall, cooler temperatures, and positive yield anomalies. Pearson correlation confirmed a statistically significant moderate-to-strong negative relationship between ONI and detrended yield anomaly (r = -0.535, p = 0.003, 95% CI: -0.754 to -0.210), with ENSO variability accounting for 28.6% of interannual yield variance. Correlations between yield and total seasonal rainfall (r = 0.215, p = 0.262) and between yield and seasonal mean temperature (r = -0.197, p = 0.304) were not statistically significant, suggesting that ONI integrates the compound climate signal more effectively than individual climate variables. Multiple linear regression confirmed ONI as the sole statistically significant predictor of detrended maize yield (β = -0.255, p = 0.012), with the overall model explaining 29.8% of yield variance (Adjusted R2 = 0.214, F = 3.541, p = 0.029). The regression coefficient implies a yield reduction of approximately 0.25 mt/ha for each unit increase in ONI, providing a locally calibrated quantitative estimate of El Niño's agricultural impact in Thyolo District. This study provides the first empirically rigorous, district-level quantification of ENSO's impact on maize yield in Thyolo District, Malawi.
The findings confirm that El Niño events significantly reduce and La Niña events moderately enhance maize productivity in the district, with the ONI serving as a statistically reliable and parsimonious seasonal yield predictor.
These results validate the scientific basis for integrating ENSO-based seasonal climate forecasts into district agricultural planning, early warning systems, and food security contingency frameworks. The study recommends that Thyolo District agricultural authorities use ONI forecasts to pre-position drought-tolerant seed varieties, adjust fertilizer subsidy targeting, and activate food security response plans during declared El Niño seasons, while future research should incorporate intra-seasonal rainfall distribution metrics and heat stress degree-day calculations to further refine climate–yield modelling in the district.