Assessing the Impact of Weather Variables on Crop Yield Variability

Thesis Abstract

Climate variability and changing weather patterns significantly influence agricultural productivity, yet the quantification of how specific weather variables affect crop yield variability remains insufficiently explored, particularly in regions heavily reliant on rain-fed agriculture. This study aims to assess the impact of key weather variables—temperature, rainfall, humidity, and solar radiation—on the variability of maize crop yields in the Central Plains region over a ten-year period. The specific objectives include identifying correlations between weather parameters and crop yields, determining the relative contribution of each weather variable to yield variability, and developing predictive models to enhance yield forecasting accuracy under varying climate scenarios. Employing a quantitative research design, the study utilizes secondary data sources comprising monthly weather records from the National Meteorological Agency and agricultural yields obtained from the Regional Agricultural Development Authority. The population encompasses all maize farms within the region, and a stratified random sampling technique is applied to select 150 farms representing diverse agro-ecological zones, ensuring representativeness. Data collection involves compilation of weather data at farm-level coordinates disaggregated to monthly intervals, and corresponding maize yield records validated through field surveys and farm reports. Instrumentation includes structured data extraction templates and calibration of meteorological data ensuring standardization. Analytical methods employed include descriptive statistics to characterize weather trends and yield patterns, Pearson's correlation analysis to identify initial relationships, and multiple linear regression models to quantify the influence of each weather variable on crop yield variability. To address multicollinearity, variance inflation factors are analyzed, and stepwise regression enhances model parsimony. Additionally, time series decomposition using Seasonal-Trend decomposition procedures and ARIMA models will be applied to assess temporal dependencies. Model validation involves residual analysis, cross-validation with holdout datasets, and measures such as Root Mean Square Error (RMSE) to evaluate predictive performance. It is anticipated that the findings will demonstrate statistically significant associations between increased temperature, erratic rainfall patterns, and reduced maize yields, with humidity and solar radiation exerting moderate effects. The regression models are expected to identify temperature and rainfall as primary predictors, accounting for over 65% of the yield variability observed across the study period. Furthermore, the analysis will reveal seasonal and interannual fluctuations linked to climate anomalies such as El Niño and La Niña episodes. This research contributes novel empirical evidence to the existing body of knowledge by quantifying the relative impacts of specific weather variables on crop productivity within a tropical agrarian context, thus advancing theoretical understanding of climate-agriculture linkages grounded in the adaptive capacity framework. The study also provides a valuable predictive model that can inform regional crop forecasting and climate resilience planning. The main conclusion underscores the critical importance of integrating meteorological data into agricultural decision-making processes and highlights the need for climate-smart agricultural practices and early warning systems. Recommendations include policy emphasis on weather-based crop insurance schemes, targeted irrigation interventions, and the adoption of resilient crop varieties suited to predicted climate scenarios. Finally, the study advocates for further research incorporating socio-economic factors and climate change projections to enhance adaptive strategies for sustainable farming systems in climate-sensitive regions.