crop yield forecasting

Rice remains one of Nigeria’s most important staple crops, serving as both a major source of food and a significant contributor to national agricultural output. However, inconsistencies in production statistics and yield estimates have made it difficult to accurately assess the country’s progress toward self-sufficiency. This study, therefore, focuses on developing a rigorous mathematical framework for estimating and analyzing rice yield in Nigeria from 1990 to 2022. The research integrates statistical modeling and mathematical reasoning to provide a more objective and quantifiable understanding of yield dynamics, while addressing uncertainties associated with data collection, reporting errors, and environmental variability. The study utilizes secondary data from the Food and Agriculture Organization’s FAOSTAT database, which provides national figures on rice production and harvested area. The mathematical model adopts the classical yield equation � = �/� where � denotes yield (t/ha), � represents production (tonnes), and � is harvested area (hectares). To estimate the reliability of calculated yields, the propagation of uncertainty formula �� = � �� � 2 + 𝛿 � 2 was applied, allowing error terms in production and area to be combined mathematically. Statistical regression models (linear, exponential, and polynomial) were used to evaluate long-term yield trends and to test the hypothesis of yield improvement over time. In addition, stochastic simulation techniques and correlation analyses were introduced to capture the variability and interdependence between production and land-use parameters. Findings indicate that Nigeria’s rice yield followed a fluctuating but generally upward trend, rising from an average of 1.5 �/ℎ� in the early 1990s to about 2.8 �/ℎ� by 2022. The regression analysis revealed a statistically significant positive trend, confirming gradual improvements in productivity over the years. However, the propagated error analysis showed that yield uncertainties ranged between 5 –10% depending on data completeness and measurement precision. This highlights persistent limitations in the reliability of agricultural data collection systems. The study concludes that mathematical modeling provides a robust foundation for understanding agricultural yield trends and recommends the integration of error analysis and predictive modeling into national data reporting frameworks. By combining quantitative rigor with empirical agricultural data, the research establishes a replicable approach for improving the precision of yield estimation in Nigeria and other developing economies.