Abstract

This study proposes a novel application of regression-based control charts (RCC) to monitor agricultural yield performance in Pakistan. While control charts are widely utilized in manufacturing and healthcare, their use in agriculture remains limited. By integrating statistical process control techniques with agronomic data, we aim to identify abnormal variations in crop yields that deviate from expected behavior after accounting for key influencing factors. Using 25 years (1999–2023) of data on wheat and rice yields, rainfall, and fertilizer consumption—along with optional variables such as temperature and irrigation—we develop multiple linear regression models to estimate yield based on climatic and input variables. Residuals from these models are then plotted on regression control charts to flag years where actual yield significantly diverged from predicted levels. These "out-of-control" signals are further investigated through qualitative analysis of historical events, including pest outbreaks, policy changes, and natural disasters. This mixed-method approach offers a robust framework for agricultural monitoring and can support timely decision-making for food security and resource management in South Asia.