Modeling Regional Variation in Wheat Production Dynamics Using Trend-Based Forecasting Techniques

The current study is being carried out to take a close look examining the patterns in yield, production, and area of important crops like wheat from all provinces of Pakistan in detail. The proposed research will involve the determination of the best forecasting models to be used in fundamental crop growing. The time series data were analyzed using six different forecasting models applied to the data between the years 1981 and 2023 and they included: single exponential smoothing model, double exponential smoothing model, exponential growth model, S-curve trend model, linear trend model and quadratic trend model. A thorough assessment by the method of error minimization was done to determine the best model to predict both area and production and this includes the Mean Squared Deviation (MSD), Mean Absolute Deviation (MAD) and Mean Absolute Percentage Error (MAPE). Due to the relatively low error estimates, the results of the calculations showed that the Double Exponential Smoothing model was significantly more useful in matching the calibration of area and production of wheat across several provinces. However, the S-curve and quadratic trend models were found to be much better performing in some geographic areas compared to the other models; they were able to describe the local growth trends of the crops. It is projected that the results of this research would significantly contribute to the policymakers in coming up with informed policies to manage agricultural areas, optimizing production, pricing policies, and analysis of consumption trends which are all vital towards the food security of Pakistan.