Data-driven Prediction of Crop Yield Over Germany

Abstract

This research aims to create a system that predicts agricultural yields using an all-encompassing system that integrates Numerical Weather Prediction (NWP) with machine learning. We want to know if combining NWP and ML models improves crop production forecasts over either NWP or ML models alone. We collect historical crop yield data, weather, and soil parameters data. Using the meteorological parameters and yield data, we will train an ML model to predict crop production and find out which model gives the best prediction. Finally, we will combine the ML and NWP models to improve forecast accuracy and reliability. The system is assessed using multiple metrics. The results will show if the technology can predict crop yields for different crops and regions. For agricultural decision-making, the method may identify the most important meteorological and soil parameters that affect crop yields. Farmers, policymakers, and agricultural stakeholders may benefit from accurate crop output forecasts using the proposed method. Ultimately, the goal is to contribute to NWP and ML model agricultural production prediction research. Combining these two methods can improve crop production estimates, helping farmers make better decisions and improve food security.