Deep Learning Meets Process-Based Models: A Hybrid Approach to Agricultural Challenges

Process-based models (PBMs) and deep learning (DL) are two key approaches in agricultural modelling, each offering distinct advantages and limitations. PBMs provide mechanistic insights based on physical and biological principles, ensuring interpretability and scientific rigour. However, they often struggle with scalability, parameterisation, and adaptation to heterogeneous environments. In contrast, DL models excel at capturing complex, nonlinear patterns from large datasets but may suffer from limited interpretability, high computational demands, and overfitting in data-scarce scenarios. This study presents a systematic review of PBMs, DL models, and hybrid PBM-DL frameworks, highlighting their applications in agricultural and environmental modelling. We classify hybrid PBM-DL approaches into DL-informed PBMs, where neural networks refine process-based models, and PBM-informed DL, where physical constraints guide deep learning predictions. Additionally, we conduct a case study on crop dry biomass prediction, comparing hybrid models against standalone PBMs and DL models under varying data quality, sample sizes, and spatial conditions. The results demonstrate that hybrid models consistently outperform traditional PBMs and DL models, offering greater robustness to noisy data and improved generalisation across unseen locations. Finally, we discuss key challenges, including model interpretability, scalability, and data requirements, alongside actionable recommendations for advancing hybrid modelling in agriculture. By integrating domain knowledge with AI-driven approaches, this study contributes to the development of scalable, interpretable, and reproducible agricultural models that support data-driven decision-making for sustainable agriculture.

Comparison of Machine Learning Methods for Predicting Winter Wheat Yield in Germany

This study analyzed the performance of different machine learning methods for winter wheat yield prediction using extensive datasets of weather, soil, and crop phenology. To address the seasonality, weekly features were used that explicitly take soil moisture conditions and meteorological events into account. Our results indicated that nonlinear models such as deep neural networks (DNN) and XGboost are more effective in finding the functional relationship between the crop yield and input data compared to linear models. The results also revealed that the deep neural networks often had a higher prediction accuracy than XGboost. One of the main limitations of machine learning models is their black box property. As a result, we moved beyond prediction and performed feature selection, as it provides key results towards explaining yield prediction (variable importance by time). The feature selection method estimated the individual effect of weather components, soil conditions, and phenology variables as well as the time that these variables become important. As such, our study indicates which variables have the most significant effect on winter wheat yield.