NeuralCrop: Combining physics and machine learning for improved crop yield predictions

2025-12-23

AI Summary

Global gridded crop models (GGCMs) exhibit explicit process-based mechanisms but suffer from high uncertainty, whereas purely data-driven models lack generalizability—especially under climate change—limiting their reliability for yield forecasting. To address this, we propose a hybrid crop model that synergistically integrates physical mechanisms with data-driven learning, introducing the novel two-stage paradigm of “GGCM pre-training followed by observational data fine-tuning.” Leveraging knowledge distillation and transfer learning, our approach jointly enhances mechanistic interpretability and deep-learning predictive capacity. Evaluated across European wheat-growing regions and the U.S. Corn Belt (2000–2019), the model significantly outperforms state-of-the-art GGCMs: it reduces prediction error by over 30% in drought-extreme years and demonstrates markedly superior extrapolative generalization stability compared to purely machine-learning-based methods. This work establishes a new paradigm for climate-resilient food security assessment—one that concurrently delivers accuracy, robustness, and scientific interpretability.

Technology Category

• Machine Learning: Deep Neural Architectures and Foundation Models
• Computer Vision: Large Vision Models
• Game Theory and Economic Paradigms: Adversarial Learning

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Web Mining and Content Analysis: Large pretrained models with web data
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

Global gridded crop models (GGCMs) simulate daily crop growth by explicitly representing key biophysical processes and project end-of-season yield time series. They are a primary tool to quantify the impacts of climate change on agricultural productivity and assess associated risks for food security. Despite decades of development, state-of-the-art GGCMs still have substantial uncertainties in simulating complex biophysical processes due to limited process understanding. Recently, machine learning approaches trained on observational data have shown great potential in crop yield predictions. However, these models have not demonstrated improved performance over classical GGCMs and are not suitable for simulating crop yields under changing climate conditions due to problems in generalizing outside their training distributions. Here we introduce NeuralCrop, a hybrid GGCM that combines the strengths of an advanced process-based GGCM, resolving important processes explicitly, with data-driven machine learning components. The model is first trained to emulate a competitive GGCM before it is fine-tuned on observational data. We show that NeuralCrop outperforms state-of-the-art GGCMs across site-level and large-scale cropping regions. Across moisture conditions, NeuralCrop reproduces the interannual yield anomalies in European wheat regions and the US Corn Belt more accurately during the period from 2000 to 2019 with particularly strong improvements under drought extremes. When generalizing to conditions unseen during training, NeuralCrop continues to make robust projections, while pure machine learning models exhibit substantial performance degradation. Our results show that our hybrid crop modelling approach offers overall improved crop modeling and more reliable yield projections under climate change and intensifying extreme weather conditions.

Problem

• Improves crop yield predictions by combining physics-based models with machine learning
• Addresses uncertainties in simulating biophysical processes under changing climate conditions
• Enhances generalization for robust projections beyond training data distributions

Innovation

• Hybrid model combines process-based and machine learning components
• First trains to emulate GGCM then fine-tunes on observational data
• Outperforms existing models and generalizes better to unseen conditions