Process-guidance improves predictive performance of neural networks for carbon turnover in ecosystems
M. Wesselkamp, N. N. Moser, M. Kalweit, J. Boedecker, C. F. Dormann
The paper improves neural network predictions for carbon turnover in ecosystems by integrating process-based models, showing better performance than data-only or mechanism-only models.
bioRxiv · Sep 28, 2022 · PreprintEcology