Process-guidance improves predictive performance of neural networks for carbon turnover in ecosystems

Abstract

Despite deep-learning being state-of-the-art for data-driven model predictions, it has not yet found frequent application in ecology. Given the low sample size typical in many environmental research fields, the default choice for the modelling of ecosystems and its functions remain process-based models. The process understanding coded in these models complements the sparse data and neural networks can detect hidden dynamics even in noisy data. Embedding the process model in the neural network adds information to learn from, improving interpretability and predictive performance of the combined model towards the data-only neural networks and the mechanism-only process model. At the example of carbon fluxes in forest ecosystems, we compare different approaches of guiding a neural network towards process model theory. Evaluation of the results under four classical prediction scenarios supports decision-making on the appropriate choice of a process-guided neural network. O_TEXTBOXSignificance StatementDeep-learning is the state-of-the-art for data-driven model predictions. Given the low sample size typical in many environmental research fields, these approaches can rarely be applied. When data are complemented by process understanding, as coded in physical or empirical models, both predictions and generalisations can be substantially improved over both data-only neural networks and mechanism-only process models. Comparing different approaches of such process-guidance helps decide on how to best combine process models and neural networks. C_TEXTBOX