干旱区生态环境知识资源中心(Arid): When ancient numerical demons meet physics-informed machine learning: adjoint-based gradients for implicit differentiable modeling

Arid

DOI	10.5194/hess-28-3051-2024
	When ancient numerical demons meet physics-informed machine learning: adjoint-based gradients for implicit differentiable modeling
	Song, Yalan; Knoben, Wouter J. M.; Clark, Martyn P.; Feng, Dapeng; Lawson, Kathryn; Sawadekar, Kamlesh; Shen, Chaopeng
通讯作者	Song, YL ; Shen, CP
来源期刊	HYDROLOGY AND EARTH SYSTEM SCIENCES
ISSN	1027-5606
EISSN	1607-7938
出版年	2024
卷号	28 期号:13 页码:3051-3077
英文摘要	Recent advances in differentiable modeling, a genre of physics-informed machine learning that trains neural networks (NNs) together with process-based equations, have shown promise in enhancing hydrological models' accuracy, interpretability, and knowledge-discovery potential. Current differentiable models are efficient for NN-based parameter regionalization, but the simple explicit numerical schemes paired with sequential calculations (operator splitting) can incur numerical errors whose impacts on models' representation power and learned parameters are not clear. Implicit schemes, however, cannot rely on automatic differentiation to calculate gradients due to potential issues of gradient vanishing and memory demand. Here we propose a discretize-then-optimize adjoint method to enable differentiable implicit numerical schemes for the first time for large-scale hydrological modeling. The adjoint model demonstrates comprehensively improved performance, with Kling-Gupta efficiency coefficients, peak-flow and low-flow metrics, and evapotranspiration that moderately surpass the already-competitive explicit model. Therefore, the previous sequential-calculation approach had a detrimental impact on the model's ability to represent hydrological dynamics. Furthermore, with a structural update that describes capillary rise, the adjoint model can better describe baseflow in arid regions and also produce low flows that outperform even pure machine learning methods such as long short-term memory networks. The adjoint model rectified some parameter distortions but did not alter spatial parameter distributions, demonstrating the robustness of regionalized parameterization. Despite higher computational expenses and modest improvements, the adjoint model's success removes the barrier for complex implicit schemes to enrich differentiable modeling in hydrology.
类型	Article
语种	英语
收录类别	SCI-E
WOS记录号	WOS:001267236600001
WOS关键词	SENSITIVITY-ANALYSIS ; HYDROLOGIC MODEL ; CALIBRATION ; UNIVERSAL ; NETWORKS
WOS类目	Geosciences, Multidisciplinary ; Water Resources
WOS研究方向	Geology ; Water Resources
资源类型	期刊论文
条目标识符	http://119.78.100.177/qdio/handle/2XILL650/404129
推荐引用方式 GB/T 7714	Song, Yalan,Knoben, Wouter J. M.,Clark, Martyn P.,et al. When ancient numerical demons meet physics-informed machine learning: adjoint-based gradients for implicit differentiable modeling[J],2024,28(13):3051-3077.
APA	Song, Yalan.,Knoben, Wouter J. M..,Clark, Martyn P..,Feng, Dapeng.,Lawson, Kathryn.,...&Shen, Chaopeng.(2024).When ancient numerical demons meet physics-informed machine learning: adjoint-based gradients for implicit differentiable modeling.HYDROLOGY AND EARTH SYSTEM SCIENCES,28(13),3051-3077.
MLA	Song, Yalan,et al."When ancient numerical demons meet physics-informed machine learning: adjoint-based gradients for implicit differentiable modeling".HYDROLOGY AND EARTH SYSTEM SCIENCES 28.13(2024):3051-3077.