Revealing Causal Controls of Storage-Streamflow Relationships With a Data-Centric Bayesian Framework Combining Machine Learning and Process-Based Modeling

被引:4
作者
Tsai, Wen-Ping [1 ]
Fang, Kuai [1 ,2 ]
Ji, Xinye [1 ,3 ]
Lawson, Kathryn [1 ]
Shen, Chaopeng [1 ]
机构
[1] Penn State Univ, Civil & Environm Engn, University Pk, PA 16802 USA
[2] Stanford Univ, Earth Syst Sci, Stanford, CA 94305 USA
[3] Shenzhen State High Tech Ind Innovat Ctr, Shenzhen, Peoples R China
来源
FRONTIERS IN WATER | 2020年 / 2卷
关键词
Machine Learning (ML); process-based model (PBM); streamflow-storage relationships; data-centric; Bayes law; classification tree; soil texture; LAND-SURFACE MODEL; SOIL-MOISTURE; GUIDED DATA; FLOW; WATER; CALIBRATION; SIMULATION; EVOLUTION; PATTERNS; IMPACTS;
D O I
10.3389/frwa.2020.583000
中图分类号
TV21 [水资源调查与水利规划];
学科分类号
081501 ;
摘要
Some machine learning (ML) methods such as classification trees are useful tools to generate hypotheses about how hydrologic systems function. However, data limitations dictate that ML alone often cannot differentiate between causal and associative relationships. For example, previous ML analysis suggested that soil thickness is the key physiographic factor determining the storage-streamflow correlations in the eastern US. This conclusion is not robust, especially if data are perturbed, and there were alternative, competing explanations including soil texture and terrain slope. However, typical causal analysis based on process-based models (PBMs) is inefficient and susceptible to human bias. Here we demonstrate a more efficient and objective analysis procedure where ML is first applied to generate data-consistent hypotheses, and then a PBM is invoked to verify these hypotheses. We employed a surface-subsurface processes model and conducted perturbation experiments to implement these competing hypotheses and assess the impacts of the changes. The experimental results strongly support the soil thickness hypothesis as opposed to the terrain slope and soil texture ones, which are co-varying and coincidental factors. Thicker soil permits larger saturation excess and longer system memory that carries wet season water storage to influence dry season baseflows. We further suggest this analysis could be formulated into a data-centric Bayesian framework. This study demonstrates that PBM present indispensable value for problems that ML cannot solve alone, and is meant to encourage more synergies between ML and PBM in the future.
引用
收藏
页数:16
相关论文
共 57 条
[1]  
[Anonymous], Artif. Intell. Mach. Learn.
[2]   On Pixel-Wise Explanations for Non-Linear Classifier Decisions by Layer-Wise Relevance Propagation [J].
Bach, Sebastian ;
Binder, Alexander ;
Montavon, Gregoire ;
Klauschen, Frederick ;
Mueller, Klaus-Robert ;
Samek, Wojciech .
PLOS ONE, 2015, 10 (07)
[3]   THE FUTURE OF DISTRIBUTED MODELS - MODEL CALIBRATION AND UNCERTAINTY PREDICTION [J].
BEVEN, K ;
BINLEY, A .
HYDROLOGICAL PROCESSES, 1992, 6 (03) :279-298
[4]  
Breiman L., 1984, WADSWORTH INT GROUP, DOI DOI 10.1785/0120150058
[5]   The Community Climate System Model version 3 (CCSM3) [J].
Collins, William D. ;
Bitz, Cecilia M. ;
Blackmon, Maurice L. ;
Bonan, Gordon B. ;
Bretherton, Christopher S. ;
Carton, James A. ;
Chang, Ping ;
Doney, Scott C. ;
Hack, James J. ;
Henderson, Thomas B. ;
Kiehl, Jeffrey T. ;
Large, William G. ;
McKenna, Daniel S. ;
Santer, Benjamin D. ;
Smith, Richard D. .
JOURNAL OF CLIMATE, 2006, 19 (11) :2122-2143
[6]   The Community Land Model and its climate statistics as a component of the Community Climate System Model [J].
Dickinson, Robert E. ;
Oleson, Keith W. ;
Bonan, Gordon ;
Hoffman, Forrest ;
Thornton, Peter ;
Vertenstein, Mariana ;
Yang, Zong-Liang ;
Zeng, Xubin .
JOURNAL OF CLIMATE, 2006, 19 (11) :2302-2324
[7]  
Dingman SL., 2015, Physical hydrology, V3
[8]   Near-Real-Time Forecast of Satellite-Based Soil Moisture Using Long Short-Term Memory with an Adaptive Data Integration Kernel [J].
Fang, Kuai ;
Shen, Chaopeng .
JOURNAL OF HYDROMETEOROLOGY, 2020, 21 (03) :399-413
[9]   Combining a land surface model with groundwater model calibration to assess the impacts of groundwater pumping in a mountainous desert basin [J].
Fang, Kuai ;
Ji, Xinye ;
Shen, Chaopeng ;
Ludwig, Noel ;
Godfrey, Peter ;
Mahjabin, Tasnuva ;
Doughty, Christine .
ADVANCES IN WATER RESOURCES, 2019, 130 :12-28
[10]   The Value of SMAP for Long-Term Soil Moisture Estimation With the Help of Deep Learning [J].
Fang, Kuai ;
Pan, Ming ;
Shen, Chaopeng .
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2019, 57 (04) :2221-2233