Evaluation of water flux predictive models developed using eddy-covariance observations and machine learning: a meta-analysis

[摘要] With the rapid accumulation of water flux observations from globaleddy-covariance flux sites, many studies have used data-driven approaches tomodel water fluxes, with various predictors and machine learning algorithmsused. However, it is unclear how various model features affect predictionaccuracy. To fill this gap, we evaluated this issue based on records of 139developed models collected from 32 such studies. Support vector machines (SVMs; average R -squared  =  0.82) and RF (random forest; average R -squared  =  0.81) outperformed other evaluatedalgorithms with sufficient sample size in both cross-study and intra-study(with the same data) comparisons. The average accuracy of the model appliedto arid regions is higher than in other climate types. The average accuracyof the model was slightly lower for forest sites (average R -squared  =  0.76) than for croplands and grasslands (average R -squared  =  0.8 and0.79) but higher than for shrubland sites (average R -squared  =  0.67).Using R n / R s , precipitation, T a , and the fractionof absorbed photosynthetically active radiation (FAPAR) improved the model accuracy. Thecombined use of T a and R n / R s is very effective, especially in forests, whilein grasslands the combination of W s and R n / R s is also effective. Randomcross-validation showed higher model accuracy than spatial cross-validationand temporal cross-validation, but spatial cross-validation is moreimportant in spatial extrapolation. The findings of this study are promisingto guide future research on such machine-learning-based modeling.

[学科分类] 妇产科学