Predicting carbon and water vapor fluxes using machine learning and novel feature ranking algorithms

Cui, Xia; Goff, Thomas; Cui, Song; Menefee, Dorothy; Wu, Qiang; Rajan, Nithya; Nair, Shyam; Phillips, Nate; Walker, Forbes

doi:10.1016/j.scitotenv.2021.145130

Cited by 15 publications

(18 citation statements)

References 72 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Numerous studies conducted in recent years have capitalized on machine learning‐based algorithms to simulate water fluxes, mostly using RF, SVM and ANN methods (Cui et al, 2021). Across simulation studies, the optimal machine learning models are diverse, likely due to differences in modelling data and test methods.…”

Section: Discussionmentioning

confidence: 99%

“…However, MLR may be useful for variable selection, and model accuracy when based on only important variables was basically the same as when using all variables. Previous studies have only utilized data from single sites to build models simulating water fluxes, which requires a large amount of data for each site (Cui et al, 2021). This study built a more general model (across sites) using machine learning to improve the simulation accuracy of some sites, such as those containing relatively few samples or irregular data.…”

Section: Discussionmentioning

confidence: 99%

“…Machine learning is the core of data science. With the recent increase in available flux data, machine learning models have been gradually applied to simulate water fluxes (Cui et al, 2021). The most widely used machine learning models are supervised learning methods (Jordan & Mitchell, 2015), such as the random forest (RF), support vector machine (SVM) and back propagation neural network (BPNN) algorithms.…”

Section: Introductionmentioning

confidence: 99%

“…These types of models have more extensive applicability for analysing geospatial data. SVM and RF models predicted water fluxes with high accuracy in the agroecosystem (Cui et al, 2021). Though SVM and RF models produced similar results for simulating water fluxes in grassland, rice paddy field and forest ecosystems, the SVM model is the most recommended model (Hang & Hsieh, 2020).…”

Section: Introductionmentioning

confidence: 99%

“…To create site‐scale simulation models of water fluxes, most studies have built machine learning models based on data from a single site (Cui et al, 2021; Dou & Yang, 2018; Hang & Hsieh, 2020); this typically requires a large amount of data from one EC tower, and makes the resultant model difficult to apply to other sites. It is an effective way to solve this issue by creating a universal simulation model using the data of multiple flux sites.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Simulation of site‐scale water fluxes in desert and natural oasis ecosystems of the arid region in Northwest China

Xie

Luo

Hellwich

et al. 2021

Hydrological Processes

View full text Add to dashboard Cite

The study of water fluxes is important to better understand hydrological cycles in arid regions. Data‐driven machine learning models have been recently applied to water flux simulation. Previous studies have built site‐scale simulation models of water fluxes for individual sites separately, requiring a large amount of data from each site and significant computation time. For arid areas, there is no consensus as to the optimal model and variable selection method to simulate water fluxes. Using data from seven flux observation sites in the arid region of Northwest China, this study compared the performance of random forest (RF), support vector machine (SVM), back propagation neural network (BPNN), and multiple linear regression (MLR) models in simulating water fluxes. Additionally, the study investigated inter‐annual and seasonal variation in water fluxes and the dominant drivers of this variation at different sites. A universal simulation model for water flux was built using the RF approach and key variables as determined by MLR, incorporating data from all sites. Model performance of the SVM algorithm (R2 = 0.25–0.90) was slightly worse than that of the RF algorithm (R2 = 0.41–0.91); the BPNN algorithm performed poorly in most cases (R2 = 0.15–0.88). Similarly, the MLR results were limited and unreliable (R2 = 0.00–0.66). Using the universal RF model, annual water fluxes were found to be much higher than the precipitation received at each site, and natural oases showed higher fluxes than desert ecosystems. Water fluxes were highest during the growing season (May–September) and lowest during the non‐growing season (October–April). Furthermore, the dominant drivers of water flux variation were various among different sites, but the normalized difference vegetation index (NDVI), soil moisture and soil temperature were important at most sites. This study provides useful insights for simulating water fluxes in desert and oasis ecosystems, understanding patterns of variation and the underlying mechanisms. Besides, these results can make a contribution as the decision‐making basis to the water management in desert and oasis ecosystems.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%