Estimating the Net Ecosystem Exchange at Global FLUXNET Sites Using a Random Forest Model

Huang, Ni; Wang, Li; Zhang, Yuelin; Gao, Shuai; Niu, Zheng

doi:10.1109/jstars.2021.3114190

Cited by 19 publications

(12 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Huang et al. (2021) used the variable importance measurement method of RF to construct the most suitable set of candidate predictors for predicting net ecosystem exchange, including 13 variables such as temperature, precipitation, EVI, and leaf area index, and realized the upscaling of net ecosystem exchange at global FLUXNET sites. Yao et al.…”

Section: Methodsmentioning

confidence: 99%

“…Moreover, RF can provide a measure of the importance of variables and identify significant explanatory factors among numerous predictors. Huang et al (2021) used the variable importance measurement method of RF to construct the most suitable set of candidate predictors for predicting net ecosystem exchange, including 13 variables such as temperature, precipitation, EVI, and leaf area index, and realized the upscaling of net ecosystem exchange at global FLUXNET sites. Yao et al (2021) constructed a two-layer gap-filling framework based on RF for site-level carbon fluxes and comprehensively selected 12 carbon flux driving factors such as solar radiation, temperature, relative humidity, and NDVI as input variables for the ML algorithm through importance evaluation.…”

Section: Random Forest Modelmentioning

confidence: 99%

“…Although ML techniques are primarily data‐driven, that is, black box models, highly accurate simulation results can inform mechanistic studies. ML technology has been applied in upscaling eddy flux data (Huang et al., 2021; Xiao et al., 2014; Zhu et al., 2020). However, the mechanistic study of carbon flux footprints is still in the early stage, especially in semi‐arid regions.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Quantifying the Spatial Representativeness of Carbon Flux Footprints of a Grassland Ecosystem in the Semi‐Arid Region

Gong

Wang

et al. 2023

JGR Atmospheres

View full text Add to dashboard Cite

Anthropogenic carbon dioxide (CO 2 ) emissions have been increasing rapidly since the industrial revolution, causing global warming and environmental change. CO 2 emissions in arid regions are only 30% of those in humid regions, but the warming rate is 20%-40% higher than that of humid regions (Yao et al., 2020). Such a significant spatial heterogeneity in warming rate and CO 2 emissions causes arid regions to face greater climate hazards than humid regions (Huang et al., 2012(Huang et al., , 2019Yao et al., 2020), and increasingly frequent and severe droughts threaten the ability of dryland ecosystems to sequester carbon and maintain biodiversity (Yao et al., 2020). Therefore, it is

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Random Forest Modelmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Quantifying the Spatial Representativeness of Carbon Flux Footprints of a Grassland Ecosystem in the Semi‐Arid Region

Gong

Wang

et al. 2023

JGR Atmospheres

View full text Add to dashboard Cite

show abstract

“…• Adaptive boosting-ADA [59,60] • Decision tree-DT [61,62] • K-nearest neighbor-KNN [63,64] • Multi-layer perceptron-MLP (artificial neural network) [65][66][67] • Random forest-RF [30,[68][69][70] • Support-vector regressor-SVR [71][72][73] • Extreme gradient boosting-XGB [74][75][76] These ML algorithms apply distinctive methodologies, making it useful to compare their results when applied to complex datasets. They can be categorized as regression-based (SVR), single-tree (DT), ensemble-tree (ADA, RF, XGB), data-matching (KNN) and neural-network (MLP) algorithms.…”

Section: Machine Learning Methods Appliedmentioning

confidence: 99%

Machine Learning and Regression Analysis Reveal Different Patterns of Influence on Net Ecosystem Exchange at Two Conifer Woodland Sites

Wood

2022

Res. in Ecol.

View full text Add to dashboard Cite

Variations in net ecosystem exchange (NEE) of carbon dioxide, and the variables influencing it, at woodland sites over multiple years determine the long term performance of those sites as carbon sinks. In this study, weekly-averaged data from two AmeriFlux sites in North America of evergreen woodland, in different climatic zones and with distinct tree and understory species, are evaluated using four multi-linear regression (MLR) and seven machine learning (ML) models. The site data extend over multiple years and conform to the FLUXNET2015 pre-processing pipeline. Twenty influencing variables are considered for site CA-LP1 and sixteen for site US-Mpj. Rigorous k-fold cross validation analysis verifies that all eleven models assessed generate reproducible NEE predictions to varying degrees of accuracy. At both sites, the best performing ML models (support vector regression (SVR), extreme gradient boosting (XGB) and multi-layer perceptron (MLP)) substantially outperform the MLR models in terms of their NEE prediction performance. The ML models also generate predicted versus measured NEE distributions that approximate cross-plot trends passing through the origin, confirming that they more realistically capture the actual NEE trend. MLR and ML models assign some level of importance to all influential variables measured but their degree of influence varies between the two sites. For the best performing SVR models, at site CA-LP1, variables air temperature, shortwave radiation outgoing, net radiation, longwave radiation outgoing, shortwave radiation incoming and vapor pressure deficit have the most influence on NEE predictions. At site US-Mpj, variables vapor pressure deficit, shortwave radiation incoming, longwave radiation incoming, air temperature, photosynthetic photon flux density incoming, shortwave radiation outgoing and precipitation exert the most influence on the model solutions. Sensible heat exerts very low influence at both sites. The methodology applied successfully determines the relative importance of influential variables in determining weekly NEE trends at both conifer woodland sites studied.

show abstract

“…Considering a large amount of 15‐min interval flux data during the long experimental period will cause overlapping in the regression plot, the original data in each group were generated into 10 average data points of λE according to the ascending order of measured λE when doing the regression plot between measured and simulated λE . The data points were not generated by weekly or monthly average as related flux studies did (e.g., Huang et al, 2021; Tsuruta et al, 2016) because our classification depends on wetness, and the original data were not equally distributed in each week or month. The regression based on original 15‐min interval data is provided in Appendix .…”

Section: Methodsmentioning

confidence: 99%

Does interception evaporation occur from both sides of leaves at the wet Japanese cypress canopy during and after rainfall?

et al. 2022

View full text Add to dashboard Cite

The interception by the lower (abaxial) side of needle leaves not only contributes to the forest evapotranspiration but will also cause a diffusion barrier of CO 2 over the stomata and depresses the plant gas exchange. This study combined the eddy covariance (EC) technique and leaf wetness measurement with a soil-vegetationatmosphere transfer (SVAT) multilayer model to examine the occurrence of interception evaporation from both sides of leaves at a Japanese cypress forest canopy relating to rainfall intensity and different wetting periods. We compared the measured latent heat flux (λE) with the simulated wet canopy λE with two models that interception evaporation only happens from the upper (adaxial) side of leaves and both sides of leaves. Both models showed a low λE during the rainfall as the EC data did. The simulated λE at the wet period after rainfall indicates that the interception evaporation from both sides of leaves is more likely to happen after heavy rainfall (>15 mm/12 h). However, for the most frequent small rainfall events (0-5 mm/12 h) at this site, interception evaporation is more likely to occur only from the adaxial side than both sides, which helps the wet leaves to maintain stomata opening and process CO 2 uptake after rainfall.

show abstract

Estimating the Net Ecosystem Exchange at Global FLUXNET Sites Using a Random Forest Model

Cited by 19 publications

References 46 publications

Quantifying the Spatial Representativeness of Carbon Flux Footprints of a Grassland Ecosystem in the Semi‐Arid Region

Quantifying the Spatial Representativeness of Carbon Flux Footprints of a Grassland Ecosystem in the Semi‐Arid Region

Machine Learning and Regression Analysis Reveal Different Patterns of Influence on Net Ecosystem Exchange at Two Conifer Woodland Sites

Does interception evaporation occur from both sides of leaves at the wet Japanese cypress canopy during and after rainfall?

Contact Info

Product

Resources

About