Forecasting hourly $${\hbox {NO}_{2}}$$ concentrations by ensembling neural networks and mesoscale models

Valput, D.; Navares, Ricardo; Aznarte, José Luis

doi:10.1007/s00521-019-04442-z

Cited by 12 publications

(5 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…RF is an ensemble decision tree model, which performs a split according to a given splitting criterion, and the variable importance can be obtained by weighing the improvements in the splitting criterion in all the nodes where the variable appears as a splitter. Therefore, the RF model is more explanatory than other machine learning models and is a natural model for variable sensitivity analysis [ 52 , 53 ]. We built an RF model and removed variables in ascending order based on variable importance; sample-based CV10 was used to evaluate the model performance after removing one variable each time.…”

Section: Resultsmentioning

confidence: 99%

New Deep Learning Model to Estimate Ozone Concentrations Found Worrying Exposure Level over Eastern China

Wang

Jiang

et al. 2022

IJERPH

View full text Add to dashboard Cite

Ozone (O3), whose concentrations have been increasing in eastern China recently, plays a key role in human health, biodiversity, and climate change. Accurate information about the spatiotemporal distribution of O3 is crucial for human exposure studies. We developed a deep learning model based on a long short-term memory (LSTM) network to estimate the daily maximum 8 h average (MDA8) O3 across eastern China in 2020. The proposed model combines LSTM with an attentional mechanism and residual connection structure. The model employed total O3 column product from the Tropospheric Monitoring Instrument, meteorological data, and other covariates as inputs. Then, the estimates from our model were compared with real observations of the China air quality monitoring network. The results indicated that our model performed better than other traditional models, such as the random forest model and deep neural network. The sample-based cross-validation R2 and RMSE of our model were 0.94 and 10.64 μg m−3, respectively. Based on the O3 distribution over eastern China derived from the model, we found that people in this region suffered from excessive O3 exposure. Approximately 81% of the population in eastern China was exposed to MDA8 O3 > 100 μg m−3 for more than 150 days in 2020.

show abstract

Section: Resultsmentioning

confidence: 99%

New Deep Learning Model to Estimate Ozone Concentrations Found Worrying Exposure Level over Eastern China

Wang

Jiang

et al. 2022

IJERPH

View full text Add to dashboard Cite

show abstract

“…As such, the results of their base learners were geographically-weighted instead of the traditional approach of using constant weights for each base learner. Similarly, Valput et al (2019) used an ensemble approach to provide local predictions using regional numerical AP predictions.…”

Section: Ensemble Approachmentioning

confidence: 99%

Methods used for handling and quantifying model uncertainty of artificial neural network models for air pollution forecasting

Cabaneros

Hughes

2022

Environmental Modelling & Software

View full text Add to dashboard Cite

“…To estimate P M 2.5 , deep belief network and CNN-LSTM were applied with considering spatiotemporally correlation [33], [34]. Cabaneros et al [35] and Valput et al [36] achieved forecasting NO 2 concentration by neural networks.…”

Section: Related Workmentioning

confidence: 99%

Prediction of Vertical Profile of NO₂ Using Deep Multimodal Fusion Network Based on the Ground-Based 3-D Remote Sensing

Zhang

Liu

et al. 2022

IEEE Trans. Geosci. Remote Sensing

View full text Add to dashboard Cite

The vertical distribution profiles of NO 2 are essential for understanding the mechanisms, detecting near-surface emissions, and tracking pollutant transportation at high altitude. However, most of the published NO 2 studies are based on the surface 2-D measurements. The ground-based 3-D remote-sensing stations were recently built to measure vertical distribution profiles of NO 2. However, the stations were spatially sparse due to the high cost and could not make the measurements without sunlight. In this study, we first developed a multimodel fusion network (MF-net) based on the sparse vertical observations from the Jing-Jin-Ji region. We achieved the 3-D profile prediction of NO 2 in the range of 39.005-41.405N and 115.005-117.905E with 24-h coverage. The MF-net significantly surpassed the conventional WRF-CHEM model and provided a more accurate evaluation of the NO 2 transmission between Beijing and the neighboring cities. Besides, the MF-net covers the monitoring of NO 2 to the whole study area and extends the monitoring time to the entire day (24 h), making it serviceable for continuous spatial-temporal estimation of NO 2 and its transmission in pollution events. The MF-net provides more robust data support to formulate reasonable and effective pollution prevention and control measures. Index Terms-3-D prediction of NO 2 , deep learning neural network, multiaxis differential optical absorption spectroscopy (MAX-DOAS), multimodal information fusion, remote sensing. I. INTRODUCTION I N RECENT years, people have increasingly been concerned about air pollution with the development of urban Manuscript

show abstract

Forecasting hourly $${\hbox {NO}_{2}}$$ concentrations by ensembling neural networks and mesoscale models

Cited by 12 publications

References 27 publications

New Deep Learning Model to Estimate Ozone Concentrations Found Worrying Exposure Level over Eastern China

New Deep Learning Model to Estimate Ozone Concentrations Found Worrying Exposure Level over Eastern China

Methods used for handling and quantifying model uncertainty of artificial neural network models for air pollution forecasting

Prediction of Vertical Profile of NO₂ Using Deep Multimodal Fusion Network Based on the Ground-Based 3-D Remote Sensing

Contact Info

Product

Resources

About