1. Key Laboratory of Environment Optics and Technology, Anhui Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, Hefei 230031, China;
2. University of Science and Technology of China, Hefei 230026, China;
3. Center for Excellence in Regional Atmospheric Environment, Chinese Academy of Sciences, Xiamen 361021, China
A gaseous nitrous acid (HONO) prediction model based on Stacking ensemble learning was proposed. The concentrations of HONO in Beijing urban area were obtained using incoherent broadband cavity enhanced absorption spectroscopy (IBBCEAS). Combined with the HONO sources, O3, CO, SO2, NO, NO2, NOy, temperature (T), relative humidity (RH), wind speed (WS), j(HONO), j(NO2), j(O1D) were selected as characteristic data. By analyzing the average diurnal variation of HONO, the measurement time was converted into a new feature hour by hour. The base model was constructed by utilizing Extreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM) and Random Forest (RF) algorithm. The training set was partitioned by 5-fold cross-validation method. The output of the base model was taken as a new feature set and as the input of second-level linear regression model. HONO prediction model was finally obtained via training the models in these two layers. Through the feature importance analysis and calculating the contribution of direct emission of vehicles at night, it showed that CO was an important impact factor in the prediction model, and that the direct emission of vehicles was a major source of HONO in the winter period at the region. The prediction performance of the base model and the Stacking ensemble model were evaluated by the test set respectively. The correlation coefficients between forecast results and measured values for the three base models were above 0.91. The performance of the Stacking ensemble model was the best, whose correlation coefficients reached 0.94. The average absolute error and root mean square error were 0.307×10-9 and 0.453×10-9, respectively. Explanability and applicability of the HONO prediction model based on Stacking ensemble learning.
Huang R J, Zhang Y, Bozzetti C, et al. High secondary aerosol contribution to particulate pollution during haze events in China[J]. Nature, 2014,514(7521):218-222.
[2]
Li G, Bei N, Cao J, et al. A possible pathway for rapid growth of sulfate during haze days in China[J]. Atmospheric Chemistry and Physics, 2017,17(5):3301-3316.
[3]
Sun Y L, Wang Z F, Fu P Q, et al. Aerosol composition, sources and processes during wintertime in Beijing, China[J]. Atmospheric Chemistry and Physics, 2013,13(9):4577-4592.
[4]
Deng W, Liu T, Zhang Y, et al. Secondary organic aerosol formation from photo-oxidation of toluene with NOx and SO2:chamber simulation with purified air versus urban ambient air as matrix[J]. Atmospheric Environment, 2017,150:67-76.
[5]
Tan Y, Lim Y B, Altieri K E, et al. Mechanisms leading to oligomers and SOA through aqueous photooxidation:insights from OH radical oxidation of acetic acid and methylglyoxal[J]. Atmospheric Chemistry and Physics, 2012,12(2):801-813.
[6]
Hou S Q, Tong S R, Ge M F, et al. Comparison of atmospheric nitrous acid during severe haze and clean periods in Beijing, China[J]. Atmospheric Environment, 2016,124:199-206.
[7]
Zhang J, Chen J, Xue C, et al. Impacts of six potential HONO sources on HOx budgets and SOA formation during a wintertime heavy haze period in the North China Plain[J]. Science of the Total Environment, 2019,681:110-123.
[8]
Fu X, Wang T, Zhang L, et al. The significant contribution of HONO to secondary pollutants during a severe winter pollution event in southern China[J]. Atmospheric Chemistry and Physics, 2019,19(1):1-14.
[9]
Liu Y, Lu K, Li X, et al. A comprehensive model test of the HONO sources constrained to field measurements at rural North China Plain[J]. Environment Science Technology, 2019,53(7):3517-3525.
[10]
Maljanen M, Yli-Pirilä P, Hytönen J, et al. Acidic northern soils as sources of atmospheric nitrous acid (HONO)[J]. Soil Biology and Biochemistry, 2013,67:94-97.
[11]
杨闻达,程鹏,田智林,等.广州市夏秋季HONO污染特征及白天未知源分析[J]. 中国环境科学, 2017,37(6):2029-2039. Yang W D, Cheng P, Tian Z L, et al. Study on HONO pollution characteristics and daytime unknown sources during summer and autumn in Guangzhou China[J]. China Environmental Science, 2017, 37(6):2029-2039.
[12]
Wall K J, Harris G W. Uptake of nitrogen dioxide (NO2) on acidic aqueous humic acid (Ha) solutions as a missing daytime nitrous acid (HONO) surface source[J]. Journal of Atmospheric Chemistry, 2016,74(3):283-321.
[13]
Zhou X, Zhang N, TerAvest M, et al. Nitric acid photolysis on forest canopy surface as a source for tropospheric nitrous acid[J]. Nature Geoscience, 2011,4(7):440-443.
[14]
Wauters M, Vanhoucke M. Support vector machine regression for project control forecasting[J]. Automation in Construction, 2014, 47:92-106.
[15]
Svetnik V, Liaw A, Tong C, et al. Random forest:a classification and regression tool for compound classification and osar modeling[J]. Journal of Chemical Information and Computer Sciences, 2003,43(6):1947-1958.
[16]
Santana D, Borges W, Poppi R J. Random forest as one-class classifier and infrared spectroscopy for food adulteration detection[J]. Food Chem, 2019,293:323-332.
[17]
Ma X, Tao Z, Wang Y, et al. Long short-term memory neural network for traffic speed prediction using remote microwave sensor data[J]. Transportation Research Part C:Emerging Technologies, 2015,54:187-197.
[18]
黄婕,张丰,杜震洪,等.基于RNN-CNN集成深度学习模型的PM2.5小时浓度预测[J]. 浙江大学学报(理学版), 2019,46(3):370-379. Huang J, Zhang F, Du Z H, et al. Hourly concentration prediction of PM2.5 based on RNN-CNN ensemble deep learning model[J]. Journal of Zhejiang University (Science Edition), 2019,46(3):370-379.
[19]
沈路路,王聿绚,段雷.神经网络模型在O3浓度预测中的应用[J]. 环境科学, 2011,32(8):2231-2235. Shen L L, Wang Y X, Duan L. Application of artificial neural networks on the prediction of surface ozone concentrations[J]. Environmental Science, 2011,32(8):2231-2235.
[20]
Duan J, Qin M, Ouyang B, et al. Development of an incoherent broadband cavity-enhanced absorption spectrometer for in situ measurements of HONO and NO2[J]. Atmospheric Measurement Techniques, 2018,11(7):4531-4543.
[21]
Dietterich T G. An experimental comparison of three methods for constructing ensembles of decision trees:bagging, boosting, and randomization[J]. Machine Learning, 2000,40(2):139-157.
[22]
Wu Y C, Qi S F, Hu F, et al. Guestrin C. Recognizing activities of the elderly using wearable sensors:a comparison of ensemble algorithms based on boosting[J]. 2019,39(6):743-751.
[23]
Gounaridis D, Koukoulas S. Urban land cover thematic disaggregation, employing datasets from multiple sources and random forests modeling[J]. International Journal of Applied Earth Observations & Geoinformation, 2016,51:1-10.
[24]
Ju Y, Sun G Y, Chen Q H, et al. A model combining convolutional neural network and lightGBM algorithm for ultra-short-term wind power forecasting[J]. IEEE Access, 2019,7(4):28309-28318.
[25]
Gao X, Luo H, Wang Q, et al. A human activity recognition algorithm based on stacking denoising autoencoder and lightgbm[J]. Sensors (Basel), 2019,19(4).
[26]
Lu X, Wang Y, Li J, et al. Evidence of heterogeneous HONO formation from aerosols and the regional photochemical impact of this HONO source[J]. Environmental Research Letters, 2018,13(11):114002.
[27]
孟凡浩,秦敏,梁帅西,等.合肥市典型交通干道大气苯系物的特征分析[J]. 环境科学, 2011,32(8):2231-2235. Meng F H, Qin M, Liang S X, et al. Characteristics of atmospheric BTX near a main road in Hefei city[J]. Environmental Science, 2011,32(8):2231-2235.
[28]
Su H, Cheng Y F, Cheng P, et al. Observation of nighttime nitrous acid (HONO) formation at a non-urban site during PRIDE-PRD2004 in China[J]. Atmospheric Environment, 2008,42(25):6219-6232.
[29]
Zhang W Q, Tong S R, Ge M F, et al. Variations and sources of nitrous acid (HONO) during a severe pollution episode in Beijing in winter 2016[J]. Science of the Total Environment, 2019,648:253-262.
