A new method was developed for classifying aerosol types involving a machine-learning approach to the use of satellite data. An Aerosol Robotic NETwork (AERONET)-based aerosol-type dataset was used as a target variable in a random forest (RF) model. The contributions of satellite input variables to the RF-based model were quantified to determine an optimal set of input variables. The new method, based on inputs of satellite variables, allows the classification of seven aerosol types: pure dust, dust-dominant mixed, pollution-dominant mixed aerosols, and pollution aerosols (strongly, moderately, weakly, and non-absorbing). The performance of the model was statistically evaluated using AERONET data excluded from the model training dataset. Model accuracy for classifying the seven aerosol types was 59%, improving to 72% for four types (pure dust, dust-dominant mixed, strongly absorbing, and non-absorbing). The performance of the model was evaluated against an earlier aerosol classification method based on the wavelength dependence of single-scattering albedo (SSA) and fine-mode-fraction values from AERONET. Typical wavelength dependences of SSA for individual aerosol types are consistent with those obtained for aerosol types by the new method. This study demonstrates that an RF-based model is capable of satellite aerosol classification with sensitivity to the contribution of non-spherical particles.

An investigation into the diurnal characteristics of vertical formaldehyde (HCHO) profiles was conducted based on multi-axis differential optical absorption spectroscopy (MAX-DOAS) measurements in Beijing during the CAREBEIJING campaign, covering a month-long period through August and September 2006. Vertical HCHO profiles were retrieved based on a combined differential optical absorption spectroscopy (DOAS) technique and an optimal estimation method (OEM). The HCHO volume-mixing ratio (VMR) was found to be highest in the layer from the surface up to an altitude of 1 km and to decrease with altitude above this layer. In all retrieved profiles, HCHO was not detected in the layer from 3-4 km. Over the diurnal cycle, the HCHO VMR values were generally highest at 15:00 local time (LT) and were lower in the morning and late afternoon. The mean HCHO VMRs were 6.17, 1.82, and 0.80 ppbv for the 0-1, 1-2, and 2-3-km layers, respectively, at 15:00 LT, whereas they were 3.54 (4.79), OPEN ACCESSAtmosphere 2015, 6 1817 1.06 (1.43), and 0.46 (0.63) ppbv for the 0-1, 1-2, and 2-3-km layers, respectively, at 09:00 (17:00) LT. The HCHO VMRs reached their highest values at 15:00 LT on August 19, which were 17.71, 5.20, and 2.31 ppbv for the 0-1, 1-2, and 2-3-km layers, respectively. This diurnal pattern implies that the photo-oxidation of volatile organic compounds (VOCs) was most active at 15:00 LT for several days during the campaign period. In a comparison of the derived HCHO VCDs with those obtained from the Ozone Monitoring Instrument (OMI) measurements, the HCHO vertical column density (VCD) values obtained from the OMI measurements tend to be smaller than those from the MAX-DOAS.

The spatial coverage of satellite aerosol classification was improved using a random forest (RF) model trained with observational data including target (aerosol type) and input (satellite measurement) variables. The AErosol RObotic NETwork (AERONET) aerosol-type dataset was used for the target variables. Satellite input variables with many missing data or low mean-decrease accuracy were excluded from the final input variable set, and good performance in aerosol-type classification was achieved. The performance of the RF-based model was evaluated on the basis of the wavelength dependence of single-scattering albedo (SSA) and fine-mode-fraction values from AERONET. Typical SSA wavelength dependence for individual aerosol types was consistent with that obtained for aerosol types by the RF-based model. The spatial coverage of the RF-based model was also compared with that of previously developed models in a global-scale case study. The study demonstrates that the RF-based model allows satellite aerosol classification with improved spatial coverage, with a performance similar to that of previously developed models.

Abstract:In the present study, we investigate the effects of signal to noise (SNR), slit function (FWHM), and aerosol optical depth (AOD) on the accuracy of formaldehyde (HCHO) vertical column density (HCHO VCD ) using the ground-based direct-sun synthetic radiance based on differential optical absorption spectroscopy (DOAS). We found that the effect of SNR on HCHO retrieval accuracy is larger than those of FWHM and AOD. When SNR = 650 (1300), FWHM = 0.6, and AOD = 0.2, the absolute percentage difference (APD) between the true HCHO VCD values and those retrieved ranges from 54 (30%) to 5% (1%) for the HCHO VCD of 5.0 × 10 15 and 1.1 × 10 17 molecules cm −2 , respectively. Interestingly, the maximum AOD effect on the HCHO accuracy was found for the HCHO VCD of 3.0 × 10 16 molecules cm −2 . In addition, we carried out the first ground-based direct-sun measurements in the ultraviolet (UV) wavelength range to retrieve the HCHO VCD using Pandora in Seoul. The HCHO VCD was low at 12:00 p.m. local time (LT) in all seasons, whereas it was high in the morning (10:00 a.m. LT) and late afternoon (4:00 p.m. LT), except in winter. The maximum HCHO VCD values were 2.68 × 10 16 , 3.19 × 10 16 , 2.00 × 10 16 , and 1.63 × 10 16 molecules cm −2 at 10:00 a.m. LT in spring, 10:00 a.m. LT in summer, 1:00 p.m. LT in autumn, and 9:00 a.m. LT in winter, respectively. The minimum values of Pandora HCHO VCD were 1.63 × 10 16 , 2.23 × 10 16 , 1.26 × 10 16 , and 0.82 × 10 16 molecules cm −2 at around 1:45 p.m. LT in spring, summer, autumn, and winter, respectively. This seasonal pattern of high values in summer and low values in winter implies that photo-oxidation plays an important role in HCHO production. The correlation coefficient (R) between the monthly HCHO VCD values from Pandora and those from the Ozone Monitoring Instrument (OMI) is 0.61, and the slope is 1.25.

Surface NO 2 volume mixing ratio (VMR) at a specific time (13:45 Local time) (NO 2 VMR ST ) and monthly mean surface NO 2 VMR (NO 2 VMR M ) are estimated for the first time using three regression models with Ozone Monitoring Instrument (OMI) data in four metropolitan cities in South Korea: Seoul, Gyeonggi, Daejeon, and Gwangju. Relationships between the surface NO 2 VMR obtained from in situ measurements (NO 2 VMR In-situ ) and tropospheric NO 2 vertical column density obtained from OMI from 2007 to 2013 were developed using regression models that also include boundary layer height (BLH) from Atmospheric Infrared Sounder (AIRS) and surface pressure, temperature, dew point, and wind speed and direction. The performance of the regression models is evaluated via comparison with the NO 2 VMR In-situ for two validation years (2006 and 2014). Of the three regression models, a multiple regression model shows the best performance in estimating NO 2 VMR ST and NO 2 VMR M . In the validation period, the average correlation coefficient (R), slope, mean bias (MB), mean absolute error (MAE), root mean square error (RMSE), and percent difference between NO 2 VMR In-situ and NO 2 VMR ST estimated by the multiple regression model are 0.66, 0.41, −1.36 ppbv, 6.89 ppbv, 8.98 ppbv, and 31.50%, respectively, while the average corresponding values for the other two models are 0.75, 0.41, −1.40 ppbv, 3.59 ppbv, 4.72 ppbv, and 16.59%, respectively. All three models have similar performance for NO 2 VMR M , with average R, slope, MB, MAE, RMSE, and percent difference between NO 2 VMR In-situ and NO 2 VMR M of 0.74, 0.49, −1.90 ppbv, 3.93 ppbv, 5.05 ppbv, and 18.76%, respectively.