Performance comparisons of the three data assimilation methods for improved predictability of PM2·5: Ensemble Kalman filter, ensemble square root filter, and three-dimensional variational methods.

To improve the predictability of concentrations of atmospheric particulate matter, a data assimilation (DA) system using ensemble square root filter (EnSRF) has been developed for the Community Multiscale Air Quality (CMAQ) model. The EnSRF DA method is a deterministic variant of the ensemble Kalman filter (EnKF) method, which means that unlike the EnKF method, it does not add random noise to the observations. To compare the performances of the EnSRF with those of other DA methods, such as EnKF and 3DVAR (three-dimensional variational), these three methods were applied to the same CMAQ model simulations with identical experimental settings. This is the first attempt in the field of chemical DA to compare the EnKF and EnSRF methods. An identical set of surface fine particulate matter (PM2.5) were assimilated every 6 h by all the DA methods over a CMAQ domain of East Asia, during the period from 01 May to 11 June 2016. In parallel with 'reanalysis experiments', we also carried out '48 h prediction experiments' using the optimized initial conditions produced by the three DA methods. Detailed analyses among the three DA methods were then carried out by comparing both the reanalysis and the prediction outputs with the observed surface PM2.5 over four regions (i.e., South Korea, the Beijing-Tianjin-Hebei (BTH) region, Shandong province, and Liaoning province). The comparison results revealed that the EnSRF produced the best reanalysis and prediction fields in terms of several statistical metrics. For example, when the 3DVAR, EnKF, and EnSRF methods were used, averaged normalized mean biases (NMBs) decreased by (57.6, 85.6, and 91.8) % in reanalyses and (39.7, 87.6, and 91.5) % in first-day predictions, compared to the CMAQ control experiment (i.e., without DA) over South Korea, respectively. Also, over the three Chinese regions, the EnSRF method outperformed the EnKF and 3DVAR methods.

Demonstration of aerosol profile measurement with a dual-wavelength high-spectral-resolution lidar using a scanning interferometer.

Simple dual-wavelength high-spectral-resolution lidar at 355 and 532 nm with a scanning interferometer was developed for continuous observations of aerosol profiles. Scanning the interferometer periodically over a range of one fringe at 532 nm (1.5 fringes at 355 nm) enabled recording of range-resolved interference signals at these two wavelengths. Reference signals taken from the transmitted laser were used to correct the interference phase shift due to laser frequency variation for every scan. Profiles of aerosol backscatter and extinction coefficients were retrieved from range-resolved interference data. One month of continuous measurements demonstrated the robustness of the system.

Impacts of COVID-19 lockdown, Spring Festival and meteorology on the NO2 variations in early 2020 over China based on in-situ observations, satellite retrievals and model simulations.

The lockdown measures due to COVID-19 affected the industry, transportation and other human activities within China in early 2020, and subsequently the emissions of air pollutants. The decrease of atmospheric NO2 due to the COVID-19 lockdown and other factors were quantitively analyzed based on the surface concentrations by in-situ observations, the tropospheric vertical column densities (VCDs) by different satellite retrievals including OMI and TROPOMI, and the model simulations by GEOS-Chem. The results indicated that due to the COVID-19 lockdown, the surface NO2 concentrations decreased by 42% ± 8% and 26% ± 9% over China in February and March 2020, respectively. The tropospheric NO2 VCDs based on both OMI and high quality (quality assurance value (QA) ≥ 0.75) TROPOMI showed similar results as the surface NO2 concentrations. The daily variations of atmospheric NO2 during the first quarter (Q1) of 2020 were not only affected by the COVID-19 lockdown, but also by the Spring Festival (SF) holiday (January 24-30, 2020) as well as the meteorology changes due to seasonal transition. The SF holiday effect resulted in a NO2 reduction from 8 days before SF to 21 days after it (i.e. January 17 - February 15), with a maximum of 37%. From the 6 days after SF (January 31) to the end of March, the COVID-19 lockdown played an important role in the NO2 reduction, with a maximum of 51%. The meteorology changes due to seasonal transition resulted in a nearly linear decreasing trend of 25% and 40% reduction over the 90 days for the NO2 concentrations and VCDs, respectively. Comparisons between different datasets indicated that medium quality (QA ≥ 0.5) TROPOMI retrievals might suffer large biases in some periods, and thus attention must be paid when they are used for analyses, data assimilations and emission inversions.

Paradigm shift in aerosol chemical composition over regions downwind of China.

A rapid decrease in PM2.5 concentrations in China has been observed in response to the enactment of strong emission control policies. From 2012 to 2017, total emissions of SO2 and NOx from China decreased by approximately 63% and 24%, respectively. Simultaneously, decreases in the PM2.5 concentration in Japan have been observed since 2014, and the proportion of stations that satisfy the PM2.5 environmental standard (daily, 35 µg/m3; annual average, 15 µg/m3) increased from 37.8% in fiscal year (FY) 2014 (April 2014 to March 2015) to 89.9% in FY 2017. However, the quantitative relationship between the PM2.5 improvement in China and the PM2.5 concentration in downwind regions is not well understood. Here, we (1) quantitatively evaluate the impacts of Chinese environmental improvements on downwind areas using source/receptor analysis with a chemical transport model, and (2) show that these rapid emissions reductions improved PM2.5 concentrations both in China and its downwind regions, but the difference between SO2 and NOx reduction rates led to greater production of nitrates (e.g., NH4NO3) due to a chemical imbalance in the ammonia-nitric acid-sulfuric acid-water system. Observations from a clean remote island in western Japan and numerical modeling confirmed this paradigm shift.

Current state of the global operational aerosol multi-model ensemble: An update from the International Cooperative for Aerosol Prediction (ICAP).

Since the first International Cooperative for Aerosol Prediction (ICAP) multi-model ensemble (MME) study, the number of ICAP global operational aerosol models has increased from five to nine. An update of the current ICAP status is provided, along with an evaluation of the performance of ICAP-MME over 2012-2017, with a focus on June 2016-May 2017. Evaluated with ground-based Aerosol Robotic Network (AERONET) aerosol optical depth (AOD) and data assimilation quality MODerate-resolution Imaging Spectroradiometer (MODIS) retrieval products, the ICAP-MME AOD consensus remains the overall top-scoring and most consistent performer among all models in terms of root-mean-square error (RMSE), bias and correlation for total, fine- and coarse-mode AODs as well as dust AOD; this is similar to the first ICAP-MME study. Further, over the years, the performance of ICAP-MME is relatively stable and reliable compared to more variability in the individual models. The extent to which the AOD forecast error of ICAP-MME can be predicted is also examined. Leading predictors are found to be the consensus mean and spread. Regression models of absolute forecast errors were built for AOD forecasts of different lengths for potential applications. ICAP-MME performance in terms of modal AOD RMSEs of the 21 regionally representative sites over 2012-2017 suggests a general tendency for model improvements in fine-mode AOD, especially over Asia. No significant improvement in coarse-mode AOD is found overall for this time period.

Dust Vortex in the Taklimakan Desert by Himawari-8 High Frequency and Resolution Observation.

The Taklimakan Desert is known to be one of the world's major sources of aeolian dust particles. Continuous images with 10-min temporal and 2-km spatial resolutions from a new-generation geostationary meteorological satellite captured the lifecycle (generation, evolution and outflow) of a previously unrecognized type of Taklimakan dust storm. The dust storm showed an anti-clockwise spiral structure and a clear core and behaved like a "dust vortex". From image analysis, the horizontal scale and temporal lifetime of the dust vortex were estimated to be 600 km and 36 hours, respectively. We found that a strong pressure trough (cut-off low), along with a cold air mass located on the northwestern side of the Taklimakan Desert and the high mountains surrounding the Taklimakan Desert, played important roles in the formation and evolution of the dust vortex.

Chinese province-scale source apportionments for sulfate aerosol in 2005 evaluated by the tagged tracer method.

Appropriate policies to improve air quality by reducing anthropogenic emissions are urgently needed. This is typified by the particulate matter (PM) problem and it is well known that one type of PM, sulfate aerosol (SO42-), has a large-scale impact due to long range transport. In this study we evaluate the source-receptor relationships of SO42- over East Asia for 2005, when anthropogenic sulfur dioxide (SO2) emissions from China peaked. SO2 emissions from China have been declining since 2005-2006, so the possible maximum impact of Chinese contributions of SO42- is evaluated. This kind of information provides a foundation for policy making and the estimation of control effects. The tagged tracer method was applied to estimate the source apportionment of SO42- for 31 Chinese province-scale regions. In addition, overall one-year source apportionments were evaluated to clarify the seasonal dependency. Model performance was confirmed by comparing with ground-based observations over mainland China, Taiwan, Korea, and Japan, and the model results fully satisfied the performance goal for PM. We found the following results. Shandong and Hebei provinces, which were the largest and second largest SO2 sources in China, had the greatest impact over the whole of East Asia with apportionments of around 10-30% locally and around 5-15% in downwind receptor regions during the year. Despite large SO2 emissions, the impact of south China (e.g., Guizhou, Guangdong, and Sichuan provinces) was limited to local impact. These results suggest that the reduction policy in south China contributes to improving the local air quality, whereas policies in north and central China are beneficial for both the whole of China and downwind regions. Over Taiwan, Korea, and Japan, the impact of China was dominant; however, local contributions were important during summer.

Inverse modeling of the 137Cs source term of the Fukushima Dai-ichi Nuclear Power Plant accident constrained by a deposition map monitored by aircraft.

The amount of 137Cs released by the Fukushima Dai-ichi Nuclear Power Plant accident of 11 March 2011 was inversely estimated by integrating an atmospheric dispersion model, an a priori source term, and map of deposition recorded by aircraft. An a posteriori source term refined finer (hourly) variations comparing with the a priori term, and estimated 137Cs released 11 March to 2 April to be 8.12 PBq. Although time series of the a posteriori source term was generally similar to those of the a priori source term, notable modifications were found in the periods when the a posteriori source term was well-constrained by the observations. Spatial pattern of 137Cs deposition with the a posteriori source term showed better agreement with the 137Cs deposition monitored by aircraft. The a posteriori source term increased 137Cs deposition in the Naka-dori region (the central part of Fukushima Prefecture) by 32.9%, and considerably improved the underestimated a priori 137Cs deposition. Observed values of deposition measured at 16 stations and surface atmospheric concentrations collected on a filter tape of suspended particulate matter were used for validation of the a posteriori results. A great improvement was found in surface atmospheric concentration on 15 March; the a posteriori source term reduced root mean square error, normalized mean error, and normalized mean bias by 13.4, 22.3, and 92.0% for the hourly values, respectively. However, limited improvements were observed in some periods and areas due to the difficulty in simulating accurate wind fields and the lack of the observational constraints.

Long-term inverse modeling of Chinese CO emission from satellite observations.

Carbon monoxide (CO) emissions in China in 2005-2010 were estimated by inversion, using the Green's function method from vertical CO profiles derived from MOPITT Version 5 satellite data and a tagged CO simulation, and validated with independent in situ observations from the World Data Centre for Greenhouse Gases. Modeling with a posteriori emission successfully reproduced CO outflow from the continent to the East China Sea, Sea of Japan, and Japanese islands during winter and spring, and compensated for underestimates in central and eastern China in summer. A posteriori emissions showed large seasonal variations in which December and March emissions were on average 23% larger than August emissions, consistent with other studies. Estimated Chinese CO emissions were 184.4, 173.1, 184.6, 158.4, 157.4, and 157.3 Tg/year for 2005-2010, respectively. The decrease after 2007 is partly attributed to Chinese socioeconomic conditions and improved combustion efficiency.