Aerosol optical depth assimilation for a modal aerosol model: Implementation and application in AOD forecasts over East Asia
Graphical abstract
Introduction
Atmospheric aerosol has important impacts on visibility issues, regional air quality, climate change, and human health (e.g., Lowe and Zealand, 2007; Pope et al., 2002; Ramanathan et al., 2008; Wan Mahiyuddin et al., 2013). Therefore, it is essential to estimate aerosol accurately. However, aerosol simulation by using numerical models remain contains large uncertainties due to the imperfect on model parametrization, emission inventories, and aerosol processes. Data assimilation (DA) has been utilized in many works to improve the simulation precision, which try to reduce the uncertainties of the input data for a model by assimilating aerosol-related observations (e.g., Chen et al., 2019a; Cheng et al., 2019; Liu et al., 2011; Ma et al., 2019; Pang et al., 2018; Peng et al., 2018; Saide et al., 2014). Due to the widely coverage and long period observing, aerosol optical depth (AOD) has been widely applied in data assimilation. DA was used to reduce the uncertainties of model inputs, including the initial conditions (ICs) (e.g., Elbern and Schmidt, 2001; Feng et al., 2018; Pang et al., 2018; Peng et al., 2018; Saide et al., 2014), the boundary conditions (BCs) (e.g., Roustan and Bocquet, 2006), emission rate (e.g., Elbern et al., 2007; Ma et al., 2019; Peng et al., 2018; Peng et al., 2017; Qu et al., 2017; Wang et al., 2012; Wang et al., 2016; Xu et al., 2013), and model parameters (e.g., Barbu et al., 2009; Bocquet, 2012). In this study, AOD DA was applied to improve the accuracy of aerosol ICs and the subsequent model forecast by using Weather Research and Forecasting/Chemistry (WRF/Chem) model (Grell et al., 2005), which has been shown a probable improvement in some works (e.g., Chen et al., 2017; Liu et al., 2011; Pang et al., 2018; Saide et al., 2013; Saide et al., 2014; Schwartz et al., 2012).
Within Gridpoint Statistical Interpolation (GSI) 3DVAR DA system (Kleist et al., 2009; Wu et al., 2002), based on the community radiative transfer model (CRTM) (Han et al., 2006; Liu and Weng, 2006), Liu et al. (2011) developed CRTM-AOD module for the Goddard Chemistry Aerosol Radiation and Transport (GOCART) aerosol scheme (Chin et al., 2002; Chin et al., 2000; Ginoux et al., 2001) in WRF/Chem. This DA framework has been utilized over East Asia (Jiang et al., 2013; Pang et al., 2018; Peng et al., 2017; Xia et al., 2019), the United States (Chen et al., 2014; Schwartz et al., 2012), and Atlantic Ocean (Chen et al., 2017). However, GOCART is a bulk aerosol model, missing secondary organic aerosols, nitrate, and ammonium interactions, which might produce uncertainties on aerosol simulations (McKeen et al., 2009; Volkamer et al., 2006; Zhang et al., 2007). Based on the GSI system developed by Liu et al. (2011) and Schwartz et al. (2012), Saide et al. (2013) extended it for a sectional aerosol model in WRF/Chem, the Model for Simulating Aerosol Interactions and Chemistry (MOSAIC) (Zaveri et al., 2008). And it showed the probable improvement of aerosol predictions by assimilating satellite AOD retrieval. In that AOD DA framework, the Fast-J optical module within WRF/Chem was incorporated into GSI as the AOD observation operator, which is highly nonlinear. The generation of the corresponding tangent linear (TL) and adjoint (AD) code were very complex and difficult, which was conducted with the automatic differentiation tool TAPENADE v3.6 (Hascoët and Pascual, 2004). But it is not easy to apply to the other aerosol models nor the other optical modules. Also, based on the GSI system developed for GOCART AOD DA, Tang et al. (2017) and Cheng et al. (2019) extended it to suit for the Community Multi-scale Air Quality Model (CMAQ) and MOSAIC respectively, through “variables matching” way. They transfer the control variables between the original scheme (GOCART) and target scheme (CMAQ and MOSAIC), according to the similar aerosol species and bin sizes. Even though this approach was efficient for coding, it might be inappropriate when the control variables of the target scheme are significantly different from the original scheme, failed to converge during the minimization procedure.
Moreover, another most commonly used aerosol mechanism in WRF/Chem haven't been applied on AOD DA with 3DVAR method, the Modal Aerosol Dynamics Model for Europe (MADE) (Ackermann et al., 1998) with the secondary organic aerosol model (SORGAM) of Schell et al. (2001) (MADE/SORGAM). Different aerosol mechanisms have different aerosol processes, which affect the aerosol simulation. Also, it might affect the impacts of aerosol DA on aerosol prediction. Currently, there are many works studying the sensitivities among different aerosol mechanisms (e.g., Chen et al., 2019b; Kim et al., 2011; Yang et al., 2018; Zhang et al., 2016). But, the sensitivity analysis of the aerosol DA impacts on aerosol predictions with different aerosol mechanisms was rarely studied. Werner et al. (2019) studied the sensitivity of ground PM2.5 with two chemical schemes by assimilating ground PM2.5 observations. However, AOD assimilation is more complicated than ground aerosol concentrations assimilation.
In this study, a new AOD DA module, named FastJ/CRTM-AOD module, was developed through an efficient way within GSI-3DVAR system. In order to utilize the FastJ/CRTM-AOD module to as many chemical options in WRF/Chem as possible, this new developed module was applied to MADE/SORGAM scheme. This AOD DA module can be easily and efficiently expanded to suit to the other aerosol mechanisms and optical modules. Moreover, on the purpose of comparing the AOD DA impacts on AOD forecast between different aerosol mechanisms, the CRTM-AOD DA module applied with GOCART mechanism, the original AOD DA scheme in GSI, was employed for the comparisons during January (regarded as winter) and April (regarded as spring) of 2014.
Section 2 describes the methods and data, including the new AOD DA module developed in this study, data used for assimilation and evaluation, and the experimental design. The assimilation results and discussions are presented in Section 3, before a brief summary in Section 4.
Section snippets
Configurations of WRF/Chem model
GOCART and MADE/SORGAM aerosol schemes in WRF/Chem were chosen to study the applicability of AOD predictions over a complicated polluted area, East Asia. The model configuration was based on Pang et al. (2018). A succinct summary of follows and significant differences are also pointed out here. Some physical parameterizations employed in this study include Goddard shortwave radiation (Chou et al., 1998), RRTM longwave radiation scheme (Mlawer et al., 1997), and YSU Boundary layer scheme (Hong
DA impacts on aerosol ICs and AOD forecasts
The domain-averaged concentration as a function of the height for the total aerosol mass and the main components are shown in Fig. 3. It shows that AOD DA affected the aerosol ICs significantly. And, the characteristics of DA impacts were consistent with the characteristics of BEC. AOD DA mainly affected “so4aj”, “no3aj”, and “nh4aj” in MADE/SORGAM, while the impacts on the other components were excessively small. For GOCART, AOD assimilation mainly affected “sulf” and “OC2”. Meanwhile, AOD was
Conclusions
In this study, the FastJ/CRTM-AOD module was constructed in GSI-3DVAR system, and applied to the MADE/SROGAM aerosol mechanism in WRF/Chem. The Fast-J optical module in WRF/Chem was used as the AOD forward operator. The corresponding Jacobian code was modified from the one in CRTM-AOD module. Meanwhile, the performance on AOD DA with MADE/SORGAM, was compared with the one of CRTM-AOD module applied with GOCART, by evaluating with AOD observations from ten AERONET sites in January and April 2014.
CRediT authorship contribution statement
Jiongming Pang: Conceptualization, Methodology, Software, Validation, Writing - original draft. Xuemei Wang: Supervision, Resources, Writing - review & editing. Min Shao: Investigation, Writing - review & editing. Weihua Chen: Data curation, Validation. Ming Chang: Data curation, Validation.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This study was supported by the National Key Research and Development Program of China (grant nos. 2017YFC0210100 and 2016YFC0202206), the National Natural Science Foundation of China (grant nos. 41425020 and 91644215), Special Fund Project for Science and Technology Innovation Strategy of Guangdong Province (grant no. 2019B121205004), and the Jiangsu Collaborative Innovation Center for Climate Change, China. We would like to thank Zhiquan Liu for patient guidance and useful suggestion
References (72)
- et al.
Modal aerosol dynamics model for Europe: development and first applications
Atmos. Environ.
(1998) - et al.
A multi-component data assimilation experiment directed to sulphur dioxide and sulphate over Europe
Atmos. Environ.
(2009) - et al.
Lidar data assimilation method based on CRTM and WRF-Chem models and its application in PM2.5 forecasts in Beijing
Sci. Total Environ.
(2019) - et al.
Impact of 3DVAR assimilation of surface PM2.5 observations on PM2.5 forecasts over China during wintertime
Atmos. Environ.
(2018) - et al.
Fully coupled “online” chemistry within the WRF model
Atmos. Environ.
(2005) - et al.
Assimilating AOD retrievals from GOCI and VIIRS to forecast surface PM2.5 episodes over Eastern China
Atmos. Environ.
(2018) - et al.
Short-term effects of daily air pollution on mortality
Atmos. Environ.
(2013) - et al.
Top-down estimate of dust emissions through integration of MODIS and MISR aerosol retrievals with the GEOS-Chem adjoint model
Geophys. Res. Lett.
(2012) - et al.
Reductions in sulfur pollution in the Pearl River Delta region, China: assessing the effectiveness of emission controls
Atmos. Environ.
(2013) - et al.
Assimilation of PM2.5 ground base observations to two chemical schemes in WRF-Chem – the results for the winter and summer period
Atmos. Environ.
(2019)
Aerosol analysis and forecast in the European centre for medium-range weather forecasts integrated forecast system: 2. Data assimilation
J. Geophys. Res. Atmos.
The regional particulate matter model: 1. Model description and preliminary results
J. Geophys. Res. Atmos.
Parameter-field estimation for atmospheric dispersion: application to the Chernobyl accident using 4D-Var
Q. J. R. Meteorol. Soc.
Early on-orbit performance of the visible infrared imaging radiometer suite onboard the Suomi National Polar-Orbiting Partnership (S-NPP) satellite
IEEE Trans. Geosci. Remote Sens.
The impact of aerosol optical depth assimilation on aerosol forecasts and radiative effects during a wild fire event over the United States
Geosci. Model Dev.
Dust radiative effects on atmospheric thermodynamics and tropical cyclogenesis over the Atlantic Ocean using WRF-Chem coupled with an AOD data assimilation system
Atmos. Chem. Phys.
Retrospective analysis of 2015–2017 wintertime PM2.5 in China: response to emission regulations and the role of meteorology
Atmos. Chem. Phys.
MICS-Asia III: multi-model comparison and evaluation of aerosol over East Asia
Atmos. Chem. Phys.
Atmospheric sulfur cycle simulated in the global model GOCART: comparison with field observations and regional budgets
J. Geophys. Res. Atmos.
Tropospheric aerosol optical thickness from the GOCART model and comparisons with satellite and sun photometer measurements
J. Atmos. Sci.
Parameterizations for Cloud Overlapping and Shortwave Single-scattering Properties for Use in General Circulation and Cloud Ensemble Models
Ozone episode analysis by four-dimensional variational chemistry data assimilation
J. Geophys. Res. Atmos.
Emission rate and chemical state estimation by 4-dimensional variational inversion
Atmos. Chem. Phys.
Evolution of ozone, particulates, and aerosol direct radiative forcing in the vicinity of Houston using a fully coupled meteorology-chemistry-aerosol model
J. Geophys. Res. Atmos.
PREP-CHEM-SRC – 1.0: a preprocessor of trace gas and aerosol emission fields for regional and global atmospheric chemistry models
Geosci. Model Dev.
Sources and distributions of dust aerosols simulated with the GOCART model
J. Geophys. Res. Atmos.
JCSDA Community Radiative Transfer Model (CRTM) - Version 1. Technical Report
TAPENADE 2.1 user’s Guide
Multi-resolution Emission Inventory for China (MEIC): Model Framework and 1990–2010 Anthropogenic Emissions. AGU Fall Meeting Abstracts
An emerging ground-based aerosol climatology: aerosol optical depth from AERONET
J. Geophys. Res. Atmos.
A New Vertical Diffusion Package With an Explicit Treatment of Entrainment Processes
Suomi-NPP VIIRS aerosol algorithms and data products
J. Geophys. Res. Atmos.
Probing into the impact of 3DVAR assimilation of surface PM10 observations over China using process analysis
J. Geophys. Res. Atmos.
Comparison of different gas-phase mechanisms and aerosol modules for simulating particulate matter formation
J. Air Waste Manage. Assoc.
Introduction of the GSI into the NCEP global data assimilation system
Weather Forecast.
Primary anthropogenic aerosol emission trends for China, 1990-2005
Atmos. Chem. Phys.
Cited by (6)
The impacts of background error covariance on particulate matter assimilation and forecast: An ideal case study with a modal aerosol model over China
2021, Science of the Total EnvironmentCitation Excerpt :Thus, data assimilation (DA) became an essential tool for improving the accuracy of many atmospheric environment forecasting and analysis systems, which assimilating the related observations (Chen et al., 2019; Cheng et al., 2019; Liu et al., 2011; Ma et al., 2019; Pang et al., 2018; Peng et al., 2017; Saide et al., 2014) to decrease the uncertainties of the input data for the numerical model, including the initial conditions (ICs) (Feng et al., 2018; Pang et al., 2020; Saide et al., 2014), the boundary conditions (BCs) (Roustan and Bocquet, 2006), emission rate (Elbern et al., 2007; Ma et al., 2019; Peng et al., 2020; Qu et al., 2017; Wang et al., 2012), and model parameters (Barbu et al., 2009; Bocquet, 2012). Here, aerosol DA was used to reduce the uncertainties of aerosol ICs and the subsequent forecast by applying Weather Research and Forecasting/Chemistry (WRF/Chem) model (Grell et al., 2005), which has exhibited a probable enhancement in some works (Chen et al., 2019; Cheng et al., 2019; Pang et al., 2020; Peng et al., 2020). The background error covariance (BEC) plays a vital role in data assimilation.
A comprehensive evaluation of aerosol extinction apportionment in Beijing using a high-resolution time-of-flight aerosol mass spectrometer
2021, Science of the Total EnvironmentCitation Excerpt :Black carbon, dust and brown carbon are typical absorbing aerosols and sulfate, nitrate, ammonium and most organics are typical scattering aerosols (Seinfeld and Pandis, 2016). Quantifying the contribution of chemical component to extinction is an important step in pollution control, and helpful for improving parameterization schemes of the atmospheric chemical model and beneficial for improving the simulation accuracy of aerosol optical properties (Pang et al., 2020). IMPROVE (Interagency Monitoring of PROtected Visual Environments) formula were often utilized to do such study (Cao et al., 2012; Chen et al., 2014; Pitchford et al., 2007; Tao et al., 2017), but the initial formula may not be suitable for China due to the differences of aerosol chemical component and size distribution (Hand and Malm, 2007; Huang et al., 2017; Xin et al., 2015).
Bayesian Inference Approach to Quantify Primary and Secondary Organic Carbon in Fine Particulate Matter Using Major Species Measurements
2023, Environmental Science and TechnologyImproved Modeling of Spatiotemporal Variations of Fine Particulate Matter Using a Three-Dimensional Variational Data Fusion Method
2021, Journal of Geophysical Research: AtmospheresHandling missing data in large-scale modis aod products using a two-step model
2020, Remote Sensing