See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/346954679

Advances in air quality modeling and forecasting

Article in Global Transitions · December 2020

DOI: 10.1016/j.glt.2020.11.001

CITATIONS READS

90 684

2 authors, including:

Alexander Baklanov
World Meteorological Organization
567 PUBLICATIONS 11,737 CITATIONS

SEE PROFILE

All content following this page was uploaded by Alexander Baklanov on 11 December 2020.

The user has requested enhancement of the downloaded file.

 Global Transitions 2 (2020) 261e270

Contents lists available at ScienceDirect

Global Transitions
journal homepage: www.keaipublishing.com/en/journals/global-transitions/

Advances in air quality modeling and forecasting

Alexander Baklanov a, b, *, Yang Zhang c
a
Science and Innovation Department, World Meteorological Organization (WMO), 7 bis, Avenue de la Paix, BP2300, 1211, Geneva 2, Switzerland
b
Climate and Geophysics, Niels Bohr Institute, University of Copenhagen, 2100, København Ø, Denmark
c
Department of Civil and Environmental Engineering, Northeastern University, Boston, MA, 02115, USA

a r t i c l e i n f o a b s t r a c t

Article history: The importance of and interest to research and investigations of atmospheric composition and its
Received 13 April 2020 modeling for different applications are substantially increased. Air quality forecast (AQF) and assessment
Received in revised form systems help decision makers to improve air quality and public health, mitigate the occurrence of acute
1 November 2020
air pollution episodes, particularly in urban areas, and reduce the associated impacts on agriculture,
Accepted 1 November 2020
ecosystems and climate. Advanced approaches in AQF combine an ensemble of state-of-the-art models,
high-resolution emission inventories, satellite observations, and surface measurements of most relevant
chemical species to provide hindcasts, analyses, and forecasts from global to regional air pollution and
Keywords:
Aerosols
downscaling for selected countries, regions, and urban areas. Based on published reviews and recent
Air pollution analyses, the article discusses main gaps, challenges, applications and advances, main trends and
Air quality research needs in further advancements of atmospheric composition and air quality modeling and
Atmospheric chemistry forecasting.
Dispersion © 2020 The Authors. Production and hosting by Elsevier B.V. on behalf of KeAi Communications Co., Ltd.
This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

1. Implications for sustainable development. The United Nations 17 Sustainable

Development Goals (SDGs) and 169 Targets aim to end extreme
Proceeding from global challenges and based on published re- poverty and create a healthy, sustainable world by the year 2030
views and recent analyses, the article discusses main gaps, chal- [1]. At their core lies the health and well-being of people and our
lenges, applications and advances, main trends and research needs planet. It means air pollution - the deaths and disabilities it causes
in further advancements of atmospheric composition and air and its close links to climate change - is a huge threat to delivering
quality modeling and forecasting. on the vision of a better world. Air pollution is specifically
mentioned under at least the following four SDGs: Goal 3 focuses
on good health and well-being; Goal 7 targets access to clean and
2. Introduction affordable energy; Goal 11 focuses on sustainable cities and com-
munities; and Goal 13 tackles climate change. Strategy and actions
Global challenges, such as climate change, environmental to improve air quality - such as switching to cleaner energy, green
degradation, rapid urbanization, or pandemics, require a review of economy, smart cities, cooking and transport solutions - will also
the current understanding and traditional methods of assessing address the climate emergency.
anthropogenic impacts on the environment, and new multi- During the last decades the importance of and interest to
disciplinary approaches of integrated assessments and methods research and investigations of atmospheric composition and its
modeling have substantially increased. There are several reasons
for these, including increased risks of air pollution for human
* Corresponding author. Science and Innovation Department, World Meteoro-
health, dramatically increased studies of climate change, much
logical Organization (WMO), 7 bis, Avenue de la Paix, BP2300, 1211, Geneva 2,
Switzerland. better computational capabilities of modern supercomputers and
E-mail addresses: abaklanov@wmo.int, abaklanov@nbi.ku.dk (A. Baklanov). better understanding of physical and chemical processes in the
atmosphere. The World Health Organization [2] reported that
about 7 million people deaths (1/8 of total annual global deaths) are
anticipated due to exposure to outdoor and indoor air pollution (3.7
Production and Hosting by Elsevier on behalf of KeAi
and 3.3 million, respectively).

https://doi.org/10.1016/j.glt.2020.11.001
2589-7918/© 2020 The Authors. Production and hosting by Elsevier B.V. on behalf of KeAi Communications Co., Ltd. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).
A. Baklanov and Y. Zhang Global Transitions 2 (2020) 261e270

It is important to stress that key air pollutants, particularly dynamic, and chemical processes. We will not thus provide a his-
black carbon and methane, also have near-term climate impacts torical overview of existing modeling experience, but based on
that accelerate the pace of global warming. Most of the leading published reviews and recent analysis, will focus in this article on
sources of air pollution are among the most significant emitters of advances, main trends and research needs in further developments
longer-lived greenhouse gases, particularly carbon dioxide and of atmospheric composition and air quality modeling and
nitrous oxide. Therefore, there is a critical need to consider the forecasting.
problem in a complex manner with co-benefits for air quality and For example [15], reviewed existing experience and identified
climate as well as interactions of climate change and atmospheric several inaccuracies in AQF and their possible causes, including
pollution. different factors related to:
Additionally to annual average air quality assessments for
health impacts, it is also important to realize short-term fore- - meteorology, e.g., inaccurate characterization of the transport
casting of peak pollutants concentrations to prevent health ef- and planetary boundary layer (PBL) meteorological processes
fects from acute episodes. To protect citizens from unhealthy air, such as turbulent mechanisms and vertical convection, cloud
many countries have real-time air quality forecasting (AQF) attenuation of photolysis, local drainage and sea-breeze circu-
programs in place to forecast the concentrations of pollutants of lations and PBL heights;
special health concerns such as ozone (O3), nitrogen dioxide - boundary conditions (BCONs) including inadequate represen-
(NO2), particulate matter with diameters less than and equal to tations of BCONs of O3, PM2.5, and PM10);
2.5 mm (PM2.5, and PM10, respectively) e.g., Ref. [3e11]. Such in- - emissions including uncertainties in anthropogenic emissions of
formation has been used to issue early air quality alerts that sulfur dioxide (SO2) and volatile organic compounds (VOCs),
allow government and people to take precautionary measures wildfire emissions, particular matter (PM), dust, pollen and
such as temporarily banning major emission sources, favoring car other meteorology-dependent emissions;
pooling or taking public transportation to reduce air pollution - model process treatments, e.g., inaccurate model treatments
and minimize their exposures to unhealthy levels of air pollution such as urban processes, gas-phase chemistry, in-cloud oxida-
[12]. Air quality forecasting has been used to schedule and plan tion of SO2, secondary organic aerosol (SOA) formation, and dry
numerous field campaigns to effectively track pollutant plume and wet deposition.
transport and sample pollutant concentrations, which maximizes
the usage of expensive instrumented platforms such as airplanes As [24] stressed, to reduce short- and long-term exposure of
and other limited measurement resources e.g., Ref. [13,14]. Ac- public to air pollution, especially in developing countries, an
curate AQF can therefore offer tremendous societal and economic internationally coordinated and integrated air quality monitoring,
benefits by enabling advanced planning for individuals, organi- analysis and prediction system must be developed by effectively
zations, and communities in order to avoid exposure, and reduce combining scientific and technological progress in observing ca-
pollutant emissions and their adverse health impacts. Driven by pabilities (in situ and space), mathematical models, data analyses,
crucial regulations, societal and economic needs, scientific ad- and computer science.
vancements, and increasing availability of high performance There are the following trends in the development of modern
computing capacity, AQF has evolved from weather forecasting atmospheric composition modelling and AQF systems:
and developed into a new discipline that integrates science and
technology from several disciplines including meteorology, at- - Seamless prediction of the Earth system approach;
mospheric chemistry/air quality, mathematics, physics, environ- - Online coupling of atmospheric dynamics and chemistry
mental statistics, and computer sciences/engineering. models;
Two main approaches can be distinguished in AQF: empirical/ - Multi-scale prediction approach;
statistical methods and chemical transport modelling. The empir- - Subseasonal to seasonal forecast;
ical/statistical approaches have several common drawbacks for AQF - Emission modeling for improved emission data;
(see review in Refs. [15e17]; and [18]. For example, they cannot - Multi-platform observations and data assimilation;
predict concentrations during periods of unusual emissions (e.g., - Data fusion, machine learning methods and bias correction
substantial emission reductions that have happened during the techniques;
2020 lockdown period due to the COVID-19 pandemic) and/or - Ensemble approach;
meteorological conditions that deviate significantly from the his- - Fit for purpose approach;
torical record [19]. These statistical models provide neither the - Impact based forecast.
direct linkages between precursor emissions and resultant pollu-
tion nor the interrelationships among multiple pollutants (i.e., the Given the large scope and many focuses of this field, the paper
interactions among pollutants that may potentially exacerbate one focuses on selected topics. The advances in several important
pollution problem while alleviating another problem). Explicit domains, e.g., in meteorological processes and atmospheric
treatments for such linkages and interactions in AQF models are chemical mechanisms for air quality (AQ) modeling, as well as
essential to the enhancement of understanding of the physical- emission modeling for improved emission data are discussed in
chemical system, the improvement of short- and long-term AQF other recent publications: by Refs. [25e27] and by Ref. [28].
skill, and the development of integrated emission control strategies Therefore we will not touch these issues in this article and will
for multi-pollutants. Therefore chemical transport models (CTMs) concentrate on the following: (i) Seamless prediction of the Earth
are much more commonly used in AQF systems and in this article system, including online coupled chemistry-meteorology
we will consider mostly chemical transport modelling-based modelling, multi-scale prediction systems and seasonal fore-
approaches. casts; (ii) utilizing modern observational data in models,
There are several comprehensive review papers, e.g., including data assimilation and data fusion algorithms, machine
Refs. [7,15,17,20e22,23]; and [18] analyzing current major 3-D learning methods, bias correction techniques and ensemble
global and regional real-time air quality forecasting models and methods; (iii) fit for purpose and impact based forecast
identifying areas of improvement in meteorological forecasts, approaches.
chemical inputs, and model treatments of atmospheric physical,
262
A. Baklanov and Y. Zhang Global Transitions 2 (2020) 261e270

3. Toward seamless prediction of the earth system studying and modelling of physical-chemical-biological processes
at multi-scales and integrated assessment for environment, popu-
3.1. Seamless environmental prediction systems lation, ecosystems, etc. In particular, a hierarchy framework of such
modern multi-scale models for different elements of the Earth’s
The new generation of integrated atmospheric dynamics and system, integrated with the observation system, is being developed
composition models is based on the seamless Earth System within the Pan-Eurasian Experiment (PEEX) multi-disciplinary
Modelling (ESM) approach [29] to evolve from separate model programme [30,31]. Moreover, the models will inform the devel-
components to seamless meteorology-composition-environment opment of the in-situ monitoring component of the PEEX Research
models to address challenges in weather, climate, and atmo- Infrastructure by providing information on regions, where specific
spheric composition fields whose interests, applications, and processes or interactions may be important to measure with new
challenges are now overlapping. “Seamless” is considered in rela- observational capability.
tion to, at least, two aspects. Firstly, at the process-scale, it refers to,
for example, the coupling within a model of meteorology and
composition processes to represent the two-way interactions be- 3.2. Coupled chemistry-meteorology modeling (CCMM)
tween composition and radiative processes or microphysics, or the
consistent treatment of water vapor. Secondly, to be considered in The use of a coupled meteorology-chemistry model, e.g., WRF-
terms of time-space-scales, it refers to the absence of discontinu- Chem [32] or Enviro-HIRLAM [33], for AQF represents a signifi-
ities in model behavior when used at multiple temporal or spatial cant advancement in routine operational AQFs and would greatly
resolutions to have, for example, consistent treatment of black enhance understanding of the underlying complex interplay of
carbon for air quality and climate applications or consistent meteorology, emission, and chemistry. Model evaluation demon-
coupling interval between land, ocean and atmosphere. strates that a modeling approach based on CTMs has skills consis-
In a more general sense, the approach considers several di- tent with or better than many statistical forecasting tools [34,35].
mensions of the seamless coupling, including: Online-coupled meteorology atmospheric chemistry models
(CCMM) have greatly evolved in recent decades [36e40]. Although

Time scales: from seconds and nowcasting to decadal and mainly developed by the air quality modeling community, these
centennial (climate) time-scale; integrated models are also of interest for numerical weather pre-

Spatial scales: from street-level to global scale (downscaling and diction and climate modeling as they can consider both the effects
upscaling); of meteorology on air quality, and the potentially important effects

Processes: physical, chemical, biological, social; of atmospheric composition on weather. Migration from offline to

Earth system elements/environments/components: atmo- online integrated modeling and seamless environmental prediction
sphere, hydrosphere, lithosphere/pedosphere, ecosystems/ systems [41] are recommended for consistent treatment of pro-
biosphere; cesses and allowance of two-way interactions of physical and

Different types of observations and modelling as tools: chemical components, particularly for AQ and numerical weather
observations-model fusion, data processing and assimilation, prediction (NWP) communities.
validation and verification; Regarding AQF and atmospheric composition modelling, the

Links with health and social consequences, impact, assessment, CCMM approach will certainly improve forecast capabilities as it
and services and end-users. allows a correct way of jointly and consistently describing meteo-
rological and chemical processes within the same model time steps
Different aspects of methodology and research needs for real- and grid cells. Applications that may benefit from CCMM are
isation of the Seamless Prediction Systems, as a part of WMO numerous and include [42]: chemical weather forecasting (CWF),
research strategy, are presented and discussed in the book numerical weather prediction for precipitation, visibility, thun-
“Seamless Prediction of the Earth System: From Minutes to derstorms, etc., integrated urban meteorology, environment and
Months” [29]. climate services, sand and dust storm modeling and warning sys-
The ensemble approach integrates modeling results from tems, wildfire atmospheric pollution and effects, volcano ash
different models, participants and countries, or at one center with forecasting, warning and effects, high impact weather and disaster
one model, e.g., of the European Center for Medium-Range risk, effects of short-lived climate forcers, earth system modeling
Weather Forecasts (ECMWF), but with perturbations to initial and projections, data assimilation for CWF and NWP, and weather
conditions and physical parameterizations (stochastic physics). modification and geo-engineering. Online integrated models,
This allows information on model process diversity to inform un- however, need harmonized formulations of all processes influ-
certainty in our understanding of Earth system linkages, responses, encing meteorology and chemistry.
and impact assessments. It is possible to utilize the full potential of Based on the EuMetChem activities [43], CCMM symposium [41]
a hierarchy of models: scenario analysis, forward and inverse and further discussions (e.g. Ref. [29,42], the following recom-
modeling, modeling based on measurement needs and processes. mendations for future research have been identified.
The models are evaluated and constrained by available in-situ and For air quality forecasting and atmospheric composition studies
remote sensing data of various spatial and temporal scales using the following research needs are stressed:
data assimilation and top-down modeling. The analyses of the
anticipated large volumes of data produced by available models and
Better representation of aerosol processes and their formation
sensors will be supported by a dedicated virtual research platforms (especially in terms of chemical speciation), heterogeneous
developed for these purposes. chemistry and interactions with cloud;
Linking between the temporal and spatial scales for modelling
Experiments that are specifically defined to look at chemistry-
and observations is a key aspect of the seamless prediction cloud-microphysics at different scales;
approach. Both short- and long-term measurements from different
More field experimental data to evaluate online coupled
observational platforms including field campaigns, laboratory, models;
chamber experiments, satellite, etc. represent valuable contribu-
Improved numerical and computational efficiency of the models
tions to evaluation and verification of different models as well as for as the complexity of applications grows (e.g., scales);
263
A. Baklanov and Y. Zhang Global Transitions 2 (2020) 261e270

Intercomparisons both at global and regional/urban scale for AQ, addition transmitted across the boundaries to the coarser
NWP, and climate should continue; intercomparisons that are resolution.
cutting across all 3 fields should be considered. For example, the European Copernicus Atmosphere Monitoring
Service (CAMS) provides such AQF from global to regional [46,47]
For meteorological studies and specifically NWP the following and possibly urban scales (FUMAPEX, MEGAPOLI) for European
research is needed: countries [6,48e50]; GURME [51,52], MarcoPolo & PANDA projects
[53] developed and tested similar AQF systems for Asian countries

Developing diagnostics and validation methodologies to more (China, India) and selected cities (e.g., Shanghai, New Delhi, etc.).
explicitly separate different effects of the intertwined feedback The development of models dedicated to address urban AQ is-
processes; sues requires also some urbanization of meteorological/NWP

More collaboration between operational centers and research models [29] and a good set of data regarding urban structure and
communities. This needs to be focused on providing schemes many other characteristics. There are important initiatives in that
with an impact that is proven to be valuable enough to justify direction providing a more detailed atmospheric dataset, e.g.,
the cost of their implementation (even for relatively modest pollutant fluxes and concentrations, temperature, wind, pressure,
increases in CPU); and moisture, and also providing important parameters to describe

More evaluation of aerosol properties routinely, not only for the urban features in an aggregate manner that can be introduced into
indicators PM10 and PM2.5 but also for optical, chemical and urban scale models with resolutions of 200e2000 m and corre-
microphysical properties; sponding physical options in single and multi-layer urban canopy

The treatment of the indirect effect of aerosols is one of the key models (Ching et al., 2018).
uncertainties; ice nucleation processes and parameterizations To forecast street level air quality the downscaling with specific
are less well defined than CCN formation processes and (e.g., obstacle-resolved computational fluid dynamics (CFD) type or
parameterizations; parameterized) microscale models is needed. CFD models coupled

Further research is needed to better understand the importance with mesoscale models can provide this high spatial resolution (up
of including more accurate representation of aerosol properties to 1e10 m) because they explicitly resolve the turbulent flow
in satellite retrievals; around buildings (see e.g., Refs. [49,54,55]. However, the required

Research on the impact of online modeling of aerosols on visi- computational time increases and it would be difficult, at a
bility forecasting, observational constraints on the causes of reasonable CPU time, to use a direct coupling between mesoscale
light extinction and on parameterizations for calculating and CFD models for forecasting purposes. In the near future, CFD
extinction given model parameters. models could become an appropriate tool for forecasting due to the
increase of computational resources. Nowadays, a database created
For climate research the following main developments in with CFD simulation of several scenarios represents the best
CCMMs are needed: compromise to perform the downscaling of the mesoscale outputs.
Another alternative is the use of a parametrized microscale model

Improve our understanding of indirect effects (e.g., BC on (e.g., Ref. [56,57]. These models do not resolve explicitly the tur-
clouds); bulent flow but use parametrized relations between urban

Develop CCCMs with prognostic aerosols to assess the tradeoff configuration and flow and dispersion variables within the urban
between a more complex aerosol representation on the one side canopy. This fact allows a simplified representation of concentra-
and model resolution, or the atmosphere-ocean coupling, on the tions within the urban canopy in reasonable computational time
other side; for forecasting purposes.

Test model performance in terms of relevant physical, chemical, Seamless unified modelling system that allows a single platform
and radiative processes and mechanisms (in contrast to just to operate over the full scale (i.e., across-scale) will represent a
testing mean performance); substantial advancement in both the science and the computa-

Test model performance in terms of tropospheric dynamics/ tional efficiency. The Model for Prediction Across Scales for Atmo-
meteorology and their effect on composition (and vice-versa). sphere (MPAS-A) being developed by the U.S. National Center for
Atmospheric Research is a good example of such a modelling sys-
tem (mpas-dev.github.io; [58]. MPAS-A uses the unstructured
3.3. Multi-scale prediction approach Voronoi meshes (hexagons) and C-grid discretization. The former
allows for smoothly-varying mesh transitions and local refinement.
The AQ monitoring, analysis and forecasting systems should The latter solves for normal velocities on cell edges and is well-
operate at different spatial scales from the global scale to the suited for higher-resolution, mesoscale atmosphere and ocean
regional, national, urban and sub-urban scales. In previous decades, simulations. Two other examples of recently developed
AQ models usually were applied only for a specific spatial scale: community-based, coupled, comprehensive multi-scale Earth
global, regional, meso-, local, urban or street scales. During last modeling systems are the MUlti-Scale Infrastructure for Chemistry
decades research achievements, high-performance computational and Aerosols e MUSICA [59] and the Unified Forecast System [60]
resources, and real time data access possibilities make it realistic to to support the Weather Enterprise and to be the source system for
build a full chain of multi-scale AQ modelling and forecasting. NOAA’s operational numerical weather prediction applications
Zooming or special nesting grid techniques are usually required including the atmospheric composition.
to transfer model information between the scales. Grid nesting is Major challenges include globalization/downscaling with
the most common method, being employed in most CTMs. Many of consistent model physics and two-way nesting with mass conser-
the currently used CTMs (e.g., WRF-Chem, CAMx [44], CMAQ [45]) vation and consistency. Such a unified global-to-urban scale
allow for grid nesting. Grid nesting is further classified in one-way modelling systems will provide a new scientific capability for
nesting (sometimes referred to as downscaling), when values of the studying important problems that require a consideration of multi-
modelled variables at a coarse resolution are used as boundary scale feedbacks.
conditions for finer (subscale) resolution runs, and two-way nest-
ing, when information from the higher resolution scale is in
264
A. Baklanov and Y. Zhang Global Transitions 2 (2020) 261e270

3.4. Subseasonal to seasonal forecast filter (EnKF) or hybrid techniques combining the advantages of
both variational and EnKF techniques are all applicable in CTMs
Significant progress has been made in recent decades on [76]. Other methodologies such as inverse modelling of emission
medium-range weather forecasts and seasonal climate predictions. fields appear as a promising technique to improve the skill of AQF
Because of a difficult time range that is not as well defined as and may have a stronger impact for short-lived pollutants than
weather and seasonal forecasting, subseasonal forecasting has not chemical data assimilation (CDA) has on initial conditions. CDA has
received as much attention as weather or seasonal forecasting. been implemented in many operational AQF and some CCMM
Advancements in real-time weather and air quality forecasting models, such as the ECMWF/Copernicus Integrated Forecast System
models and the data analysis techniques have made reasonably (IFS-CTM) [47], the Weather Research and Forecasting Model
accurate predictions and data analytics for extended timeframes coupled with chemistry (WRF-Chem) [77], and the Goddard Earth
possible. Seasonal and subseasonal forecasts are generated at Observing System Model, version 5 (GEOS-5) [78]. CDA is used to
monthly intervals out to 7 months and weekly intervals out to 5 initialize air quality forecasts and for retrospective analysis of at-
weeks, respectively. The subseasonal to seasonal forecast will mospheric composition. It has been shown that corrections to
bridge the gap between weather and climate and bring the weather emissions as part of the CDA procedure can help to improve the
and climate communities together to tackle the intervening time impact of the observations on the predictions.
range, harnessing shared and complementary experience and A main challenge of CDA is the limited information content of
expertise in forecasting, research and applications, toward more the atmospheric composition observations [41]. More routine
seamless weather/climate prediction systems and more integrated evaluation of aerosol properties, not only for the indicators PM2.5
weather and climate services [61,62]. Seasonal and subseasonal and PM10, but also for optical, chemical, and microphysical prop-
forecasts deliver easy-to-use, predictive analysis that anticipates erties, are needed. Spatial and temporal coverage of important
market behavior and potential risks for scientific communities, parameters such as aerosol composition is limited because of
government, and many sectors such as agriculture, aviation, media, network design, viewing geometry, or cloud cover masking [29].
energy, and insurance. For example, they can be used for market Also, vertically integrated observations, such as aerosol optical
shifts that may affect profitability in the 3- to 5-week, 1- to 4- depth or total column, have to be distributed to model levels.
month, and 5- to 7-month timeframes. In particular seasonal Furthermore, biases between different observing systems (e.g.,
forecast for wildfires and smog pollution is underdeveloped within satellite data, lidars, ceilometers, ground observations) should be
the VFSP-WAS [63]. removed before the assimilation [23].
Subseasonal weather and air quality forecasts have been iden- Although the importance of assimilating atmospheric compo-
tified as research and operation priorities by WMO [62,64] and sition data into CCMM to get realistic chemical fields has been
several countries such as the U.S [65,66]. Subseasonal forecasts will demonstrated, there is little experience in assimilation of both
have large societal benefits, as many market behaviors and man- meteorological and chemical observations into CCMM. Further
agement decisions in agriculture and food security, water, risk research is needed to better understand the importance of
reduction of weather disaster and hazards (e.g., drought, wildfire, including more accurate representation of aerosol properties in
heatwave, cold wave) and health fall into the subseasonal to sea- satellite retrievals.
sonal time range.
4.2. Measurement-model fusion, machine learning and bias
4. Improvements by utilizing modern observation data and correction techniques
ensembles
Tremendously growing number and different types of obser-
One of the key modern trends to improve AQF systems is vations became available during the last decade, require and give a
developing new methods of utilizing modern observational data in strong impulse for development of new methods for measurement-
models, including data assimilation and data fusion algorithms, model fusion to improve AQF [69]. We have considered in previous
machine learning methods and bias correction techniques. sections first of all most applicable 3D numerical atmospheric
transport models and classical data assimilation methods, however
4.1. Multi-platform observations and data assimilation other types of data fusion algorithms, such as the statistical
methods, optimal interpolation, objective analysis, bias correction,
During last decades the techniques and possibilities of multi- as well as relatively new artificial intelligence, neural network,
platform (in-situ, ground, aircraft and satellite remote sensing) machine learning and hybrid methods, were also actively devel-
observations of air pollution and atmospheric parameters, as well oped during last decades. Several recent publications considered
as their near real time (NRT) or real time (RT) availabilities and such methods and analysed their applicability (e.g., Ref. [16,18,69].
citizen science opportunities [67], are dramatically improved and Statistical methods are simple, but require a large amount of his-
increased. This leads to a revision of concepts of AQF and impor- torical data and highly depend on them. Artificial intelligence,
tance of data assimilation (both chemical and meteorological neural network, and machine learning methods have better per-
measurements) in AQF systems [23,68] and measurement-model formance, but can be unstable and also depend on data. Hybrid or
fusion for air quality and deposition assessments [69]. combined methods have a better quality. Such methods can also
In particular, the new generation of geostationary satellites, improve AQF utilizing additional observational data. For example
consisting of GEMS (Asia), Sentinel-4 (Europe) and TEMPO (USA), [79], applied machine learning methods for ozone ensemble fore-
expects to provide an unprecedented view of air quality from space cast, they performed sequential aggregation based on ensemble
[70e72]. Given these planned missions, it is even more imperative simulations and past observations. Further development of ma-
to develop effective data assimilation systems for AQF that opti- chine learning methods provides a potential way to bridge the
mally incorporate satellite information. resolution gap between global, low-resolution model output and
Correspondingly, experience in the assimilation of chemical and local, high-resolution information requested by end users. How-
aerosol observations from ground-based and satellite instruments ever, further research is needed to identify applications where
into AQF and CCMM has grown significantly [23,29,73e75]. machine learning may have particular advantages.
Optimal interpolation, variational approaches, Ensemble Kalman Another broadly used approach to improve the AQF model
265
A. Baklanov and Y. Zhang Global Transitions 2 (2020) 261e270

performance is bias correction techniques. They have been proven the-art numerical air quality models developed in Europe. Under
to be effective in improving the forecasting accuracy and applied to the MarcoPolo-PANDA EU FP7 Projects (http://www.marcopolo-
the routine O3 and PM2.5 forecasting in many forecasting models panda.eu/), nine different models were used for multi-model
such as the U.S. NOAA’s operational forecasting model, the National ensemble over China [53,97]. The most recent International Coop-
Air Quality Forecast Capability (NAQFC) [80e84]. Several bias erative for Aerosol Prediction (ICAP) multi model ensemble (ICAP
correction methods have been implemented in air quality fore- MME) consists of nine global models [98]. Multi-model ensemble
casting models, e.g., for wildfire plumes. The simplest and also most results provide a range and an indication of the robustness of the
commonly-used method is the mean subtraction method, in which forecasts and help to improve the accuracy of chemical weather and
the mean bias is subtracted from the forecasted values at each air quality forecasting.
monitoring site. This method, however, does not always guarantee The ensemble forecast is usually a weighted linear combination
a positive value. The second approach is the multiplicative ratio of the individual ensemble members, in which the weight factors
adjustment method, in which the forecasted value is multiplied by may be equal or unequal for individual members. As such, it is
the mean ratio of the sum of the observed value to the sum of the subject to the inherent limitations of individual ensemble members
forecasted value at each monitoring site. This alternative correction and is sensitive to the selected weighting factors. More advanced
guarantees that the concentrations will remain positive and in- ensemble forecast approaches have been developed. One such
creases forecasting skill, but the ratio-adjustment method provides method is the ensemble forecast of analyses (EFA) of [79]. EFA
additional improvement over the mean subtraction method for couples an ensemble forecasting approach (i.e., sequential aggre-
models with the highest biases. The third method is to force the gation) with CDA techniques to forecast an analysis from data
zero differences between observed and simulated seasonal means assimilation, instead of observations.
by using an empirical linear fit between forecasted and bias cor-
rected values. More advanced bias correction methods have also 5. Fit for purpose approach and impact based forecast
been developed and applied for improvement of AQF, e.g., the
Kalman Filter (KF) bias correction [85,80e82,86,87]. KF has been The Earth system modelling approach provides a broad AQF
combined with the Kolmogorov-Zurbenko (KZ) filter to correct platform for different applications, but the modelling system
biases in air quality forecasting [80]. KZ filter can be used to sepa- should fit for purposes and be optimized for specific tasks. For
rate observed and simulated time series data into the short-term example, for smog pollution from wildfires the problem is more
and baseline components to discern major systematic model er- complex than for AQF with fixed anthropogenic emission sources
rors [11,88e90]. The former is influenced by prevailing weather and includes first of all the forecasting of fire danger and sources of
conditions and the latter is influenced mainly by emissions, wildfire emission, including detection of sources, mass of emitted
boundary conditions, and other slow-varying processes. The KF bias species, plume rise, fire behaviour, etc. Research challenges facing
correction technique can be applied to the baseline component of fire and smoke prediction were evaluated by the Interdisciplinary
the time series at a monitoring location, which can significantly Biomass Burning Initiative (IBBI) [23,63,99].
improve the forecasting accuracies. The KF or the coupled KF-KZ As mentioned, there are also several user communities (e.g.,
bias correction method can be potentially extended to locations NWP, climate) and specialized applications of AQF system de-
where no monitor information is available for these methods to velopments for long-term prediction and specific episodes of at-
generate bias-corrected spatial maps of forecast products. mospheric harmful contamination, affecting not only health but
The performance of bias-corrected forecast depends on the many other sectors of economics. In particular, sand and dust
performance of the raw model forecast to which the bias-correction storms pose a major challenge to sustainable development in arid
method is applied. While bias correction methods have shown and semi-arid regions of the planet. Airborne dust presents serious
obvious short-term benefits, the improvements of model inputs risks for human health [100]. Sand and dust storms are also detri-
such as emissions and boundary conditions and model represen- mental for ecosystems and diverse socio-economic sectors. Surface
tations of key chemical and physical processes such as secondary dust deposits are a source of micro-nutrients for both continental
organic aerosol formation are necessary to address the mechanistic and maritime ecosystems. Although dust can work as fertilizer, it
deficiencies of the AQF models for long-term benefits. has many negative impacts on agriculture, including reducing crop
yields by burying seedlings, causing loss of plant tissue, reducing
4.3. Ensemble approach photosynthetic activity, and increasing soil erosion. Reductions in
visibility due to airborne dust also have an impact on air and land
Ensemble forecasting is a numerical prediction method that is transport. Poor visibility conditions are a hazard during aircraft
used to produce a representative sample of the possible future landing and taking off operations e flights landings may be diver-
states of a model system. Ensemble forecasting can be imple- ted and departures may be delayed. Dust can also scour aircraft
mented using multiple models or one model but with different surfaces and damage engines. Reduced radiation at the surface has
inputs (e.g., varying meteorological input forcings, emission sce- an impact on the output from solar power plants, especially those
narios, chemical initial conditions) or different process parameters that rely on direct solar radiation. Dust deposits on solar panels are
(e.g., varying chemical reaction rates) or different model configu- a main concern of plants operators. Volcanic eruptions contribute
rations (e.g., varying grid spacings) or different models. Ensemble to natural aerosols directly via emissions of ash and other partic-
forecasting has shown significant statistical improvements for both ulates, or indirectly through the release of sulfur gases that sub-
O3 (e.g. Ref. [86,91e93], and PM2.5 forecasts over any individual sequently condense to form sulfate aerosols. Volcanic aerosols are
forecast [94]. Multi-model ensemble air quality forecasting has particularly important for aviation safety and climate modelling, as
been emerging for AQF on global scale and regional scales (e.g., over was witness by the eruption of the Icelandic Volcano Eyjafjal-
Europe, U.S. and China) (e.g., Ref. [86,87,91e96]; Solazzo et al., € kull, that paralyzed for several days air traffic throughout
lajo


2012; Monteiro et al., 2013; Zabkar et al., 2013; Im et al., 2015; Europe.
[53,97]. For example, Delle Monache et al. (2004) performed one of Society is impacted by both long-term and short-term changes
the first multi-model real-time O3 forecasts over the U.S. The to atmospheric composition. Volcanic ash, desert dust, natural and
regional air quality production of the Copernicus Atmosphere anthropogenic gas emissions are good examples demonstrating
Monitoring Service (CAMS) is based on ensemble of seven state-of- both immediate impacts, for example, on aviation and human
266
A. Baklanov and Y. Zhang Global Transitions 2 (2020) 261e270

health e and longer-term, including climate forcing and impacts e Better representation of aerosol processes and feedbacks, their
from the changing atmospheric composition. While both present a interactions with clouds and radiation,
range of challenges, the former places considerable additional de- e Improved data assimilation and fusion, machine learning, and
mands in terms of data timeliness and temporal and spatial reso- artificial intelligent methods
lutions. This near-real-time need for observations is indeed a e Toward seamless Earth system modelling,
common requirement across a range of impacts, but one that is not e Multi-scale prediction approach,
always conducive to the significant processing involved in pro- - Subseasonal to seasonal forecasting,
ducing fully assured atmospheric composition data. However, - Impact based forecast and fit for purposes systems.
timely data can be produced and has the potential to be of
considerable use for a wide range of applications. In recognition of Air quality forecast and assessment systems help decision
this, the WMO Global Atmosphere Watch Programme [101] has makers to improve air quality and public health, mitigate the
identified the need for increased support for the development and occurrence of acute air pollution episodes, particularly in urban
expanded use of services and research activities concerning the areas, and reduce the associated impacts on agriculture, transport,
forecasting of atmospheric composition and its induced environ- economy, ecosystems and climate.
mental phenomena. Seamless modelling is a prospective way for future single-
So, the AQF should not be limited only to the concentrations of atmosphere modelling systems with advantages for applications
air pollutants but should also consider their impacts on different at all space- and timescales for multi applications not only for air
sectors (e.g., health, agriculture, land transport, aviation, energy, quality, but also for NWP, climate and other atmospheric compo-
etc.) and provide warning and recommendations proceeding from sition models.
different thresholds, risks, and cost functions (Fig. 1). Such impact- The advance approach combines an ensemble of state-of-the-art
based forecast and assessment systems will help stakeholders and models, high-resolution emission inventories, space observations,
responsible agencies to improve air quality and public health, and surface measurements of most relevant chemical species
mitigate the occurrence of acute harmful air pollution episodes. The (coarse and fine PM, ozone, reduced and oxidized nitrogen, etc.) to
suggested approach combines state-of-the-art models, high- provide hindcasts, analyses and forecasts of from global to regional
resolution emission inventories, space observations and surface air pollution and downscaling for selected countries and urban
measurements of most relevant chemical species to provide hind- areas.
casts, analyses, and forecasts of regional air pollution in a specific Dramatically increasing possibilities of multi-platform (in-situ,
region (using boundary and initial conditions from global AQF ground, aircraft and satellite remote sensing) observations in real or
system such as the CAMS) and downscaling for selected countries near-real time, as well as citizen science opportunities, lead to a
and urban areas. revision of concepts of AQF, increase importance of data assimila-
The still unsolved challenge to make the produced AQF infor- tion techniques. New measurement-model data fusion, machine
mation available to end users in a simple way fitting for purpose. So, learning and artificial intelligence methods have a good potential to
such a common platform for air quality forecasts would increase bridge the resolution gap between global, low-resolution model
the usability of these systems and make it much easier to integrate output and local, high-resolution information required by users.
them into decision-making processes. Improved data assimilation, both in terms of techniques and
choice of aerosol variables to be assimilated. Key questions for the
future are whether there is a benefit to move from assimilating AOD
to assimilating clear-sky radiances in the shortwave spectrum and
6. Summary and perspectives how to make the best possible use of vertical profiles from lidar
observations.
Further improvements to AQF and modeling systems will likely Although we did not focus deeply on emission modelling
follow several directions: methods in this paper, it is important to highlight the importance of
emission data and models as one of the key uncertainties in AQF. In
e Online coupling of atmospheric dynamics and chemistry particular, dramatic changes of emissions from COVID-19
models,

Fig. 1. Example of impact-based forecast and assessment systems for the WMO Sand and Dust Storm Warning Advisory and Assessment System (SDS-WAS) (after [102].

267
A. Baklanov and Y. Zhang Global Transitions 2 (2020) 261e270

lockdown-related reduced activities, as well challenges with modeling to support air quality policies, in: C. Mensink, G. Kallos (Eds.), Air
Pollution Modeling and its Application XXV. ITM 2016. Springer Proceedings
meteorology-dependent emissions (e.g., wildfires, dust storms,
in Complexity, Springer, Cham, 2018, pp. 25e29, https://doi.org/10.1007/
pollen, VOC, secondary aerosols and climate forcers), required 978-3-319-57645-9_4.
further research efforts in online monitoring and assessments of [12] R.A. Wayland, J.E. White, P.G. Dickerson, T.S. Dye, Communicating real-time
emissions using remote sensing observations, crowdsourcing data, and forecasted air quality to the public, Environ. Manag. 28e36 (2002)
(December).
modern inverse models, and machine learning and artificial intel- [13] A.M. Lee, G.D. Carver, M.P. Chipperfield, J.A. Pyle, Three dimensional chemical
ligent methods. forecasting: a methodology, J. Geophys. Res. 102 (1997) 3905e3919.
[14] F. Flatøy, O. Hov, H. Schlager, Chemical forecasts used for measurement flight
planning during POLINAT 2, Geophys. Res. Lett. 27 (2000) 951e954.
CRediT author statement [15] Y. Zhang, C. Seigneur, M. Bocquet, V. Mallet, A. Baklanov, Real-time air
quality forecasting, Part I: history, techniques, and current status, Atmos.
The corresponding author is responsible for ensuring that the Environ. 60 (2012a) 632e655, https://doi.org/10.1016/
j.atmosenv.2012.06.031.
descriptions are accurate and agreed by both authors. [16] Y. Zhang, A. Baklanov (Eds.), Best Practices and Training Materials for
The role of each of the two authors are almost equal: review of Chemical Weather/Air Quality Forecasting, The First Edition ETR-26, World
the problem, analysis and writing the manuscript. Meteorological Organization (WMO), Geneva, Switzerland, 2020, p. 565.
[17] W.F. Ryan, The air quality forecast rote: recent changes and future chal-
Alexander Baklanov and Yang Zhang. lenges, J. Air Waste Manag. Assoc. 66 (6) (2016) 576e596, https://doi.org/
10.1080/10962247.2016.115146.
Acknowledgements [18] L. Bai, J.-Z. Wang, X.-J. Ma, H.-Y. Lu, Air pollution forecasts: an overview, Int. J.
Environ. Res. Publ. Health 15 (2018) 780, https://doi.org/10.3390/
ijerph15040780.
YZ acknowledges support from the Agreement No. RD835871 [19] W.R. Stockwell, R.S. Artz, J. Meagher, R.A. Petersen, K.L. Schere, G.A. Grell,
awarded by the U.S. Environmental Protection Agency to Yale S.E. Peckham, A.F. Stein, R.V. Pierce, J.M. O’Sullivan, P.Y. Whung, The scientific
basis of NOAA’s air quality forecast program, Environ. Manager, December 8
University (with a subaward to Northeastern University) and the (2002) 20e27.
grant #NA20OAR4590259 awarded by NOAA Office of Weather and [20] L. Menut, B. Bessagnet, Atmospheric composition forecasting in Europe, Ann.
Air Quality to Northeastern University. It has not been formally Geophys. 28 (2010) 61e74, 2010.
[21] Y. Zhang, C. Seigneur, M. Bocquet, V. Mallet, A. Baklanov, Real-time air
reviewed by the U.S. EPA. The views expressed in this document are quality forecasting, Part II: state of the science, current research needs, and
solely those of the authors and do not necessarily reflect the views future prospects, Atmos. Environ. 60 (2012b) 656e676, https://doi.org/
and policies of the U.S. Environmental Protection Agency. EPA does 10.1016/j.atmosenv.2012.02.041.
[22] A. Baklanov, K.H. Schlünzen, P. Suppan, J. Baldasano, D. Brunner,
not endorse any products or commercial services mentioned in this S. Aksoyoglu, G. Carmichael, J. Douros, J. Flemming, R. Forkel, S. Galmarini,
publication. The authors thank their colleagues and two anony- M. Gauss, G. Grell, M. Hirtl, S. Joffre, O. Jorba, E. Kaas, M. Kaasik, G. Kallos,
mous reviewers for their very useful comments and suggestions. X. Kong, U. Korsholm, A. Kurganskiy, J. Kushta, U. Lohmann, A. Mahura,
A. Manders-Groot, A. Maurizi, N. Moussiopoulos, S.T. Rao, N. Savage,
C. Seigneur, R.S. Sokhi, E. Solazzo, S. Solomos, B. Sørensen, G. Tsegas,
References E. Vignati, B. Vogel, Y. Zhang, Online coupled regional meteorology chemistry
models in Europe: current status and prospects, Atmos. Chem. Phys. 14
[1] UN SDGs, Sustainable Development Goals. United Nations, 2015, 2015, web- (2014) 317e398, https://doi.org/10.5194/acp-14-317-2014.
site: https://sustainabledevelopment.un.org/. [23] A. Benedetti, J.S. Reid, P. Knippertz, J.H. Marsham, F. Di Giuseppe, S. Re my,
[2] WHO, Burden of Disease from the Joint Effects of Household and Ambient Air S. Basart, O. Boucher, I.M. Brooks, L. Menut, L. Mona, P. Laj, G. Pappalardo,
Pollution for 2012, World Health Organization, 2014, 25, March 2014, http:// A. Wiedensohler, A. Baklanov, M. Brooks, P.R. Colarco, E. Cuevas, A. da Silva,
www.who.int/phe/health_topics/outdoorair/databases/FINAL_HAP_AAP_ J. Escribano, J. Flemming, N. Huneeus, O. Jorba, S. Kazadzis, S. Kinne, T. Popp,
BoD_24March2014.pdf?ua¼1. P.K. Quinn, T.T. Sekiyama, T. Tanaka, E. Terradellas, Status and future of
[3] P. Manins, Air Quality Forecasting for Australia’s Major Cities: 1st Progress numerical atmospheric aerosol prediction with a focus on data re-
Report, CSIRO Atmospheric Research: Aspendale, Australia, 1999. SB/1/407 quirements, Atmos. Chem. Phys. 18 (2018) 10615e10643, https://doi.org/
25, http://www.dar.csiro.au/publications/Manins_1999a.pdf. 10.5194/acp-18-10615-2018.
[4] U.S. EPA, Guideline for Developing an Ozone Forecasting Program, U.S. [24] R. Kumar, V.-H. Peuch, J.H. Crawford, G. Brasseur, Five steps to improve air-
Environmental Protection Agency Rep., 1999. EPA-454/R-99-009, 88 pp. quality forecasts, Nature, 6 September 561 (2018) 27e29, 2018.
(Available from Office of Air Quality Planning and Standards, EPA, Research [25] R.T. McNider, A. Pour-Biazar, Meteorological modeling relevant to mesoscale
Triangle Park, NC 27711.). and regional air quality applications: a review, J. Air Waste Manag. Assoc. 70
[5] J.A. Pudykiewicz, A.S. Koziol, The application of Eulerian models for air (1) (2020) 2e43, https://doi.org/10.1080/10962247.2019.1694602.
quality prediction and the evaluation of emission control strategies in Can- [26] M. Mircea, G. Calori, G. Pirovano, C.A. Belis, European Guide on Air Pollution
ada, Int. J. Environ. Pollut. 16 (2001) 425e438. Source Apportionment for Particulate Matter with Source-Oriented Models
[6] A. Baklanov, O. Ha €nninen, L.H. Slordal, J. Kukkonen, N. Bjergene, B. Fay, and Their Combined Use with Receptor Models, JRC technical report
S. Finardi, S.C. Hoe, M. Jantunen, A. Karppinen, A. Rasmussen, A. Skouloudis, JRC119067, 2020, p. 66, https://doi.org/10.2760/470628.
R.S. Sokhi, j.H. Sorensen, V. Ødegaard, Integrated systems for forecasting [27] W.R. Stockwell, E. Saunders, W.S. Goliff, R.M. Fitzgerald, A perspective on the
urban meteorology, air pollution and population exposure, Atmos. Chem. development of gas-phase chemical mechanisms for Eulerian air quality
Phys. 7 (2007) 855e874. www.atmos-chem-phys.net/7/855/2007/. models, J. Air Waste Manag. Assoc. 70 (1) (2020) 44e70, https://doi.org/
[7] J. Kukkonen, T. Olsson, D.M. Schultz, A. Baklanov, T. Klein, A.I. Miranda, 10.1080/10962247.2019.1694605.
A. Monteiro, M. Hirtl, V. Tarvainen, M. Boy, V.-H. Peuch, A. Poupkou, [28] V. Matthias, J.A. Arndt, A. Aulinger, J. Bieser, H.D. van der Gon, R. Kranenburg,
I. Kioutsioukis, S. Finardi, M. Sofiev, R. Sokhi, K.E.J. Lehtinen, K. Karatzas, J. Kuenen, D. Neumann, G. Pouliot, M. Quante, Modeling emissions for three-
R. San Jose , M. Astitha, G. Kallos, M. Schaap, E. Reimer, H. Jakobs, K. Eben, dimensional atmospheric chemistry transport models, J. Air Waste Manag.
A review of operational, regional-scale, chemical weather forecasting models Assoc. 68 (8) (2018) 763e800, https://doi.org/10.1080/
in Europe, Atmos. Chem. Phys. 12 (2012) 1e87, https://doi.org/10.5194/acp- 10962247.2018.1424057.
12-1-2012. [29] WWRP, Seamless Prediction of the Earth System: from Minutes to Months,
[8] M.-T. Chuang, Y. Zhang, D.-W. Kang, Application of WRF/Chem-MADRID for vol. 1156, WMO-No., 2015, ISBN 978-92-63-11156-2. Available from: http://
real-time air quality forecasting over the southeastern United States, Atmos. library.wmo.int/pmb_ged/wmo_1156_en.pdf.
Environ. 45 (34) (2011) 6241e6250, https://doi.org/10.1016/ [30] M. Kulmala, H.K. Lappalainen, T. Pet€ aj€
a, T. Kurten, V.-M. Kerminen,
j.atmosenv.2011.06.071. Y. Viisanen, P. Hari, S. Sorvari, J. Ba€ck, V. Bondur, N. Kasimov, V. Kotlyakov,
[9] Y. Zhang, C.-P. Hong, K. Yahya, Q. Li, Q. Zhang, K.-B. He, Comprehensive G. Matvienko, A. Baklanov, H.D. Guo, A. Ding, H.-C. Hansson, S. Zilitinkevich,
evaluation of multi-year real-time air quality forecasting using an online- Introduction: the Pan-Eurasian Experiment (PEEX) e multidisciplinary,
coupled meteorology-chemistry model over southeastern United States, multiscale and multicomponent research and capacity-building initiative,
Atmos, Environ. Times 138 (2016) 162e182, https://doi.org/10.1016/ Atmos. Chem. Phys. 15 (2015) 13085e13096, https://doi.org/10.5194/acp-
j.atmosenv.2016.05.006. 15-13085-2015.
[10] P. Lee, J. McQueen, I. Stajner, J. Huang, L. Pan, D. Tong, H. Kim, Y. Tang, [31] M. Kulmala, Build a global Earth observatory, Nature 553 (2018) 21e23,
S. Kondragunta, M. Ruminski, S. Lu, E. Rogers, R. Saylor, P. Shafran, H.- https://doi.org/10.1038/d41586-017-08967-y, 2018.
C. Huang, J. Gorline, S. Upadhayay, R. Artz, NAQFC developmental forecast [32] G.A. Grell, S.E. Peckham, R. Schmitz, et al., Fully coupled “online” chemistry
guidance for fine particulate matter (PM2.5), Weather Forecast. 32 (1) (2017) within the WRF model, Atmos. Environ. 39 (2005) 6957e6975.
343e360, https://doi.org/10.1175/WAF-D-15-0163.1. [33] A. Baklanov, U.S. Korsholm, R. Nuterman, A. Mahura, K.P. Nielsen, B.H. Sass,
[11] S.T. Rao, H. Luo, M. Astitha, C. Hogrefe, R. Mathur, N. Kumar, On regional A. Rasmussen, A. Zakey, E. Kaas, A. Kurganskiy, B. Sørensen, I. Gonza lez-

268
A. Baklanov and Y. Zhang Global Transitions 2 (2020) 261e270

Aparicio, The Enviro-HIRLAM online integrated meteorologyechemistry description and implementation of the MarcoPoloepanda prediction system,
modelling system: strategy, methodology, developments, and applications version 1, geosci, Model Dev 12 (1) (2019) 33e67, https://doi.org/10.5194/
(v. 7.2), Geosci. Model Dev. (GMD) 10 (2017b) 2971e2999, https://doi.org/ gmd-12-33-2019.
10.5194/gmd-10-2971-2017. [54] K.H. Schlünzen, D. Grawe, S.I. Bohnenstengel, I. Schlüter, R. Koppmann, Joint
[34] J.N. McHenry, W.F. Ryan, N.L. Seaman, C.J. Coats Jr., J. Pudykiewics, modelling of obstacle induced and mesoscale changes e current limits and
S. Arunachalam, J.M. Vukovich, A real-time Eulerian photochemical model challenges, Journal of Wind Engineering and Industrial AerodynamicsJ. Wind
forecast system: overview and initial ozone forecast performance in the Eng. Ind. Aerodyn. 99 (2011) 217e225, https://doi.org/10.1016/
Northeast U.S. corridor, Bull. Am. Meteorol. Soc. 85 (2004) 525e548. j.jweia.2011.01.009.
[35] A.M.M. Manders, M. Schaap, R. Hoogerbrugge, Testing the capability of the [55] J.L. Santiago, R. Borge, F. Martin, D. de la Paz, A. Martilli, J. Lumbreras,
chemistry transport model LOTOS-EUROS to forecast PM10 levels in The B. Sanchez, Evaluation of a CFD-based approach to estimate pollutant dis-
Netherlands, Atmos, Environ. Times 43 (2009) 4050e4059, https://doi.org/ tribution within a real urban canopy by means of passive samplers, Sci. Total
10.1016/j.atmosenv.2009.05.006. Environ. 576 (2017) 46e58.
[36] Y. Zhang, Online coupled meteorology and chemistry models: history, cur- [56] R. Berkowicz, OSPM - a parameterised street pollution model, Environ.
rent status, and outlook, Atmos. Chem. Phys. 8 (2008) 2895e2932, https:// Monit. Assess. 65 (2000) 323e331.
doi.org/10.5194/acp-8-2895-2008. [57] Y. Kim, Y. Wu, C. Seigneur, Y. Roustan, Multi-scale modelling of urban air
[37] A. Baklanov, Chemical weather forecasting: a new concept of integrated pollution: development and application of a Street-in-Grid model (v1.0) by
modelling, Adv. Sci. Res. 4 (2010) 23e27, 2010. coupling MUNICH (v1.0) and Polair3D (v1.8.1), Geosci. Model Dev. (GMD) 11
[38] G.A. Grell, A. Baklanov, Integrated modelling for forecasting weather and air (2018) 611e629.
quality: a call for fully coupled approaches, Atmos. Environ. 45 (2011) [58] D. Ahijevych, M. Duda, L. Fowler, S.-Y. Ha, J. Jang, J. Klemp, S.-H. Park,
6845e6851, https://doi.org/10.1016/j.atmosenv.2011.01.017. B. Skamarock, W. Wang, M. Wong, C. Zarzycki, MPAS-A: WRF Numerics and
[39] K. Alapaty, R. Mathur, J. Pleim, Ch Hogrefe, S.T. Rao, V. Ramaswamy, Physics with a Height Coordinate on a Centroidal Voronoi Mesh, the 18th
S. Galmarini, M. Schaap, R. Vautard, P. Makar, A. Baklanov, G. Kallos, B. Vogel, WRF Users’ Workshop, 2017. Boulder, Colorado, U.S.A., June 12 e 16.
R. Sokhi, New directions: understanding interactions of air quality and [59] G.G. Pfister, S.D. Eastham, A.F. Arellano, B. Aumont, K.C. Barsanti, M.C. Barth,
climate change at regional scales, Atmos. Environ. 49 (2012) 419e421. A. Conley, N.A. Davis, L.K. Emmons, J.D. Fast, et al., A MUlti-scale infra-
[40] J. Pleim, R. Mathur, S.T. Rao, J. Fast, A. Baklanov, Integrated meteorology and structure for chemistry and aerosols - MUSICA, Bull. Am. Meteorol. Soc. 101
chemistry modeling: evaluation and research needs, Bull. Am. Meteorol. Soc. (10) (2020) E1743eE1760, https://doi.org/10.1175/BAMS-D-19-0331.1.
(2014) ES81eES84, https://doi.org/10.1175/BAMS-D-13-00107.1. [60] UFS, The Unified Forecast System (UFS), NOAA-NCAR, 2020. http://ufs-dev.
[41] CCMM, Coupled Chemistry-Meteorology/Climate Modelling (CCMM): Status rap.ucar.edu/.
and Relevance for Numerical Weather Prediction, Atmospheric Pollution and [61] G. Brunet, M. Shapiro, B. Hoskins, M. Moncrieff, R. Dole, G.N. Kiladis,
Climate Research (Symposium Materials), WMO GAW Report #226, WMO, B. Kirtman, A. Lorenc, B. Mills, R.M.S. Polavarapu, D. Rogers, J. Schaake,
Geneva, Switzerland, 2016. March 2016. Available from: https://www.wmo. J. Shukla, Collaboration of the weather and climate communities to advance
int/pages/prog/arep/gaw/documents/Final_GAW_226_10_May.pdf. subseasonal-to-seasonal prediction, Bull. Am. Meteorol. Soc. 91 (2010)
[42] A. Baklanov, D. Brunner, G. Carmichael, J. Flemming, S. Freitas, M. Gauss, 1397e1406, https://doi.org/10.1175/2010BAMS3013.1.
Ø. Hov, R. Mathur, K.H. Schlünzen, C. Seigneur, B. Vogel, Key issues for [62] F. Vitart, A.W. Robertson, D.L.T. Anderson, Subseasonal to seasonal prediction
seamless integrated chemistryemeteorology modeling, Bull. Am. Meteorol. project: bridging the gap between weather and climate, World Meteorol.
Soc. 98 (2017) 2285e2292, https://doi.org/10.1175/BAMS-D-15-00166.1. Organ. Bull. 61 (2) (2012) 2012. https://public.wmo.int/en/resources/
[43] EuMetChem, Coupled chemistryemeteorology modelling: status and rele- bulletin/subseasonal-seasonal-prediction-project-bridging-gap-between-
vance for numerical weather prediction, air quality and climate commu- weather-and-climate.
nities, in: A. Baklanov, B. Vogel, S. Freitas (Eds.), Special Issue Jointly [63] WMO, Vegetation Fire and Smoke Pollution Warning and Advisory System
Organized between Atmospheric Chemistry and Physics and Geoscientific (VFSP-WAS): Concept Note and Expert Recommendations, World Meteoro-
Model Development, 2016, p. 370. http://www.atmos-chem-phys.net/ logical Organization, 2018. GAW Report No235, https://library.wmo.int/doc_
special_issue370.html. num.php?explnum_id¼4519.
[44] Ramboll Environ, User’s Guide - Comprehensive Air Quality Model with [64] F. Vitart, C. Ardilouze, A. Bonet, A. Brookshaw, M. Chen, C. Codorean,
Extensions. Version 6.40, Ramboll Environ, Novato, California, 2016, p. 283. M. De que, L. Ferranti, E. Fucile, M. Fuentes, et al., The subseasonal to seasonal
[45] R. Mathur, J. Xing, R. Gilliam, G. Sarwar, C. Hogrefe, J. Pleim, G. Pouliot, (S2S) prediction project database, Bull. Am. Meteorol. Soc. 98 (2017)
S. Roselle, T.L. Spero, D.C. Wong, J. Young, Extending the Community Mul- 163e173, https://doi.org/10.1175/BAMS-D-16-0017.1.
tiscale Air Quality (CMAQ) modeling system to hemispheric scales: overview [65] J. Cortinas, NOAA/OAR Contributions to Subseasonal to Seasonal Research
of process considerations and initial applications, Atmos. Chem. Phys. 17 that Will Benefit National Capabilities, Office of Weather & Air Quality NOAA
(2017) 12449e12474, https://doi.org/10.5194/acp-17-12449-2017. Office of Oceanic and Atmospheric Research (OAR), 2017. January 25, ftp://
[46] A. Hollingsworth, R.J. Engelen, C. Textor, A. Benedetti, O. Boucher, ftp.oar.noaa.gov/ESPC/Program%20Briefs/AMS%202017/Cortinas%20AMS%
F. Chevallier, A. Dethof, H. Elbern, H. Eskes, J. Flemming, C. Granier, 20presentation%202_S2S.pdf.
J.W. Kaiser, J.-J. Morcrette, P. Rayner, V.-H. Peuch, L. Rouil, M.G. Schultz, [66] NASEM National Academies of Sciences, Engineering, and Medicine, Thriving
A.J. Simmons, Toward a monitoring and forecasting system for atmospheric on Our Changing Planet: A Decadal Strategy for Earth Observation from
composition: the GEMS project, Bull. Am. Meteorol. Soc. 89 (2008) Space. Chapter: 7 Weather and Air Quality: Minutes to Subseasonal, The
1147e1164, https://doi.org/10.1175/2008BAMS2355.1. National Academies Press, Washington, DC, 2018, https://doi.org/10.17226/
[47] J. Flemming, A. Inness, H. Flentje, V. Huijnen, P. Moinat, M.G. Schultz, O. Stein, 24938.
Coupling global chemistry transport models to ECMWF’s integrated forecast [67] S. Munir, M. Mayfield, D. Coca, S.A. Jubb, Structuring an integrated air quality
system, Geosci, Model Dev 2 (2009) 253e265. monitoring network in large urban areas e discussing the purpose, criteria
[48] A. Baklanov, M. Lawrence, S. Pandis, A. Mahura, S. Finardi, N. Moussiopoulos, and deployment strategy, Atmos. Environ. X 2 (2019) 100027, https://
M. Beekmann, P. Laj, L. Gomes, J.-L. Jaffrezo, A. Borbon, I. Coll, V. Gros, doi.org/10.1016/j.aeaoa.2019.100027.
J. Sciare, J. Kukkonen, S. Galmarini, F. Giorgi, S. Grimmond, I. Esau, A. Stohl, [68] J. Flemming, A. Benedetti, A. Inness, R.J. Engelen, L. Jones, et al., The CAMS
B. Denby, T. Wagner, T. Butler, U. Baltensperger, P. Builtjes, D. van den Hout, interim reanalysis of carbon monoxide, ozone and aerosol for 2003e2015,
H.D. van der Gon, B. Collins, H. Schluenzen, M. Kulmala, S. Zilitinkevich, Atmos. Chem. Phys. 17 (2017) 1945e1983, https://doi.org/10.5194/acp-17-
R. Sokhi, R. Friedrich, J. Theloke, U. Kummer, L. Jalkinen, T. Halenka, 1945-2017.
A. Wiedensholer, J. Pyle, W.B. Rossow, MEGAPOLI: concept of multi-scale [69] WMO, Global Atmosphere Watch Expert Meeting on Measurement-Model
modelling of megacity impact on air quality and climate, Adv. Sci. Res. 4 Fusion for Global Total Atmospheric Deposition (MMF-GTAD) (26-27
(2010) 115e120, https://doi.org/10.5194/asr-4-115-2010. February 2019; Geneva, Switzerland), 2020. WMO GAW Report No 250,
[49] R. Nuterman, B. Amstrup, A. Mahura, A. Baklanov, A. Zakey, D. Banciu, https://library.wmo.int/index.php?lvl¼author_see&id¼12920.
O. Diaconu, S. Tascu, M. Pietrisi, R. Dumitrache, Core-downstream processing [70] J. Kim, U. Jeong, M.H. Ahn, J.H. Kim, R.J. Park, H. Lee, C.H. Song, Y.S. Choi, New
chain test cases. Running the full downscaling model systems based on the era of air quality monitoring from space: geostationary environment moni-
MACC forecasts, Report D_OINT_2.7.1-2 (2011) 27. October 2011. toring spectrometer (GEMS), Bull. Am. Meteorol. Soc. (2019), https://doi.org/
[50] J. Benavides, M. Snyder, M. Guevara, A. Soret, C. Pe rez García-Pando, 10.1175/BAMS-D-18-0013.1.
F. Amato, X. Querol, O. Jorba, CALIOPE-Urban v1.0: coupling R-LINE with a [71] P. Ingmann, B. Veihelmann, J. Langen, D. Lamarre, H. Stark, G.B. Courre ges-
mesoscale air quality modelling system for urban air quality forecasts over Lacoste, Requirements for the GMES atmosphere service and ESA’s imple-
Barcelona city (Spain), Geosci, Model Dev 12 (2019) 2811e2835, https:// mentation concept: sentinels-4/-5 and5p, Remote Sens. Environ. 120
doi.org/10.5194/gmd-12-2811-2019. (2012) 58e69, https://doi.org/10.1016/j.rse.2012.01.023.
[51] X. Tang, Managing disaster risk in a mega-city, WMO Bull 55 (4) (2006). [72] P. Zoogman, X. Liu, R.M. Suleiman, W.F. Pennington, D.E. Flittner, J.A. Al-
October 2006. Saadi, et al., Tropospheric emissions: monitoring of pollution (TEMPO),
[52] G. Beig, D.M. Chate, S.K. Sahu, N.S. Parkhi, R. Srinivas, A. Kausar, S.D. Ghude, J. Quant. Spectrosc. Radiat. Transf. 186 (2017) 17e39, https://doi.org/
S. Yadav, H.K. Trimbake, System of Air Quality Forecasting and Research 10.1016/j.jqsrt.2016.05.008.
(SAFAR e India) vol. 217, WMO GAW Report No., Geneva, 2015, p. 51. [73] M. Bocquet, H. Elbern, H. Eskes, M. Hirtl, R. Zabkar,
G.R. Carmichael,
[53] G.P. Brasseur, Y. Xie, A.K. Petersen, I. Bouarar, J. Flemming, M. Gauss, F. Jiang, J. Flemming, A. Inness, M. Pagowski, J.L. Pe rez Caman ~ o, P.E. Saide, R. San Jose,
R. Kouznetsov, R. Kranenburg, B. Mijling, V.-H. Peuch, M. Pommier, A. Segers, M. Sofiev, J. Vira, A. Baklanov, C. Carnevale, G. Grell, C. Seigneur, Data
M. Sofiev, R. Timmermans, R. van der A, S. Walters, J.-M. Xu, G.-Q. Zhou, assimilation in atmospheric chemistry models: current status and future
Ensemble forecasts of air quality in eastern China e Part 1: model prospects for coupled chemistry meteorology models, Atmos. Chem. Phys. 15

269
A. Baklanov and Y. Zhang Global Transitions 2 (2020) 261e270

(2015) 5325e5358, https://doi.org/10.5194/acp-15-5325-2015. North America during the summer of 2004, J. Geophys. Res. 110 (2005)
[74] D. Tong, L. Pan, W. Chen, L. Lamsal, P. Lee, Y. Tang, H. Kim, S. Kondragunta, D21307, https://doi.org/10.1029/2005JD005858.
I. Stajner, Impact of the 2008 Global Recession on Air Quality over the, 2016. [92] M. Pagowski, G.A. Grell, S.A. McKeen, D. De ve
nyi, J.M. Wilczak, V. Bouchet,
[75] R.S. Sokhi, A. Baklanov, K.H. Schlünzen (Eds.), Mesoscale Modelling for W. Gong, j. McHenry, S. Peckham, J. McQueen, R. Moffet, Y. Tang, A simple
Meteorological and Air Pollution Applications, Anthem Press, 2018, ISBN method to improve ensemble ebased ozone forecasts, Geophys. Res. Lett. 32
9781783088263, p. 380. November 2018. (2005) L07814, https://doi.org/10.1029/2004GL022305.
[76] G.I. Marchuk. Adjoint Equations and Analysis of Complex Systems, Springer [93] M. Pagowski, G.A. Grell, D. Devenyi, S.E. Peckham, S.A. McKeen, W. Gong,
Netherlands, 1995, p. 468p. L. Delle Monache, J.N. McHenry, J. McQueen, P. Lee, Application of dynamic
[77] G.R. Carmichael, A. Sandu, T. Chai, D. Daescu, E. Constantinescu, Y. Tang, linear regression to improve the skill of ensemble-based deterministic ozone
Predicting air quality: improvements through advanced methods to inte- forecasts, Atmos. Environ. 40 (2006) 3240e3250, https://doi.org/10.1016/
grate models and measurements, J. Comp. Physiol. 227 (2008) 3540e3571. j.atmosenv.2006.02.006.
[78] J.E. Nielsen, S. Pawson, A. Molod, B. Auer, A.M. da Silva, A.R. Douglass, [94] S. McKeen, S.H. Chung, J. Wilczak, G. Grell, I. Djalalova, S. Peckham, W. Gong,
B.N. Duncan, Q. Liang, M.E. Manyin, L.D. Oman, W.M. Putman, S.E. Strahan, V. Bouchet, R. Moffet, Y. Tang, G.R. Carmichael, R. Mathur, S. Yu, Evaluation of
K. Wargan, Chemical mechanisms and their applications in the goddard several PM2.5 forecast models using data collected during the ICARTT/
earth observing system (GEOS) earth system model, J. Adv. Model. Earth Syst. NEAQS 2004 field study, J. Geophys. Res. 112 (2007) D10S20, https://doi.org/
(2017), https://doi.org/10.1002/2017MS001011. 10.1029/2006JD007608.
[79] V. Mallet, Ensemble forecast of analyses: coupling data assimilation and [95] S. Potempski, S. Galmarini, R. Addis, P. Astrup, S. Bader, R. Bellasio,
sequential aggregation, J. Geophys. Res. 115 (2010) D24303, https://doi.org/ R. Bianconi, et al., Multi-model ensemble analysis of the ETEX-2 experiment,
10.1029/2010JD014259. Atmos. Environ. 42 (2008) 7250e7265, https://doi.org/10.1016/
[80] D. Kang, R. Mathur, S.T. Rao, S. Yu, Bias adjustment techniques for improving j.atmosenv.2008.07.027.
ozone air quality forecasts, J. Geophys. Res. 113 (2008) D23308, https:// [96] I. Djalalova, Coauthors, Ensemble and bias-correction techniques for air
doi.org/10.1029/2008JD010151. quality model forecasts of surface O3 and PM2.5 during the TEXAQS-II
[81] D. Kang, R. Mathur, S.T. Rao, Real-time Bias-adjusted O3 and PM2.5 air quality experiment of 2006, Atmos. Environ. 44 (2010) 455e467, https://doi.org/
index forecasts and their evaluations over the continental United States, 10.1016/j.atmosenv.2009.11.007.
Atmos, Environ. Times 44 (2010a) 2203e2212. [97] A.K. Petersen, G.P. Brasseur, I. Bouarar, J. Flemming, M. Gauss, F. Jiang,
[82] D. Kang, R. Mathur, S.T. Rao, Assessment of bias-adjusted PM2.5 air quality R. Kouznetsov, R. Kranenburg, B. Mijling, V.-H. Peuch, M. Pommier, A. Segers,
forecasts over the continental United States during 2007, Geosci, Model Dev M. Sofiev, R. Timmermans, R. van der A, S. Walters, Y. Xie, J.-M. Xu, G.-
3 (2010b) 309e320. Q. Zhou, Ensemble forecasts of air quality in eastern China e Part 2. Evalu-
[83] J. Huang, J. McQueen, J. Wilczak, I. Djalalova, I. Stajner, P. Shafran, D. Allured, ation of the MarcoPolo-panda prediction system, version 1, geosci, Model
P. Lee, P. Li, D. Tong, H.-C. Huang, G. DiMego, S. Upadhayay, L. Monache, Dev. Discuss (2019), https://doi.org/10.5194/gmd-2018-234.
Improving NOAA NAQFC PM2.5 predictions with a bias correction approach, [98] W.R. Sessions, J.S. Reid, A. Benedetti, P.R. Colarco, A. da Silva, S. Lu,
Weather Forecast. 32 (2017) 407e421, https://doi.org/10.1175/WAF-D-16- T. Sekiyama, T.Y. Tanaka, J.M. Baldasano, S. Basart, M.E. Brooks, T.F. Eck,
0118.1,2017. M. Iredell, J.A. Hansen, O.C. Jorba, H.-M.H. Juang, P. Lynch, J.-J. Morcrette,
[84] I. Stajner, J. McQueen, P. Lee, J. Huang, H.-C. Huang, L. Pan, Y. Tang, D. Tong, S. Moorthi, J. Mulcahy, Y. Pradhan, M. Razinger, C.B. Sampson, J. Wang,
A. Stein, J. Wilzak, I. Djalalova, D. Allured, P. Dickerson, J. Tirado, National Air D.L. Westphal, Development towards a global operational aerosol consensus:
Quality Forecast Capability Prediction Updates: Bias Correction for Ozone, basic climatological characteristics of the International Cooperative for
Oral Presentation at the 17th Annual CMAS Conference, Chapel Hill, North Aerosol Prediction Multi-Model Ensemble (ICAP-MME), Atmos. Chem. Phys.
Carolina, U.S.A, 2018. Oct. 22-24. 15 (2015) 335e362, https://doi.org/10.5194/acp-15-335-2015.
[85] L. Delle Monache, J. Wilczak, S. Mckeen, G. Grell, M. Pagowski, S. Peckham, [99] J.W. Kaiser, M. Keywood, Preface for atmospheric environment special issue
R. Stull, J. Mchenry, J. McQueen, A Kalman filter bias correction method on IBBI, Atmos. Environ. 121 (2015) 1e3.
applied to deterministic, ensemble averaged,and probabilistic forecasts of [100] A. Tobías, M. Stafoggia, Modeling desert dust exposures in epidemiologic
surface ozone, Tellus, Ser. B 60 (2008) 238e249, https://doi.org/10.1111/ short-term health effects studies, Epidemiology 31 (6) (2020) 788e795,
j.1600-0889.2007.00332.x. https://doi.org/10.1097/EDE.0000000000001255. November 2020.
[86] L. Delle Monache, X. Deng, Y. Zhou, R. Stull, Ozone ensemble forecasts: 1. A [101] GAW, WMO Global Atmosphere Watch (GAW) Implementation Plan: 2016-
new ensemble design, J. Geophys. Res. 111 (2006a) D05307, https://doi.org/ 2023. WMO GAW Report No.228, World Meteorological Organization,
10.1029/2005JD006310. Geneva, 2017, p. 81.
[87] L. Delle Monache, T. Nipen, X. Deng, Y. Zhou, R. Stull, Ozone ensemble [102] S. Nickovic, E. Cuevas Agullo  , J.M. Baldasano, E. Terradellas, T. Nakazawa,
forecasts: 2. A Kalman filter predictor bias correction, J. Geophys. Res. 111 A. Baklanov, Sand and Dust Storm Warning Advisory and Assessment Sys-
(2006b) D05308, https://doi.org/10.1029/2005JD006311. tem, SDS-WAS) Science and Implementation Plan, 2015, pp. 2015e2020.
[88] L. Delle Monache, J. Wilczak, S. Mckeen, G. Grell, M. Pagowski, S. Peckham,
R. Stull, J. Mchenry, J. McQueen, A Kalman filter bias correction method
applied to deterministic, ensemble averaged,and probabilistic forecasts of Alexander Baklanov is science officer of the Science and Innovation Department, World
surface ozone, Tellus, Ser. B 60 (2008) 238e249, https://doi.org/10.1111/ Meteorological Organization (WMO), 7 bis, Avenue de la Paix, BP2300, 1211 Geneva 2,
j.1600-0889.2007.00332.x. Switzerland; and also affiliated as professor at the Physics of Ice, Climate and Earth
[89] R.E. Eskridge, J.-Y. Ku, S.T. Rao, P.S. Porter, I.G. Zurbenko, Separating different department, Niels Bohr Institute, University of Copenhagen, 2100 København Ø,
scales of motion in time series of meteorological variables, Bull. Am. Mete- Denmark, and lead scientist at the A.M. Obukhov Institute of Atmospheric Physics of
orol. Soc. 78 (1997) 1473e1483, https://doi.org/10.1175/1520-0477. Russian Academy of Sciences. He can be contacted at: abaklanov@wmo.int and
[90] D. Kang, C. Hogrefe, K. Foley, S. Napelenok, R. Mathur, S.T. Rao, Application of abaklanov@nbi.ku.dk
the Kolmogorov-zurbenko filter and the decoupled direct 3D method for the
dynamic evaluation of a regional air quality model, Atmos. Environ. 80
(2013) 58e69. Yang Zhang is professor at the Department of Civil and Environmental Engineering,
[91] S. McKeen, J. Wilczak, G. Grell, I. Djalova, S. Peckham, E.-Y. Hsie, W. Gong, Northeastern University, Boston, MA 02115, USA, She can be contacted at: ya.zhang@
V. Bouchet, S. Me nard, R. Moffet, J. McHenry, J. McQueen, Y. Tang, northeastern.edu
G.R. Carmichael, M. Pagowski, A. Chan, t. Dye, G. Frost, P. Lee, R. Mathur,
Assessment of an ensemble of seven real-time ozone forecasts over eastern

270

View publication stats