Available via license: CC BY 4.0
Content may be subject to copyright.
1
Satellite-based evaluation of AeroCom model bias in biomass
burning regions
Qirui Zhong1, Nick Schutgens1, Guido van der Werf 1, Twan van Noije2, Kostas Tsigaridis3,4, Susanne
E. Bauer4,3, Tero Mielonen5, Alf Kirkevåg6, Øyvind Seland6, Harri Kokkola5, Ramiro Checa-Garcia7,
David Neubauer8, Zak Kipling9, Hitoshi Matsui10, Paul Ginoux11, Toshihiko Takemura12, Philippe Le 5 Sager2, Samuel Rémy13, Huisheng Bian14,15, Mian Chin15, Kai Zhang16, Jialei Zhu17, Svetlana G. Tsyro6,
Gabriele Curci18,19, Anna Protonotariou20, Ben Johnson21, Joyce E. Penner22, Nicolas Bellouin23,
Ragnhild B. Skeie24, and Gunnar Myhre24
1Department of Earth Sciences, Vrije Universiteit, Amsterdam, The Netherlands. 10 2Royal Netherlands Meteorological Institute, De Bilt, the Netherlands.
3Center for Climate Systems Research, Columbia University, 2880 Broadway, New York, NY 10025, USA.
4NASA Goddard Institute for Space Studies, 2880 Broadway, New York, NY 10025, USA.
5Finnish Meteorological Institute, Kuopio, Finland.
6Norwegian Meteorological Institute, Oslo, Norway. 15 7Laboratoire des Sciences du Climat et de l'Environnement, IPSL, Gif-sur-Yvette, France.
8Institute for Atmospheric and Climate Science, ETH Zurich, Zurich, Switzerland.
9European Centre for Medium-Range Weather Forecasts, Reading, UK.
10Graduate School of Environmental Studies, Nagoya University, Nagoya, Japan.
11NOAA, Geophysical Fluid Dynamics Laboratory, Princeton, NJ, USA. 20 12Research Institute for Applied Mechanics, Kyushu University, Fukuoka, Japan.
13HYGEOS, Lille, France.
14University of Maryland, Baltimore County (UMBC), Baltimore, MD, USA.
15NASA Goddard Space Flight Center, Greenbelt, MD, USA.
16Pacific Northwest National Laboratory, Richland, WA, USA. 25 17Institute of Surface-Earth System Science, School of Earth System Science, Tianjin University, Tianjin 300072, China.
18Department of Physical and Chemical Sciences, University of L’Aquila, L’Aquila, Italy.
19Center of Excellence in Telesening of Environment and Model Prediction of Severe Events (CETEMPS), University of
L’Aquila, L’Aquila (AQ), Italy.
20Department of Physics, University of Athens, Athens, Greece. 30 21Met Office, Exeter UK.
22Department of Climate and Space Sciences and Engineering, University of Michigan, Ann Arbor, MI, USA.
23Department of Meteorology, University of Reading, Reading, UK.
24Center for International Climate and Environmental Research-Oslo (CICERO), Oslo, Norway.
Correspondence to: Qirui Zhong (q.zhong@vu.nl) 35
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
2
Abstract. Global models are widely used to simulate biomass burning aerosols (BBA). Exhaustive evaluations on model
representation of aerosol distributions and properties are fundamental to assess health and climate impacts of BBA. Here we
conducted a comprehensive comparison of Aerosol Comparisons between Observation project (AeroCom) model
simulations with satellite observations. A total of 59 runs by 18 models from three AeroCom Phase III experiments (i.e., 40
Biomass Burning Emissions, CTRL16, and CTRL19) and 14 satellite products of aerosols were used in the study. Aerosol
optical depth (AOD) at 550 nm was investigated during the fire season over three key fire regions reflecting different fire
dynamics (i.e., deforestation-dominated Amazon, Southern Hemisphere Africa where savannas are the key source of
emissions, and boreal forest burning on boreal North America). The 14 satellite products were first evaluated against
AErosol RObotic NETwork (AERONET) observations, with large uncertainties found. But these uncertainties had small 45
impacts on the model evaluation that was dominated by modeling bias. Through a comparison with Polarization and
Directionality of the Earth’s Reflectances (POLDER-GRASP) observations, we found that the modeled AOD values were
biased by -93–152%, with most models showing significant underestimations even for the state-of-art aerosol modeling
techniques (i.e., CTRL19). By scaling up BBA emissions, the negative biases in modeled AOD were significantly mitigated,
although it yielded only negligible improvements in the correlation between models and observations, and the spatial and 50
temporal variations of AOD biases did not change much. For models in CTRL16 and CTRL19, the large diversity in
modeled AOD was in almost equal measures caused by diversity in emissions, lifetime, and mass extinction coefficient
(MEC). We found that in the AEROCOM ensemble, BBA lifetime correlated significantly with particle deposition (as
expected) and in turn correlated strongly with precipitation. Additional analysis based on Cloud-Aerosol LIdar with
Orthogonal Polarization (CALIOP) aerosol profiles suggested that the altitude of the aerosol layer in the current models was 55
generally too low, which also contributed to the bias in modeled lifetime. Modeled MECs exhibited significant correlations
with the Ångström Exponent (AE, an indicator of particle size). Comparisons with the POLDER-GRASP observed AE
suggested that the models tended to overestimate AE (underestimated particle size), indicating a possible underestimation of
MECs in models. The hygroscopic growth in most models generally agreed with observations and might not explain the
overall underestimation of modeled AOD. Our results imply that current global models comprise biases in important aerosol 60
processes for BBA (e.g., emissions, removal, and optical properties) that remain to be addressed in future research.
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
3
1 Introduction
Biomass burning (BB) injects large amounts of aerosols into the atmosphere every year. It is estimated that BB is responsible
for 26−73% and 27−41% of global organic carbon (OC) and black carbon (BC) emissions, respectively (Bond, 2004; Andrea 65
and Rosenfeld, 2008; Wiedinmyer et al., 2011; Wang et al., 2014; Huang et al., 2015). As a result, BB aerosol (BBA) has a
considerable impact on human health and the global climate. For example, numerous studies have shown that exposure to
BBA can cause cardiovascular diseases and subsequently lead to premature death (Johnston et al., 2012; Lelieveld et al.,
2015). In addition, BBA can also alter the global and regional energy budgets by interacting with solar radiation directly, and
indirectly by modifying the lifetime and albedo of cloud through their role as cloud condensation nuclei and ice-nucleating 70
particles (Engelhart et al., 2012; Jahl et al., 2021). On a global scale, assessments of these health and climate impacts rely
directly or indirectly on model simulations regarding BBA’s distributions, compositions, and properties (Martins et al., 2009;
Lin et al., 2014; Dong et al., 2019).
One of the frequently used variables to define model representation for BBA is aerosol optical depth (AOD) which
depends on both aerosol abundance and optical properties in the atmosphere. Previous studies have reported that global 75
models produced substantial underestimations of AOD over BB regions with highly varying extent despite using different
emission inventories (Kaiser et al., 2012; Veira et al., 2015; Johnson et al., 2016; Reddington et al., 2016; Mallet et al.,
2021). For example, Kaiser et al (2012) showed the global Monitoring Atmospheric Composition and Change (MACC)
aerosol model driven by emissions from the Global Fire Assimilation System (GFAS) underestimated AOD by a factor of 2–
4 for BBA. While Johnson et al (2016) found that the AOD was underestimated by a factor of 1.6–2 in the simulations by 80
Hadley Centre Global Environment Model version 2 and 3 (HadGEM2 and HadGEM3) based on the Global Fire Emission
Database version 3 (GFED3). The systematic underestimation of AOD in global models suggests a potential negative bias in
current BB emission inventories (Reddington et al., 2016). Several factors could contribute to producing such bias in
emission inventories based on either satellite-detected burned areas (e.g., van der Werf et al., 2017) or fire radiative power
(FRP, e.g., Ichoku and Ellison, 2014). The burned-area-based emission inventories comprise uncertainties in satellite 85
detection of burned areas and fuel load (Randerson et al., 2012; Andela et al., 2016), while FRP-based emission datasets are
largely affected by the translation of FRP into rates of biomass combustion (Kaiser et al., 2012). In addition, both emission
datasets rely on uncertain emission factors converting burned biomass to trace gas or aerosol emissions (Stockwell et al.,
2015). Moreover, when these emission inventories are used to run models, the OC emissions will be converted to emissions
of organic aerosols (OA) based on the assumed OA/OC ratio which differs extensively among models (Gliß et al., 2021). It’s 90
thus expected to see large diversities in simulated AOD from models driven by varying BBA emission inventories.
In addition to emissions, model performance for simulating BBA also depends on model configurations. This has been
reported for individual models. Reddington et al (2019) showed that increasing the aerosol hygroscopicity can reduce AOD
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
4
errors simulated by Global Model for Aerosol Processes (GLOMAP) over tropical BB regions. A similar impact of
hygroscopicity was also observed in Johnson et al (2016) by comparing the modeled AOD errors between two aerosol 95
schemes in HadGEM3 model. Schill et al (2020) found that the large BBA biases in the remote troposphere could be
eliminated by increasing wet removal strength. Additional configurations that can alter model performance include, for
example, model resolution (Bian et al., 2009), particle size distribution (Chin et al., 2009), complex refractive index (Brown
et al., 2021), aerosol lifetime (Bauer et al., 2013), and aerosol mixing state (Cappa et al., 2012; Brown et al., 2021). With
different assumptions, methodologies, and parameterizations selected for aerosol processes in models, model evaluations can 100
be very different even when the same emission inventory is used.
Apart from the issues in emissions and model configurations, the uncertainty in observations is another factor affecting
model evaluations. AErosol RObotic NETwork (AERONET) is frequently used as a solid observation dataset for aerosols
(Tombette et al., 2008; Smirnov et al., 2011). However, the AERONET network is not particularly well aligned with BBA
regions and available observations are limited (e.g., in Africa, Siberia). Over specific BB regions, flight campaign 105
measurements are applied to be compared with models for certain periods (e.g., Myhre et al., 2003; Johnson et al., 2016).
But the temporal coverage of these campaigns is limited given the large inter-annual variability of fires (van der Werf et al.,
2017), and the observations suffers from uncertainties due to sampling instruments (Pistone et al., 2019). In comparison,
satellite datasets provide more continuous observations in space and time. Unfortunately, satellite remote sensing, conducted
by either a polar-orbiting or geo-stationary satellite, suffers from a series of uncertainties and noise that can originate from 110
radiance calibration, cloud screening, the effects of strong surface reflection, and the variation in aerosol particle sizes and
components (Li et al., 2009; Schutgens et al., 2020; Falah et al., 2021). As a result, the satellite retrieved AOD displays
significant variations. For example, Schutgens et al (2020) found that the diversities of individual satellite products can reach
up to 100% on regional scales. It is therefore necessary to understand the uncertainties in the satellite products prior to the
model validation. 115
To better quantify and interpret the model bias of BBA, we conducted a comprehensive inter-comparison between
various global models and observations. The aim of this work is to provide a satellite-based assessment of global models in
representing BBA and to see if models have been improved regarding to knowing issues for BBA models for more than ten
years. This study focusses on AOD at 550 nm−a basic optical property used to measure the abundance of aerosols in the
atmosphere−during fire seasons. A model ensemble was built from three phase-III experiments of the Aerosol Comparisons 120
between Observation (AeroCom) project. Such a comparison between models and satellite observation ensembles will
provide more robust results than individual comparisons, and the spread of individual models allows an in-depth
interpretation of the modeled diversities. Additional modeled variables and observations (e.g., total emissions, aerosol load,
precipitation, plume height, Ångström Exponent, hygroscopic growth) were also used to further aid in the interpretation.
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
5
Prior to the model validation, we assessed a total of 14 satellite products to identify the possible uncertainties induced by 125
observations of AOD. The paper is organized as follows. The details of the methodology and data sources are presented in
Section 2. Section 3 evaluates satellite observation uncertainties over the selected fire regions and their impacts on model
validations. Section 4 quantifies the model bias in AOD. Section 5 presents the diversity in modeled AOD which is further
interpreted through three aspects of the modeling processes.
2 Data and Methods 130
2.1 Models and variables
This study evaluated the AOD at 550 nm simulated by models from three AeroCom Phase-III experiments: the biomass
burning emissions experiment (BBE), control experiment 2016 (CTRL16), and control experiment 2019 (CTRL19). Table 1
provides an overview of all the models involved in our evaluation, more details are provided in the appendix questionnaire
and listed references. A total of 18 different models were investigated in our study, and some models participated in multi 135
experiments with different versions (Table 1). The general settings of the three experiments were as follows.
The aim of the BBE was to quantify the impact of BBA emissions on AOD simulations. All the participating models
presented simulations for the year 2008 using the prescribed BB emission input (GFED3). In addition, simulations with
scaling factors of 0, 0.5, 2, and 5 (referred to as BBE0, BBE0.5, BBE2, BBE5) adapted to GFED3 emissions were also
provided. These scaling factors were based on a preliminary simulation by the Goddard Chemistry Aerosol Radiation and 140
Transport (GOCART) aerosol model, which found that using default GFED3 emissions would lead to AOD
underestimations over most fire regions (Petrenko et al., 2012). The perturbations in emissions would allow a quantitative
analysis of the AOD-emission response.
The models in CTRL16 adopted the standard diagnostics and presented simulations for 2006, 2008, and 2010. The
modelers were advised to nudge the meteorology to (or drive the models by) their preferred datasets (see Table 1). The 145
standard outputs mainly included 2-D fields at a monthly frequency, which were extended by several other experiments
launched subsequently (e.g., the remote sensing experiment). High-frequency (3-h) AOD data together with other
information (e.g., 3-D fields of the AOD) are currently available. In this study, we examined 12 models with an AOD output
at a 3-h frequency for 2006, 2008, and 2010.
The state-of-art of aerosol modeling for 1850 (pre-industrial era) and 2010 (present day) was assembled in CTRL19. All 150
models were nudged to (or driven by) a fixed sea surface temperature and 2010 meteorology using different data sources
(see Table 1). Emissions from Coupled Model Intercomparison Project Phase 6 (CMIP6) were used when applicable. The
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
6
model AOD was output at a daily or monthly frequency. In this study, we selected 12 models that provided a daily output for
2010.
In addition to AOD, additional variables from the models were used to interpret model diversity when available. These 155
additional variables included emissions, total deposition (both dry and wet deposition), aerosol column load (with aerosol
species resolved), the vertical profile of the extinction coefficient (EC), precipitation, and Ångström Exponent (AE, which
was calculated using the AOD at 440 and 550 nm, the AE-based interpolation was adopted if AOD at 440 nm was not
available for some models).
We also prepared a questionnaire filled by modelers to acquire information on the model configuration details (see 160
Appendix). Information was collected for models in CTRL16 and CTRL19.
2.2 Fire regions
Based on the models considered, three key BB regions were selected in this study: Amazon (AMAZ), Southern Hemisphere
Africa (SHAF), and boreal North America (BONA). Figure 1 shows the domains of these three regions and the
corresponding OC emissions from BB. In terms of their aerosol emission, different fire types could be identified in each 165
region. The BB emissions in AMAZ were dominated by tropical forest fires and deforestation, whereas emission from
savanna grassland fires was the major source in SHAF. In BONA, BB aerosols were mainly emitted from boreal forest fires.
Regions with agricultural waste burning or temperate forest fires were not considered due to their small contribution on a
global scale (van der Werf et al., 2010). Using the satellite observation of AOD, we defined the fire seasons (dry seasons)
over the three regions (see Figure 1b) that were investigated in this study. 170
2.3 Observation data
A total of 14 satellite AOD datasets were used in this study. Table 2 provides an overview of the datasets. The AOD data at
550 nm wavelength were obtained by either direct retrieval or interpolation/extrapolation from the AOD at nearby
wavelengths.
The ground-based remote sensing data were taken from AERONET DirectSun L2 v3 (Dubovik et al., 2000). The 175
locations of the AERONET sites within the three fire regions are shown in Figure 1a. Given that the sparse distribution of
AERONET sites results in poor spatial data coverage, especially in SHAF and BONA, we mainly used the AERONET data
to evaluate the satellite datasets, while model validations relied on satellite data.
For the vertical profiles, we used the Cloud-Aerosol LIdar with Orthogonal Polarization (CALIOP) L2 Layer 5 km v4.20
product. The EC data at 532 nm were compared with models (at 550 nm) where the vertical data were available. For 180
CALIOP data, we only considered columns that had at least one aerosol retrieval based on the cloud-aerosol discrimination
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
7
(CAD) scores (CAD < -20) (Watson-Parris et al., 2018). Columns with extreme CAD scores (< -100) were also excluded
because they might have been the result of bad shots (Watson-Parris et al., 2018). To ensure data quality, we only used the
most reliable retrievals that had extinction quality control (QC) flags of 0, 1, 2, 16, or 18. In addition to the direct
comparison of vertical extinction profiles, we calculated the weighted mean plume extinction height (PEH) based on the 185
vertical EC and layer height (hi) for the aerosol layers below 6 km (Koffi et al., 2016), as shown by Eq. (1):
PEH =
(1)
In addition, we evaluated the modeled precipitation as it is the cause of a major deposition process. The precipitation data
were taken from the Global Precipitation Climatology Project (GPCP), which incorporates precipitation from low-orbit
satellite data, geosynchronous satellite data, and surface raindrop observations (Adler et al., 2003). 190
2.4 Data analysis
To avoid sampling issues, we conducted strict collocations before the data were evaluated (Schutgens et al., 2016a; b). Both
model and observation data were firstly re-gridded into the 1° × 1° spatial grid-boxes. The temporal resolution was
aggregated into 3-h or daily intervals according to the model output frequency (see Table 1). For the satellite validation
against AERONET, we compared satellite data with AERONET at the resolution of 1° × 1° × 3-h. Specially, the plume 195
height in models was validated against CALIOP on monthly basis since CTRL19 models only provided data at such a
resolution. Vertically, the CALIOP data were aggregated into 100-m intervals and all the extinction profiles from models
were linearly interpolated into the same resolution for validation.
The data aggregation and collocation were processed via a command-line tool called Community Intercomparison Suite
(CIS, Watson-Parris, et al., 2016). To quantitatively evaluate the model performance and satellite observation uncertainties, 200
we utilized Taylor diagrams to present the statistics, including the Pearson correlation coefficient (R), standard deviation
(SD) and centered root mean square error (CRMSE) (Taylor, 2001). Taylor diagrams are presented in polar coordinates with
the polar axis showing the SD of evaluated data and cosine of the polar angle showing the R-value between evaluated and
‘reference’ data. The distance between the evaluated and ‘reference’ data shows the CRMSE according to the law of cosines.
Both evaluated and ‘reference’ data were normalized by the SD of ‘reference’ data so that the ‘reference’ was always located 205
at [1, 0] (see Figure 2a for an example). A Taylor diagram is a convenient way to visualize the performance of models or
observations versus a reference data set. However, bias is not shown by Taylor diagrams, and we accompanied each Taylor
diagram with a plot showing the normalized mean bias (NMB, defined as the mean bias divided by the mean value of
observation) to provide a comprehensive evaluation.
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
8
3 Evaluation of satellite products 210
A large number of satellite AOD datasets have become available, and it is important to use the dataset that can adequately
serve the specific research goal. In light of the uncertainties in satellite observations, we evaluated individual satellite
datasets against AERONET observations before model validation in the three fire regions. The evaluation was only
conducted for data during the fire seasons, and most observations were collected over AMAZ.
Figure 2 shows the evaluation of 14 satellite datasets against AERONET observations for the three fire regions during 215
the fire season. The data points in the Taylor diagram were normalized by AERONET data with a different sampling for
each product (Figure 2a), while NMBs are shown in the scatter diagram (Figure 2b). All the satellite datasets agreed with
AERONET observations over AMAZ better than the other two regions, with stronger correlations (R = 0.85–0.95) and lower
normalized CRMSE (< 0.5). For AMAZ, all the datasets had similar correlations and CRMESs but very different biases. The
POLDER-GRASP dataset and two algorithms adopted to Moderate Resolution Imaging Spectroradiometer (MODIS) data 220
(BAR and DarkTarget) tended to overestimate AOD (3–13%), while the others resulted in underestimations (-1–-20%).
Unlike in AMAZ, individual satellite products agreed less well with AERONET and there were strong variations within each
of them over SHAF and BONA (R = 0.31–0.91, CRMSE = 0.51–1.71). All products except Aqua-MODIS-BAR
underestimated AOD over SHAF (-7–-73%), whereas most products overestimated AOD over BONA by up to +73%. Both
the spatial and temporal data coverages in BONA and SHAF were much lower than in AMAZ, which suggested that there 225
might be more cloud contamination issues in the two regions and higher biases could be expected (Schutgens et al., 2020).
Generally, we found that MODIS products agreed well with the AERONET data, although details vary by the retrieval
algorithm. For example, the MODIS-BAR products were the best in AMAZ and SHAF, while the MODIS-MAIAC product
was better than the others in BONA. From the perspective of bias, we found that the variations among satellite products were
affected more by the algorithm than the instrument, which was related to the amount of spectral information used in the 230
retrieval. For example, the data spread of the four instruments that adopted the DeepBlue algorithm (i.e., Aqua-MODIS-DB,
Terra-MODIS-DB, AVHRR-DB, and SeaWiFS-DB) was smaller than that for the MODIS products that used four different
algorithms (i.e., BAR, DB, DT, and MAIAC) for all three regions.
It should be noted that the evaluation was affected by representation issues. As shown by Figure 1, there were more
AERONET sites located in fire areas in AMAZ. While in SHAF, the AERONET sites were far from the fire emission sites 235
and the downwind area and only captured a small part of the BB aerosol signals. In BONA, the temporal coverages of both
AERONET and satellites were poor. Due to the stratocumulus and low broken cumulus cloud contamination, satellite
retrievals of AOD were enhanced, which could lead to unexpected overestimations when compared with the ground-based
observations over BONA (Toth et al., 2013).
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
9
In this study, we proposed to use POLDER-GRASP to evaluate all the models. AOD from POLDER-GRASP has been 240
validated in a previous study which suggests POLDER-GRASP is superior to other products globally (Schutgens et al.,
2021). The AE data has also been validated before, showing a good agreement with AERONET (Chen et al., 2020). In our
study, we also investigated how the observation uncertainties mentioned above may affect model validations, which were
indicated by the interquartile ranges of the R, CRMSE, and NMB based on validations using different satellite products. The
interquartile values were further compared with the statistics (i.e., R, CRMSE, and NMB) when using POLDER-GRASP to 245
show the uncertainty range when using the specific dataset. Before calculating the difference in model validation (i.e., R,
CRMSE, NMB) due to different satellite products, each model was collocated with satellite products either individually (i.e.,
all the models were collocated with the different sampling of each satellite product, see Figure 3a-c) or synchronously (i.e.,
the model data were collocated with the same sampling where all satellite products could provide data, see Figure 3d-f). In
the latter case, only products that had a similar overpass time with POLDER-GRASP were considered (i.e., with an overpass 250
time in the afternoon, excluding datasets onboard Terra and ENVISAT). For comparison, the uncertainty ranges of 25%,
50%, and 100% for the relative uncertainties to POLDER-GRASP were also shown. We found that for R and CRMSE using
an individual collocation (Figure 3a-c), the uncertainties due to the different satellite products were generally lower than
25%, indicating a small impact when using different satellite datasets. The impact on CRMSE was slightly stronger than that
on R, which suggested more agreements among the different satellite products (or better performance) in capturing the 255
spatiotemporal trends of AOD than the magnitude. In the case of NMB, the impacts of different were large only when the
modeled NMB was small (<20%). The majority of simulations had an NMB higher than ± 40%, suggesting the uncertainties
among the different satellite products were less important for NMB and the modeled bias was dominated by the biases in the
model instead of the difference in satellite products. For the synchronous collocation which eliminated the sampling
differences (Figure 3d-f), similar results were obtained with even much smaller satellite uncertainties. In this case, all the 260
satellite products were collocated, which greatly reduced the frequency of cloud contamination issues and provided more
reliable results. Due to the synchronous collocation, a large portion of the original observations was filtered and statistical
noise may stand out. We then conducted a 10,000-time bootstrap sampling with replacement to examine the potential effects
of such noise. In each time, we randomly excluded 20% of the data to test the robustness of our evaluations. The coefficient
of variation for the satellite observation uncertainties from the 10,000-time bootstrap sampling was 1–10% for R, 1–12% for 265
CRMSE, and 3–27% for NMB. For the stronger variation in NMB, over 85% of simulations were subject to an NMB
variation of less than 10%, suggesting very robust results for the above analysis. All this indicated that although there were
different errors in these satellite products, only a small part (accounting for < 25% of the modeled errors) could be expected
to affect the model validation. Given the small impacts, we decided to validate models against POLDER-GRASP product for
both AOD and AE, which provided a degree of consistency for the whole research. 270
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
10
However, the small validation impacts of using different satellite products were only found in our validations over BB
regions during BB seasons. Such a conclusion cannot be directly applied to other areas/periods. For example, for the same
fire regions outside the fire seasons, we found that the uncertainties due to different satellite products could be as high as
50% in most cases (not shown).
4. Evaluation of AeroCom models 275
We then evaluated AOD in AeroCom models in three experiments using the POLDER-GRASP product. All model data were
collocated with POLDER-GRASP sampling. The model evaluation is shown in Figure 4 via Taylor diagrams and bias plots.
The R-values ranged from 0.1 (INCA over BONA in BBE) to 0.78 (ECMWF-IFS over SHAF in CTRL19) for all models
and regions, with a median value of 0.63. Over 80% of the model simulations had an R-value higher than 0.5, but only 24%
of simulations had correlations stronger than 0.70, suggesting a generally moderate capability for capturing the 280
spatiotemporal variation in aerosol data. For CRMSE, the modeled variation (defined as the inter-quantile range divided by
the median value, 51%) was stronger than that for R (22%), indicating a higher modeled disparity of the AOD magnitude
than the spatiotemporal trends. Based on an analysis of variance for R and CRMSE (Figure 4a-c), we found that the models
showed similar performance over the three regions as there was no significant difference found. The median NMB of models
(Figure 4d) for AMAZ, SHAF, and BONA were -28% (-6%–-54% as inter-quartile), -54% (-30%–-63%), and -54% (-46%–285
-57%), respectively. Models produced significantly smaller NMB over AMAZ than over the other two regions, though the
inter-model variation was also found to be the highest among the three regions. More than half of the simulations showed an
underestimation of AOD by a factor of > 2, consistent with previous studies (e.g., Kaiser et al., 2012; Veira et al., 2015;
Johnson et al., 2016).
In addition to the overall model evaluations, we also evaluated the modeled temporal (time series for the whole fire 290
regions) and spatial patterns (temporal averages for individual grid-boxes during fire seasons). In Figure 5, we compared the
temporal and spatial correlations of modeled AOD with observations. Most models showed similar temporal and spatial
correlations ranging from 0.6 to 0.9 which were slightly higher than the overall correlations shown in Figure 4 due to data
averaging. Both the spatial and temporal correlations in most models clustered in this range, which partly explained the
similarity of the overall correlations mentioned above. We found there was no significant difference between the temporal 295
and spatial correlations in individual models from the three experiments. Although the AOD errors differed substantially per
model, the spatial and temporal variation among models tended to be small. For most models, we did not observe significant
improvements in spatial and temporal correlations following the time sequence from BBE to CTRL16 to CTRL19. We also
compared the variations of temporal and spatial AOD biases AOD biases, as shown in Figure 6. Here the variations were
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
11
defined as the ratio of interquartile to median values of the time series (temporal variations) or spatial averages (spatial 300
variations) of absolute modeled AOD bias. The spatial variations were significantly smaller than temporal variations for all
three experiments, suggesting the different temporal evolution of AOD biases was the leading cause of the large NMB
diversity in Figure 4. It partly suggested that current emission inventories had a better representation of BBA emissions over
space than over time.
Since the modeled AOD bias is strongly affected by input emissions (Kaiser et al., 2012; Johnson et al, 2016), we also 305
investigated the model response to the changes in emissions based on BBE experiment. This scaling-up procedure has been
used to fix overall AOD errors for BBA in previous studies (e.g., Kaiser et al., 2012; Johnson et al, 2016; Veira et al. 2015).
Figure 7 shows the evaluation of these models for R, CRMSE, and NMB. As expected, NMB increased monotonously with
the increase of emissions. Most models would produce significant positive bias when the scaling factors to GFED3 reached
5, but more than half of models still underestimated AOD when BBA emissions were doubled. Such trends were also found 310
for CRMSE with a much weaker sensitivity. Similar phenomena were also found in the other two experiments. For example,
we found the ECHAM-HAM model agreed well with observation in CTRL16 experiment which used 3.4 × GFAS
emissions, while it produced large underestimation when the CMIP6 emissions (much lower than 3.4 × GFAS) were used
(see Figure 4d). Given the metrics of CRMSE and NMB, the ensemble of models in BBE experiment showed the best
agreement with observations when the emissions were scaled by a factor of 2. This systematic response of modeled bias also 315
suggested a possible underestimation of emissions in the applied inventory (GFED3). However, correlations in most models
did not improve along with the increased emissions, since there was no further spatiotemporal information added into the
emissions.
The modeled AOD bias during fire seasons could be due to both BBA and background sources (e.g., anthropogenic,
biogenic, dust, and sea salt aerosols). However, it is difficult to isolate BBA errors from the background based on existing 320
simulations. Since we found that most models underestimated AOD in the BBE1 simulations, it was not possible to
determine the real BBA impacts by comparing BBE1 and BBE0 simulations. Instead, we compared the collocated BBE0
AOD (background) with POLDER-GRASP observation during fire seasons. The modeled AOD in BBE0 varied
substantially by a factor of 9 in the three regions. Compared with observations, the background averagely accounted for only
14%, 12%, and 11% of total AOD over AMAZ, SHAF, and BONA, respectively. We also compared the modeled AOD 325
biases during non-fire seasons with those during fire seasons, with the former showing much smaller magnitude compared
with the latter (0.04 vs 0.35, for the absolute mean bias). This analysis supports the notion that AOD bias over the fire
regions was dominated by the BBA rather than background sources.
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
12
5. Model diversity and its interpretation
In this section, we explore the diversities of AOD which lead directly to the differences in bias discussed above (see Figure 330
4). Understanding the model diversities and the drivers would improve our knowledge of model bias, which will enable
further development of the models. Our strategy is to first evaluate the variation in modeled AOD and the possible causes
that could lead to such model variability, and then compare those causes with observations to understand the model
variability and therefore bias. Unless stated otherwise, data in this section are presented as area averages for the whole fire
season based on the raw model outputs without any collocation. The aim is to determine the general drivers of variation in 335
AOD for model ensembles rather than individual models, although evaluations for specific models are also presented where
sufficient information is available.
The diversities of AOD were decomposed into three factors, i.e., total aerosol emissions, aerosol lifetime, and the MEC,
as described by the following function:
AOD =Emission × Lifetime × MEC (2) 340
where emission indicates the total emissions of OA (including secondary organic aerosols which were treated as emitted
aerosols given the fast transformation), BC, sulfur dioxide (SO2), sulfate (SU), mineral dust (DU), and sea salt (SS) within
the fire regions; lifetime is defined as the average total aerosol load divided by total emissions within the fire regions; and
the MEC is defined as AOD divided by total aerosol load, which is strongly associated with the modeled aerosol optical
properties (e.g., size distribution, refractive index, hygroscopicity, etc.). Emissions, aerosol load, and AOD were first 345
calculated as regional and seasonal averages so that the lifetime and MEC were determined on a seasonal level for the
focused regions. Note that the definition of lifetime in this study is different from the usual one as we are considering open
systems. However, the time scale here called lifetime is still determined by the same relevant process (e.g., depositions). This
is discussed in detail in Sec. 5.1.
Figure 8 shows the diversities of the three factors. The slope of the line between each dot and the origin indicates the 350
aerosol lifetime (Figure 8a) and MEC (Figure 8b) for a specific model averaged for the whole fire season, respectively. The
emissions varied by a factor of 10 among the models. Such large deviations resulted from different emission inventories
(mainly for CTRL16 models) and the different schemes for estimating non-BB aerosols (e.g., dust, biogenic sources). For
CTRL19 experiment with its prescribed emission inventory (CMIP6), the input emissions were altered mainly by the
different OA/OC ratios, and to a lesser extent by the different mechanisms of DU production and biogenic sources. For 355
example, the OA/OC ratios were set as 1.4, 1.8, and 2.6 in ECHAM-HAM, GEOS, and SPRINTARS, leading to emissions
being 34% and 88% higher in the latter two models, respectively. The difference in these ratios is a consequence of the
different assumptions regarding the oxidation of freshly emitted OC. The widely used ratio of 1.4 was established based on
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
13
field measurements over urban regions (Turpin and Lim, 2001) and was therefore more representative for anthropogenic OC
emissions. More recent investigations of the BB plume have suggested that the oxidation levels are higher for both fresh and 360
aged BB OC particles (Aiken et al., 2008; Brito et al., 2014; Tiitta et al., 2014). Increasing the OA/OC ratio can directly lead
to an elevated AOD in models and a higher ratio than 1.4 has been suggested for BB aerosols in some previous studies (e.g.,
Reid et al., 2005; Aiken, et al., 2008; Johnson et al., 2016). Omitting GISS-MATRIX and GISS-OMA which produced an
unexpected positive AOD bias (see Figure 4), we found that the modeled NMB generally decreased with an increase in the
OA/OC ratio for CTRL19 models. For example, the average NMB of AOD for the model group that used the ratio of 1.4 365
(i.e., CAM5-ATRAS, ECHAM-HAM, and ECHAM-SALSA) was -61%, whereas the value was only -22% for the group
using a ratio of 2.6 (i.e., CAM-Oslo, NorESM, OsloCTM, and SPRINTARS). The NMB of the models using a ratio of 1.6–
1.8 was within an intermediate range (-43% to -46%). This shows the importance of determining realistic values of the
OA/OC ratio. However, it does not necessarily mean that higher OA/OC ratios can address the underestimated AOD. For
example, both SPRINTARS and OsloCTM produced significant overestimations in AMAZ using an OA/OC ratio of 2.6 that 370
was higher than many in-situ observations (e.g., Brito et al., 2014; Zheng et al., 2017).
When all three fire regions were considered simultaneously, there was a general linear response of the aerosol load to
aerosol emissions, and of AOD to aerosol loads. Nevertheless, significant diversities in lifetime and MEC were found. For
the three regions, the relative variation (i.e., interquartile value divided by the median value) was found to be the lowest for
the MEC (49%, 41%, and 40% for AMAZ, SHAF, and BONA, respectively), moderate for aerosol lifetime (62%, 49%, and 375
26%, respectively), and highest for emissions (62%, 95%, and 64%, respectively). For the aerosol lifetime and MEC which
were mainly affected by other model aspects than emissions, there was no significant difference found among the three fire
regions for the same model. HadGEM3 over BONA presents an outlier case for lifetime which is probably related to high
local DU emissions. We also noticed that the DU emission in HadGEM3 covered a much wider area than in the other models
due to the use of different mechanisms (Woodward, 2001; Mulcahy et al., 2020). 380
The contributions of aerosol emissions, aerosol lifetime, and the MEC (which were found to be statistically independent
on each other) to the overall variation in AOD were evaluated. We used Eq. 2 to investigate such contributions. In the case
of the AOD variation induced by emissions, we calculated the AOD variation (i.e., the standard deviation) over all modeled
emissions and a random combination of aerosol lifetime and MEC values from the model ensemble. This calculation was
repeated for all the combinations of aerosol lifetime and the MEC, and the variation in AOD attributable to emissions was 385
then quantified as the average value of all the standard deviations. Similar calculations were also applied to aerosol lifetime
and MEC values. It was estimated that aerosol emissions, aerosol lifetime, and the MEC accounted for 38%, 33%, and 29%
of the variation in AOD, respectively, suggesting only small differences in determining the overall variation, although
emissions might be slightly more important than aerosol lifetime and the MEC. We also applied this evaluation to individual
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
14
fire regions and similar conclusions were obtained. This suggests that reducing the uncertainties associated with emissions 390
uncertainties might have only a moderate impact on the accuracy of the BBA simulations, and uncertainties in lifetime and
MEC should also be considered.
5.1 Diversity of aerosol lifetime
In this section, we discuss the potential factors that contribute to the diversity of aerosol lifetime. Because we focused on
three separate open systems, we described the aerosol budget of each region as a simple box model, as shown by Eq. (3): 395
= + +
=
(3)
where B, E, D, I, O, P, and L indicate the average of total aerosol burden, emission, deposition, inflow, outflow, chemical
production, and chemical loss of a focused region. For a closed system and a steady state without chemistry I = O = P = L =0
and a lifetime can be defined as E/B = D/B. For an open system, steady-state and with on-going chemistry, E/B does not
equal to D/B but both still are time-scales defining the system. Here we show that for these fire regions, E/B correlates with 400
D/B. Figure 9 displays the linear dependence of the modeled aerosol lifetime on the time scale of total deposition and all
other processes. For most models, the reciprocal of aerosol lifetime (E/B) responded linearly to the time scale of deposition
(D/B), except for INCA from CTRL16. This suggests that the difference in deposition is a leading contributor to the
variation in aerosol lifetime. HadGEM3 simulations in BONA (the outliers of the aerosol lifetime trend in Figure 8) still
followed the same linear trend, confirming that the short aerosol lifetime is a direct result of the strong deposition of coarse 405
mineral dust. For INCA, the simulated aerosol load was much lower than other models, and the modeled aerosol
composition was very different with OA contributing less than 20% of the total aerosol load. As a result, the coarse mode
aerosols dominated the total aerosol composition, resulting in a relatively short aerosol lifetime. When the INCA model was
omitted, the correlation between the reciprocal of aerosol lifetime and the deposition timescale was 0.95, suggesting that
90% of the modeled variation in aerosol lifetime could be explained by deposition. The variation in regional transport and 410
the chemical budget together only contributed around 10% of the variation in aerosol lifetime and was therefore much less
important to the overall difference in AOD. The timescale of the total deposition had a variation of 72% (i.e., the
interquartile value divided by the median) which was slightly higher than the aerosol lifetime (62%).
The modeled deposition was primarily a consequence of wet deposition (61% of the total deposition on average) even
during the dry fire season. The modeled wet deposition, which occurred mainly due to below-cloud scavenging (Andronache 415
2003; Zhang et al., 2004), was related to the size distribution of aerosols and raindrops as well as the precipitation intensity
(Seinfeld and Pandis, 2006). Figure 10 compares the modeled timescale of total deposition and precipitation strength. Note
that not all models provided both deposition and precipitation outputs, the conclusion of the evaluation may need to be re-
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
15
examined when more data becomes available in the future. The modeled precipitation differed among the models by factors
of 3.8, 13.6, and 2.2 for AMAZ, SHAF, and BONA, indicating a substantial model discrepancy. When all regions were 420
considered, there is a significant positive correlation between the modeled precipitation and the timescale of total deposition.
For comparison, we also compared models with GPCP data. GEOS, SPRINTARS, and TM5 were among the models with an
overestimated precipitation in all three fire regions, which suggested systematic errors in the modeled lifetime. On a regional
basis, models exhibited large regional variations. Almost all models tended to overestimate the precipitation over BONA by
up to 69% (ECHAM-HAM from CTRL19), which might partly explain the underestimated AOD in this region. There were 425
large disparities in precipitation simulation over AMAZ, ranging from -21% to 130%. In contrast, we found that most
models underestimated both AOD and precipitation in SHAF, suggesting other important sources of AOD bias in addition to
precipitation. However, we did not observe a clear dependence of AOD biases on precipitation biases. For example, a bias of
6% and -9% were found for precipitation and AOD, respectively, over AMAZ in CAM5-ATRAS from CTRL19, whereas
the corresponding AOD biases were 14% and -86% over BONA. This suggests that other factors than precipitation affect 430
AOD biases significantly.
In addition to precipitation, we also examined the impacts of aerosol plume height on the aerosol lifetime. Figure 11a
compares the modeled plume height (as represented by PEH) and aerosol lifetime. Based on the limited number of models
with data available, there was a generally increasing trend in the aerosol lifetime as the plume height increased (r = 0.65)
except for one outlier (IMPACT over BONA), suggesting that plume height could also affect the modeled aerosol lifetime. 435
Generally, the modeled PEH varied by a factor of 4, partly due to the model assumption in the fire injection height for BB
emissions. For example, ECHAM-HAM and ECHAM-SALSA, which allowed 25% of BB aerosol emissions to be emitted
above the planetary boundary layer (PBL), generally had a higher plume height than models that distributed emissions within
the PBL (e.g., GEOS, GISS-MATRIX). For validation, we further compared the aerosol vertical profiles between models
and CALIOP observations (see Figure 11b1-3). To highlight the aerosol layer, we normalized each vertical profile based on 440
the maximum (EC_max) and minimum extinction coefficients (EC_min) to remove the magnitude difference. The
normalized EC was calculated as (EC_model – EC_min)/(EC_max-EC_min). Over AMAZ and SHAF, only a few models
(ECHAM-HAM, ECHAM-SALSA, CAM5-ATRAS, GISS-OMA, and GISS-MATRIX) could capture the peak aerosol
extinction at 2–4 km, whereas other models tended to show the strongest extinction at lower altitudes or the surface. Over
BONA, the observed extinction peaked at ~ 4 km, but no models were found with a similar profile. Compared with PEH 445
from CALIOP, the simulated BBA plume tended to be too low for all the models. A similar underestimation was also
reported elsewhere for AeroCom models with the bias being attributed to wet deposition being too strong in the models
(Koffi et al., 2016).
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
16
5.2 Diversity of MEC
Modeled MECs are affected by several factors (e.g., particle size, complex refractive index, and hygroscopicity). As BBA is 450
dominated by OA and very similar refractive indices are used in models (see Appendix), the choice of refractive indices is
not discussed. Here we mainly examined the impacts of particle size and hygroscopicity.
Because particle size information was missing for the AeroCom models, we used modeled AE as it is an indicator of
particle size (Shuster et al., 2006). Figure 12a shows the dependence of modeled MECs on AEs. The modeled AE varied
from 0.21 to 2.2. Ambient particle size is the result of emitted particle sizes and particle processing after emission (see 455
Appendix). Among all models, the lowest AE was found in INCA from CTRL16 due to the large contribution from coarse-
mode SS, which also led to lower extinctions because of the lower MECs for SS than OA. When omitting INCA, a
significant negative correlation was found between MECs and AEs (r = -0.58), although there were large variations between
models. The correlation for CTRL19 models that were driven by the same emission inventory was even stronger (r = -0.73).
The negative correlation suggested that a larger size (smaller AE) resulted in a stronger extinction per mass unit for typical 460
BB aerosols, which agreed well with the observations (Laing et al., 2016; Kleinman et al., 2020). This can also be explained
by the Mie-scattering theory. In Figure 12a, we show the relation between MECs and AEs for pure OA aerosol based on the
Mie-scattering theory. We assumed that the radius of dry OA particle ranged from 0.02 to 0.5 μm. The lower edge of the
radius corresponded to the smallest emitted particle assumed in all the models examined (see Appendix) and the latter
indicated the upper edge of accumulation mode (Tegen et al., 2019). Hygroscopic growth was considered to occur based on 465
the Kappa-Köhler theory under an RH of 50% and the kappa value for OA was set as 0.06 referring to Zhang et al (2012). A
series of sensitive tests suggested that hygroscopic growth did not affect the calculation much. The refractive index was set
to 1.53-0.0055i as assumed in most models. The extinction cross-section was retrieved from the look-up table from
ECHAM-HAM, based on which MECs and AEs were calculated. The calculated MEC increased with increasing particle
size (decreasing AE), which agreed with the modeled relations. 470
The negative correlation between AEs and MECs suggested the possibility of evaluating and subsequently constraining
MEC by AE. In Figure 12b we validated modeled AEs against observations from POLDER-GRASP for the fire season.
Because most of the AE data for CTRL19 models had a monthly resolution, we collocated all the model data with
observations on monthly basis. Compared with POLDER-GRASP observations, the majority of models tended to
overestimate AEs by up to 0.85. BONA had the highest overestimation on average (0.27), followed by AMAZ and SHAF. 475
Given the previous analysis of the MEC dependence on AE, the underestimation of particle size may have led to a
considerable underestimation of MECs and thereby AOD. Similar to the impacts of precipitation, no strong correlation was
found for AOD biases with AE biases, which was largely due to the interaction of multiple factors and non-linear model
response.
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
17
The hygroscopicity was quantified as the extinction enhancement factor (EEF) which was defined as the ratio of AOD at 480
the ambient relative humidity (RH) to AOD at zero RH (dry AOD). Figure 13 illustrates the relation between MECs and
EEFs in models. For most models, a small EEF (< 2) was observed, and we did not observe clear patterns between EEFs and
MECs, probably because the hygroscopic growth was not significant given the low hydrophilicity of OA and the dry air
condition during fire seasons. For comparison, we also collected the EEF for BBA from in-situ measurements (see Table 3).
The observed EEF ranges from 1 to 2.1, which was consistent with the values calculated by most models. Note that the ‘dry’ 485
condition (RH = 20%−30%) in field measurements was different from that in models which assumed an RH of 0. The
ambient RH in models over the three regions (47%−75%) also differed a lot from the RH in observations. However, given
that BBA was dominated by OA and BC which were moderately hygroscopic or hydrophobic, the difference in the reference
RH to calculate EEF might have small impacts (Tito et al., 2016). Apart from these models, there were a few models that
showed pretty strong hygroscopic growth, accompanied by a positive correlation with MECs for each model. Since EEF was 490
related to the ambient RH and composition of hydrophilic aerosols (i.e., SU and SS), we further compared these values
between the two groups of models with either low or high hygroscopicity. Both the modeled RH (averaged for layers from
the surface to 650 hPa) and the percentage of hydrophilic aerosols show much smaller differences in the same fire regions
between the two model groups compared with the large disparities in EEF. This suggested that the difference of EEF in the
two model groups was linked to the BBA properties, which might result from the modeled particle size, mixing state, and 495
hygroscopicity parameterizations for OA (Burgos et al., 2020). For example, SPRINTARS assumed a similarly strong
hygroscopicity for BBA compared with SU (Takemura, 2005). In addition, we also found the modeled relations between
MEC and EEF were closely related to the treatment of ‘clear-sky/all-sky’ assumptions. For example, the clear-sky data from
GISS-OMA model showed similar EEFs to other models, whereas the all-sky data exhibited much higher EEFs. For those
models with higher hygroscopicity (i.e., GEOS-Chem, GISS-OMA, IMPACT, and SPRINTARS), the predicted MECs under 500
the same EEF varied substantially by a factor of 5 per model, suggesting that the modeled MEC diversity was controlled by
other factors. When all models were considered together, there was no clear pattern between EEFs and MECs found.
6. Conclusions
In this paper, we conducted a comprehensive evaluation and interpretation of AOD errors in AeroCom models over three
key BB regions. We first evaluated 14 satellite AOD datasets against AERONET and identified their errors. These errors in 505
satellite observations were then compared with model errors, with a much larger magnitude for the latter found in most
models. We noticed that such a small impact from different satellite products only applied for our validations over BB
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
18
regions during fire seasons. Specially, we found that the errors due to different satellite observations were comparable to the
model errors for the non-fire seasons over the three BB regions.
Detailed model validations against POLDER-GRASP observations suggested that most of the models still largely 510
underestimated AOD, especially when using the standard emission inventories (e.g., GFED3, CMIP6). We did not observe
significant improvements of modeled AOD in the latest experiment (CTRL19) compared with previous ones. The model
ensembles from the three AeroCom experiments exhibited a smaller inter-model spread of AOD correlation with
observations than AOD errors (e.g., CRMSE, NMB). Models seem to have a similar capability to model the spatiotemporal
variation of BBA, probably due to the similarity of input emissions as we found pretty strong correlations (~ 0.7) among the 515
emission inventories used by these models (see Appendix). Most of the diversity in model errors is due to a season-wide
bias. That said, temporal biases seem larger than spatial biases. We also provided evidence that AOD errors during the fire
season were dominated by BBA errors, with only a small contribution from the background. Based on BBE simulations, we
found negative biases could be reduced by scaling up BBA emissions. However, we showed that simulations with scaled
emissions did not thoroughly increase model performance. 520
We further analyzed the large diversity in fire AOD as resulting from emissions, lifetimes, and MECs which all exhibited
large diversities too. When all models were considered, we showed that the contributions of these three factors to the overall
AOD diversities were similar, though emissions exhibited slightly higher importance. In spite of the large inter-model
diversities, individual models show very similar lifetime and MEC over different BB regions, suggesting that basic model
assumptions underlie lifetime and MEC for each model. We suspect that relatively simple changes in these assumptions may 525
produce significant improvement in BBA simulations.
Modeled lifetime was correlated with modeled precipitation strength. Comparisons with observations suggested diverse
and region-specific precipitation errors. Modeled lifetime was also related to plume height which was found to be strongly
underestimated by models. We found MECs depended on how models simulate AE (or particle size). We further compared
modeled AE with POLDER-GRASP observations where general AE overestimations were found in most models. Most 530
models produced acceptable hygroscopicity compared with observations. These findings can provide useful information for
future model improvement and development.
There are several uncertainties in our evaluation and analysis. One is the uncertainties in POLDER-GRASP satellite
observation. Although we showed that satellite errors did not affect our evaluations very much, we still found that POLDER-
GRASP had un-ignorable retrieval errors over the focused regions (13%). However, the retrieval error was difficult to be 535
precisely defined due to the lack of sufficient samplings in SHAF and BONA by AERONET. On a global scale, POLDER-
GRASP was found to be superior to other satellite products used in this study. The other uncertainty stems from the
assumption of clear-sky conditions. As we evaluate model AOD against satellite data which are always clear-sky
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
19
observations, clear-sky model AOD should be used for comparison. However, models have very different treatments of the
‘clear-sky’ assumption (see Appendix). Although strict collocation can partly address this issue, uncertainties may still exist. 540
Such an issue should be investigated more in further model validations.
Data availability
All the model data can be accessed at AeroCom Wiki (https://aerocom.met.no). POLDER-GRASP dataset can be found at
https://download.grasp-cloud.com/download/polder/. All the other observations can be found in their references as listed.
The data processing in this work was done via CIS (http://www.cistools.net/). Codes to create individual figures can be 545
obtained from the corresponding author upon request (q.zhong@vu.nl).
Competing interests
The authors declare that they have no conflict of interest.
Acknowledgements 550
This work was financially supported by Netherlands Organization for Scientific Research (NWO; ALWGO.2018.052). K.T.
and S.E.B. acknowledge NASA MAP for support. Resources supporting this work were provided by the NASA High-End
Computing (HEC) Program through the NASA Center for Climate Simulation (NCCS) at Goddard Space Flight Center.
H.M. was supported by the Ministry of Education, Culture, Sports, Science and Technology of Japan and the Japan Society
for the Promotion of Science (MEXT/JSPS) KAKENHI Grant Numbers JP19H04253, JP19H05699, JP19KK0265, 555
JP20H00196, and JP20H00638, MEXT Arctic Challenge for Sustainability phase II (ArCS-II; JPMXD1420318865) project,
and the Environment Research and Technology Development Fund 2–2003 (JPMEERF20202003) of the Environmental
Restoration and Conservation Agency. We thank all the modelers that have submitted AeroCom model data used in this
work.
560
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
20
References
Adler, R. F., Huffman, G. J., Chang, A., Ferraro, R., Xie, P. P., Janowiak, J., Rudolf, B., Schneider, U., Curtis, S., Bolvin,
D., Gruber, A., Susskind, J., Arkin, P., and Nelkin, E.:The version-2 global precipitation climatology project (GPCP)
monthly precipitation analysis (1979–present). J. Hydrometeorol., 4, 1147-1167, https://doi.org/10.1175/1525-
7541(2003)004<1147:TVGPCP>2.0.CO;2, 2003. 565
Aiken, A. C., Decarlo, P. F., Kroll, J. H., Worsnop, D. R., Huffman, J. A., Docherty, K. S., Ulbrich, I. M., Mohr, C.,
Kimmel, J. R., Sueper, D., Sun, Y., Zhang, Q., Trimborn, A., Northway, M., Ziemann, P. J., Canagaratna, M. R., Onasch, T.
B., Rami Alfarra, M., Prevot, A. S. H., Dommen, J., Duplissy, J., Metzger, A., Baltensperger, U. and Jimenez, J. L.: O/C and
OM/OC ratios of primary, secondary, and ambient organic aerosols with high-resolution time-of-flight aerosol mass
spectrometry. Environ. Sci. Technol., 42, 4478-4485, https://doi.org/10.1021/es703009q, 2008. 570
Andela, N., van der Werf, G. R., Kaiser, J. W., van Leeuwen, T. T., Wooster, M. J., and Lehmann, C. E. R.: Biomass
burning fuel consumption dynamics in the tropics and subtropics assessed from satellite, Biogeosciences, 13, 3717-3734,
https://doi.org/10.5194/bg-13-3717-2016, 2016.
Andreae, M. O., and Rosenfeld, D.: Aerosol–cloud–precipitation interactions. Part 1. The nature and sources of cloud-active
aerosols, Earth Sci. Rev., 89, 13-41, https://doi.org/10.1016/j.earscirev.2008.03.001, 2008. 575
Andronache, C.: Estimated variability of below-cloud aerosol removal by rainfall for observed aerosol size distributions,
Atmos. Chem. Phys., 3, 131–143, https://doi.org/10.5194/acp-3-131-2003, 2003.
Balkanski, Y., Schulz, M., Claquin, T., Moulin, C., and Ginoux, P.: Global Emissions of Mineral Aerosol: Formulation and
Validation using Satellite Imagery, in: Advances in Global Change Research, Springer Netherlands, 239−267,
https://doi.org/10.1007/978-1-4020-2167-1_6, 2004. 580
Bauer, S. E., Bausch, A., Nazarenko, L., Tsigaridis, K., Xu, B. Q., Edwards, R., Bisiaux, M., and McConnell, J.: Historical
and future black carbon deposition on the three ice caps: Ice core measurements and model simulations from 1850 to 2100, J.
Geophys. Res. Atmos., 118, 7948-7961, https://doi.org/doi:10.1002/Jgrd.50612, 2013.
Bauer, S. E., Tsigaridis, K., Faluvegi, G., Kelley, M., Lo, K. K., Miller, R. L., Nazarenko, L., Schmidt, G. A., and Wu, J.:
Historical (1850–2014) Aerosol Evolution and Role on Climate Forcing Using the GISS ModelE2.1 Contribution to CMIP6, 585
J. Adv. Model. Earth Syst., 12, https://doi.org/10.1029/2019ms001978, 2020.
Bauer, S. E., Wright, D. L., Koch, D., Lewis, E. R., McGraw, R., Chang, L.-S., Schwartz, S. E., and Ruedy, R.: MATRIX
(Multiconfiguration Aerosol TRacker of mIXing state): an aerosol microphysical module for global atmospheric models,
Atmos. Chem. Phys., 8, 6003–6035, https://doi.org/10.5194/acp-8-6003-2008, 2008.
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
21
Bellouin, N., Mann, G. W., Woodhouse, M. T., Johnson, C., Carslaw, K. S., and Dalvi, M.: Impact of the modal aerosol 590
scheme GLOMAP-mode on aerosol forcing in the Hadley Centre Global Environmental Model, Atmos. Chem. Phys., 13,
3027-3044, https://doi.org/10.5194/acp-13-3027-2013, 2013.
Bey, I., Jacob, D. J., Yantosca, R. M., Logan, J. A., Field, B. D., Fiore, A. M., Li, Q., Liu, H. Y., Mickley, L. J., and Schultz,
M. G.: Global modeling of tropospheric chemistry with assimilated meteorology: Model description and evaluation, J.
Geophys. Res. Atmos., 106, 23073-23095, https://doi.org/10.1029/2001jd000807, 2001. 595
Bevan, S. L., North, P. R., Los, S. O., and Grey, W.M.: A global dataset of atmospheric aerosol optical depth and surface
reflectance from AATSR. Remote Sens. Environ., 116, 199-210, https://doi.org/10.1016/j.rse.2011.05.024, 2012.
Bian, H., Chin, M., Rodriguez, J. M., Yu, H., Penner, J. E., and Strahan, S.: Sensitivity of aerosol optical thickness and
aerosol direct radiative effect to relative humidity, Atmos. Chem. Phys., 9, 2375–2386, https://doi.org/10.5194/acp-9-2375-
2009, 2009. 600
Bond, T. C.: A technology-based global inventory of black and organic carbon emissions from combustion, J. Geophys.
Res., 109, https://doi.org/10.1029/2003jd003697, 2004.
Brito, J., Rizzo, L. V., Morgan, W. T., Coe, H., Johnson, B., Haywood, J., Longo, K., Freitas, S., Andreae, M. O., and
Artaxo, P.: Ground-based aerosol characterization during the South American Biomass Burning Analysis (SAMBBA) field
experiment, Atmos. Chem. Phys., 14, 12069-12083, https://doi.org/10.5194/acp-14-12069-2014, 2014. 605
Brown, H., Liu, X., Pokhrel, R., Murphy, S., Lu, Z., Saleh, R., Mielonen, T., Kokkola, H., Bergman, T., Myhre, G., Skeie, R.
B., Watson-Paris, D., Stier, P., Johnson, B., Bellouin, N., Schulz, M., Vakkari, V., Beukes, J. P., van Zyl, P. G., Liu, S., and
Chand, D.: Biomass burning aerosols in most climate models are too absorbing, Nat. Commun., 12, 277,
https://doi.org/10.1038/s41467-020-20482-9, 2021.
Burgos, M. A., Andrews, E., Titos, G., Benedetti, A., Bian, H., Buchard, V., Curci, G., Kipling, Z., Kirkevåg, A., Kokkola, 610
H., Laakso, A., Letertre-Danczak, J., Lund, M. T., Matsui, H., Myhre, G., Randles, C., Schulz, M., van Noije, T., Zhang, K.,
Alados-Arboledas, L., Baltensperger, U., Jefferson, A., Sherman, J., Sun, J., Weingartner, E., and Zieger, P.: A global
model–measurement evaluation of particle light scattering coefficients at elevated relative humidity, Atmos. Chem. Phys.,
20, 10231–10258, https://doi.org/10.5194/acp-20-10231-2020, 2020.
Cappa, C. D., Onasch, T. B., Massoli, P., Worsnop, D. R., Bates, T. S., Cross, E. S., Davidovits, P., Hakala, J., Hayden, K. 615
L., Jobson, B. T., Kolesar, K. R., Lack, D. A., Lerner, B. M., Li, S., Mellon, D., Nuaaman, I., Olfert, J. S., Petäjä, T., Quinn,
P. K., Song, C., Subramanian, R., Williams, E. J., and Zaveri, R. A.: Radiative Absorption Enhancements Due to the Mixing
State of Atmospheric Black Carbon, Science, 337(6098), 1078-1081, https://doi.org/10.1126/science.1223447, 2012.
Chen, C., Dubovik, O., Fuertes, D., Litvinov, P., Lapyonok, T., Lopatin, A., Ducos, F., Derimian, Y., Herman, M., Tanré,
D., Remer, L. A., Lyapustin, A., Sayer, A. M., Levy, R. C., Hsu, N. C., Descloitres, J., Li, L., Torres, B., Karol, Y., Herrera, 620
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
22
M., Herreras, M., Aspetsberger, M., Wanzenboeck, M., Bindreiter, L., Marth, D., Hangler, A., and Federspiel, C.: Validation
of GRASP algorithm product from POLDER/PARASOL data and assessment of multi-angular polarimetry potential for
aerosol monitoring, Earth Syst. Sci. Data, 12, 3573-3620, https://doi.org/10.5194/essd-12-3573-2020, 2020.
Chin, M., Diehl, T., Dubovik, O., Eck, T. F., Holben, B. N., Sinyuk, A., and Streets, D. G.: Light absorption by pollution,
dust, and biomass burning aerosols: a global model study and evaluation with AERONET measurements, Ann. Geophys., 625
27, 3439–3464, https://doi.org/10.5194/angeo-27-3439-2009, 2009.
Colarco, P., da Silva, A., Chin, M., and Diehl, T.: Online simulations of global aerosol distributions in the NASA GEOS-4
model and comparisons to satellite and ground-based aerosol optical depth, J. Geophys. Res., 115,
https://doi.org/10.1029/2009jd012820, 2010.
Dong, X., Fu, J. S., Huang, K., Zhu, Q., and Tipton, M.: Regional Climate Effects of Biomass Burning and Dust in East 630
Asia: Evidence From Modeling and Observation, Geophys. Res. Lett., 46, 11490-11499,
https://doi.org/10.1029/2019gl083894, 2019.
Donner, L. J., Wyman, B. L., Hemler, R. S., Horowitz, L. W., Ming, Y., Zhao, M., Golaz, J.-C., Ginoux, P., Lin, S.-J.,
Schwarzkopf, M. D., Austin, J., Alaka, G., Cooke, W. F., Delworth, T. L., Freidenreich, S. M., Gordon, C. T., Griffies, S.
M., Held, I. M., Hurlin, W. J., Klein, S. A., Knutson, T. R., Langenhorst, A. R., Lee, H.-C., Lin, Y., Magi, B. I., Malyshev, 635
S. L., Milly, P. C. D., Naik, V., Nath, M. J., Pincus, R., Ploshay, J. J., Ramaswamy, V., Seman, C. J., Shevliakova, E.,
Sirutis, J. J., Stern, W. F., Stouffer, R. J., Wilson, R. J., Winton, M., Wittenberg, A. T., and Zeng, F.: The Dynamical Core,
Physical Parameterizations, and Basic Simulation Characteristics of the Atmospheric Component AM3 of the GFDL Global
Coupled Model CM3, J. Climate, 24, 3484–3519, https://doi.org/10.1175/2011JCLI3955.1, 2011.
Dubovik, O., Herman, M., Holdak, A., Lapyonok, T., Tanré, D., Deuzé, J. L., Ducos, F., Sinyuk, A., and Lopatin, A.: 640
Statistically optimized inversion algorithm for enhanced retrieval of aerosol properties from spectral multi-angle polarimetric
satellite observations, Atmos. Meas. Tech., 4, 975-1018, https://doi.org/10.5194/amt-4-975-2011, 2011.
Dubovik, O., Smirnov, A., Holben, B. N., King, M. D., Kaufman, Y. J., Eck, T. F., and Slutsker, I.: Accuracy assessments of
aerosol optical properties retrieved from Aerosol Robotic Network (AERONET) Sun and sky radiance measurements, J.
Geophys. Res. Atmos., 105, 9791-9806, https://doi.org/10.1029/2000jd900040, 2000. 645
Dumka, U.C., Kaskaoutis, D.G., Sagar, R., Chen, J., Singh, N. and Tiwari, S.: First results from light scattering enhancement
factor over central Indian Himalayas during GVAX campaign, Sci. Total Environ., 605, 124-138,
https://doi.org/10.1016/j.scitotenv.2017.06.138, 2017.
EMEP status report 1/2012, Chapter 10 https://emep.int/publ/reports/2012/status_report_1_2012.pdf.
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
23
Engelhart, G. J., Hennigan, C. J., Miracolo, M. A., Robinson, A. L., and Pandis, S. N.: Cloud condensation nuclei activity of 650
fresh primary and aged biomass burning aerosol, Atmos. Chem. Phys., 12, 7285-7293, https://doi.org/10.5194/acp-12-7285-
2012, 2012.
Falah, S., Mhawish, A., Sorek-Hamer, M., Lyapustin, A. I., Kloog, I., Banerjee, T., Kizel, F., and Broday, D. M.: Impact of
environmental attributes on the uncertainty in MAIAC/MODIS AOD retrievals: A comparative analysis, Atmos. Environ.,
262, https://doi.org/10.1016/j.atmosenv.2021.118659, 2021. 655
Gras, J.L., Jensen, J.B., Okada, K., Ikegami, M., Zaizen, Y., Makino, Y.: Some optical properties of smoke aerosol in
Indonesia and Tropical Australia. Geophys. Res. Lett. 26, 1393-1396, https://doi.org/10.1029/1999GL900275, 1999.
Gliß, J., Mortier, A., Schulz, M., Andrews, E., Balkanski, Y., Bauer, S. E., Benedictow, A. M. K., Bian, H., Checa-Garcia,
R., Chin, M., Ginoux, P., Griesfeller, J. J., Heckel, A., Kipling, Z., Kirkevåg, A., Kokkola, H., Laj, P., Le Sager, P., Lund,
M. T., Lund Myhre, C., Matsui, H., Myhre, G., Neubauer, D., van Noije, T., North, P., Olivié, D. J. L., Rémy, S., Sogacheva, 660
L., Takemura, T., Tsigaridis, K., and Tsyro, S. G.: AeroCom phase III multi-model evaluation of the aerosol life cycle and
optical properties using ground- and space-based remote sensing as well as surface in situ observations, Atmos. Chem. Phys.,
21, 87-128, https://doi.org/10.5194/acp-21-87-2021, 2021.
Hsu, N. C., Jeong, M. J., Bettenhausen, C., Sayer, A. M., Hansell, R., Seftor, C. S., Huang, J., and Tsay, S. C.: Enhanced
Deep Blue aerosol retrieval algorithm: The second generation, J. Geophys. Res. Atmos., 118, 9296-9315, 665
https://doi.org/10.1002/jgrd.50712, 2013.
Hsu, N. C., Lee, J., Sayer, A. M., Kim, W., Bettenhausen, C., and Tsay, S. C.: VIIRS Deep Blue Aerosol Products Over
Land: Extending the EOS Long‐Term Aerosol Data Records, J. Geophys. Res. Atmos., 124, 4026-4053,
https://doi.org/10.1029/2018jd029688, 2019.
Huang, Y., Shen, H., Chen, Y., Zhong, Q., Chen, H., Wang, R., Shen, G., Liu, J., Li, B., and Tao, S.: Global organic carbon 670
emissions from primary sources from 1960 to 2009, Atmos. Environ., 122, 505-512,
https://doi.org/10.1016/j.atmosenv.2015.10.017, 2015.
Ichoku, C., and Ellison, L.: Global top-down smoke-aerosol emissions estimation using satellite fire radiative power
measurements, Atmos. Chem. Phys., 14, 6643-6667, https://doi.org/10.5194/acp-14-6643-2014, 2014.
Jahl, L.G., Brubaker, T.A., Polen, M.J., Jahn, L.G., Cain, K.P., Bowers, B.B., Fahy, W.D., Graves, S., and Sullivan, R.C.: 675
Atmospheric aging enhances the ice nucleation ability of biomass-burning aerosol. Sci. Adv., 7, eabd3440,
https://doi.org/10.1126/sciadv.abd3440, 2021.
Johnson, B. T., Haywood, J. M., Langridge, J. M., Darbyshire, E., Morgan, W. T., Szpek, K., Brooke, J. K., Marenco, F.,
Coe, H., Artaxo, P., Longo, K. M., Mulcahy, J. P., Mann, G. W., Dalvi, M., and Bellouin, N.: Evaluation of biomass burning
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
24
aerosols in the HadGEM3 climate model with observations from the SAMBBA field campaign, Atmos. Chem. Phys., 16, 680
14657-14685, https://doi.org/10.5194/acp-16-14657-2016, 2016.
Johnston, F. H., Henderson, S. B., Chen, Y., Randerson, J. T., Marlier, M., DeFries, R. S., Kinney, P., Bowman, D. M. J. S.,
and Brauer, M.: Estimated global mortality attributable to smoke from landscape fires. Environ. Health Perspect. 120, 695-
701, https://doi.org/10.1289/ehp.1104422, 2016.
Jung, J., and Kim, Y. J.: Tracking sources of severe haze episodes and their physicochemical and hygroscopicproperties 685
under Asian continental outflow: Long‐range transport pollution, postharvest biomass burning, and Asian dust, J. Geophys.
Res.,116, D02206, https://doi.org/10.1029/2010JD014555, 2011.
Kaiser, J. W., Heil, A., Andreae, M. O., Benedetti, A., Chubarova, N., Jones, L., Morcrette, J. J., Razinger, M., Schultz, M.
G., Suttie, M., and van der Werf, G. R.: Biomass burning emissions estimated with a global fire assimilation system based on
observed fire radiative power, Biogeosciences, 9, 527-554, https://doi.org/10.5194/bg-9-527-2012, 2012. 690
Kirkevåg, A., Grini, A., Olivié, D., Seland, Ø., Alterskjær, K., Hummel, M., Karset, I. H. H., Lewinschal, A., Liu, X.,
Makkonen, R., Bethke, I., Griesfeller, J., Schulz, M., and Iversen, T.: A production-tagged aerosol module for Earth system
models, OsloAero5.3 – extensions and updates for CAM5.3-Oslo, Geosci. Model Dev., 11, 3945-3982,
https://doi.org/10.5194/gmd-11-3945-2018, 2018.
Kleinman, L. I., Sedlacek Iii, A. J., Adachi, K., Buseck, P. R., Collier, S., Dubey, M. K., Hodshire, A. L., Lewis, E., Onasch, 695
T. B., Pierce, J. R., Shilling, J., Springston, S. R., Wang, J., Zhang, Q., Zhou, S., and Yokelson, R. J.: Rapid evolution of
aerosol particles and their optical properties downwind of wildfires in the western US, Atmos. Chem. Phys., 20, 13319-
13341, https://doi.org/10.5194/acp-20-13319-2020, 2020.
Koffi, B., Schulz, M., Breon, F. M., Dentener, F., Steensen, B. M., Griesfeller, J., Winker, D., Balkanski, Y., Bauer, S. E.,
Bellouin, N., Berntsen, T., Bian, H., Chin, M., Diehl, T., Easter, R., Ghan, S., Hauglustaine, D. A., Iversen, T., Kirkevag, A., 700
Liu, X., Lohmann, U., Myhre, G., Rasch, P., Seland, O., Skeie, R. B., Steenrod, S. D., Stier, P., Tackett, J., Takemura, T.,
Tsigaridis, K., Vuolo, M. R., Yoon, J., and Zhang, K.: Evaluation of the aerosol vertical distribution in global aerosol models
through comparison against CALIOP measurements: AeroCom phase II results, J. Geophys. Res. Atmos., 121, 7254-7283,
https://doi.org/10.1002/2015JD024639, 2016.
Kokkola, H., Kühn, T., Laakso, A., Bergman, T., Lehtinen, K. E. J., Mielonen, T., Arola, A., Stadtler, S., Korhonen, H., 705
Ferrachat, S., Lohmann, U., Neubauer, D., Tegen, I., Siegenthaler-Le Drian, C., Schultz, M. G., Bey, I., Stier, P., Daskalakis,
N., Heald, C. L., and Romakkaniemi, S.: SALSA2.0: The sectional aerosol module of the aerosol–chemistry–climate model
ECHAM6.3.0-HAM2.3-MOZ1.0, Geosci. Model Dev., 11, 3833-3863, https://doi.org/10.5194/gmd-11-3833-2018, 2018.
Kotchenruther, R.A., Hobbs, P.V.: Humidification factors of aerosols from biomass burning in Brazil. J. Geophys. Res. 103,
32081-32089, https://doi.org/10.1029/98JD00340, 1998. 710
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
25
Laing, J. R., Jaffe, D. A., and Hee, J. R.: Physical and optical properties of aged biomass burning aerosol from wildfires in
Siberia and the Western USA at the Mt. Bachelor Observatory, Atmos. Chem. Phys., 16, 15185-15197,
https://doi.org/10.5194/acp-16-15185-2016, 2016.
Lelieveld, J., Evans, J. S., Fnais, M., Giannadaki, D., and Pozzer, A.: The contribution of outdoor air pollution sources to
premature mortality on a global scale, Nature, 525, 367-371, https://doi.org/10.1038/nature15371, 2015. 715
Li, Z., Zhao, X., Kahn, R., Mishchenko, M., Remer, L., Lee, K.-H., Wang, M., Laszlo, I., Nakajima, T., and Maring, H.:
Uncertainties in satellite remote sensing of aerosols and impact on monitoring its long-term trend: a review and perspective,
Ann. Geophys., 27, 2755–2770, https://doi.org/10.5194/angeo-27-2755-2009, 2009.
Lin, N. H., Sayer, A. M., Wang, S. H., Loftus, A. M., Hsiao, T. C., Sheu, G. R., Hsu, N. C., Tsay, S. C., and Chantara, S.:
Interactions between biomass-burning aerosols and clouds over Southeast Asia: current status, challenges, and perspectives, 720
Environ. Pollut., 195, 292-307, https://doi.org/10.1016/j.envpol.2014.06.036, 2014.
Lipponen, A., Mielonen, T., Pitkänen, M. R. A., Levy, R. C., Sawyer, V. R., Romakkaniemi, S., Kolehmainen, V., and
Arola, A.: Bayesian aerosol retrieval algorithm for MODIS AOD retrieval over land, Atmos. Meas. Tech., 11, 1529-1547,
https://doi.org/10.5194/amt-11-1529-2018, 2018.
Liu, X., Easter, R. C., Ghan, S. J., Zaveri, R., Rasch, P., Shi, X., Lamarque, J. F., Gettelman, A., Morrison, H., Vitt, F., 725
Conley, A., Park, S., Neale, R., Hannay, C., Ekman, A. M. L., Hess, P., Mahowald, N., Collins, W., Iacono, M. J.,
Bretherton, C. S., Flanner, M. G., and Mitchell, D.: Toward a minimal representation of aerosols in climate models:
description and evaluation in the Community Atmosphere Model CAM5, Geosci. Model Dev., 5, 709-739,
https://doi.org/10.5194/gmd-5-709-2012, 2012.
Liu, X., Penner, J. E., and Herzog, M.: Global modeling of aerosol dynamics: Model description, evaluation, andinteractions 730
between sulfate and nonsulfate aerosols, J. Geophys. Res., 110, D18206, https://doi.org/10.1029/2004JD005674, 2005.
Lyapustin, A., Wang, Y., Korkin, S., and Huang, D.: MODIS Collection 6 MAIAC algorithm, Atmos. Meas. Tech., 11,
5741-5765, https://doi.org/10.5194/amt-11-5741-2018, 2018.
Magi, B.I., and Hobbs, P.V.: Effects of humidity on aerosols in southern Africa during the biomass burning season. J.
Geophys. Res., 108, 8495, http://dx.doi.org/10.1029/2002JD002144, 2003. 735
Mallet, M., Nabat, P., Johnson, B., Michou, M., Haywood, J.M., Chen, C. and Dubovik, O.: Climate models generally
underrepresent the warming by Central Africa biomass-burning aerosols over the Southeast Atlantic. Sci. Adv., 7, eabg9998,
https://doi.org/10.1126/sciadv.abg9998, 2021.
Martins, J. A., Silva Dias, M. A. F., and Gonçalves, F. L. T.: Impact of biomass burning aerosols on precipitation in the
Amazon: A modeling case study, J. Geophys. Res., 114, https://doi.org/10.1029/2007jd009587, 2009. 740
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
26
Matsui, H.: Development of a global aerosol model using a two-dimensional sectional method: 1. Model design, J. Adv.
Model. Earth Syst., 9, 1921-1947, https://doi.org/10.1002/2017MS000936, 2017.
Matsui, H., and Mahowald, N.: Development of a global aerosol model using a two-dimensional sectional method: 2.
Evaluation and sensitivity simulations, J. Adv. Model. Earth Syst., 9, 1887-1920, https://doi.org/10.1002/2017ms000937,
2017. 745
Mulcahy, J. P., Johnson, C., Jones, C. G., Povey, A. C., Scott, C. E., Sellar, A., Turnock, S. T., Woodhouse, M. T., Abraham,
N. L., Andrews, M. B., Bellouin, N., Browse, J., Carslaw, K. S., Dalvi, M., Folberth, G. A., Glover, M., Grosvenor, D. P.,
Hardacre, C., Hill, R., Johnson, B., Jones, A., Kipling, Z., Mann, G., Mollard, J., O'Connor, F. M., Palmiéri, J., Reddington,
C., Rumbold, S. T., Richardson, M., Schutgens, N. A. J., Stier, P., Stringer, M., Tang, Y., Walton, J., Woodward, S., and
Yool, A.: Description and evaluation of aerosol in UKESM1 and HadGEM3-GC3.1 CMIP6 historical simulations, Geosci. 750
Model Dev., 13, 6383-6423, https://doi.org/10.5194/gmd-13-6383-2020, 2020.
Myhre, G., Bellouin, N., Berglen, T. F., Berntsen, T. K., Boucher, O., Grini, A., Isaksen, I. S. A., Johnsrud, M., Mishchenko,
M. I., Stordal, F., and Tandre, D.: Comparison of the radiative properties and direct radiative effect of aerosols from a global
aerosol model and remote sensing data over ocean, Tellus B, 59, 115–129, https://doi.org/10.1111/j.1600-
0889.2006.00226.x, 2007. 755
Myhre, G., Berglen, T. F., Johnsrud, M., Hoyle, C. R., Berntsen, T. K., Christopher, S. A., Fahey, D. W., Isaksen, I. S. A.,
Jones, T. A., Kahn, R. A., Loeb, N., Quinn, P., Remer, L., Schwarz, J. P., and Yttri, K. E.: Modeled radiative forcing of the
direct aerosol effect with multi-observation evaluation, Atmos. Chem. Phys., 9, 1365–1392, https://doi.org/10.5194/acp-9-
1365-2009, 2009.
Myhre, G., Berntsen, T. K., Haywood, J. M., Sundet, J. K., Holben, B. N., Johnsrud, M., and Stordal, F.: Modeling the solar 760
radiative impact of aerosols from biomass burning during the Southern African Regional Science Initiative (SAFARI-2000)
experiment, J. Geophys. Res. Atmos., 108, n/a-n/a, https://doi.org/10.1029/2002jd002313, 2003.
North, P. R. J.: Estimation of aerosol opacity and land surface bidirectional reflectance from ATSR-2 dual-angle imagery:
Operational method and validation, J. Geophys. Res., 107, https://doi.org/10.1029/2000jd000207, 2002.
North, P. R. J., Briggs, S. A., Plummer, S. E., and Settle, J. J.: Retrieval of Land Surface Bidirectional Reflectance and 765
Aerosol Opacity from ATSR-2 Multiangle Imagery, IEEE T. Geosci. Remote Sens., 37, 526–537, 1999.
Petrenko, M., Kahn, R., Chin, M., Soja, A., Kucsera, T., and Harshvardhan: The use of satellite-measured aerosol optical
depth to constrain biomass burning emissions source strength in the global model GOCART, J. Geophys. Res. Atmos., 117,
D18212, https://doi.org/10.1029/2012jd017870, 2012.
Pistone, K., Redemann, J., Doherty, S., Zuidema, P., Burton, S., Cairns, B., Cochrane, S., Ferrare, R., Flynn, C., Freitag, S., 770
Howell, S. G., Kacenelenbogen, M., LeBlanc, S., Liu, X., Schmidt, K. S., Sedlacek III, A. J., Segal-Rozenhaimer, M.,
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
27
Shinozuka, Y., Stamnes, S., van Diedenhoven, B., Van Harten, G., and Xu, F.: Intercomparison of biomass burning aerosol
optical properties from in situ and remote-sensing instruments in ORACLES-2016, Atmos. Chem. Phys., 19, 9181–9208,
https://doi.org/10.5194/acp-19-9181-2019, 2019.
Randerson, J. T., Chen, Y., van der Werf, G. R., Rogers, B. M., and Morton, D. C.: Global burned area and biomass burning 775
emissions from small fires, J. Geophys. Res.: Biogeosci., 117, G04012, https://doi.org/10.1029/2012jg002128, 2012.
Reddington, C. L., Spracklen, D. V., Artaxo, P., Ridley, D. A., Rizzo, L. V., and Arana, A.: Analysis of particulate emissions
from tropical biomass burning using a global aerosol model and long-term surface observations, Atmos. Chem. Phys., 16,
11083-11106, https://doi.org/10.5194/acp-16-11083-2016, 2016.
Reddington, C. L., Morgan, W. T., Darbyshire, E., Brito, J., Coe, H., Artaxo, P., Scott, C. E., Marsham, J., and Spracklen, D. 780
V.: Biomass burning aerosol over the Amazon: analysis of aircraft, surface and satellite observations using a global aerosol
model, Atmos. Chem. Phys., 19, 9125-9152, https://doi.org/10.5194/acp-19-9125-2019, 2019.
Reid, J. S., Koppmann, R., Eck, T. F., and Eleuterio, D. P.: A review of biomass burning emissions part II: intensive physical
properties of biomass burning particles, Atmos. Chem. Phys., 5, 799–825, https://doi.org/10.5194/acp-5-799-2005, 2005.
Remer, L., Kaufman, Y., Tanre, D., Mattoo, S., Chu, D., Martins, J., Li, R.-R., Ichoku, C., Levy, R., Kleidman, R., Eck, T., 785
Vermote, E., and Holben, B.: The MODIS Aerosol Algorithm, Products, and Validation, J. Atmos. Sci., 62, 947–973,
https://doi.org/10.1175/JAS3385.1, 2005.
Rémy, S., Kipling, Z., Flemming, J., Boucher, O., Nabat, P., Michou, M., Bozzo, A., Ades, M., Huijnen, V., Benedetti, A.,
Engelen, R., Peuch, V.-H., and Morcrette, J.-J.: Description and evaluation of the tropospheric aerosol scheme in the
European Centre for Medium-Range Weather Forecasts (ECMWF) Integrated Forecasting System (IFS-AER, cycle 45R1), 790
Geosci. Model Dev., 12, 4627-4659, https://doi.org/10.5194/gmd-12-4627-2019, 2019.
Sayer, A. M., Hsu, N. C., Lee, J., Carletta, N., Chen, S. H., and Smirnov, A.: Evaluation of NASA Deep Blue/SOAR aerosol
retrieval algorithms applied to AVHRR measurements, J. Geophys. Res. Atmos., 122, 9945-9967,
https://doi.org/10.1002/2017JD026934, 2017.
Sayer, A. M., Hsu, N. C., Lee, J., Kim, W. V., and Dutcher, S. T.: Validation, Stability, and Consistency of MODIS 795
Collection 6.1 and VIIRS Version 1 Deep Blue Aerosol Data Over Land, J. Geophys. Res. Atmos., 124, 4658-4688,
https://doi.org/10.1029/2018jd029598, 2019.
Schill, G. P., Froyd, K. D., Bian, H., Kupc, A., Williamson, C., Brock, C. A., Ray, E., Hornbrook, R. S., Hills, A. J., Apel, E.
C., Chin, M., Colarco, P. R., and Murphy, D. M.: Widespread biomass burning smoke throughout the remote troposphere,
Nat. Geosci., 13, 422-427, https://doi.org/10.1038/s41561-020-0586-1, 2020. 800
Schulz, M., Cozic, A., and Szopa, S.: LMDzT-INCA dust forecast model developments and associated validation efforts,
IOP Conference Series: Earth and Environmental Science, 7, 12014, https://doi.org/10.1088/1755-1307/7/1/012014, 2009.
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
28
Schuster, G. L., Dubovik, O., and Holben, B. N.: Angstrom exponent and bimodal aerosol size distributions. J. Geophys.
Res. Atmos., 111, D7, https://doi.org/10.1029/2005JD006328, 2006.
Schutgens, N., Dubovik, O., Hasekamp, O., Torres, O., Jethva, H., Leonard, P. J. T., Litvinov, P., Redemann, J., Shinozuka, 805
Y., de Leeuw, G., Kinne, S., Popp, T., Schulz, M., and Stier, P.: AEROCOM and AEROSAT AAOD and SSA study – Part
1: Evaluation and intercomparison of satellite measurements, Atmos. Chem. Phys., 21, 6895–6917,
https://doi.org/10.5194/acp-21-6895-2021, 2021.
Schutgens, N., Gryspeerdt, E., Weigum, N., Tsyro, S., Goto, D., Schulz, M., and Stier, P.: Will a perfect model agree with
perfect observations? The impact of spatial sampling, Atmos. Chem. Phys., 16, 6335-6353, https://doi.org/10.5194/acp-16-810
6335-2016, 2016a.
Schutgens, N. A. J., Partridge, D. G., and Stier, P.: The importance of temporal collocation for the evaluation of aerosol
models with observations, Atmos. Chem. Phys., 16, 1065–1079, https://doi.org/10.5194/acp-16-1065-2016, 2016b.
Schutgens, N., Sayer, A. M., Heckel, A., Hsu, C., Jethva, H., de Leeuw, G., Leonard, P. J. T., Levy, R. C., Lipponen, A.,
Lyapustin, A., North, P., Popp, T., Poulsen, C., Sawyer, V., Sogacheva, L., Thomas, G., Torres, O., Wang, Y., Kinne, S., 815
Schulz, M., and Stier, P.: An AeroCom–AeroSat study: intercomparison of satellite AOD datasets for aerosol model
evaluation, Atmos. Chem. Phys., 20, 12431-12457, https://doi.org/10.5194/acp-20-12431-2020, 2020.
Seinfeld, J.H. and Pandis, S.N.: Atmospheric Chemistry and Physics: From Air Pollution to Climate Change. 2nd Edition,
John Wiley & Sons, New York, 2006.
Seland, Ø., Bentsen, M., Olivié, D., Toniazzo, T., Gjermundsen, A., Graff, L. S., Debernard, J. B., Gupta, A. K., He, Y.-C., 820
Kirkevåg, A., Schwinger, J., Tjiputra, J., Aas, K. S., Bethke, I., Fan, Y., Griesfeller, J., Grini, A., Guo, C., Ilicak, M., Karset,
I. H. H., Landgren, O., Liakka, J., Moseid, K. O., Nummelin, A., Spensberger, C., Tang, H., Zhang, Z., Heinze, C., Iversen,
T., and Schulz, M.: Overview of the Norwegian Earth System Model (NorESM2) and key climate response of CMIP6
DECK, historical, and scenario simulations, Geosci. Model Dev., 13, 6165-6200, https://doi.org/10.5194/gmd-13-6165-
2020, 2020. 825
Sheridan, P. J., Jefferson, A., and Ogren, J. A.: Spatial variability of submicrometer aerosol radiative properties over the
Indian Ocean during INDOEX, J. Geophys. Res., 107, 8011, https://doi.org/10.1029/2000JD000166, 2002.
Smirnov, A., Holben, B. N., Giles, D. M., Slutsker, I., O'Neill, N. T., Eck, T. F., Macke, A., Croot, P., Courcoux, Y.,
Sakerin, S. M., Smyth, T. J., Zielinski, T., Zibordi, G., Goes, J. I., Harvey, M. J., Quinn, P. K., Nelson, N. B., Radionov, V.
F., Duarte, C. M., Losno, R., Sciare, J., Voss, K. J., Kinne, S., Nalli, N. R., Joseph, E., Krishna Moorthy, K., Covert, D. S., 830
Gulev, S. K., Milinevsky, G., Larouche, P., Belanger, S., Horne, E., Chin, M., Remer, L. A., Kahn, R. A., Reid, J. S., Schulz,
M., Heald, C. L., Zhang, J., Lapina, K., Kleidman, R. G., Griesfeller, J., Gaitley, B. J., Tan, Q., and Diehl, T. L.: Maritime
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
29
aerosol network as a component of AERONET – first results and comparison with global aerosol models and satellite
retrievals, Atmos. Meas. Tech., 4, 583-597, https://doi.org/10.5194/amt-4-583-2011, 2011.
Sogacheva, L., Kolmonen, P., Virtanen, T. H., Rodriguez, E., Saponaro, G., and de Leeuw, G.: Post-processing to remove 835
residual clouds from aerosol optical depth retrieved using the Advanced Along Track Scanning Radiometer, Atmos. Meas.
Tech., 10, 491-505, https://doi.org/10.5194/amt-10-491-2017, 2017.
Stockwell, C. E., Veres, P. R., Williams, J., and Yokelson, R. J.: Characterization of biomass burning emissions from
cooking fires, peat, crop residue, and other fuels with high-resolution proton-transfer-reaction time-of-flight mass
spectrometry, Atmos. Chem. Phys., 15, 845-865, https://doi.org/10.5194/acp-15-845-2015, 2015. 840
Takemura, T.: Simulation of climate response to aerosol direct and indirect effects with aerosol transport-radiation model, J.
Geophys. Res., 110, https://doi.org/10.1029/2004jd005029, 2005.
Taylor, K. E.: Summarizing multiple aspects of model performance in a single diagram, J. Geophys. Res. Atmos., 106, 7183-
7192, https://doi.org/10.1029/2000jd900719, 2001.
Titos, G., Cazorla, A., Zieger, P., Andrews, E., Lyamani, H., Granados-Muñoz, M., Olmo, F., and Alados-Arboledas, L.: 845
Effect of hygroscopic growth on the aerosol light scattering coefficient: A review of measurements, techniques and error
sources, Atmos. Environ., 141, 494-507, https://doi.org/10.1016/j.atmosenv.2016.07.021, 2016.
Tegen, I., Neubauer, D., Ferrachat, S., Siegenthaler-Le Drian, C., Bey, I., Schutgens, N., Stier, P., Watson-Parris, D.,
Stanelle, T., Schmidt, H., Rast, S., Kokkola, H., Schultz, M., Schroeder, S., Daskalakis, N., Barthel, S., Heinold, B., and
Lohmann, U.: The global aerosol–climate model ECHAM6.3–HAM2.3 – Part 1: Aerosol evaluation, Geosci. Model Dev., 850
12, 1643–1677, https://doi.org/10.5194/gmd-12-1643-2019, 2019.
Thomas, G. E., Carboni, E., Sayer, A. M., Poulsen, C. A., Siddans, R., and Grainger, R. G.: Oxford-RAL Aerosol and Cloud
(ORAC): aerosol retrievals from satellite radiometers, in: Satellite remote sensing over land, edited by: Kokhanovsky, A. and
de Leeuw, G., 193–224, Springer, Chichester, UK, 2009.
Tiitta, P., Vakkari, V., Croteau, P., Beukes, J. P., van Zyl, P. G., Josipovic, M., Venter, A. D., Jaars, K., Pienaar, J. J., Ng, N. 855
L., Canagaratna, M. R., Jayne, J. T., Kerminen, V. M., Kokkola, H., Kulmala, M., Laaksonen, A., Worsnop, D. R., and
Laakso, L.: Chemical composition, main sources and temporal variability of PM1 aerosols in southern African grassland,
Atmos. Chem. Phys., 14, 1909-1927, https://doi.org/10.5194/acp-14-1909-2014, 2014.
Tombette, M., Chazette, P., Sportisse, B., and Roustan, Y.: Simulation of aerosol optical properties over Europe with a 3-D
size-resolved aerosol model: comparisons with AERONET data, Atmos. Chem. Phys., 8, 7115–7132, 860
https://doi.org/10.5194/acp-8-7115-2008, 2008.
Toth, T. D., Zhang, J., Campbell, J. R., Reid, J. S., Shi, Y., Johnson, R. S., Smirnov, A., Vaughan, M. A., and Winker, D.
M.: Investigating enhanced Aqua MODIS aerosol optical depth retrievals over the mid-to-high latitude Southern Oceans
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
30
through intercomparison with co-located CALIOP, MAN, and AERONET data sets, J. Geophys. Res. Atmos., 118, 4700-
4714, https://doi.org/10.1002/jgrd.50311, 2013. 865
Turpin, B. J., and Lim, H.-J.: Species Contributions to PM2.5 Mass Concentrations: Revisiting Common Assumptions for
Estimating Organic Mass, Aerosol Sci. Technol., 35, 602-610, https://doi.org/10.1080/02786820119445, 2001.
van der Werf, G. R., Randerson, J. T., Giglio, L., Collatz, G. J., Mu, M., Kasibhatla, P. S., Morton, D. C., DeFries, R. S., Jin,
Y., and van Leeuwen, T. T.: Global fire emissions and the contribution of deforestation, savanna, forest, agricultural, and
peat fires (1997–2009), Atmos. Chem. Phys., 10, https://doi.org/11707-11735, 10.5194/acp-10-11707-2010, 2010. 870
van der Werf, G. R., Randerson, J. T., Giglio, L., van Leeuwen, T. T., Chen, Y., Rogers, B. M., Mu, M., van Marle, M. J. E.,
Morton, D. C., Collatz, G. J., Yokelson, R. J., and Kasibhatla, P. S.: Global fire emissions estimates during 1997–2016,
Earth Syst. Sci. Data, 9, 697-720, https://doi.org/10.5194/essd-9-697-2017, 2017.
van Noije, T. P. C., Le Sager, P., Segers, A. J., van Velthoven, P. F. J., Krol, M. C., Hazeleger, W., Williams, A. G., and
Chambers, S. D.: Simulation of tropospheric chemistry and aerosols with the climate model EC-Earth, Geosci. Model Dev., 875
7, 2435-2475, https://doi.org/10.5194/gmd-7-2435-2014, 2014.
van Noije, T., Bergman, T., Le Sager, P., O'Donnell, D., Makkonen, R., Gonçalves-Ageitos, M., Döscher, R., Fladrich, U.,
von Hardenberg, J., Keskinen, J.-P., Korhonen, H., Laakso, A., Myriokefalitakis, S., Ollinaho, P., Pérez García-Pando, C.,
Reerink, T., Schrödner, R., Wyser, K., and Yang, S.: EC-Earth3-AerChem: a global climate model with interactive aerosols
and atmospheric chemistry participating in CMIP6, Geosci. Model Dev., 14, 5637-5668, https://doi.org/10.5194/gmd-14-880
5637-2021, 2021.
Veira, A., Kloster, S., Schutgens, N. A. J., and Kaiser, J. W.: Fire emission heights in the climate system – Part 2: Impact on
transport, black carbon concentrations and radiation, Atmos. Chem. Phys., 15, 7173-7193, https://doi.org/10.5194/acp-15-
7173-2015, 2015.
Wang, R., Tao, S., Shen, H., Huang, Y., Chen, H., Balkanski, Y., Boucher, O., Ciais, P., Shen, G., Li, W., Zhang, Y., Chen, 885
Y., Lin, N., Su, S., Li, B., Liu, J., and Liu, W.: Trend in global black carbon emissions from 1960 to 2007, Environ. Sci.
Technol., 48, 6780-6787, https://doi.org/10.1021/es5021422, 2014.
Watson-Parris, D., Schutgens, N., Cook, N., Kipling, Z., Kershaw, P., Gryspeerdt, E., Lawrence, B., and Stier, P.:
Community Intercomparison Suite (CIS) v1.4.0: a tool for intercomparing models and observations, Geosci. Model Dev., 9,
3093-3110, https://doi.org/10.5194/gmd-9-3093-2016, 2016. 890
Watson-Parris, D., Schutgens, N., Winker, D., Burton, S. P., Ferrare, R. A., and Stier, P.: On the Limits of CALIOP for
Constraining Modeled Free Tropospheric Aerosol, Geophys. Res. Lett., 45, 9260-9266,
https://doi.org/10.1029/2018gl078195, 2018.
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
31
Wiedinmyer, C., Akagi, S. K., Yokelson, R. J., Emmons, L. K., Al-Saadi, J. A., Orlando, J. J., and Soja, A. J.: The Fire
INventory from NCAR (FINN): a high resolution global model to estimate the emissions from open burning, Geosci. Model 895
Dev., 4, 625-641, https://doi.org/10.5194/gmd-4-625-2011, 2011.
Woodward, S.: Modeling the atmospheric life cycle and radiative impact of mineral dust in the Hadley Centre climate
model, J. Geophys. Res. Atmos., 106, 18155-18166, https://doi.org/10.1029/2000jd900795, 2001.
Zhang, K., O'Donnell, D., Kazil, J., Stier, P., Kinne, S., Lohmann, U., Ferrachat, S., Croft, B., Quaas, J., Wan, H., Rast, S.,
and Feichter, J.: The global aerosol-climate model ECHAM-HAM, version 2: sensitivity to improvements in process 900
representations, Atmos. Chem. Phys., 12, 8911–8949, https://doi.org/10.5194/acp-12-8911-2012, 2012.
Zhang, L., Michelangeli, D. V., and Taylor, P. A.: Numerical studies of aerosol scavenging by low-level, warm stratiform
clouds and precipitation, Atmos. Environ., 38, 4653-4665, https://doi.org/10.1016/j.atmosenv.2004.05.042, 2004.
Zheng, J., Hu, M., Du, Z., Shang, D., Gong, Z., Qin, Y., Fang, J., Gu, F., Li, M., Peng, J., Li, J., Zhang, Y., Huang, X., He,
L., Wu, Y., and Guo, S.: Influence of biomass burning from South Asia at a high-altitude mountain receptor site in China, 905
Atmos. Chem. Phys., 17, 6853-6864, https://10.5194/acp-17-6853-2017, 2017.
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
32
Table 1. The details of the AeroCom Phase-III models evaluated in this study.
Model
AeroCom experimenta
Lat./Lon./Lev. Meteorology Reference
BBE
CTRL16
CTRL19
CAM-Oslo
(CAM-Nor)
√ √
192×288×30 (CTRL16)
192×288×32 (CTRL19)
ERA-Interim
Kirkevåg et al, 2018;
Seland et al., 2020
CAM5
√
√
96×144×30
ERA-Interim
Liu et al., 2012
CAM5-ATRAS √ 96×144×30 MERRA2 reanalysis
Matsui, 2017; Matsui and
Mahowald, 2017
ECHAM-HAM
√
√
√
96×192×47
ERA-Interim
Tegen et al., 2019
ECHAM-
SALSA
√ √ √ 96×192×47 ERA-Interim Kokkola et al., 2018
ECMWF-IFS
√
√
√
256×512×60
ECMWF IFS forecasts
Rémy et al., 2019
EMEP
√
360×720×20
ECMWF IFS forecasts
EMEP, 2012
GEOS
√
181×360×72
MERRA-2 reanalysis
Colarco et al., 2010
GEOS-Chem √ √
46×72×47 (BBE)
91×144×47 (CTRL16)
MERRA-2 reanalysis Bey et al., 2001
GFDL
√
180×360×33
NCEP/NCAR reanalysis
Donner et al., 2011
GISS-MATRIX
√
√
90×144×40
NCEP/NCAR reanalysis
Bauer et al., 2008
GISS-OMA
√
√
90×144×40
NCEP/NCAR reanalysis
Bauer et al., 2020
HadGEM3 √ 144×192×38 ERA-Interim
Bellouin et al., 2013;
Mulcahy et al., 2020
IMPACT
√
96×144×30
Liu et al., 2005
INCA √ √ 143×144×79 ECMWF
Balkanski et al., 2004;
Schulz et al., 2009
OsloCTM √ √ √
64×128×60 (BBE)
80×160×60 (CTRL16/19)
ECMWF Myhre et al., 2007; 2009
SPRINTARS √ √ √ 320×640×40
ERA-Interim (BBE, CTRL16)
ERA5 (CTRL19)
Takemura et al., 2005
TM5
√
√
90×120×34
ERA-Interim
van Noije et al., 2014; 2021
a. Note that different versions of models have participated in the 3 experiments.
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
33
Table 2. Details of the satellite datasets used in this study. 910
Platform
Instrument
Algorithms/products
Dataset name
Reference
Aqua
MODIS
BAR v1.0
Aqua-MODIS-BAR
Lipponen et al., 2018
Deep Blue C6.1 Aqua-MODIS-DB
Hsu et al., 2013; 2019;
Sayer et al., 2019
Dark Target C6.1
Aqua-MODIS-DT
Remer et al., 2005
MAIAC v2.0
Aqua-MODIS-MAIAC
Lyapustin et al., 2018
Terra
MODIS
BAR v1.0
Terra-MODIS-BAR
Lipponen et al., 2018
Deep Blue C6.1 Terra-MODIS-DB
Hsu et al., 2013; 2019;
Sayer et al., 2019
Dark Target C6.1
Terra-MODIS-DT
Remer et al., 2005
MAIAC v2.0
Terra-MODIS-MAIAC
Lyapustin et al., 2018
ENVISAT
AATSR
ADV/ASV v2.30
AATSR-ADV
Sogacheva et al., 2017
ORAC v3.20
AATSR-ORAC
Thomas et al., 2009
SU v4.21 AATSR-SU
North et al., 1999; North,
2002; Bevan et al., 2012
noaa18
AVHRR
Deep Blue
AVHRR-DB
Sayer et al., 2017
SeaSTAR
SeaWiFS
Deep Blue
SeaWiFS-DB
Hus et al., 2013
PARASOL
POLDER
GRASP v2.1
POLDER-GRASP
Dubovik et al., 2011
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
34
Table 3. The extinction enhancement factor (EEF) for BBA at 550 nm wavelength from in-situ measurements.
Region
Dry RH, %
Reference RH, %
EEF
Reference
Brazil
30
80
1.01−1.51
Kotchenruther and Hobbs, 1998
Australia
20
80
1.1−1.7
Gras et al., 1999
Indonesia
20
80
1.2−2.1
Gras et al., 1999
Southern Africa 30 80
1.66 ± 0.08 (fresh)
1.42 ± 0.05 (aged)
Magi and Hobbs, 2003
India
40
85
1.58 ± 0.21
Sheridan et al., 2002
China
30
80
1.64
Jung and Kim, 2011
India
≤40
85
1.32 ± 0.14
Dumka et al., 2017
915
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
35
Figure 1. Three focused fire regions in this study. a) Global maps of BB OC emissions averaged for 2006, 2008, and 2010
based on GFED4.1s (https://www.globalfiredata.org/). The domains of the three fire regions are shown by the red boxes
together with the AERONET sites (purple dots). b) The monthly evolutions of BB OC emissions from six fire types in
AMAZ (b1), SHAF (b2), and BONA (b3), respectively. The un-collocated regional mean AOD observations from 14 920 satellite datasets are shown by the light-red shaded areas as inter-quartile ranges (only the grid-boxes with more than 20 data
available in a month are included). Emissions for BB were considered in terms of the biome/fire type: tropical forest and
deforestation (DEFO), savanna (SAVA), temperate forest (TEMF), boreal forest (BORF), peat (PEAT), and agricultural
waste (AGRI).
925
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
36
Figure 2. Comparison of the 14 satellite AOD products against AERONET observations as shown by a Taylor
diagram (a) and scattering diagram of NMB (b). The colors and shapes of dots indicate different satellite datasets and fire
regions. All the satellite data were individually collocated with AERONET data during the fire seasons. POLDER-GRASP
and SeaWiFS products over BONA are not shown because the available sample size was too small (< 5) after collocation. 930
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
37
Figure 3. Variation of model AOD evaluation due to different choices of satellite products in terms of the correlation
coefficient (a, d), centered root mean square error (b, e), and normalized mean bias (c, f). The results are shown as
comparisons between the values using POLDER-GRASP (horizontal axis) and interquartile ranges (vertical axis) when 935 validating each model with different satellite products. The top (a-c) and bottom panels (d-f) show the results for individual
and synchronous collocation, respectively. The color, shape, and size of dots indicate different models, three fire regions,
and three AeroCom experiments respectively. The dashed lines show the 25%, 50%, and 100% slopes (interquartile/median).
The GISS-OMA data for BBE over BONA is not shown in (b) due to the very high CRMSE.
940
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
38
Figure 4. Validation of AeroCom models against POLDER-GRASP observations for AOD during fire seasons. The
validations of models from the three AeroCom experiments are shown as Taylor diagrams for BBE (a), CTRL16 (b), and
CTRL19 (c). The NMB for all the models is shown in panel d. The colors and shapes of dots indicate different models and 945 fire regions. All the model data are collocated with POLDER-GRASP data. The evaluation is for 2008 in BBE, for 2006,
2008, and 2010 in CTRL16, and for 2010 in CTRL19. The GISS-OMA data for BBE over BONA is not shown in (a) due to
the very large normalized CRMSE.
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
39
950 Figure 5. Comparison of the temporal and spatial correlations between modeled AOD and POLDER-GRASP
observations. Results were shown for the three experiments individually (a-BBE, b-CTRL16, c-CTRL19). All the model
data are collocated with POLDER-GRASP during fire seasons. The correlations are then calculated using either the time-
series of the regional averages (horizontal axis) or the spatial averages for the whole fire seasons (vertical axis). The red
dashed lines show the 1:1 range. 955
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
40
Figure 6. Comparison of the temporal and spatial variations of modeled AOD errors. Results are shown for the three
experiments individually (a-BBE, b-CTRL16, c-CTRL19). All the model data were collocated with POLDER-GRASP. The
variation is calculated as the ratio of interquartile to median values of the absolute bias for time series (temporal variations) 960 and spatial averages (spatial variations). The red dashed lines show the 1:1 range.
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
41
Figure 7. Changes of correlation (a), centered root mean square error (b), and normalized mean bias (c) in the BBE 965 experiment in responding to different scaling factors adopted to BBA emissions (0, 0.5, 2, 5). The colors and shapes of
dots indicate different models and fire regions. All data are collocated with POLDER-GRASP for 2008 fire seasons. The
BBE5 CRMSE and NMB for several models are not shown given the extremely large values (up to 19).
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
42
Figure 8. The dependence of total aerosol load on total aerosol emissions (a) and dependence of AOD on aerosol load 970 (b), indicating the aerosol lifetime and MECs, respectively. The data are average values for the whole fire seasons based
on the raw model output without collocation, and the light-colored error bars indicate the corresponding temporal variations
(as standard deviation). The color, shape, and size of dots indicate different models, three fire regions, and two AeroCom
control experiments, respectively. The red dashed lines show the linear trends, with a regression function and correlation
coefficients (r) also shown. Note that some CTRL16 models provide data for different years (2006, 2008, and 2010) which 975 are illustrated separately.
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
43
Figure 9. Dependence of the modeled aerosol lifetime on the time scale of total deposition. The color, shape, and size of
dots indicate different models, three fire regions, and two AeroCom control experiments, respectively. The embedded 980 diagram shows the same results zooming to a smaller scale (excluding INCA and HadGEM3 in BONA). The Person
correlation (r) and p-value (p) are shown.
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
44
Figure 10. Dependence of the modeled timescale of total deposition on precipitation strength during fire season in 985 2010. The color, shape, and size of dots indicate different models, three fire regions, and two AeroCom control experiments,
respectively. The three dashed lines indicate the GPCP data averaged for each region.
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
45
Figure 11. Variation in modeled plume height (a) and validation of modeled aerosol vertical profile against CALIOP 990 for AMAZ (b-1), SHAF (b-2), BONA (b-3). The color scheme in Fig. 11b1-3 is the same as in Fig. 11a, with the solid and
dashed lines showing the model data from CTRL16 and CTRL19 experiments (if both are available for the same model),
respectively. The gray shaded areas in Fig. 11b show the ±σ ranges for the CALIOP observation.
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
46
Figure 12. Dependence of modeled MECs on AE (a) and the validation of modeled AE against POLDER-GRASP data 995 (b). Data in Fig. 12a are model original output without collocation. The dashed line in Fig. 12a shows the relation calculated
based on Mie-scattering theory and ECHAM-HAM lookup table. Modeled AE in Fig. 12b are collocated with POLDER-
GRASP (shown as red lines) on monthly basis during fire seasons.
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
47
Figure 13. Dependence of modeled MECs on extinction enhancement factor (EEF) in models for 2010. The grey 1000 shaded area shows the EEF range from in-situ observations according to previous studies (see Table 3). Both clear-sky and
all-sky results are shown for GISS-OMA data.
https://doi.org/10.5194/acp-2022-96
Preprint. Discussion started: 28 February 2022
c
Author(s) 2022. CC BY 4.0 License.
