High-time-resolution measurements of in situ aerosol and cloud properties
provide the ability to study regional atmospheric processes that occur on
timescales of minutes to hours. However, one limitation to this approach is
that continuous measurements often include periods when the data collected
are not representative of the regional aerosol. Even at remote locations,
submicron aerosols are pervasive in the ambient atmosphere with many
sources. Therefore, periods dominated by local aerosol should be identified
before conducting subsequent analyses to understand aerosol regional
processes and aerosol–cloud interactions. Here, we present a novel method to
validate the identification of regional baseline aerosol data by applying a
mathematical algorithm to the data collected at the U.S. Department of
Energy's (DOE) Atmospheric Radiation Measurement (ARM) user facility in the
eastern North Atlantic (ENA). The ENA central facility (C1) includes an aerosol observing system (AOS) for the measurement of aerosol physical,
optical, and chemical properties at time resolutions from seconds to
minutes. A second temporary supplementary facility (S1), located
First, we investigate the local aerosol at both locations. We associate
periods of high submicron number concentration (
Due to the complexity of high-
Ambient aerosols interact with clouds by acting as cloud condensation nuclei and affecting cloud radiative properties, with significant implications for global climate change (Anderson et al., 2003; IPCC, 2014). Currently, climate forcing associated with aerosol–cloud interactions represents one of the largest uncertainties in the climate system (Carslaw et al., 2013) and in future climate projections (Simpkins, 2018). Compounding the effect on climate, regions dominated by clean atmospheric conditions, such as those observed in marine environments with low-lying clouds, are the most susceptible to aerosol perturbations (Rosenfeld et al., 2014). Recently, increases in larger longer-lasting cloud cover and cooling have been correlated with enhanced concentrations of aerosols in ultraclean regimes (Goren and Rosenfeld, 2015).
The eastern North Atlantic (ENA) Ocean is a remote region characterized by a clean marine environment and persistent subtropical marine boundary layer (MBL) clouds (Wood et al., 2015). Throughout the year, transported air masses from North and Central America, Europe, the Arctic, and North Africa (O'Dowd and Smith, 1993; Hamilton et al., 2014; Logan et al., 2014) periodically impact ENA, leading to perturbations in aerosol properties and cloud condensation nuclei concentrations. As a result, ENA is one of the regions in the world with the strongest aerosol indirect forcing and, as a result, has one of the highest associated uncertainties in terms of the aerosol impact on cloud formation, albedo, and lifetime (Carslaw et al., 2013). In the past few decades, major efforts have focused on improving the knowledge of atmospheric processes in the ENA region. Since 1991, several campaigns including the Atlantic Stratocumulus Transition Experiment (ASTEX) (Albrecht et al., 1995), the North Atlantic Regional Experiment (NARE) field mission (Penkett et al., 1998), the International Consortium for Atmospheric Research on Transport and Transformation (ICARTT) (Fehsenfeld et al., 2006), and the BORTAS campaign (Parrington et al., 2012) were conducted in the North Atlantic, studying cloud structure and long-range-transport patterns over the region.
Starting in 2009, the U.S. Department of Energy's (DOE) Atmospheric Radiation Measurement (ARM) user facility has deployed campaigns at ENA to improve comprehensive long-term measurements of marine boundary layer aerosol and low clouds in high-latitude marine environments. In 2009, the 21-month field campaign (from April 2009 until December 2010) – Clouds, Aerosol, and Precipitation in the Marine Boundary Layer (CAP-MBL) on Graciosa Island (Azores archipelago) – provided the most extensive characterization of MBL clouds in the North Atlantic (Rémillard et al., 2012; Rémillard and Tselioudis, 2015). The observations collected during the 21 months of the deployment also highlighted a strong synoptic meteorological variability associated with seasonal variations of aerosol properties (Logan et al., 2014; Wood et al., 2015; Pennypacker and Wood, 2017; Wood et al., 2017).
Following the outstanding uncertainties identified during CAP-MBL and to
continue the research on aerosol–cloud–precipitation interactions on
marine stratocumulus clouds, in 2013, ARM established a fixed site, known as
the ENA ARM facility (Mather and Voyles, 2013; Dong et al., 2014; Logan
et al., 2014; Feingold and McComiskey, 2016). The ENA fixed site is located
on the north side of Graciosa Island, which is the northernmost island
within the central group of islands in the Azores. Graciosa is the second
smallest in size with an area of
The ENA central facility (C1) is equipped with an aerosol observing system (AOS). The AOS provides a unique dataset of high-temporal-resolution measurements of in situ aerosol optical, physical, and chemical properties and their associated meteorological parameters (Uin et al., 2019). Most recently, motivated by the need of a characterization of the horizontal variability and the vertical structure of aerosol and clouds over ENA, ARM deployed the Aerosol and Cloud Experiments in the Eastern North Atlantic (ACE-ENA) field campaign (J. Wang et al., 2019). In July 2017 during ACE-ENA, ARM established a temporary supplementary facility (S1), approximately 0.75 km from the central ENA site (C1), to understand the regional representativeness of the AOS data at the ground level. A subset of AOS instruments was deployed for a period of approximately 1 year to identify the local impacts at C1 and to add additional constraints for the development of algorithms to mask local aerosol influences. During two intensive operating periods (IOPs), in June–July 2017 and January–February 2018, the ARM Aerial Facility (AAF) Gulfstream-159 (G-1) research aircraft flew over ENA, providing high-quality measurements of the marine boundary layer and lower free troposphere (FT) structure, as well as the vertical distribution and horizontal variability of low clouds and aerosol over ENA (J. Wang et al., 2016, 2019). We use the AAF and S1 data to constrain periods of time when the ENA AOS data were regionally representative of aerosol concentrations at the ground level and when they represented aerosol concentrations that were well mixed within the boundary layer.
The impact of local sources on aerosol and trace gas measurements is a common issue for continuous ambient datasets (Drewnick et al., 2012). Even at remote sites such as ENA, local sources can be pervasive and unavoidable. At ENA, the location for C1 was selected by ARM to minimize local aerosol and trace gas sources since they can interfere with regional and large-scale atmospheric aerosol processes. However, competing needs of instruments, logistics, and operations (e.g., requirement of large flat surface areas for the radars, power, and infrastructure to operate the facility) constrained the site selection. As a consequence, episodes of local aerosols are sampled by the AOS and can be observed in the high-time-resolution data. Thus, we identify all known local sources and develop a mask to isolate the regionally representative data (Aiken and Gallo, 2020; Gallo and Aiken, 2020a, b).
One method to estimate the regionally representative concentrations at sites affected by local aerosol is with meteorological filters (Giostra et al., 2011; Gao et al., 2019; Y. Wang et al., 2019). This approach masks all data related to air masses coming from wind directions associated with sources. However, meteorological filters rely upon accurate knowledge of the local sources and the availability of high-quality meteorological data (Giostra et al., 2011). This method has limited use at locations where local sources originate from a wide range of wind directions and vary with time, such as seasonal sources, as well as at locations with complex meteorology, terrain, and high wind speeds.
With high-time-resolution continuous data, it is possible to implement post-data-processing methods using statistics to identify and mask high-concentration aerosol events without removing a large fraction of the data
or relying on observational data to identify nearby sources. Smoothing
methods based on robust nonlinear data smoothing algorithms have been used
historically to improve the signal-to-noise ratio for data that includes
occasional high signals due to random noise and other events that can bias
the measurements (Beaton and Tukey, 1974; Velleman, 1977; Goring and
Nikora, 2002). Smoothing algorithms separate data into a smoothed sequence
that can be used to represent the baseline and a residual sequence composed
of the noise. Recently, Liu et al. (2018) used a smoothing algorithm
based on a 24 h running median to mask short-term local events with an
average duration of
The first aerosol filter applied to ENA AOS data by
Zheng et al. (2018) was used to study
seasonal aerosol–cloud interactions. The authors used AOS CPC data to mask
time periods when the first derivative of the submicron aerosol number
concentration exceeded 60 particles cm
We present data from two facilities at ENA, C1 and S1, during ACE–ENA to identify the local aerosol sources at ENA and to determine their influence on the AOS data. Submicron aerosol concentrations, size distributions, and meteorological data are presented. We develop a new aerosol mask at ENA using AOS data to identify periods of short-duration high-concentration submicron particle events. Our mathematical algorithm and the determination of a regional baseline for submicron aerosol is validated using the data from C1 and S1. After determining the regional baseline, we compare AOS masked data with the AAF data collected during ACE-ENA flights over C1 to understand the vertical distribution of aerosol at ENA.
The ENA central facility (C1) is located on Graciosa Island within the
Azores archipelago at 39
The aerosol supplementary site (S1) was deployed at 39
Satellite image of ENA C1 and S1 on Graciosa Island, Azores, Portugal (© Google Earth).
Three instruments, duplicate models of those used within the AOS at C1, were
deployed at S1. Two aerosol instruments were selected for their ability to
measure submicron aerosol concentrations in high time resolution. The third
instrument was included to associate the measurements with meteorological
parameters as is done in the AOS. The aerosol instruments were powered and
located inside a converted garage in an unoccupied house, with the computer
for data acquisition. The meteorology sensor was mounted above the inlet at
Briefly, the fine-mode condensation particle counter (CPC) (TSI, Inc.,
Shoreview, MN, USA; model 3772) measures the submicron number concentration
(
The ARM Aerial Facility (AAF) Gulfstream-159 (G-1) research aircraft flew
from Terceira Island (
We present and evaluate different strategies to identify periods when the AOS data are impacted by high submicron aerosol concentrations and associate them with nearby potential aerosol sources. The impacts of local aerosol sources at ENA C1 are evaluated by comparing data collected at C1 and S1. We analyzed two 1-month time periods that represent two seasons: summer (22 July–20 August 2017) and winter (1–30 December 2017).
Measurements from the UHSAS and CPC are combined to describe the submicron
aerosol size distribution by dividing the data into three optical size
modes. Zheng et al. (2018) used lognormal fitting of the submicron aerosol
size distributions from the UHSAS to define three modes to study
aerosol–cloud interactions at ENA. The lognormal fittings gave three
parameters: mode diameter, mode number concentration, and mode
One way to determine statistical outliers in the data is by comparing the
difference between the median and the mean. Time periods when the median and
mean
ENA Aerosol Mask (ENA-AM) is a standard deviation algorithm that was
parameterized for the ENA
Flow chart to apply the standard deviation algorithm to high-time-resolution aerosol data.
We determined the standard deviation of the data below the median (
Whenever a data point is identified above the threshold, the next point in
the time series is evaluated using a random walk method (threshold
We tested six different parameterizations of the algorithm, which included
two
Standard deviation algorithm input parameters tested at C1 and S1 in the summer.
Wind directions can be used with aerosol measurements to determine aerosol
sources (Zhou et al., 2016; Cirino et al., 2018). To understand the
frequency and direction from which local aerosols originate at ENA, we
present mean aerosol
In Fig. 3, 1 min
The largest mean
Polar graphs of the mean
While mean
The results of the wind direction analysis indicate that the main sources
of
Number concentrations from three aerosol modes that we defined in Sect. 3.1 are presented in Fig. 4 from C1 and S1 in the summer and winter. The
smallest mode number concentration,
For the three submicron size modes analyzed at C1 and S1,
The higher observed mean
Box and whisker plot of At-, Ac- and LA-mode aerosol number
concentrations at C1 (orange) and S1 (pink) in the
Mean and median
Time series of
In Fig. 5, we present two 1 d periods sampled at C1 and S1 during the
summer.
The 1 s
Graciosa Airport on average hosts two flights a day, the first typically in
the late morning/early afternoon and the second in the late afternoon. The
airport time tables for 2017 and 2018 reported that planes landed and took
off from Graciosa Island during three distinct time periods throughout the
day:
To further understand the potential influence of the airport operations on
Throughout the day, abrupt changes in wind direction were observed. Winds
from the south, southwest, and west dominated until 17:58 UTC. Starting at
18:00 UTC, the dominant wind directions were northwest, north, and east.
Analysis of the video from the AOS camera at C1 showed that diesel trucks
were on the runway from 09:07 to 09:27 UTC for daily maintenance. At two
times during the afternoon, 13:42 to 15:02 and 18:46 to 19:51 UTC, the
aircraft was idling near the airport terminal (Fig. 6). During the first
part of the day, when the wind directions were from the south and west,
The high-
While the influence of the airport operations may not be readily apparent
from the short-duration high concentrations observed at C1 and S1 (Fig. 5),
further information constraining this influence was obtained by looking at
the diurnal cycle of mean and median
The diurnal variation observed in the mean
At ENA, the periods with the largest deviation between the median and mean
Box and whisker diurnal profile of
To apply a mathematical algorithm to mask high-
We present, in Fig. 8, the results from the application of the algorithm
over the same 24 h period that we analyzed in Sect. 4.3, Fig. 6. The first three parameterizations selected,
Original (orange points) and masked
Similar results were obtained when we tested the six parameterizations of
the algorithm on
Due to the diverse high-
Application of the RW threshold generated
Scatter plot of
Furthermore, we evaluated the ability the
An episode of long-range-transported continental aerosol at C1
determined
Therefore, we used the
After the application of ENA-AM, we observed that mean, deviation between
mean and median, and standard deviation
Contrarily to
Mean, median, and standard deviations (
To estimate the influence of local aerosol events on daily
Original (black) and ENA-AM masked (green)
We tested using wind direction to mask local aerosols by applying a
meteorological mask to remove C1 and S1
Masking AOS data at ENA utilizing the associated metadata, such as AOS motion-activated cameras and airport flight logs, was not able to identify all of the periods impacted by local aerosol sources. However, analysis of videos and airport flight logs were useful to confirm the presence of the aircraft at the airport to validate the application of an aerosol mask. These observations and metadata were therefore used to validate the application of ENA-AM as discussed in Sect. 3.2.
Application of smoothing algorithms has been shown to be effective in
filtering measurements affected by events lasting less than 1 h
(Liu et al., 2018) and that are associated
with rapid increases in
We tested a different mathematical algorithm to filter aerosol data based on previous work by Hagler et al. (2012). The authors applied the coefficient of variation algorithm to ultra-fine-particle concentrations and greenhouse gas data. At ENA, this method masked the dominant fraction of the data, which is 72 % at C1 in the summer. We were not able to validate the additional reductions in comparison to ENA-AM with other observations or collocated measurements. Periods with known long-range-transported aerosol were also removed. Therefore, the application of this method was not pursued further at ENA.
Comparison was also made between ENA-AM and the 1 s time base filter
developed by Zheng et al. (2018) at ENA. Conducted at C1 over two
3-month periods in the summer (June to August 2017) and winter (December 2016 to February 2017), the authors found similar baseline values for
After determining the regional baseline for
The high
Scatter plot of ENA-AM masked C1
High-concentration aerosol events were observed in the AOS data at the ENA central facility. Analysis of the submicron aerosol concentrations and size distributions were used with collocated meteorological data (wind direction) to associate high-concentration aerosol events with potential local aerosol sources. Total submicron and Aitken-mode aerosol were the most affected as determined by wind direction and should be masked before conducting ambient aerosol process studies at ENA. Accumulation-mode aerosol was less impacted, especially in the summer. Ac mode might then be used without applying an aerosol mask as representative of the regional aerosol.
We developed a novel aerosol mask at ENA called ENA-AM and validated its application by using two measurement locations located within 1 km of each other. The temporary supplementary site was deployed to validate the new aerosol mask at the central facility with the AOS. Time periods impacted by high-concentration aerosol events were removed, and we were able to define a regional baseline for the submicron number concentration data at ENA during the summer and winter. The masked submicron aerosol number concentrations from the ground site were compared with the AAF aircraft data during flights over the facility. It was possible to determine a well-mixed regional aerosol within the first 500 m of the marine boundary layer for the data presented here collected during the summer ACE-ENA IOP.
Application of ENA-AM required measurements in which (1) the time resolution
of the dataset was shorter than the typical length of the event and (2) the
variation within the baseline data was smaller than the variation during the
periods containing local aerosol. The CPC 1 min submicron number
concentration data satisfied these requirements at ENA. Therefore, we
developed an algorithm using the CPC data at ENA that could be applied to
the AOS data for studying regional aerosol processes. After the application
of ENA-AM, 26 % of the 1 min data at C1 and 15 % at S1 were masked
in the summer. ENA-AM masked a lower percentage of the data than the wind
direction mask, which masked 39 % of the data at C1 data and 43 % at S1.
Compared to the meteorological method, ENA-AM removed approximately half of
the data compared to the mask based on wind direction and, more importantly,
resulted in a higher
Data were obtained from the Atmospheric Radiation
Measurement (ARM) user facility, a U.S. Department of Energy Office of
Science user facility sponsored by the Office of Biological and
Environmental Research (available at
The supplement related to this article is available online at:
ACA and FG conceptualized the analysis and wrote the manuscript. FG led the analyses with additional input from JU, StS, RW, and FM. ACA was the project administrator. All authors were involved in helpful discussions and contributed to the manuscript.
The authors declare that they have no conflict of interest.
This article is part of the special issue “Marine aerosols, trace gases, and clouds over the North Atlantic (ACP/AMT inter-journal SI)”. It is not associated with a conference.
The work was supported by the Atmospheric Radiation Measurement (ARM) program, funded by the U.S. Department of Energy (DOE), Office of Science, Office of Biological and Environmental Research. We acknowledge the ARM Research Facility, a user facility of the U.S. DOE, Office of Science, sponsored by the Office of Biological and Environmental Research for providing data. We also acknowledge the ENA ARM site manager, Heath Powers; operations manager, Paul Ortega; and site operators, Carlos Sousa, Tercio Silva, and Bruno Cunha.
This research has been supported by the U.S. Department of Energy (Atmospheric Radiation Measurement (ARM) program).
This paper was edited by Lynn M. Russell and reviewed by two anonymous referees.