COMMIT in 7-SEAS/BASELInE: Operation of and Observations from a Novel, Mobile Laboratory for Measuring In-Situ Properties of Aerosols and Gases

Trace gases and aerosols (particularly biomass-burning aerosols) have important implications for air quality and climate studies in Southeast Asia (SEA). This paper describes the purpose, operation, and datasets collected from NASA Goddard Space Flight Center’s (NASA/GSFC) Chemical, Optical, and Microphysical Measurements of In-situ Troposphere (COMMIT) laboratory, a mobile platform designed to measure trace gases and optical/microphysical properties of naturally occurring and anthropogenic aerosols. More importantly, the laboratory houses a specialized humidification system to characterize hygroscopic growth/enhancement, a behavior that affects aerosol properties and cloud-aerosol interactions and is generally underrepresented in the current literature. A summary of the trace gas and optical/microphysical measurements is provided, along with additional detail and analysis of data collected from the hygroscopic system during the 2015 Seven South-East Asian Studies (7-SEAS) field campaign. The results suggest that data from the platform are reliable and will complement future studies of aerosols and air quality in SEA and other regions of interest.


INTRODUCTION
Aerosols (e.g., dust, soot, black carbon, organic carbon, sea salt) and trace gases such as carbon dioxide (CO2), carbon monoxide (CO), ozone (O3), nitric oxide (NO), nitrogen dioxide (NO2), and sulfur dioxide (SO2) are of special interest to air quality and climate studies because they play an active role in modulating Earth's radiative forcing (via atmospheric absorption/scattering), long-term climate processes (Booth et al., 2012;Lin et al., 2014), as well as human health and life expectancy (Brook et al., 2010;Brauer et al., 2012).Ground-based measurements provide an in-situ source of aerosol and gas information for direct analysis and for calibration/validation of observations derived from satellite, suborbital, and other in-situ or remote sensing platforms.
For Southeast Asia (SEA), aerosol and air quality studies are of particular interest.In past studies, the area around northern Thailand has been described as a major active source for biomass burning (BB), air pollution, and fires in South Asia (Chan et al., 2003;Chantara et al., 2012;Chuessard et al., 2014;Lin et al., 2014;Wang et al., 2015).Gautam et al. (2013) used satellite and ground-based radiometric observations over five years to specify the Thailand-Myanmar border region as a BB hotspot during the springtime dry (pre-monsoon) season.BB aerosols are particulates emitted and suspended in the atmosphere as a result of anthropogenic fires started to clear land, cook, or eliminate waste.These aerosols may also be generated by naturally occurring wildfires.Additional studies have demonstrated that BB aerosols generated in SEA undergo long-range transport to the north and east at upper levels and settle in areas as far away as Taiwan, some 2400 kilometers (km) northeast of western Thailand (Chan et al., 2003;Lin et al., 2009;Huang et al., 2013;Yen et al., 2013;Lin et al., 2014).Bell et al. (2013) identified regions within the South China Sea, east of Mainland SEA, as an area where the effects of BB aerosols are combined with those of naturally occurring mineral dust and maritime aerosols.Lin et al. (2014) used satellite and ground-based observations and simulations to demonstrate that source aerosols generated over SEA had a net negative radiative effect at ground level and at top of the atmosphere, as well as a positive effect in between.In addition, they isolated a zone stretching from southern China to Taiwan where surface shortwave radiation was reduced as a result of these aerosols.Thus it is evident that natural and anthropogenic (particularly BB) aerosols play a major role in SEA air quality.
Aerosol optical/microphysical properties and trace gas concentrations associated with BB have been the subject of several studies on air quality.Trace gases act as strong markers and help to differentiate types and sources of pollution regimes (Chan et al., 2003;Poncanart and Akimoto, 2003;Andreae and Merlet, 2011;Kondo et al., 2011;Tsai et al., 2013;Ou-Yang et al., 2016).The same is true for aerosol spectral single scattering albedo (SSA), the ratio of aerosol scattering to total (absorption plus scattering) extinction, which also affects climate (Eck et al., 1999;Petzold et al., 2013, and references therein).Microphysical properties of aerosols (e.g., particle diameter, particle mass concentration, particle size distribution [PSD]) correlate well with respiratory and cardiovascular health (Kristensson et al., 2013;Lelieveld et al., 2015) but also have implications for cloud-aerosol interaction studies (Lance et al. 2006), among others.
Hygroscopic properties remain an important factor in understanding aerosol behavior.Depending on the composition and size of aerosol particles, as well as the relative humidity (RH) of the ambient air, aerosols can absorb water and increase in size as RH increases beyond a certain value, called the deliquescence point.Moistened particles can display different optical properties (Tao et al., 2014), exhibit PSDs shifted toward larger particle diameters (Rissler et al., 2006;Hu et al., 2013), and experience enhanced scattering and radiative effects (Hervo et al., 2014) compared to their dry counterparts.These effects are different for drying particles, since recrystallization does not necessarily occur at the deliquescence point, leading to complicated aerosol hysteresis effects, which were first demonstrated by Orr et al. (1958). Englehard et al. (2012) concluded that hygroscopic properties also differ and act as a marker between primary and aged aerosols, where primary aerosols that have not been aged by photochemical oxidation show higher hygroscopic growth variability than secondary organic aerosols.Climate models may suffer from not parameterizing these effects (e.g., Khylstov et al., 2005;Bell et al., 2013), and lack of hysteresis data creates a "significant gap" in aerosol studies according to Reid et al. (2005).Analysis of hygroscopic growth is therefore a primary interest in many recent aerosol studies and is a main focus of the most recent observations.Since man-made and naturally occurring aerosols generated within mainland SEA can travel long distances and affect the radiation budget, climate, and human health along the pathway, the development and deployment of a comprehensive, reliable suite of instruments for measuring hygroscopic, optical, and microphysical properties of aerosols and trace gases will have a positive impact on the understanding of air quality and climate in SEA.It will serve as a platform to increase the temporal resolution of data offered by polar-orbiting satellites, and to intercompare with data products retrieved from geostationary satellites.Over the course of the past ten years, the NASA Goddard Space Flight Center's (NASA/GSFC) Chemical, Optical, and Microphysical Measurements of In-situ Troposphere (COMMIT) ground-based mobile laboratory has participated in field campaigns at various international locations, most recently concentrated in SEA.During each experiment, a broad range of air quality/aerosol instruments operate continuously and archive quality-controlled, high temporal resolution, and calibrated data products.
In this paper, we provide a brief description of the most recent sites to which COMMIT has been deployed.Next we detail facility operations, then provide a brief overview of trace gas and optical/microphysical aerosol measurements, followed by a detailed description of a hygroscopic analysis system as it applies to optical scattering and aerosol PSD measurements.Finally, we summarize results from recent campaigns and conclude with suggestions for future work.The primary intent of this study is therefore to provide a point of reference for COMMIT's standard operating procedures (SOP) for field campaigns and to highlight a unique method of measuring hygroscopic properties of BB aerosols in SEA.Detailed analyses using the results obtained during campaigns are presented in other papers within this special issue (Tsay et al., 2016 and references therein).

Deployment Sites
Measurements from the most recent field campaign were collected at the meteorological station in Doi Ang Khang (DAK), Chiang Mai Province, Thailand (19.93°N, 99.05°E, 1518 meters [m] above sea level).This intensive operation period was part of the larger Seven South-East Asian Studies/ Biomass-burning Aerosols and Stratocumulus Environment: Lifecycles and Interactions Experiment (7-SEAS/BASELInE) field campaign, which has been ongoing since 2013, and since 2010 as a precursor, the 7-SEAS/Dongsha Experiment (Lin et al. 2013).The site is located in the mountainous northwestern terrain of Thailand, around 125 km north of Chiang Mai.Data were collected from February to April 2015, during the peak of the BB source activities in SEA.Prior to the experiment it was assumed that a large portion of the aerosol loading would be generated from local BB activities, which was at least qualitatively confirmed by observation of daily burning events for clearing of forested land and agricultural fields.
Previous 7-SEAS experiments involving the COMMIT mobile laboratory were conducted at the meteorological station in Son La (SL), Vietnam, (21.33°N, 103.90°E, 660 m above sea level) during Spring 2012 and 2013 (Tsay et al., 2013).Investigations completed here provide continuity in instrument data sets and contribute to the study of the evolution of atmospheric pollutants as they move downstream and age from their source regions.Local anthropogenic BB activities have a strong influence on the region's air quality as well.During 2013, the Aerosol-Cloud-Humidity Interactions Exploring and Validating Enterprise (ACHIEVE) mobile laboratory was deployed downstream of the SL site, and it provided the remote-sensing data detailed in Loftus et al. (2016).ACHIEVE's mission is focused on understanding aerosol-cloud interactions, to which COMMIT's aerosol and hygroscopic growth data will provide helpful modeling constraints.

Facility Operations
In total, over 25 instruments operate simultaneously inside COMMIT.Instrument data streams include aerosol observations at up to 1 Hertz (Hz) temporal resolution.Table 1 provides more information on the individual instruments available with the COMMIT platform.All instruments referenced in this paper will be referred to using the abbreviations listed there.
The system is housed in a 20-foot weatherproof sea container (Seabox, Inc., USA) that can be made fully operational from a packed state in less than one day.All For most instruments, ambient air is sampled at 7 m above the laboratory roof and is drawn into a 15-centimeter (cm) diameter metal stack (Fig. 1) via a blower.Air flows through the stack such that the flow across the cross-sectional area of the stack is ten times higher than the flow across the individual sampling ports so that flow is approximately isokinetic.At the base of the main stack, several stainless steel pipes are installed to direct the airflow into various rackmounted instruments via conductive tubing with minimal bend.Tube diameters and flow pathways were chosen to minimize particle loss and to balance the flow among the different instruments.Beta Attenuation Monitors (BAMs) and a Tapered Element Oscillating Microbalance (TEOM) sample air independently from the main stack via vacuum pumps.Nephelometer (NEPH) flow is reported in the DAK campaign with a PM2.5 size-cut for this particular study.However, the cyclone may be replaced/removed if particular/full-range particle sampling is desired.
All instruments, where possible, undergo flow checks, zero-air (i.e., particle-free air) checks, and intercomparisons before and after the experiment to ensure quality data and consistent operation of the equipment (Table 2).Flows are monitored using a Gillian bubble meter (Sensidyne LP, USA), which was crossed-checked with an independently calibrated flow meter during the most recent campaign.Zero-air is generated using a chemical scrubbing method, in which compressed air is passed at ~30 psi through a dehumidifier (Twin Towers Engineering, Inc., USA) and a series of chemical scrubbers containing Purafil (Purafil, Inc., USA), Drierite (W. A. Hammond Drierite Co. Ltd., USA), sodalime, and activated charcoal.This routine has demonstrated effectiveness in past studies (Rutter et al., 2012).All scrubbers are changed before and after each campaign, with more frequent changes to Drierite to ensure dehumidification of the incoming air.For cases where quick, filtered air is required, a fresh high-efficiency particulate arrestance (HEPA) filter is placed in the sampling line.Data streams are collected and archived using a centralized command and data handling system that provides a virtualized platform and a consistent manner in which to interact with the independent instruments.It offers large  1 sample from the main sampling stack, with NEPH sample air passing through a PM2.5 sharp cut cyclone (orange line).All flows are listed, with units in liters per minute (L min -1 ).Vacuum pumps, labeled 1-4, actively pump air through instruments and outside the laboratory.Instruments not connected to vacuum pumps sample air via internal pumps and exhaust it passively outside the laboratory.

Trace Gas Measurements
COMMIT houses six trace gas analyzers that measure gas concentration (Fig. 1, dark red lines) using a variety of methods including chemiluminescence and infrared absorption.Ambient air passes through a particle filter to remove large particles prior to sampling.All gas monitors are regularly calibrated using a dynamic gas calibrator and predefined spans of a National Institute of Standards and Technology (NIST)-traceable, known gas mixture containing CO2, CO, SO2, and NO.The plumbing has been modified to allow seamless switching between calibration and sampling modes using a three-way valve.The zero-air generation system described above is periodically used to challenge the monitors with purified air to test their baseline readings.However, a traceable ozone source is required to calibrate the ozone monitor.
Air Quality Design, Inc. (AQD) upgraded a 42C Chemiluminescence NO-NO2-NOx Analyzer in 2011.The upgrade includes a redesigned sampling inlet containing a blue light converter that converts NO2 to NO in the presence of O3 at a ~50% conversion rate.Various valves direct sample flow to the blue light converter and to a molybdenum converter (for conversion of NOx + other nitrogen species [NOy] to NO), and then the 42C for direct measurement of NO by chemiluminescence (Fig. 1, light blue/light green lines).The modified AQD sampling inlet is located close to the convertors to minimize the physical distance between them.The monitor undergoes the same multi-point span calibrations as the other analyzers, with an additional calibration using n-propyl nitrate (nPN) to qualitatively assess the performance of the NOy converter.
Ou-Yang et al. ( 2016) analyze several trace gas data sets collected from COMMIT at DAK during 2015 and also make comparisons to observations taken at SL during 2012-2013.They use the data to characterize the age and composition of the air masses of the region and also compare the trace gas ratios from BB-dominated regions and urban pollution dominated regions.Sayer et al. (2016) illustrate how peaks in the gas concentration data align well in time with peaks in aerosol loading observed by NASA's Aerosol Robotic Network (AERONET, Holben et al., 1998) of sun photometers.

Optical Measurements
During the DAK and SL campaigns, aerosol loading due to BB was often high enough to cause loss of sensitivity in the optical instruments.For example, aerosol deposition on the filter-based Particle Soot Absorption Photometer (PSAP) would decrease optical transmission of light to below 70% of reference transmission within one hour during poor air quality days, a threshold below which Bond et al. (2010) showed that instrument calibrations are no longer valid.The Cavity-Attenuated Phase Shift Monitor (PMEx), an optical extinction monitor, measures the time that light spends inside a sample cell bound by two mirrors.The time decreases as particle concentration increases due to forward scattering of the light by the particles (Petzold et al., 2013), so contamination of the mirrors directly affects the efficiency at which this occurs, particularly when the loss is greater than 1500 Mm -1 , and results in periods of missing data.Changing filters and cleaning mirrors frequently using a positive pressure clean room and filtered air helps avoid downtime.
An Aethalometer (AE31, Fig. 1, blue line) measures black carbon concentrations by passing ambient air through a quartz filter and performing continuous optical analysis on the deposited sample at seven wavelengths from 370 to 950 nanometers (nm).The black carbon concentrations at each wavelength are directly proportional to absorption coefficients (σa), which are then corrected according to Arnott et al. (2005) to account for multiple scattering effects.Using the corrected AE31 data strengthens the correlation with σa derived from PSAP (Fig. 1, blue line), which provides additional absorption information at three wavelengths.
Optical extinction coefficients (σe) are measured using three PMEx monitors (Fig. 1, blue lines) operating at different visible wavelengths (Table 1).Due to the design of the instrument, a small amount of filtered, purged air is required inside the chamber to prevent the mirrors from becoming contaminated.This slightly reduces path length and can be corrected in post-mission analyses (Petzold et al., 2013).
Optical scattering measurements are provided at three wavelengths, measured by a pair of NEPHs.Described in detail below, the pair consists of two identical NEPHs, one sampling ambient air (NEPHA) and one sampling humidified air (NEPHW), to extract hygroscopic aerosol effects on optical scattering.Instrument intercomparisons are conducted with data from NEPHA so that RH is similar between absorption, extinction, and scattering coefficient measurements.
Fig. 2 illustrates a closure study with data from DAK during 2015, using measurements interpolated to the blue channel (470 nm).During March 2015, high correlations (r ≥ 0.97) between σ e, measured by PMEx, and the sum of σa and σs, measured by an AE31 and NEPHA, respectively, indicate that the measurements are consistent between platforms.There is a linear relationship between PMEx extinction vs. AE31 and NEPH A with a gradient of ~0.94 and small intercept, meaning larger values of extinction are estimated by summing AE31 and NEPHA than are estimated by directly observing extinction using PMEx.Deviation from a slope of unity could be attributed to the use of a PM2.5 cyclone for the NEPH, however in a BB atmosphere, particles larger than PM2.5 contribute comparatively less than smaller particles to extinction in the visible and are not expected to be present in large concentrations (during the period from 1-20 March 2015 at DAK, the mean and standard deviation of the PM2.5/PM10 ratio was 89.5% ± 8.8%).Performing the analysis for the same time frame using PSAP σa also gives high correlations (> 0.97), but the linear relationship is weaker (gradient ~0.84) than that seen using AE31.Sayer et al. (2016) provide a detailed analysis and comparison between Moderate Resolution Imaging Spectroradiometer (MODIS) Deep Blue aerosol products, COMMIT optical in-situ data, and SSA observed by a collocated sun photometer operated by AERONET at DAK.They determined that surface SSA was well correlated with AERONET retrievals of column SSA, but that a low offset (more strongly absorbing aerosols) was present in the in-situ COMMIT data, attributed to a combination of discrepancies in aerosol loading throughout the atmospheric column versus at the surface, and potential instrumental/ retrieval artifacts.
An Aerodynamic Particle Sizer (APS, Fig. 1, green line) and a Fast Mobility Particle Sizer (FMPS, Fig. 1, light green line) provide additional PSD information to extend that of SMPS and to complement the particle mass information from TEOM and BAMs.APS reports aerodynamic particle diameters, geometric means, and geometric modes; FMPS and SMPS report the same information, but for electrical mobility diameter.All particle sizers were configured to report number concentrations [dN/dlogdP] at identical time intervals to simplify intercomparisons.FMPS operates behind a PM1 size-cut cyclone along with TEOM to provide high temporal resolution information on PSD of fine-mode particles.
Following Khylstov et al. (2005) and Bell et al. (2013), SMPS and APS data are post-processed and merged to create a continuous PSD ranging from 20 nm to 20 µm, accounting for differences in measurements of aerodynamic diameter mobility and electrical mobility.This technique shifts the APS PSD to match a power fit of SMPS data in the overlap region (~520 nm-900 nm) and provides a visualization of a wide range of ambient aerosol particles that is not commonly observed with any single instrument, particularly when mineral dust and BB aerosols are mixed together in a region.
A Cloud Condensation Nuclei (CCN) counter provides CCN number concentration at various supersaturation ratios (Fig. 1, pink line).These data are further explored by Hsiao et al. (2016), who analyzed the diurnal patterns of CCN and determined that fresh BB aerosols were composed mainly of particulate organic matter during the DAK campaign.Calibration of the CCN counter is conducted by sampling atomizer-generated ammonium sulfate particles of a known diameter.

Hygroscopic Measurements and Analysis
COMMIT collects continuous observations of hygroscopic effects on aerosol scattering, PSD, and mean geometric particle diameter.These provide two robust datasets that are unique but correlated, and thus valuable for furthering the understanding of aerosol behavior in high RH environments; they will be of particular interest to studies of BB aerosols in SEA.
A humidification system, similar to the water-based system described by Titos et al. (2014), is the critical component for the hygroscopic analysis.The humidifier was developed and tested in the laboratory (Fig. 3(a)).The unit consists of a 30 cm long Gore-Tex tube with a 1.5 cm inner diameter, through which a sample of ambient-RH air passes with minimal particle loss or modification assumed.Stainless steel coils are embedded inside the tube to maintain its structure at high flow rates.A plastic cylinder surrounds the tube and is filled with temperature-controlled distilled water that is heated and recirculated through a water bath.Compared to a diffusion membrane humidifier, this design allows the desired sample RH to be achieved quickly by adjusting the temperature of the recirculating water.As shown in Fig. 3(b), the humidifier can generate an ambient air sample with a temporally stable RH.Sharp changes of ambient RH are dampened by passage through the humidifier, which ensures consistent measurements of hygroscopicity.By adjusting the humidified air flow and temperature of the water from 2-20 liters per minute (L min -1 ) and 20-50°C, respectively, a sample RH of anywhere from 35-90% can be maintained, although it is generally maintained at or above 85% for hygroscopic growth studies, a level suggested by Titos et al. (2004).
The performance of this system was validated by checking the hygroscopic growth of laboratory-generated 100 nm sodium chloride particles.The critical humidity (the RH at which particles grow rapidly) of the experimental particles was observed at 75.3%, compared to literature values ~75%, which confirms the capability of the humidification system.Temperature and RH sensors (PTB210s, Vaisala, Finland) continuously monitor conditions upstream and downstream of the humidifiers to ensure stable performance.
During normal operations, the humidification system is operated with two identical instruments (one sampling ambient air and the other sampling humidified air) to obtain simultaneous measurements at different RH.This setup has benefits compared to operating a single instrument with switched air flows, since uncertainties related to sampling time, instrument response, and disruptions of flow are all avoided by maintaining continuous measurements.However, characterization and calibration of the individual instruments to account for their systematic differences is required prior to data analysis to ensure high quality results.
Hygroscopic enhancement of light scattering coefficients (σs) of aerosols is measured with two integrating NEPH instruments operating in series (Fig. 1, orange line).Total scattering and backscattering coefficients at three wavelengths (Table 1) are reported throughout the deployment and corrected for angular nonidealities as described by Anderson and Ogren (1998).
As shown in Fig. 4, ambient air exhausted by NEPHA passes through the humidification system and is redirected as input for NEPHW to keep flow conditions identical between instruments.Anderson and Ogren (1998) demonstrated that NEPHs operating in series have no significant particle loss beyond the uncertainty of the instruments.The original NEPH blowers were removed and replaced with an external pump to control the fixed flow rate at 16.7 L min -1 required by the upstream PM2.5 size-cut cyclone.Other cyclones or selective inlets may be placed upstream to further categorize the sample, particularly if particle diameter/human health is of interest.
Before and after each campaign, the paired NEPH system is tested and calibrated rigorously.Most commonly, users perform a gaseous calibration using CO2, which has a well-known σs, although other gases have been proposed (Anderson and Ogren 1998;Abu-Rahman et al. 2006).Another method is to use white (non light-absorbing) aerosol particles to characterize the NEPH's performance at higher σs than is normally observable for particle-free, gaseous air (Abu-Rahman et al., 2006).In addition to calibrating the NEPHs with filtered air and CO2, the effects of the humidification system on the measurements are also characterized.First, scattering measurements between NEPHA and NEPHW are compared with the humidifier removed from the sampling line to determine the baseline performance of each instrument.Before and after the DAK campaign, there was good agreement in slope (1.03-1.12pre-campaign, 0.98-1.03post-campaign) and correlation (r = 0.97-0.98pre-campaign, 0.85-0.86post-campaign) at all wavelengths between instruments sampling ambient air for 24 hours.As an alternate test, the humidifier in front of NEPHW is replaced, but with the flow of water stopped.In this sense, the degree of particle loss across the Gore-tex tube, which is assumed to be minimal, is determined without the additional effect of hygroscopic growth due to the warm water.Titos et al. (2014) conducted similar tests of their humidification system, during periods when the reference NEPH closely matched the humidity of the RH-modified NEPH.Completion of these tests before and after each campaign provides important baselines and flow calibrations, which are taken into account during subsequent analyses.
Hygroscopic effects on aerosol microphysical properties are assessed using a tandem SMPS system with a humidifier like that used for the NEPH system (Fig. 1, pink lines).Unlike the NEPHs, however, the humidified SMPS (SMPSW) and ambient SMPS (SMPSA) run in parallel.Each SMPS consists of an X-ray neutralizer (TSI3087) to obtain Boltzmann charge equilibrium, a differential mobility analyzer (DMA3081) to classify the aerosols by size, and a condensation particle counter (CPC3787) to determine number concentrations (Table 1).The sheath and aerosol flow rates are set at 2.0 and 0.6 L min -1 , respectively, and they were specifically chosen to extend the PSD range up to 1000 nm.
SMPSA samples ambient air, and its sheath and excess flows are recirculated between it and the classifier (Fig. 4) in a closed-loop system.SMPSW samples air downstream of a humidifier, and sheath flow passing through the DMA is pulled by a mass flow controller from the humidified sample stream in a "flow-through" mode.This design eliminates a potential imbalance in RH between the humidified aerosol flow from the classifier and the otherwise unhumidified sheath/excess flow (comparable to studies by Massling et al., 2007).Lab testing indicates that the difference in humidified sheath and humidified sample flows can be reduced to within ± 1%.The same sensors used for the humidified NEPH system monitor RH throughout the SMPS system.Likewise, RH is fully adjustable based on the temperature of the associated water bath, although flow rates and temperature ranges are adjusted to meet the requirements of the SMPS flow.
Extensive inter-comparisons are conducted before and after the campaign to assess characteristics of the paired SMPS system (cf.Table 2).Traditionally, inorganic salt particles like ammonium sulfate (Svenningsson et al., 2006;Mikhailov et al., 2009;Suda and Petters, 2013), a calibration dust standard (Denhean et al., 2014), or polystyrene latex (PSL) particles with known diameters (Hennig et al., 2005) are used for calibration and hygroscopic analysis of particle sizers.In this operation, the performance and effect of the humidification system is characterized using a series of different plumbing schemes.First, with the humidifier bypassed, the baseline measurements of each instrument operating under a closed-loop sheath flow mode are recorded.Next, SMPSW is switched to "flow-through" mode with the humidifier bypassed.Finally, the humidifier is added back into the SMPSW sampling line, without water flow, to determine the combined effect of the "flow-through" sheath modification and particle loss across the humidifier.Suda and Petters (2013) calibrate their tandem DMAs in a similar way, characterizing flow and voltage differences before further analysis of their data.
A calibration was conducted at the end of the DAK campaign in April 2015 (Fig. 5).The results showed that the number concentration was consistently greater in SMPSA compared to SMPSW even during calibration tests, with SMPS W reporting concentrations around 80% of those observed by SMPSA at the middle of the observable PSD.This result was also observed prior to the start of the campaign, with little change in overall concentrations between instruments during pre-and post-mission calibrations.The differences may arise from counting uncertainties or flow rate differences in the CPCs for particles at low concentrations.To make further comparisons for this analysis, the number-weighted PSDs for each instrument are normalized by the corresponding total concentrations.The normalized PSD eliminates the concentration discrepancy and allows for the direct comparison of the sizing changes (i.e., particle growth) due to the humidification process For the calibration test where SMPSW operated in closed-loop mode with no humidifier (Fig. 5(A)), the mode particle diameter of the PSD for SMPSW occurred at 113.5 nm, whereas it was 109.4 nm for SMPSA.Using "flowthrough" mode (Fig. 5(B)), the mode occurred at 145.9 nm for SMPSA and 151.2 nm for SMPSW.After adding the waterless humidifier back in line (Fig. 5(C)), the mode occurred at 145.9 nm for SMPSA and 135.8 nm for SMPSW.For each calibration test, the mode particle diameter were within one or two bins of each other (< 2% of the full 114 bin range of the SMPS), which reveals that the PSDs were relatively stable between SMPS A and SMPSW within each of the various calibration tests.The difference in particle modes between tests is due to the fact that they were conducted over several hours and air masses changed between sampling periods.Therefore, it appears that the presence of the humidifier does not substantially modify the sample flow.
Based on data collected during several days at DAK, the results demonstrate that hygroscopic growth was successfully obtained using the setup described above for the NEPH and SMPS systems.Fig. 5(D) shows the averaged, normalized PSDs for SMPSW and SMPSA from 15-20 March 2015 at DAK.For this study, all times when SMPSW RH was greater than 85% and when SMPSA RH was less than 40% were considered.After normalization of the PSDs, the SMPSW PSD appears noticeably shifted rightward, with the mode particle diameter for SMPSW at 168.5 nm compared to 126.3 nm for SMPSA, a difference of eight bins (7.0% of the full 114 bin range of the SMPS).
Scattering enhancement, f [RH], describes the increase in optical scattering of aerosols due to their hygroscopic growth.It is normally reported at 85% RH and given by: [ ] ( ) where σs wet (σs dry) describes optical scattering at high (low) RH, a describes the enhancement as RH approaches 100%, and γ describes the magnitude of the enhancement.The NEPHs at DAK demonstrated that f [RH = 85%] ranged from 1.20 for the blue wavelength to 1.28 for the red wavelength (Fig. 6(A)).The degree of scattering enhancement, γ, is related to wavelength, with longer wavelengths exhibiting higher γ.Previous reviews have shown a wide range in f [RH = 85%], from 1.1-2.2 for BB aerosols (Reid et al., 2005) and from 1.2-3.5 for non-BB aerosols (Titos et al., 2014).This variability is attributed to differences in aerosol type,

CONCLUSIONS
7-SEAS/BASELInE marked an important deployment of the COMMIT mobile laboratory into a BB-rich aerosol environment, complementing other 7-SEAS observations from 2012-2013.After analysis of various measurements, we conclude that our SOPs are conducive to continuous, quality controlled data streams of value to the aerosol community.Our extensive observations of anthropogenic and naturally occurring trace gases and optical/microphysical properties of aerosols (particularly BB aerosols) may be applied to derive further information about the air quality of a particular region.Several related studies demonstrated that the calibrated trace gas measurements have the potential to characterize the age and composition of observed aerosols.Our closure analysis validated the scattering/absorption/ extinction measurements and suggests that they can be reliably used to pursue studies on aerosol effects on SSA and the radiation budget.Finally, microphysical measurements of particle size, mass, and concentration have obvious implications to human health and longevity.
The key finding of this analysis stems from the hygroscopic growth setup for the paired NEPH and SMPS instruments.The system is well characterized, and both hygroscopic growth and scattering enhancement were observed during the latest campaign in DAK.Calibration of the systems before and after operations ensures that the humidification itself does not introduce major artifacts into to the data streams.Hygroscopic measurements will help characterize aerosol behavior in high RH environments and is expected to play a major role in the analysis of BB and aerosols in SEA since they have not been extensively studied in the literature.In combination with remote sensing data on internal cloud structure obtained with ACHIEVE, COMMIT hygroscopic monitoring will advance the understanding of aerosol-cloud interactions in a region where those effects are poorly known.
Future deployments will continue to emphasize wherever BB aerosols play a major role in pollution and climate forcing.However, focus may shift to megacity regions, where urban pollution and smog outweigh the effects of BB.Such studies would provide an interesting contrast to the recent studies of SEA, and have been preliminary pursued in campaigns such as Deriving Information on Surface Conditions from Column and Vertically Resolved Observations Relevant to Air Quality (Discover-AQ), when COMMIT collected several months of observations near GSFC in Maryland.
More extensive analysis of hygroscopic growth is currently in progress.In this study, systematic differences were observed in the measurements between the CPCs that comprise the SMPS systems.For this particular analysis, number-weighted PSDs from SMPS were normalized by total concentration and used to compare the size changes between humidified and ambient particles to reduce the effect of the discrepancy.Through a complete examination instrument performance and parameters, including the comparison of the raw particle counts of the PSDs from the CPCs, we hope to determine the source and resolve the concentration discrepancy before the next field deployment, and therefore eliminate the need for normalization of the SMPS data.
Additionally, a logic-controlled feedback loop is being integrated into the NEPH/SMPS humidification system.This will allow the recirculated water to not only be heated, but to be chilled as well.This upgrade will ensure a more stable RH of the humidified instrument flow, and one that is resistant to transient changes in ambient air temperature.
Finally, measurements of dried particles and volatile organic compounds would complement the suite of COMMIT observations, extending the ability to access hygroscopic effects on aerosols and air quality.After the current humidification system is upgraded, tested, and completed, focus will shift to these types of measurements and the hardware required to obtain them.
Data are already available upon request through direct communications with the NASA SMARTLabs team.However, a new web interface is in development, and it will allow community access to various levels of qualitycontrolled/post-processed datasets and visualizations from current and past field campaigns.

Fig. 1 .
Fig. 1.Schematic of sample flow within COMMIT.Instruments sample ambient (or humidified) air via six separate pathways.Each BAM samples air individually, downstream of PM10 size selective inlets (red, purple lines).TEOM and FMPS sample air downstream of a PM1 sharp cut cyclone (light green line).Gas monitors sample air independently of all other instruments (brown line), with NOxy flow passing through the modified AQD inlet (cyan line).Other instruments from Table1sample from the main sampling stack, with NEPH sample air passing through a PM2.5 sharp cut cyclone (orange line).All flows are listed, with units in liters per minute (L min -1 ).Vacuum pumps, labeled 1-4, actively pump air through instruments and outside the laboratory.Instruments not connected to vacuum pumps sample air via internal pumps and exhaust it passively outside the laboratory.

Fig. 3 .
Fig. 3. a: Diagram showing the design of the humidifier for the tandem NEPH and SMPS hygroscopic systems.Air flows through a central tube made of Gore-Tex that is surrounded by a cylinder with recirculating warm water.b: Time series of 16 hours of RH data from various sensors mounted through the SMPSW system, upstream and downstream of the humidification system, during April 2015.

Fig. 4 .
Fig. 4. Simplified schematic of the sample flow within the humidified/ambient SMPS and NEPH systems.Red arrows indicate ambient/non-humidified air sampled by SMPS A and NEPHA.Blue arrows indicate humidified air sampled by SMPSW and NEPHW.Note difference in plumbing (green arrows) between SMPSW and SMPSA sheath flows, which is necessary in order to maintain a humidified sheath flow for SMPSW.Flows along the individual lines are listed in L min -1 .

Fig. 5 .
Fig. 5. A: Comparison of number concentration between SMPSW (blue) and SMPSA (green), with SMPSW operating in closed-loop mode with humidification bypassed (calibration test), for 5 consecutive scans.Each PSD is normalized as described in the text.Mode particle diameters for each PSD are displayed for reference.B: Same as A, but with SMPSW in flow-through mode.C: Same as B, but with the humidifier in line (no water).D: Normalized PSDs averaged over all times during normal operations from 15-20 March 2015, whenever SMPSW RH > 85% and SMPSA RH < 40%.

Fig. 6 .
Fig. 6.A: f [RH] vs. RH curves calculated from daily-averaged, corrected NEPHA and NEPHW data from DAK from 15 -20 March 2015 at three wavelengths (450 nm, 550 nm, 700 nm), demonstrating scattering enhancement at high RH during the period.Only data when humidified RH is > 80% (wet) and ambient RH is < 40% (dry) are included.Humidified (ambient) RH is the average of RH observed by sensors located upstream and downstream of NEPHW (NEPHA), downstream (upstream) of the humidifier.B: Same as upper panel, but for GF [RH] vs. RH curves calculated using dailyaveraged, SMPS geometric mean data.
instrument setup, daily operating, maintenance, and calibration routines are well documented in a set of SOPs, which are publically available at http://www.smartlabs.gsfc.nasa.gov.

Table 2 .
Listing of instrument zero-air test lengths and intercomparison tests completed before, during, and after each campaign.
decreased, which allows COMMIT to perform in challenging environments with less manpower.Following a deployment, raw data files are quality-controlled and processed into Network Common Data Format (netCDF) for archiving.