Lake Spray Aerosol: A Chemical Signature from ... - ACS Publications

Aug 22, 2016 - Herein, lake spray aerosol (LSA) is defined as aerosol generated from freshwater through bubble bursting, analogous to SSA from seawate...
0 downloads 0 Views 7MB Size
Article pubs.acs.org/est

Lake Spray Aerosol: A Chemical Signature from Individual Ambient Particles Jessica L. Axson,† Nathaniel W. May,‡ Isabel D. Colón-Bernal,‡ Kerri A. Pratt,*,‡,§ and Andrew P. Ault*,†,‡ †

Department of Environmental Health Sciences, ‡Department of Chemistry, and §Department of Earth and Environmental Sciences, University of Michigan, Ann Arbor, Michigan 48109, United States S Supporting Information *

ABSTRACT: Aerosol production from wave breaking on freshwater lakes, including the Laurentian Great Lakes, is poorly understood in comparison to sea spray aerosol (SSA). Aerosols from freshwater have the potential to impact regional climate and public health. Herein, lake spray aerosol (LSA) is defined as aerosol generated from freshwater through bubble bursting, analogous to SSA from seawater. A chemical signature for LSA was determined from measurements of ambient particles collected on the southeastern shore of Lake Michigan during an event (July 6−8, 2015) with wave heights up to 3.1 m. For comparison, surface freshwater was collected, and LSA were generated in the laboratory. Single particle microscopy and mass spectrometry analysis of field and laboratory-generated samples show that LSA particles are primarily calcium (carbonate) with lower concentrations of other inorganic ions and organic material. Laboratory number size distributions show ultrafine and accumulation modes at 53 (±1) and 276 (±8) nm, respectively. This study provides the first chemical signature for LSA. LSA composition is shown to be coupled to Great Lakes water chemistry (Ca2+ > Mg2+ > Na+ > K+) and distinct from SSA. Understanding LSA physicochemical properties will improve assessment of LSA impacts on regional air quality, climate, and health.



INTRODUCTION Wave breaking and subsequent bubble bursting on large bodies of water produce atmospheric aerosols, with the ocean being one of the greatest global sources of atmospheric particulate matter.1,2 While the production and properties of sea spray aerosol (SSA) have been an important research topic for more than 60 years,3−12 only two studies have examined the production of aerosols from large bodies of freshwater.13,14 These studies specifically focus on the Laurentian Great Lakes and have implications for regional air quality, climate, and health.15 The aerosol produced from freshwater environments, defined here as lake spray aerosol (LSA), is expected to have different physicochemical properties than SSA due to major differences in water salinity and chemical concentrations. Specific differences include lower total salt concentrations, by ∼250 times, and differing composition, with [Ca2+] > [Mg2+] > [Na+] > [K+] common for freshwater, compared to [Na+] > [Mg2+] > [Ca2+] > [K+] for seawater. Freshwater total organic carbon (TOC) concentrations are also of the same order of magnitude as freshwater inorganic ion concentrations, whereas TOC is 102−104 times lower than inorganic ions in seawater.16 Thus, Great Lakes LSA are expected to have different heterogeneous reactivity17 and abilities to act as cloud condensation nuclei (CCN)18−20 and ice nuclei (IN),21−24 in comparison with SSA. Additionally, LSA may lead to the © XXXX American Chemical Society

emission of species harmful to human health, such as heavy metals25 or toxins during harmful algal blooms.26,27 Additional freshwater studies in lakes have shown that cyanobacteria and the toxin microcystin can be ejected from lakes during recreational activities.26−28 Determining the size and chemical composition of LSA particles is crucial for understanding potential regional climate and health effects. To date, only one study has identified the ambient presence of LSA. Slade et al.13 observed LSA with an aerosol number size distribution mode of 15−40 nm during low altitude flights over Lake Michigan. Strong evidence from wind speed and vertical profiles indicated that these particles were produced from wave breaking, particularly since they only occurred when wind speeds were high enough to produce whitecaps (>3.5 m s−1).13 A second peak may have been present in the accumulation mode during these measurements (∼200 nm),13 but there was not conclusive data to determine its origin. Regional chemical transport modeling incorporating the ultrafine lake-derived aerosol mode showed that wave breaking particle production could account for 5−20% of particles over the Great Lakes.15 A Received: April 4, 2016 Revised: August 19, 2016 Accepted: August 22, 2016

A

DOI: 10.1021/acs.est.6b01661 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Figure 1. (A) Beach sampling site at Van Buren State Park (star) (Adapted from Google Map data. Copyright 2016). (B) Photograph of waves taken from the shore on July 7, 2015, between 14:00 and 15:00 EST (Photo credit: N.W.M.). (C) HYSPLIT backward trajectories of air masses at three starting altitudes (100 (red), 500 (blue), and 1000 (green) m) (Imagery adapted from Landsat, NOAA, Data SIO, NOAA, U.S. Navy, NGA, GEBCO. Copyright 2016. Map data adapted from Google. Copyright 2016). Trajectory air mass heights are provided in Figure S1. (D) Measured wave height mapping over Lake Michigan during sampling (Wave height data NOAA GLERL. Imagery adapted from Landsat, NOAA, Data SIO, NOAA, U.S. Navy, NGA, GEBCO. Copyright 2016. Map data adapted from Google. Copyright 2016).

to the presence of other materials such as organic or biological material in the Lake Michigan sample.14 In this study, atmospheric particles were collected during a wave breaking event on the southeastern shore of Lake Michigan in July 2015. A co-located surface freshwater sample was collected and used to generate LSA in the laboratory using the recently constructed LSA generator,14 providing a direct comparison of the ambient LSA physicochemical properties. A chemical signature for ambient LSA was identified from both microscopy/spectroscopy and mass spectrometry, which was compared with the chemical composition of laboratorygenerated LSA of co-located lake water. This information will be used for future identification of atmospheric LSA during field measurements, modeling of particle mixing state,36 and for improved treatment in regional models.15

limitation of the modeling study was the utilization of a SSA source function due to the lack of an available LSA source function. Given the differences in bubble size distributions between freshwater and seawater,14,29−32 these SSA source functions are likely a poor representation for LSA generation. Chung et al.15 noted that further information regarding the generation and properties of LSA is needed to ensure more accurate modeling of these particles and evaluation of their potential regional impacts. Previous laboratory studies have examined the bubble plume and initial droplets produced from freshwater wave breaking.14,33−35 However, only recently did May et al.14 examine both bubble bursting and subsequent LSA production through the use of a laboratory-based LSA generator. In that study, a freshwater sample from Lake Michigan and solutions of inorganic ions at concentrations ranging from that of the Great Lakes to seawater were examined for bubble and aerosol size distributions.14 It was found that LSA particles from surface freshwater exhibited a bimodal size distribution, with a mode at 46 nm, which is slightly larger than that observed by Slade et al.,13 and a mode at 180 nm in the accumulation mode, which is near their second possible mode.14 The presence of two modes was primarily observed when low ion concentrations representative of the Great Lakes (0.05−0.15 g L−1) were used, as opposed to one mode for higher seawater concentrations.14 The variation in the size distributions between the synthetic and collected freshwater was attributed



EXPERIMENTAL METHODS Atmospheric Particle Sampling. Atmospheric particle sampling was conducted from July 6, 2015 23:45 to July 8, 2015 12:45 EST (Table S1) in Van Buren State Park, South Haven, MI, at a beach site along Lake Michigan (42.333206 N, −86.309986 W; 197 m) during a wave event (Figures 1A and S1). An AeroTrak hand-held airborne particle counter (AeroTrak 9306; TSI, Inc.) and a microanalysis particle sampler (MPS; California Measurements, Inc.) were used to measure size-resolved aerosol number concentration and impact ambient particles, respectively. The AeroTrak measured B

DOI: 10.1021/acs.est.6b01661 Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

particles in the 0.7−5.0 μm size range, scanning electron microscopy (SEM-EDX) and computer controlled scanning electron microscopy (CCSEM-EDX) were run on an FEI Nova (JEOL). The FEI Nova contains an environmental dual FIB/ SEM with Schottky field emitters operating at 15 kV and secondary electron detectors used to collect images of individual particles. An EDAX UTW detector was used to collect EDX spectra of individual particles. The elements used for CCSEM-EDX were C, O, Na, Mg, Si, P, S, Cl, K, Ca, and Fe. Although Al may be present in the sample, it was not examined using the CCSEM-EDX due to overwhelming signal from the Al foil substrate. Contributions of Al are marked appropriately in the raw spectra. A total of 185 ambient and 99 generated particles were chemically analyzed. Computer controlled techniques were unable to be used on particles 10 μm. The three stage MPS was operated at 2 L min−1 with aerodynamic diameter (Da) size ranges of 2.5−5.0 μm (stage 1), 0.7−2.5 μm (stage 2), and