University of Florida Thesis Or Dissertation

Home , Bioprecipitation, Cumulonimbus cloud, Hymenobacter aerophilus, Ice pellets, Perlucidibaca

BIOPRECIPITATION: THE CONNECTION BETWEEN MICROBIOLOGY AND METEOROLOGY

RACHEL ELAINE JOYCE

A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2020

To Mom and Dad

ACKNOWLEDGMENTS

I want to first and foremost thank my advisor, Dr. Brent Christner, who has made this research possible. His mentorship and support gave me the confidence to carry out my dissertation research. Without a doubt, I would not be the scientist that I am today, with such an unwavering passion for scientific research, if I had not gotten the encouragement that I received from Dr. Christner, and for that I am incredibly grateful. I would also like to thank all of my committee members for agreeing to oversee my PhD and offering me guidance and advice: Dr.

Bryan Kolaczkowski, Dr. Ana Conesa, and Dr. Corene Matyas. I want to acknowledge the friends who proved to be an incredible support system over the last six years: To James

Ramsden, who has worked with me in the lab for over a year and is always willing to go above and beyond to get work done; To Christina Davis, who was my first friend and confidant at UF; and Rachel Moore, whose sheer excitement about bioaerosols and microbiology is enough to keep me going even on my worst days; To the ladies of LSU--Heather Lavender, Dr. Noelle

Bryan, and Dr. Amanda Achberger—who even after four years of moving apart, are always there to offer me a shoulder to lean on; To my two biggest role models, my mom and dad, who showed me how far hard work and dedication can get you in life; To all three of my siblings,

Alex, Chris, and Luke, because I love them more than anything; And to my rock, Christopher

Wilson, who has kept me grounded and sane and smiling even through the toughest of times.

This research was funded by The National Science Foundation (DEB-1241161 and -1643288).

TABLE OF CONTENTS

page

ACKNOWLEDGMENTS ...... 4

LIST OF TABLES ...... 8

LIST OF FIGURES ...... 10

LIST OF OBJECTS ...... 12

LIST OF ABBREVIATIONS ...... 13

ABSTRACT ...... 15

CHAPTER

1 INTRODUCTION ...... 17

Ice Nucleation ...... 17 The Physics of Water and Ice Nucleation ...... 17 Heterogeneous Ice Nucleation in the Troposphere ...... 20 Discovery of Biological Ice Nucleation ...... 24 Biological Ice Nucleation ...... 26 Characterization of the Bacterial Ice Nucleation Protein ...... 26 Ice Nucleation in Fungi, Lichens, Pollen, and Algae ...... 31 Applications of Biological Ice Nucleation ...... 33 Aeromicrobiology ...... 34 Airborne Microorganisms and Disease Transmission ...... 34 Emissions of Airborne Microorganisms ...... 38 Hight Altitude and Cloud Microbiology ...... 41 Potential Influence of Biological Ice Nuclei on Meteorology ...... 43 Biological INPs in Air, Cloud, and Precipitation Samples ...... 43 Challenges and Limitations to Studying Biological INPs ...... 48 Use of Biological INPs in Numerical Cloud and Climate Models ...... 51 Purpose of This Research ...... 56

2 CHARACTERIZATION AND SOURCE IDENTIFICATION OF BIOLOGICAL ICE NUCLEATING PARTICLES DEPOSITED YEAR-ROUND IN SUBTROPICAL PRECIPITATION ...... 60

Overview ...... 60 Methods ...... 62 Precipitation Sampling ...... 62 Quantification of INPs and Cells ...... 63 Amplification and Sequencing of 16S rRNA Genes ...... 64 Inorganic and Organic Chemistry ...... 66

Analysis of Meteorological Data and Ecoregions ...... 68 Statistical Analyses ...... 72 Results...... 73 Total, Biological, and Bacterial INPs in Louisiana Precipitation ...... 73 Exploratory Factor Analysis of the INP Data ...... 75 INP Factor Concentrations Correlate With the Physical, Chemical, and Microbiological Data ...... 76 INP Factors Correlate with Season, Cloud Type, and the Air Mass History ...... 77 Correlations Between Abundances of Bacterial Operational Taxonomic Units and INPs...... 79 Discussion ...... 80 INP “Classes” Identified in Louisiana Precipitation ...... 80 Potential Geographic Origins of INP Classes in Louisiana Precipitation ...... 82 INP Concentration as a Function of Season and Meteorology ...... 84 Potential Phylogenetic and Geographic Sources of the Ice Nucleating Bacteria ...... 85 Potential Implications for Biological INPs on Meteorological Processes ...... 87 Concluding Remarks ...... 89

3 INDUCTION OF ICE NUCLEATION ACTIVITY IN NOVEL ICE+ BACTERIA ...... 100

Overview ...... 100 Methods ...... 102 Sample Collection ...... 102 Bacterial Culturing ...... 102 Immersion Freezing Assays for Ice Nucleation Activity ...... 104 Nutrient Limitation Experiments ...... 105 Identification of the Bacterial Isolates ...... 106 Results...... 107 Primary Enrichments of Arid Topsoil and Sleet Samples ...... 107 Dependence of Ice Nucleation Activity on Culture Age ...... 107 Induction of IN Activity in Isolates AZ_82Pink and AZ_82Red ...... 110 Characterization of the Bacterial INP in AZ_82Pink ...... 112 Discussion ...... 113 Concluding Remarks ...... 121

4 SIZE-RESOLVED BIOLOGICAL INPS IN NIMBOSTRATUS PRECIPITATION ...... 129

Overview ...... 129 Methods ...... 130 Precipitation Collection ...... 130 Quantification of INPs and Cells ...... 131 Analysis of Meteorological Data ...... 133 Results...... 133 Stratiform Storm Morphology and Origin ...... 133 Spatial Differences in INP Concentrations from Baton Rouge, LA Stratiform Precipitation ...... 135 Temporal Differences in INP Concentrations from Baton Rouge, LA Precipitation ....136

Size-Resolved INP Concentrations ...... 136 Discussion ...... 137 Concluding Remarks ...... 141

5 INFLUENCE OF BIOLOGICAL ICE NUCLEATING PARTICLES ON PRECIPITATION FORMATION IN NIMBOSTRATUS CLOUDS ...... 148

Overview ...... 148 Methods ...... 150 Numerical Model ...... 150 Idealized Tests ...... 152 Empirical Biological INP Parameterization ...... 153 Results...... 155 Discussion ...... 156 Concluding Remarks ...... 160

6 CONCLUSIONS ...... 163

APPENDIX

A SUPPLEMENTAL INFORMATION FOR CHAPTER 2 ...... 174

B SUPPLEMENTARY INFORMATION FOR CHAPTER 3 ...... 185

LIST OF REFERENCES ...... 187

BIOGRAPHICAL SKETCH ...... 221

LIST OF TABLES Table page

1-1 Variations in the basic habits of ice crystals with temperature, adapted from Wallace and Hobbs, 2006...... 57

1-2 Documented IN-active bacteria with sequenced ina genes...... 59

2-1 Average differential concentrations of total, biological, and bacterial INPs ...... 93

2-2 Descriptive statistics of numerical variables measured in this study...... 94

2-3 Results of exploratory factor analysis for total, biological, and bacterial INPsa ...... 95

2-4 Correlations of INP factors with physical and chemical measurements of the precipitation, shown as Pearson correlations coefficients calculated between factors of EFA and measured variables of precipitation ...... 96

3-1 Collection sites of samples from which bacteria were cultured in this study...... 122

3-2 Broth recipes used for nutrient deprivation experiments...... 123

3-3 Broth recipes used for nutrient deprivation experiments...... 123

3-4 Results of sequence identification of potential Ice+ bacterial isolates...... 125

3-5 Bacterial isolates received from Boris Vinatzer (Virginia Polytechnic Institute) that demonstrated IN activity at -4oC...... 126

4-1 Sampling locations and times, with respective precipitation types collected and the ecoregions that the storm origins interacted with...... 142

A-1 Results of Multiple Imputation for missing INP data...... 174

A-2 MANOVA results of INP concentrations, interactions of air masses, and ecoregions. .174

A-3 Characteristics of fluorescent dissolved organic matter PARAFAC components in precipitation from air masses interacting with distinct ecoregions...... 175

A-4 Results of multivariate analysis of variance (MANOVA) for all INP (total, biological, and bacterial) concentrations as a function of season, cloud type, and precipitation type...... 176

A-5 Correlations between ice nucleating particle (INP) factors and local meteorological conditions...... 176

A-6 Significant Spearman’s rank correlation coefficients (for which rho ρ≥0.40; and significance p<0.05) between ice nucleating particle (INP) factors and taxon abundance...... 177

A-7 Spearman correlations between ice nucleating particle (INP) factors...... 181

LIST OF FIGURES Figure page

1-1 Simplified diagram depicting the microphysical processes that can occur within a mixed-phase cloud...... 57

1-2 The various types of clouds that occur in the earth’s troposphere...... 58

2-1 Location and extent of the source ecoregions relevant in this study...... 92

2-2 Concentrations of total, biological, and bacterial ice-nucleating particles (INPs) at activation temperatures of –4 to –15°C...... 92

2-3 Results of ﬂuorescent dissolved organic matter (DOM) excitation-emission matrices data investigated using parallel factor (PARAFAC) analysis...... 95

2-4 Average INP factor concentrations as a function of ecoregion classiﬁcation...... 97

2-5 Average INP factor concentrations as a function of cloud type, season, and precipitation phase...... 98

2-6 Summary of trends in ice-nucleating particle (INP) data with source ecoregion and physical, chemical, and microbiological properties of the precipitation...... 99

3-1 Dependence of culture age on IN activity of bacterial isolates. Temperature activity shown is the warmest temperature that freezing was observed for that isolate...... 124

3-2 Gram stains of the two most IN-active bacterial isolates...... 126

3-3 Induction of ice nucleation activity in AZ_82Pink using nutrient deprivation...... 127

3-4 Nutrient deprivation experiments conducted on AZ_82Red...... 128

3-5 Inhibition of ice nucleation activity in AZ_82Pink (Kocuria sp), at 14 days of incubation...... 128

4-1 120-hour backward air mass trajectories for Chicago, IL (Dec. 28, 2015); Atlanta, GA (Jan 6-7, 2017); Baton Rouge, LA (Oct. 24-25, 2015); Gainesville, FL (Jan. 7, 2017) precipitation collections...... 143

4-2 120-hour backwards air mass trajectories for the temporal precipitation collections made in Baton Rouge, LA, on Dec. 19-20, 2015...... 144

4-3 0.22 μm size-cut for INP concentrations taken from Hurricane run-off during RAINS campaign in Baton Rouge, LA...... 145

4-4 Temporal variability of Total, Biological, and Bacterial INPs...... 146

4-5 Size-resolved concentrations of biological INPs detected in mixed-phase precipitation collected during the Chicago winter storm on Dec 27, 2015...... 147

4-6 Concentration of biological INPs from mixed-phase precipitation samples collected from Atlanta, GA...... 147

5-1 Vertical cross section of the cloud produced by the 2D idealized simulation of the mixed-phase precipitation collected outside of Chicago, IL, on Dec. 28, 2015...... 161

5-2 Total amounts of non-convective rain and snow (mm) plotted as a function of biological INP content. The dust INPs-only simulation is represented by “0 Bio INPs”...... 162

A-1 Cloud formation mechanisms and HYSPLIT trajectory analysis...... 182

A-2 PARAFAC Components fluorescence intensity profiles based on The North American Ecoregion classifications used in this study...... 183

A-3 Significant differences in DNA operational taxonomic unit (OTU) abundances as a function of cloud type and season...... 183

B-1 Results of preliminary nutrient deprivation experiments for isolate AZ_122...... 186

B-2 Results of preliminary nutrient deprivation experiments for isolate AZ_8...... 186

LIST OF OBJECTS Object page

2-1 Excel file, 67KB in size, in tabular format, of all data collected and analyzed for Chapter 2...... 71

LIST OF ABBREVIATIONS

CCL Convective condensation level: Altitude within the troposphere at which condensation of water vapor occurs when air is lifted due to convection (heating and rising of a parcel of air in an unstable atmosphere)

CEC Commission for Environmental Cooperation: An agency established by Canada, Mexico, and The United States of America as part of the North American Agreement on Environmental Cooperation. This agency serves the purpose of protecting the environment by promoting cooperation between these nations and the public.

DOC Dissolved organic carbon: The carbon fraction of DOM

DOM Dissolved organic matter: Soluble organic molecules, predominantly composed of carbon, nitrogen, or phosphorous, that are dissolved within aqueous solutions

EEMs Excitation emission matrices: A three-dimensional scan produced by fluorescence spectroscopy which produces a contour plot with excitation and emission wavelengths on the x- and y-axis and intensity on the z-axis. This method of chemical analysis provides qualitative and quantitative information on the presence fluorescent molecules within an aqueous solution.

EFA Exploratory factor analysis: A multivariate statistical procedure which groups together similar variables based on the pattern of matrix data across all observations and all variables

EPA Environmental Protection Agency: of The United States of America which serves the purpose of protecting the environment

HYSPLIT Hybrid Single-Particle Lagrangian Integrated Trajectory: Atmospheric model created by The National Oceanic and Atmospheric Administration that can calculate the history of an air mass at any particular altitude within the atmosphere.

IN Ice-nucleating: The act of nucleating ice

INP Ice nucleating particle: A particle which catalyzes the conversion of liquid or gaseous water to ice

LCL Lifted condensation level: Altitude within the troposphere at which condensation of water vapor occurs when air is forcibly lifted by mechanical means (frontal systems, convergence of air masses, or orographic lifting)

mAGL Meters above ground level

MANOVA Multivariate analysis of variance: A multivariate statistical procedure which accounts for variances within and between group values in order to determine whether there is any significant difference between the groups in question

MBL Mixed boundary layer: Lower layer of the troposphere where turbulent mixing with the Earth’s surface occurs.

OTU Operational taxonomic unit: A group of related organisms based on a threshold of similarity among their DNA sequences. The taxonomic level of sampling, which is defined by the user, can correspond to bacterial populations, species, or genera.

PARAFAC Parallel factor analysis: A statistical method used to analyze the output of EEMs and identify recurring areas of high fluorescent intensities. PARAFAC profiles aid in characterizing the dissolved organic matter present within an aqueous sample.

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

BIOPRECIPITATION: THE CONNECTION BETWEEN MICROBIOLOGY AND METEOROLOGY

Rachel Elaine Joyce

May 2020

Chair: Brent C. Christner Major: Microbiology and Cell Science

At subzero temperatures warmer than -36oC, the phase change of liquid water or vapor to ice requires the presence of an ice nucleating particle (INP). In the atmosphere, this process is

critical to the formation of precipitation, and the INPs thought to be most important to the

process are mineral dusts. However, at temperatures > -15oC, mineral dusts are not ice nucleation

(IN)-active--only certain IN-active microorganisms can catalyze freezing at these temperatures.

Despite the detection of microbial INPs in clouds and precipitation, the lack of information on their diversity, spatiotemporal occurrence, and tropospheric concentrations has made deciphering their role in precipitation formation difficult. This study aims to bridge these gaps in knowledge by studying the patterns of biological INP occurrence in precipitation over a two year period; investigating the potential source environments of biological INPs; and numerically simulating their effects in mixed-phase clouds to determine their contribution to precipitation formation.

The results show that the highest concentrations of biological INPs in precipitation collected in the southeastern USA were sourced from distant terrestrial environments such as the high northern latitudes and East Asia, and were positively correlated with bacterial taxa not previously

known to be IN-active, including genera from the Bacteroidetes and Firmicutes. Novel IN-active

species from bacterial phyla not previously recognized as harboring IN-active organisms were 15

isolated from Mojave Desert soils and Virginia precipitation. Several of these novel strains significantly increased their IN activity following nitrogen- and/or carbon-limitation and incubation at 4oC. Wintertime, nimbostratus clouds produced precipitation with the highest

concentrations of biological INPs, and a size-resolved analysis indicated that the majority were

<0.1 μm in diameter. Numerical cloud simulations indicated that the cumulative concentrations

of biological INPs active at ≥ -10oC estimated within nimbostratus clouds (~1 biological INP m-

3) enhanced precipitation formation in the events analyzed. Taken together, these results

demonstrate that the ecology and diversity of IN-active bacteria may be much broader than

previously thought, and that arid regions and/or high northern latitudes may be significant

sources of biological INPs that have the potential to influence precipitation formation thousands of kilometers from their emission source.

CHAPTER 1 INTRODUCTION

Ice Nucleation

The Physics of Water and Ice Nucleation

The process by which water molecules transform from a less ordered state to a more ordered state is called nucleation (1). When water vapor is nucleated into liquid water, it is called condensation nucleation (2). Likewise, when water vapor or liquid water is nucleated into ice, the process is called ice nucleation (2). At the molecular level, the transitioning between these phases requires bringing the thermodynamics, or heat and energy, of the system from a stable state to a temporarily unstable state (1). The changing of thermodynamic instability can, and does, happen at a random rate due to thermal vibrations, which can lead to the spontaneous formation and deformation of small water and ice molecule clusters, called embryos (3–5).

Embryos are not thermodynamically stable, and do not ensure the permanent transition of phases from vapor to liquid or solid, or from liquid to solid (6). In order to permanently change states, an embryo must undergo more clustering to reach a “critical” nucleus size, and only then has nucleation to a new phase begun (7). Nucleation may therefore be looked at as the first irreversible formation of a new equilibrium phase.

Reaching a critical nucleus size, however, requires surmounting the “nucleation” barrier, which is the amount of Gibbs’ free energy that needs to be overcome by the water embryos to cluster together into a nucleus (1). Considering first the nucleation of vapor to liquid, the kinetics of this process has been shown experimentally to be directly related to the radius of the embryo formed, the free energy of the system, and the degree of supersaturation of the system (6). An appreciable amount of work would need to be put into the system to force high energy water

1Erwinia herbicola has since been reclassified as Pantoea agglomerans, and more recently, Enterobacter agglomerans 2Erwinia ananas has since been reclassified as Pantoea anantis 3Erwinia uredovora has since been reclassified as Pantoea ananas

vapor molecules into a lower energy liquid or solid state (8). So much so, in fact, that this spontaneous condensation process—formally referred to as homogeneous nucleation--would only be possible in the earth’s atmosphere at supersaturations of 2000% (9). Such a degree of supersaturation is not physically possible in the earth’s atmosphere. Yet, we observe clouds, a direct product of condensation nucleation, in our atmosphere every day.

This is because there are other nucleation processes relevant to the atmosphere, which are dependent upon the presence of airborne, micrometer to nanometer-sized particles called aerosols (2). Aerosols assist in the nucleation of water vapor molecules to liquid water by acting as substrates upon which water vapor can collect and condense, in a process called heterogeneous nucleation (2). In the context of cloud formation, these aerosols are termed cloud condensation nuclei, or CCN (Figure 1-1; 10, 11). CCN aid in lowering the nucleation barrier by decreasing the exposed surface area of the forming nucleus, which directly effects the free energy of the system (1). CCN that are good at heterogeneous nucleation were first thought to be caused only by insoluble aerosol particles, however research over the last century has shown that almost any aerosol particle, soluble and insoluble, can act as a CCN (2). The efficacy of the aerosol particle to act as a CCN is only dependent upon the supersaturation of the atmosphere, the size of the particle, and the curvature of the aerosol particle, which directly effects the amount of exposed surface area of the forming embryo (1, 2, 12). While potentially any aerosol particle can act as a CCN, water-soluble and hygroscopic particles are the most efficient at initiating droplet formation. A plethora of field research has shown that the particles which most commonly initiate condensation of water vapor include combustion products such as black carbon or soot, sulfates derived from volcanic activity, sea salt from oceans, or sulfur dioxide

and other secondarily produced organic matter that result from the oxidation of volatile organic

compounds.

The same fundamental concepts of condensation nucleation can be applied to ice

nucleation in the atmosphere. Homogeneous freezing of pure liquid water will not occur until

temperatures below approximately -36oC. This is because the change of phase from liquid to ice

requires the breaking of water-to-water hydrogen bonds and the formation of water-to-ice

hydrogen bonds (2). This process, much like the condensation process, requires that a water

molecule move from its liquid equilibrium energy state to a new equilibrium energy state of solid

ice (13). These two positions of equilibrium energy are again separated by a nucleation barrier

(7), and at lower temperatures, this barrier is decreased (8). Thus, at sufficiently low

temperatures (<-36oC), the probability of pure water freezing spontaneously is increased significantly, and homogeneous nucleation of ice can occur (13).

While the process of homogeneous ice nucleation does occur in the earth’s atmosphere, it

cannot account for the substantial amounts of ice that are formed at temperatures above -36oC

(14). Indeed, heterogeneous ice nucleation, much like heterogeneous condensation nucleation, is

made possible by impurities that act as substrates upon which freezing can occur. These

impurities are called ice nucleating particles (INPs) (Figure 1-1; 15). However, whereas

essentially any aerosol can act as a CCN, the number of aerosols which can act as INPs are far

fewer (2). Long-standing theoretical considerations of ice nucleation suggested that INPs must

be insoluble and contain “active sites” which initiate freezing (2). Experimental results suggested

that the number of active sites on a given particle increased or decreased in scale with the size of

the particle, thus making larger particles (> ~500 nm) more effective INPs (16–18). However,

research over the last several decades has indicated that soluble, nanoscale particles can also act

as effective INPs (19–24).

The physical structure of the INP is thought to be of importance as well. Theory suggests

that a crystal lattice surface which mimics that of ice would make a good template upon which

water molecules can organize (2). Molecular dynamics and empirical studies performed on

graphitic soot (25), silver iodide (26), and volcanic ash (27) appeared to confirm this theoretical

prerequisite. However, other evidence suggested that the chemical nature of the INP active sites

may be of more importance. Indeed, an INP that possesses ligands capable of forming hydrogen

bonds which are similar in strength and polarity to the water molecule is more likely to interact

with an available water molecule (2). Experimental and observational data suggests that

functional groups such as hydroxyl (-OH) groups play an important role in ice formation (28–

30). While, there is still much uncertainty on what, exactly, makes a molecule or particle an

effective INP, scientists do know that ice itself is the best ice nucleus. However, since

homogeneous nucleation doesn’t occur at temperatures above -36oC, the presence of INPs are

very important for ice formation within the earth’s lowest layer, the troposphere (2, 9).

Heterogeneous Ice Nucleation in the Troposphere

Ice formation in the earth’s troposphere has a significant impact on a number of physical processes, including precipitation formation, cloud electrification, and radiative forcing (31).

While precipitation can and does form through the collision and coalescence of liquid water droplets (Figure 1-1, (2)), most precipitation, including rain, starts off as ice in the cloud (32, 33).

At subzero temperatures warmer than -36oC, ice may form through several modes of heterogeneous ice nucleation (Figure 1-1): Contact nucleation, whereby contact of supercooled liquid with an INP initiates its freezing; Condensation nucleation, whereby liquid water

condenses onto an INP and freezes in the process;. Immersion nucleation, whereby an INP is

immersed in liquid and at cooling temperatures initiates ice formation; and deposition nucleation,

whereby water vapor deposits directly onto the INP (2).

Ice particles in clouds with significant vertical development (convective clouds, Figure 1-

2), can collect other ice particles in a process called aggregation, which can lead to snowflake

formation (Figure 1-1, (9)). However, the process of aggregation depends strongly on

temperature (34). Indeed, the likelihood of freezing is increased at subzero temperatures above -

5oC, which makes the ice stickier and more likely to adhere to other ice particles (34). The ice crystal shapes, which are also dictated by temperature (Table 1-1), play an important role in aggregation as well. Delicate crystals, such as dendrites, can easily become aggregated and entwined with one another (34). In fact, the shape of crystals that form through primary ice nucleation is also dependent on temperature (Table 1-1). Thin hexagonal plates are the dominant ice shape formed at -4oC and warmer; temperatures between -4oC and -10oC usually lead to the formation of needles and hollow columns; temperatures between -10oC and -22oC tend to see

sector plates and dendrites forming; and temperatures colder than -22oC lead to formation of hollow columns (2).

The majority of the precipitation-producing clouds that exist in the earth’s troposphere are mixed-phase, meaning that they contain both ice particles and liquid water. In mixed-phase clouds, supercooled liquid water, which is liquid water that exists below the melting point, can collect and freeze onto existing ice particles in a process called riming (Figure 1-1). Riming typically leads to the production of precipitation in the form of graupel or sleet (9). However, when sufficient vertical velocities are present, as can be the case in convective clouds, riming can lead to hail formation through supercooled liquid accretion onto the ice surface in multiple

rounds of vertical ascent and descent (9). Similar to aggregation processes, riming efficiencies

can also be affected by the shape of the ice crystals being rimed, with branched crystals, such as

needles and dendrites, grow quicker and more efficiently than unbranched crystals (Table 1-1).

Since the phase-change of liquid or vapor to ice leads to a release in energy known as

latent heat, ice formation on a large scale within a cloud in the troposphere can also lead to

invigoration of updrafts and increased vertical extent of the cloud, further promoting

development of precipitation (9). Additionally, when sufficient updrafts are present within a

developing cloud, several processes of secondary ice formation, or ice enhancement, can occur

(Figure 1-1 (35–39)). Ice in convective clouds can undergo collisional (38, 40, 41) or thermal

shock (42–45) fracturing, whereby the ice particles that form break apart due to collision with

other ice particles or due to severe temperature changes, respectively (Figure 1-1). Clouds with

significant vertical velocities (>1.4 m s-1) and temperatures in between -3oC and -8oC can also

undergo rime-splintering, or the Hallett-Mossop process (Figure 1-1; 36, 46–48). The Hallett-

Mossop process occurs when small splinters of ice break off from the ice particle during the

riming process after collision with large (>23 μm) supercooled droplets (9, 36). Lastly, dendritic

ice crystals that are exposed to dry air can break apart as they evaporate, producing more

fragmented ice crystals, in a process called evaporative fracture (Figure 1-1; 31, 49, 50). All the ice crystals produced in these secondary processes may go on to aggregate, rime, or produce more ice crystals in secondary ice processes. Thus, secondary ice formation can lead to an exponential increase in the number of ice particles within a cloud.

Most of the ice enhancement processes described above require clouds with sufficient depth and vertical velocities to occur. Stratiform clouds, on the other hand, typically do not undergo the explosive secondary ice formation processes described above and are thus likely to

require higher concentrations of INPs for precipitation formation. Indeed, stratiform clouds have less vertical development when compared to convective clouds, and as such, typically have warmer (≥−20°C) temperatures (9). As such, these clouds would require INPs active at warmer temperatures to allow for ice crystal formation. Additionally, due to a general lack of strong vertical velocities--which would normally allow for sufficient ice formation via secondary mechanisms such as the Hallet-Mossop process—stratiform clouds may produce their precipitable fractions in the form of ice via the Wegner-Bergeron-Findieson (WBF) mechanism

(51–53). The WBF mechanism proposes that in a mixed-phase cloud with updrafts ≤ 3 ms−1 and e1 > e > ei (where e = vapor pressure of air; e1 = saturation vapor pressure of over liquid; and ei = saturation vapor pressure over ice), rapid ice crystal growth occurs at the expense of liquid cloud droplets (54, 55). In fact, global models suggest that the WBF process may drastically increase the ratio of stratiform/convective precipitation produced on a global scale, while also significantly altering the net cloud radiative effect (54).

Ice formation in the troposphere also has a major effect on the global radiative budget

(31). Precipitation-containing clouds with significant vertical extent, such as convective clouds

(Figure 1-2) can alter radiative transfer within the atmosphere by reflecting incoming solar radiation (31, 56). Likewise, low to mid altitude stratiform clouds (Figure 1-2) can cool the earth by reflecting shortwave radiation out of the atmosphere, a process termed “the solar albedo effect” (57). At the same time, these thicker clouds can warm the earth through a “blanketing” effect that keeps outgoing longwave radiation in the atmosphere, particularly at night, when geothermal radiation from the earth’s surface is otherwise lost to space (58). On the other hand, high altitude and optically-thin ice clouds, such as cirrus clouds (Figure 1-2), have a net warming effect on the earth because their greenhouse effect is greater than their solar albedo

effect (56, 59). However, the degree to which various types of ice-containing clouds affect the global climate depends on many factors, including the altitude of the cloud, the thickness of the cloud, and the temperature of the cloud which directly affects ice crystal shapes and as a result, albedo (60–64).

Research on INPs over the last century has indicated that insoluble soil dusts, volcanic dusts, silicates and clay minerals are likely the most important INPs in clouds due to their high abundance in the troposphere (65–70). However, mineral dusts have not been shown to initiate ice nucleation at temperatures above -15oC (71). Thus, as early investigators began examining clouds in-situ, they discovered a perplexing paradox: in many instances, ice crystals were detected at numbers that far exceeded INP concentrations, even in clouds with temperatures warmer than -15oC (35, 72, 73). Such discrepancies at an early age of ice nucleation research led

many atmospheric scientists to believe that INP measurements were inaccurate, or that INPs

were irrelevant for the formation of ice in clouds (74). As described above, research undertaken

in the last half century has shown that for clouds with temperatures well below -15oC, this

phenomenon could be explained by secondary ice formation processes (Figure 1-1). However,

secondary ice formation processes could not explain the finding of ice particles in clouds with

warm subzero temperatures. This remained unexplained until the discovery of an unsuspecting

group of highly active INPs.

Discovery of Biological Ice Nucleation

In the early 1960s, a group of researchers at The University of Wisconsin were trying to

figure out why corn crops within the area were experiencing increased rates of leaf blight

infections (75). Their initial hypothesis was that the rust fungus pathogen, Helminthosporium

turcicum, was causing leaf blight infection to the crops. Thus, the goal of their work was to

identify corn plants that were naturally resistant to the pathogen, which would allow for

subsequent breeding of the immune corn plants. The first step in identifying plants with natural

resistance was to inoculate an experimental crop with powdered corn leaves from infected plants.

However, after several days of incubation, it was evident that the experiment had been

complicated by an unseasonable bout of frost (76). To the researcher’s surprise, the control corn

plants that had not been inoculated with diseased corn leaves had not been subject to the frost.

Further research demonstrated that powdered leaves from healthy field-grown corn, when dusted

onto the leaves of healthy corn seedlings, significantly increased frost damage of the corn

seedlings at -4oC—a temperature far warmer than the typical onset of frost damage to plants

(76). As the researchers studied this phenomenon, they discovered that there was another agent on the diseased plants which had been inducing early frost damage. The addition of antibiotics during these experiments inhibited warm-temperature frost damage, indicating that a bacterium might be to blame for the increased frost sensitivity of the corn crops (77).

Around the same time, another group of researchers were examining organic sources of highly-active INPs for a completely unrelated reason—to identify INPs that could affect weather.

An atmospheric scientist by the name of Gabor Vali, with the help of his graduate student,

Russell Schnell, was analyzing samples rich in organic molecules and found that decomposing vegetation (78) and soils high in organic content contained much more active INPs than pure clays or sands (79). Specifically, they found that actively decaying leaves, especially those undergoing aerobic decay, produced very high concentrations of warm-temperature INPs (78).

Indeed, at just -10oC, these leaf-derived nuclei were occurring at concentrations as high as 1010

INPs per gram of decaying leaf, which was much warmer than any other known INP at the time.

Subsequent follow-up studies showed that highly active leaf-derived INPs occurred in locations

across the globe (80). It appeared, that the INP content of these global samples was strongly

correlated to the climatological conditions of the environment from which the sample was taken,

regardless of plant species, leading Vali and Schnell to believe that it was the microbiota of the

plants which were causing ice nucleation, as opposed to any plant-related molecule (80). And in

fact, the hypothesis that these leaf-derived INPs were microbiological in origin was validated not

long after.

By the mid-1970s, another research group had analyzed the samples of decaying alder

leaves from Vali and Schnell’s studies, and determined that the molecules responsible for

freezing at such warm temperatures were being produced by a bacterium, Pseudomonas syringae

C-9 (81). In the laboratory, they showed that this bacterial strain could induce freezing at -1.9oC.

The only compound with comparable activity at the time was silver iodide, which was shown to

induce freezing at -6oC (26). The ability of an epiphytic plant pathogen, such as P. syringae, to

induce ice formation would certainly confer a selective advantage to the species which possesses

it, as frost damage to plant walls and surrounding tissues would give the bacterium easier access

to nutrients within the plant cell than it would have gotten on the plant surface (76). The impact

that bacterial INPs could have on meteorology (82) and plant disease (83), led other researchers

to quickly join in the hunt to discover novel biological INPs and characterize the extent to which

they existed in nature. It was thus that the era of research on biological ice nucleation was born.

Biological Ice Nucleation

Characterization of the Bacterial Ice Nucleation Protein

The discovery of ice nucleation activity in P. syringae invigorated research on bacterial

INPs. Initial investigations indicated that the molecule responsible for conferring IN activity to

P. syringae was heat-labile and associated with an intact bacterial cell wall, suggesting that it

was of a proteinaceous nature and attached to or associated with the outer membrane of the

bacterial cell (81). Subsequent studies confirmed that the IN activity of P. syringae was indeed a protein, encoded in a single open reading frame (ORF) of approximately 3.6 kb (84). This gene

was subsequently named inaZ, and its protein product, InaZ (Table 1-2). The inaZ gene encodes

a 1200 amino acid protein, approximately 150 kDa in size (20), with a very distinct primary

structure of 122 octapeptide repeats and two higher order periodicities of 16 and 48 amino acid

repeats (85). Deletion analysis of the octapeptides showed that while they were not all necessary

for IN activity, each unit added to the efficacy of the overall activity (85, 86). Further,

aggregation of the full proteins into multimers led to a logarithmic increase in IN activity from -

2oC to -12oC with increasing INP size, until maximum activity was achieved after aggregation of

the proteins to a total size of 19,000 kDa (20). Localization studies confirmed that the protein

was attached to the outer membrane (87), and that at peak activity, large amounts of the protein

clustered together in non-uniform patches on the outside of the cell (88).

As molecular research on the InaZ protein progressed, other species of

Gammaproteobacteria were also shown to possesses the IN-phenotype (Ice+), and to produce IN

proteins that were homologous to the InaZ protein (89) (Table 1-2). Several other species of

Pseudomonas, including Pseudomonas fluorescens (90, 91), Pseudomonas viridiflava (92, 93),

and Pseudomonas borealis (94), as well as species from two other genera of known gram-

negative phytopathogens, including Erwinia herbicola1 (95), Erwinia anana2 (96), Erwinia

uredovora3 (93), and Xanthamonas campestris (97), all demonstrated IN activity. Much like

InaZ, the E. herbicola IN protein product (IceE) and P. fluorescens protein product (InaW) is encoded by a single ORF, approximately 3.8 and 3.6 kb long (respectively), and contained considerable homology to the primary protein structure of the InaZ protein, with octapeptide and

16-48 residue repeats (89). Similarly, the other species and their protein products (Table 1-2) contained homology to the InaZ protein as well but appeared to lack the higher order 16- and 48- residue periodicities. Unlike the Pseudomonas species in which IN activity is demonstrated, several of the Erwinia species, including E. herbicola and E. uredovora are able to shed their IN proteins in outer membrane vesicles (98–100).

While characterization of the primary structures of the proteins was possible following identification of the encoding gene, purification of the proteins for crystallographic studies proved to be a challenge. Given that these bacterial proteins are embedded within the outer membrane or membrane vesicles produced by the bacteria, the tertiary structure of these proteins are not well characterized. However, several studies have used alternative methods to give a best estimate of the folding of the protein. Predictive modeling of the P. borealis protein, based on the crystallographic structures of antifreeze proteins, indicated that it likely folded into beta- helices that can undergo multimerization on the outer membrane of the bacterium (101). The exposed surfaces of the proteins are hypothesized to contain amino acids that hydrogen-bond with water molecules, leading to the formation of an ice-like clathrate coating of the protein, as is commonly seen in antifreeze proteins (101). Studies using sum frequency generation spectroscopy and molecular dynamics show that P. syringae’s InaZ protein exposes active sites on the outer membrane surface that consist of repeating hydrophilic-hydrophobic amino acid- residues (102). The protein’s repetitive amino acid residues give it the ability to hydrogen bond with water molecules and force them into a highly ordered alignment similar to the ice lattice.

Additionally, as theorized by previous studies, the multimerization of the protein on the surface of the bacterium leads to more effective freezing at warmer temperatures (102). In fact, it appeared that this multimerization of the protein could be enhanced by manipulating the

conditions in which the bacteria are grown, however the mechanism behind this emergent

property is poorly understood.

Since the discovery of Ice+ P. syringae, dozens of its strains have been shown to possess

the phenotype—however not all strains within P. syringae, or within other Ice+ species, possess

the phenotype (76). Additionally, the nucleation frequency (that is, the number of cells within a

population that can nucleate ice) vary significantly across strains and species. Nevertheless,

certain strains have been demonstrated to have very high nucleation frequencies, with almost

every single cell in the population able to nucleate ice. Indeed, using P. syringae strain T1, a

pathovar of tomato plants, it was shown that the highest IN activity of the bacterium was

obtained after limiting it for nitrogen, phosphorous, sulfur, and iron, as well as incubation in the

cold (103). Additionally, detectable IN activity of P. syringae T1 was only obtained when cell

cultures were at a concentration of 106 cells mL-1 or higher (103), with the highest activity

obtained after the cells had entered stationary phase (104). Similar to P. syringae, P. viridiflava

also requires cells to be in a concentration of at least 105 cells mL-1 in order for IN activity to be

detected at warmer temperatures (93, 105), and achieved its highest IN activity when cultured at lower temperatures, such as 18oC, which induced IN activity at -2.8oC (93). However, unlike P.

syringae, P. viridiflava appeared to produce INPs following aerobic culturing in both TSA and

LB broth, without the need for starvation for certain nutrients (93). P. fluorescens, on the other

hand, was shown to produce IN proteins regardless of cell concentration, and at all stages of its

growth cycle (104). In Pseudomonas borealis, which was more recently reported to contain a

shorter IN gene, IN activity was successfully induced after cold conditioning at 4 to 8oC (94,

106). The activity of INPs produced by E. uredovora was shown to be significantly affected by

the pH and media composition that it was being grown in (107, 108), with highest activities

when grown at cold temperatures (100). However, the effects of nutrient deprivation and

induction conditions for IN activity in P. fluorescens and P. borealis have not been investigated.

Similar to P. syringae, IN activity in E. herbicola was induced at cold temperatures and when

starved for phosphate (109). However, E. herbicola is the only species demonstrated to increase

IN activity after exposure to DNA-damaging agents (110). For example, exposure of E.

herbicola cultures to UV, mitomycin-C, nalidixic acid, and thymine deprivation, followed by a

2-hour incubation, led to a significant increase in IN activity (103). This induced activity was inhibited by incubation with transcription and translation-blocking drugs, indicating that increased production of the IN protein in E. herbicola is necessary for higher IN activity, as also reported in P. syringae (103).

It was originally thought that the increase in IN activity in these Gammaproteobacterial species was due to multimerization of the proteins on the surface of the cell. Whether this phenomenon is a direct result of an increase in transcription of the protein, or a reorganization of proteins that have already been produced remains unclear (20, 87). Certain studies suggest that

post-translational modifications to the INP are a reason for changes in IN-activity (111–113).

Turner et al (111, 113) demonstrated that the IN proteins produced by P. syringae had three

“classes” of proteins, which were characterized by their temperature of activation and distinct

chemical and pH-sensitivities. This study hypothesized that these differences occurred as a result

of various post-translational modifications, thus changing their hydrophobic or hydrophilic

characteristics, and ultimately affecting their ability to self-aggregate and interact with water

molecules. The potential posttranslational addition of a lipid or carbohydrate, they argued, were

at the root of why the efficacy and temperature activity of the INPs produced by those species

changed with varying conditions encountered by the bacterium.

While a number of publications have claimed IN activity in a range of bacteria, only one

study has successfully demonstrated IN activity in bacterial species not belonging to the

Gammaproteobacteria. Lysinibacillus, a gram-positive genus of the Firmicutes phylum, has been shown to possess at least two species of IN-active bacteria (114). However, no inaZ-like gene or

InaZ-like protein has been reported in this species. In fact, the nanometer-sized INPs produced

and secreted by Lysinibacillus do not appear to be proteinaceous at all and are resistant to both

heat and proteinase (114). The ability of a non-pathogenic bacterial species to possess IN-activity

is significant, given that the Ice+ phenotype in the Gammaproteobacteria is speculated to have evolved as a way to increase pathogenicity on plant surfaces. Indeed, a number of studies have reported that frost injury to plants is a predisposing factor for subsequent infection by P. syringae

(115–120), and studies on E. herbicola showed that corn seedlings which were sprayed with E.

herbicola suspensions and cooled to -4oC and -5oC were substantially more damaged than their

controls which had no E. herbicola (121) or X. campestris (97) on sprayed on them. These

observations would make sense, as many pathogenic epiphytes cannot easily invade uninjured

plant tissues (122). Such a disruption of plant tissues would allow for easier uptake of nutrients

from the plant by the bacterial pathogens. However, IN activity has also been described in non- pathogenic species of fungi, lichen, pollen, and algae.

Ice Nucleation in Fungi, Lichens, Pollen, and Algae

Organisms outside of the bacterial kingdom, including certain fungi, lichens, plants, and algae, have also been shown to possess IN activity. Fusarium acuminatum, Fusarium

avenaceum, and Fusarium oxysporum, all fungal pathogens of various types of plants, as well as

non-pathogenic fungi such as Mortierella alpina, Isaria farinosa, and Acremonium implicatum,

have been shown to produce INPs active at temperatures as warm as -2.5oC (23, 123–125). The

INPs produced by species of Fusarium appeared to be proteinaceous, given their sensitivity to

proteinase-K, secreted from the cell, much smaller (~50 kDa) and more heat and pH-stable than

those related to the InaZ protein family (123, 126–128). Additionally, a variety of epilithic lichen

including Rhizoplaca chrysolecua, Caldonia species, Lecanora dispersa, Pertursaria flavicans,

Xanthoria species, and Xanthoparmelia species, showed IN activity at a range of temperatures,

with some displaying activity as warm as -2.3oC (129, 130). The INPs produced by these lichen

were concluded to not be bacterial in origin and not sensitive to heat (70oC) or sonication (129).

Additionally, rust fungi, a group of fungal pathogens of a wide variety of plants, have been

shown to possess IN active urediospores that may be of a polysaccharidic nature (131). However,

despite the discovery of fungal INPs over 30 years ago, very little is known about their genetic

basis. Further, exposure to low temperatures and nutrient deprivation was not shown to enhance

IN activity in Fusarium acuminatum, as it did in the Ice+ bacteria (127). Other studies have

indicated that the INPs produced by fungi may be of a nanoscale, soluble nature (132, 133), and

may make up a significant fraction of the IN activity associated with many minerals and soils

(24, 133).

IN activity has also been demonstrated in algae (134, 135) and several plant species (22,

136–139). Although not capable of inducing freezing at temperatures as warm as Ice+ fungi or

bacteria, the spores of Polytrichum commune, a species of haircap moss, have demonstrated heat-

sensitive IN activity at -10oC and colder (139). Birch and conifer pollens, including those of the

Silver birch, Scotts Pine, and the Common Juniper, have been shown to produce non- proteinaceous, soluble INPs, capable of nucleating ice at -18oC and colder (22, 136). INPs

produced by winter rye showed IN activity at temperatures as warm as -7oC, and were thought to

be composed of a mixture of proteins, carbohydrates, and phospholipids (137). The potential for

plant species to produce INP propagules may be advantageous when considering the potential for

such INPs to induce precipitation formation. However, very few articles on the nature of plant

INPs and their propagules exist, and as such, not much is known about the mechanisms leading

to their production. But, while the mechanism of ice nucleation and chemical makeup of

biological INPs may differ across species, it is clear that biological INPs are much more effective

than their abiotic, mineral dust counterparts.

Applications of Biological Ice Nucleation

The ability to initiate and control freezing with the IN protein makes Ice+ microorganisms attractive to bioprospecting. Indeed, by the end of the 20th century, researchers had come up with

a range of ideas for biological INPs and their potential applications in other industries. Given the

propensity for the IN protein to cause frost damage and disease to a number of important

agricultural crops (121), bioengineering efforts were made to manipulate the activity of naturally

occurring epiphytic populations Ice+ bacteria. Indeed, several studies were published that

demonstrated the attempt to alter epiphytic communities through the introduction of naturally

antagonistic bacteria to the Ice+ bacteria (140–146), through the introduction of genetically

engineered Δina (Ice-) mutant strains (105, 147, 148), IN-inhibiting chemicals (149), or

bacteriophages specific to Ice+ bacteria (150). Additionally, use of the Ice+ phenotype was

attempted in molecular applications, such as gene reporter systems, as the IN assay is

inexpensive to perform and very sensitive (151, 152). Ice+ bacteria have also been investigated

for their use in the food industry during the initial process of freezing for various types of

foodstuffs (148, 153). Use of biological INPs to induce freezing at warmer temperatures was

shown to improve the texture of frozen foods, such as raw egg white, bovine blood, soybean

curd, soybean protein isolate, agar hydrogel, cornstarch paste, hydrogels, and meat (76, 153),

while also lowering production costs because of the energy saved on warmer-subzero freezing

(76, 154). Bacterial INPs have also been the subject of research for use in the artificial snow making industry (84). The use of high-temperature INPs for the production of snow allows for the generation of snow at much warmer temperatures, which significantly decreases the costs of production. Today, artificial snow making is still practiced by many ski-resorts with the help of the product Snomax, which is a patented mixture composed of desiccated and sterilized P. syringae cell fragments and proteins (155).

Ice+ bacteria have also been considered for their use in weather modification schemes,

which came about after the realization that warm-temperature INPs found in the atmosphere and

precipitation were biological in origin (156, 157). Because other ice nucleation active agents,

such as silver iodide and dry ice, had been shown to be somewhat successful in cloud seeding

(158, 159), researchers proposed the use of naturally occurring bacterial INPs in weather

modification schemes. However, there were many uncertainties associated with how downwind

environments would respond to the introduction of potentially plant-pathogenic bacteria. Further, the public perception surrounding the release of bacteria, whether alive or dead, was likely negative. Thus, the idea of cloud seeding with biological INPs was abandoned.

Aeromicrobiology

Airborne Microorganisms and Disease Transmission

The study of airborne, IN-active microorganisms may not have come to be had it not

been for the work accomplished on airborne microorganisms in the century beforehand. In the

early 20th century, the field of aerobiology emerged as a distinct scientific discipline, coined by

Fred C. Meier as the study of life in the air. However, the development of this field of science can be credited to the pioneering work of the earliest “aerobiologists”, such as Louis Pasteur,

who famously disproved the theory of spontaneous generation in his swan-neck flask

experiments, and John Tyndall, who considered the movement of organic material through the

air as a “cloud” or “system” (160, 161). Since the 19th century, many advancements have been made in this area of science, leading to increases in our understanding of how airborne microorganisms could be used in industrial and agricultural applications, space exploration, biological warfare, and more importantly, the implications that they have on public health and ecosystem dynamics. Indeed, the study of airborne disease-causing organisms has received much attention over the last century, as it has implications not only for human and animal health, but for agricultural crops, plants, and entire ecosystems as well.

However, the potential spread of disease is limited by the size of the disease-causing agent. Indeed, the small sizes of viruses (20-400 nm in diameter on average), make them prime candidates for dissemination over long distances, as their small sizes and weight make them less likely to be removed from the atmosphere by dry deposition (162). Indeed, airborne viruses that become suspended within the free troposphere may remain airborne for days to weeks (162).

However, atmospheric residence times of viral particles are not well constrained, and in fact, bacteria are thought to remain airborne longer than viral particles, despite being larger (163). In fact, the average aerodynamic size of planktonic bacteria, which is assumed to be ~1.0 μm in diameter, is within a size range that is thought to escape the normal physics of wet and dry deposition of all other particles—a gap which is referred to as the “Greenfield Gap” (164).

Further, many microorganisms, bacteria in particular, have developed tolerance and/or resistance to UV radiation, desiccation, cold temperatures, and low nutrients. As such, their residence time in the atmosphere can range from several days to weeks, allowing them to travel thousands of kilometers (165–171).

Such a phenomenon may explain the ease with which bacterial plant pathogens appear to

disseminate to new environments. Indeed, the long-distance dispersal of plant pathogens is a

significant factor in the spread of disease, and one that is very difficult to control. Pseudomonas

syringae, for example, has over 50 pathovars able to infect a wide variety of plant species, and

can cause not only frost damage, but bacterial canker, bacterial blight, and generalized plant

tissue necrosis (172), and is considered the most important bacterial plant pathogen in terms of

scientific and economic importance (173). P. syringae has been detected in air (157),

precipitation (157, 174, 175), and cloud water (176, 177), exemplifying its capacity to survive

aerosolization, atmospheric transport, and subsequent dissemination via the Earth’s water cycle

(175). Long-distance dispersal is also a well-documented mode of transmission for pathogenic fungi that use wind to disperse their spores to distances hundreds or thousands of kilometers from their source (178). Rust fungi, which consist of approximately 7,000 different fungal species, can produce infective spores which are also easily disseminated by aerial transport through the atmosphere (179, 180). Outbreaks of plant diseases such as the Irish Potato Famine, which was caused by the fungal pathogen Phytophthora infestans (181, 182), or coffee leaf rust

(183) and sugar cane rust (184), demonstrate the destructive effects of airborne plant pathogens,

and the inability to constrain their aerial transport (178).

The long-distance transport of microbial pathogens also has implications for the spread of

animal and human disease. Outbreaks of foot and mouth disease, a highly contagious viral

infection in livestock, has repeatedly demonstrated long distance spread over thousands of

kilometers and subsequent infection of animals at distant sources (185–190). Indeed, modeling

and epidemiological studies have indicated that the transmission of this virus by wind under

favorable meteorological conditions can lead to a rapid and extensive dissemination, making it

nearly impossible to quarantine and prevent (191). In fact, several other viral diseases in animals have more recently been implicated in long range aerial dispersal, including the lumpy skin disease virus (192), the Bluetongue virus (193), and the porcine reproductive and respiratory syndrome virus (194).

Long-distance dissemination of viruses may be an overlooked piece in most common epidemiological models which examine the spread of human respiratory viruses, such as the influenza virus (195). The potentially long residence times of small viral particles in the free troposphere would allow for the travel of viral particles over thousands of kilometers (195), and as such has become a more recent concern for the containment of human viruses (196). Indeed, outbreaks of measles have been linked to the occurrence of dust storms in western China (197),

Niger (198), and the United States during the dust bowl of 1935 (199).

Seasonal outbreaks of meningococcal meningitis in sub-Saharan Africa, which kills thousands of individuals every year, repeatedly coincides with the start of the dry season and increases in dust storm activity (200–203). Scientists surmise that the ability of the disease- causing bacterium, Neisseria meinigitidis, to travel over long distances is made possible by the bacterium’s ability to attach to dust particles (171, 200–203). The arid regions of the meningitis belt see their largest increases in meningococcal meningitis during the dry seasons, when topsoil becomes loose and windblown, and dust storms known as the Harmattan winds become incessant

(201). The marked increase in disease during the dry seasons end abruptly when the first rain arrives, further indicating that increase in dust loadings into the atmosphere is tied with spread of the disease. Topsoil from desert regions are the single largest sources of aerosols to the atmosphere, and the major avenue by which microorganisms enter the atmosphere as well (165).

Indeed, it is estimated that up to 5 billion tons of desert dust enters the atmosphere every year,

the majority of which comes from the Sahara and Sahel desert of North Africa and the Gobi and

Takla Makan regions of Asia (204), a number which has increased significantly over the last

twenty years due to climate change and desertification (7, 205–207). Additionally, estimations

indicate that dust clouds produced from arid environments can attenuate UV rays by almost 50%

(208). With the protection from harsh environmental conditions, the likelihood of eventual

inhalation by a human host downwind is increased. Given that inhalation of dust particles can lead to irritation of the human airways, and make them more susceptible to infection, the likelihood of infection from a disease-causing bacterium during dusty conditions is of concern

(209–211).

In fact, the ability of microbial pathogens to attach to dust particles has raised concerns

over increases in disease in other parts of the world. For example, increases in pediatric asthma

hospital admissions in the Caribbean is thought to be a result of interhemispheric movement of

bacteria and fungi with dust from Africa (207). Likewise, enterotoxin-producing pathogens,

including Bacillus and Staphylococcus species, have been detected as airborne agents in Kosas,

the well-documented Asian dust storms that transport dust from Mongolia to China and North

America (212). The transmission of Kawasaki Disease, of which the causative agent is believed

to be bacterial or viral, has also been linked to synoptic scale wind patterns and the movement of

aerosols from central Asia to the north Pacific (213).

Emissions of Airborne Microorganisms

The densities of airborne microorganisms vary spatially and temporally, making general

assumptions about their occurrences in the atmosphere difficult. For example, the number of

bacteria estimated during the day can differ significantly from that at night, due to changes in

boundary layer turbulence (214). Similarly, upward fluxes of bacteria above plant canopies have

been shown to increase during rain (215). Further, concentrations in airborne microbial

abundance have a strong seasonal dependence, as demonstrated by a number of studies (215–

219). While observations on the concentrations of culturable airborne microorganisms exist (157,

215, 217, 218, 220, 221), relatively few studies that enumerate their abundance through culture-

independent techniques have been performed, making estimations of microbial fluxes to the

atmosphere a challenge (222).

Nevertheless, sources environments or airborne microorganisms have been studied

numerous times. While microorganisms can be aerosolized from virtually any environment, arid

dusts and soils have been implicated as the leading sources of microbial biomass to the

atmosphere (170). Indeed, with an estimated 5 billion tons of arid topsoil emitted to the

atmosphere every year, and up to 1010 microorganisms estimated in a single gram, the potential

number of microorganisms emitted to the atmosphere every year from arid soils alone could be

on the order of 4.5x1022. While culturable fractions of microorganisms underestimate the true

numbers of microorganisms in the environment, studies done on Saharan dust storms did indicate

that airborne CFUs during the storm increased from 1,100 CFUs m-3 of air at ambient conditions

to over 15,700 CFUs m-3 of air during the storm (223). Further, studies have shown that while

wind-borne bacteria typically travel less than a single kilometer from their source, those

associated with dust have been shown to travel over 5000 km, from Africa all the way to the

Caribbean (207, 224), the Amazon (225), the Mediterranean (169, 171, 226), the Pyrenees

Mountains (227) and the Alps Mountains (228) of Europe. The ability of bacteria to travel longer distances with dust is likely due to the protection that dust particles offer to bacteria, including attenuation of UV rays, availability of minerals and nutrients, and protection from desiccation.

Similarly, microbially laden dust from Asia has been shown to travel across the Pacific Ocean to

the Arctic and the West Coast of North America as well (168, 229). But dust is not the only source of microorganisms to the atmosphere.

Given that there are an estimated 1026 prokaryotic cells residing on leaf surfaces worldwide (Morris and Kinkel 2002), agricultural crops and plant canopies are thought to be significant sources of bacteria and fungi to the atmosphere as well. Airborne bacteria have been detected over agricultural crops (215, 230, 231); rural fields and grasslands (218, 221, 230–234), and forest ecosystems (220, 235). However, despite their detection over plant canopies, increased concentrations of microorganisms over plants and their direct sources are difficult to constrain. Nevertheless, rough comparisons of culturable to unculturable fractions of microorganisms estimate that approximately of 104 to 106 m-3 microorganisms are present in the air over plant canopies, croplands, or grasslands at any given time (236). Marine and freshwater environments, through the action of wave breaking and sea spray, can also source microorganisms to the atmosphere (220, 237, 238). Some estimations of total loads of prokaryotic cells over tropical and subtropical oceans are on the order of 1021 cells, with approximately 25% of them originating from marine environments, and 42% originating from terrestrial environments (239).

The majority of observations on bioaerosol concentrations occur near the surface, in the troposphere’s lowest layer, which is referred to as the mixed boundary layer (MBL). While point observations taken within the MBL may indicate high concentrations of bioaerosols at any given moment, the actual amount of microorganisms that remain airborne within the MBL for longer than an hour are thought to be negligible, as turbulent mixing within the MBL leads to an almost equal upward and downward flux of microorganisms (240). Indeed, the vertical distribution of aerosols is thought to be highly concentrated in the lowest 10-100 meters of the troposphere,

with movement out of the MBL and into the free troposphere occurring mainly through moist

convection and frontal systems (240). However, while larger sized bioaerosols such as fungi may

not commonly make it out of the MBL, the small sizes of bacteria may make them prime

candidates for ejection into the free troposphere (241).

Hight Altitude and Cloud Microbiology

In a high-altitude study conducted over the southwestern United States, cell abundances

averaged ~106 cells m-3 in the MBL and ~105 cells m-3 in the free troposphere, with cell

concentrations remaining relatively similar between much of the free troposphere and the lower stratosphere (242). Several other papers have reported viable microorganisms in the stratosphere, however the possibility of contamination from sources other that high altitudes cannot be ruled out (243–246). Further, the recovery of viable microorganisms from several high-altitude samples indicates that the vertical extent of microbial life may be well into the stratosphere, however survival may be limited to only those microbial aerosols in lower portions of the stratosphere that receive sufficient protection from UVC (242, 246–249). Indeed, while these studies imply that bacterial survival in the stratosphere may be possible, the harsh conditions

encountered in the stratosphere, (e.g. DNA-damaging UVC, low temperatures, and low

humidity) make it unlikely to be an environment conducive to active microbial metabolism. As

such, the free troposphere and stratosphere have yet to be definitively considered an environment

where microorganisms can survive and metabolize.

Clouds, on the other hand, may very well be a microbial habitat (176, 177, 250–253).

Studies have shown that bacteria collected from cloud droplets can be present in concentrations

of up to 104 cells m-3 of cloud water (177), can actively divide at temperatures below 0oC (250),

and that they typically contain characteristics that make them well adapted to the harsh

conditions of the atmosphere (177). In fact, studies have shown that bacteria might actually metabolize within cloud water, and actively participate in biochemical transformations of cloud chemistry (252–256). The strains of bacteria that have been isolated from cloud water have been shown to metabolize the most abundant organic compounds typically found in cloud water, including formate, acetate, lactate, and succinate (253). Such microbial processing has many implications for cloud dynamics, given that the salts of the aforementioned organic acids can act as effective CCN in the troposphere, thereby aiding in cloud droplet formation (257).

Microorganisms have also demonstrated an ability to metabolize other common compounds found in cloud water, including methanol and formaldehyde, which have important roles in cloud chemical reactions (177, 254). Studies suggest that the up to 37% of the oxidation of carbonaceous compounds in a cloud can be attributed to bacteria (253).

Microorganisms that have made their way from the MBL and into the free troposphere have the potential to be dispersed globally through synoptic scale meteorological formations, such as hurricanes (258). In a study conducted on two different hurricanes that crossed the

Atlantic Ocean, viable bacterial cells were found to account for 20% of the total particles samples within the 0.25-1μm size range (258). The small size of bacterial cells may allow them to be more easily lofted from the boundary layer into the free troposphere. The consistent detection of viable microorganisms in clouds, coupled with their active in-situ metabolic activity, indicates that airborne microorganisms may be more than just passive passengers caught in the winds of the troposphere. As such, the hypothesis of bioprecipitation, which was set forth by

David Sands in the 1980s and states that airborne microorganisms are critical components to precipitation formation (156), may be proved to be true.

Potential Influence of Biological Ice Nuclei on Meteorology

Biological INPs in Air, Cloud, and Precipitation Samples

Biological INPs are a general term used to define an INP of biological or biogenic origin.

However, “biological” as a designation of a type of INP can vary widely from study to study.

Most often, biological INPs refer to those INPs which are sensitive to gentle heat treatment (60 to 100oC), which can lead to protein denaturation. Other agents that may be used to identify the presence of biological INPs include enzymes such as lysozyme which targets the cell wall of bacteria and causes them to lyse; or proteinases that cleave proteins; or antimicrobial chemicals such as urea or antibiotics. However, other methods that to not employ chemical or physical disruption to the INP are also used, including the detection of fluorescent amino acids or organic matter specific to biological organisms. Further, biological INPs may refer to the entire cell which produces them, or it could refer to the biogenic propagule, such as a secreted protein or a pollen grain (15). Biological INPs can range in size from several nanometers (20, 22, 98, 123) to several micrometers (15, 81, 82, 95), and they are ubiquitous in the environment. They are found in soils (23, 24, 106, 133, 134, 259), fresh water (175, 260–262), marine water (135, 263–265), on plant surfaces (95, 121, 122, 175, 221, 266, 267), in decomposing vegetation (78, 80, 82), in the air (125, 225, 268–271), in clouds (176, 229, 272), and in precipitation (114, 174, 229, 241,

266, 273–279). Their ubiquitous presence in the atmosphere, clouds, and precipitation, combined with their high temperature--anywhere from -1.9oC for P. syringae (81) to -10oC for moss spores

(139)--IN activity, suggests that they may play a role in weather and climate (156, 276, 280).

Indeed, during the 1970s, David Sands, a plant pathologist at Montana State University, was tasked with studying disease outbreak on wheat fields in Montana found P. syringae at cloud altitudes and surmised that these plant pathogens may also play a role in weather. This led to the

formal proposal of “bioprecipitation”, whereby microorganisms create precipitation in the clouds by inducing ice nucleation (156). Subsequent studies identified P. syringae in rain and air samples over soybean and bean crops (157, 215). However, this idea of bioprecipitation did not catch on until the early 2000s (276). Further, the prolific research period on biological ice nucleation during the late twentieth century appeared to end rather abruptly in the nineties, with interest on the subject matter lying dormant for almost ten years (74). However, a more recent resurgence in interest on biological INPs and their effects on meteorological processes may have abounded after a study published in 2007, which reviewed the potential role that biological particles could play in atmospheric processes (281), and another study published a year later, which indicated that the most active INPs in fresh precipitation samples were biological (275).

Surprisingly, despite such a prolific 30 years of research on biological INPs in the 20th century, this was the first study to describe biological INPs in precipitation samples. Previous measurements of INPs in precipitation had been done on hail and rain (Vali 1970), however this study did not examine the nature of the INPs. Indeed, Christner et al 2008 were the first to examine precipitation samples for biological INPs, and ultimately showed that at temperatures ≥-

7oC, up to 100% of the INPs were inferred to be biological, and up to 85% of the INPs were inferred to be bacterial (273, 275).

Over the last decade, major improvements in measuring systems for analyzing INPs in environmental samples has also allowed for a renewed interest in studying the connections between biological INPs and meteorological processes. Since 2007, multiple studies have established the presence of highly active biological INPs in precipitation samples (174, 241, 274,

277–279, 282). Experimental designs that coupled biogeochemical data to microbiological data allowed researchers to draw inferences on the occurrence of biological INPs and their connection

to precipitation formation. For example, Michaud et al (282) found biological INPs in the centers

of hailstones. Combined with stable isotope analysis, they were able to infer that the hailstones

formed at the same warm temperatures (around -5oC) at which the biological INPs were

demonstrated to initiate freezing, suggesting the biological INPs initiated ice and eventual

hailstone formation. Similarly, Stopelli et al (278) used stable oxygen isotopes to show that the

first raindrops which fell from a cloud contained biological INPs, indicating that they could be

significant in initiating the original ice particles that form within a cloud.

Biological INPs have also been reported in near surface and boundary-layer air (125, 270,

271, 283). With the development of portable in-situ INP and bioaerosol sampling instruments,

such as the continuous flow diffusion chamber (284) and the wideband integrated bioaerosol

sensor (285), measurements of biological INPs in air samples became more attainable. The

combination of these types of tools allowed researchers to study the flux of bioaerosols and

warm-temperature INPs, and show that the two concentrations typically correlated well with one

another (225, 271). Likewise, the development of modern-day nucleic acid sequencing

techniques also permitted a novel way to detect Ice+ microorganisms. Several studies have used

PCR primers for identification of ina-gene containing bacteria (266, 270), and determined that the number of Ice+ microorganisms increased in air samples from undetectable at background

concentrations to up to 19 Ice+ bacteria L-1 air during crop harvesting season in the western

united states (270). They also found that there was a significant contribution of ice nucleation

from biological INPs that did not contain the ina gene, further demonstrating that there are IN

active microorganisms that induce freezing using mechanisms not related to that of the Ice+

Gammaproteobacteria.

The use of portable, in-situ sampling equipment allowed researchers to unearth

interesting temporal trends in the occurrence of airborne biological INPs. Indeed, several studies

indicated a significant increase in biological INPs during and after rain (125, 283). Prenni et al

(283), showed that ground-level INP concentrations were enhanced by a factor of 40 during rain,

and that a fraction of those were biological. Similarly, Huffman et al (125) showed that the

concentration of airborne biological particles in the size range of 2-6 μm increased significantly

during a rain event, and that these bioparticles correlated significantly with INP concentrations.

The identity of these biological INPs remain unclear, however several studies have suggested

that increases in relative humidity during a precipitation event might lead to bursting of pollens

or fungal spores, both of which have been implicated as sources of biological INPs (22, 132, 268,

269).

Although limited in their number, several studies have confirmed the presence of

biological INPs in clouds as well (176, 272, 286, 287). A study conducted in the free troposphere

over Wyoming confirmed the presence of biological INPs in ice crystal residues taken directly

from clouds (272). Likewise, another study confirmed the presence of Ice+ bacteria isolated from

cloud water, including Ice+ Pseudomonas and Xanthomonas species. Their estimations indicated

that the Ice+ bacteria were active at -3oC and present at concentrations of 1 bacterium per mL-1 of

cloud water, and at concentrations of ~500 bacteria mL-1 at -10oC (176). It was further

demonstrated that biological INPs were present at concentrations of up to 220 biological INPs

mL-1 active at -10oC (286, 287). Whether these biological INPs are bacterial or fungal in origin is unknown, however more recent studies have indicated that nanoscale fungal INPs may be an important component for precipitation formation (288, 289).

In fact, there is now evidence that mineral dust INPs, such as montmorillonite clays, may

get their IN activity from microbial and organic particles (24, 133, 259). In some soils, up to

94% of the IN activity of organic matter may be conferred by the INPs produced by M. alpina

alone (133). Indeed, heat treatment of fertile soil dusts indicate that the majority of IN activity

measured at temperatures of -6oC and colder were heat-labile, indicating that they are likely of biological origin (24), and that the organic matter of agricultural soils conferred the majority of

IN activity as well (271). Further, soluble, proteinaceous INPs, such as those produced by fungi, have been shown to preferentially attach to kaolinite, which is assumed to be one of the most dominant abiotic INPs in the atmosphere. In light of the fact that mineral dusts, such as kaolinite, are thought to be one of the most important INPs for ice formation in the atmosphere, such a finding that their activity might be conferred by microorganisms is significant. (290). Indeed, many modern climate models use montmorillonite clays to represent mineral dusts, and are assumed to have IN activity at temperatures as warm as -12oC (69, 158).

However, the detection of high concentrations of biological INPs in terrestrial

environments does not necessarily equate to an equal abundance in the atmosphere.

Aerosolization of biological INPs from surface environments, and their subsequent lifting to

altitudes above the MBL are both processes that that need to occur in order for biological INPs to

affect cloud ice formation but are poorly characterized. Further, there are a number of challenges associated with studying airborne biological INPs in clouds. As such, the scientific community has yet to come to an agreement on whether or not biological INPs are even present in abundances sufficient to affect precipitation formation and cloud processes on a larger scale

(291).

Challenges and Limitations to Studying Biological INPs

Even though biological INPs have been detected in air, clouds, and precipitation,

relatively few studies have attempted to quantify their concentrations in the free troposphere.

Detecting and quantifying biological INPs in the environment is hindered by the simple fact that

the chemical makeup and sizes of biological INPs are very diverse, and thus the full extent of

their existence in the environment remains hard to define (292). It is likely that their presence in

the atmosphere is dependent upon season and land type, yet the extent of their temporal and

spatial variability remains mostly uncharacterized (241, 273, 274). Studies have shown explosive

increases in the concentrations of bioaerosols and biological INPs based on seasonal and diurnal

changes (219), in addition to meteorological events (125) or agricultural practices (270), which

may be missed in solitary, point observations. Aerosolization of bioaerosols may occur during

dust storms (165, 165, 167, 168, 229), rain events (125, 269, 283), or from marine surfaces

through wave breaking mechanisms (264, 293). However, the mechanisms through which

biological INPs are aerosolized and subsequently lofted to cloud heights is still an area of active

research.

The inability to study biological INP aerosolization and flux in the MBL and free

troposphere is due, in part, to their overall low abundance in the atmosphere (292). Indeed,

detection of the biological INPs is hindered by the fact that they occur below the limit of

detection for most sampling equipment. Further, since most (99%) environmental

microorganisms are unculturable, culture-independent techniques are sorely needed for quantifying biological INP concentrations. However, low biomass in aerosol samples limits the types of culture-independent techniques that can be carried out. Studies which have sought to enumerate biological INPs through the use of culture-independent techniques, such as PCR for

the detection of the ina gene (266), have significant biases associated with them. The last several decades of research has demonstrated that Ice+ genotype is limited to a small group of

Gammaproteobacteria. Thus, other mechanisms of inducing ice formation by microorganisms are likely overlooked by ina-gene amplification studies (266), making the use of a universal primer for genetic identification of IN-active microorganisms an ineffective way to determine their presence in the atmosphere.

Most studies which examine biological INPs do so with downstream laboratory analyses following sample collections in the field, which may lead to a change of the biological INP in- situ state. For example, the storage of microbial samples, even at 4oC, may lead to increases or decreases in biomass (292). Concentrating field samples (e.g. cloud water, rainwater, or air samples), which is often necessary due to low biomass, may lead to a disruption of the original state of the sample as well. Further, the proper choice of bioaerosol sampling equipment is important for obtaining bioaerosols that are still viable following inertial impaction. In general, high-flow rate impactors, which are often necessary for collecting sufficient biomass, can lead to lysing upon impaction or desiccation during sampling.

Sampling of biological INPs from higher altitudes (such as those that would occur at cloud height), is also a very difficult endeavor to undertake. Dedicated research flights are expensive and must be planned months in advance, without any foreknowledge of whether the days for sampling will contain clouds. Further, flights through mixed-phase clouds which contain updrafts and downdrafts are dangerous, as they can lead to aircraft icing and/or aircraft failure from an encounter with strong vertical winds. As such, the ability to fly through the cloud which is to be studied is often times not even possible from a safety standpoint. The study of biological

INPs in clouds that occur over mountain top observatories, such as the Puy de Dome in France,

or Storm Peak Laboratory in Colorado, USA, make sampling of clouds more feasible. However,

these types of clouds, termed orographic clouds, represent a very small subset of all clouds that

occur globally. Additionally, they have more extensive interactions with the surface, making

inferences about their relationship to clouds that occur in the free troposphere difficult (71, 74).

Another mechanism for studying free tropospheric air samples is the use of weather balloons,

upon which payloads may be attached (242). However, this method must utilize passive sampling, as pumps are in general too heavy, and do not work at higher altitudes as they encounter lower pressures. Thus, while passive sampling is a method for high-altitude bioaerosol, it is also limited in the amount of biomass that it can collect.

Understanding the role that biological INPs play in cloud processes is further limited by the fact that atmospheric scientists do not yet have a full grasp on all of the different mechanisms that may lead to ice formation within a cloud (74). Theoretical considerations dictate that there are a number of different pathways which can lead to ice formation, including primary ice formation and secondary ice enhancement processes (Figure 1-1). However, without a firm understanding of ice formation in the cloud, deciphering the ways in which biological INPs can affect precipitation formation remains somewhat elusive (74). Nevertheless, the influence of biological INPs in ice formation in clouds is still an active question among the research community. As such, other methods have been used to study their meteorological effects. Indeed, with an increase in computational power over the last several decades, it has become possible to simulate their involvement in microphysical processes in a cloud through the use of numerical cloud resolving and climate model simulations.

Use of Biological INPs in Numerical Cloud and Climate Models

Mixed phase stratiform clouds occur globally and are significant contributors to the

earth’s radiative budget (294) and production of precipitation (9). However, the microphysical processes that govern the evolution of mixed‐phase stratiform clouds and their production of precipitation are not well understood (31, 295–297), which has led to uncertainties in numerical weather models and climate projections (296, 298, 299). There are a few studies that have examined in situ ice numbers in warm stratiform clouds and reported a similar paradox in high ice numbers as compared to INP concentrations (35, 72, 73, 300). Importantly, remote‐sensing has also shown that mixed‐phase stratiform clouds with temperatures > ‐8oC are a global

phenomenon (295).

Several studies have shown that mixed‐phase clouds can have ice concentrations which

far exceed the number of INPs present (72, 73, 301). For convective clouds, it is theorized that

this discrepancy is due to primary ice formation by INPs and subsequent secondary ice formation

processes, such as the Hallet‐ Mossop (HM) process (302), which occurs at temperatures of ‐3 to ‐8oC and in clouds with substantial vertical wind velocities. The relative activities of the

various mechanisms responsible for primary and secondary ice formation differ substantially

between nimbostratus and convective clouds (31, 302–304). Stratiform clouds at subzero

temperatures have weak vertical wind velocities that can readily cause cloud droplets to

evaporate, reducing the chance of coalescence and making the HM process less likely. Thus, low

altitude, mixed‐phase nimbostratus clouds possess conditions highly favorable for interacting

with biological INPs because cold‐phase microphysical processes are the dominant pathway of

precipitation production. However, the long lifetimes of stratiform clouds can allow breakup in

ice‐ice collisions to prevail for overall ice concentrations, as observed by Schwarzenboeck et al.

(305) in the Arctic. Given that a large fraction of the earth’s clouds are stratiform (306–308), and

that they are significant producers of global precipitation, and are thought to be major

modulators in global climate through their UV-attenuation capabilities (33, 63), the effect of

biological INPs on stratiform clouds is an area of research that requires further scrutinization.

Several studies have attempted to estimate global biological INP concentrations and

determine their effect on ice formation using numerical models. However, the results of such

modeling efforts thus far have varied. One such study indicated that on a global level, biological

INPs are trivial (291, 309). Using a classical nucleation theory, they found that the contribution

of biological INPs to freezing in the atmosphere at temperatures in between 0oC and -38oC was insignificant, and that the dominant pathway of heterogeneous freezing was by mineral dust INPs

(309). As mentioned earlier, however, scientists now know that a significant fraction of the IN

activity of mineral dusts and soils can be conferred by microbial IN activity (24, 259, 290).

Further, a limitation to early numerical cloud models which examined biological INP

contributions to ice formation was the use of outdated and under representative datasets to

estimate biological INP emissions. Indeed, Hoose et al (291) were restricted to estimates of

bioaerosol fluxes to the atmosphere from a handful of studies (236, 310), some of which only

measured the culturable fraction of bacteria to the atmosphere. This, in addition to the lack of

measurements which study the flux of biological INPs from the earth’s surface to the free

troposphere makes global estimations of ice made by biological INPs very rough. It is necessary

to consider point emissions of biological INPs to the troposphere, in addition to seasonal,

diurnal, and situational variations (e.g., increases in biological INP emissions during and after

rain, or during Midwest harvesting season).

On the other hand, other studies have indicated that the influence of biological INPs on

global ice formation may be important. Indeed, another global climate model concluded that the

presence IN-active fungal spores are enough to modulate the amount of ice and water found in

stratiform clouds over boreal and tundra regions (311). Several smaller-scale sensitivity analyses indicated that biological INPs might affect cloud ice formation under certain meteorological conditions, mainly at warmer subzero temperatures (225, 311–314). More recent studies have attempted to include more than just one species or type of biological INP in their models. One such study modeled the effect of bacterial INPs, as whole cells and as fractions of cells (315).

While this study did not include all types of biological INPs, the inclusion of bacterial INPs indicated that they may have an impact on the production of ice within a cloud, and may trigger precipitation formation, but that their concentrations are too low to enhance precipitation generation. However, this study, like many of the others that came before it, only examined the impact that biological INPs may have on convective clouds.

Additionally, very few of the cloud modeling studies mentioned above specifically alter

size distributions and sensitivity parameters for biological INPs (15, 17, 64). Indeed, the few cloud modeling studies mentioned that incorporate bacterial or biological INPs into their simulations do not take into account bacterial and biological surface properties, and generally assume that biological or bacterial INPs occur in amounts identical to those quantified in the laboratory. For example, one study attempted to parameterize bacterial INPs in clouds based on empirical evidence collected on several Ice+ strains (e.g., P. syringae and E. herbicola). The

numbers of biological INPs used in this study were taken from the results of early laboratory

studies on the IN activity of P. syringae, which indicated that on average, lab-grown P. syringae

cultures produce INPs at a frequency of 10-4 INPs per cell. This average was then applied to an

estimate of the number of bacteria that are present within a cubic meter of air—however these

bacterial concentrations were, again, estimated based on culturable fractions. While modeling

efforts such as these are moving in the correct direction for deciphering biological influence on

precipitation formation, they make very rough and potentially inaccurate estimations on Ice+

bacteria and overall bioaerosol concentrations. Further, studies such as these leave out the

potential contribution of IN activity from bacteria other than the Ice+ Gammaproteobacteria, as

well as IN activity conferred by fungi or other organisms. As such, studies that focus only on one

type of IN-active microorganism likely lead to an underestimation of the actual number of biological INPs that are in the cloud.

In general, the incorporation of aerosols that can act as INPs into weather and climate models is a rather recent phenomenon. Part of the reason that aerosol interactions in weather models is a relatively newer addition, is because inclusion of aerosols in numerical models

comes at much higher computational costs. Further, aerosol-cloud interactions is a very active

area of research, and in fact, has been labeled by the IPCC as the area of greatest uncertainty in

climate models. Several newer approaches to incorporating the IN activity of aerosols into

numerical models have become quite popular. One such approach approximates INP numbers

based on the assumption that mineral dust INPs and aerosols with aerodynamic diameters larger

than 0.5 mm are positively correlated (316). While this scheme may accurately predict mineral

dust INP concentrations, this size‐constraint may lead to an underestimation of biological INPs

that are smaller 0.5 mm, such as the extracellular vesicles of Erwinia herbicola that are INPs

(Phelps et al. 1986) and those commonly observed in precipitation. Furthermore, data acquired

for the DeMott et al. (317) parameterization was restricted to particle sizes < 1.6 μm due to

constraints imposed by the sampling method, which may lead to underestimations of biological

INP concentrations. Additionally, in situ biological INP measurements in surface air may not

accurately represent the number of biological INPs that are active while in cloud. Indeed, studies

have shown that bacteria can actively divide and metabolize in cloud water (252, 253, 318–320).

Ice nucleation active bacteria are also known to modulate INP production on the scale of hours to days based on environmental conditions. For example, INP production increases by several orders of magnitude when E. herbicola is subjected to UV radiation (110) and P. syringae is

subjected to cold temperatures and nutrient deprivation (103). Hence, it may be necessary to

consider such physiological changes to accurately portray biological INPs in numerical

simulations. However, inclusion of complex microbial processes within a numerical simulation

is no simple task, as it is computationally expensive. Indeed, microphysical parameterizations, let

alone detailed microphysical parameterizations that investigate chemical and size differences in

INPs, require heavy amounts of computing power, which makes the application of biological

INPs in simulations on a large scale nearly impossible (321, 322).

As such, the use of small-scale, single-cloud simulations, referred to as cloud resolving

models (CRMs), have become an integral part of aerosol-cloud interaction research. Indeed, as

computational availability continues to increase, so too does the use and accuracy of cloud

CRMs, which allow for the study of ice nucleation processes at very fine grid resolutions, on the

order of hundreds of meters (322). The use of CRMs in studying microphysical schemes allows

researchers to study a variety of small-scale atmospheric processes, ranging from those that

specify aerosol types (including hydrometeors, particles, size distributions, and mass content), to

those that vary the types of processes taken into account (such as secondary ice formation

mechanisms), to those that study the mathematical propagation of these processes within the

cloud (296). However, the number of studies which have utilized CRMs to investigate biological

INP activity in clouds are few. And those which have, have only investigated biological INP

effects in convective clouds. As such, the effect that biological INPs have on other types of

important precipitation-producing clouds, including nimbostratus and stratus clouds, is unknown.

Purpose of This Research

The distribution and concentration of biological INPs in the atmosphere, as well

as their source environments, and the diversity of microorganisms that can act as INPs, remains

poorly characterized. Furthermore, the current state of knowledge on whether biological INPs

can affect precipitation formation is equivocal. Thus, the purpose of the research carried out in

this study was to characterize the occurrence of biological INPs in precipitation and apply that

knowledge to understanding whether their concentrations and IN activity could play a role in the

generation of precipitation. To that end, the research presented here aims to bridge these gaps in

knowledge by studying the occurrence of biological INPs in precipitation, identifying their

potential source environments, examining their bacterial diversity, and lastly, their ability to

affect precipitation formation. As a result, this study presents evidence that high northern latitude and east Asian terrestrial source environments can produce high concentrations of biological

INPs that can affect precipitation formation thousands of kilometers away. Further,

investigations into arid soils as potential source environments of biological INPs led to the

discovery that the diversity of bacteria which possess IN activity might be much more expansive

than previously thought. Lastly, this study shows that the contribution of biological INPs to

precipitation formation in nimbostratus clouds is likely significant.

Figure 1-1. Simplified diagram depicting the microphysical processes that can occur within a mixed-phase cloud. Text and arrows in red indicate warm (above 0oC) temperature processes, while text and arrows in blue indicate cold (below 0oC) temperature processes. Dashed arrows indicate falling precipitation, whereas solid arrows indicate a process leading to the eventual development of precipitation. The bolded and starred boxes indicate heterogeneous, primary ice nucleation processes.

Table 1-1. Variations in the basic habits of ice crystals with temperature, adapted from Wallace and Hobbs, 2006.

Temperature (oC) Basic Habit Ice Crystal Shape

0 to -4 Plate-like Thin hexagonal plates

-4 to -10 Prism-like Needles (-4 to -6) Hollow columns (-5 to - 10)

-10 to -22 Plate-like Sector plates (-10 to -12) Dendrites (-12 to -16) Sector plates (-16 to -22)

-22 to -25 Prism-like Hollow columns

Figure 1-1. The various types of clouds that occur in the earth’s troposphere. The precipitation producers are the cumulonimbus and cumulus clouds, which are generally referred to as convective clouds. The other precipitation producers are the nimbostratus and stratus clouds, which are generally referred to as stratiform clouds. This diagram is a rough estimate of average altitudes and temperatures of the troposphere, based on the vertical profile of the troposphere in the subtropics.

Table 1-2. Documented IN-active bacteria with sequenced ina genes.

Organism INA gene Protein characteristics Localization

inaZ, inaV, Size: 1200 amino acids Outer Pseudomonas inaK, iceC 1o stucture: 122 octapeptide repeats membrane- syringae 3600 bp with 16- and 48-residue higher bound single ORF orders of periodicity Size: 1210 amino acids inaW Outer Pseudomonas 1o structure: 120 octapeptide 3630 pb membrane- fluorescens repeats, with 16- and 48-residue single ORF bound higher orders of periodicity Outer Pseudomonas N/A Size: 1248 amino acids membrane- viridiflava bound Size: 1244 residues Outer Pseudomonas inaPb 1o structure: 65 16-amino acid membrane- borealis repeats bound Size: 1258 amino acids Outer iceE Erwinia 1o structure: 124 octapeptide membrane- 3774 bp herbicolaa repeats, with 16- and 48-residue bound; shed in single ORF higher orders of periodicity vesicles Size: 1322 amino acids inaA Erwinia ananasb 1o structure: 70 16-amino acid N/A 4294 bp repeats Size: 1034 amino acids Erwinia inaU 1o structure: 52 16-amino acid Shed in vesicles uredovorac 3102 bp repeats Size: 1567 amino acids inaX X. campestris 1o structure: 154 octapeptide N/A 4701 bp repeats a Erwinia herbicola has since been reclassified as Pantoea agglomerans or Enterobacter agglomerans b Erwinia ananas has since been reclassified as Pantoea anantis c Erwinia uredovora has since been reclassified as Pantoea ananas

CHAPTER 2 CHARACTERIZATION AND SOURCE IDENTIFICATION OF BIOLOGICAL ICE NUCLEATING PARTICLES DEPOSITED YEAR-ROUND IN SUBTROPICAL PRECIPITATION

Overview

At temperatures warmer than -36 °C, the phase change from water to ice requires the

presence of impurities that serve as sites for ice nucleation (2, 15, 281). Dust aerosols have

important roles in the troposphere by serving as ice-nucleating particles (INPs) that contribute to

ice formation in clouds; a prerequisite for snow and most rainfall (2, 323). However,

atmospherically relevant mineral dust aerosols do not initiate freezing at temperatures warmer

than -15 °C (71), whereas certain microorganisms and biogenic molecules are very effective

INPs and have ice-nucleating (IN) activities > -10 oC (22, 76, 131, 139, 273, 323). For instance,

some strains of Pseudomonas syringae express an outer membrane protein (InaZ) that

structurally orders water molecules into an ice-like configuration, allowing the phase transition at

temperatures as warm as -1.8 °C (81, 87, 102, 324). Since its initial discovery, the IN phenotype

has been demonstrated in several other Gammaproteobacteria (76, 106, 266, 325–327) and a

species of Firmicutes (114), as well as in certain fungi (23, 123), algae (135, 328), and pollens

(22). High IN activities are associated with epiphytic bacterial communities on the leaves of

deciduous plants [~105 IN bacteria cm-2; (121)] and the microbiological decomposition of

detritus in a variety of plant (80, 81, 122, 139), soil (24, 259), and aquatic (135, 263, 264) ecosystems. While various environments are recognized to harbor biological INPs and their presence in precipitation is well documented (176, 273, 329, 330), longitudinal data sets are lacking and there is a need for studies that identify sources of atmospheric biological INPs and the conditions affecting their distribution in precipitation (71).

The warm temperatures at which biological INPs initiate freezing coupled with their ubiquity in near-surface air (235, 270, 329, 331), cloud water (176, 272, 332), and precipitation

(176, 273, 329, 330) implies that they could contribute to cloud ice formation under certain meteorological conditions (276). At cloud temperatures warmer than -10 °C, biological INPs could affect ice formation directly or via secondary ice formation processes (2, 31, 302, 312,

333–336). The interaction between biological INPs and cloud water may also indirectly affect climate by influencing cloud albedo, and consequently, the global radiative budget implies that they could contribute to cloud ice formation under certain meteorological conditions (276). At cloud temperatures warmer than -10 °C, biological INPs could affect ice formation directly or via secondary ice formation processes (2, 31, 302, 312, 333–336). The interaction between biological INPs and cloud water may also indirectly affect climate by influencing cloud albedo, and consequently, the global radiative budget (43, 303, 304, 337). Thus far, assessing the meteorological effects of biological INPs has been hampered by the shortage of data on their abundance and ice nucleating activity in the atmosphere, which are important parameterizations in cloud modeling studies (276). There is also evidence that the ice nucleation phenotype could provide a selective advantage to aerially transported microbial populations, facilitating their removal from the atmosphere and delivery to the surface in precipitation (333). Well-studied IN- active bacteria such as P. syringae live epiphytically on plants and are phytopathogens, suggesting that IN activity could play an important role in dissemination to new hosts (175).

To improve understanding of the environmental factors affecting the type, abundance, and sources of biological INPs in precipitation, we collected microbiological, geochemical, and meteorological data from 65 precipitation events in Louisiana (USA) over a two-year period. The purpose of this study was to determine if discernable patterns in the composition and

concentration of biological INPs deposited in precipitation were associated with season, air mass

history, cloud type, meteorological conditions, and precipitation chemistry. Multivariate

statistical analysis revealed temperature-specific characteristics of the INP populations, identified

explanatory variables that provide information on the atmospheric sources of the biological

INPs, and detected specific meteorological conditions where they had high abundances in the

precipitation. Our analysis implies that the bioaerosol sources could be quite distant from the

deposition site (i.e., continental Asia and high-latitude North America) and indicated that

terrestrial environments were the main source of warm-temperature biological INPs in the

precipitation sampled. Moreover, our results show that the abundance of biological and bacterial

INPs were strongly associated with the season, storm type, and the presence of specific bacterial

taxa in the precipitation.

Methods

Precipitation Sampling

With the exception of one sleet storm sampled in Alexandria, LA (31.3113 °N, 92.4451

°W) during March 2015, 64 precipitation events (61 rain, 1 sleet, and 2 snow events) were

sampled between May 2013 and July 2015 from the roof of a six floor building (~20 meters

above ground level) on the Louisiana State University campus in Baton Rouge, LA, USA

(30.4145 °N, 91.1783 °W). The precipitation was sampled at ambient temperatures by direct

collection in ten 120 L galvanized cans that were lined with clean, sterile 94 x 122 cm

polypropylene bags (Fisher Scientific, Pittsburgh, PA). A total of eleven cans (ten for samples

and one as a procedural control) were used to collect samples for each precipitation event. The

procedural control remained sealed for the duration of the precipitation event, then 3 L of sterile

deionized water that was poured into the can was collected in a manner identical to the samples.

A minimum of 3 L of rain or snow water equivalent was amalgamated from three collection cans

for each sample. Precipitation from Alexandria was collected using seven cans (six for samples

and one as a procedural control). For fourteen of the low accumulation events, the precipitation

collected in all 10 cans was pooled for DNA extraction. Immediately following each

precipitation event, the material collected was transferred to sterile 9 L carboys and stored at 4

°C in the dark until processed. Processing typically occurred within ~1 h, but in certain cases the

samples were processed up to 48 h after collection.

Precipitation samples for the measurement of dissolved organic carbon (DOC),

- 2- + - fluorescent dissolved organic matter (DOM), and major ions (NO3 , SO4 , Na and Cl ) were

collected separately from those used for microbiological analysis in a borosilicate glass funnel

and bottle that was thoroughly cleaned by washing with detergent, soaking in 10% HCl for 30

min, rinsing with ultrapure deionized water (18.2 M), and combusting at 400 °C for 4 h.

Following collection, the samples were filtered through pre-combusted GF/F filters (Whatman,

Inc.), stored frozen in the dark at -20 °C, and analyzed within 6 months.

Quantification of INPs and Cells

Immersion freezing assays were performed as described previously (273) but with the following modifications: no filter concentration was performed and 200 µL aliquots (rain or snow water equivalent) of the precipitation sample were placed into each well of a 96-well plate and sealed with adhesive film. Triplicates of each sample and experimental treatment were tested over a temperature range of -4 to -15 oC in 0.5 oC increments using a Neslab RTE 7 series

refrigerated ethylene glycol bath (Thermo Scientific, Waltham, MA). The number of wells

frozen at each temperature were recorded and the differential (the number of INPs activated at a

specific temperature) and cumulative concentrations (the number of INPs activated at all

temperatures warmer than a given temperature) of INPs were calculated by the method of Vali

(338). Each precipitation sample was tested by analyzing the triplicate preparations that were either untreated, heated for 10 min at 95 oC, or incubated with 3 mg mL-1 of lysozyme for 1 h

prior to the immersion freezing assay. The data and related calculations were used to assess the

total (i.e., untreated), biological (i.e., heat-sensitive), and bacterial (i.e., lysozyme-sensitive) INP

content of each sample. Given that only proteinaceous INPs may be sensitive to heat

denaturation and that not all bacteria are sensitive to lysozyme, this method should be viewed as

a conservative estimate for INPs of biological or bacterial origin.

DNA-containing cells were stained with SYBR Gold (Invitrogen, Carlsbad, CA) and

counted using epifluorescence microscopy (Olympus BX51-TRF, Center Valley, PA) according

to the method of Christner et al. (339). Cells in the precipitation were preserved by adding

sodium borate-buffered formalin (pH=8.2, stored at room temperature) to a final concentration of

5% v/v. Triplicate samples from each precipitation event were processed by filtering 10 mL of

sample onto 0.22µm, 25mm black polycarbonate filters (Millipore) and staining with a final

concentration of 25X SYBR-Gold (Invitrogen) for 15 min. in the dark. Cell density estimates

were obtained using an epifluorescence microscope (Olympus bx51) and data from 60 fields of

view (1 field of view=34636 μm2).

Amplification and Sequencing of 16S rRNA Genes

Initial testing showed that reliable DNA amplification typically required a minimum of 3

L of precipitation (data not shown). Therefore, at least 3 L of precipitation or sterile deionized water (for the procedural controls) was collected, filtered onto sterile 0.2 μm, 47 mm Supor PES membrane filters (Pall Corp., Port Washington, NY), and stored at -80 °C until processed.

The thawed 47mm Supor PES membrane filters were transferred to a laminar flow hood

and cut into small pieces using sterile scissors. The filter pieces were transferred to a bead

beating tube from the FastDNATM SPIN Kit for Soil (MP Biomedicals, Santa Ana, CA). The

DNA was extracted according to the manufacturer’s protocol, with the following modifications: in step 4, a mini bead beater was used to homogenize extracts for 70 seconds; step 5 was performed for 8 minutes; a 15 mL tube was used for step 7; and 100 µL of DNase/pyrogen-free water was used to elute DNA in step 16. The extracts obtained were further purified using steps

14-22 of the manufacturer’s protocol for the MoBio Power Soil Kit (MoBio Laboratories,

Carlsbad, CA). In these steps, 160 µL of solution C4 was added to the DNA extraction. The manufacturer’s protocol was followed for all other steps with the exception of adding 25 µL of solution C6 and incubating at room temperature prior to centrifugation. The DNA was stored at

−20 °C prior to polymerase chain reaction (PCR) amplification.

The V4 region of the bacterial 16S rRNA gene was PCR amplified in triplicate from each

DNA extract for sequencing on the Illumina MiSeq (Illumina, Inc., San Diego, CA, USA). The

PCR was carried out using the barcoded primers and methods of Caporaso et al. (340). The 25

µL reaction contained the following components: 5 Prime Master Mix (1X), 0.5 µM 515F, 0.5

µM 806R and nuclease free water (13 µL) and 2.0 µL of DNA. Thermal-cycling was carried out in an Eppendorf PRO S Master Cycler (Eppendorf North America, Hauppauge, NY, USA) under the following conditions: initial denaturation at 94 °C for 3 min. followed by 32 cycles of denaturation at 94 °C for 1 min., annealing at 50 °C for 1 min. extension at 72 °C for 1 min. 45 s, and followed by a final extension at 72 °C for 10 min. The amplicons obtained were evaluated by electrophoretic separation on a 1% agarose gel buffered with Tris-acetate-EDTA and stained with ethidium bromide. PCR products from the triplicate reactions were pooled, purified, and

concentrated using the Qiagen MinElute PCR Cleanup kit (Qiagen, Valencia, California, USA) with the optional 35% guanidine-HCl wash step to ensure removal of large primer-dimers.

Amplicons were stored at −20 °C until sequencing at Idaho State University’s Molecular

Research Core Facility (Pocatello, ID, USA) on the Illumina MiSeq Platform (Illumina Inc., San

Diego, CA) using the V2 500 bp kit.

Contamination was assessed by attempting PCR amplification with the V4 primers on extracts from the procedural controls, as described by Aho et. al. (341). The sequences obtained were analyzed using mothur software (342) and aligned to a SILVA bacterial 16S rRNA gene reference alignment (343). Contigs were assembled and parsed on the basis of unique barcodes attached to the 806R primer (340). Sequences that did not contain exact matches to the primer and barcodes utilized in the PCR amplification were discarded. Sequence library size was normalized by randomly sampling 49,263 sequences from each library, which was the number of sequences in the smallest library. Sequences were clustered into operational taxonomic units

(OTUs) based on a sequence dissimilarity of ≤ 0.03. OTUs were classified using the Ribosomal

Database Project classifier implemented in the mothur software and assigned to a particular taxon if they classified with ≥ 80% confidence. Sequences were filtered for quality with a 50- base sliding window and a minimum average quality score of 25, and those containing ambiguous bases, homopolymers (> 7 bases), or having lengths > 259 bases were eliminated from the dataset.

Inorganic and Organic Chemistry

The conductivity and pH of the precipitation samples were measured using a multi- parameter PCSTest probe (Oakton Instruments, Vernon Hills, IL). The concentration of major ions and DOC were determined on aqueous samples that were filtered through pre-combusted 25

mm, 0.7 μm GF/F Whatman filters (GE Healthcare, Chicago, IL). Samples of deionized water

were routinely analyzed and served as procedural blanks. A Dionex ICS-3000 ion

chromatography system was used to determine the concentration of major ions in the

precipitation samples. The system was equipped for anion separation with a 4×250 mm RFIC™

IonPac® AS18 column (Dionex Corporation, CA, USA), using a water:potassium hydroxide

eluent, and a 4×250 mm RFIC™ IonPac® CS16 column (Dionex Corporation, CA, USA) with a water:methanesulfonic acid eluent. Guard columns (Dionex Corporation, CA, USA) preceded each column (4×50 mm RFIC™ IonPac® AG18 guard column was used for the anion channel

and a 4×50 mm RFIC™ IonPac® CG16 guard column for the cation channel). Each sample was analyzed in triplicate. With this method, 2 carboxylic acids (formate and oxalate) and 13

− − 2− − 3− 2− 3− + + + + 2+ 2+ inorganic ions (F , Cl , NO , Br , NO , SO4 , PO4 , Li , Na , NH4 , K , Mg and Ca )

could be quantified. The limit of detection, calculated as three times the standard deviation of the

field blanks, was between 0.1 to 0.8 µM for all ions reported.

DOC concentrations were obtained from a GE Sievers 900 Total Organic Carbon

Analyzer. An average DOC measurement was calculated from three measurements of organic

carbon concentrations for each precipitation filtrate sample (sample size ~25 mL). Blank

samples of Milli-Q Water were measured between each sample to monitor successive sample-to-

sample contamination throughout instrument use. Acidification was not necessary prior to

experimentation due to an internal acidification step within the instrument. A Horiba Jobin Yvon

Fluoromax-4 Spectrofluorometer generated the Excitation Emission Matrices (EEMs) of the fluorescent dissolved organic matter (DOM) in the precipitation samples. This instrument is equipped with a Xenon lamp light source and a 1 cm path length quartz cuvette was used for all measurements. Excitation (Ex) wavelengths were scanned from 240-450 nm in 10 nm intervals

and emission (Em) was recorded between 300-560 nm in 2 nm increments. Data integration time was 0.25 s and data acquisition was carried out in signal/reference mode using a 5 nm bandpass on both Ex and Em monochromators, normalizing the fluorescence Em signal with the Ex light intensity. Absorbance spectra (190-1100nm) was incorporated into the spectral correction calculations of primary and secondary inner filter effects for post-processing the fluorescence data to generate EEMs (344, 345). Spectra were blank corrected against purified water from a

Milli-Q system each day. A Parallel factor analysis (PARAFAC) model was generated in

MATLAB by drEEM and the N-way toolbox scripts (346) to determine individual DOM fluorescing components in the EEMs.

Analysis of Meteorological Data and Ecoregions

Storm classification was based on cloud top data retrieved from The National Weather

Service (NWS) archive of Geostationary Operational Environmental Satellite-East (GOES-East) infrared and visible satellite imagery and Next Generation Radar (NEXRAD) Level III radar reflectivity provided by the National Climatic Data Center’s (NCDC) website

(https://www.ncdc.noaa.gov/nexradinv/). Backward trajectories (120-168 h) of air masses over the site at the time of each precipitation event were determined using the NOAA Air Resources

HYSPLIT (Hybrid Single-Particle Lagrangian Integrated Trajectory) Model, accessed via the

NOAA ARL READY website (http://arl.noaa.gov/HYSPLIT.php). Troposphere temperature profiles were retrieved from The NWS radiosonde data archive of stations located in Lake

Charles, Louisiana (LCH, station number 72240) and Slidell Muni, Louisiana (LIX, station number 72233). On-site meteorological data was collected continuously with a weather station

(Vantage Pro, Davis Instruments), and supplemented with data obtained locally through the

Louisiana Agriclimatic Information System (automated weather station at Ben Hur Agricultural

Fields, ~5 km SE of the main sampling location). Classification of convective and nimbostratus

precipitation based on radar reflectivity and satellite imagery was carried out according to

previous methods (347, 348). Tropospheric stability indices from NWS soundings (349, 350) were used to confirm the presence or lack of Convective Available Potential Energy (CAPE), which indicates the presence of convection. Cloud top heights were estimated using the

Equilibrium Levels (EL) and Maximum Parcel Levels (MPL), in addition to NCDC’s Level III echo top data, which estimates cloud height based on recorded pressure and temperature levels.

Herein, stratiform precipitation is defined specifically as precipitation that was collected from stratus and nimbostratus-like cloud systems independent of trailing stratiform regions from convective storm formations (9). Once cloud formation type was determined, six unique altitudes were chosen to be analyzed for 120-168 h backward trajectory analysis based on the CCL or

LCL of the precipitation event (Figure A-1 a-d). Trajectories were analyzed for previous interactions with the surface and/or mixed boundary layer (MBL) (Figure A-1e). This was accomplished by downloading the “tdump.csv” files produced by HYSPLIT, which detailed the recorded height above ground level of the trajectory being analyzed, in addition to the height of the MBL.

Convection occurs when the surface of the earth is heated unevenly, leading to the warming of air directly above the heated surface. This warmer air is more buoyant than the surrounding air and begins to rise. Once this “parcel” of warm air rises to the “convective condensation level” (CCL), water vapor will begin to condense and form water droplets, and subsequently, a cloud develops. If precipitation came from a cloud which was formed in the presence of convection the Convective Condensation Level (CCL) was used to estimate the height of the cloud base (Figure A-1 a). Convergence (Figure A-1 b) occurs when a low pressure

system is present. Air in high pressure regions moves towards lower pressure regions, leading to the convergence of air masses, forcing the air to move up in the atmosphere. Warm front (Figure

A-1 c) lifting occurs when a warm front advances and the less dense, warm air within that warm front is displaced upward over cooler, denser air ahead of it. Cold front (Figure A-1 d) lifting occurs when a cold front advances and displaces the warmer air ahead of it upward. Orographic lifting was not observed in this study and is not depicted. If precipitation came from a cloud which was formed in the absence of convection and by one of the prior three methods listed, the

Lifted Condensation Level (LCL) was used to estimate the height of the cloud base (Figure A-1 b-d).

As an example of how trajectories were analyzed, panels f and e in Figure A-1 show the altitudes and trajectories used for a particular rain event that occurred on August 25, 2014. The cloud base for this event was at approximately 1160 mAGL, and the cloud top was at approximately 16800 mAGL. Given that the cloud system developed through convection (Figure

A-1 a), the six altitudes chosen (depicted in panels a-d as the gray dotted lines) for HYSPLIT backward trajectory analysis were below the cloud at the time of precipitation in Baton Rouge.

The six backward trajectories were then examined for previous interactions with the MBL or surface using the tdump.csv files generated by HYSPLIT, in which trajectory height, MBL height, and surface height are listed at hourly intervals for each trajectory. If at any point along the trajectory history (histories of which ranged from 120-168 hours) the air masses descended into the MBL or interacted with the ground, the geographic coordinates these interactions were recorded for that event and plotted in R to determine the corresponding ecoregion (Figure A-1 f).

Panel f shows the geographic coordinates of the trajectories for this precipitation event where

interactions with the surface of MBL occurred. These coordinates were then mapped to the ecoregions outlined in Figure 2-1. Trajectory and ecoregion interactions are listed in Object 2-1.

Object 2-1. Excel file, 67KB in size, in tabular format, of all data collected and analyzed for Chapter 2. Available at the University of Florida’s Institutional Repository at this link: https://ufdc.ufl.edu//IR00011135/00001

Trajectory interactions with the surface or MBL were categorized based on North

American Level 1 ecoregions (351, 352). Briefly, an ecoregion is defined by the Environmental

Protection Agency and Commission for Environmental Cooperation as an ecological region that contains similar plant and animal communities, climate, geology, topography, hydrology, soil, land use, and natural resources (352). The most recent classification system defined 15 broad categories for Level 1 ecoregions. To increase homoscedasticity and the number of degrees of freedom for downstream statistical analyses, several regions of similar ecology, geography, and climate were combined. For this analysis, the following Level I ecoregions were combined: the

Arctic Cordillera, Tundra, Taiga, Hudson Plain, and Northern Forests were combined and designated as “High Northern Latitudes”; the Northwestern Forested Mountains and Marine

West Coast Mountains were combined and designated as “Northwest Forested Mountains”; the

North American Deserts, Mediterranean California, Southern Semi-Arid Highlands, and

Temperate Sierras were combined and designated as “Deserts and Semi-Arid Highlands”; and the “Tropical Dry Forests” and “Tropical Wet Forests” were combined and designated as

“Tropical Forests” (Figure 2-1). Backward trajectories that interacted with the MBL in Asia occurred at latitudes above approximately 35 oN and longitudes higher than 90 oE, and were

therefore classified into a single ecoregion called “East Asia” (Figure 2-1). The Great Plains and

eastern US ecoregions were not combined with other ecoregions and are referred to as “The

Great Plains” and “Eastern Woodlands and Wetlands,” respectively (Figure 2-1). Marine regions

were divided into Pacific Maritime, North Atlantic Maritime, and South Atlantic Maritime

(Figure 2-1).

Statistical Analyses

All statistical procedures were performed using Version 9.4 of the SAS System for

Windows. Graphs and plots were produced using R Software Version 3.2.1 (The R Core Team

2015). The statistical procedures (Exploratory Factor Analysis, Multiple Imputation, Analysis of

Variance, Multivariate Analysis of Variance, Mann-Whitney U-Test, Welch’s Tests, Kruskall-

Wallis Test, Pearson’s and Spearman’s Correlations, and Tukey’s Honest Significant Difference

post-hoc analysis) were performed using SAS software, Version 9.4 of the SAS System for

Windows. Graphs and plots were produced using R Software Version 3.2.1 (The R Core Team

2015).

Prior to hypothesis testing with analysis of variance (ANOVA) and multivariate analysis of variance (MANOVA), the raw data were screened for univariate and multivariate normality using The Shapiro-Wilk Test and through visual inspection of Q-Q plots. The assumption of homoscedasticity was verified using Levene’s test, univariate outliers were examined based on z- score distributions, multivariate outliers identified using Mahalanobis distance, and linearity/collinearity evaluated through visual inspection of bivariate scatter plots. Log and arcsine transformations were used on distributions that violated assumptions of normality. For distributions not corrected by data transformations, hypothesis tests that assume non-normality or heteroscedasticity were used (i.e., Mann-Whitney U-test, Kruskall-Wallis one-way ANOVA, and

Welch’s Test). For extreme outliers (values more than three times the interquartile range), raw data values were adjusted according to the method of Tabachnick and Fidell (353). Multiple imputation was used to provide missing INP concentration data prior to Exploratory Factor

Analysis (EFA) and hypothesis testing. The differential INP concentrations were grouped and

summed based on the results of the EFA. For example, the INP concentrations used to represent

o Bio-5 to -10 is the summed differential concentrations of INPs active between -5 and -10 C.

For hypothesis testing, the dependent variables were always the summed differential INP

concentrations for each INP category determined by EFA, which were continuous variables.

When the independent variables were continuous (Cell abundance, pH, conductivity, major ion

concentrations, DOC concentrations, PARAFAC Components C1-C3 intensities, OTU sequence

reads), Pearson’s R and Spearman’s Rank correlational analyses were used. Note that for

correlations between INP concentrations and bacterial taxa, the number of OTU sequence reads

was used for the analysis. When the independent variables were categorical (ecoregion

classification, cloud type, season, and precipitation type), ANOVA and MANOVA were used.

For post-hoc analysis of ANOVA and MANOVA, Tukey’s Honest Significant Difference (HSD) analysis was used. Tukey’s HSD test is a post-hoc analysis that compares the means of each group to find significant differences between groups (353).

Results

Total, Biological, and Bacterial INPs in Louisiana Precipitation

The data collected and analyzed during this study are provided in Object 2-1. Following

Christner et al. (273), INPs in untreated samples are referred to as “total”; those sensitive to

heating at 95 °C for 10 min are inferred to be proteinaceous in origin and classified as

“biological”; and those inactivated by digestion with 3 mg mL-1 lysozyme at 22 °C for 60 min.

are classified as “bacterial”. The warmest temperature at which ice nucleation occurred in all of

the precipitation samples was -10 °C, but freezing was detected at -4 °C in six of the precipitation samples. The number and activity temperature of total, biological, and bacterial

INPs varied significantly between precipitation events, with concentration values for a given

temperature ranging two to three orders of magnitude (Figure 2-2 a-c). The cumulative concentrations of total, biological, and bacterial INPs active at -15 °C averaged 16,200, 9,000, and 6,600 INPs L-1 precipitation, respectively, with biological and bacterial INPs comprising

56% and 41%, respectively, of the total cumulative INPs (Figure 2-2d, Table 2-1). Based on the total differential INP data, >95% of the INPs active at warmer than -10 °C were inferred to be biological in origin (Table 2-1). At temperatures colder than -10 °C, biological INPs were a

maximum of 66% of the total differential INPs and 35% of the total differential INPs at -15 °C

(Table 2-1). On average, less than half of the INPs at temperatures colder than -8 °C were

bacterial, but they represented the vast majority of INPs that were active at > -8 °C (Table 2-1).

The concentrations for each category of INP increased exponentially with decreasing

temperature to approximately -14 °C (Figure 2-2 a-c). For fifteen of the precipitation samples, all

of the 96 wells being tested were frozen before cooling to -15 °C, indicating high concentrations

of INPs in the sample. Since at least one of the wells must remain unfrozen to calculate the INP

concentration (338), abundances at the lowest temperatures could not be directly determined for

these samples. Accordingly, the data reported at –13 to –15 oC in Figure 2-2 are missing

observations for samples that had the highest INP concentrations at these temperatures.

Specifically, the total differential INP data (N=65) lacked values at -15 °C (N=3), ≤ -14 °C

(N=8), and ≤ -13 °C (N=4); the biological differential INP data (N=65) lacked values at -15 °C

(N=4), ≤ -14 °C (N=6), and ≤ -13 °C (N=6); and the bacterial differential INP data (N=56) lacked values at -15 °C (N=3), ≤ -14 °C (N=7), and ≤ -13 °C (N=7). To use observations with missing INP data in subsequent statistical analyses, multiple imputation was performed to estimate the values at these colder temperatures based on the pattern of available data [Table A-

1, (354, 355)]. Satisfactory relative efficiency values were produced for all INP types using 5

imputations and a monotone regression method (355). The descriptive statistics for the INP

concentration data are listed in Table 2-2.

Exploratory Factor Analysis of the INP Data

Exploratory factor analysis (EFA) was used to examine correlations between variables

and the differential INP concentrations active from -4 to -15 °C (Table 2-3). EFA is a statistical method that identifies patterns in multivariate data sets by searching for latent variables (referred to as factors), which are identified based on shared variance between the measured variables

(356). In other words, the factors identified by EFA are measurements of intercorrelation between measured variables in the data set. Factor “loading” is the correlation between the factor and measured variable. A high positive loading value (1.0 ≥ x ≥ 0.5) of a variable indicates that its variance in the dataset is mostly explained by the factor it is loading high onto. As such, a loading cut-off value of > 0.50 was used to determine the variables retained for each factor

(Table 2-3). The total variance of the factor is an estimation of how well the factor explains variance in the dataset, which can be viewed as an indication of the influence the latent variable had on the measured variables.

The analysis for total and bacterial INPs produced two significant factors, whereas biological INPs produced three significant factors (Table 2-3). Total INP concentrations between

-5 and -11 °C positively correlated with factor 1, and those between -11 and -14 °C positively correlated with factor 2; however, the factor loadings for -4 and -15 °C were not significant. The biological INP factor pattern produced similar results to the total INP factor pattern: INP concentrations between -5 and -10 °C loaded onto factor 1 for biological INPs; -13 and -14 °C loaded onto factor 2; -11 and -12 °C loaded onto factor 3; and -4 and -15 °C did not load

significantly onto any factor. For bacterial INPs, factor 1 positively correlated with INP

concentrations between -5 and -10 °C, whereas factor 2 positively correlated with INP

concentrations at -4 and -12 °C and negatively correlated with INP concentrations at -14 °C. Due

to the small number of observations for freezing at -4 °C (n=6), bacterial factor 2 (-4 and -12 °C)

was dropped from subsequent analyses. For simplicity, each factor is referred to hereafter by INP

type (total, biological, or bacterial), with the temperature range as subscript (e.g., factor 1 for the

total INPs active between -5 to -11 °C is ‘total-5 to -11’).

INP Factor Concentrations Correlate With the Physical, Chemical, and Microbiological

Data

The average concentration of DNA-containing cells in the 65 precipitation events was

6.4±9.5 x 105 cells L-1 (±SEM), and the statistical average and data range for the chemical and

physical measurements are provided in Table 2-2. Cell concentrations correlated positively and

significantly with the warm INP concentrations (total-5 to -11, bio-5 to -10, and bac-5 to -10; Table 2-4).

Precipitation pH correlated significantly with total-5 to -11 (Pearson’s r = 0.45) but not with any

- 2- other factor. While NO3 and SO4 concentrations did not significantly correlate with any factor,

+ - bac-5 to -10 correlated negatively and significantly with Na (Pearson’s r < -0.42), and Cl

correlated negatively and significantly with total-5 to -11 (Table 2-4). DOC concentration and

precipitation conductivity (Object 2-1) did not correlate with any of the INP factors.

PARAFAC modeling revealed three components in the excitation emission matrices of

all precipitation samples (Figure 2-3a-c). PARAFAC component one (C1) correlated

significantly (Pearson’s r > 0.39) with all the warmer INP factors (Table 2-4) and showed

maximum fluorescence in two regions: one typically associated with the humic-like DOM of

terrestrial soils and plants (excitation and emission wavelength: 250/400-440 nm) and the other

associated with biologically-influenced humic-like DOM of marine and freshwater environments. PARAFAC C2 correlated with bio-11 to -12 while PARAFAC C3 did not correlate

with any factors (Table 2-4). The tyrosine- and tryptophan-like fluorophore components (C2 and

C3, respectively) are amino acid-like constituents of the dissolved organic carbon pool that are

indicative of dissolved organic matter produced through microbial processes (357–359). There

- 2- were also significant correlations of PARAFAC C1 fluorescence with NO3 and SO4 , but not

with Cl- or Na+ concentrations.

INP Factors Correlate with Season, Cloud Type, and the Air Mass History

Multivariate analysis of variance (MANOVA) was conducted to investigate if INP

concentrations varied based on the history of air mass interactions with the MBL. The results

(Table A-2, Figure 2-4) indicated that total, biological, and bacterial INP concentrations were

significantly different based on air mass origin and interactions with the MBL (p<0.05). Air mass

trajectory interactions with the East Asia ecoregions produced the highest concentrations of INPs

for all factors (Figure 2-4). While most of the East Asia trajectories interacted with the temperate

and taiga regions of Asia, one event (March 5, 2015) did interact with the arid deserts and xeric

shrub lands of China and Mongolia. Tukey’s Honest Significant Difference post-hoc analysis

indicated that for the total-5 to -11, bio-5 to -10, and bac-5 to -10 factors, the High Northern Latitudes

ecoregion had statistically similar INP concentrations as those for the East Asia ecoregion (Table

A-2). The Eastern Woodlands and Wetlands ecoregion and Deserts and Semi-Arid Highlands ecoregion produced the lowest concentrations of total-11 to -14, bio-11 to -12, and bac-5 to -10 (Figure 2-

4). In general, the Pacific, North Atlantic, and South Atlantic Maritime ecoregions were inferred

to be minor sources of all INP classes. Trajectory analysis indicated that none of the air masses

originating from the Tropical Forest ecoregion during this study interacted with the MBL or

surface.

To validate trends unmasked from the INP and air mass data analysis, DOM PARAFAC

component intensities were analyzed with backward air mass trajectories (Table A-3). Overall, the fluorescent DOM components showed differences between air mass trajectories that interacted with central continental environments (i.e., the Great Plains, Eastern Woodlands and

Wetlands, and Deserts and Semi-Arid Highlands), northern and more distant continental environments (East Asia, the High Northern Latitudes, and the Northwest Forested Mountains), and marine environments (Pacific Maritime and South Atlantic Maritime) (Figure A-2). Since none of the storm air mass histories analyzed originated from North Atlantic Maritime environments, no PARAFAC data are available from this ecoregion. Significant correlations were identified for DOM PARAFAC components and select major ion concentrations: C1 to

- 2- - NO3 (r=0.46 and =0.61) and SO4 (r=0.58 and =0.66); C2 to NO3 (r=0.36 and =0.42) and

2- - SO4 (r=0.30 and =0.54); and C3 to NO3 (r=0.33). None of the components correlated with

Cl- or Na+ concentrations.

The concentration of all INP types correlated with season (MANOVA, p<0.001; Table A-

4) and precipitation in winter contained the highest concentrations of all INPs (Figure 2-5b).

Highly significant differences in INP concentrations were associated with the type of mesoscale

cloud system (i.e., stratiform vs. convective) responsible for the precipitation (p <0.05; Table A-

4), with stratiform systems containing the highest concentrations for all INP factors (Figure 2-

5a). A subset of the stratiform events with ice-containing precipitation (sleet or snow, N=4) had the highest concentrations of all INP classes (Figure 2-5c), and a MANOVA showed these

differences were significant from those observed in rain samples (p= <0.05; Table A-4). Several

of the locally measured meteorological variables, including surface temperature and surface wind

speed, correlated well (Pearson’s r > 0.30) with INP concentrations (Table A-5).

Correlations Between Abundances of Bacterial Operational Taxonomic Units and INPs

Of the 65 samples collected for DNA extraction, 20 of the samples were excluded from further analysis because amplification of the 16S rRNA gene failed or they were associated with trials in which spurious amplicons were generated in the procedural controls. The composition of bacterial assemblages in 45 of the precipitation events was assessed by comparing the amplified

16S rRNA genes from the samples, and a total of 60,289 operational taxonomic units (OTUs) were identified. This analysis revealed that 1,425 OTUs had statistically significant positive correlations to the INP factors. Of these 1,425 OTUs, only those which had reads accounting for

>0.1% of the total number of reads across all samples were retained for further analysis

(Supplementary Table S6). Spearman’s ρ correlations indicated that OTUs classifying within the

Bacteroidetes had the most significant correlations with all INP factors—especially those within the Cytophagales order. Several of the Cytophagales taxa (Unidentified genus from

Cytophagaceae, Hymenobacter, and Flexibacter) showed the highest and most significant correlations with total-5 to -11, bio-5 to -10 and bac-5 to -10 INPs (Table A-7). Additionally, the family

Rikenellaceae from the order Bacteroidales, the family env.OPS_17 from the order

Sphingobacteriales, and the genus Segetibacter from the order Chitinophagales contained highly

significant correlations with all INP classes except for bio-11 to -12.

Firmicutes and unclassified divisions also had a relatively large number of significant

positive correlations with the INP factors (Table A-6). Within the Firmicutes, OTUs significantly correlating with all warm classes of INPs (total-5 to -11, bio-5 to -10, and bac-5 to -10) were limited to

the Carnobacteriaceae. A broader phylogenetic range of OTUs positively correlated with the

colder INP factors (total-11 to -14, bio-11 to -12, and bio-13 to -14), and most of these taxa affiliated with

the Bacteroidetes, Firmicutes and Proteobacteria, but there were also representatives from the phyla Cyanobacteria, Planctomycetes, Spirochaetes, and Verrucomicrobia (A-6). Overall, a smaller number of OTUs correlated with the bacterial INP factors, and those that did belonged to the Bacteroidetes, Proteobacteria, or unclassified divisions. Many of the OTUs correlating with

INP factors had significantly higher abundances in precipitation from nimbostratus clouds

(N=16) and during the winter (N=16; Figure A-3, Table A-7).

Discussion

During this two-year study, observations from a range of storm origins and types provided the opportunity to examine linkages between biological INPs in precipitation and geographic source, season, cloud lifetime and precipitation development, and the meteorologically-influenced dispersal of microorganisms in the atmosphere. The key results from this analysis can be summarized as follows: (i) the sampled precipitation contained several

“classes” of INPs with distinct behaviors; the highest concentrations of INPs likely originated from (ii) terrestrial regions geographically distant from Louisiana and (iii) nimbostratus clouds, ice-phase precipitation, and winter storms; (iv) certain bacterial taxa correlated significantly with the INP concentrations; and (v) the concentrations of INPs observed in the precipitation implied sufficient abundances in the atmosphere to affect precipitation production in certain cloud types.

The significance and implications for each of these findings are discussed in greater detail below.

INP “Classes” Identified in Louisiana Precipitation

The EFA grouped differential INP concentrations by temperature of activation (Table 2-

3), implying the sampled precipitation contained distinct types or classes of INPs. Since the total-

5 to -11, bio-5 to -10, and bac-5 to -10 factors were highly correlated to each other (r=0.83 to 0.96,

ρ=0.83 to 0.94; Table S9), they are likely capturing similar aspects of the warm temperature IN

activities. Based on the properties of known IN species and biological INPs (81, 323, 360), the total-5 to -11, bio-5 to -10, and bac-5 to -10 factors may represent activities conveyed by intact bacterial

cells that harbor IN proteins or heat-sensitive compounds. Indeed, cell concentration correlated

significantly with the warm INP factors (total-5 to -11, bio-5 to -10, and bac-5 to -10, Table 2-4). All of

the warm INP factors also correlated significantly with PARAFAC component C1 (Table 2-4),

which is known to be associated with terrestrially-derived DOM (358, 359, 361, 362). Taken

together, these results suggest that the warmest temperature INPs in the precipitation were

largely bacterial in origin, conveyed by proteinaceous IN activity, and aerosolized from

terrestrial environments. That no correlation existed for bio-11 to -12 or bio-13 to -14 with pH is

interesting considering proteinaceous IN activity in certain fungi is heat-sensitive but unaffected by low pH (23). As such, the “colder” classes of biological INPs identified could correspond to fungi possessing IN proteins (123) or DOM with IN activity (24, 259).

It is important to note that our method for classifying INPs provides a conservative

estimate of those from biological and bacterial sources (273). For example, the non-

proteinaceous IN material of some pollens and bacteria are resistant to heating at 100 °C (22,

114, 132), and therefore, would not be identified as “biological” by our method. Likewise, the

use of lysozyme sensitivity to diagnose bacterial INPs should underestimate their concentration

because disparate cell wall compositions allow some bacteria to resist its hydrolytic activity.

Nevertheless, the significant correlation (Pearson’s r=0.54) between cell abundance and

lysozyme-sensitive INPs (Table 2-4) implies that any bias introduced by this approach was

relatively uniform across samples.

Potential Geographic Origins of INP Classes in Louisiana Precipitation

Previous studies have detected high abundances of biological INPs in plant and soil environments (80, 363), and the emission of IN strains of P. syringae and Erwinia herbicola from agricultural crops (157, 221). Our analysis revealed that air masses interacting with the

MBL in mid-latitude continental regions replete with agricultural, woodland, and herbaceous ecosystems (Eastern Woodland and Wetlands and the Great Plains ecoregions; Figure 2-1) did not subsequently produce precipitation with the highest concentrations of total-5 to -11, bio-5 to -10,

and bac-5 to -10 INPs observed (Figure 6b, d, and g). In fact, air masses from the Eastern

Woodland and Wetland ecoregion produced the lowest INP concentrations in precipitation of all

non-maritime locations (Figure 2-6c, f, and g). East Asia, the Northwest Forested Mountains,

and/or the High Northern Latitudes ecoregions were the largest sources of all INP classes (Figure

2-6). Although few data are available, a recent study implicates high northern latitude aerosols

in North America as sources of biological INPs (364). The identification of soil dwelling taxa

[for example, Hymenobacter are ubiquitous in soils, especially desert soils (365–373)] that

correlate well with the warm INP factors also suggests that terrestrial soils were the most likely

sources of the warm temperature INPs.

Nearly all the INP factors correlated significantly (p<0.05) with DOM PARAFAC C1

(Table 2-4, Figure 2-6h). The region of maximum fluorescence in C1 (Ex/Em maximum at

250/410 nm) is known to represent constituents of terrestrially-derived DOM from plants, soils,

estuaries, wastewater, and agricultural catchments (358, 359, 361, 362). Fluorescence at higher

Ex/Em wavelengths maximum of C1 (300/410 nm) exhibited intensities similar to the “M peak”

described by Coble et al. (357) in shallow eutrophic marine waters from the Gulf’s of Mexico

and Maine, transitional waters of Puget Sound (374), and freshwater environments (362, 375,

376). DOM associated with C1 typically consists of higher molecular weight aromatic chemical

species, characteristic of material that has been highly processed by biological and geochemical

activities (362, 375, 376). However, that the C1 component did not correlate with the Cl-

concentration in precipitation suggests its source was not marine. This contention is further

- supported by the significant positive correlation of C1 with NO3 (r=0.46 and =0.61), which is

an indicator of a terrestrial aerosol source (377) . Together these results support the contention

that continental environments were the main source of warm INPs in the precipitation analyzed.

Aircraft-based measurements in air masses above the Sierra Nevadas have shown high

concentrations of biological INPs in clouds and precipitation containing desert dust from Asia

(229). Additionally, studies have also shown that the intercontinental transport of

microorganisms as aerosols is possible (165–167, 229). This together with the observation of

terrestrial DOM (i.e., PARAFAC component C1) in air masses originating from Asia implies an

extracontinental source for the INPs in these precipitation samples (Figure 2-6). The same can be

said for the High Northern Latitudes ecoregion, however, we are not aware of prior studies that

have examined geographic regions in the tundra or northern forests of North America as

atmospheric sources of biological INPs.

The bio-11 to -12 factor is the only class of INPs that has a significant correlation (r=0.40) with the tyrosine-like DOM chemical species associated with PARAFAC C2. Previous studies have identified C2 in microbially-produced DOM from terrestrial (362, 378) and freshwater ecosystems (362). Additionally, the bio-11 to -12 factor is the only one that did not have the highest

concentrations when air masses interacted with the surface in Eastern Asia or the High Northern

Latitudes. Instead, its highest concentrations were observed in storms that originated from the

Forested Mountains ecoregion (Figure 2-6e).

INP Concentration as a Function of Season and Meteorology

Winter storms contained the highest concentrations of nearly all the INP factors (Figure

2-5b and Object 2-1), supporting the hypothesis that seasonal changes in meteorological patterns

and/or the source ecosystems of bioaerosols influenced the abundance of INPs in the

precipitation. A considerable fraction of OTUs that correlate with the INP data had higher

abundances in winter precipitation (Figure A-3; Table A-7) and are related to psychrotolerant or

psychrophilic species, suggesting a linkage between cold adaptation, lower winter temperatures,

and IN activity. This deduction is also supported by surface temperature being inversely

correlated with the warmest INP classes (total-5 to -11, bio-5 to -10, and bac-5 to -10; Table A-5). An

alternative explanation for this finding is the possible inactivation of biological INPs while

descending through the atmosphere or in the collection cans at the surface during summer. If this

was the case, then the actual number and activity of INPs at cloud heights may be

underestimated based on inferences from summer rain samples. However, negative correlations

between INP concentration and the sample collection time (amount of time precipitation samples

sat in collection cans before being processes) during the summer months were not observed,

suggesting the data were not affected by this potential sampling artifact. In fact, the only

significant correlation to collection time observed was positive and associated with the bio-13 to-14

data (r =0.648 p=0.005). Some IN active pollens release their ice nucleating macromolecules into solution when immersed in water (22), providing one possible explanation for this curious result.

A subset of the wintertime nimbostratus events contained all the extreme outlying INP data (Figure 2-1), had the highest concentrations for all factors, had significantly higher concentrations of INPs compared to winter convective storms (Figure 2-5b), and were

significantly enriched for OTUs that correlated with the INP data (Figure A-3). There were only

four observations of ice-containing precipitation (snow or sleet) during this study, and these

samples had the highest INP concentration and activity observed (Figure 2-5c). While it is

tempting to suggest this result as evidence for the involvement of biological INPs in ice phase

precipitation, it may also be explained by differences in below-cloud scavenging efficiencies between snowflakes and raindrops. However, previous work documented distinct microbial assemblages in snow versus the air mass it precipitated from (329), suggesting that snowflakes may be relatively inefficient at scrubbing bacterial-sized aerosols from the atmosphere. Although there are a number of uncertainties regarding aerosol scavenging in snow versus rain (379–381), cell concentrations in the snow and sleet were not significantly different from those observed in rain samples, implying similar scavenging processes regardless of precipitation phase.

Potential Phylogenetic and Geographic Sources of the Ice Nucleating Bacteria

The 16S rRNA gene sequences from 45 of the 65 precipitation events were successfully amplified, sequenced, and analyzed (341). Of the OTUs that correlate significantly with INP concentrations (Table A-6), many are related to bacterial taxa documented in soils, plant ecosystems, air, and precipitation (231, 341, 382). Although there were no significant trends between INPs and the known ice nucleating genera Pseudomonas or Pantoea, there are significant correlations between OTUs from Xanthomonadaceae (i.e., containing the IN species

Xanthomonas campestris) and the total-5 to -11, total-11 to -14, and bio-5 to -10 factors. Taxa in the

phylum Bacteroidetes had the highest correlations with INP concentrations and were abundant in the samples (Table A-6), including OTUs from the Sphingobacteriales and Cytophagales.

Several genera from these orders have been reported in desert soils (Hymenobacter; (371, 383,

384)); snow and hail (Hymenobacter and genera from Sphingobacteriales; (228, 385)); and air

samples collected over the Northwest Forested Mountain/Desert and Semi-Arid Highlands ecoregions and Asia (Flexibacter, Segetibacter, Hymenobacter and genera from

Sphingobacteriales; (228, 331, 386–388)). The strong correlations between bacteroidetal OTUs and the warm INP factors (total-5 to -11, bio-5 to -10, and bac-5 to -10) is noteworthy because ice

nucleating activity has not been documented within this phylum. In particular, Hymenobacter

and Segetibacter taxa were strongly correlated with the total-5 to -11, total-11 to -14, bio-5 to -10, and

bac-5 to -10 INP data, and members of these genera have also been reported in air and precipitation

containing Saharan and Asian dust (228, 270, 331). In fact, Meola et al. (228) showed that

similar Cytophagaceae phylotypes in European snow sequences coincided exclusively with the

presence of Saharan dust. Several of the OTUs strongly correlated with the INP data (Table A-6)

classified within groups of human and animal-associated bacteria (e.g., Rikenellaceae, Blautia,

and Roseburia, (389–391)), indicating an anthropogenic and/or agricultural influence on the

microbial composition of the precipitation analyzed. There were also strong correlations between

the INP data and numerous unclassified taxa (Table A-6), supporting the possibility for the

presence of ice nucleating activity in these poorly characterized bacterial lineages. Considering

that correlations may be illusory, exploration of species in these genera requires further

investigation to determine if they possess species with the IN-phenotype.

The taxonomic data and chemical composition of the precipitation also provide

information to aid in assessing the potential ecological sources of the INPs sampled. For

instance, the warmer classes of INPs (total-5 to -11, bio-5 to -10, and bac-5 to -10 factors) positively

correlated with taxa associated with desert and arctic soils (228, 325, 366, 371, 383, 384) and

DOM typically associated with terrestrial plant, soil, and freshwater ecosystems (362, 376), but was negatively correlated with the Cl- concentration. This implies that the most efficient INPs

may have been sourced from desert and high northern latitude continental environments as

opposed to marine habitats (Figure 2-6), which is supported by HYSPLIT back-trajectory

analysis that showed air masses interacting with the East Asia and High Northern Latitude

continental surfaces contained the highest INP concentrations (Figure 2-6 b-g, Table A-2). In contrast, OTUs related to known marine taxa (e.g., Oceanospirillaceae) correlated with colder classes of INPs (total-11 to -14, bio-11 to -12, and bio-13 to -14), suggesting marine environments as their

source.

Potential Implications for Biological INPs on Meteorological Processes

While ice-phase clouds produce the majority of precipitation on a global scale (392), there are large uncertainties on ice formation in mixed-phase clouds, the abundances of biological INPs at cloud altitudes, and their involvement in precipitation generation (71). To investigate the possibility that biological INPs have sufficient abundances to affect ice formation in cloud water droplets, we used the precipitation data to estimate their in-cloud abundances.

Below-cloud scavenging complicates efforts to directly relate precipitation to cloud water composition, and therefore, we estimated boundary conditions following the approach of Petters and Wright (393).

Modeled simulations of summer and springtime convective clouds over North America imply that as few as 1 INP m-3 (of air) active at -10 °C could produce observed rates of precipitation (317, 337). At temperatures ≥-10 °C, the Louisiana precipitation samples averaged cumulative concentrations of 1,300 total INPs L-1, 1,100 biological INPs L-1, and 600 bacterial

INPs L-1 of precipitation (Figure 2-2). Based on these data, we estimated the relative abundance

of INPs within a cubic meter of developing cumulonimbus cloud (i.e., convective cloud) and

stratocumulus cloud (i.e., stratiform). Over continental regions, cumulonimbus clouds typically

contain cloud droplet sizes in the range of 12 to 16 μm at concentrations of approximately 500

cm-3 (394, 395). Assuming a median cumulonimbus cloud droplet size of 14 μm (312), the

average cumulative concentration of INPs active at -10 °C in precipitation would represent

approximately 0.5 – 1 INP m-3 of air, which is similar to modeled values affecting precipitation

rates [1 INP m-3;(337)]. When considering the highest observed INP abundances (16,100 total;

15,900 biological; and 5,900 bacterial INPs L-1 of precipitation, active at ≥-10 °C; Figure 2-2),

concentrations as high as 5-10 INPs m-3 of cumulonimbus cloud are inferred.

The mechanisms responsible for primary and secondary ice formation in nimbostratus

clouds differ substantially from those of convective clouds (31, 302–304). The Wegener-

Bergeron-Findeisen (WBF) process is thought to be the dominant process of ice formation in

mixed-phase clouds with updraft speeds of <2 m s-1 (convective clouds typically contain updraft

speeds >2 m s-1), which is consistent with nimbostratus cloud conditions (55, 333, 396). Hence,

it is possible that the WBF process allows efficient removal of warm temperature INPs from a

nimbostratus cloud via ice formation and subsequent precipitation initiation. Although INPs

active at temperatures warmer than -9 °C are also important for the Hallet-Mossop mechanism,

this process requires updraft speeds of >3 m s-1 for riming to work effectively and may only be

relevant to convective clouds (302). Stratus clouds that form over continents typically contain

fewer (~250 cm-3) and smaller cloud droplets than convective clouds, with average diameters of

~10 μm (395). Based on these meteorological parameters, we infer concentrations in stratus

clouds of 0.1 – 0.2 INP m-3 based on the average biological INP concentrations, and 0.8 – 2 INP

m-3 using the highest INP concentrations observed. To our knowledge, INP concentration

thresholds for precipitation formation in stratus or stratus-like clouds have not been constrained within an aerosol-cloud modeling framework. As our data show that biological INPs were at

their highest concentrations in precipitation from stratus clouds, and ice-phase stratiform precipitation in particular (Figure 2-5), such modeling efforts represent fertile territory for further study.

It is important to note that there are a number of uncertainties inherent to our estimates of cloud INP concentrations based on data from precipitation, including the actual cloud droplet sizes, the cloud thermodynamic properties, and the contribution of below-cloud scavenging.

With respect to the latter, aerosols the size of most bacteria (0.5-1.5 μm) are typically referred to as the “scavenging gap” in aerosol size distributions because particles of this size are not efficiently scrubbed by raindrops regardless of the rainfall rate (397). Previous studies have shown similar concentrations of INPs in cloud water compared to precipitation at ground level

(176, 225, 398), suggesting that below-cloud scrubbing may have a negligible role in altering the composition of bacterial-sized INPs in raindrops.

Concluding Remarks

Despite the fact that biological INPs have been studied for nearly 50 years, little is known about the microbiological and ecological sources of these particles to the atmosphere. In this study, we analyzed precipitation from 65 storms originating from air masses that had traversed regions of Asia, North America, the Pacific Ocean, the Atlantic Ocean, and the Gulf of Mexico

(Figure 2-1), providing new observations to assess the geographic sources of INPs deposited with precipitation in the southeastern USA. Various studies have detected biological INPs in marine waters and concluded that the ocean is a major atmospheric source of these bioaerosols

(135, 263, 264, 363). However, we observed that pre-storm air masses that interacted with marine environments generated precipitation with the lowest concentrations of INPs observed

(Figure 2-6). In contrast, the concentrations of all INP classes were highest when air masses had

interacted with continental ecoregions. The continental origin of these bioaerosols was supported

by positive correlation with the abundance of DOM and bacterial taxa typically associated with

terrestrial ecosystems, and negative correlation to marine aerosol proxies (Figure 2-6h). These

results lead to the conclusion that forest and soil ecosystems are important year-round atmospheric sources of biological INPs to precipitation in this subtropical region. Unexpectedly, our analysis implicated the High Northern Latitudes and Eastern Asia ecoregions as sources that produced the highest concentrations of INPs in Louisiana precipitation (Figure 6), corroborating similar observations in the western USA (229).

Multiple studies have confirmed the ubiquity of biological INPs in the atmosphere and precipitation (71, 272, 273), but the meteorological conditions under which they may have roles in affecting cloud ice formation and precipitation has remained speculative. We observed the highest concentrations of total, biological, and bacterial INPs in precipitation associated with winter storms comprised of nimbostratus-like cloud formations that originated from the western

US (Figure 2-6). Since nimbostratus are low level, relatively warm clouds (9), biological INPs would be in a favorable position to affect precipitation processes if sufficiently abundant. Indeed, our estimates from precipitation samples imply that INPs originating from warm clouds (tops warmer than -15 oC) were at concentrations predicted to influence ice formation and

precipitation. The role of biological INPs in meteorological process may be especially relevant to

stratus-like clouds where updrafts are minimal and the WBF process is significant for precipitation production. Considering that low temperature and nutrient limitation affect the IN activity of P. syringae (103), metabolically active microorganisms are present in cloud droplets

(250), and liquid water can persist for days in nimbostratus clouds at temperatures conducive to

metabolism [10 °C to -25 °C (9, 399)], conditions within a nimbostratus cloud might even promote in situ expression of the ice nucleation phenotype.

Bacterial ice nucleation has been thought to be limited to certain species of

Gammaproteobacteria (i.e., P. syringae, P. viridiflava, P. fluorescens, Pantoea agglomerans, and

Xanthomonas campestris), but recently, was expanded to include a member of the phylum

Firmicutes (114). This raises the possibility that the ice nucleation phenotype may be distributed in a broader phylogenetic range of bacteria than previously appreciated. We identified a number of phyla (Acidobacteria, Bacteriodetes, Chlorobi, Cyanobacteria, Firmicutes, Planctomycetes,

Proteobacteria, Spirochaetes, and Verrucomicrobia) that may contain unrecognized lineages of ice nucleating bacteria, providing a motivation for future studies to explore the presence of this phenotype in these taxa. Improved understanding of the diversity of ice nucleating bacteria and their atmospheric sources, transport, and role in hydrologic cycling would lend valuable insight on their capacity to disperse aerially and participate in landscape-atmospheric feedbacks that have meteorological consequences (400).

Figure 2-1. Location and extent of the source ecoregions relevant in this study. These ecoregions of North America were defined by the EPA, USGS, and CEC, as described in the text. The source ecoregion was used to classify interaction patterns between backwards air mass trajectories and the mixed boundary layer for each precipitation event (Sif. S1). Location of the primary sampling site (Baton Rouge, LA) is indicated by the red star. EA, eastern Asia; PM, Pacific maritime; NFM, northwest forested mountains; HNL, high northern latitudes, DSAH, desert and semiarid highlands; DP, Great Plains; EWW, eastern woodlands and wetlands; NAM, North Atlantic maritime; SAM, South Atlantic maritime.

Figure 2-2. Concentrations of total, biological, and bacterial ice-nucleating particles (INPs) at activation temperatures of –4 to –15°C. The median and 95% conﬁdence interval (CI) values are depicted by the black horizontal lines and dashed vertical lines, respectively; boxes represent the interquartile range; whiskers represent maximum and minimum values, excluding outliers; and open circles represent outliers. (a) Differential total INP concentrations, with sample sizes as follows: -4 to -12°C, n=61; -13°C, n=56; -14°C, n=48; and -15°C, n=44. (b) Differential biological INP concentrations, with sample sizes as follows: -4 to-12°C, n=61; -13°C, n=56; -14°C, n=48; and -15°C, n=44. (c) Differential bacterial INP concentrations, with sample sizes as follows: -4 to-12°C, n=54; -13°C, n=44; -14°C, n=37; and -15°C, n=34. (d)

Cumulative concentrations of INPs active at temperatures of ≥-15°C for total, biological, and bacterial INPs.

Table 2-1. Average differential concentrations of total, biological, and bacterial INPs % of % of Temp. of total INP (L- biological INP bacterial INP Total Total Activation 1) (L-1) (L-1) INPa INPa -4°C 2 ± 1 2 ± 1 100% 2 ± 1 100% -5°C 11 ± 6 11 ± 6 100% 10 ± 5 91% -6°C 61 ± 11 60 ± 11 98% 35 ± 6 57% -7°C 174 ± 41 169 ± 41 97% 108 ± 31 62% -8°C 231 ± 64 223 ± 64 97% 118 ± 42 51% -9°C 349 ± 103 332 ± 101 95% 160 ± 72 46% -10°C 520 ± 161 460 ± 155 88% 211 ± 54 41% -11°C 1030 ± 156 679 ± 121 66% 523 ± 119 51% -12°C 2700 ± 341 1530 ± 227 57% 941 ± 153 35% -13°C 5200 ± 464 2890 ± 326 56% 2060 ± 277 40% -14°C 5650 ± 496 2800 ± 359 50% 2830 ± 480 50% -15°C 2090 ± 190 724 ± 119 35% 1020 ± 218 49% Cumulative 16200 ± 954 9020 ± 773 56% 6620 ± 660 41% Concentrations are per L of precipitation Total IN: -4 to -12°C, N=61; -13°C, N=56; -14°C, N=48; -15°C, N=44 Biological IN: -4 to -12°C, N=61; -13°C, N=56; -14°C, N=48; -15°C, N=44 Bacterial IN: -4 to -12°C, N=54; -13°C, N=44; -14°C, N=37; -15°C, N=34 a Estimated percentages of total IN based on mean concentrations

Table 2-2. Descriptive statistics of numerical variables measured in this study. Total, biological, and bacterial INPs (INPs L-1 precipitation); cells (cells L-1 precipitation); conductivity (µS cm-1); DOC (ppm); chloride, nitrate, sulfate, sodium (µM). - + statisti total biolo bacteri cells pH condu DO Cl NO3 SO4 Na ca INPsb gical al ctivit C - 2- INPsb INPsc y n 65 65 56 63 56 55 47 32 32 32 32 Mean 16200 9020 6620 6.41E+ 6.3 15.4 1.59 108 23.1 13 162 05 9 SD 6349 5108 3628 9.54E+ 0.8 8.95 1.12 104 14.1 7.8 101 05 3 Upper 16232 9271 7154 8.81E+ 6.6 17.82 1.92 145 28.2 17.8 198 95% 05 1 CI Lower 16133 9192 7094 4.00E+ 6.1 12.98 1.26 70.1 18.0 12.2 125 95% 05 7 CI Media 17268 8914 6689 225012 6.4 14.3 1.32 65.4 18.1 12.3 143 n 5 Varian 40316 2608 131596 9.10E+ 0.6 80.15 1.27 1083 199 60 101 ce 117 6560 23 11 9 1 92 IQR 11911 8560 4805 549505 1.0 12.42 1.37 157 17.8 6.5 145 8 aCI, confidence interval; IQR, interquartile range; SD, standard deviation bCumulative concentration at -15oC

Table 2-3. Results of exploratory factor analysis for total, biological, and bacterial INPsa Temperatu re of total INPs biological INPs bacterial INPs Activation Factor 1: Factor Factor Factor Factor 3: Factor 1: Factor 2: Total−5 to 2: 1: 2: Bio−11 to Bac−5 to Bac−4 and −11 Total−11 Bio−5 to Bio−13 to −12 −10 −12 to −14 −10 −14 -4°C ------0.69 -5°C 0.51 - 0.50 - - 0.57 - -6°C 0.77 - 0.80 - - 0.69 - -7°C 0.79 - 0.76 - - 0.72 - -8°C 0.91 - 0.91 - - 0.70 - -9°C 0.79 - 0.76 - - 0.52 - -10°C 0.75 - 0.67 - - 0.65 - -11°C 0.60 0.60 - - 0.68 - - -12°C - 0.72 - - 0.78 - 0.54 -13°C - 0.85 - 0.87 - - - -14°C - 0.75 - 0.92 - - -0.56 -15°C ------Percent of 35.4% 21.7% 32.7% 15.4% 13.2% 30.2% 13.8% variance aFactors at each temperature were interpreted by examining the factor loadings of each variable. Variable loadings of ≥ |0.50| (i.e., accounts for ≥ 25% of overlap between variable and factor variance) were retained. bIndicates the proportion of variance each factor accounts for in the INP data

Figure 2-3. Results of ﬂuorescent dissolved organic matter (DOM) excitation-emission matrices data investigated using parallel factor (PARAFAC) analysis. (a to c) The three ﬂuorescent DOM PARAFAC components.

Table 2-4. Correlations of INP factors with physical and chemical measurements of the precipitation, shown as Pearson correlations coefficients calculated between factors of EFA and measured variables of precipitation Fluorescent Cell organic abundanc intensity (RU) e Concn (μM) of: DOC by componentb -1 2 INP (L Conductivit NO3 SO4 (ppm Factor precip) pH y (μS cm-1) Cl- Na+ - - ) C1 C2 C3 - - total−5 .45* .39 .48 .26 .41 -.34 -.03 .29 -.01 .23 .0 to −11 * * * 1 - total−1 .25 .19 .23 .07 .07 .17 .23 -.002 .27 -.01 .0 1 to −14 8 - bio−5 to .47 .43* .35 .20 -.35 -.31 -.08 .30 -.01 .14 .0 −10 * 7 bio−13 .0 .06 -0.1 .27 .25 .34 .26 .30 .24 .30 .01 to −14 3 bio−11 .39 .40 .1 .29 .25 .22 -.20 -.21 .18 .21 .14 to −12 * * 5 - - bac−5 .46 .54** .29 .17 -.37 .42 -.18 .20 -.07 .20 .0 to −10 * * 8 aSignificance levels of Pearson correlation coefficients: *p < .05, **p < .01, ***p < .001. The number of samples of each were the following: cell abundance, n=61; pH and conductivity, n=52; ionic concentrations, n=31; DOC and C1 to C3, n=43. bRU, Raman units

Figure 2-4. Average INP factor concentrations as a function of ecoregion classiﬁcation. Multivariate analysis of variance indicated that average INP concentrations differed based on the interactions of air masses with the various ecoregions. This plot is showing the average of summed differential INP concentrations (y axis) for each INP factor for each ecoregion. For example, bac (–5 to –10) is the average value of the summed differential INP concentrations at –5, –6, –7, –8, –9, and –10°C for each ecoregion

Figure 2-5. Average INP factor concentrations as a function of cloud type, season, and precipitation phase. (a to c) Multivariate analysis of variance indicated that average INP concentrations (Conc.) differed based on the cloud type (a), season (b), and precipitation type (c). The average of summed differential (Diff.) INP concentrations were determined as described in the Fig. 3 legend.

Figure 2-6. Summary of trends in ice-nucleating particle (INP) data with source ecoregion and physical, chemical, and microbiological properties of the precipitation. Ecoregions as depicted in Fig. 2-1 are outlined by black solid lines. (a) Key for INP concentrations in the ecoregion for panels b to g and Pearson correlation coefficents for panel h are shown. (b to g) Ecoregions that correlated significantly with total–5 to–11 (b), total–11 to – 14 (c), bio–5 to–10 (d), bio–11 to –12 (e), bio–13 to –14 (f), and bac–5 to–10 (g) INP concentrations are shown.(h) Heatplot of Pearson correlation coefficients between INP concentrations and various precipitation measurements, as follows (units): cell abundance (cells liter-1 precipitation); conductivity (cond. μS cm-1); chloride, sodium, nitrate, and sulfate (μM); DOC, dissolved organic carbon (parts per million); C1, PARAFAC component C1 (Raman units [R.U.]); C2, PARAFAC component C2 (A.U.); and C3, PARAFAC component C3 (A.U.).

CHAPTER 3 INDUCTION OF ICE NUCLEATION ACTIVITY IN NOVEL ICE+ BACTERIA

Overview

Bacterial INPs are active at temperatures warmer (> −10°C) than most mineral aerosols

(10, 19, 20) and their presence in the atmosphere could have a significant effect on meteorological processes (1, 14, 15, 21-24). However, the diversity of bacteria that possess the ability to nucleate ice (a phenotype referred to as Ice+) is not well understood, nor are the

environments that serve as sources of Ice+ bacteria to the atmosphere. As discussed in Chapter 1,

the known mechanism of bacterial ice nucleation is conveyed by a homologous protein (360),

and no other sequenced bacterial genomes have been found to contain the ina gene. Most known

Ice+ bacteria are members of the Gammaproteobacteria, and species from other phyla may

possess novel IN proteins or produce non-proteinaceous IN-active macromolecules, as has been demonstrated in several species of pollen (22, 136, 268). Indeed, recent work has shown IN activity in isolates of gram-positive Lysinibacillus (114) are able to nucleate ice using a non- proteinaceous molecule.

The lack of information on the diversity of IN-active microorganisms prohibits a full understanding of their ecological sources and function, influence on meteorology and climate, role in frost damage to crops, and their potential applications for bioprospecting (76). Adding to the challenge in studying IN-active bacteria is the fact that an estimated 99% of microbes are uncultivable using standard methods, because identification of the Ice+ phenotype has required

cultivation under specific conditions (103, 107, 109). Therefore, it is highly likely that the

diversity of IN-active bacteria exceeds that which have been detected in characterized bacteria to

100

date. If expression of the Ice+ phenotype is similar to that in P. syringae, specific environmental and nutrient conditions may be required to promote the induction of this activity (103).

Various studies have concluded that certain plant and soil environments are significant sources of highly active biological INPs to the atmosphere (25-28). However, results from chapter 2 show that air masses which had previously interacted with the mixed boundary layer

(MBL) in continental US regions replete with agricultural, woodland, and herbaceous ecosystems did not produce precipitation with the highest concentrations of warm-temperature

INPs (active -4 to -10oC). In fact, the highest concentrations of warm-temperature INPs were sourced from high-latitude North American terrestrial regions (Figure 2-1). Additionally, previous studies have indicated that desert dusts from Asia are significant sources of INPs to the troposphere over North America (29), and that long distance transport of viable microorganisms is possible (29-32). Thus, the Asian continent and high northern latitudes may be significant sources of biological INPs to the atmosphere. Members of the Actinobacteria, and in particular, those in the genus Hymenobacter (Supplementary Table S6), were highly positively correlated with INP concentrations and were present in samples with the highest documented INP concentrations. Taken together, these results suggest that arid soils may harbor undiscovered IN- active bacteria. Indeed, a large fraction of species from the bacterial genus Hymenobacter have previously been isolated from and/or detected in desert soils (33-37), Saharan aeolian deposits in snowpack from high-altitude mountain tops (45, 46), as well as in air samples (270). Thus, the objective of this study was to use a selective enrichment approach to isolate and characterize strains of Hymenobacter from precipitation and arid soil samples to test their IN activities. The effect of incubation conditions and nutrient limitation on the IN activity of isolates of

Hymenobacter and other UVC-tolerant bacteria are presented and discussed.

101

Methods

Sample Collection

Soil samples were collected at various locations in Colorado and Nevada, USA (Table 3-

1), placed in sterile 4 oz specimen containers (Fisher Scientific, Pittsburgh, Pennsylvania) or 15 mL conical tubes (Corning, New York City, New York), and stored at room temperature until processing (~3-7 days after collection). Sleet samples were collected in four 120-liter galvanized cans that were lined with clean and sterile 94 by 122-cm polypropylene bags (Fisher Scientific,

Pittsburg, PA). Sleet samples were collected outside of Chicago, Illinois, at ground level in an open park (41.898oN, 87.9536oW) on December 27, 2015, from 4am until 11am. Sleet samples were stored on dry ice during overnight shipment, and subsequently placed at -80oC until processed (up to 1 month). Bacterial strains were obtained from the precipitation culture collection at Virginia Polytechnic Institute, which were isolated as described in Failor et al (114).

Bacterial Culturing

Soil slurries were created (5% w/v) using sterile 1X phosphate buffered saline (PBS) and vortexing for 15 min on medium speed. After suspension, large soil particles were allowed to settle and 100 μL of the supernatant was spread plated in triplicate onto culture media. Sleet samples were removed from storage at -80oC and allowed to thaw at room temperature. Once thawed, 1 L of the melted sleet was concentrated onto sterile 0.05 μm Nucleopore filters (Fisher

Scientific, Pittsburg, Pennsylvania). The retentate was reconstituted in 5 mL of sterile 1X PBS by vortexing the filter on low speed for 10 min and 100 μL of the retentate was plated in triplicate onto solid media. The media used for these enrichments included Tryptic Soy Agar

(TSA), Reasoner’s 2 Agar (R2A), Nutrient Agar (NA), and Marine Agar (MA) (Appendix B).

102

Within a laminar flow hood, the inoculated media was subsequently exposed to UVC

(254 nm) for 1 to 9 min (199.8 to 1798.2 W m-2). The lamp was a distance of 64.5 cm from the

exposure surface and a flux of 3.33 W m-2 was measured using a model 3UV-38 lamp (Part no.:

95-0341-01, UVP LLC). Following UVC exposure, the lids were placed on each plate and they

were incubated upside down in the dark at the following temperatures: 37oC for the TSA plates;

4oC, 18oC, and 25oC for the R2A plates; 18oC, 25oC, and 37oC for the NA plates; and 18oC and

25oC for the MA plates. Following incubation and colony growth on the media, pink- and red-

pigmented bacterial colonies with morphologies resembling that typical for Hymenobacter (366,

369, 372, 383, 384) were picked and isolated in pure culture using a standard three-phase

streaking technique. Bacterial cultures from the Virginia Tech precipitation culture collection

were provided on R2A slants; they were sub-cultured and demonstrated pure by isolation on

R2A using a standard three-phase streaking technique.

Individual isolates were grown aerobically at 30oC in liquid R2A with shaking at 250 rpm

until the populations reached late exponential phase. The cultures were harvested and

resuspended into sterile 1X PBS, and then aliquoted into nutrient deprived media (Tables 3-2 and

3-3). For the R2A nutrient-limited variations (Table 3-2), “R2A-3N” refers to the nitrogen

deprived media; “R2A-3C” refers to the carbon deprived media; “R2A-CN” refers to the carbon

and nitrogen deprived media; and R2A refers to the complete medium. R2A-3N was prepared with 10% of the normal amount of nitrogen-containing constituents, which included proteose peptone, casamino acids, and yeast extract (Table 3-1). R2A-3C media was prepared with 10% of the normal amount of carbon constituents, which included dextrose, soluble starch, and sodium pyruvate (Table 3-2). R2A-CN was prepared with 10% of the normal amounts of both the carbon and nitrogen constituents (Table 3-2). The same system of nomenclature was applied

103

to the minimal media variations, which were designated as “M9”, “M9-C”, and “M9-N” (Table

3-3). The M9-C media was prepared with 10% the amount of the carbon source dextrose, and the

M9-N media was prepared with 10% the amount of nitrogen source ammonium chloride (Table

3-3). All variations on minimal media were supplemented with a vitamin stock (Appendix B).

Glycerol stocks of the isolates were prepared by growing cultures aerobically at 30oC

overnight with shaking at 350 rpm. An aliquot of 500 μL from the overnight culture was combined with 500 μL of a sterile 50% (v/v) glycerol solution into 1.5 mL cryovial tubes (Fisher

Scientific, Hampton, NH, USA). The cryovials were then flash frozen in liquid nitrogen and stored at -80oC.

Immersion Freezing Assays for Ice Nucleation Activity

Pure cultures of the isolates were screened for IN activity by scraping a loopful of the

colony from the agar using sterile inoculating loops (Fisher Scientific, Hampton, NH, USA) and

suspending the material into 1 mL of 1X PBS in sterile 1.5 μL microcentrifuge tubes (Fisher

Scientific, Hampton, NH, USA). Samples from all the isolates were prepared in triplicate. The

colonies were suspended by vortexing the microcentrifuge tubes for 5-10 seconds and the

samples were subsequently placed in a refrigerated circulating ethylene glycol bath (Thermo

Fisher Scientific, Waltham, MA, USA). Freezing of suspensions was observed during decrease of the temperature at 1oC intervals from -4oC to -13oC. Potential Ice+ bacterial isolates were

considered to be those in which two of the three tubes froze at -13oC and warmer. Isolates

demonstrating IN at temperatures ≥ -13oC were subjected to further analysis, as described

below.

Bacterial isolates demonstrating IN activity were used in nutrient limitation experiments

(described below) and INP concentrations were quantified using an immersion freezing assay.

104

One hundred μL of the culture was added to 3.5 mL of 1X PBS and 35 μL of fluorescein (50

ppm), and vortexed to mix. Solutions were then poured into 5 mL sterile reagent basins (Fisher

Scientific, Hampton, NH, USA) and 100 μL was placed into each well of a 96-well plate

(Thermo Fisher Scientific, Waltham, MA, USA) in triplicate, covered with adhesive films

(Thermo Fisher Scientific, Waltham, MA, USA), and the sides secured with tape. The plates were floated in the refrigerated ethylene glycol bath and cooled in 1oC intervals from -4oC to -

13oC. The number of wells that froze at each interval were recorded. The number of ice

nucleating particles (INPs) were estimated according to the procedure of Vali et al (338).

In several experiments, there were additional treatments done prior to the immersion freezing assays. Heat treatment consisted of heating cell cultures in a water bath at 95oC for 10

min. Lysozyme- and proteinase K-treated cells were incubated with 18 mg mL-1 of lysozyme

(Fisher) or 100 μg mL-1 of proteinase K (Sigma), respectively, for 1 h at 25oC. To inhibit protein

and RNA synthesis, cells were incubated with 50 μg mL-1 of chloramphenicol and 50 μg mL-1 rifampicin, respectively, for 1 h at 25oC. Cultures were centrifuged at 3000xg for 5 min to test IN activity associated with the supernatant and cells; the pellet by resuspending to the original

concentration in 1X PBS and the supernatant was collected and filtered through a 0.2 μm syringe filter for the analysis.

Nutrient Limitation Experiments

Fresh cultures of the bacterial isolates were grown from glycerol stocks on R2A. Isolated

colonies were picked from the agar plates and cultured in 15 mL of R2A broth. Cultures were

incubated at 30oC with shaking in sterile 50 mL conical tubes, glass culture tubes, or flasks with

aeration at 250 rpm and. Cells were harvested from the late exponential phase by centrifugation

at 3000 x g for 10 min and subsequently rinsed three times with 1X PBS. The washed cell pellets

105

were resuspended in 15 mL of 1X PBS and aliquoted into various nutrient deprived media described in Tables 3-2 and 3-3. For each experimental media, 500 μL of the washed cell suspension was added to 500 μL of 2X concentrations of the respective media (Tables 3-2 and 3-

3) in 1.5 mL microcentrifuge tubes. The cultures were incubated at 4oC, 25oC, and 30oC for a

duration of up to 5 weeks. Each isolate and condition was tested in triplicate.

The samples were tested for IN activity at approximate weekly intervals by conducting

the immersion freezing assay protocol in 96-well plates (described above). At each experimental time point, the optical densities at 600 nm were recorded for each culture. In all experiments,

Ice+ Pseudomonas syringae DC3000 and Escherichia coli were used as positive and negative

controls, respectively. Blanks consisting of 1X PBS and 2X media were used as procedural

controls and tested identically to the experimental treatments.

Identification of the Bacterial Isolates

Genomic DNA was extracted from cell pellets harvested from the isolates using the lysis and phenol-chloroform extraction method of Sambrook et al. (401). The 16S rRNA gene was

PCR-amplified from the DNA using 27f and 1492r universal primers (402) and, the size of the

amplicons were evaluated by agarose gel electrophoresis, and purified using the MP Biomedicals

PCR Purification kit (Irvine, CA, USA). The PCR products were bidirectionally sequenced at

Eton Bioscience Inc. (San Diego, California, USA) using Sanger Sequencing. The nucleotide

sequences obtained were analyzed using MEGA X software. Sequences were trimmed based on

the quality of chromatograms, and reverse and forward reads were overlapped to obtain a contig

of the 16S rRNA gene that ranged from 577 to 1297 bp in size. Sequences were aligned by

secondary structures using the SILVA database, and pairwise distance matrices were constructed

in MEGA X to identify their nearest neighbors.

106

Results

Primary Enrichments of Arid Topsoil and Sleet Samples

Plates exposed to UVC for less than 2 min (399.6 W m-2) led to lawn growth and too many colonies to count. At UVC exposures greater than 2 min, fewer colonies were observed. In general, less than 10 colonies grew on the Chicago sleet plates, and the colony morphologies and pigments were not very diverse. On the other hand, plates grown from the arid topsoil samples had more CFUs than the Chicago spread plates, and a wider variety of bacterial growth of both pigmented and non-pigmented colonies. Samples taken from the continental divide in Colorado,

Hoosier Pass and Monarch Pass (approximately 4000 and 3500 mAGL, respectively), had the

largest number of colonies at higher UV fluxes (>999 W m-2), whereas samples taken from lower

altitude sites in Colorado had less growth after 5 min of UVC irradiation (999 W m-2), and fewer

pigmented colonies overall. Similar to the high elevation Colorado sample sites, samples

collected from Mojave Desert topsoil in Nevada also produced more colonies at longer UV-

exposures (999 to 1798.2 W m-2), as well as more pigmented colonies overall. All of the cultures

were examined for sticky, pink or red-pigmented colonies, which are colony morphological

characteristics of Hymenobacter species (37), and these colonies were picked and streaked for

isolation. From each plate, several non-pigmented colonies were also selected for isolation.

Individual isolates (~200) were grown on the same media and temperature of isolation.

Dependence of Ice Nucleation Activity on Culture Age

Fresh cultures of the isolates were screened for IN activity immediately following

isolations. A loopful of colonies were scraped from the agar and screening for initial IN activity

as described in the methods. Of the ~200 isolates initially screened, the vast majority remained

unfrozen above -10oC, with freezing occurring in one out of the three replicates for a few of the

107

isolates. Based on these results, it was concluded that none of the bacteria tested had IN activity under the conditions tested.

The isolates were stored at 4oC for approximately 4-6 months before being IN activity

was tested again. At this point, several of the isolates demonstrated IN activity, with at least two

out of three triplicate microcentrifuge tubes frozen at temperatures as warm as than -5oC (Figure

3-1). After confirmatory experiments, the overall trend after 6-9 months of incubation on R2A at

4oC was an increase in IN activity in 7 of the ~200 isolates tested. The highest IN activity

observed was for isolates AZ_82Pink (Figure 3-1) and AZ_82Red, which were co-isolated from

a Mojave desert soil sample. Two bacterial cultures isolated from Colorado soil samples, CO_1

and CO_2, demonstrated IN activity at temperatures as warm as -6oC after 6-9 months, however the activity of CO_5 was highly variable with time (Figure 3-1). Three bacterial isolates from mixed-phase precipitation collected in Chicago, IL showed IN activity after approximately 2-6 months of cold storage, with the warmest temperature IN activity displayed by IL_9 at -7oC

(Figure 3-1).

Amplification and sequencing of the 16S rRNA gene from the isolates demonstrating IN-

activity revealed that most were Actinobacteria, with representatives of the Firmicutes,

Deinococcus-Thermus, and Alphaproteobacteria also identified (Table 3-4). The only sample

location from Colorado that IN-active bacteria were isolated was from Hoosier Pass, located on

the continental divide at approximately 4000 mAGL. The remainder of the IN active strains

were isolated from samples of Chicago sleet or Mojave Desert soil. Most of the bacteria stained

gram-positive and were cocci, but variable gram-staining characteristics and pleomorphic cell

morphologies were also observed (Figure 3-2c). Only four of the isolates stained gram-negative,

all of which were members of the Proteobacteria (Table 3-3).

108

The isolates that had IN activity after 4 to 6 months at 4oC were subsequently screened

every month, and the majority were shown to steadily increase the temperatures at which they

induced ice nucleation (Figure 3-1). Many of the other isolates listed in Table 3-4 also showed

IN activity after incubations in at 4oC for several months but were difficult to cultivate to

sufficient cell densities in liquid, and were not able to be induced using nutrient deprivation.

Therefore, these isolates were not investigated further in this study (Table 3-3). Fifteen isolates

were obtained from a culture collection maintained by collaborators at Virginia Polytechnic

Institute and were also screened for IN activity (VA isolates, Table 3-3). Preliminary screening

results indicated that a number of these isolates had IN activity after 6 weeks of incubation in

nutrient deprived media (Table 3-4). Because all isolates which demonstrated freezing froze

completely at -4oC, the number of wells on the 96-well plate that had frozen are listed here, as opposed to the number of INPs. Table 3-5 shows IN activity after 6 weeks of incubation in the various nutrient deprived media and temperatures. Two of these Virginia isolates, VA_2512

(Kocuria sp.) and VA_2470 (Paenibacillus sp.), were related to two potentially IN-active species isolated in this study from Mojave desert soils (Isolate AZ_82Pink, a Kocuria sp.; and Isolate

AZ_8, a Paenibacillus sp.). Several other Virginia isolates appeared to possess IN activity at -

4oC as well (Table 3-4). The two other bacterial cultures isolated from Mojave desert soil,

AZ_122 (Arthrobacter sp.) and AZ_8 (Paenibacillus sp.), appeared to exhibit IN activity (Figure

B-1 and B-2), although to a lesser degree than that of AZ_82Pink and AZ_82Red.

The isolates with the warmest temperature IN activity were AZ_82Pink (-4oC) and

AZ_82Red (-5oC), which were co-isolated from Mojave Desert soil samples. Both were

examined more extensively and were the primary focus for specific experiments performed in

this study. According to phylogenetic analysis of the 16S rRNA gene, AZ_82Pink was a member

109

of the phylum Actinobacteria within the genus Kocuria, most closely related to the type species

Kocuria polaris (Table 3-3). AZ_82Red belonged to the Actinobacterial phylum and the

Arthrobacter genus, and appeared most closely related to Arthrobacter agilis (Table 3-3). Gram-

staining indicated that the Kocuria sp. was a gram-positive coccus, approximately 0.5 μm in

diameter (Figure 3-2a), whereas the Arthrobacter sp. was gram-variable and pleomorphic,

presenting as a gram-negative rod at early exponential phase and a gram-positive coccus at late

exponential-stationary phase (Figure 3-2b).

Induction of IN Activity in Isolates AZ_82Pink and AZ_82Red

A series of experiments were performed on the two most IN active bacterial isolates from

Mojave desert soils (AZ_82Pink and AZ_82Red) to better understand the conditions that induce their IN activity. In an attempt to accelerate conditions experienced by the bacterial isolates during long term incubation at 4oC, experiments were conducted in which the media

formulations deprived the isolates of specific C/N nutrients and at different temperatures

(Figures 3-3 and 3-4). While both isolates were able to grow in M9 media, when supplemented

with a vitamin stock, neither isolate had IN activity induced when using variations based on M9

minimal media (Table 3-3). However, variations in R2A media compositions (Table 3-2) did show an effect on IN activity in repeat experiments (Fix. X). For the Kocuria isolate AZ_82Pink, the most significant influence on its IN activity was nitrogen-limited conditions, with the highest

fraction of Ice+ cells in the populations observed after 14-21 days of incubation at 4oC (Figure 3-

3). At its maximum activity, AZ_82Pink displayed a nucleation frequency (INPs per cell) of

approximately 1 INP per 100,000 CFU at -5oC. The number of INPs per CFU were significantly

higher in N-limited R2A when compared to the other conditions, with all of the wells frozen

(n=32) before the assay reached a temperature of -6oC. However, cell suspensions incubated in

110

full strength and C-limited R2A also had increased IN activity at 4oC after 14-21 days. Further, the temperature spectrum at which IN activity was observed in the 4oC incubations, for all

media, was much warmer than the temperature spectrum for the 25oC and 30oC incubations.

Incubations at 25oC in R2A and C-limiting conditions showed increases in IN activity, however, not as high as those observed at 4oC and with lower warm temperature (>-8oC)

nucleation frequencies of ~ 1.00-7 INPs per CFU. Under nitrogen-limitation for the 25oC incubations, there were not significantly higher ice nucleation frequencies. In addition, the carbon-limited media appeared to have a larger influence on ice nucleation frequencies at warmer temperatures than the nitrogen-limited media (Figure 3-3). Overall, carbon- and nitrogen-limitation at 25oC had a more significant effect on IN activity than standard R2A.

Incubations at 30oC in R2A had no effect on IN activity in AZ_82Pink (Figure 3-3), but this was

not the case under carbon- and nitrogen-limitation. After 7 to 21 days, the carbon-limitation led

to higher ice nucleation frequencies, whereas after 28 days, nitrogen-limitation resulted in higher

ice nucleation frequencies. After incubation for 28 days, the IN activity was lower and only

detected at temperatures of -10oC and colder.

The Arthrobacter isolate AZ_82Red did not grow to high densities in the liquid media, with ODs were constantly low throughout the duration of experiments (on average, Absorbance at 600nm was <0.01). Nevertheless, IN activity was demonstrated to be dependent on nutrient- limitation and temperature for this isolate (Figure 3-4). For all experiments, at 0 days, there was no IN activity. After 11 days, INP concentrations of ~103 per L were detected at -12oC for R2A

and the nitrogen-limited conditions. After 35 days, only populations in R2A showed IN activity

(~103 INPs L-1) at -8oC, increasing to >104 L-1 at -9oC and all samples frozen by -10oC. Unlike

AZ_82Pink, the incubation temperature for AZ_82Red did not appear to be influence IN activity.

111

In samples incubated for 11 d, IN activity at -8oC was detected, with concentrations of 102 to 103

L-1 observed at -13oC. After 35 d, INPs at -5oC were 102 INPs L-1, and up to 104 INPs L-1 by -

13oC. Similar to AZ_82Pink, the combination of R2A and 30oC incubation led to no production

of detectable INPs in AZ_82Red. On the other hand, after 11 d, the carbon-limited media saw a

significant increase in INP production, with 103 INPs L-1 produced at -4oC and 104 INPs

produced by -13oC. At 13 d, the nitrogen-limited media did not have as much of an effect

compared to normal R2A, but still led to the production of 102 INPs L-1 by -10oC. After 13 d, the

nitrogen-limited media surpassed the carbon-limited media in INP concentrations, with 103 INPs

L-1 produced at -6oC and 104 INPs L-1 produced by -13oC.

Characterization of the Bacterial INP in AZ_82Pink

For AZ_82Pink, IN activity was decreased after heat treatment at almost all temperatures,

for all incubations and media compositions (Figure 3-5), which is consistent with a proteinaceous

source for the activity. Additionally, the 0.22 μm filtrate had no detectable INPs, indicating that

the IN molecule was intracellular or associated with the cell. IN activity in the supernatant of

AZ_82Pink cultures, following centrifugation, showed mixed results (Figure 3-5). For the 4oC

incubations, the supernatant, which was assumed to be the “cell-free” INP fraction, was statistically the same as the “cell-associated” INP fraction for the carbon-limited media. For the nitrogen-limited media, the “cell-free” INPs made up a majority of the INPs at -4oC, but

significantly less of the total INP fraction at all other temperatures. Unlike the nutrient-limited

media, the IN activity in the supernatant of the normal R2A media showed much more

significant differences, with the “cell-associated” INPs making up significantly more of the total

INPs than the “cell-free” INPs (Figure 3-5).

112

For the 25oC incubations, the carbon-limited media had significantly higher amounts of

“cell-free” INPs at the warmer end of the temperature spectrum (-9 to -10oC), but at the colder

temperatures (-11 to -13oC), the “cell-associated” INPs were higher in concentration. For the

nitrogen-limited and normal R2A media, the “cell-associated” INPs made up a larger fraction of the total INPs. For the 30oC incubations, all types of media had significantly lower

concentrations of “cell-free” INPs as compared to the “cell-associated INPs.”

Although the results of these experiments were successfully replicated three times,

subsequent trials under the same presumed conditions lead to much lower induction of IN

activity than initially observed for AZ_82Pink and AZ_82Red. To investigate the influence of

culturing conditions for the initial populations used in these experiments, the isolates were grown

using several different approaches. In initial experiments where induction of IN activity was

successful, cultures of AZ_82Pink and AZ_82Red were grown by shaking 15 mL of media in

capped 50 mL conical tubes at 250 rpm and 30oC. Because inadequate aeration and the presence of oxygen may affect the experiments, the cultures were grown in 50mL conical tubes (as

previously), 15 mL glass culture tubes, and 100 mL Erlenmeyer flasks at 300 rpm (the latter two

with vented caps) for comparison. IN activity did not differ significantly between cultures which

had been grown in 50mL conical tubes, versus glass culture tubes, versus glass aeration flasks.

Induction of IN activity was observed, to a lesser degree than previous experiments, within the

first week of subsequent nutrient deprivation experiments (Figure B-3). However, this IN activity gradually decreased with each day, until it was undetectable after 7 days (Figure B-2).

Discussion

The known Ice+ bacteria are primarily Gammaproteobacteria (81, 95, 96, 106, 107, 403),

with one example in the Firmicutes (114). The conserved primary structure of the Ina protein

113

(89, 404) coupled to the fact that most Ice+ strains are phytopathogens supports the hypothesis that the phenotype evolved to promote bacterial pathogenicity. Indeed, frost damage would facilitate intrusion of pathogenic bacteria into healthy plant tissues. In fact, several experiments

have confirmed that Ice+ strains of P. syringae have increased pathogenicity when compared to

their Ice- counterparts (83, 119).

For microbes that are not phytopathogenic, the advantage of the Ice+ phenotype requires

an alternative explanation. For example, two species of Lysinibacillus, neither of which are

known to be plant-associated, have demonstrated IN activity via a nanometer-sized, non-

proteinaceous molecule; a fundamentally different mechanism than that described in the Ice+

Gammaproteobacteria (114). Further, there exist lichen, fungi, and plants which possess non-

proteinaceous IN activity (22, 129–131, 136, 139). Data from chapter 2 indicated that the

concentrations of biological INPs did not correlate with the number of reads from known Ice+

genera, including Pseudomonas, suggesting that other bacteria were responsible for the

biological INP data observed. As such, we hypothesized that novel species of IN-active bacteria have the capacity to nucleate ice by mechanisms different from those known in certain

Gammaproteobacteria. The results of this study are the first to demonstrate IN activity within

members of the Actinobacteria.

Precipitation from air masses that interacted with the surface in high northern latitudes

and the Asian continent contained significantly higher concentrations of biological INPs than

from the other locations examined (Chapter 2). This combined with the large quantities of

microorganisms aerosolized from arid topsoil globally (165, 170, 223), motivated efforts to

examine locations with arid topsoil that could represent material that is easily windblown.

Significant correlations between biological INP concentrations and Hymenobacter OTUs

114

(chapter 2), led us to design a selective strategy based on UVC tolerance and shown previously

to be successful for the enrichment of Hymenobacter species (370, 371). While a species of

Hymenobacter was successfully isolated, it did not appear to possesses IN activity. However, in

the attempt to isolate Hymenobacter species, several other UVC-tolerant bacterial species

cultured fortuitously possesses the Ice+ phenotype under certain conditions. The results of this

study are the first to demonstrate IN activity within members of the Actinobacteria. (Table 3-3).

The bacterial isolates examined in this study were initially found to possess IN activity after

streak cultures were aged for several months at 4oC. Emphasis was placed on inducing IN

activity in AZ_82Pink and AZ_82Red because they nucleated ice at the warmest temperatures

(Table 3-3). Given the correlation between culture age and increase in IN activity (Figure 3-1),

we hypothesized that nutrient limitation in non-growing populations together with cold

temperature induces the IN activity, as has been observed in P. syringae (103) and E. herbicola

(109, 110).

Early experiments performed on P. syringae indicated that the frequency of IN-active

cells was most highest when cells were in stationary phase, and only detectable at concentrations above 106 CFU mL-1 (103). Additionally, our fresh cultures had no or very low IN activity. As

such, we grew our cultures to late exponential/early stationary phase, at concentrations of 109

CFU mL-1 to ensure enough biomass for INP detection—however, cultivation of AZ_82Red

proved difficult in R2A, and concentrations rarely got above 106 CFU mL-1. Nevertheless,

experiments to induce IN activity in AZ_82Red were still attempted. Initial results indicated that

when the fraction of CFUs that were INPs was considered, incubation at 4oC and nitrogen-

limitation for 14 to 21 days had the most pronounced effect on increasing IN activity in

AZ_82Pink, with freezing detected as warm as -4oC (Figure 3-3). In fact, for several of the IN

115

assays done with AZ_82Pink, the entire 96-well plate of triplicate measurement had frozen by -

5oC, indicating very high INP concentrations and warm temperature activity. A similar trend,

although to a lesser degree, was observed for AZ_82Red, except 25oC incubation as opposed to

4oC had the highest IN activity. Indeed, after 35 days of incubation in nitrogen-limited R2A,

AZ_82Red increased its INP production per CFU tenfold and showed IN activity at temperatures as warm as -5oC. Further, for AZ_82Pink the INP temperature spectrum shifted from colder temperatures to warmer temperatures as time in incubation increased, and as incubation

temperature decreased (Figure 3-4). While nutrient limitation had a common effect on IN

activity, different responses were found based on temperature.

The conditions that induce IN activity in AZ_82Pink and AZ_82Red make it likely that

this phenotype is linked to a physiological response to nutrient deprivation and cold

conditioning, as is the case for other Ice+ bacteria (103, 109, 110). It is interesting to note that the

induction of IN activity in AZ_82Pink was a factor of both cold-temperature incubation and

nutrient-deprivation. Incubation at 30oC in normal R2A, had no effect on IN activity, whereas

cultures grown in nutrient deprived media had increased IN activity. This supports the hypothesis

that nutrient limitation is a deciding factor in the induction of IN activity in this species—

however, this induction likely depends on a balance of other factors. Indeed, the same

experiments were carried out in minimal media, however no IN activity was observed, even

though cell concentrations were as high as those carried out in the R2A experiments.

Heat treatment of AZ_82Pink cells reduced IN, consisting with the an activity conveyed

by a proteinaceous compound. Additionally, IN activity was preferentially associated with the cell depending on whether it had been starved for carbon or nitrogen, and whether it had been incubated at 4oC, 25oC, and 30oC. Indeed, whereas IN activity was mostly cell-associated when

116

AZ_82Pink had incubated at 30oC for 14 days, the number of cell-free INPs appeared to contribute an equal amount to IN activity when cultures had been incubated at 4oC. These

variations in IN activity may be associated with the pleomorphic characteristics of this species.

This could indicate a change in the way the bacterial cell produces the INP in response to

varying environmental conditions. Indeed, previous studies have indicated that there may be

several different “classes” of bacterial IN proteins within the same species, which display

varying temperature activities and chemical makeups (111, 113). It is possible that the variations

in IN protein activity and chemical makeup are a result of post-translational modifications, which

could be a result of changes in transcriptional regulation due to bacterial stress responses (405).

However, the authors of this publication only speculate at post-translational modifications, and

thus it remains unknown whether IN proteins may undergo post-translational modifications.

Relatively few studies have investigated the molecular mechanisms which control the expression of IN proteins in bacteria. Changes in IN activity may reflect regulation at the level of

transcription within the cell. Indeed, Nemecek-Marshall et al (103) showed that inhibition of

cellular protein and RNA synthesis blocked the induction of IN activity, indicating that active

transcription and translation of the inaZ gene was necessary for IN activity. Induced IN activity

in E. herbicola cells was also shown to be blocked by inhibition of DNA metabolism, which

indicates that IN activity in that species is regulated at the transcriptional level (110). Given that

the IN protein in Gammaproteobacteria is relatively large (89), it is thus likely that it is

preferentially produced only when necessary – for example, when a cell experiences nutrient

limitation, it may be able to increase nutrient uptake by inciting frost damage to the plant on

which it resides. Such a phenomenon could explain the similar induction of IN activity in

AZ_82Pink. A search through published Kocuria genomes (n=28) indicated that at least to date,

117

none contain an annotated ina gene. Thus, it may be that the AZ_82Pink produces a protein that

is not phylogenetically related to genetic form found in Gammaproteobacteria.

The induction of IN activity following nutrient starvation and cold conditioning led to

experiments designed to determine if IN activity in AZ_82Pink and AZ_82Red is regulated at

the transcriptional or translational level. However, subsequent experiments in which we attempted to investigate transcriptional or translational regulation of the INP produced by using antibiotics were unsuccessful for technical reasons. The induction of IN activity observed in early trials was observed to be at lower frequencies in subsequent experiments. Initially, it was thought that growing the cultures in capped conical tubes, which may not allow for sufficient exchange of oxygen into the cultures, may have caused low oxygen or anaerobic conditions during growth. Since higher concentrations of biological INPs are found in aerobically decomposing vegetation when compared to anaerobically decomposing vegetation (78, 80), availability of oxygen might be important for expression of the Ice+ phenotype. We tested

cultures capped in 50 mL conical tubes with those in flasks that allowed vigorous aeration of the

cultures and observe no significant differences. In fact, we found that IN activity was observed

within the first week of culturing for AZ_82Pink and AZ_82Red in both capped conical tubes

and well aerated culture tubes and flasks. However, the number of INPs per cell were not as high

as they had been in previous experiments (Figure 3-3 and 3-4), and decreased within the span of

7 days, until it was undetectable by day 8. Monitoring of culture ODs and CFUs indicated that

this was not due to a loss in biomass. The culturing of AZ_82Pink and AZ_82Red populations in

sealed tubes during initial trials inadvertently affected IN activity in ways that could not be

repeated. Establishment of anoxic conditions in aged colonies and poorly ventilated culture

118

vessels may have led to our initial observations, and unexplainable losses of IN activity in bacteria have been reported in similar studies (114).

Isolates AZ_122 and AZ_8, species of Arthrobacter and Paenibacillus, respectively,

were also isolated from the Mojave Desert topsoil (Table 3-3) and initially appeared to possess

IN activity (Figure B-1 and B-2). Indeed, AZ_122 demonstrated constant IN activity from day 0

to day 21. Interestingly, the nutrient deprived media did not appear to influence the IN activity of

AZ_122 and different than standard R2A (Figure B-1), which was also observed for other

Arthrobacter species investigated in this study, AZ_82Red (Figure 3-4). Isolate AZ_8, a species

of Paenibacillus demonstrated IN activity only after several days of incubation, and also

appeared to not be influenced by media composition (Figure B-2). However, AZ_8 appeared to

lose its IN activity after only several days in the experiments, suggesting depletion of nutrients

did not control expression of the IN activity. Additionally, growth of the AZ_8 cultures in liquid

media was difficult, due to the bacterium’s tendency to aggregate, which complicated

homogeneous suspension and subsequent CFU and INP quantifications.

We obtained fifteen of these cultures to determine if nutrient deprivation and variations in

temperature incubations affected their IN activity (Table 3-4). Only twelve cultures were able to

be grown in the lab, and of these, only eight showed IN activity in preliminary screening (Table

3-4). However, freezing was only detected in one of the isolates, namely VA_2512, a species of

Kocuria, earlier than 6 weeks, albeit at temperatures colder than -9oC. It was not until after at

least 6 weeks of incubation at 4oC that many of the cultures showed IN activity at -4oC. Due to

the fact that the immersion freezing assay requires at least one well to remain unfrozen in order

to quantify the concentration of INPs in the sample, we were unable to estimate INP

concentrations as all of the wells had frozen at -4oC (Table 3-4).

119

There were a number of closely related bacteria observed from different environmental sources that displayed similar patterns for the induction of IN activity. Although we were unable to induce IN activity in isolate VA_2470, it was interesting to note that it was a species of

Paenibacillus, as we had separately isolated a different species of potentially IN-active species of

Paenibacillus from Mojave Desert soil. In fact, our dataset indicated a handful of recurring IN active species from the same genera: Two species of Kocuria, one isolated from Mojave desert soil (AZ_82Pink), and the other isolated from precipitation collected in Blacksburg, VA

(VA_2512); two species of Paenibacillus (AZ_8 and VA_2470), one isolated from Mojave desert soil and the other isolated from precipitation collected outside of Blacksburg, VA; and four species of Arthrobacter, two isolated a year apart from Mojave desert soil (AZ_82Red, and

AZ_122), one from the top of the continental divide at Hoosier Pass in Colorado (CO_1), and one from sleet collected from a nimbostratus cloud outside of Chicago, IL (IL_9). Of the four

Arthrobacter species isolated, three of them (AZ_82Red, AZ_122, and IL_9) were most closely related to Arthrobacter agilis. However, Arthrobacter species are common inhabitants of soil environments and readily cultured from environmental samples due to their nutritional and tolerance to stress, which may explain their repeated cultivation (406, 407). Nevertheless, the fact that they were chosen for investigation only after they exhibited the IN phenotype is significant.

The ability to induce IN activity under specific environmental and nutrient-limiting conditions indicates that this phenotype is not exclusive to members of the Gammaproteobacteria and Firmicutes. Indeed, the inherent difficulties associated with culturing environmental isolates in vitro, in addition to the challenges of inducing a phenotype of which we have very little understanding, complicates efforts to define novel IN-active species. Collaborators at Virginia

120

Polytechnic Institute had created a library of potentially IN-active bacterial isolates from precipitation samples that showed IN activity during initial trials, and then appeared to lose their activity. In fact, initial experiments carried out on IN-active P. syringae strains were generally only able to induce IN activity in 1 in a million cells—it wasn’t until researchers shifted nutrient concentrations and temperatures that they began to observe an increase in IN activity to every single cell in the P. syringae population. However, the data presented in this chapter should make clear that there are likely many other species of bacteria which can nucleate ice, and probably do so by means different from that of the Ice+ Gammaproteobacteria. Lastly, the detection of species from the same genera in both terrestrial environments and precipitation may be an indication of potential source and sink environments, further illustrating the ability of species with this phenotype to cycle between new environments by participating in meteorological processes.

Concluding Remarks

Our initial discovery of IN activity in old cultures that had been incubating at 4oC for several months ultimately led to the design of experiments which might expediate this induction phenomenon. The IN activity of several isolates was demonstrated, none of which have previously been recognized as IN-active. Further, the seemingly proteinaceous INP produced by

AZ_82Pink emphasizes the need for investigations into the possibility that microorganisms may possess a variety of IN proteins which differ from those of the Gammaproteobacteria. While the nutrient-limited experiments appeared to be successful for several of the isolates, there were many challenges associated with attempting to induce IN activity in a repeatable. Much time was spent on optimizing the experimental protocol and the testing of various conditions under which the potential IN-active bacteria should be grown. As such, more care needs to be taken when

121

investigating the IN activity of cultured bacteria, as their phenotype may not be displayed without implementing conditions such as nutrient limitation that induce it. The results presented in this chapter underscore the need for studies which examine the extent of bacterial IN activity and their potential source environments and sink environments.

Table 3-1. Collection sites of samples from which bacteria were cultured in this study. Sample Location Sample Type Date Maroon Bells trail, Aspen, Soil 6/7/2016 Colorado, USA Maroon Bells Trail, Aspen, Soil 6/7/2016 Colorado, USA Snodgrass Mountain Trail, Soil 6/8/2016 Crested Butte, Colorado, USA

Continental Divide at Soil 6/9/2016 Monarch Pass, Colorado, USA

Continental Divide at Hoosier Soil 6/9/2016 Pass, Colorado, USA Mojave Desert, Nevada, USAa Soil 6/15/2017 Mojave Desert, Nevada, USAa Soil 6/23/2018 Chicago, IL, USA Precipitation 12/28/2015 Blacksburg, VA, USAb Precipitation 3/21/2013 - 4/19/2014 a A total of 4 samples were collected for each sampling event b These bacteria isolates were obtained from Boris Vinatzer (VT) and were cultured from precipitation samples collected over one year.

122

Table 3-2. Broth recipes used for nutrient deprivation experiments. Reasoner’s 2 Agar (R2A) broth was manipulated to have 10% Carbon constituents (R2A-3C), 10% Nitrogen constituents (R2A-3N), and 10% of both Carbon and Nitrogen constituents (R2A- CN). Ingredient R2Aa R2A-3Ca R2A-3Na R2A-CNa Proteose peptone 0.5 g 0.5 g 0.05 g* 0.05 g* Casamino acids 0.5 g 0.5 g 0.05 g* 0.05 g* Yeast extract 0.5 g 0.5 g 0.05 g* 0.05 g* Dextrose (R-Glucose) 0.5 g 0.05 g* 0.5 g 0.05 g* Soluble starch 0.5 g 0.05 g* 0.5 g 0.05 g* Dipotassium phosphate 0.3 g 0.3 g 0.3 g 0.3 g

Magnesium sulfate 0.05 g 0.05 g 0.05 g 0.05 g Sodium pyruvate 0.3 g 0.03 g* 0.3 g 0.03 g*

Table 3-3. Broth recipes used for nutrient deprivation experiments. Minimal media broth (M9) was manipulated to have 10% Carbon constituents (M9-C) and 10% Nitrogen constituents (M9-N). Ingredient M9a M9-Ca M9-Na Disodium phosphate 33.9 g 33.9 g 33.9 g (Anhydrous) Monopotassium 15.0 g 15.0 g 15.0 g phosphate Sodium chloride 2.5 g 2.5 g 2.5 g Ammonium chloride 5.0 g 5.0 g 0.05 g* 1.0 M Magnesium 2 mL 2 mL 2 mL sulfate 1.0 M Calcium 100 μL 100 μL 100 μL chloride Dextrose (R- 4.0 g (0.4% w/v) 0.4 g (0.04% w/v)* 4.0 g (0.4% w/v) Glucose) aAmounts listed per liter of water *Indicates media constituent that was decreased

123

Figure 3-1. Dependence of culture age on IN activity of bacterial isolates. Temperature activity shown is the warmest temperature that freezing was observed for that isolate.

124

Table 3-4. Results of sequence identification of potential Ice+ bacterial isolates. Isolated Isolate_IDa Nearest Neighbor Genus/Class Gram- IN from of Type Species stain activity Soil AZ_122 Arthobacter agilis Actinobacteria Variable -4.5oC Soil AZ_8 Paenibacillus Firmicutes Positive -8oC Soil AZ_82Pink Kocuria polaris Actinobacteria Positive -4°C (99.9%) Soil AZ_82Red Arthrobacter Actinobacteria Variable -5°C agilis (99.2%) Soil CO_1 Arthrobacter Actinobacteria Variable -6°C tectii (99.5%) Soil CO_5 Sphingomonas Alphaproteobacteria Negative -6°C humi (99.73%) Sleet IL_9 Arthrobacter Actinobacteria Variable -7°C agilis (99.9%) Sleet IL_10 Methylobacterium Alphaproteobacteria Negative -8°C brachiatum (99%) Sleet IL_15 Deinococcus sp. Deinococcus- Positive -10°C (98%) Thermus Rain VA_2512 Kocuria sp. Actinobacteria Positive -6oC Rain VA_2527 Sphingomonas sp. Alphaproteobacteria Negative -4oC Rain VA_2548 Plantibacter sp. Actinobacteria Positive -4oC Rain VA_2530 Rathayibacter Actinobacteria Positive -4oC festucae Rain VA_2470 Paenibacillus sp. Firmicutes Positive N/A Rain VA_2500 Leifsonia Actinobacteria Positive -4oC shinshuensis Rain VA_2506 Frigoribacterium Actinobacteria Positive -4oC sp. Rain VA_2538 Acidovorax sp. Gammaproteobacter Negative -4oC ia Rain VA_2503 Lysinibacillus sp. Firmicutes Positive -4oC a Isolate_ID designated by the state it was isolated in (AZ, Arizona; CO, Colorado; IL, Illinois; or VA, Virginia). VA bacterial isolates were isolated in a separate study (114).

125

Table 3-5. Bacterial isolates received from Boris Vinatzer (Virginia Polytechnic Institute) that demonstrated IN activity at -4oC. Isolate ID 4oC Incubations 25oC Incubations 30oC Incubations R2A-3C R2A-3N R2A-3C R2A-3N R2A-3C R2A-3N VA_2527 0/8 8/8 0/8 8/8 8/8 8/8 VA_2530 8/8 8/8 0/8 0/8 0/8 8/8 VA_2548 8/8 8/8 0/8 0/8 0/8 0/8 VA_2512 8/8 8/8 8/8 0/8 0/8 8/8 VA_2470 0/8 0/8 0/8 0/8 0/8 0/8 VA_2500 0/8 8/8 0/8 8/8 0/0 0/8 VA_2506 8/8 8/8 0/8 0/8 0/8 8/8 VA_2538 8/8 8/8 0/8 8/8 8/8 8/8 VA_2503 8/8 0/8 0/8 0/8 0/8 0/8 Bolded values indicate successful freezing events and number of wells frozen at -4oC

Figure 3-2. Gram stains of the two most IN-active bacterial isolates, a AZ_82Pink (a), AZ_82Red (b), and CO_1 (c). Scale bar in lower right hand corner indicates 10 μm.

126

Figure 3-3. Induction of ice nucleation activity in AZ_82Pink using nutrient deprivation. At timepoint 0 (0 days), there was no IN activity. Graphs that possess curves ending before -13oC had assays which had completely frozen prior to the end of the assay (at -13oC), and therefore do not have any data points beyond the temperature at which they completely froze at.

127

Figure 3-4. Nutrient deprivation experiments conducted on AZ_82Red. Graphs that possess curves ending before -13oC had assays which had completely frozen prior to the end of the assay (at -13oC), and therefore do not have any data points beyond the temperature at which they completely froze at.

Figure 3-5. Inhibition of ice nucleation activity in AZ_82Pink (Kocuria sp), at 14 days of incubation. Various media are as follows: Black – Normal R2A; Red – Nitrogen- limited R2A; Blue – Carbon-limited R2A. Treatments are as follows: Circles – No treatment; Squares – Supernatant; Triangles – Heat-treatment. Y-axis shows INP concentrations per liter of liquid.

128

CHAPTER 4 SIZE-RESOLVED BIOLOGICAL INPS IN NIMBOSTRATUS PRECIPITATION

Overview

A biological INP may consist of nanoscale proteinaceous (84), polysaccharide (131), or lipid (408) material (98); they may also be represented by intact microbial cells (81). Information on the size and numbers of biological INPs in the troposphere is paramount for understanding their atmospheric transport and effects on precipitation formation and climate (71). However, few studies have attempted to collect the relevant information necessary to estimate their influence on meteorological and climatological processes. Those that have, did not account for the variety of biological INPs which have varying temperatures of activation and sizes of biological INPs that might occur in the troposphere (136, 289, 291, 312–314). Disregard for these variations in biological INP sizes and temperatures of activation likely lead to seasonal, diurnal, and geographic location biases in INP concentration estimates. Indeed, the concentration of biological INPs in the troposphere is a result of multiple factors that likely include seasonal and diurnal sources, geographic source influences, and meteorological phenomena that can cause their aerosolization from the earth’s surface (214, 241, 273, 274, 382). Additionally, numerous studies have indicated that biological INPs occur in a range of sizes [nanometers (114, 132) to micrometers (90)], and possess diverse surface molecules (15, 23, 82, 90, 130, 131, 133, 134).

The array of biological INPs documented from bacteria (90), fungi (23), algae (134), and plants

(22) together with their varied sizes have made quantification of their occurrence in precipitation, clouds, and air challenging (292).

In addition to the fact that most numerical atmospheric simulations of biological INPs do not account for biological INP diversity (291, 311, 312), there have been few attempts to specifically investigate their impact on stratiform precipitation—that is, the precipitation that

129

falls from stratus and nimbostratus clouds. In fact, with the exception of one study on fungal

INPs, every simulation that has been carried out on biological INPs in cloud processes has been

on convective clouds (136, 289, 291, 312–314). Given what is known about ice formation processes within convective clouds, it is likely that biological INPs have minimal effects on overall precipitation formation (Chapter 1). Nimbostratus clouds, on the other hand, are thought to be far more sensitive to the concentration of INPs present. This is because of the general lack of vertical movements within the cloud, which prevents the formation of secondary ice via ice redistribution and secondary ice enhancement processes (Figure 1-1; (9)). Given that nimbostratus clouds account for a significant portion of global cloud coverage and are major precipitation producers in temperate and polar regions, there is a need to understand the microphysical processes which may lead to their formation, dissipation, and precipitation production.

The results from chapter 2 indicated that the interactions of air masses with distant geographic locations may lead to higher concentrations of biological INPs within the troposphere. Additionally, our data detected seasonal trends in biological INPs, with winter precipitation producing significantly higher concentrations of biological INPs than spring or summer. Of these winter events, precipitation that originated from nimbostratus clouds have the highest concentrations of biological INPs. Thus, this current chapter focuses on research efforts to collect additional data on the nature of biological INPs deposited in nimbostratus precipitation.

Methods

Precipitation Collection

Precipitation was collected in four locations, outlined in Table 4-1: A suburb outside of

Chicago, IL (41.8398oC, 87.9536oW); A suburb outside of Atlanta, GA (33.5832oN,

130

84.3394oW); Gainesville, FL (29.6516oN, 82.3248oW); and Baton Rouge, LA (30.4515oN,

91.1871oW). All precipitation was sampled at ambient temperatures by direct collection in 120 L

galvanized cans that were lined with clean, sterile 94 x 122 cm polypropylene bags (Fisher

Scientific, Pittsburgh, PA). A total of four cans (three for samples and one as a procedural

control) were used to collect samples in Chicago, IL on Dec. 27, 2015; and six cans (five for

samples and one as a procedural control) were used to collect samples on January 6-7, 2017, in

Atlanta, GA and Gainesville, FL. In Baton Rouge, LA, precipitation was sampled temporally and

spatially October 25-28, 2015. For the temporal collections, a total of eleven cans (10 for

samples and one as a procedural control) were used and were collected on the roof of the Life

Sciences building on the Baton Rouge campus, which is a six-story building (~20 mAGL). For

the spatial collections, four cans (3 for samples and 1 for procedural control) were used at Ben

Hur Farm, located approximately 4 miles from the Life Sciences Building (LSB); four cans (3

for samples and 1 for procedural control) were used in an open soccer field, the Burbank Soccer

Complex, located approximately 7 miles from LSB; and four cans (3 for samples and 1 for

procedural control) were used at LSB.

Precipitation samples collected in Chicago were stored on dry ice for overnight shipping and subsequently stored at -80oC until analyzed. For all other samples, immediately following

each precipitation event, the material collected was transferred to sterile 9 L carboys and stored

at 4 °C in the dark until processed. Processing typically occurred within ~1 h, but in certain cases

the samples were processed up to 48 h after collection.

Quantification of INPs and Cells

To size fractionate particles, the precipitation samples were filtered using a peristaltic pump (Geotech, Denver, CO, USA) to progressively concentrate the particles onto sterile 47 mm

131

nucleopore membrane filters, with pore sizes of either 0.10 μm, 1.0 μm, and 3.0 μm, and 100

kDa. Thus size fractions that were analyzed for INP concentrations were as follows: >3.0 μm,

3.0 μm–1.0 μm, 1.0 μm-0.10 μm, 0.10 μm-100 kDa, and < 100 kDa. For the size fractions < 100 kDa, ultrafiltration centrifugal concentrator with a PES membrane (Sartorius, Göttingen,

Germany) was used, and the filtrate analyzed. Immersion freezing assays were performed as described previously (273) but with the following modifications: no filter concentration was performed and 200 µL aliquots (rain or snow water equivalent) of the precipitation sample were placed into each well of a 96-well plate and sealed with adhesive film. Triplicates of each sample and experimental treatment were tested over a temperature range of -4 to -15 oC in 0.5 oC increments using a Neslab RTE 7 series refrigerated ethylene glycol bath (Thermo Scientific,

Waltham, MA). The number of wells frozen at each temperature were recorded and the differential (the number of INPs activated at a specific temperature) and cumulative concentrations (the number of INPs activated at all temperatures warmer than a given temperature) of INPs were calculated by the method of Vali (338).

Each precipitation sample was tested by analyzing triplicate preparations that were either untreated, heated for 10 min at 95 oC, or incubated with 3 mg mL-1 of lysozyme for 1 h prior to

the immersion freezing assay. The data and related calculations were used to assess the total (i.e.,

untreated), biological (i.e., heat-sensitive), and bacterial (i.e., lysozyme-sensitive) INP content of each sample. Given that only proteinaceous INPs may be sensitive to heat denaturation and that not all bacteria are sensitive to lysozyme, this method should be viewed as a conservative estimate for INPs of biological or bacterial origin. DNA-containing cells were stained with

SYBR Gold (Invitrogen, Carlsbad, CA) and counted using epifluorescence microscopy

132

(Olympus BX51-TRF, Center Valley, PA) according to the method of Christner et al. (339).

Triplicate measurements from each precipitation event were analyzed.

Analysis of Meteorological Data

Each storm was classified as a “stratus” or “convective” event based on cloud structure.

Cloud structure, height, and depth were estimated based on radar reflectivity, infrared satellite

imagery, and radiosonde data (347, 348). Backward trajectories (120 h) of air masses over the site at the time of each precipitation event were determined using the NOAA Air Resources

HYSPLIT (Hybrid Single-Particle Lagrangian Integrated Trajectory) Model, accessed via the

NOAA ARL READY website (http://arl.noaa.gov/HYSPLIT.php). For each precipitation event, six altitudes were chosen for backward trajectory analysis based on the lifting mechanism responsible for cloud formation as well as the level of cloud base and height (Supplementary

Figure S1). Cloud base and height were determined using Lifted Condensation Levels (LCL) and/or Convective Condensation Levels (CCL), depending on whether the cloud was formed by mechanical or convective lifting mechanisms, respectively (Supplementary Methods,

Supplementary Figure S1a-d). Geographic locations where backward trajectories interacted with the Earth’s surface or the mixed boundary layer (MBL, i.e., lowest layer of the troposphere where turbulent mixing with the Earth’s surface occurs) were graphed in R by using data from the “tdump.csv” hourly data files produced by HYSPLIT for each backward trajectory

(Supplementary Figure S1e-f). Ecoregions were classified as previously described (Chapter 2).

Results

Stratiform Storm Morphology and Origin

The precipitation collected in all events listed in this study came from stratiform clouds, and their details are listed in Table 4-1. The precipitation collected in Baton Rouge, LA, during

133

October 24-25, 2015, was rain. The precipitation came from stratiform clouds that were

produced as an indirect result of a hurricane which had traveled up the west coast of the United

States. Approximately 50 Liters of rainwater were collected in total. Backward trajectory

analysis indicated that aerosols mainly originated from the SAM and NAM ecoregions (Figure 4-

1a), and that cloud top temperatures reached a low of approximately -6oC. The precipitation

collected outside of Chicago, IL, on Dec. 28, 2015, was mixed-phase but the majority of it was

sleet and freezing rain. The backward air mass trajectories indicated that air masses had last

interacted with the EA and HNL ecoregions (Figure 4-1b). The coldest cloud temperature

reached by these nimbostratus clouds was estimated to be -15oC. The precipitation collected

outside of Atlanta, GA, on January 6-7, 2017, was mixed-phase, and mostly freezing rain

(Appendix C), and obtained cloud tops with minimums of -10oC. Precipitation was collected in

Gainesville, FL, at the same time as the collections in Atlanta, GA, on January 6-7, 2017, however, the precipitation from Gainesville was all rain, without any evidence for ice or mixed- phase precipitation according to weather radars and meteorological reports. The coldest cloud temperature reached by the stratiform clouds in Gainesville, FL was approximately -15oC.

HYSPLIT trajectory analysis indicated that the air masses above Atlanta, GA had last interacted

with the HNL, NFM, and DSAH ecoregions (Figure 4-1c), whereas the airmasses above

Gainesville, FL, at time of precipitation, had last interacted with the SAM ecoregion (Figure 4-

1d). The precipitation collected in Baton Rouge, LA, on Dec. 19-20, 2015, was rain, with cloud top temperature minimums reaching approximately -4oC. This collection was split into four

segments to analyze potential temporal variations in biological INP content for the duration of

the storm. Segment 1 was collected from 0900-1600 UTC on 12/19/2015; Segment 2 was collected from 1600-2100 UTC on 12/19/2015; Segment 3 was collected from 2100 UTC-0100

134

UTC on 12/19/2015 to 12/20/2015; and Segment 4 was collected from 0300–1000 UTC on

12/20/2015. The coldest cloud top temperature observed was approximately -4oC, and backward

airmass trajectories indicated that air mass origins stayed relatively the same throughout the

duration of the entire storm, with previous interactions with HNL, NWFM, and PM ecoregions

(Figure 4-2).

Spatial Differences in INP Concentrations from Baton Rouge, LA Stratiform Precipitation

The overall total and biological INP content were very low in the Baton Rouge, LA

events (Figure 4-3). For the spatial collections, the warmest temperature of IN activity detected

for untreated INPs (total INPs) was at -6oC, at approximately 17 INPs L-1 precipitation, from

samples collected at LSB. This warm-temperature activity was heat-sensitive (biological) and

associated with the 0.22 μm filtrate (Figure 4-3). For the LSB samples, at temperatures above -

10oC, 100% of the 0.22 μm filtrate INPs were biological. Additionally, there was no significant

differences between the concentrations of total and biological INPs for all temperatures,

indicating that most of the total INPs were biological in origin.

A similar trend was observed in the samples collected from Ben Hur farm, however the

onset of freezing for the bulk, total INPs was not detected until -8oC. Likewise, the onset for 0.22

μm filtrate freezing wasn’t observed until -9oC. For the bulk concentrations, at temperatures

above -11oC, 100% of the INPs were biological. However, for the 0.22 μm filtrate samples,

biological INPs were more numerous than total INPs until -12oC. At -12oC and colder, the

biological INPs of the 0.22 μm filtrate made up approximately 50% of the total INPs <0.22 μm.

The INP spectra for samples collected from the soccer fields indicated that bulk INPs, for

both total and biological, generally consisted of significantly more INPs than those which were

135

<0.22 μm. The warmest IN activity detected was at -7oC, from the bulk samples. The INPs <0.22

μm were mostly biological at -8oC, and mostly non-biological at temperatures below -9oC.

Temporal Differences in INP Concentrations from Baton Rouge, LA Precipitation

Precipitation was collected in intervals from a long-lived, stratiform precipitation event

that occurred over Baton Rouge, LA, on Dec. 19-20, 2015. Overall, the majority of warm-

temperature INPs (>-10oC) were biological and bacterial in origin (Figure 4-4). A MANOVA

indicated that the overall differential temperature spectrum of total, biological, and bacterial

INPs (Figure 4-4 a-c) did not differ significantly between intervals (Wilks’ Lambda, p=0.87 for

total INPs; p=0.82 for biological INPs; and p=0.88). Interval 4 had the highest temperature of IN

activity detected at -5oC, of which were bacterial in origin (Figure 4-4c). At temperatures above approximately -10oC, interval 2 appeared to produce significantly higher concentrations of

bacterial INPs, when compared to the other intervals, whereas interval 3 produced significantly

lower concentrations of bacterial INPs. At colder temperatures (<-10oC), biological INPs were

significantly higher in collections from interval 2 when compared to the other intervals.

However, multivariate analysis of variance indicated that overall, IN activity at these

temperatures did not differ significantly between the intervals sampled.

Size-Resolved INP Concentrations

The concentration of INPs collected from the Chicago sleet storm contained very high concentrations of both total and biological INPs when compared to the average concentrations of total and biological INPs from stratiform events collected in Baton Rouge, LA (Chapter 2). At temperatures above -10oC, all INPs were inferred to be biological, and thus only the biological

INPs are shown in Figure 4-5. Overall, the size fractions larger than 3.0 μm made up the largest fraction of all biological INPs at temperatures above -10oC (Figure 4-5), however the biological

136

INPs smaller than 0.1 μm also comprised a large fraction of all biological INPs. At temperatures

colder than -10oC, the size fraction smaller than 0.1 μm contributed to more of the overall

biological INP concentrations when compared to the size fractions larger than 3.0 μm (Figure 4-

5). The size fractions in between 3 μm and 1.0 μm were the next most numerous fractions of

biological INPs, however their contribution to bulk biological INP concentrations were

significantly less than those greater than 3.0 μm and smaller than 0.1 μm. The size fractions in

between 1.0 μm and 0.1 μm had the second to last highest concentrations of biological INPs. The

100 kDa filtrate biological INPs contributed the least to the bulk amount of biological INPs

(Figure 4-5). The warmest temperature that biological IN activity was detected at was -5oC, in

the bulk INP assays. The size fractions greater than 3.0 μm showed IN activity at -6oC, with approximately 160 INPs L-1 detected (Figure 4-4). The size fractions in between 3.0 μm and 1.0

μm displayed IN activity at -6oC as well, at 15 INPs L-1. The size fraction less than 0.1 μm

displayed approximately 65 biological INPs L-1 at -6oC.

The concentrations of INPs differed significantly between the mixed-phase samples

collected in Atlanta versus the rain collected in Gainesville on the same day (Figure 4-6). While

the cumulative concentrations of biological INPs collected in the Atlanta mixed-phase precipitation were not as high as those collected in the Chicago, IL mixed-phase precipitation, they were still significantly higher than the cumulative concentrations of biological INPs collected in the convective precipitation events from chapter 2 (Figure 2-2).

Discussion

Relating the concentrations of biological INPs in precipitation to concentrations in the source clouds has several caveats. First, there is the issue of below-cloud scavenging, in which

INPs and other aerosols are removed from below the cloud by falling precipitation (410).

137

Scrubbing of biological INPs from the atmosphere would therefore lead to inflated estimates of

the concentrations of biological INPs in clouds. However, previous studies have shown similar

concentrations of INPs in cloud water compared to precipitation at ground level (176, 225, 398),

suggesting that below-cloud scrubbing may have a negligible role in altering the composition of

bacterial-sized INPs in raindrops (411). Another potential limitation to studying biological INPs

in precipitation is the spatial and temporal differences in the concentrations of biological INPs

that could occur throughout the duration of and in different regions of the storm. Certain studies

indicate that biological INPs are preferentially precipitated out within the first portion of a storm

(278). If the start of precipitation event was missed, the potential concentration of biological

INPs could be biased in the collected precipitation samples. Further, if the entirety of the event

was collected, the concentration of biological INPs could be diluted by further precipitation

collection should they originated in the earliest portion of the precipitation event.

Our analysis on the temporal distribution of biological INPs in precipitation indicated that

there was no significant difference in biological or bacterial INP concentrations collected

throughout the precipitation event, when compared to the beginning of the precipitation event. In

fact, there appeared to be higher concentrations of bacterial INPs during the middle intervals

(Figure 4-4). However the results of a MANOVA indicated that there was not a significant difference in the total, biological, or bacterial INPs for any interval at the temperatures tested.

Given that the air mass trajectories for the temporal collections appeared to stay relatively the

same throughout the duration of the event (Figure 4-2), this is consistent with air mass history as

the biggest determinant of biological INP concentrations in clouds, and subsequently,

precipitation. However, the data are limited to analysis of one precipitation event and therefore

may not be representative of all conditions and locations.

138

To examine spatial variations in biological INP data collected in close proximity sites

(within ~7 km?), we collected precipitation from three different locations in Baton Rouge, LA during October, 2015. No significant differences in INP concentrations were detected between the three locations. Size resolution of these samples indicated that size fractions < 0.22 μm of both total and biological INPs were consistently lower than those above 0.22 μm (Figure 4-3).

The lack of significant differences in spatial and temporal collections for biological INP concentrations in precipitation implies that within the same precipitation event, INP concentrations are relatively uniform. Thus, point collection of precipitation is likely to be representative of biological INP content from a cloud, at least on the scale of tens of kilometers.

Further, a point collection from a long-lived precipitation event that experiences diurnal variations may also still be representative from the beginning to the end of the storm. Based on this data, these remarkably similar measurements from different points in time would imply that collections of precipitation at any point during the event would be representative of the entire storm. In other words, the use of biological INP concentrations based on samples from a given time and location may constitute a representative precipitation sample.

The size-resolved biological INP concentrations for the Chicago and Atlanta events had varying results. Collections from stratiform, mixed-phase precipitation indicated that the sizes of biological INPs sampled differed significantly between precipitation events and locations

(Figures 4-4 and 4-5). Samples collected from the Chicago winter storm on Dec. 28, 2015, harbored biological INPs > 3.0 μm and < 0.1 μm, both of which contributed a high amount to the total biological INP concentrations (Figure 4-4). The sizes of biological INPs in between 3.0 μm and 0.1 μm, which is in the size range of most bacteria, did not contribute much to the total number of biological INPs analyzed. On the other hand, biological INPs larger than 3.0 μm made

139

up the smallest fractions of all biological INPs collected in the Atlanta, GA, precipitation event

(Figure 4-5). Of the two stratiform events analyzed, biological INPs < 0.1 μm contributed

significantly to both. Backwards trajectory analysis indicated that both the Chicago and Atlanta

precipitation events had previously had interactions with the high northern latitudes (Figure 4-1).

Their similarities in the concentrations of biological INPs <0.1 μm may be a result of the

collection of similar bioaerosols that had been sourced from the high northern latitudes. Several

studies have implicated Arctic regions as potential sources of airborne, biologically-derived INPs

(412, 413). In fact, a study done at Prudhoe Bay in Alaska indicated that airborne INPs from 3 to

12 μm in size constituted the most active fraction of INPs detected, and that they were likely

sourced from fresh sea ice melt near phytoplankton blooms (412). The overlap between that

study’s area of sampling and the air mass trajectories of the precipitation collected in Chicago

are evident (Figure X). As such, it is possible that the arctic regions of Alaska are significant

sources of biological INPs > 3 μm in size. However, aerosols larger than 3.0 μm are not thought

to be able to travel long distances due to their weight, and are more likely to be scrubbed from

the atmosphere by falling precipitation (2). Therefore, the biological INPs > 3.0 μm in the

Chicago precipitation may have been sourced from a more local environment. Similarly, the size of the INP that interacts with water species in the clouds can influence the rate of ice nucleation.

Indeed, theory and experimental results suggest that larger particles are likely more efficient

INPs (2).

Biological INPs < 0.1 μm, on the other hand, do have the potential to travel long distances due to their small size (2, 414). Backward trajectory analysis indicated that both the

Chicago and the Atlanta airmasses had interactions in similar regions of the HNL ecoregion

(Figure 4-1b, c). A study conducted in Lake Erie, which is near the geographic locations where

140

both the Atlanta and Chicago storms’ airmasses had interacted with, showed that the majority of

the biological INP content analyzed in the surface waters of the lake were submicron in size, and

active between -3 and -15oC (409). In fact, this same study indicated that the source of the highly

active, submicron-sized biological INPs was terrestrial in origin, due to the signature detected

from soil-dwelling fungal species. It is possible that soil-dwelling biological INPs that are located within a watershed may eventually become aerosolized through bubble bursting and wave breaking, as would occur in large bodies of water such as Lake Erie (415–417). However, more likely they originated directly from soil (133, 259)

Concluding Remarks

Variations in the sizes of biological INPs that are present within precipitation could help to elucidate the types of biological INPs (e.g., pollen, bacteria, fungi, other organic matter) that affect precipitation generation in the cloud. Indeed, the few cloud modeling studies which have taken biological INPs into account when examining their potential effects on ice formation, rarely examine the variations in sizes in which they may occur (303, 309, 312, 314). The sizes of the INPs being used are important for transport processes (e.g., larger INPs are less common in clouds due to their weight and inability to escape the MBL), as well as their effects on IN rates.

In fact, almost all numerical models assume a single concentration and size of all biological INPs and are usually based on a single species of biological INPs. There are no studies which consider the various types and sizes of biological INPs that occur, i.e., pollen, fractions of bacteria, clumps of bacteria, fungal spores, etc. As such, it is important that measurements regarding the concentrations as well as the various compositions of the surfaces of the biological INPs that can occur within the atmosphere be examined much more closely.

141

Table 4-1. Sampling locations and times, with respective precipitation types collected and the ecoregions that the storm origins interacted with. Location Geographic Time and Date Type of Cloud Cloud Ecoregi Coordinates Precipitati altitudes temperat on on ures LSB, Baton 30.4109oN, Dec. 19-Dec Rain 400 - 4000 13 to -4 HNL, Rouge, LAa 91.1770oW 20, 2014 NFM, 09Z (12/19) – PM 1000Z (12/20) LSB, Baton 30.4109oN, Oct. 24-25, Rain 300 – 19 to -6 NAM, Rouge, LA 91.1770oW 2015 6000 SAM 20Z (10/24) – 23Z (10/25) Ben Hur 30.3740oN, Oct. 24-25, Rain 300 - 6000 19 to -6 NAM, Farm, Baton 91.1535oW 2015 SAM Rouge, LA 20Z (10/24) – 23Z (10/25) Burbank 30.3459oN, Oct. 24-25, Rain 300 - 6000 19 to -6 NAM, Soccer 91.0969oW 2015 SAM Complex, 20Z (10/24) – Baton 23Z (10/25) Rouge, LA Chicago, IL 41.8396oN, Dec. 28, 2015 Mixed- 600-915; -2 to -6; HNL, 87.9649oW phase and 1700-5500 1 to -15 EA sleet Atlanta, GA 33.5796oN, Jan. 6-7, 2017 Mixed- 600 - 3600 5 to -10 HNL, 84.3417oW 18Z (1/6) – 06Z phase and NFM, (1/7) freezing DSAH rain

Gainesville, 29.6399oN, Jan. 7, 2017 Rain 500 - 5400 5 to -15 SAM FL 82.3623oW 00Z (1/7) – 10Z (1/70) aCollections made on Dec 19-20 were collected in four segments from the same storm to study temporal variation: Segment 1 09Z (12/19)-16Z (12/29); Segment 2 16Z (12/19)-21Z (12/19); Segment 3 21Z (12/19) – 01Z (12/20); Segment 4 03Z (12/20) – 10Z (12/20)

142

Figure 4-1. 120-hour backward air mass trajectories for Chicago, IL (Dec. 28, 2015); Atlanta, GA (Jan 6-7, 2017); Baton Rouge, LA (Oct. 24-25, 2015); Gainesville, FL (Jan. 7, 2017) precipitation collections. Red stars indicate sampling locations.

143

Figure 4-2. 120-hour backwards air mass trajectories for the temporal precipitation collections made in Baton Rouge, LA, on Dec. 19-20, 2015. Red trajectories indicate Segment 1 (0900-1600 UTC on 12/19); Yellow trajectories indicate segment 2 (1600-2100 UTC on 12/19); Green trajectories indicate segment 3 (2100 – 0100 UTC on 12/19-12/20); and blue trajectories indicate segment 4 (0300-1000 UTC on 12/20). Sample location, Baton Rouge, LA, is indicated by the white star.

144

Figure 4-3. 0.22 μm size-cut for INP concentrations taken from Hurricane run-off during RAINS campaign in Baton Rouge, LA. Top graph: INP concentrations from the soccer fields. Middle graph: INP concentrations from Ben Hur. Bottom graph: INP concentrations from LSB. Closed markers are bulk concentrations. Open markers are 0.22 μm filtrate concentrations. Circles are total concentrations. Triangles are biological concentrations.

145

Figure 4-4. Temporal variability of (a) Total INPs, (b) Biological INPs, and (c) Bacterial INPs collected in Baton Rouge, LA, on Dec. 19-20, 2015. Interval 1 was collected from 0900-1600 UTC on 12/19; Interval 2 was collected from 1600-2100 UTC on 12/1); Interval 3 was collected from 2100 – 0100 UTC on 12/19-12/20; and interval 4 was collected from 0300-1000 UTC on 12/20.

146

Figure 4-5. Size-resolved concentrations of biological INPs detected in mixed-phase precipitation collected during the Chicago winter storm on Dec 27, 2015.

Figure 4-6. Concentration of biological INPs from mixed-phase precipitation samples collected from Atlanta, GA.

147

CHAPTER 5 INFLUENCE OF BIOLOGICAL ICE NUCLEATING PARTICLES ON PRECIPITATION FORMATION IN NIMBOSTRATUS CLOUDS Overview

In the late 1950s, atmospheric scientists began to explore the sources of warm temperature (<-10oC) ice nucleation in samples of snow, rain, and hailstones. As more

sophisticated sampling equipment for in-situ cloud research became available, investigators

discovered that in some instances, ice crystal numbers exceeded INP numbers (72, 73, 301). For

clouds with temperatures well below -10oC, this phenomenon could be explained by secondary

ice formation processes, whereby INPs such as mineral dust aerosols nucleate the first ice

crystals of a cloud that subsequently multiply through splintering and collisions with other ice

particles. However, the presence of mineral dust aerosols, which are thought to be the dominant

type of INP present in the atmosphere, do not initiate freezing at temperatures warmer than -15oC

(71), and thus cannot explain this ice multiplication phenomenon in clouds with warmer subzero

temperatures. However, certain microorganisms and biogenic molecules are very effective INPs

and have ice-nucleating activities > -10oC (22, 76, 131, 139, 323). The simple fact that bioaerosols are highly efficient INPs coupled with their ubiquity in the atmosphere and precipitation has fueled speculation on their roles in meteorological processes. While various environments are recognized to harbor biological INPs (80) and their presence in precipitation is widely documented (176, 273–275, 329, 330), very little is known about their sources to the atmosphere. Despite the critical need for new information on the atmospheric sources and distribution of biological INPs (71), deciphering the seasons, geographical locations, and atmospheric conditions under which these bioaerosols can exert meteorological effects has been hampered by the paucity of available data.

148

Analysis of two years of precipitation data (Chapter 3; (274)) showed that winter storms in Louisiana contained the highest concentrations of total, biological, and bacterial INPs, supporting the hypothesis that seasonal changes in meteorological patterns and/or the source ecosystems of bioaerosols influenced the abundance of INPs in the precipitation. Additionally, a subset of the wintertime nimbostratus events contained all the extreme outlying INP datapoints, had the highest concentrations of INPs, had significantly higher concentrations of INPs compared to winter convective storms, and were significantly enriched for OTUs (e.g.,

Hymenobacter spp.) that positively correlated with the INP data. There were only four observations of ice-containing precipitation (snow or sleet) during this study, and these samples had the highest INP concentration and activity observed. To further investigate the types of biological INPs precipitating out of wintertime nimbostratus storm systems, several targeted wintertime nimbostratus precipitation events were collected as a follow-up study, which were described in the previous chapter.

The mechanisms responsible for primary and secondary ice formation in nimbostratus clouds differ substantially from those of convective clouds (31, 302–304). In mixed-phase nimbostratus clouds, cold-phase microphysical processes are the dominant pathway of precipitation production. This is in part due to fact that the ice crystals grow quickly, with humidity close to water saturation, and partly because supercooled cloud-droplets are too small for liquid coalescence when the cloud-base is near 0oC. Additionally, low altitude, mixed-phase nimbostratus clouds possess conditions highly favorable for interacting with biological INPs.

Since nimbostratus clouds are low level, relatively warm clouds (9), biological INPs would be in a favorable position to affect precipitation processes if sufficiently abundant. Importantly, it is speculated that ice and precipitation formation in these types of stratiform clouds are highly

149

dependent on INP number concentrations (418). Given the common occurrence of stratiform

clouds, and their potential to significantly affect global climate and precipitation, there is a need

to investigate the effects of biological INPs on stratiform cloud lifetime, albedo, and

precipitation formation. The objective of this study was to usebiological INP data from stratiform

clouds (chapter 4) to examine the effect of their concentrations on the properties of stratiform

clouds inWeather Research and Forecasting (WRF) cloud simulations.

Methods

Numerical Model

Simulations were performed using the Weather Research and Forecasting (WRF) Model, version 4.1 (419).The simulations carried out in this study were performed using the Weather

Research and Forecasting (WRF) Model, version 4.1 (419). The WRF model, developed and

maintained by the National Center for Atmospheric Research (NCAR) is a mesoscale numerical

weather prediction system that uses two dynamic solvers, the Advanced Research WRF (ARW)

core and the Nonhydrostatic Mesoscale Core (NMM). The ARW core used in this study is fully

compressible and uses Eulerian and nonhydrostatic equations with run-time hydrostatic options.

It uses Arakawa C-grid staggering, and Runge-Kutta 2nd and 3rd order time integration schemes,

and 2nd to 6th order advection schemes in both the vertical and horizontal, and its dynamics

conserve scalar variables. For the simulations carried out in this study, nonhydrostatic equations

and the 3rd order Runge-Kutta time-integration scheme was used. The turbulence and mixing

option (diff_opt) was set to option 2, which allows for the evaluation of mixing terms in physical

space; and the eddy coefficient option was set to a constant K (option 1). Upper level damping

was set with w-Rayleigh damping, and the damping was set to 20 km from the model top, with a

damping coefficient of 0.1. The horizontal and vertical diffusion constants were set to 0 m2 s-1.

150

Divergence damping was set to 0.1. Time off-centering for vertical sound waves was set to 0.1,

and 6 sound steps per time-step were used. The horizontal momentum and scalar advection

orders were set to 5, and the vertical momentum and scalar advection orders were set to 3.

Although a suite of physics parameterizations are available for use in WRF, only the

microphysics scheme was turned on and set to the Thompson Aerosol‐Aware microphysical

parameterization (Option 28), which considers “water‐friendly” and “ice‐friendly” aerosols

(420). All other microphysical schemes were turned off and the requirement for a cumulus

parameterization was bypassed by using high-resolution grid spacing. The Thompson Aerosol-

Aware Scheme (420) is an adaptation of the bulk microphysical parameterization scheme

described in Thomson et. al. (421, 422). This scheme uses bulk microphysics and five separate

water species: cloud water, cloud ice, rain, snow, and a hybrid graupel-hail category. The more

recent adaptation of this aerosol-aware scheme (420) includes the five water species and

explicitly predicts CCN and INP number concentrations, as well as the droplet number

concentration of cloud water (Nc). The CCN, or hygroscopic aerosols, are referred to as “water

friendly” aerosols (Nwfa) and the INPs, are referred to as “ice friendly” aerosols (Nifa). The equations for mass mixing ratios or number concentrations of each species (vapor, liquid, or solid) are described in Thompson et al (420). The process rate terms for existing species are described in Thompson et al (421, 422), while the process rate terms for new variables of Nc,

Nwfa, and Nifa are described in equations 2-4 of Thompson et al (420). The modifications to the

Thompson microphysics module were altered as described below.Although a suite of physics

parameterizations are available for use in WRF, only the microphysics scheme was turned on,

and was set to the Thompson Aerosol‐Aware microphysical parameterization (Option 28), which

considers “water‐friendly” and “ice‐friendly” aerosols (420). All other physics were turned off,

151

and the need for a cumulus parameterization was bypassed by using a high-resolution grid

spacing. The Thompson Aerosol-Aware Scheme (420), is an adaptation of the bulk

microphysical parameterization scheme described in Thomson et. al. (421, 422). This scheme

uses bulk microphysics and uses five separate water species: cloud water, cloud ice, rain, snow,

and a hybrid graupel-hail category. The newer aerosol-aware scheme includes the five water

species, and explicitly predicts CCN and INP number concentrations, as well as the droplet

number concentration of cloud water (Nc). The CCN, or hygroscopic aerosols, are referred to as

“water friendly” aerosols (Nwfa) and the INPs, are referred to as “ice friendly” aerosols (Nifa). The

equations for mass mixing ratios or number concentrations of each species (vapor, liquid, or

solid) are described in Thompson et al (420). The process rate terms for existing species are

described in Thompson et al (421, 422), while the process rate terms for new variables of Nc,

Nwfa, and Nifa are described in equations 2-4 of Thompson et al (420). The modifications to the

Thompson microphysics module was altered as described below.

Idealized Tests

A sensitivity analysis was conducted using WRF’s two-dimensional flow over a hill idealized simulation, to mimic the movement of air over a cold front. The only meteorological input used was radiosonde sounding data from the NWS station at Green Bay, WI (GRB, Station

ID 72645), which was launched at 12Z on 12/28/2015. The sounding was modified slightly to increase linearly from 0 m s-1 at the surface to a constant 15 m s-1 at 1 km and above (420). Other than modifications to horizontal wind components, all other meteorological variables were kept the same. The model contained 120 grid points in the horizontal, each spaced 100 meters apart.

The hill height and half-width were derived based on the estimated depth and of the cold front that caused lifting of air to produce the stratiform cloud. This was done by analyzing atmospheric

152

soundings from multiple weather stations and drawing lines of isentropes between weather

stations (423). Thus, the simulation contained a 1 km high mountain with a 25 km half-width,

and a model top of 15 km, with 41 vertical levels spaced with sigma coordinates. caused lifting

of air to produce the stratiform cloud. This was done by analyzing atmospheric soundings from

multiple weather stations and drawing lines of isentropes between weather stations (423). Thus,

the simulation contained a 1 km high mountain with a 25 km half-width, and a model top of 15

km with 41 vertical levels spaced with sigma coordinates. The simulation was run for a total of

10 hours, with 72 second time steps.

Empirical Biological INP Parameterization

The Thompson Aerosol-Aware microphysics scheme is described in detail in Thompson

et al (420), and the accompanying Fortran code is (“module_mp_thompson.F”) is available free to the public on the WRF website. Only the portions of the microphysical scheme that were modified in this study are described in detail here. The Thompson aerosol aware scheme has the option for the “aerosol-aware” function to be on or off. When turned on, the scheme uses a global aerosol dataset derived from a six-year global model simulation (424), which uses information on aerosols emitted from natural and anthropogenic sources modeled by the

Goddard Chemistry Aerosol Radiation and Transport (GOCART) model (425). This global aerosol dataset includes mass mixing ratios of sulfates, sea salts, organic carbon, dust, and black carbon with a 0.5o x 1.25o (longitude x latitude) grid spacing. The Thompson Aerosol-Aware

scheme identifies INP number concentrations from the 7-yr dataset by the parameterization of

DeMott et al (317), which expresses INP numbers based on the mass concentration of dust

aerosols larger than 0.5 μm. Thus, the Nifa number concentration is calculated prior to model run

as the number of dust particles larger than 0.5um. All other aerosols besides black carbon are

153

included as internally mixed cloud-droplet CCN (Nwfa). Cloud ice activation then proceeds based

on mineral dust parameterizations set forth in several previous studies (309, 316, 323, 426, 427).

The DeMott (317) 0.5 μm parameterization is used to determine the number of dust particles that cause primary ice nucleation in the cloud when above water saturation (i.e. for immersion and condensation freezing). When below water saturation (i.e. deposition freezing), the heterogeneous ice nucleation parameterization described in Phillips et al (428) is used. The homogeneous freezing of existing water droplets follows the parameterization of Bigg et al

(429).

For these idealized simulations, the Aerosol-Aware scheme was turned on by inputting the WRF namelist option “use_aero_icbc” and setting it to “.true.”. The Thompson microphysics module was altered such that the concentrations of “biological INPs” were represented by warm temperature ice nucleation. This was done by manually setting the concentration of INPs active from ‐4oC to -10oC to 0.1, 1, and 10 biological INPs m‐3. Line 4938 in Object 5-1 would

normally enumerate the concentration of INPs based on mineral dusts >0.5 μm, but in the

modified version, it contains the concentration of biological INPs (xni) instead. The temperature

of activation of the biological INPs is adjusted in line 2285, where the temperature of activation

would normally need to be less than 253.15 K (-20oC) to allow for heterogeneous ice nucleation.

For the simulations in this study, that temperature was changed to 269.15 K (-4oC) to allow for

ice nucleation in the temperature range where ice nucleation is mostly biological. This method

directly bypasses the inclusion of Nifa aerosols based on the 7-yr climatological dataset, and

instead sets the initial number concentration of Nifa to that of 0.1, 1, and 10 biological INPs. Thus the “biological INP” parameterization directly assesses the effect of concentrations of warm-

temperature INPs, whereas the original scheme does not allow for immersion or condensation

154

freezing at temperatures above -15oC, and is based only on the number of mineral dust particles.

The default scheme, which allows only for dust INPs, was also run, and as it does not include

biological INPs, is referred to as the scheme with “0 bio INPs”.

Results

Precipitation was sampled in an open field in the western suburbs of Chicago, IL

(41.8396oN, 87.9649oW). The duration of the precipitation event lasted several days, from

December 26-28, 2015, and affected a large portion of the Midwest. The sleet and snow total

recorded at the Chicago Rockford International Airport was 3.5 in., with sleet comprising the

total of anywhere from 1.9 in. to 2.1 in. across northern Illinois. Samples were collected on Dec.

28, 2015 from 1000-1800 UTC. Assuming cloud droplet diameters of 10 μm, concentrations of

250 cm-3 based on average continental stratus cloud data (2), and that below-cloud scavenging was negligible, the cumulative concentrations of biological INPs active at -10oC were estimated

in cloud water and at a concentration of approximately 1 biological INP m-3. For the simulations, biological INPs were assumed to be immersed homogeneously in cloud droplets prior to freezing, as the immersion freezing assays only measure this mode of freezing. Additionally, homogeneous distribution of biological INPs throughout the cloud were also assumed.

The results of the sensitivity analysis indicated that the inclusion of biological INPs significantly affected ice formation within the cloud (Figure 5-1). Figure 5-1 shows a vertical cross section of the cloud after 2 hours of the simulation, which was a sufficient period of simulation to allow for cloud water and ice development. Both snow and ice mixing ratios were significantly lower in the simulation that only had 0.1 biological INPs m-3 (Figure 5-1). While the simulations that used 1versus 10 biological INPs m-3 contained similar snow and ice mixing ratios, the simulation with 1 biological INP m-3 typically had slightly higher mixing ratios. For

155

all three simulations, precipitable water content stayed relatively equal, with minimums of

approximately 2.5 kg m2 and maximums of approximately 6.2 kg m2.

The use of 1 biological INP m‐3 of cloud increased the total amount of non‐convective

snow and rain production, when compared to only 0.1 biological INPs m‐3 (Figure 5-2). With

only 0.1 biological INPs m‐3, cloud top temperatures did not exceed –12oC, whereas the

simulations with 1 and 10 biological INPs m‐3 had cloud top temperature maximums at –39 and –

45oC, respectively. Additionally, total non‐convective rain and snow amounts were only 0.85

mm and 0.656 mm, respectively, for the 0.1 biological INP m‐3 simulation, whereas they were

1.89 mm and 1.88 mm, respectively, for the 1 biological INP m‐3 simulation. Overall, when the

model was run with 10 biological INPs m‐3, the cloud top temperature and precipitable water

content remained similar to the 1 biological INP m‐3 simulation, but the total amounts of rain and

snow produced for the entirety of the simulation decreased. The 0.1 and 10 biological INP

simulations had smaller maximum radar reflectivities, of 10.7 and 10.9 dBZ, respectively, when

compared to the maximum radar reflectivity (13.8 dBZ) of the simulation which used 1

biological INP (Figure 5-2). The simulation which only examined dust INPs (“0 bio INPs”),

produced no snow, and 0.1 mm of rain, and overall produced the least amount of precipitation for all of the simulations (Figure 5-2).

Discussion

Conclusions from existing modeling studies that have evaluated the role of biological

INPs in ice nucleation and precipitation formation are equivocal, with some studies implying

they are insignificant on a global scale (291), while others have found them to be potentially

important (271, 313). One of the drawbacks in modeling biological INP effects on clouds and

climate is the lack of field observations for biological INP concentrations. In fact, most

156

numerical modeling efforts are based on limited data (e.g., (314, 317, 421)) and did not consider

the seasonal and spatial trends of biological INP sources (274), and thereby, the ecological

associations of the airborne microorganisms (241). For example, Phillips et al. (314) used

averages of bacterial INP concentrations per bacterial cell (approximately 10‐4 INPs per cell)

and averages of total airborne bacterial concentrations based on the culturable fraction, which

likely underrepresent the number of biological INPs by not resolving the various species of

bioaerosol separately.

Simulating the effects of biological INPs on clouds has also been limited by

technological and logistical limitations associated with obtaining the parameters necessary for

microphysical modeling. For example, the most commonly used heterogeneous ice nucleation

parameterization, which relates INP number concentrations to aerosols with aerodynamic sizes >

0.5 mm (317), is based on the assumption that mineral dust INPs and aerosols with aerodynamic diameters larger than 0.5 mm are positively correlated (316). While this scheme may accurately predict mineral dust INP concentrations, this size‐constraint may lead to an underestimation of biological INPs that are smaller than 0.5 mm, such as the extracellular vesicles of Erwinia herbicola that are INPs (98) and those commonly observed in precipitation (Figures 4-5 and 4-

6). Furthermore, data acquired for the DeMott et al. (317) parameterization was restricted to particle sizes < 1.6 mm due to constraints imposed by the sampling method, which may lead to underestimations of biological INP concentrations. Additionally, in situ biological INP measurements in surface air may not accurately represent the number of biological INPs that are active while in cloud. Indeed, studies have shown that bacteria can actively divide and metabolize in cloud water (253, 318–320). Ice nucleation active bacteria are also known to modulate INP production on the scale of hours to days based on environmental conditions. For

157

example, INP production increases by several orders of magnitude when E. herbicola is

subjected to UV radiation (110) and P. syringae is subjected to cold temperatures and nutrient deprivation (103). Hence, it may be necessary to consider such physiological changes to accurately portray biological INPs in numerical simulations.

This study aimed to determine whether the concentration of biological INPs enumerated

from precipitation collected from a nimbostratus cloud occurred in concentrations that would

have affected precipitation generation in that cloud. To simplify the model and clearly determine

whether changes in precipitation amounts and cloud structure were due to biological INPs, a 2D

idealized simulation was used. While the nimbostratus event for which the sounding was derived lasted for several days and covered several hundred kilometers, the idealized simulation was run on a much smaller scale, for a total of 10 hours and with a total domain of 12 km. Additionally, many microphysical options are available for use in WRF, but this study examined only whether the Thompson “Aerosol Aware” scheme (420) could be manipulated to express biological INP number concentrations, as it is one of the few existing microphysical parameterizations that allowed for the explicit a priori treatment of INP number concentrations.

The results of this simulation indicated that biological INPs, defined in this

context as those that are active at -10oC and warmer, influenced cloud ice mixing ratios (Figure

5-1) and the total amount of precipitation that was produced (Figure 5-2). The simulation that

included 1 biological INP m-3 produced the most precipitation, which was significant for several

reasons. Firstly, 1 biological INP m-3 was the estimated concentration of biological INPs to have

been present in the cloud based on the number quantified from its precipitation samples (Chapter

4). Second, the simulation clearly indicated that the production of significant amounts of

precipitation from this type of cloud required the presence of biological INPs, as the simulation

158

which only used 0.1 biological INPs m-3 produced significantly less rain and snow. The

simulation which used dust INPs only (0 biological INPs), produced only 0.1 mm of rain and no

snow. These results confirm that ice formation in mixed-phase stratiform clouds, which in this scenario is directly influenced by biological INPs, is an important step in the subsequent production of precipitation. Given that the formation of ice in stratiform clouds is thought to occur mostly through primary mechanisms (9, 430), it is not surprising that INP concentrations would influence ice formation. However, no studies have shown a direct connection between biological INP number concentrations and precipitation production. Previous studies have indicated that while biological INPs can cause ice formation in clouds, their concentrations due not occur in high enough numbers to subsequently affect precipitation production (315). It is thought that the more numerous INPs, such as mineral dusts, are the most important factors leading to precipitation formation (2). Thus, this is the first study to demonstrate numerically the

effect of biological INPs on precipitation production in a mixed-phase nimbostratus cloud.

Interestingly, while the differences between the inclusion of 1 biological INP versus 10 biological INPs m-3 were small, this simulation did indicate that too many biological INPs in a

cloud may have the reverse effect on precipitation production. High concentration of INPs in a

cloud could lead to a decreases in precipitation, due to the formation of many small ice crystals,

none of which are heavy enough to become precipitable (2). Indeed, the WBF mechanism, which can lead to the production of precipitation in some instances, can also cause cloud glaciation

(55). Such glaciation processes occur through the formation of ice crystals at the expense of water vapor or liquid water, which can, in some instances, cause the cloud to dissipate or prohibit the production of precipitation (55).

159

The significant differences in cloud top temperatures between the three simulations

should also be noted. The simulation which used 0.1 biological INPs m-3 produced a cloud with significantly warmer temperature minimums than those of the simulations which used 1 and 10 biological INPs m-3. This result may be a reflection of the importance of ice nucleation in the

vertical development of the cloud. Indeed, the release of latent heat during the freezing process is

a significant component to the invigoration of updrafts in convective clouds, allowing for their

extensive vertical development, and the production of more precipitation (9).

Concluding Remarks

The results presented in this study indicate that biological INPs are important for the

formation of precipitation in nimbostratus clouds. However, this numerical study is a simplistic

representation of what might occur in the real troposphere. As such, subsequent studies should

attempt to include a similar microphysics scheme in more realistic numerical simulations.

Further, this study did not attempt to create a scalable biological INP parameterization that could

be used by other researchers. It simply indicated that biological INPs should not be left out in

simulations which investigate stratiform precipitation. Thus, a particular area of future research

would be to create a parameterization, similar to that of DeMott et al (317), but which

specifically treats biological INPs separately from the common abiotic INPs, such as mineral

dusts. Lastly, the potential for biological INPs to influence nimbostratus cloud vertical depth,

cloud ice and water content (and thus albedo), and cloud dissipation should be investigated more

closely, as these clouds occur on a global scale and are important components to the global

radiative budget. While much work can still be done on the dynamics of biological INPs in

nimbostratus clouds, this study provided necessary preliminary evidence that exemplifies their

potential significance to precipitation formation.

160

Figure 5-1. Vertical cross section of the cloud produced by the 2D idealized simulation of the mixed-phase precipitation collected outside of Chicago, IL, on Dec. 28, 2015. Top panels show cloud snow mixing ratios (g/kg), and the bottom panels show cloud ice mixing ratios (g/kg) at hour 2 of the simulation for 0.1 biological INPs (left column), 1 biological INP (middle column), and 10 biological INPs (right column). Vertical grid levels are numbered on y-axes (total of 41 vertical levels, and 15 km top), and horizontal grid spacings are numbered on x-axes (total of 120 grid points, spaced 100 meters apart, for a total distance of 12 km).

161

Figure 5-2. Total amounts of non-convective rain and snow (mm) plotted as a function of biological INP content. The dust INPs-only simulation is represented by “0 Bio INPs”.

162

CHAPTER 6 CONCLUSIONS

Since the discovery of the Ice+ phenotype in P. syringae in the 1970s, biological INPs

have been detected in air (125, 157, 225, 229, 270, 271, 412), clouds (176, 256, 272, 286, 287,

319), and precipitation (114, 174, 229, 241, 266, 273–275, 278, 279), leading to speculation on

their role in precipitation formation (156, 276, 280). However, the inherent difficulties associated

with studying airborne biological INPs (74, 292) have resulted in a lack of data regarding their

abundance in the troposphere, thus making their meteorological roles difficult to assess. The parameterization of biological INPs in numerical cloud models have produced equivocal results, with several studies concluding that the low number of biological INPs in the troposphere are too low to have more than a negligible influence on precipitation formation (291, 309, 312), whereas other studies have made the opposite conclusion (136, 311, 315). While these studies represent pioneering attempts at elucidating the role that microorganisms may play in weather, few have considered mixtures of bacterial, fungal and other biological or biogenic INPs in their simulations. In the few that do (291, 309), biological INP concentrations were inferred from models (236, 310) that obtained global bioaerosol estimations from studies (218, 220, 230, 431–

433) that relied on culture-dependent methodologies to quantify bioaerosols.

Indeed, a full understanding of the various types and sizes of aerosols that can atmospheric processes has yet to be achieved; in fact it is one of the largest uncertainties in climate projections produced by the IPCC (71, 434). Further, there is a lack of information on the environmental sources to the atmosphere, and no studies have systematically quantified the temporal and spatial variability of airborne biological INPs. This is of critical importance, as studies have shown that diurnal and seasonal changes in atmospheric bioaerosol concentrations can be highly variable (125, 219, 270, 435). Additionally, the flux behavior of biological INPs

163

from terrestrial surfaces to the MBL, and from the turbulent MBL up to an altitude where they

can interact with clouds, remains poorly understood. Despite the discovery of biological INPs

over half a century ago, only one study (272) has explicitly attempted to quantify biological INP

concentrations in free tropospheric clouds. The handful of other studies which have quantified

biological INP occurrence in clouds are limited to a single location (the Puy de Dome in France),

of which only a certain type of cloud (orographic) can be examined (176, 286, 287, 319). These gaps in knowledge, while mostly a result of the technical difficulties associated with conducting bioaerosol research, can also be credited to the disciplinary gap between the fields of microbiology and atmospheric sciences and paucity of collaboration studies on this topic.

Without a coordinated effort to merge research from these disciplines, the capacity to realistically model biological INP to examine their influence on precipitation is likely to remain elusive. As such, the work presented in this thesis aimed to investigate several of these deficiencies by using a combination of microbiological and meteorological approaches to address a widely debated question: do microorganisms have important roles in precipitation formation?

We began our investigations by working backwards, examining the potential product of biological ice nucleation in clouds—that is, precipitation that has reached the earth’s surface.

The results of this two-year study were the first to use meteorological, chemical, and microbiological data to identify the predictors of biological INP concentrations within freshly fallen precipitation. Contrary to expectations, the geographic regions that were inferred to be the major sources of biological INPs to precipitation in Louisiana, were not rich in crops (83, 95,

215, 270) or vegetation (125, 266, 271), and there were no significant correlations between biological INPs and many of the known genera of Ice+ Gammaproteobacteria, including

164

Pseudomonas (81, 91, 403) and Erwinia (95, 96, 107) species. Instead, the highest

concentrations of biological INPs observed in precipitation appeared to be sourced from high

northern latitudes and arid regions of the Asian continent (274), and biological INPs correlated

most significantly with genera of arid soil-dwelling taxa from the Bacteroidetes and Firmicutes

(274). While previously underappreciated as a source of biological INPs, several recent studies

have indicated that terrestrial environments in the high northern latitudes are sources for high

concentrations of aerosolized biological INPs (364, 412, 413). Additionally, given the high

amounts of bioaerosols that are aerosolized from the desert regions of the Asian and African

continents every year, it is likely that they are significant sources of biological INPs. Indeed,

Saharan and Asian dusts have previously been shown to be associated with significant sources of

microorganisms (165, 168, 224, 227–229) and biological INPs (229) in precipitation and air

samples that occur thousands of kilometers away. In fact, soil environments in general may be

significant sources of biological INPs, as more recent studies have indicated that fertile soils

contain high concentrations of fungal INPs (23, 133), as well as unidentified biological INPs (24,

259). Taken together, these results support the contention that arid and semi-arid regions may be

significant sources of novel IN-active microorganisms.

Although desert regions may be significant sources of biological INPs, no studies have

attempted to identify the species of bacteria or fungi that produce biological INPs in these soil

ecosystems. As such, this study attempted to culture IN-active bacteria from desert

environments, and identified several novel species that possess IN activity following incubation

in carbon or nitrogen-limited conditions for after several weeks at 4oC. The fact that the bacteria

identified as IN-active in this study were tolerant to the UV-radiation, nutrient deprivation, and cold-conditioning, highlights the need for more specialized enrichment and induction culturing

165

techniques to identify novel IN-active species from the environment. Further, the repeatability of inducing the Ice+ phenotype in these species became difficult after subsequent rounds of experiments, which is a phenomenon reported by other researchers, and underscores the inherent difficulties associated with making a positive identification of Ice+ bacterial species. While this

phenomenon is not easy to explain, it may be that our manipulations in the laboratory to induce

IN activity cannot accurately replace the natural culture aging process that initially lead to the

induction of IN activity. The limitations of using conventional media in in vitro experiments are

likely to lead to severe underestimations of the number of bacterial and/or biological INPs that

exist in the environment, as conventional media does not limit nutrients.

An important limitation of relying on culturing alone to identify an environmental phenotype is that 99.9% of microorganisms have not culturable been cultured in the laboratory

(436). While culture-independent techniques have been attempted, such as using PCR with ina gene primers to estimate the number of Ice+ bacteria in the environment (266, 270), this

approach is biased to the limited number of species which possess the ina gene. Further, the

existing ina gene primers do not effectively cover all existing Ice+ bacteria. As such, other

microbiological techniques should be considered. For example, one could create a fosmid library

from the digested DNA of an environmental sample possessing high concentrations of biological

INPs and use heterologous expression in E. coli to identify novel gene products that serve as

efficient ice nucleators. A method such as this could allow for the identification of novel IN

activity in microorganisms that may not be culturable.

The data presented herein displayed very warm IN activity in a number of bacterial

species that were previously not known to possess this phenotype. Further, while we were not

successful in inducing high IN activity in all of the species that were tested in this study, it is

166

interesting to note that several of the same genera with IN activity came from very different environmental sources. Indeed, species of Kocuria, Paenibacillus, and Arthrobacter were

identified as potentially IN-active in both the Mojave Desert soil samples, as well as

precipitation collected in Blacksburg, VA. Additionally, these genera are commonly observed in

air, snow, ice, and cloud water samples. While the collections of Mojave Desert soil samples and

Virginia precipitation samples did not occur concurrently, this information indicates a potential

connection between IN-active species from soil environments and deposition in precipitation at

distance sites. Indeed, given the Westerly flow of tropospheric air over the coterminous United

States, it is more than likely that dusts sourced from the American southwest, as well as the

Asian deserts and arid regions, would at some point occur in precipitation to the East. The

frequent occurrence of precipitation in Virginia would increase the likelihood of IN-active

microorganisms returning to the surface through ice-nucleation and subsequent wet deposition.

Such a mechanism of microbial dispersal is not unlikely, and in fact, the preferential deposition

of IN-active bacteria in precipitation has been demonstrated (278).

The movement of microorganisms through the free troposphere raises concerns over the

ability to control the potential spread of plant, animal, and human disease. Indeed, many studies

have shown evidence for the global dispersal of microorganisms, however, none of them

explicitly examine a microorganism’s ability to return to the surface once suspended in the free

troposphere. Several studies have suggested that aerosol particles with aerodynamic diameters

between 0.1-10um that are suspended between ~1.5 km to the mid-troposphere are most

efficiently removed by precipitation (2, 437–439). Aerosols which reach higher altitudes, such as

the upper troposphere or stratosphere, may have significantly longer residence times (10-1000

days, respectively) than those that are able to be removed by wet deposition (2, 439, 440). As

167

such, it is possible that possessing a phenotype that ensures removal by precipitation from the

atmosphere, where harsh conditions such as ultraviolet radiation, low relative humidity, and cold

temperatures abound, would lend a selective advantage to the microorganisms that possess it.

Such a mechanism for microbial dispersal is yet another aspect of this research that needs to be

investigated more closely.

The occurrence of specific microorganisms in wet deposition may be highly dependent

on their ability to effect precipitation generation within a cloud. To understand their role in

precipitation generation, a closer look at their influence on the microphysical level is needed. a

recent study attempted to examine the microphysical influence of biological INPs using

laboratory-derived data on the IN-active fractions of P. syringae to make empirical

parameterizations for the cloud model (314). While this study outlined the most detailed

biological INP parameterization to date, it relied solely on the average active fraction of

laboratory-cultured P. syringae cells and their estimated occurrence in the air. Therefore, it did

not consider the contribution of IN activity from other bacterial species or other biological

sources. Further, the induction of IN activity in P. syringae under specific conditions can lead to

every single cell in the population being IN active (103)—on the other hand, there are

environmental conditions under which no cells are IN active. However, the physiological and

molecular mechanisms, together with biochemical interactions between the cell, ice, and water

are not well characterized. Further, while the inclusion of such a detailed process may be possible in a high-resolution cloud resolving model, it would likely be computationally exhaustive in larger models.

Indeed, the microphysical interactions between aerosols and clouds are complex. Thus, an important feature of any numerical atmospheric model that includes heterogeneous ice

168

nucleation is its ability to parameterize it on a scale that is not computationally exhaustive (296).

In general, most research and operational weather models do not specifically diagnose the

various categories of INPs a priori. Instead, parameterizations that can quickly and

inexpensively derive INP concentrations from aerosol datasets produced on a daily basis, such as

The Goddard Chemistry Aerosol Radiation and Transport (GOCART) model, are favored. One

of the most widely used parameterizations estimates global INP concentrations based on aerosols

with aerodynamic diameters > 0.5 μm (317). This scheme has demonstrated its accuracy at

enumerating concentrations of mineral dust INPs, which are active at temperatures colder than approximately -15oC, but underestimates INP concentrations active at warmer temperatures. This

flaw has not gone unnoticed by the developers of the parameterization, as they have explicitly

stated that the temperature regime warmer than -15oC still requires further evaluation, as the

nature of INPs active in this range are likely much more specialized (71). Indeed, many studies

have demonstrated that INPs active at temperatures warmer than -15oC are biological in origin

(90, 273–275). In fact, our data explicitly showed that the sizes and activity of biological INPs

active at -15oC and warmer can vary greatly between precipitation samples, and frequently occur

in sizes smaller than 0.1 μm (Figures 4-X and 4-X). This, in combination with the body of

literature that demonstrates the potential influence of nanoscale biological INPs on cloud

processes (288), emphasizes the need for a more specialized biological INP parameterization.

Additionally, there are no studies that have examined the effects that biological INPs may

have on different species of clouds. The results of chapter 2 indicated that compared to

convective clouds, the concentration of biological INPs in precipitation which had fallen from

nimbostratus or stratus-like clouds was significantly higher. This observation may be a direct

result of the differences in how precipitation forms in stratiform versus convective clouds.

169

Indeed, while a convective cloud can produce much of its precipitation through secondary ice

formation and ice enhancement mechanisms (Figure 1-2), stratiform clouds lack sufficient vertical movement to allow for those processes (9). As such, their ability to produce precipitation is likely dependent on the concentration of INPs present in the cloud. Further, the lack of vertical depth of stratiform clouds compared to convective clouds generally leads to warmer cloud temperatures, and the need for INPs which can nucleate ice at these warmer temperatures.

The simulated effects of biological INPs on stratiform precipitation development,

presented in this study, showed that INPs active at -4oC and colder were indeed necessary to

explain observed rates and amounts of precipitation. As expected, the INP parameterization

developed by DeMott et al. (317) did not allow for the production of precipitation in these

stratiform clouds. It should be emphasized, however, that these simulations are idealized, and

thus not wholly representative of what might occur in the troposphere. Nevertheless, they do

underscore the need for special attention to be paid to the effects that biological INPs likely have

on stratiform precipitation. While many studies have focused on the dynamics of convective

clouds, stratus and nimbostratus clouds are major producers of precipitation on a global scale and

significant modulators of global albedo (306–308).

The dynamics of cloud-aerosol-climate feedback are not well understood, and has been

cited as the area of highest uncertainty in future climate projections (63). Changes in landscapes

due to a warming climate will likely affect the number of aerosols released to the atmosphere,

thus affecting cloud properties. If arid and semi-arid soils are a significant source of biological

INPs, one must consider how an increase in their numbers to the atmosphere as a direct result of

desertification might introduce a negative rainfall and cloud cover feedback mechanism. During

glacier periods, dust levels are much higher in the atmosphere and conditions are more arid.

170

Indeed, our idealized models indicated that increases in biological INPs could decrease the amount of precipitation and cloud depth. Other studies have indicated that the introduction of too many INPs to the atmosphere could result in cloud and precipitation dissipation (55). Thus, in the polar regions, where stratiform cloud cover is significant, the increased upward flux of biological INPs could result in stratiform cloud dissipation, which in turn would amplify the rate of thawing that is already occurring in polar ice sheets as they become more exposed to solar radiation and higher temperatures in the absence of cloud cover. As such, the degree to which biological INPs affect cloud lifetimes and precipitation generation needs to be investigated with more scrutiny.

An understanding of the influence of biological INPs on clouds will continue to be limited by the lack of available information regarding the sources and types of biological INPs, as well as their spatial and temporal variability in the atmosphere. These data presented in this study took steps in the direction of closing these gaps in knowledge by attempting to define the most significant geographic sources of biological INPs in precipitation, attempting to culture novel Ice+ species from those environments, and characterizing the meteorological and seasonal patterns of biological INP dissemination. Longitudinal studies such as these are needed in order to determine whether these patterns the similar in other locations or restricted simply to the southeastern United states.

Additionally, studies that focus on unearthing geographic sources of airborne biological

INPs, such as those carried out by Creamean et al (364, 412, 413), are excellent steps in the right direction. The design of empirical microphysical parameterizations for biological INPs, such as those made by Phillips et al (cite), are important, but require input from coordinated field campaigns that can enumerate biological INP concentrations in air, clouds, and precipitation,

171

such as those described in Creamean et al (229). The development of such a parameterization would be useful if it was designed in a way that did not increase computing costs so much that it is unable to be used in large-scale numerical models. For example, a parameterization such as the one produced by DeMott et. al. (317) would be invaluable for global and climate studies if it were refined for applications with bioaerosols.

The data presented herein provides a framework upon which future studies can build. For example the potential for IN-active microorganisms to travel long distances through the atmosphere and remove themselves through the initiation of precipitation begs the question of whether the IN phenotype can play a role in the evolution of airborne microorganisms. Further, it is not unrealistic to ponder over how the microbial diversity of distant ecosystems could be influenced by the introduction of microorganisms from distant locations. The dissemination of plant pathogens through rain could is likely a worldwide phenomenon. As such, perhaps the dispersal of pathogens of other organisms, such as humans or animals, should be considered.

From a meteorological perspective, there is still much to be learned. The most logical next step in characterizing the significance of biological INPs on meteorology would be to include them in more advanced numerical models, such as those produced by WRF’s “real.exe” program. From there, one might better be able to understand how much of an influence biological INPs have on stratiform precipitation on a global scale. A clearer understanding of their role in clouds can inform climate projection models. If biological INPs have a notable influence on cloud microphysics, then there is potential for them to be harnessed in a way that promotes global cooling through cloud cover, or even precipitation formation in drought-stricken regions.

Regardless of the direction in which this research goes, we may be able to say, for now, that biological INPs are likely much more important for precipitation formation than previously

172

thought. Thus, the idea of bioprecipitation, proposed over 40 years ago by David Sands, may now have enough evidence to suggest that it is more than just an idea.

173

APPENDIX A SUPPLEMENTAL INFORMATION FOR CHAPTER 2

Table A-1. Results of Multiple Imputation for missing INP data. Missing data column lists the variables which did not contain an observation. All missing data followed a monotone missing data pattern. Missing data Sample Missing Data Pattern Number of Relative Pr > |t|c Size Frequencya Imputations Efficiencyb Performed total INPs ≤−13°C 4 6.56% 5 0.98 <0.0001 ≤−14°C 8 13.11% 5 0.92 <0.0001 −15°C 3 4.92% 5 0.92 <0.0001 biological INPs ≤−13°C 6 9.84% 5 0.95 <0.0001 ≤−14°C 6 9.84% 5 0.93 <0.0001 −15°C 4 6.56% 5 0.89 <0.0001 bacterial INPs ≤−13°C 3 13.46% 5 0.98 <0.0001 ≤−14°C 7 13.46% 5 0.92 <0.0001 −15°C 7 5.77% 5 0.95 <0.0001 aThe percent of observations containing the specified missing data bMeasure of how well the imputation calculations converged, as described in Li et al., JAMA 314:1966–1967, 2015. c P-value for t-test of H0: mean=0

Table A-2. MANOVA results of INP concentrations, interactions of air masses, and ecoregions. PM, Pacific Maritime; NAM, North Atlantic Maritime; SAM, South Atlantic Maritime; NFM, Northwest Forested Mountains; DSAH, Desert and Semi-Arid Highlands; HNL, High Northern Latitudes; GP, Great Plains; EWW, Eastern Woodlands and Wetlands; EA, East Asia. INP class Ecoregions of significance ANOVA test results Highest INP Lowest INP Prob > F concentrationsa concentrationsa total−5 to −11 EA, HNL NAM, SAM P < 0.0001 NAM, SAM, total−11 to −14 EA P < 0.001 EWW, DSAH bio−5 to −10 EA, HNL NAM, SAM P < 0.0001 NAM, SAM, bio−13 to −14 EA P < 0.0001 EWW, DSAH bio−11 to −12 NFM NAM P < 0.05 SAM, EWW, bac−5 to −10 EA, HNL P < 0.001 DSAH

174

Table A-3. Characteristics of fluorescent dissolved organic matter PARAFAC components in precipitation from air masses interacting with distinct ecoregions. Ecoregions are based on Level 1 Ecoregions defined by the EPA and CEC. The PARAFAC component means are shown as Raman Units. Numbers following ecoregion name correspond to numbers listed in the ecoregion column of Supplementary Dataset S1. Ecoregion Avg. Fluorescence Intensity Maximum C1: 0.0035 Pacific Maritime (1) C2: 0.0072 C3: 0.0017 N/A North Atlantic Maritime (3)

C1: 0.0036 South Atlantic Maritime (4) C2: 0.0062 C3: 0.0040 C1: 0.0041 Northwest Forested Mountains (5) C2: 0.0065 C3: 0.0035 C1: 0.0043 Desert and Semi-Aric Highlands (6) C2: 0.0066 C3: 0.0037 C1: 0.0071 High Northern Latitudes (7) C2: 0.0083 C3: 0.0032 C1: 0.0035 Great Plains (8) C2: 0.0053 C3: 0.0015 C1: 0.0041 Eastern Woodlands and Wetlands (9) C2: 0.0062 C3: 0.0036 C1: 0.0071 East Asia (11) C2: 0.0083 C3: 0.0032

175

Table A-4. Results of multivariate analysis of variance (MANOVA) for all INP (total, biological, and bacterial) concentrations as a function of season, cloud type, and precipitation type.

Meteorological Parameter MANOVA Test Results Season F(15, 122) = 3.12 p = 0.0003 Cloud Type F(5, 46) = 2.47 p=0.0462 Precipitation Type F(5, 46) = 2.45 p=0.0472 Bolded p-values indicate statistically significant differences between air masses tested at alpha=0.05 level Whole Model, F Test, Prob>F

Table A-5. Correlations between ice nucleating particle (INP) factors and local meteorological conditions. Pearson correlation coefficients (r) calculated between INP factors and locally recorded meteorological data. Relative humidity (RH%). Significance levels of Pearson correlation coefficients: *p < .05, **p < .01, ***p < .001. Cloud top temperature, surface temperature, surface wind speed N=61; Relative humidity, rain amount N=60. Cloud top Surface Rain Amount Surface wind INP Factor temperature Temperature RH % (mm h−) speed (mph) (oC) (oC) Total total−5 to −11 .07 -.53*** -.24 .06 .45*** total−11 to −14 .01 .30* -.08 -.18 -.17 Biological bio−5 to −10 .06 -.53*** -.24 .05 .45*** bio−13 to −14 -.05 -.01 -.03 -.07 -.12 bio−11 to −12 .18 -0.29* -.005 -.14 .04 Bacterial bac−5 to −10 -.01 -0.51*** -.30* -.12 .32*

176

Table A-6. Significant Spearman’s rank correlation coefficients (for which rho ρ≥0.40; and significance p<0.05) between ice nucleating particle (INP) factors and taxon abundance. Significance levels of Spearman’s rank correlation coefficients: *p < .05, **p < .01, ***p < .001. Only taxa with relative abundance >0.1% for total number of sequence reads across all precipitation events were analyzed. Total number of sequence reads across all precipitation events are listed in last column

Taxon (Order; total−5 total−11 to bio−5 to −10 bio−11 to bio−13 bac−5 to No. Family; to −11 −14 −12 to −14 −10 Sequenc Genus)umn1 e Reads Spear Spearma Spearman’ Spear Spear Spearma man’s n’s rho s rho (ρ) man’s man’s n’s rho rho (ρ) (ρ) rho (ρ) rho (ρ) (ρ) Acidobacteria Acidobacteriales; -- -- 0.41* ------5674 Acidobacteriacea e; Candidatus Chloracidobacter ium Bacteroidetes Bacteroidales; 0.62** 0.43* 0.59*** -- 0.55* 0.54** 81 Rikenellaceae; * ** N/A Sphingobacterial 0.51** -- 0.53** ------67269 es; N/A; N/A Cytophagales; -- 0.40* ------514 Cyclobacteriacea e; N/A Cytophagales; ------0.45* -- -- 69 Cyclobacteriacea e; Algoriphagus Cytophagales; 0.51** -- 0.55*** -- -- 0.43* 28658 Cytophagaceae; N/A Cytophagales; 0.56** 0.42* 0.59*** -- -- 0.48** 19161 Hymenobacterace * ae; Hymenobacter Cytophagales; -- -- 0.41* ------4486 Cytophagaceae; Spirosoma Chitinophagales; 0.55** 0.45* 0.53** -- 0.44* 0.55** 294 Chitinophagacea * e; Segetibacter

177

Table A-6. Continued. Taxon (Order; total−5 total−11 to bio−5 to −10 bio−11 to bio−13 bac−5 to No. Family; to −11 −14 −12 to −14 −10 Sequenc Genus)umn1 e Reads Sphingobacterial 0.52** 0.45* 0.44* -- 0.52* 0.46* 672 es; env.OPS_17; * N/A Cytophagales; 0.42* 0.53** 0.44* 0.48** -- -- 277 Cytophagaceae; Flexibacter Sphingobacterial 0.41* -- 0.40* ------20330 es; Sphingobacteriac eae; N/A Candidate ------0.42* -- 81 Division TM6 Chlorobi Chlorobiales; 0.42* ------144 N/A; N/A Cyanobacteria -- -- 0.40* ------707954 Firmicutes Bacillales; -- 0.53** ------888 Planococcaceae; Planococcus Bacillales; -- 0.57*** -- -- 0.41* -- 271 Staphylococcacea e; Macrococcus Clostridiales; 0.43* 0.56*** 0.43* ------75 Lachnospiraceae; Blautia Clostridiales; -- 0.41* ------53 Lachnospiraceae; Roseburia Erysipelotrichale -- 0.47** ------679 s; Erysipelotrichace ae; Turicibacter Lactobacillales; 0.47** 0.49** 0.46* -- -- 0.40* 642 Carnobacteriacea e; N/A Lactobacillales; 0.41* 0.48** 0.43* ------157 Carnobacteriacea e; Carnobacterium

178

Table A-6. Continued. Taxon (Order; total− total−11 to bio−5 to −10 bio−11 to bio−13 bac−5 to No. Family; 5 to −11 −14 −12 to −14 −10 Sequenc Genus)umn1 e Reads Lactobacillales; -- -- 0.40* ------91 Carnobacteriaceae; Desemzia Lactobacillales; -- 0.42* ------16067 Lactobacillaceae; N/A Lactobacillales; ------0.47* -- 457 Leuconostocaceae; * Leuconostoc Planctomycetes

Planctomycetales; -- 0.47** ------105 Planctomycetaceae ; Planctomyces Proteobacteria

Campylobacterales -- 0.45* -- 0.43* -- -- 300 ; Campylobacteracea e; Arcobacter Bacteriovoracales; -- 0.45* ------770 Bacteriovoraceae; N/A Bacteriovoracales; -- 0.47** ------730 Bacteriovoraceae; Peredibacter Rhizobiales; ------0.45* -- -- 523 Methylocystaceae; Methylosinus Sphingomonadales; ------0.41* -- 1190 Erythrobactereacea e; N/A Oceanospirillales; -- 0.44* -- -- 0.42* -- 66 Oceanospirillaceae ; N/A Pseudomonadales; -- 0.41* ------2013 Moraxellaceae; Perlucidibaca

179

Table A-6. Continued. Taxon (Order; total−5 total−11 to bio−5 to −10 bio−11 to bio−13 bac−5 to No. Family; to −11 −14 −12 to −14 −10 Sequenc Genus)umn1 e Reads Burkholderiales; -- -- 0.42* 108 Comamonadacea e; Polaromonas Chromatiales; -- 0.50* ------912 Chromatiaceae; Rheinheimera Rhodocyclales; 0.47** ------0.52* 0.48* 8105 Rhodocyclaceae; * N/A Rhodospirillales; ------0.40* -- 1094 wr0007; N/A Xanthomonadales 0.47** 0.44* 0.49** ------14094 ; Xanthomonadace ae; N/A Spirochaetes Spirochaetales; -- 0.41* -- -- 0.43* -- 93 N/A; N/A Verrucomicrobia Chthoniobacteral ------0.42* -- 510 es; Chthoniobacterac eae; Chthoniobacter Unclassified Unclassified; ------0.40* 425 OTU20 Unclassified; ------0.42* 175 OTU32 Unclassified; ------0.43* -- 1020 OTU43 Unclassified; ------0.43* -- -- 1051 OTU51 Unclassified; 0.52** 0.45* 0.44* -- 0.52* 0.46* 351 OTU73 * Unclassified; ------0.42* -- -- 672 OTU74 Unclassified; ------0.44* -- -- 254 OTU88 Unclassified; ------0.46* -- 120 OTU108

180

Table A-6. Continued. Taxon (Order; total−5 total−11 to bio−5 to −10 bio−11 to bio−13 bac−5 to No. Family; to −11 −14 −12 to −14 −10 Sequenc Genus)umn1 e Reads Unclassified; 0.47* -- 0.52** ------232 OTU13 Unclassified; ------0.43* -- 146 OTU18 Unclassified; 0.44* -- 0.51** ------110 OTU22

Table A-7. Spearman correlations between ice nucleating particle (INP) factors. Correlations calculated for differential concentrations of INPs between factor groupings. Significance levels of Pearson correlation coefficients (top number in each cell) and Spearman’s rho (bottom number in each cell): *p < .05, **p < .01, ***p < .001.

total−5 to −11 total−11 to −14 bio−5 to −10 bio−13 to −14 bio−11 to −12 bac−5 to −10

total−5 to −11 ------.08 total−11 to −14 ------.03 .96*** .20 bio−5 to −10 ------.94*** .16 -.01 .47** .00 bio−13 to −14 ------.07 .62*** .01 .04 .53** .00 .00 bio−11 to −12 -- -- -.02 .55*** -.05 .06 .83*** .26 .84*** .06 .10 bac−5 to −10 -- .83*** .32* .83*** .07 .15

181

Figure A-1. Cloud formation mechanisms and HYSPLIT trajectory analysis. The mechanism of lifting for each precipitation event is important in determining the altitudes for backward trajectory analysis. There are five general mechanisms of large-scale air movement that lead to cloud formation, which are dealt with as described in the supplemental methods. (a) Convection (b) Convergence (c) Warm front lifting (d) Cold front lifting. Orographic lifting was not observed in this study and is not depicted. Panels e and f show an example of how trajectories were analyzed, and are described in detail in the supplemental methods. Trajectory and ecoregion interactions are listed in Supplementary Dataset S1.

182

Figure A-2. PARAFAC Components fluorescence intensity profiles based on The North American Ecoregion classifications used in this study. Average Fluorescence Intensity was calculated based on ecoregion and is plotted on the y-axis in Raman Units (R.U.). PARAFAC Components C1-C3 are plotted as categories on the x-axis.

Figure A-3. Significant differences in DNA operational taxonomic unit (OTU) abundances as a function of cloud type and season. The mean number of sequence reads for each OTU is plotted, with bars indicating the standard error of the mean. Each taxon is represented by a single unique OTU. Top: OTUs that correlated with ice nucleating particle (INP) concentrations and had significantly different abundances in precipitation from stratiform (N=10) and convective (N=35) cloud formations. Bottom: OTUs that correlated with INP concentrations and had significantly different

183

abundances based on season (Autumn, N=12; Spring, N=6; Summer, N=14; Winter, N=13).

184

APPENDIX B SUPPLEMENTARY INFORMATION FOR CHAPTER 3

The media used for initial UVC enrichments from environmental samples was supplemented with 100 μg/mL of cycloheximide to inhibit fungal growth, and included the following: Tryptic Soy Agar (15 g/L Casein peptone, 5 g/L Soya peptone, 5 g/L Sodium

Chloride, 15 g/L Agar, pH 7.3 +/- 0.2), Reasoner’s 2 Agar (0.5 g/L Casamino acids 0.5 g/L

Proteose peptone, 0.5 g/L Yeast extract, 0.5 g/L Dextrose, 0.5 g/L soluble starch, 0.3 g/L

Dipotassium phosphate, 0.05 g/L Magnesium sulfate, 0.3 g/L Sodium pyruvate, 15 g/L Agar),

Nutrient Agar (1 g/L Meat extract, 2 g/L Yeast Extract, 5 g/L Peptone, 5 g/L Sodium Chloride,

15 g/L Agar, pH 7.4 +/- 0.2), and Marine Agar (5 g/L Peptone, 1 g/L Yeast extract, 0.1 g/L

Ferric Citrate, 19.45 g/L Sodium Chloride, 8.8 g/L Magnesium chloride, 3.24 g/L Sodium sulfate, 1.8 g/L Calcium chloride, 0.55 g/L Potassium chloride, 0.16 g/L Sodium bicarbonate,

0.08 g/L potassium bromide, 34 mg/L Strontium chloride, 22 mg/L boric acid, 4 mg/L sodium silicate, 2.4 mg/L sodium fluoride, 1.6 mg/L ammonium nitrate, 8 mg/L disodium phosphate, 15 g/L agar.

Vitamin stocks used in minimal media consisted of the following: 10 mg/L p- aminobenzoic acid; 10 mg/L Nicotinic acid; 10 mg/L Ca-pantothenate; 10 mg/L Pyridoxin

(Vitamin B6); 10 mg/L Riboflavin (Vitamin B12); 10 mg/L Thiamine (Vitamin B1); 5 mg/L

Biotin (Vitamin H); 5 mg/L Folic acid; 5 mg/L a-Lipoic acid; 5 mg/L Cyanobalamin (Vitamin

B12).

185

Figure B-1. Results of preliminary nutrient deprivation experiments for isolate AZ_122. Incubations were only done at 4oC. Top left graph shows INP concentrations at 0 days; Top right graph shows INP concentrations at 7 days; Bottom left graph shows INP concentrations at 14 days; And bottom right graph shows INP concentrations at 21 days.

Figure B-2. Results of preliminary nutrient deprivation experiments for isolate AZ_8. Incubations for this isolate were only done at 4oC. This is the INP concentrations detected after 21 days of incubation. All other time points had zero INPs.

186

LIST OF REFERENCES

1. Hurle DTJ. 1993. Nucleation theory. Handbook of Crystal Growth, 2nd ed. Elsevier.

2. Pruppacher HR, Klett JD. 2010. Microphysics of Clouds and Precipitation, 2nd ed. Springer Netherlands.

3. Volmer M, Weber Α. 1926. Keimbildung in übersättigten Gebilden. Zeitschrift für Physikalische Chemie 119U:277–301.

4. Becker R, Döring W. 1935. Kinetische Behandlung der Keimbildung in übersättigten Dämpfen. Annalen der Physik 416:719–752.

5. Frenkel J. 1939. A General Theory of Heterophase Fluctuations and Pretransition Phenomena. J Chem Phys 7:538–547.

6. Lin N, Liu XY. 2015. Correlation between hierarchical structure of crystal networks and macroscopic performance of mesoscopic soft materials and engineering principles. Chem Soc Rev 44:7881–7915.

7. Zhang Z, Liu X-Y. 2018. Control of ice nucleation: freezing and antifreeze strategies. Chemical Society Reviews.

8. Ickes L, Welti A, Hoose C, Lohmann U. 2014. Classical nucleation theory of homogeneous freezing of water: thermodynamic and kinetic parameters. Phys Chem Chem Phys 17.

9. Houze Jr. RA. 2014. Cloud Dynamics. Academic Press.

10. Squires P. 1958. The Microstructure and Colloidal Stability of Warm Clouds. Tellus 10:256–261.

11. Squires P. 1958. The Microstructure and Colloidal Stability of Warm Clouds. Tellus 10:262–271.

12. Chuang PY, Charlson RJ, Seinfeld JH. 1997. Kinetic limitations on droplet formation in clouds. Nature 390:594–596.

13. Li T, Donadio D, Russo G, Galli G. 2011. Homogeneous ice nucleation from supercooled water. Physical Chemistry Chemical Physics 13:19807–19813.

14. Kanji ZA, Ladino LA, Wex H, Boose Y, Burkert-Kohn M, Cziczo DJ, Krämer M. 2017. Overview of Ice Nucleating Particles. Meteorological Monographs 58:1.1-1.33.

15. Vali G, DeMott PJ, Möhler O, Whale TF. 2015. Technical Note: A proposal for ice nucleation terminology. Atmos Chem Phys 15:10263–10270.

187

16. Archuleta CM, DeMott PJ, Kreidenweis SM. 2005. Ice nucleation by surrogates for atmospheric mineral dust and mineral dust/sulfate particles at cirrus temperatures. Atmospheric Chemistry and Physics 5:2617–2634.

17. Kanji ZA, Abbatt JPD. 2010. Ice nucleation onto Arizona test dust at cirrus temperatures: effect of temperature and aerosol size on onset relative humidity. J Phys Chem A 114:935–941.

18. Welti A, Lüönd F, Stetzer O, Lohmann U. 2009. Influence of particle size on the ice nucleating ability of mineral dusts. Atmospheric Chemistry and Physics 9:6705–6715.

19. Mertes S, Verheggen B, Walter S, Connolly P, Ebert M, Schneider J, Bower KN, Cozic J, Weinbruch S, Baltensperger U, Weingartner E. 2007. Counterflow Virtual Impactor Based Collection of Small Ice Particles in Mixed-Phase Clouds for the Physico-Chemical Characterization of Tropospheric Ice Nuclei: Sampler Description and First Case Study. Aerosol Science and Technology 41:848–864.

20. Govindarajan AG, Lindow SE. 1988. Size of bacterial ice-nucleation sites measured in situ by radiation inactivation analysis. Proceedings of the National Academy of Sciences 85:1334–1338.

21. Hartmann S, Augustin S, Clauss T, Wex H, Šantl-Temkiv T, Voigtländer J, Niedermeier D, Stratmann F. 2013. Immersion freezing of ice nucleation active protein complexes. Atmospheric Chemistry and Physics 13:5751–5766.

22. Pummer BG, Bauer H, Bernardi J, Bleicher S, Grothe H. 2012. Suspendable macromolecules are responsible for ice nucleation activity of birch and conifer pollen. Atmos Chem Phys 12:2541–2550.

23. Fröhlich-Nowoisky J, Hill TCJ, Pummer BG, Yordanova P, Franc GD, Pöschl U. 2015. Ice nucleation activity in the widespread soil fungus Mortierella alpina. Biogeosciences 12:1057–1071.

24. O’Sullivan D, Murray BJ, Malkin TL, Whale TF, Umo NS, Atkinson JD, Price HC, Baustian KJ, Browse J, Webb ME. 2014. Ice nucleation by fertile soil dusts: relative importance of mineral and biogenic components. Atmos Chem Phys 14:1853–1867.

25. Lupi L, Hudait A, Molinero V. 2014. Heterogeneous Nucleation of Ice on Carbon Surfaces. J Am Chem Soc 136:3156–3164.

26. Vonnegut B. 1947. The Nucleation of Ice Formation by Silver Iodide. Journal of Applied Physics 18:593–595.

27. Kulkarni G, Nandasiri M, Zelenyuk A, Beranek J, Madaan N, Devaraj A, Shutthanandan V, Thevuthasan S, Varga T. 2015. Effects of crystallographic properties on the ice nucleation properties of volcanic ash particles. Geophysical Research Letters 42:3048– 3055.

188

28. Kanji ZA, Florea O, Abbatt JPD. 2008. Ice formation via deposition nucleation on mineral dust and organics: dependence of onset relative humidity on total particulate surface area. Environ Res Lett 3:025004.

29. Freedman MA. 2015. Potential Sites for Ice Nucleation on Aluminosilicate Clay Minerals and Related Materials. J Phys Chem Lett 6:3850–3858.

30. Sihvonen SK, Schill GP, Lyktey NA, Veghte DP, Tolbert MA, Freedman MA. 2014. Chemical and Physical Transformations of Aluminosilicate Clay Minerals Due to Acid Treatment and Consequences for Heterogeneous Ice Nucleation. J Phys Chem A 118:8787–8796.

31. Cantrell W, Heymsfield A. 2005. Production of Ice in Tropospheric Clouds: A Review. Bull Amer Meteor Soc 86:795–808.

32. Lau KM, Wu HT. 2003. Warm rain processes over tropical oceans and climate implications. Geophysical Research Letters 30.

33. Lohmann U, Feichter J. 2005. Global indirect aerosol effects: a review. Atmospheric Chemistry and Physics 5:715–737.

34. Hobbs PV, Chang S, Locatelli JD. 1974. The dimensions and aggregation of ice crystals in natural clouds. Journal of Geophysical Research (1896-1977) 79:2199–2206.

35. Rangno AL, Hobbs PV. 2001. Ice particles in stratiform clouds in the Arctic and possible mechanisms for the production of high ice concentrations. Journal of Geophysical Research: Atmospheres 106:15065–15075.

36. Hallett J, Mossop SC. 1974. Production of secondary ice particles during the riming process. Nature 249:26–28.

37. Chisnell RF, Latham J. 1976. Ice particle multiplication in cumulus clouds. Quarterly Journal of the Royal Meteorological Society 102:133–156.

38. Yano J-I, Phillips VTJ. 2010. Ice–Ice Collisions: An Ice Multiplication Process in Atmospheric Clouds. J Atmos Sci 68:322–333.

39. Field PR, Lawson RP, Brown PRA, Lloyd G, Westbrook C, Moisseev D, Miltenberger A, Nenes A, Blyth A, Choularton T, Connolly P, Buehl J, Crosier J, Cui Z, Dearden C, DeMott P, Flossmann A, Heymsfield A, Huang Y, Kalesse H, Kanji ZA, Korolev A, Kirchgaessner A, Lasher-Trapp S, Leisner T, McFarquhar G, Phillips V, Stith J, Sullivan S. 2016. Secondary Ice Production: Current State of the Science and Recommendations for the Future. Meteorological Monographs 58:7.1-7.20.

40. Takahashi T, Nagao Y, Kushiyama Y. 1995. Possible High Ice Particle Production during Graupel–Graupel Collisions. J Atmos Sci 52:4523–4527.

189

41. Vardiman L. 1978. The Generation of Secondary Ice Particles in Clouds by Crystal– Crystal Collision. J Atmos Sci 35:2168–2180.

42. Lauber A, Kiselev A, Pander T, Handmann P, Leisner T. 2018. Secondary Ice Formation during Freezing of Levitated Droplets. J Atmos Sci 75:2815–2826.

43. Phillips VTJ, Blyth AM, Brown PRA, Choularton TW, Latham J. 2001. The glaciation of a cumulus cloud over New Mexico. Q J Roy Meteor Soc 127:1513–1534.

44. Fridlind AM, Ackerman AS, McFarquhar G, Zhang G, Poellot MR, DeMott PJ, Prenni AJ, Heymsfield AJ. 2007. Ice properties of single-layer stratocumulus during the Mixed-Phase Arctic Cloud Experiment: 2. Model results. Journal of Geophysical Research: Atmospheres 112.

45. Lawson RP, Woods S, Morrison H. 2015. The Microphysics of Ice and Precipitation Development in Tropical Cumulus Clouds. J Atmos Sci 72:2429–2445.

46. Huang Y, Blyth AM, Brown PRA, Cotton R, Crosier J, Bower KN, Gallagher MW, Jones H, Gadian AM, Choularton TW, Cardwell J, Coe H, Mobbs SD, Hagen M. 2011. Development of ice particles in convective clouds observed over the Black Forest mountains during COPS. Quarterly Journal of the Royal Meteorological Society 137:275– 286.

47. Huang Y, Blyth AM, Brown PRA, Choularton TW, Connolly P, Gadian AM, Jones H, Latham J, Cui Z, Carslaw K. 2008. The development of ice in a cumulus cloud over southwest England. New J Phys 10:105021.

48. Connolly PJ, Choularton TW, Gallagher MW, Bower KN, Flynn MJ, Whiteway JA. 2006. Cloud-resolving simulations of intense tropical Hector thunderstorms: Implications for aerosol–cloud interactions. Quarterly Journal of the Royal Meteorological Society 132:3079–3106.

49. Dong Y, Oraltay RG, Hallett J. 1994. Ice particle generation during evaporation. Atmospheric Research 32:45–53.

50. Nagumo N, Fujiyoshi Y. 2015. Microphysical Properties of Slow-Falling and Fast-Falling Ice Pellets Formed by Freezing Associated with Evaporative Cooling. Mon Wea Rev 143:4376–4392.

51. Wegener A. 1911. Thermodynamik der atmosphäre. JA Barth.

52. Findeisen W, Peppler W, Bergeron T, Lange KO, Brooks CF. 1939. Outline of Physics of the Clouds. Bulletin of the American Meteorological Society 20:235–247.

53. Bergeron T. 1935. On the physics of cloud and precipitation. Proc 5th Assembly UGGI Lisbon 2:156.

190

54. Storelvmo T, Tan I. 2015. The Wegener-Bergeron-Findeisen process – Its discovery and vital importance for weather and climate. metz 24:455–461.

55. Korolev A. 2007. Limitations of the Wegener–Bergeron–Findeisen Mechanism in the Evolution of Mixed-Phase Clouds. J Atmos Sci 64:3372–3375.

56. Liou K-N. 1986. Influence of Cirrus Clouds on Weather and Climate Processes: A Global Perspective. Mon Wea Rev 114:1167–1199.

57. Storelvmo T, Hoose C, Eriksson P. 2011. Global modeling of mixed-phase clouds: The albedo and lifetime effects of aerosols. Journal of Geophysical Research: Atmospheres 116.

58. Kukla G, Karl TR. 1993. Nighttime warming and the greenhouse effect. Environ Sci Technol 27:1468–1474.

59. Platt CMR. 1989. The role of cloud microphysics in high-cloud feedback effects on climate change. Nature 341:428–429.

60. Tsushima Y, Ringer MA, Koshiro T, Kawai H, Roehrig R, Cole J, Watanabe M, Yokohata T, Bodas-Salcedo A, Williams KD, Webb MJ. 2016. Robustness, uncertainties, and emergent constraints in the radiative responses of stratocumulus cloud regimes to future warming. Clim Dyn 46:3025–3039.

61. Dickinson RE. 1989. Uncertainties of estimates of climatic change: A review. Climatic Change 15:5–13.

62. McCoy DT, Tan I, Hartmann DL, Zelinka MD, Storelvmo T. 2016. On the relationships among cloud cover, mixed-phase partitioning, and planetary albedo in GCMs. Journal of Advances in Modeling Earth Systems 8:650–668.

63. Boucher O, Randall D, Artaxo P, Bretherton C, Feingold G, Forster P, Kerminen V-M, Kondo Y, Liao H, Lohmann U. 2013. Clouds and aerosols, p. 571–657. In Climate change 2013: the physical science basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press.

64. Haywood J, Boucher O. 2000. Estimates of the direct and indirect radiative forcing due to tropospheric aerosols: A review. Reviews of Geophysics 38:513–543.

65. Mason BJ. 1950. The nature of ice-forming nuclei in the atmosphere. Quarterly Journal of the Royal Meteorological Society 76:59–74.

66. Mason BJ, Maybank J. 1958. Ice-nucleating properties of some natural mineral dusts. Quarterly Journal of the Royal Meteorological Society 84:235–241.

67. Mossop SC. 1963. Atmospheric ice nuclei. Zeitschrift für angewandte Mathematik und Physik ZAMP 14:456–486.

191

68. Rucklidge J. 1965. The Examination by Electron Microscope of Ice Crystal Nuclei from Cloud Chamber Experiments. J Atmos Sci 22:301–308.

69. Schaefer VJ. 1949. The Formation of Ice Crystals in the Laboratory and the Atmosphere. Chem Rev 44:291–320.

70. Schaefer VJ. 1949. The detection of ice nuclei in the free atmosphere. Journal of Meteorology 6:283–285.

71. DeMott P, Prenni A. 2010. New Directions: Need for defining the numbers and sources of biological aerosols acting as ice nuclei. Atmos Environ 44:1944–1945.

72. Mossop SC, Ono A. 1969. Measurements of Ice Crystal Concentration in Clouds. J Atmos Sci 26:130–137.

73. Hobbs PV, Rangno AL. 1985. Ice Particle Concentrations in Clouds. J Atmos Sci 42:2523–2549.

74. DeMott PJ, Möhler O, Stetzer O, Vali G, Levin Z, Petters MD, Murakami M, Leisner T, Bundke U, Klein H, Kanji ZA, Cotton R, Jones H, Benz S, Brinkmann M, Rzesanke D, Saathoff H, Nicolet M, Saito A, Nillius B, Bingemer H, Abbatt J, Ardon K, Ganor E, Georgakopoulos DG, Saunders C. 2011. Resurgence in Ice Nuclei Measurement Research. Bull Amer Meteor Soc 92:1623–1635.

75. Hoppe P, Arny DC, Martens JW. 1964. Frost susceptibility in corn increased by leaf blight infections. Plant Dis Rep 48.

76. Lee RE, Warren GJ, Gusta LV. 1995. Biological ice nucleation and its applications. APS Press.

77. Upper CD, Arny DC, Vojtik FC. 1972. Re-examination of relationship between Helminthosporium turicum infection and frost sensitivty of maize, p. 794–794. In Phytopathology. AMER PHYTOPATHOLOGICAL SOC 3340 PILOT KNOB ROAD, ST PAUL, MN 55121.

78. Schnell RC, Vali G. 1972. Atmospheric Ice Nuclei from Decomposing Vegetation. Nature 236:163–165.

79. Vali G. 1968. Ice nucleation relevant to the formation of hail. McGill University, Stormy Weather Group, Science Rep. MW-58, 51pp., 1985: Atmospheric ice nucleation—A review. J Rech Atmos 19:105–115.

80. Schnell RC, Vali G. 1973. World-wide Source of Leaf-derived Freezing Nuclei. Nature 246:212–213.

81. Maki LR, Galyan EL, Chang-Chien M-M, Caldwell DR. 1974. Ice Nucleation Induced by Pseudomonas syringae. Appl Environ Microbiol 28:456–459.

192

82. Vali G, Christensen M, Fresh RW, Galyan EL, Maki LR, Schnell RC. 1976. Biogenic Ice Nuclei. Part II: Bacterial Sources. J Atmos Sci 33:1565–1570.

83. Arny DC, Lindow SE, Upper CD. 1976. Frost sensitivity of Zea mays increased by application of Pseudomonas syringae. Nature 262:282–284.

84. Wolber PK, Deininger CA, Southworth MW, Vandekerckhove J, Montagu M van, Warren GJ. 1986. Identification and purification of a bacterial ice-nucleation protein. PNAS 83:7256–7260.

85. Green R, Warren G. 1985. Physical and functional repetition in a bacterial ice nucleation gene. Nature 317:645–648.

86. Green RL, Corotto LV, Warren GJ. 1988. Deletion mutagenesis of the ice nucleation gene from Pseudomonas syringae S203. Mol Gen Genet 215:165–172.

87. Lindow SE, Lahue E, Govindarajan AG, Panopoulos NJ, Gies D. 1989. Localization of ice nucleation activity and the iceC gene product in Pseudomonas syringae and Escherichia coli. Mol Plant Microbe In 2:262–272.

88. Mueller GM, Wolber PK, Warren GJ. 1990. Clustering of ice nucleation protein correlates with ice nucleation activity. Cryobiology 27:416–422.

89. Warren G, Corotto L. 1989. The consensus sequence of ice nucleation proteins from Erwinia herbicola, Pseudomonas fluorescens and Pseudomonas syringae. Gene 85:239– 242.

90. Maki LR, Willoughby KJ. 1978. Bacteria as Biogenic Sources of Freezing Nuclei. J Appl Meteor 17:1049–1053.

91. Corotto L v., Wolber P k., Warren G j. 1986. Ice nucleation activity of Pseudomonas fluorescens: mutagenesis, complementation analysis and identification of a gene product. The EMBO Journal 5:231–236.

92. Paulin JP, Luisetti J. 1978. Ice nucleation activity among phytopathogenic bacteria. Proceedings of the international conference on plant pathogenic bacteria, 4th.

93. Obata H, Nakai T, Tanishita J, Tokuyama T. 1989. Identification of an ice-nucleating bacterium and its ice nucleation properties. Journal of Fermentation and Bioengineering 67:143–147.

94. Wu Z, Qin L, Walker VK. 2009. Characterization and recombinant expression of a divergent ice nucleation protein from “Pseudomonas borealis.” Microbiology (Reading, Engl) 155:1164–1169.

95. Lindow SE, Arny DC, Upper CD. 1978. Erwinia herbicola: a bacterial ice nucleus active in increasing frost injury to corn. Phytopathology 68:523–527.

193

96. Abe K, Watabe S, Emori Y, Watanabe M, Arai S. 1989. An ice nucleation active gene of Erwinia ananas. FEBS Letters 258:297–300.

97. Kim HK, Orser C, Lindow SE, Sands DC. 1987. Xanthomonas campestris pv. translucens strains active in ice nucleation. Plant Dis 71:994–997.

98. Phelps P, Giddings TH, Prochoda M, Fall R. 1986. Release of cell-free ice nuclei by Erwinia herbicola. Journal of Bacteriology 167:496–502.

99. Kawahara H, Mano Y, Obata H. 1993. Purification and Characterization of Extracellular Ice-nucleating Matter from Erwinia uredovora KUIN-3. Bioscience, Biotechnology, and Biochemistry 57:1429–1432.

100. Michigami Y, Abe K, Iwabuchi K, Obata H, Arai S. 1995. Formation of Ice Nucleation- active Vesicles in Erwinia uredovora at Low Temperature and Transport of InaU Molecules into Shed Vesicles. Bioscience, Biotechnology, and Biochemistry 59:1996– 1998.

101. Garnham CP, Campbell RL, Walker VK, Davies PL. 2011. Novel dimeric β-helical model of an ice nucleation protein with bridged active sites. BMC Structural Biology 11:36.

102. Pandey R, Usui K, Livingstone RA, Fischer SA, Pfaendtner J, Backus EHG, Nagata Y, Fröhlich-Nowoisky J, Schmüser L, Mauri S, Scheel JF, Knopf DA, Pöschl U, Bonn M, Weidner T. 2016. Ice-nucleating bacteria control the order and dynamics of interfacial water. Sci Adv 2:e1501630.

103. Nemecek-Marshall M, LaDuca R, Fall R. 1993. High-level expression of ice nuclei in a Pseudomonas syringae strain is induced by nutrient limitation and low temperature. J Bacteriol 175:4062–4070.

104. Deininger CA, Mueller GM, Wolber PK. 1988. Immunological characterization of ice nucleation proteins from Pseudomonas syringae, Pseudomonas fluorescens, and Erwinia herbicola. J Bacteriol 170:669–675.

105. Orser C, Staskawicz BJ, Panopoulos NJ, Dahlbeck D, Lindow SE. 1985. Cloning and expression of bacterial ice nucleation genes in Escherichia coli. Journal of Bacteriology 164:359–366.

106. Vanderveer TL, Choi J, Miao D, Walker VK. 2014. Expression and localization of an ice nucleating protein from a soil bacterium, Pseudomonas borealis. Cryobiology 69:110–118.

107. Obata H, Tokuyama T, Kawate S, Hori H, Higashi Y. 1990. Culture Conditions of Erwinia uredovora in Reference to Its High Ice-nucleating Activity of the Culture Supernatant. Agricultural and Biological Chemistry 54:2171–2174.

108. Muryoi N, Matsukawa K, Yamade K, Kawahara H, Obata H. 2003. Purification and properties of an ice-nucleating protein from an ice-nucleating bacterium, Pantoea ananatis KUIN-3. Journal of Bioscience and Bioengineering 95:157–163.

194

109. Fall AL, Fall R. 1998. High-Level Expression of Ice Nuclei in Erwinia herbicola Is Induced by Phosphate Starvation and Low Temperature. Curr Microbiol 36:370–376.

110. Yankofsky SA, Bertold TN, Levin Z. 1983. Induction of latent freezing nucleus capability in an ice nucleation-active bacterium. Current Microbiology 9:263–267.

111. Turner MA, Arellano F, Kozloff LM. 1990. Three separate classes of bacterial ice nucleation structures. J Bacteriol 172:2521–2526.

112. Kozloff LM, Turner MA, Arellano F. 1991. Formation of bacterial membrane ice- nucleating lipoglycoprotein complexes. Journal of Bacteriology 173:6528–6536.

113. Turner MA, Arellano F, Kozloff LM. 1991. Components of ice nucleation structures of bacteria. Journal of Bacteriology 173:6515–6527.

114. Failor KC, Schmale DG, Vinatzer BA, Monteil CL. 2017. Ice nucleation active bacteria in precipitation are genetically diverse and nucleate ice by employing different mechanisms. ISME J 11:2740–2753.

115. Boelema BH. 1972. Bacterial blight (Pseudomonas pisi Sackett) of peas in South Africa, with special reference to frost as a predisposing factor. phd, Veenman, Wageningen.

116. Klement Z, Rozsnyay D, Arsenijevic M. 1974. Apoplexy of apricots. III. Relationship of wintr frost and the bacterial canker and die-back. AGRIS 35–45.

117. Lansade M. 1946. Recherches sur le chancre du Peuplier en France. Ann Epiphyt 12:23– 31.

118. Olien CR, Smith MN. 1981. Extension of localized freeze injury in barley by acute post- thaw bacterial disease. Cryobiology 18:404–409.

119. Panagopoulos CG, Crosse JE. 1964. Frost injury as a predisposing factor in blossom blight of pear caused by Pseudomonas syringae van Hall. Nature 202:1352–1352.

120. Sabet KA. 1953. Studies on the Bacterial Die-Back and Canker Disease of Poplar. Annals of Applied Biology 40:645–650.

121. Lindow SE, Arny DC, Upper CD. 1978. Distribution of ice nucleation-active bacteria on plants in nature. Appl Environ Microbiol 36:831–838.

122. Lindow SE. 1983. The Role of Bacterial Ice Nucleation in Frost Injury to Plants. Annu Rev Phytopathol 21:363–384.

123. Pouleur S, Richard C, Martin J-G, Antoun H. 1992. Ice Nucleation Activity in Fusarium acuminatum and Fusarium avenaceum. Appl Environ Microbiol 58:2960–2964.

124. Richard C, Martin J-G, Pouleur S. 1996. Ice nucleation activity identified in some phytopathogenic Fusarium species. phyto 77:83–92.

195

125. Huffman JA, Prenni AJ, DeMott PJ, Pöhlker C, Mason RH, Robinson NH, Fröhlich- Nowoisky J, Tobo Y, Després VR, Garcia E, Gochis DJ, Harris E, Müller-Germann I, Ruzene C, Schmer B, Sinha B, Day DA, Andreae MO, Jimenez JL, Gallagher M, Kreidenweis SM, Bertram AK, Pöschl U. 2013. High concentrations of biological aerosol particles and ice nuclei during and after rain. Atmospheric Chemistry and Physics 13:6151–6164.

126. Hasegawa Y, Ishihara Y, Tokuyama T. 1994. Characteristics of ice-nucleation activity in Fusarium avenaceum IFO 7158. Biosci Biotechnol Biochem 58:2273–2274.

127. Humphreys TL, Castrillo LA, Lee MR. 2001. Sensitivity of partially purified ice nucleation activity of Fusarium acuminatum SRSF 616. Curr Microbiol 42:330–338.

128. Lagzian M, Latifi AM, Bassami MR, Mirzaei M. 2014. An ice nucleation protein from Fusarium acuminatum: cloning, expression, biochemical characterization and computational modeling. Biotechnol Lett 36:2043–2051.

129. Kieft TL. 1988. Ice nucleation activity in lichens. Appl Environ Microbiol 54:1678–1681.

130. Kieft TL, Ahmadjian V. 1989. Biological ice nucleation activity in lichen mycobionts and photobionts. The Lichenologist 21:355–362.

131. Morris CE, Sands DC, Glaux C, Samsatly J, Asaad S, Moukahel AR, Gonçalves FLT, Bigg EK. 2013. Urediospores of rust fungi are ice nucleation active at > −10 °C and harbor ice nucleation active bacteria. Atmos Chem Phys 13:4223–4233.

132. Pummer BG, Budke C, Augustin-Bauditz S, Niedermeier D, Felgitsch L, Kampf CJ, Huber RG, Liedl KR, Loerting T, Moschen T, Schauperl M, Tollinger M, Morris CE, Wex H, Grothe H, Pöschl U, Koop T, Fröhlich-Nowoisky J. 2015. Ice nucleation by water- soluble macromolecules. Atmos Chem Phys 15:4077–4091.

133. Conen F, Yakutin MV. 2018. Soils rich in biological ice-nucleating particles abound in ice-nucleating macromolecules likely produced by fungi. Biogeosciences 15:4381–4385.

134. Worland MR, Lukešová A. 2000. The effect of feeding on specific soil algae on the cold- hardiness of two Antarctic micro-arthropods (Alaskozetes antarcticus and Cryptopygus antarcticus). Polar Biol 23:766–774.

135. Knopf DA, Alpert PA, Wang B, Aller JY. 2011. Stimulation of ice nucleation by marine diatoms. Nature Geosci 4:88–90.

136. Diehl K, Matthias-Maser S, Jaenicke R, Mitra SK. 2002. The ice nucleating ability of pollen:: Part II. Laboratory studies in immersion and contact freezing modes. Atmospheric Research 61:125–133.

137. Brush RA, Griffith M, Mlynarz A. 1994. Characterization and Quantification of Intrinsic Ice Nucleators in Winter Rye (Secale cereale) Leaves. Plant Physiol 104:725–735.

196

138. Krog JO, Zachariassen KE, Larsen B, Smidsrød O. 1979. Thermal buffering in Afro- alpine plants due to nucleating agent-induced water freezing. Nature 282:300–301.

139. Weber CF. 2016. Polytrichum commune spores nucleate ice and associated microorganisms increase the temperature of ice nucleation activity onset. Aerobiologia 32:353–361.

140. Lindow SE. 1979. Frost damage to potato reduced by bacteria antagonistic to ice nucleation-active bacteria, p. 1036–1036. In Phytopathology. Amer Phytopathological Soc.

141. Lindow SE. 1981. Frost damage to pear reduced by antagonistic bacteria, bactericides, and ice nucleation inhibitors, p. 237–237. In Phytopathology. Amer Phytopathological Soc 3.

142. Lindow SE. 1982. Integrated control of frost injury and fire blight of pear with antagonistic epiphytic bacteria, p. 946–946. In Phytopathology. Amer Phytopathological Soc.

143. Lindow SE. 1982. Population dynamics of epiphytic ice nucleation active bacteria on frost sensitive plants and frost control by means of antagonistic bacteria, p. 395–416. In Plant cold hardiness and freezing stress. Elsevier.

144. Lindow SE, Arny DC, Upper CD. 1977. Control of frost damage to corn in the field by a bacterium antagonistic to ice nucleation-active bacteria. Proceedings-American Phytopathological Society (USA).

145. Lindow SE, Arny DC, Upper CD. 1977. Antagonistic bacteria reduce bacterially incited frost damage to corn (Zea mays L.). Proceedings American Phytopathological Society.

146. Lindow SE, Arny DC, Upper CD. 1983. Biological control of frost injury: an isolate of Erwinia herbicola antagonistic to ice nucleation active bacteria. Phytopathology 73:1097– 1102.

147. Lindemann J. 1987. Competition Between Ice Nucleation-Active Wild Type and Ice Nucleation-Deficient Deletion Mutant Strains of Pseudomonas syringae and P. fluorescens Biovar I and Biological Control of Frost Injury on Strawberry Blossoms. Phytopathology 77:882.

148. Warren GJ. 1987. Bacterial Ice Nucleation: Molecular Biology and Applications. Biotechnology and Genetic Engineering Reviews 5:107–136.

149. Lindow SE, Arny DC, Barchet WR, Upper CD. 1978. Control of frost damage to plants in the field with bacterial ice nucleation inhibitors. Phytopathology News.

150. Kozloff LM, Schofield MA, Lute M. 1983. Ice nucleating activity of Pseudomonas syringae and Erwinia herbicola. Journal of bacteriology 153:222–231.

197

151. Drainas C, Vartholomatos G, Panopoulos NJ. 1995. The Ice Nucleation Gene from Pseudomonas syringae as a Sensitive Gene Reporter for Promoter Analysis in Zymomonas mobilis. Appl Environ Microbiol 61:273–277.

152. Arvanitis N, Vargas C, Tegos G, Perysinakis A, Nieto JJ, Ventosa A, Drainas C. 1995. Development of a gene reporter system in moderately halophilic bacteria by employing the ice nucleation gene of Pseudomonas syringae. Appl Environ Microbiol 61:3821–3825.

153. Arai S, Watanabe M. 1986. Freeze Texturing of Food Materials by Ice-nucleation with the Bacterium Erwinia ananas. Agricultural and Biological Chemistry 50:169–175.

154. Margaritis A, Bassi AS. 1991. Principles and biotechnological applications of bacterial ice nucleation. Critical reviews in biotechnology 11:277–295.

155. Ward PJ, DeMott PJ. 1989. Preliminary Experimental Evaluation of Snomax (TM) Snow Inducer, Nucleus Pseudomonas syringae, as an Artificial Ice for Weather Modification. The Journal of Weather Modification 21:9–13.

156. D.c S, V.e L, A.l S, G DS. 1982. The association between bacteria and rain and possible resultant meteorological implications. Idojaras.

157. Constantinidou HA, Hirano SS, Baker LS, Upper CD. 1990. Atmospheric dispersal of ice nucleation-active bacteria: the role of rain. Phytopathology 80:934–937.

158. Schaefer VJ. 1948. Experiments with Cloudy Skies. Weatherwise 1:3–18.

159. Leopold LB, Halstead MH. 1948. First Trials of the Schaefer-Langmuir Dry-Ice Cloud Seeding Technique in Hawaii. Bull Amer Meteor Soc 29:525–534.

160. Pepper IL, Gerba CP. 2015. Chapter 5 - Aeromicrobiology, p. 89–110. In Pepper, IL, Gerba, CP, Gentry, TJ (eds.), Environmental Microbiology (Third Edition). Academic Press, San Diego.

161. Comtois P. 2001. John Tyndall and the floating matter of the air. Aerobiologia 17:193– 202.

162. Wolff GT. 1980. Mesoscale and synoptic scale transport of aerosols. Annals of the New York Academy of Sciences 379–388.

163. Reche I, D’Orta G, Mladenov N, Winget DM, Suttle CA. 2018. Deposition rates of viruses and bacteria above the atmospheric boundary layer. ISME J 12:1154–1162.

164. Greenfield SM. 1957. Rain scavenging of radioactive particulate matter from the atmosphere. J Meteor 14:115–125.

165. Griffin DW, Kellogg CA, Garrison VH, Shinn EA. 2002. The Global Transport of Dust: An Intercontinental river of dust, microorganisms and toxic chemicals flows through the Earth’s atmosphere. Am Sci 90:228–235.

198

166. Kellogg CA, Griffin DW. 2006. Aerobiology and the global transport of desert dust. Trends Ecol Evol 21:638–644.

167. Smith DJ, Timonen HJ, Jaffe DA, Griffin DW, Birmele MN, Perry KD, Ward PD, Roberts MS. 2013. Intercontinental Dispersal of Bacteria and Archaea by Transpacific Winds. Appl Environ Microbiol 79:1134–1139.

168. Smith DJ, Jaffe DA, Birmele MN, Griffin DW, Schuerger AC, Hee J, Roberts MS. 2012. Free Tropospheric Transport of Microorganisms from Asia to North America. Microb Ecol 64:973–985.

169. Schlesinger P, Mamane Y, Grishkan I. 2006. Transport of microorganisms to Israel during Saharan dust events. Aerobiologia 22:259.

170. Griffin DW. 2007. Atmospheric movement of microorganisms in clouds of desert dust and implications for human health. Clin Microbiol Rev 20:459–477, table of contents.

171. Favet J, Lapanje A, Giongo A, Kennedy S, Aung Y-Y, Cattaneo A, Davis-Richardson AG, Brown CT, Kort R, Brumsack H-J, Schnetger B, Chappell A, Kroijenga J, Beck A, Schwibbert K, Mohamed AH, Kirchner T, de Quadros PD, Triplett EW, Broughton WJ, Gorbushina AA. 2013. Microbial hitchhikers on intercontinental dust: catching a lift in Chad. ISME J 7:850–867.

172. Lamichhane JR, Varvaro L, Parisi L, Audergon J-M, Morris CE. 2014. Chapter Four - Disease and Frost Damage of Woody Plants Caused by Pseudomonas syringae: Seeing the Forest for the Trees, p. 235–295. In Sparks, DL (ed.), Advances in Agronomy. Academic Press.

173. Mansfield J, Genin S, Magori S, Citovsky V, Sriariyanum M, Ronald P, Dow MAX, Verdier V, Beer SV, Machado MA. 2012. Top 10 plant pathogenic bacteria in molecular plant pathology. Molecular plant pathology 13:614–629.

174. Stopelli E, Conen F, Guilbaud C, Zopfi J, Alewell C, Morris CE. 2017. Ice nucleators, bacterial cells and Pseudomonas syringae in precipitation at Jungfraujoch. Biogeosciences 14:1189–1196.

175. Morris CE, Sands DC, Vinatzer BA, Glaux C, Guilbaud C, Buffière A, Yan S, Dominguez H, Thompson BM. 2008. The life history of the plant pathogen Pseudomonas syringae is linked to the water cycle. ISME J 2:321–334.

176. Joly M, Attard E, Sancelme M, Deguillaume L, Guilbaud C, Morris CE, Amato P, Delort A-M. 2013. Ice nucleation activity of bacteria isolated from cloud water. Atmos Environ 70:392–400.

177. Amato P, Ménager M, Sancelme M, Laj P, Mailhot G, Delort A-M. 2005. Microbial population in cloud water at the Puy de Dôme: Implications for the chemistry of clouds. Atmospheric Environment 39:4143–4153.

199

178. Brown JKM, Hovmøller MS. 2002. Aerial Dispersal of Pathogens on the Global and Continental Scales and Its Impact on Plant Disease. Science 297:537–541.

179. Schumann GL, D’Arcy CJ. 2006. Essential plant pathology. American Phytopathological Society (APS Press).

180. Petersen RH. 1974. The rust fungus life cycle. Bot Rev 40:453–513.

181. Fry WE, Goodwin SB. 1997. Re-emergence of Potato and Tomato Late Blight in the United States. Plant Disease 81:1349–1357.

182. Dyer A, Matusralt M, Drenth A, Cohen A, Splelman L. 1993. Historical and recent migrations of Phytophthora infestans: chronology, pathways, and implications. Plant disease 77:653–661.

183. Bowden J, Gregory PH, Johnson CG. 1971. Possible wind transport of coffee leaf rust across the Atlantic Ocean. Nature 229:500–501.

184. Purdy LJ, Krupa SV, Dean JL. 1985. Introduction of sugarcane rust into the Americas and its spread to Florida. Plant Disease 69:689–693.

185. Gloster J, Blackall RM, Sellers RF, Donaldson AI. 1981. Forecasting the airborne spread of foot-and-mouth disease. Vet Rec 108:370–374.

186. Gloster J, Sellers RF, Donaldson AI. 1982. Long distance transport of foot-and-mouth disease virus over the sea. Vet Rec 110:47–52.

187. Henderson RJ. 1969. The outbreak of foot-and-mouth disease in Worcestershire. An epidemiological study: with special reference to spread of the disease by wind-carriage of the virus. J Hyg (Lond) 67:21–33.

188. Sellers RF, Forman AJ. 1973. The Hampshire epidemic of foot-and-mouth disease, 1967. J Hyg (Lond) 71:15–34.

189. Tinline R. 1970. Lee wave hypothesis for the initial pattern of spread during the 1967-68 foot and mouth epizootic. Nature 227:860–862.

190. Sørensen JH, Mackay DK, Jensen CO, Donaldson AI. 2000. An integrated model to predict the atmospheric spread of foot-and-mouth disease virus. Epidemiol Infect 124:577–590.

191. A.i D, S A. 2002. Predicting the spread of foot and mouth disease by airborne virus 21:569–575.

192. Klausner Z, Fattal E, Klement E. 2017. Using Synoptic Systems’ Typical Wind Trajectories for the Analysis of Potential Atmospheric Long-Distance Dispersal of Lumpy Skin Disease Virus. Transboundary and Emerging Diseases 64:398–410.

200

193. Burgin LE, Gloster J, Sanders C, Mellor PS, Gubbins S, Carpenter S. 2013. Investigating Incursions of Bluetongue Virus Using a Model of Long-Distance Culicoides Biting Midge Dispersal. Transboundary and Emerging Diseases 60:263–272.

194. Otake S, Dee S, Corzo C, Oliveira S, Deen J. 2010. Long-distance airborne transport of infectious PRRSV and Mycoplasma hyopneumoniae from a swine population infected with multiple viral variants. Veterinary Microbiology 145:198–208.

195. Hammond GW, Raddatz RL, Gelskey DE. 1989. Impact of Atmospheric Dispersion and Transport of Viral Aerosols on the Epidemiology of Influenza. Rev Infect Dis 11:494– 497.

196. Klausner Z, Klement E, Fattal E. 2018. Source–receptor probability of atmospheric long- distance dispersal of viruses to Israel from the eastern Mediterranean area. Transboundary and Emerging Diseases 65:205–212.

197. Ma Y, Zhou J, Yang S, Zhao Y, Zheng X. 2017. Assessment for the impact of dust events on measles incidence in western China. Atmospheric Environment 157:1–9.

198. Ferrari MJ, Grais RF, Bharti N, Conlan AJK, Bjørnstad ON, Wolfson LJ, Guerin PJ, Djibo A, Grenfell BT. 2008. The dynamics of measles in sub-Saharan Africa. Nature 451:679– 684.

199. Brown EG, Gottlieb S, Laybourn RL. 1935. Dust Storms and Their Possible Effect on Health: With Special Reference to the Dust Storms in Kansas in 1935. Public Health Reports (1896-1970) 50:1369–1383.

200. Irving TJ, Blyuss KB, Colijn C, Trotter CL. 2012. Modelling meningococcal meningitis in the African meningitis belt. Epidemiology & Infection 140:897–905.

201. Roberts L. 2008. Infectious disease. An ill wind, bringing meningitis. Science 320:1710– 1715.

202. Meningococcal meningitis. https://www.who.int/news-room/fact- sheets/detail/meningococcal-meningitis

203. Sultan B, Labadi K, Guegan, JF, Janicot, S. 2005. Climate Drives the Meningitis Epidemics Onset in West Africa. PLOS Medicine 2: e6.

204. Perkins S. 2001. Dust, the Thermostat How tiny airborne particles manipulate global climate. Science News 160:200–2002.

205. Goudie AS, Middleton NJ. 2001. Saharan dust storms: nature and consequences. Earth- Science Reviews 56:179–204.

206. Moulin C, Lambert CE, Dulac F, Dayan U. 1997. Control of atmospheric export of dust from North Africa by the North Atlantic Oscillation. Nature 387:691–694.

201

207. Prospero JM. 2004. Interhemispheric Transport of Viable Fungi and Bacteria from Africa to the Caribbean with Soil Dust, p. 127–133. In Werner, D (ed.), Biological Resources and Migration. Springer, Berlin, Heidelberg.

208. Herman JR, Krotkov N, Celarier E, Larko D, Labow G. 1999. Distribution of UV radiation at the Earth’s surface from TOMS-measured UV-backscattered radiances. Journal of Geophysical Research: Atmospheres 104:12059–12076.

209. García-Pando CP, Stanton MC, Diggle PJ, Trzaska S, Miller RL, Perlwitz JP, Baldasano JM, Cuevas E, Ceccato P, Yaka P, Thomson MC. 2014. Soil Dust Aerosols and Wind as Predictors of Seasonal Meningitis Incidence in Niger. Environmental Health Perspectives 122:679–686.

210. Jusot J-F, Neill DR, Waters EM, Bangert M, Collins M, Bricio Moreno L, Lawan KG, Moussa MM, Dearing E, Everett DB, Collard J-M, Kadioglu A. 2017. Airborne dust and high temperatures are risk factors for invasive bacterial disease. J Allergy Clin Immunol 139:977-986.e2.

211. Habybabady RH, Sis HN, Paridokht F, Ramrudinasab F, Behmadi A, Khosravi B, Mohammadi M. 2018. Effects of Dust Exposure on the Respiratory Health Symptoms and Pulmonary Functions of Street Sweepers. Malays J Med Sci 25:76–84.

212. Kobayashi F, Iwata K, Maki T, Kakikawa M, Higashi T, Yamada M, Ichinose T, Iwasaka Y. 2016. Evaluation of the toxicity of a Kosa (Asian duststorm) event from view of food poisoning: observation of Kosa cloud behavior and real-time PCR analyses of Kosa bioaerosols during May 2011 in Kanazawa, Japan. Air Qual Atmos Health 9:3–14.

213. Rodó X, Ballester J, Cayan D, Melish ME, Nakamura Y, Uehara R, Burns JC. 2011. Association of Kawasaki disease with tropospheric wind patterns. Sci Rep 1:1–7.

214. Tong Y. 1999. Diurnal Distribution of Total and Culturable Atmospheric Bacteria at a Rural Site. Aerosol Science and Technology 30:246–254.

215. Lindemann J, Upper CD. 1985. Aerial Dispersal of Epiphytic Bacteria over Bean Plants. Appl Environ Microbiol 50:1229–1232.

216. Bovallius A, Bucht B, Roffey R, Anäs P. 1978. Three-year investigation of the natural airborne bacterial flora at four localities in sweden. Appl Environ Microbiol 35:847–852.

217. Lighthart B, Shaffer BT. 1995. Airborne Bacteria in the Atmospheric Surface Layer: Temporal Distribution above a Grass Seed Field. Appl Environ Microbiol 61:1492–1496.

218. Tong Y, Lighthart B. 2000. The Annual Bacterial Particle Concentration and Size Distribution in the Ambient Atmosphere in a Rural Area of the Willamette Valley, Oregon. Aerosol Science and Technology 32:393–403.

219. Fang Z, Ouyang Z, Zheng H, Wang X, Hu L. 2007. Culturable airborne bacteria in outdoor environments in Beijing,China. Microb Ecol 54:487–496.

202

220. Shaffer BT, Lighthart B. 1997. Survey of Culturable Airborne Bacteria at Four Diverse Locations in Oregon: Urban, Rural, Forest, and Coastal. Microb Ecol 34:167–177.

221. Lindemann J, Constantinidou HA, Barchet WR, Upper CD. 1982. Plants as Sources of Airborne Bacteria, Including Ice Nucleation-Active Bacteria. Appl Environ Microbiol 44:1059–1063.

222. Womack AM, Bohannan BJM, Green JL. 2010. Biodiversity and biogeography of the atmosphere. Philosophical Transactions of the Royal Society B: Biological Sciences 365:3645–3653.

223. Kellogg CA, Griffin DW, Garrison VH, Peak KK, Royall N, Smith RR, Shinn EA. 2004. Characterization of Aerosolized Bacteria and Fungi From Desert Dust Events in Mali, West Africa. Aerobiologia 20:99–110.

224. Griffin DW, Kellogg CA, Garrison VH, Lisle JT, Borden TC, Shinn EA. 2003. Atmospheric microbiology in the northern Caribbean during African dust events. Aerobiologia 19:143–157.

225. Prenni AJ, Petters MD, Kreidenweis SM, Heald CL, Martin ST, Artaxo P, Garland RM, Wollny AG, Pöschl U. 2009. Relative roles of biogenic emissions and Saharan dust as ice nuclei in the Amazon basin. Nature Geosci 2:402–405.

226. Rosselli R, Fiamma M, Deligios M, Pintus G, Pellizzaro G, Canu A, Duce P, Squartini A, Muresu R, Cappuccinelli P. 2015. Microbial immigration across the Mediterranean via airborne dust. Sci Rep 5.

227. Barberán A, Henley J, Fierer N, Casamayor EO. 2014. Structure, inter-annual recurrence, and global-scale connectivity of airborne microbial communities. Science of The Total Environment 487:187–195.

228. Meola M, Lazzaro A, Zeyer J. 2015. Bacterial Composition and Survival on Sahara Dust Particles Transported to the European Alps. Front Microbiol 6.

229. Creamean JM, Suski KJ, Rosenfeld D, Cazorla A, DeMott PJ, Sullivan RC, White AB, Ralph FM, Minnis P, Comstock JM, Tomlinson JM, Prather KA. 2013. Dust and Biological Aerosols from the Sahara and Asia Influence Precipitation in the Western U.S. Science 339:1572–1578.

230. Harrison RM, Jones AM, Biggins PDE, Pomeroy N, Cox CS, Kidd SP, Hobman JL, Brown NL, Beswick A. 2005. Climate factors influencing bacterial count in background air samples. Int J Biometeorol 49:167–178.

231. Bowers RM, Sullivan AP, Costello EK, Collett JL, Knight R, Fierer N. 2011. Sources of Bacteria in Outdoor Air across Cities in the Midwestern United States. Appl Environ Microbiol 77:6350–6356.

203

232. Bowers RM, Clements N, Emerson JB, Wiedinmyer C, Hannigan MP, Fierer N. 2013. Seasonal Variability in Bacterial and Fungal Diversity of the Near-Surface Atmosphere. Environ Sci Technol 47:12097–12106.

233. Carotenuto F, Georgiadis T, Gioli B, Leyronas C, Morris CE, Nardino M, Wohlfahrt G, Miglietta F. 2017. Measurements and modeling of surface–atmosphere exchange of microorganisms in Mediterranean grassland.

234. Bauer H, Kasper-Giebl A, Löflund M, Giebl H, Hitzenberger R, Zibuschka F, Puxbaum H. 2002. The contribution of bacteria and fungal spores to the organic carbon content of cloud water, precipitation and aerosols. Atmospheric Research 64:109–119.

235. Bowers RM, McLetchie S, Knight R, Fierer N. 2011. Spatial variability in airborne bacterial communities across land-use types and their relationship to the bacterial communities of potential source environments. ISME J 5:601–612.

236. Burrows SM, Elbert W, Lawrence MG, Pöschl U. 2009. Bacteria in the global atmosphere – Part 1: Review and synthesis of literature data for different ecosystems. Atmospheric Chemistry and Physics 9:9263–9280.

237. Aller JY, Kuznetsova MR, Jahns CJ, Kemp PF. 2005. The sea surface microlayer as a source of viral and bacterial enrichment in marine aerosols. Journal of Aerosol Science 36:801–812.

238. Hultin KAH, Krejci R, Pinhassi J, Gomez-Consarnau L, Mårtensson EM, Hagström Å, Nilsson ED. 2011. Aerosol and bacterial emissions from Baltic Seawater. Atmospheric Research 99:1–14.

239. Mayol E, Arrieta JM, Jiménez MA, Martínez-Asensio A, Garcias-Bonet N, Dachs J, González-Gaya B, Royer S-J, Benítez-Barrios VM, Fraile-Nuez E, Duarte CM. 2017. Long-range transport of airborne microbes over the global tropical and subtropical ocean. Nat Commun 8.

240. Mason PJ, Thomson DJ. 2015. Boundary Layer (Atmospheric) and Air Pollution: Overview. p. 220–226. In North, GR, Pyle, J, Zhang, F (eds.), Encyclopedia of Atmospheric Sciences (Second Edition). Academic Press, Oxford.

241. Aho KA, Weber CF, Christner BC, Vinatzer BA, Morris CE, Joyce R, Failor KC, Werth JT, Bayless-Edwards AL, Schmale III DG. 2019. Spatiotemporal patterns of microbial composition and diversity in precipitation. Ecological Monographs e01394.

242. Bryan NC, Christner BC, Guzik TG, Granger DJ, Stewart MF. 2019. Abundance and survival of microbial aerosols in the troposphere and stratosphere. ISME J 13:2789–2799.

243. Rauf K, Hann A, Wallis M, Wickramasinghe C. 2010. Study of putative microfossils in space dust from the stratosphere. International Journal of Astrobiology 9:183–189.

204

244. Shivaji S, Chaturvedi P, Suresh K, Reddy GSN, Dutt CBS, Wainwright M, Narlikar JV, Bhargava PM. 2006. Bacillus aerius sp. nov., Bacillus aerophilus sp. nov., Bacillus stratosphericus sp. nov. and Bacillus altitudinis sp. nov., isolated from cryogenic tubes used for collecting air samples from high altitudes. Int J Syst Evol Microbiol 56:1465– 1473.

245. Narlikar JV, Lloyd D, Wickramasinghe NC, Harris MJ, Turner MP, Al-Mufti S, Wallis MK, Wainwright M, Rajaratnam P, Shivaji S, Reddy GSN, Ramadurai S, Hoyle F. 2003. A Balloon Experiment to detect Microorganisms in the Outer Space. Astrophysics and Space Science 285:555–562.

246. Wainwright M, Wickramasinghe NC, Narlikar JV, Rajaratnam P. 2003. Microorganisms cultured from stratospheric air samples obtained at 41 km. FEMS Microbiol Lett 218:161– 165.

247. Smith DJ, Griffin DW, Schuerger AC. 2010. Stratospheric microbiology at 20 km over the Pacific Ocean. Aerobiologia 26:35–46.

248. Smith DJ, Griffin DW, McPeters RD, Ward PD, Schuerger AC. 2011. Microbial survival in the stratosphere and implications for global dispersal. Aerobiologia 27:319–332.

249. Smith DJ, Ravichandar JD, Jain S, Griffin DW, Yu H, Tan Q, Thissen J, Lusby T, Nicoll P, Shedler S, Martinez P, Osorio A, Lechniak J, Choi S, Sabino K, Iverson K, Chan L, Jaing C, McGrath J. 2018. Airborne Bacteria in Earth’s Lower Stratosphere Resemble Taxa Detected in the Troposphere: Results From a New NASA Aircraft Bioaerosol Collector (ABC). Front Microbiol 9.

250. Sattler B, Puxbaum H, Psenner R. 2001. Bacterial growth in supercooled cloud droplets. Geophys Res Lett 28:239–242.

251. Kourtev PS, Hill KA, Shepson PB, Konopka A. 2011. Atmospheric cloud water contains a diverse bacterial community. Atmospheric Environment 45:5399–5405.

252. Vaïtilingom M, Deguillaume L, Vinatier V, Sancelme M, Amato P, Chaumerliac N, Delort A-M. 2013. Potential impact of microbial activity on the oxidant capacity and organic carbon budget in clouds. PNAS 110:559–564.

253. Vaïtilingom M, Amato P, Sancelme M, Laj P, Leriche M, Delort A-M. 2010. Contribution of Microbial Activity to Carbon Chemistry in Clouds. Appl Environ Microbiol 76:23–29.

254. Husárová S, Vaïtilingom M, Deguillaume L, Traikia M, Vinatier V, Sancelme M, Amato P, Matulová M, Delort A-M. 2011. Biotransformation of methanol and formaldehyde by bacteria isolated from clouds. Comparison with radical chemistry. Atmospheric Environment 45:6093–6102.

255. Husarova S. 2011. NMR study of C1 compound and saccharide metabolism of Bacillus SP. 3B6, a bacterial strain isolated from cloud water. Potential implication for atmospheric processes. PhD Thesis.

205

256. Delort A-M, Vaïtilingom M, Amato P, Sancelme M, Parazols M, Mailhot G, Laj P, Deguillaume L. 2010. A short overview of the microbial population in clouds: Potential roles in atmospheric chemistry and nucleation processes. Atmospheric Research 98:249– 260.

257. Yu S. 2000. Role of organic acids (formic, acetic, pyruvic and oxalic) in the formation of cloud condensation nuclei (CCN): a review. Atmospheric Research 53:185–217.

258. DeLeon-Rodriguez D, Lathem T, Rodriguez-R. L, Barazesh J, Anderson B, Aj B, Ld Z, M B, A N, Kt K. 2013. Microbiome of the upper troposphere: species composition and prevalence, effects of tropical storms, and atmospheric implications. Proc Natl Acad Sci U S A 110:2575–2580.

259. Conen F, Morris CE, Leifeld J, Yakutin MV, Alewell C. 2011. Biological residues define the ice nucleation properties of soil dust. Atmos Chem Phys 11:9643–9648.

260. Pietsch RB, Vinatzer BA, Schmale DGI. 2017. Diversity and Abundance of Ice Nucleating Strains of Pseudomonas syringae in a Freshwater Lake in Virginia, USA. Front Microbiol 8.

261. Moffett BF, Hill TCJ, DeMott PJ. 2018. Abundance of Biological Ice Nucleating Particles in the Mississippi and Its Major Tributaries. Atmosphere 9:307.

262. Morris CE, Sands DC, Vanneste JL, Montarry J, Oakley B, Guilbaud C, Glaux C. 2010. Inferring the evolutionary history of the plant pathogen Pseudomonas syringae from its biogeography in headwaters of rivers in North America, Europe, and New Zealand. mBio 1.

263. Schnell RC, Vali G. 1975. Freezing nuclei in marine waters. Tellus 27:321–323.

264. Wilson TW, Ladino LA, Alpert PA, Breckels MN, Brooks IM, Browse J, Burrows SM, Carslaw KS, Huffman JA, Judd C, Kilthau WP, Mason RH, McFiggans G, Miller LA, Nájera JJ, Polishchuk E, Rae S, Schiller CL, Si M, Temprado JV, Whale TF, Wong JPS, Wurl O, Yakobi-Hancock JD, Abbatt JPD, Aller JY, Bertram AK, Knopf DA, Murray BJ. 2015. A marine biogenic source of atmospheric ice-nucleating particles. Nature 525:234– 238.

265. Schnell RC. 1977. Ice nuclei in seawater, fog water and marine air off the coast of Nova Scotia: Summer 1975. Journal of the Atmospheric Sciences 34:1299–1305.

266. Hill TCJ, Moffett BF, DeMott PJ, Georgakopoulos DG, Stump WL, Franc GD. 2014. Measurement of Ice Nucleation-Active Bacteria on Plants and in Precipitation by Quantitative PCR. Appl Environ Microbiol 80:1256–1267.

267. Sauer AV, Rocha KR, Gonçalves RM, Meirelles WF, Figueiredo JEF, Marriel IE, Paccola-Meirelles LD. 2015. Survival of pantoea ananatis, causal agent of maize white spot disease in crop debris. Agron Sci Biotechnol 1:21.

206

268. Hader JD, Wright TP, Petters MD. 2014. Contribution of pollen to atmospheric ice nuclei concentrations. Atmospheric Chemistry and Physics 14:5433–5449.

269. Wright TP, Hader JD, McMeeking GR, Petters MD. 2014. High Relative Humidity as a Trigger for Widespread Release of Ice Nuclei. Aerosol Science and Technology 48:i–v.

270. Garcia E, Hill TCJ, Prenni AJ, DeMott PJ, Franc GD, Kreidenweis SM. 2012. Biogenic ice nuclei in boundary layer air over two U.S. High Plains agricultural regions. J Geophys Res Atmos 117:n/a-n/a.

271. Tobo Y, Prenni AJ, DeMott PJ, Huffman JA, McCluskey CS, Tian G, Pöhlker C, Pöschl U, Kreidenweis SM. 2013. Biological aerosol particles as a key determinant of ice nuclei populations in a forest ecosystem. Journal of Geophysical Research: Atmospheres 118:10,100-10,110.

272. Pratt KA, DeMott PJ, French JR, Wang Z, Westphal DL, Heymsfield AJ, Twohy CH, Prenni AJ, Prather KA. 2009. In situ detection of biological particles in cloud ice-crystals. Nature Geosci 2:398–401.

273. Christner BC, Cai R, Morris CE, McCarter KS, Foreman CM, Skidmore ML, Montross SN, Sands DC. 2008. Geographic, seasonal, and precipitation chemistry influence on the abundance and activity of biological ice nucleators in rain and snow. PNAS 105:18854– 18859.

274. Joyce RE, Lavender H, Farrar J, Werth JT, Weber CF, D’Andrilli J, Vaitilingom M, Christner BC. 2019. Biological Ice-Nucleating Particles Deposited Year-Round in Subtropical Precipitation. Appl Environ Microbiol 85.

275. Christner BC, Morris CE, Foreman CM, Cai R, Sands DC. 2008. Ubiquity of Biological Ice Nucleators in Snowfall. Science 319:1214–1214.

276. Morris CE, Conen F, Alex Huffman J, Phillips V, Pöschl U, Sands DC. 2014. Bioprecipitation: a feedback cycle linking Earth history, ecosystem dynamics and land use through biological ice nucleators in the atmosphere. Glob Change Biol 20:341–351.

277. Stopelli E, Conen F, Zimmermann L, Alewell C, Morris CE. 2014. Freezing nucleation apparatus puts new slant on study of biological ice nucleators in precipitation. Atmospheric Measurement Techniques 7:129–134.

278. Stopelli E, Conen F, Morris CE, Herrmann E, Bukowiecki N, Alewell C. 2015. Ice nucleation active particles are efficiently removed by precipitating clouds. Sci Rep 5:1–7.

279. Pouzet G, Peghaire E, Aguès M, Baray J-L, Conen F, Amato P. 2017. Atmospheric Processing and Variability of Biological Ice Nucleating Particles in Precipitation at Opme, France. Atmosphere 8:229.

280. Morris CE, Georgakopoulos DG, Sands DC. 2004. Ice nucleation active bacteria and their potential role in precipitation. J Phys IV France 121:87–103.

207

281. Möhler O, Demott PJ, Vali G, Levin Z. 2007. Microbiology and atmospheric processes: the role of biological particles in cloud physics. Biogeosci Discuss 4:2559–2591.

282. Michaud AB, Dore JE, Leslie D, Lyons WB, Sands DC, Priscu JC. 2014. Biological ice nucleation initiates hailstone formation. Journal of Geophysical Research: Atmospheres 119:12,186-12,197.

283. Prenni AJ, Tobo Y, Garcia E, DeMott PJ, Huffman JA, McCluskey CS, Kreidenweis SM, Prenni JE, Pöhlker C, Pöschl U. 2013. The impact of rain on ice nuclei populations at a forested site in Colorado. Geophysical Research Letters 40:227–231.

284. Rogers DC. 1994. Detecting Ice Nuclei with a Continuous—Flow Diffusion Chamber- Some Exploratory Tests of Instrument Response. J Atmos Oceanic Technol 11:1042– 1047.

285. Huffman JA, Santarpia J. 2017. Online techniques for quantification and characterization of biological aerosols. Microbiology of aerosols 83–114.

286. Joly M, Amato P, Deguillaume L, Monier M, Hoose C, Delort A-M. 2014. Quantification of ice nuclei active at near 0 °C temperatures in low-altitude clouds at the Puy de Dôme atmospheric station. Atmos Chem Phys 14:8185–8195.

287. Joly M, Amato P, Deguillaume L, Monier M, Hoose C, Delort A-M. 2014. Direct quantification of total and biological ice nuclei in cloud water. Atmos Chem Phys Discuss 14:3707–3731.

288. O′Sullivan D, Murray BJ, Ross JF, Whale TF, Price HC, Atkinson JD, Umo NS, Webb ME. 2015. The relevance of nanoscale biological fragments for ice nucleation in clouds. Sci Rep 5:1–7.

289. Šantl-Temkiv T, Sahyoun M, Finster K, Hartmann S, Augustin-Bauditz S, Stratmann F, Wex H, Clauss T, Nielsen NW, Sørensen JH, Korsholm US, Wick LY, Karlson UG. 2015. Characterization of airborne ice-nucleation-active bacteria and bacterial fragments. Atmospheric Environment 109:105–117.

290. O’Sullivan D, Murray BJ, Ross JF, Webb ME. 2016. The adsorption of fungal ice- nucleating proteins on mineral dusts: a terrestrial reservoir of atmospheric ice-nucleating particles. Atmospheric Chemistry and Physics 16:7879–7887.

291. Hoose C, Kristjánsson JE, Burrows SM. 2010. How important is biological ice nucleation in clouds on a global scale? Environ Res Lett 5:024009.

292. Šantl-Temkiv T, Sikoparija B, Maki T, Carotenuto F, Amato P, Yao M, Morris CE, Schnell R, Jaenicke R, Pöhlker C, DeMott PJ, Hill TCJ, Huffman JA. 2019. Bioaerosol field measurements: Challenges and perspectives in outdoor studies. Aerosol Science and Technology 0:1–27.

208

293. DeMott PJ, Hill TCJ, McCluskey CS, Prather KA, Collins DB, Sullivan RC, Ruppel MJ, Mason RH, Irish VE, Lee T, Hwang CY, Rhee TS, Snider JR, McMeeking GR, Dhaniyala S, Lewis ER, Wentzell JJB, Abbatt J, Lee C, Sultana CM, Ault AP, Axson JL, Diaz Martinez M, Venero I, Santos-Figueroa G, Stokes MD, Deane GB, Mayol-Bracero OL, Grassian VH, Bertram TH, Bertram AK, Moffett BF, Franc GD. 2016. Sea spray aerosol as a unique source of ice nucleating particles. Proc Natl Acad Sci U S A 113:5797–5803.

294. Matus AV, L’Ecuyer TS. 2017. The role of cloud phase in Earth’s radiation budget. Journal of Geophysical Research: Atmospheres 122:2559–2578.

295. Zhang D, Wang Z, Luo T, Yin Y, Flynn C. 2017. The occurrence of ice production in slightly supercooled Arctic stratiform clouds as observed by ground-based remote sensors at the ARM NSA site. Journal of Geophysical Research: Atmospheres 122:2867–2877.

296. Khain AP, Beheng KD, Heymsfield A, Korolev A, Krichak SO, Levin Z, Pinsky M, Phillips V, Prabhakaran T, Teller A, van den Heever SC, Yano J-I. 2015. Representation of microphysical processes in cloud-resolving models: Spectral (bin) microphysics versus bulk parameterization. Rev Geophys 53:247–322.

297. Baumgardner D, Abel SJ, Axisa D, Cotton R, Crosier J, Field P, Gurganus C, Heymsfield A, Korolev A, Kraemer M. 2017. Cloud ice properties: In situ measurement challenges. Meteorological monographs 58:9–1.

298. Fan J, Ovtchinnikov M, Comstock JM, McFarlane SA, Khain A. 2009. Ice formation in Arctic mixed-phase clouds: Insights from a 3-D cloud-resolving model with size-resolved aerosol and cloud microphysics. Journal of Geophysical Research: Atmospheres 114.

299. Korolev A, McFarquhar G, Field PR, Franklin C, Lawson P, Wang Z, Williams E, Abel SJ, Axisa D, Borrmann S, Crosier J, Fugal J, Krämer M, Lohmann U, Schlenczek O, Schnaiter M, Wendisch M. 2017. Mixed-Phase Clouds: Progress and Challenges. Meteorological Monographs 58:5.1-5.50.

300. Cooper WA. 1986. Ice initiation in natural clouds, p. 29–32. In Precipitation Enhancement—A Scientific Challenge. Springer.

301. Rangno AL, Hobbs PV. 2001. Ice particles in stratiform clouds in the Arctic and possible mechanisms for the production of high ice concentrations. Journal of Geophysical Research: Atmospheres 106:15065–15075.

302. Mossop SC. 1985. The Origin and Concentration of Ice Crystals in Clouds. Bull Amer Meteor Soc 66:264–273.

303. Phillips VTJ, Choularton TW, Illingworth AJ, Hogan RJ, Field PR. 2003. Simulations of the glaciation of a frontal mixed-phase cloud with the Explicit Microphysics Model. Q J Roy Meteor Soc 129:1351–1371.

209

304. Khain AP, BenMoshe N, Pokrovsky A. 2008. Factors Determining the Impact of Aerosols on Surface Precipitation from Clouds: An Attempt at Classification. J Atmos Sci 65:1721– 1748.

305. Schwarzenboeck A, Shcherbakov V, Lefevre R, Gayet J-F, Pointin Y, Duroure C. 2009. Indications for stellar-crystal fragmentation in Arctic clouds. Atmospheric research 92:220–228.

306. Warren SG, Eastman RM, Hahn CJ. 2007. A survey of changes in cloud cover and cloud types over land from surface observations, 1971–96. Journal of Climate 20:717–738.

307. Eastman R, Warren SG. 2013. A 39-yr survey of cloud changes from land stations worldwide 1971–2009: Long-term trends, relation to aerosols, and expansion of the tropical belt. Journal of Climate 26:1286–1303.

308. Eastman R, Warren SG, Hahn CJ. 2011. Variations in cloud cover and cloud types over the ocean from surface observations, 1954–2008. Journal of Climate 24:5914–5934.

309. Hoose C, Kristjánsson JE, Chen J-P, Hazra A. 2010. A Classical-Theory-Based Parameterization of Heterogeneous Ice Nucleation by Mineral Dust, Soot, and Biological Particles in a Global Climate Model. J Atmos Sci 67:2483–2503.

310. Burrows SM, Butler T, Jöckel P, Tost H, Kerkweg A, Pöschl U, Lawrence MG. 2009. Bacteria in the global atmosphere – Part 2: Modeling of emissions and transport between different ecosystems. Atmospheric Chemistry and Physics 9:9281–9297.

311. Sesartic A, Lohmann U, Storelvmo T. 2013. Modelling the impact of fungal spore ice nuclei on clouds and precipitation. Environ Res Lett 8:014029.

312. Diehl K, Wurzler S. 2010. Air parcel model simulations of a convective cloud: Bacteria acting as immersion ice nuclei. Atmos Environ 44:4622–4628.

313. Phillips VTJ, Andronache C, Christner B, Morris CE, Sands DC, Bansemer A, Lauer A, McNaughton C, Seman C. 2009. Potential impacts from biological aerosols on ensembles of continental clouds simulated numerically. Biogeosciences 6:987–1014.

314. Phillips VTJ, Demott PJ, Andronache C, Pratt KA, Prather KA, Subramanian R, Twohy C. 2012. Improvements to an Empirical Parameterization of Heterogeneous Ice Nucleation and Its Comparison with Observations. J Atmos Sci 70:378–409.

315. Sahyoun M, Korsholm US, Sørensen JH, Šantl-Temkiv T, Finster K, Gosewinkel U, Nielsen NW. 2017. Impact of bacterial ice nucleating particles on weather predicted by a numerical weather prediction model. Atmospheric Environment 170:33–44.

316. DeMott PJ, Cziczo DJ, Prenni AJ, Murphy DM, Kreidenweis SM, Thomson DS, Borys R, Rogers DC. 2003. Measurements of the concentration and composition of nuclei for cirrus formation. Proceedings of the National Academy of Sciences 100:14655–14660.

210

317. DeMott PJ, Prenni AJ, Liu X, Kreidenweis SM, Petters MD, Twohy CH, Richardson MS, Eidhammer T, Rogers DC. 2010. Predicting global atmospheric ice nuclei distributions and their impacts on climate. PNAS 107:11217–11222.

318. Wei M, Xu C, Chen J, Zhu C, Li J, Lv G. 2017. Characteristics of bacterial community in cloud water at Mt Tai: similarity and disparity under polluted and non-polluted cloud episodes. Atmospheric Chemistry and Physics 17:5253.

319. Amato P, Parazols M, Sancelme M, Laj P, Mailhot G, Delort A-M. 2007. Microorganisms isolated from the water phase of tropospheric clouds at the Puy de Dôme: major groups and growth abilities at low temperatures. FEMS microbiology ecology 59:242–254.

320. Amato P, Demeer F, Melaouhi A, Fontanella S, Martin-Biesse A-S, Sancelme M, Laj P, Delort AM. 2007. A fate for organic acids, formaldehyde and methanol in cloud water: their biotransformation by micro-organisms.

321. Palmer T. 2017. The primacy of doubt: Evolution of numerical weather prediction from determinism to probability. Journal of Advances in Modeling Earth Systems 9:730–734.

322. Bauer P, Thorpe A, Brunet G. 2015. The quiet revolution of numerical weather prediction. Nature 525:47–55.

323. J. Murray B, O’Sullivan D, D. Atkinson J, E. Webb M. 2012. Ice nucleation by particles immersed in supercooled cloud droplets. Chem Soc Rev 41:6519–6554.

324. Kajava AV, Lindow SE. 1993. A Model of the Three-dimensional Structure of Ice Nucleation Proteins. J Mol Biol 232:709–717.

325. Mortazavi R, Attiya S, Ariya PA. 2015. Arctic microbial and next-generation sequencing approach for bacteria in snow and frost flowers: selected identification, abundance and freezing nucleation. Atmos Chem Phys 15:6183–6204.

326. Nejad P, Ramstedt M, Granhall U, Roos S, McIvor I. 2006. Biochemical characterization and identification of ice-nucleation-active (INA) willow pathogens by means of BIOLOG® MicroPlate, INA gene primers and PCR-based 16S rRNA-gene analyses. J Plant Dis Prot 113:97–106.

327. Ponder MA, Gilmour SJ, Bergholz PW, Mindock CA, Hollingsworth R, Thomashow MF, Tiedje JM. 2005. Characterization of potential stress responses in ancient Siberian permafrost psychroactive bacteria. FEMS Microbiol Ecol 53:103–115.

328. D’souza NA, Kawarasaki Y, Gantz JD, Lee RE, Beall BFN, Shtarkman YM, Koçer ZA, Rogers SO, Wildschutte H, Bullerjahn GS, McKay RML. 2013. Diatom assemblages promote ice formation in large lakes. ISME J 7:1632–1640.

329. Bowers RM, Lauber CL, Wiedinmyer C, Hamady M, Hallar AG, Fall R, Knight R, Fierer N. 2009. Characterization of Airborne Microbial Communities at a High-Elevation Site

211

and Their Potential To Act as Atmospheric Ice Nuclei. Appl Environ Microbiol 75:5121– 5130.

330. Monteil CL, Bardin M, Morris CE. 2014. Features of air masses associated with the deposition of Pseudomonas syringae and Botrytis cinerea by rain and snowfall. ISME J 8:2290–2304.

331. Fierer N, Liu Z, Rodríguez-Hernández M, Knight R, Henn M, Hernandez MT. 2008. Short-Term Temporal Variability in Airborne Bacterial and Fungal Populations. Appl Environ Microbiol 74:200–207.

332. Deguillaume L, Leriche M, Amato P, Ariya PA, Delort AM, Poschl U, Chaumerliac N, Bauer H, Flossmann A, Morris CE. 2008. Microbiology and atmospheric processes: chemical interactions of primary biological aerosols. Biogeosciences 5:1073–1084.

333. Korolev AV, Isaac GA, Cober SG, Strapp JW, Hallett J. 2003. Microphysical characterization of mixed-phase clouds. Q J Roy Meteor Soc 129:39–65.

334. Khain A, Pokrovsky A, Pinsky M, Seifert A, Phillips V. 2004. Simulation of Effects of Atmospheric Aerosols on Deep Turbulent Convective Clouds Using a Spectral Microphysics Mixed-Phase Cumulus Cloud Model. Part I: Model Description and Possible Applications. J Atmos Sci 61:2963–2982.

335. Mossop SC. 1976. Production of secondary ice particles during the growth of graupel by riming. Q J Roy Meteor Soc 102:45–57.

336. Mignani C, Creamean JM, Zimmermann L, Alewell C, Conen F. 2019. New type of evidence for secondary ice formation at around −15°C in mixed-phase clouds. Atmos Chem Phys 19:877–886.

337. Zeng X, Tao W-K, Zhang M, Hou AY, Xie S, Lang S, Li X, Starr DO, Li X, Simpson J. 2009. An Indirect Effect of Ice Nuclei on Atmospheric Radiation. J Atmos Sci 66:41–61.

338. Vali G. 1971. Quantitative Evaluation of Experimental Results an the Heterogeneous Freezing Nucleation of Supercooled Liquids. J Atmos Sci 28:402–409.

339. Christner BC, Royston-Bishop G, Foreman CM, Arnold BR, Tranter M, Welch KA, Lyons WB, Tsapin AI, Studinger M, Priscu JC. 2006. Limnological conditions in Subglacial Lake Vostok, Antarctica. Limnol Oceanogr 51:2485–2501.

340. Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D, Huntley J, Fierer N, Owens SM, Betley J, Fraser L, Bauer M, Gormley N, Gilbert JA, Smith G, Knight R. 2012. Ultra- high-throughput microbial community analysis on the Illumina HiSeq and MiSeq platforms. ISME J 6:1621–1624.

341. Aho K, Weber C, Christner B, Vinatzer B, Morris C, Joyce R, Werth J, Bayless-Edwards A, Lavender H, Schmale D. The spatiotemporal drivers of microbial composition and diversity in precipitation. In review.

212

342. Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M, Hollister EB, Lesniewski RA, Oakley BB, Parks DH, Robinson CJ, Sahl JW, Stres B, Thallinger GG, Horn DJV, Weber CF. 2009. Introducing mothur: Open-Source, Platform-Independent, Community- Supported Software for Describing and Comparing Microbial Communities. Appl Environ Microbiol 75:7537–7541.

343. Pruesse E, Quast C, Knittel K, Fuchs BM, Ludwig W, Peplies J, Glöckner FO. 2007. SILVA: a comprehensive online resource for quality checked and aligned ribosomal RNA sequence data compatible with ARB. Nucleic Acids Res 35:7188–7196.

344. Jr WEA, Tucker SA, Fetzer JC. 1991. Fluorescence Emission Properties of Polycyclic Aromatic Compounds in Review. Polycycl Aromat Comp 2:75–105.

345. Tucker SA, Acree WE. 1992. Excitation Versus Emission Spectra as a Means to Examine Selective Fluorescence Quenching Agents. Appl Spectrosc 46:1388–1392.

346. R. Murphy K, A. Stedmon C, Graeber D, Bro R. 2013. Fluorescence spectroscopy and multi-way techniques. PARAFAC. Anal Methods 5:6557–6566.

347. Anagnostou EN. 2004. A convective/stratiform precipitation classification algorithm for volume scanning weather radar observations. Meteorol Appl 11:291–300.

348. Biggerstaff MI, Listemaa SA. 2000. An Improved Scheme for Convective/Stratiform Echo Classification Using Radar Reflectivity. J Appl Meteor 39:2129–2150.

349. Schultz P. 1989. Relationships of Several Stability Indices to Convective Weather Events in Northeast Colorado. Wea Forecasting 4:73–80.

350. Blanchard DO. 1998. Assessing the Vertical Distribution of Convective Available Potential Energy. Wea Forecasting 13:870–877.

351. Omernik JM, Griffith GE. 2014. Ecoregions of the Conterminous United States: Evolution of a Hierarchical Spatial Framework. Environmental Management 54:1249–1266.

352. 1999. A proposed framework for delineating ecologically-based planning, implementation, and evaluation units for cooperative bird conservation in the U.S. ; Ecological regions of North America: Toward a common perspective. Commission for Environmental Cooperation.

353. Tabachnick BG, Fidell LS, Ullman JB. 2019. Using multivariate statisticsSeventh edition. Pearson, Boston.

354. Li P, Stuart EA, Allison DB. 2015. Multiple Imputation: A Flexible Tool for Handling Missing Data. JAMA 314:1966–1967.

355. Yuan Y. Multiple Imputation for Missing Data: Concepts and New Development. Proceedings of the Twenty-Fifth Annual SAS Users Group International Conference 267:11.

213

356. Yong AG, Pearce S. 2013. A beginner’s guide to factor analysis: Focusing on exploratory factor analysis. Tutor Quant Methods Psychol 9:79–94.

357. Coble PG. 1996. Characterization of marine and terrestrial DOM in seawater using excitation-emission matrix spectroscopy. Mar Chem 51:325–346.

358. Coble PG, Green SA, Blough NV, Gagosian RB. 1990. Characterization of dissolved organic matter in the Black Sea by fluorescence spectroscopy. Nature 348:432–435.

359. Coble PG, Del Castillo CE, Avril B. 1998. Distribution and optical properties of CDOM in the Arabian Sea during the 1995 Southwest Monsoon. Deep Sea Research Part II: Topical Studies in Oceanography 45:2195–2223.

360. Gurian-Sherman D, Lindow SE. 1993. Bacterial ice nucleation: significance and molecular basis. FASEB J 7:1338–1343.

361. Parlanti E, Wörz K, Geoffroy L, Lamotte M. 2000. Dissolved organic matter fluorescence spectroscopy as a tool to estimate biological activity in a coastal zone submitted to anthropogenic inputs. Org Geochem 31:1765–1781.

362. Stedmon CA, Markager S. 2005. Resolving the variability in dissolved organic matter fluorescence in a temperate estuary and its catchment using PARAFAC analysis. Limnol Oceanogr 50:686–697.

363. Schnell RC, Vali G. 1976. Biogenic Ice Nuclei: Part I. Terrestrial and Marine Sources. J Atmos Sci 33:1554–1564.

364. Creamean JM, Kirpes RM, Pratt KA, Spada NJ, Maahn M, Boer G de, Schnell RC, China S. 2018. Marine and terrestrial influences on ice nucleating particles during continuous springtime measurements in an Arctic oilfield location. Atmos Chem Phys 18:18023– 18042.

365. Buczolits S, Denner EB, Kämpfer P, Busse H-J. 2006. Proposal of Hymenobacter norwichensis sp. nov., classification of ‘Taxeobacter ocellatus’,‘Taxeobacter gelupurpurascens’ and ‘Taxeobacter chitinovorans’ as Hymenobacter ocellatus sp. nov., Hymenobacter gelipurpurascens sp. nov. and Hymenobacter chitinivorans sp. nov., respectively, and emended description of the genus Hymenobacter Hirsch et al. 1999. Int J Syst Evol Microbiol 56:2071–2078.

366. Hirsch P, Ludwig W, Hethke C, Sittig M, Hoffmann B, Gallikowski CA. 1998. Hymenobacter roseosalivarius gen. nov., sp. nov. from continental Antarctic soils and sandstone: bacteria of the Cytophaga/Flavobacterium/Bacteroides line of phylogenetic descent. Syst Appl Microbiol 21:374–383.

367. Hoang V-A, Kim Y-J, Nguyen NL, Yang D-C. 2013. Hymenobacter ginsengisoli sp. nov., isolated from soil of a ginseng field. Int J Syst Evol Microbiol 63:661–666.

214

368. Jin L, Lee H-G, Kim S-G, Lee KC, Ahn C-Y, Oh H-M. 2014. Hymenobacter ruber sp. nov., isolated from grass soil. Int J Syst Evol Microbiol 64:979–983.

369. Kim K-H, Im W-T, Lee S-T. 2008. Hymenobacter soli sp. nov., isolated from grass soil. Int J Syst Evol Microbiol 58:941–945.

370. Lee J-J, Srinivasan S, Lim S, Joe M, Lee SH, Kwon SA, Kwon YJ, Lee J, Choi JJ, Lee HM. 2014. Hymenobacter swuensis sp. nov., a gamma-radiation-resistant bacteria isolated from mountain soil. Curr Microbiol 68:305–310.

371. Reddy GS, Garcia-Pichel F. 2013. Description of Hymenobacter arizonensis sp. nov. from the southwestern arid lands of the United States of America. Antonie van Leeuwenhoek 103:321–330.

372. Srinivasan S, Joo ES, Lee J-J, Kim MK. 2015. Hymenobacter humi sp. nov., a bacterium isolated from soil. Antonie van Leeuwenhoek 107:1411–1419.

373. Zhang D-C, Busse H-J, Liu H-C, Zhou Y-G, Schinner F, Margesin R. 2011. Hymenobacter psychrophilus sp. nov., a psychrophilic bacterium isolated from soil. Int J Syst Evol Microbiol 61:859–863.

374. Fellman JB, Hood E, Spencer RGM. 2010. Fluorescence spectroscopy opens new windows into dissolved organic matter dynamics in freshwater ecosystems: A review. Limnol and oceanogr 55:2452–2462.

375. Murphy KR, Stedmon CA, Waite TD, Ruiz GM. 2008. Distinguishing between terrestrial and autochthonous organic matter sources in marine environments using fluorescence spectroscopy. Mar Chem 108:40–58.

376. Balcarczyk KL, Jones JB, Jaffé R, Maie N. 2009. Stream dissolved organic matter bioavailability and composition in watersheds underlain with discontinuous permafrost. Biogeochemistry 94:255–270.

377. Wolff GT. 1984. On the nature of nitrate in coarse continental aerosols. Atmos Environ 18:977–981.

378. Milne PJ, Zika RG. 1993. Amino acid nitrogen in atmospheric aerosols: Occurrence, sources and photochemical modification. J Atmos Chem 16:361–398.

379. Kyrö E-M, Grönholm T, Vuollekoski H, Virkkula A, Kulmala M, Laakso L. Snow scavenging of ultraﬁne particles: ﬁeld measurements and parameterization. Boreal Env Res 14:12.

380. Paramonov M, Grönholm T, Virkkula A. Below-cloud scavenging of aerosol particles by snow at an urban site in Finland. Boreal Env Res 16:17.

215

381. Zhang L, Wang X, Moran MD, Feng J. 2013. Review and uncertainty assessment of size- resolved scavenging coefficient formulations for below-cloud snow scavenging of atmospheric aerosols. Atmos Chem Phys 13:10005–10025.

382. Bowers RM, McCubbin IB, Hallar AG, Fierer N. 2012. Seasonal variability in airborne bacterial communities at a high-elevation site. Atmos Environ 50:41–49.

383. Zhang L, Dai J, Tang Y, Luo X, Wang Y, An H, Fang C, Zhang C. 2009. Hymenobacter deserti sp. nov., isolated from the desert of Xinjiang, China. Int J Syst Evol Microbiol 59:77–82.

384. Zhang Q, Liu C, Tang Y, Zhou G, Shen P, Fang C, Yokota A. 2007. Hymenobacter xinjiangensis sp. nov., a radiation-resistant bacterium isolated from the desert of Xinjiang, China. Int J Syst Evol Microbiol 57:1752–1756.

385. Šantl-Temkiv T, Finster K, Dittmar T, Hansen BM, Thyrhaug R, Nielsen NW, Karlson UG. 2013. Hailstones: A Window into the Microbial and Chemical Inventory of a Storm Cloud. PLOS ONE 8:e53550.

386. Weon H-Y, Kwon S-W, Son J-A, Kim S-J, Kim Y-S, Kim B-Y, Ka J-O. 2010. Adhaeribacter aerophilus sp. nov., Adhaeribacter aerolatus sp. nov. and Segetibacter aerophilus sp. nov., isolated from air samples. Int J Syst Evol Microbiol 60:2424–2429.

387. Brodie EL, DeSantis TZ, Parker JPM, Zubietta IX, Piceno YM, Andersen GL. 2007. Urban aerosols harbor diverse and dynamic bacterial populations. PNAS 104:299–304.

388. Buczolits S, Denner EBM, Vybiral D, Wieser M, Kämpfer P, Busse H-J. 2002. Classification of three airborne bacteria and proposal of Hymenobacter aerophilus sp. nov. Int J Syst Evol Microbiol 52:445–456.

389. Graf J. 2014. The Family Rikenellaceae, p. 857–859. In Rosenberg, E, DeLong, EF, Lory, S, Stackebrandt, E, Thompson, F (eds.), The Prokaryotes. Springer Berlin Heidelberg, Berlin, Heidelberg.

390. Eren AM, Sogin ML, Morrison HG, Vineis JH, Fisher JC, Newton RJ, McLellan SL. 2015. A single genus in the gut microbiome reflects host preference and specificity. ISME J 9:90–100.

391. Kaakoush NO. 2015. Insights into the Role of Erysipelotrichaceae in the Human Host. Front Cell Infect Microbiol 5.

392. Mülmenstädt J, Sourdeval O, Delanoë J, Quaas J. 2015. Frequency of occurrence of rain from liquid-, mixed-, and ice-phase clouds derived from A-Train satellite retrievals. Geophys Res Lett 42:6502–6509.

393. Petters MD, Wright TP. 2015. Revisiting ice nucleation from precipitation samples. Geophys Res Lett 42:8758–8766.

216

394. Hess M, Koepke P, Schult I. 1998. Optical Properties of Aerosols and Clouds: The Software Package OPAC. Bull Amer Meteor Soc 79:831–844.

395. Rosenfeld D, Lensky IM. 1998. Satellite-Based Insights into Precipitation Formation Processes in Continental and Maritime Convective Clouds. Bull Amer Meteor Soc 79:2457–2476.

396. Korolev AV, Mazin IP. 2003. Supersaturation of Water Vapor in Clouds. J Atmos Sci 60:2957–2974.

397. Jung CH, Bae SY, Kim YP. 2011. Approximated solution on the properties of the scavenging gap during precipitation using harmonic mean method. Atmos Res 99:496– 504.

398. Rosinski J, Nagamoto CT, Kerrigan TC, Langer G. 1973. Freezing Nuclei Derived from Soil Particles. J Atmos Sci 30:644–652.

399. Doyle S, Dieser M, Broemsen E, Christner B. 2012. General Characteristics of Cold- Adapted Microorganisms. Polar Microbiology: Life in a Deep Freeze 103–125.

400. Morris CE, Soubeyrand S, Bigg EK, Creamean JM, Sands DC. 2016. Mapping Rainfall Feedback to Reveal the Potential Sensitivity of Precipitation to Biological Aerosols. Bull Amer Meteor Soc 98:1109–1118.

401. Sambrook J, Fritsch EF, Maniatis T. 1989. Molecular cloning: a laboratory manual. Cold spring harbor laboratory press.

402. Lane DJ. 1991. 16S/23S rRNA sequencing. Nucleic acid techniques in bacterial systematics (Stackebrandt E & Goodfellow M, eds). Wiley, New York.

403. Varvaro L, Fabi A. 1992. The Role of Ice Nucleation Active Pseudomonas Viridiflava in Frost Injury to Kiwifruit Plants. Rivista di Patologia Vegetale 2:85–90.

404. Zhao JL, Orser CS. 1990. Conserved repetition in the ice nucleation gene inaX from Xanthomonas campestris pv. translucens. Mol Gen Genet 223:163–166.

405. Panoff J-M, Thammavongs B, Guéguen M, Boutibonnes P. 1998. Cold Stress Responses in Mesophilic Bacteria. Cryobiology 36:75–83.

406. Dsouza M, Taylor MW, Turner SJ, Aislabie J. 2015. Genomic and phenotypic insights into the ecology of Arthrobacter from Antarctic soils. BMC Genomics 16:36.

407. Stackebrandt E, Schumann P. 2006. Introduction to the taxonomy of actinobacteria. Prokaryotes 3:297–321.

408. Kawahara H. 2002. The structures and functions of ice crystal-controlling proteins from bacteria. Journal of Bioscience and Bioengineering 94:492–496.

217

409. Knackstedt KA, Moffett BF, Hartmann S, Wex H, Hill TCJ, Glasgo ED, Reitz LA, Augustin-Bauditz S, Beall BFN, Bullerjahn GS, Fröhlich-Nowoisky J, Grawe S, Lubitz J, Stratmann F, McKay RML. 2018. Terrestrial Origin for Abundant Riverine Nanoscale Ice- Nucleating Particles. Environ Sci Technol 52:12358–12367.

410. Santachiara G, Prodi F, Belosi F. 2012. A Review of Termo- and Diffusio-Phoresis in the Atmospheric Aerosol Scavenging Process. Part 1: Drop Scavenging 2012.

411. Moore RA, Hanlon R, Powers C, Schmale DG, Christner BC. 2020. Scavenging of Sub- Micron to Micron-Sized Microbial Aerosols during Simulated Rainfall. 1. Atmosphere 11:80.

412. Creamean J, Solomon A, Hanlon R, Schmale D III, Kirpes R, Pratt K, Spada N, Cross J, Pickart R, McRaven L, Phillips V, Schnell R. 2018. Understanding the sources of Arctic biological ice nucleating particles 20:17882.

413. Creamean JM, Cross JN, Pickart R, McRaven L, Lin P, Pacini A, Hanlon R, Schmale DG, Ceniceros J, Aydell T, Colombi N, Bolger E, DeMott PJ. 2019. Ice Nucleating Particles Carried From Below a Phytoplankton Bloom to the Arctic Atmosphere. Geophysical Research Letters 46:8572–8581.

414. Schröder F, Kärcher B, Fiebig M, Petzold A. 2002. Aerosol states in the free troposphere at northern midlatitudes. Journal of Geophysical Research: Atmospheres 107:LAC 8-1- LAC 8-8.

415. Lewis ER, Schwartz SE. 2004. Sea Salt Aerosol Production: Mechanisms, Methods, Measurements and Models—A Critical Review. Washington DC American Geophysical Union Geophysical Monograph Series 152.

416. Leeuw G de, Andreas EL, Anguelova MD, Fairall CW, Lewis ER, O’Dowd C, Schulz M, Schwartz SE. 2011. Production flux of sea spray aerosol. Reviews of Geophysics 49.

417. Gantt B, Meskhidze N. 2013. The physical and chemical characteristics of marine primary organic aerosol: a review. Atmospheric Chemistry and Physics 13:3979–3996.

418. Yang J, Wang Z, Heymsfield AJ, DeMott PJ, Twohy CH, Suski KJ, Toohey DW. 2020. High ice concentration observed in tropical maritime stratiform mixed-phase clouds with top temperatures warmer than −8 °C. Atmospheric Research 233:104719.

419. Skamarock C, Klemp B, Dudhia J, Gill O, Liu Z, Berner J, Wang W, Powers G, Duda G, Barker D, Huang X. 2019. A Description of the Advanced Research WRF Model Version 4.

420. Thompson G, Eidhammer T. 2014. A Study of Aerosol Impacts on Clouds and Precipitation Development in a Large Winter Cyclone. J Atmos Sci 71:3636–3658.

218

421. Thompson G, Field PR, Rasmussen RM, Hall WD. 2008. Explicit Forecasts of Winter Precipitation Using an Improved Bulk Microphysics Scheme. Part II: Implementation of a New Snow Parameterization. Mon Wea Rev 136:5095–5115.

422. Thompson G, Rasmussen RM, Manning K. 2004. Explicit Forecasts of Winter Precipitation Using an Improved Bulk Microphysics Scheme. Part I: Description and Sensitivity Analysis. Mon Wea Rev 132:519–542.

423. Stull R. 2017. Fronts & Airmasses, p. 389–424. In Practical Meteorology: An Algebra- based Survey of Atmospheric Science, 2nd ed. University of British Columbia.

424. Colarco P, Silva A da, Chin M, Diehl T. 2010. Online simulations of global aerosol distributions in the NASA GEOS-4 model and comparisons to satellite and ground-based aerosol optical depth. Journal of Geophysical Research: Atmospheres 115.

425. Ginoux P, Chin M, Tegen I, Prospero JM, Holben B, Dubovik O, Lin S-J. 2001. Sources and distributions of dust aerosols simulated with the GOCART model. Journal of Geophysical Research: Atmospheres 106:20255–20273.

426. Cziczo DJ, Murphy DM, Hudson PK, Thomson DS. 2004. Single particle measurements of the chemical composition of cirrus ice residue during CRYSTAL-FACE. Journal of Geophysical Research: Atmospheres 109.

427. Richardson MS, DeMott PJ, Kreidenweis SM, Cziczo DJ, Dunlea EJ, Jimenez JL, Thomson DS, Ashbaugh LL, Borys RD, Westphal DL, Casuccio GS, Lersch TL. 2007. Measurements of heterogeneous ice nuclei in the western United States in springtime and their relation to aerosol characteristics. Journal of Geophysical Research: Atmospheres 112.

428. Phillips VTJ, DeMott PJ, Andronache C. 2008. An Empirical Parameterization of Heterogeneous Ice Nucleation for Multiple Chemical Species of Aerosol. J Atmos Sci 65:2757–2783.

429. Bigg EK. 1953. The Supercooling of Water. Proc Phys Soc B 66:688–694.

430. Boer G de, Morrison H, Shupe MD, Hildner R. 2011. Evidence of liquid dependent ice nucleation in high-latitude stratiform clouds from surface remote sensors. Geophysical Research Letters 38.

431. Lighthart B, Shaffer BT. 1994. Bacterial flux from chaparral into the atmosphere in mid- summer at a high desert location. Atmospheric Environment 28:1267–1274.

432. Tilley RI, Ho J, Eamus D. 2001. Background Bioaerosols and Aerosols at Two Sites in Northern Australia: Preliminary Measurements. Defense Science and Technology Organisation Victoria (Australia) Aeronautical and Maritime Research Lab.

219

433. Griffin DW, Westphal DL, Gray MA. 2006. Airborne microorganisms in the African desert dust corridor over the mid-Atlantic ridge, Ocean Drilling Program, Leg 209. Aerobiologia 22:211–226.

434. Boucher O, Randall D, Artaxo P, Bretherton C, Feingold G, Forster P, Kerminen V-M, Kondo Y, Liao H, Lohmann U, Rasch P, Satheesh S, Sherwood S, Stevens B, Zhang X-Y. 2013. Clouds and AerosolsClimate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, United Kingdom and New York, NY, USA.

435. Schumacher CJ, Pöhlker C, Aalto P, Hiltunen V, Petäjä T, Kulmala M, Pöschl U, Huffman JA. 2013. Seasonal cycles of fluorescent biological aerosol particles in boreal and semi-arid forests of Finland and Colorado. Atmospheric Chemistry and Physics Discussions 13:17123–17158.

436. Hugenholtz P. 2002. Exploring prokaryotic diversity in the genomic era. Genome Biol 3.

437. Jaenicke R. 1978. The role of organic material in atmospheric aerosols. PAGEOPH 116:283–292.

438. Jaenicke R, Schütz L. 1978. Comprehensive study of physical and chemical properties of the surface aerosols in the Cape Verde Islands region. Journal of Geophysical Research: Oceans 83:3585–3599.

439. Jaenicke R. 1993. Tropospheric aerosols. Aerosol-cloud-climate interactions 1.

440. Flohn H. 1973. Globale Energiebilanz und Klimaschwankungen, p. 75–124. In Die Produktionskapazität der Böden der Erde. Globale Energiebilanz und Klimaschwankungen. Springer.

220

BIOGRAPHICAL SKETCH

Rachel obtained her Bachelor of Science from the University of Illinois Champaign-

Urbana (UIUC) in 2014. While she graduated with a degree in molecular and cellular biology, she started her ungraduated career majoring in atmospheric science. After taking an introductory molecular biology course during her second year at UIUC, she decided to switch majors. During her senior year of high school, she came across an article in a popular science journal that discussed the implications of airborne microorganisms and their potential effect on meteorology.

Given her background, and passion for both biology and atmospheric science, she reached out to a microbiologist who was working on this field of aeromicrobiology, Dr. Brent Christner. At the time, Dr. Christner was a professor at Louisiana State University (LSU). After applying and being accepted into the biological sciences program at LSU, Rachel started graduate career and research in Dr. Christner’s lab in the fall of 2014. In the summer of 2015, Dr. Christner was offered a position at the Department of Microbiology and Cell Science at The University of

Florida (UF), and he accepted. Because Rachel enjoyed the research she was undertaking with

Dr. Christner as her advisor, she decided to transfer schools with him. Thus, in January 2016,

Rachel moved to Gainesville, FL, where she continued her research under Dr. Christner at UF.

Now, six years after starting graduate school, Rachel is graduating with her doctorate in microbiology and cell science from The University of Florida.

221