Assessing the effects of atmospheric nitrogen (N) deposition on surface water quality requires accurate accounts of total N deposition (wet, dry, and cloud vapor); however, dry deposition is difficult to measure and is often spatially variable. Affordable passive sampling methods are available for estimating “hot spots” and spatial variations of gaseous dry N deposition (i.e., nitrogen dioxide (NO2) and ammonia (NH3)), though few viable methods for estimating the deposition from nitric acid (HNO3) gas using passive sampling techniques exist. We consider passive sampling approaches for assessing spatial patterns of dry atmospheric N deposition across watersheds. We describe a method for constructing an inexpensive passive sampler (for less than $12 per unit) for monitoring spatial variations in the magnitude of HNO3 in the atmosphere. We demonstrate the applicability of passive samplers for use in watershed biogeochemical research and water quality management through a review of previous applications and via our own case study of the South Korean peninsula.
 Increased inputs of nitrogen (N) to watersheds from multiple sources have been linked to a host of ecosystem changes throughout the world, including forest health decline [Aber et al., 1995; Ollinger et al., 2002], disruption of soil processes [Yanai et al., 2005], acidification of lakes and streams [Driscoll et al., 2003; Kahl et al., 2004], and eutrophication of downstream estuaries [Paerl et al., 2002; Whitall et al., 2007]. Emissions of NOx (NO + NO2) to the atmosphere have been substantial in recent decades, resulting in a pool of reactive, oxidized nitrogen in the atmosphere (NOy, equal to NOx plus compounds that are the products of the oxidation of NOx). Emissions of both oxidized and reduced forms of N to the atmosphere are directly related to the loads and sources of deposition of N onto watersheds [e.g., Elliott et al., 2007], which in turn has significantly affected terrestrial and aquatic ecosystems [Aber et al., 1989; Lovett, 1994; Stoddard et al., 1999; Driscoll et al., 2001; Aber et al., 2003]. NOx emissions originate primarily from motor vehicles, electricity generating units, and other industrial and residential sources that burn fossil fuels, whereas NHx emissions originate primarily from agricultural sources [Galloway et al., 2004]. NOx and NHx proceed through a chain of reactions in the atmosphere (e.g., NOx to other forms of oxidized nitrogen, such as nitric acid (HNO3)), subsequently affecting rates of inorganic dry deposition (gaseous and particulate), wet deposition (via precipitation), and cloud vapor N deposition. Atmospheric N deposition has been shown to be a significant source of N pollution to many forested and mixed land use watersheds throughout the nation and world [Castro et al., 2001; Boyer et al., 2002].
 Effective techniques for estimating the magnitude and spatial distribution of N deposition are important for determining sources of N to watersheds and interpreting spatial variations in surface water quality. National monitoring networks, including the National Atmospheric Deposition Program/National Trends Network (NADP/NTN) and the Atmospheric Integrated Research Monitoring Network (AIRMoN), measure precipitation chemistry (including NO3−, NH4+) and wet N deposition at weekly (NADP/NTN), and daily and event (AIRMoN) time scales. However, dry (gaseous and particulate) N is often an important, yet poorly estimated, contributor to the total inorganic atmospheric N budget [Lovett, 1994]. The Clean Air Status and Trends Network (CASTNET, formerly National Dry Deposition Network) provides the most comprehensive, standardized measurements of weekly averaged dry N deposition (NH4+, NO3−, HNO3) at sites throughout the USA. These records are beneficial for analyzing long-term trends in atmospheric chemistry and deposition at specific points on the landscape and sources of dry deposition (E. M. Elliott et al., Dual nitrate isotopes in actively and passively collected dry deposition: Utility for partitioning sources, understanding reaction pathways, and comparison with isotopes in wet nitrate deposition, submitted to Journal of Geophysical Research, 2008); however, the stations are sparsely located, are costly to maintain, and potentially underestimate dry N deposition, as dry NO2, NH3, and organic oxidized N are not measured at these sites. The consequent lack of data, difficulties measuring dry deposition and high spatial variability of particulate and gaseous N deposition in the absence of precipitation pose a distinct challenge for estimating rates of dry N deposition to watersheds. As a result, dry deposition is often not evaluated in watershed-based ecosystem sensitivity assessments [Grimm and Lynch, 2004], or it is estimated using a wet to dry ratio based on a synthesis of regional observations [Lovett, 1994].
 The application of passive sampling techniques for monitoring ambient air chemistry in ecological research, including determining the spatial distribution of gaseous atmospheric N species such as NO2, NH3, and HNO3 [Krochmal and Kalina, 1997] and detecting hot spots of atmospheric N activity [Ferm and Svanberg, 1998; Plaisance et al., 2002], is a relatively recent development. Only a limited number of studies have linked passive sampling of atmospheric N to water quality and watershed processes and management [e.g., Hunsaker et al., 2007; Elliott et al., submitted manuscript, 2008]. Passive samplers provide an alternative to active atmospheric chemistry monitoring systems, which are energy, labor, and financially intensive to maintain, thereby lowering costs and labor while providing relatively accurate first-order estimates of gaseous contributions to watershed N inputs. Several commercial passive samplers are available for monitoring atmospheric levels of NO2 and NH3 (e.g., Ogawa and Company, Pompano Beach, Florida; Rupprecht and Patashnick Radiello® models). However, only a few studies [Bytnerowicz et al., 2005] have described current passive sampling techniques for detecting gaseous atmospheric nitric acid (HNO3), a principle component of photochemical smog, for water quality assessments and ecosystem sensitivity studies. This paucity of studies indicates a clear need for an easily constructed, affordable, and validated HNO3 passive monitoring technique.
 The goals of this technical note are (1) to evaluate the potential for using passive atmospheric N sampling techniques in watershed biogeochemical research, (2) to present an innovative, simple, and inexpensive sampling technique for assessing the spatial distribution of ambient atmospheric HNO3 for these purposes, and (3) to report results of a pilot project testing this new passive sampler technique. This research is delivered at a key juncture when passive sampling of atmospheric N inputs to watersheds is on the rise.
2. Passive Atmospheric N Sampling: History and Potential Use for Watershed Research
 The standard classic atmospheric contaminant passive monitor is the Palmes tube, which was originally used for indoor air quality monitoring and relies on a small area-to-length ratio to inhibit mixing and turbulence at the tube opening [Cadoff and Hodgeson, 1983; Plaisance et al., 2002]. The theoretical basis for passive sampling design is addressed in several review papers [Fowler, 1982; Rose and Perkins, 1982; Brown, 1993; Krupa and Legge, 2000; Cox, 2003; Sather et al., 2006]. Briefly, passive diffusion involves the flow of air across the volume of an open-ended cylindrical sampler onto a surface or absorbent according to Fick's first law of diffusion, which describes the movement of material from areas of high concentration to low concentrations. The diffusion path length and the internal sampler volume control rates of gaseous adsorption onto the filter medium [Cox, 2003]. Deposition estimates can be calculated from measured atmospheric N concentrations using an appropriate deposition velocity [Seinfeld and Pandis, 1998], though they are often difficult to estimate.
 Several studies provide overviews of the development and application of passive sampling techniques for ecological sensitivity and environmental assessment studies [Krupa and Legge, 2000; Cox, 2003; Namiesnik et al., 2005], but few exist for watershed processes or water quality applications. Other literature from North America, Europe, and Asia demonstrate the utility of passive samplers for providing relatively precise and accurate estimates of different N species in a variety of settings (Table 1). These published studies suggest that passive atmospheric N sampling techniques could be readily transferable to studies focusing on watershed riverine N fluxes, watershed processes, and watershed nutrient management; however, the application of passive sampling for these applications has been widely underutilized. Passive atmospheric N samplers provide a simple, cost-effective method for identifying the relative magnitude, spatial patterns, and species of N concentrations across multiscale heterogeneous watersheds, though caution is urged when translating these relative patterns to absolute values of deposition (see below). Moreover, because national monitoring networks are typically located in rural areas away from local sources of emissions, passive samplers provide a method for detecting the relative magnitude of dry N species in areas where local redeposition of NOx or NHx emissions occur, and can capture changes in local variations of atmospheric “hot spots” of N at a much finer spatial scale. This, in turn, can be linked to variations in surface water quality. The use of passive samplers for other applications, such as analyzing the isotopic composition of dry nitrate deposition (δ18O and δ15N) from passively sampled atmospheric HNO3 (Elliott et al., submitted manuscript, 2008), could assist in detecting relative sources and pathways by which important dry N species reach watersheds, particularly at refined gradients. Further, combining isotopic results from passive sampling with that of surface water chemistry could provide insights to the relative proportion of atmospheric N reaching surface water systems.
Table 1. Examples of Recent Studies Demonstrating the Effectiveness of Passive Sampling Techniques for Measuring Levels of Gaseous N Species in the Atmosphere for Ecological Applicationsa
3. A Simple Passive Gaseous Nitric Acid Monitoring Technique
 Nitric acid (HNO3) typically exists in low concentrations in the atmosphere, but when NOx emissions combine with water vapor, HNO3 is produced and quickly reacts with ammonia, forms particulate or aerosol NO3−, and is rapidly removed from the atmosphere within a few days of its production [Seinfeld and Pandis, 1998]. Bytnerowicz et al. [2002, 2005] provide one of a limited numbers of studies estimating background and elevated levels of atmospheric gaseous HNO3 for ecological research using simple, affordable passive sampling methods (see Bytnerowicz et al.  for a detailed diagram and sampler design). Because of the proven effectiveness of these samplers [Bytnerowicz et al., 2005; Hunsaker et al., 2007], we attempted to design replicates for a study monitoring spatial gradients in the concentration of HNO3. As part of this process, we discovered that one of the important components (the thick Teflon insert rings measuring 47.8 mm) was difficult to purchase and required custom fitting, which was fairly expensive. In response to this challenge, we sought a more affordable method and designed a passive HNO3 sampler from commercially available monitoring cassettes that are specified for vacuum sampling of particulate matter and gases for industrial hygienic sampling. Our sampler conceptually follows Bytnerowicz et al.  and is similar in appearance to Rabaud et al.  (Figure 1). Table 2 provides a complete listing of sampler components and prices as of June 2008. Using the least expensive vendors included in the tables, we constructed HNO3 samplers for under $12 per unit.
Table 2. Sampler Parts and Sources for 37-mm Passive HNO3 Sampling Units
Part numbers and prices are specified from Pall Life Sciences, www.pall.com or VWR scientific, www.vwrsp.com. Prices are current as of June 2008. Using the least expensive vendor prices above, we built samplers (including storage containers) for a price of $11.78 each.
NA means not available.
Air monitoring cassettes, 37 mm, three-piece unit, unassembled, 100 per package
Housing for HNO3 sampler
Spacer cassettes: extra spacer ring for air monitoring cassettes, pack of 250
Expands system above to four pieces, to house two filters in sequence
Zefluor PTFE membrane disc filters, 2.0 μm, 37 mm, with PTFE support pads, 50 per pack
Outer prefilter; the PTFE support pads are used as base for Nylasorb filters
P5P JO37 ($149)
Nalgene polypropylene straight-sided jars, 125mL; case of 3 packs of 12 (36 units)
Container for storage of assembled samplers.
Total for all parts
Total for one unit
 The housing for our HNO3 monitoring system consists of a fitted four piece 37-mm styrene acrylonitrile (SAN) (4.1 cm diameter) air monitoring cassette. We have adjusted the standard commercial “three-piece” cassette with a second inner spacer ring to accommodate an extra filter. Each HNO3 unit can be mounted vertically outdoors for passive monitoring of HNO3. Ambient air first diffuses through a 37-mm polytetrafluoroethylene (PTFE) Zefluor™ membrane 2.0 μm prefilter, which is separated approximately 13 mm (using a SAN spacer ring) from a 37-mm (33 mm effective sampling diameter; 855.3 mm2 effective surface area) Pall® Nylasorb™ (1.0 μm) nylon filter. Nitric acid is selectively absorbed onto this nylon filter. A 37-mm PTFE support pad (purchased with the prefilters; see Table 2) is placed at the base of the sampler directly beneath the nylon filter. One sample of HNO3 (analyzed as NO3−) can be extracted from each sampler. Sampling unit assembly from the top of the sampler to open sampler face is as follows: SAN base (outlet) with plug, PTFE support pad, Nylasorb™ filter, SAN spacer ring, Zefluor™ membrane filter, SAN spacer ring, and SAN end cap (inlet) with plugs. The end, or inlet, cap with plug simply protects the filters during transport, and is taken off (carefully; we typically use a prying tool) prior to field exposure of the filters. All of the SAN pieces are precleaned and dried with deionized water prior to assembly, and care during assembly is required so as not to contaminate the filters (e.g., from fingers, lab benches, and ambient gases in the air). We recommend a clean laboratory space, and the use of gloves and forceps. We store the completed samplers in individual airtight Nalgene™ containers and refrigerate them until deployment. If deployed in an outdoor field setting, a protective barrier is needed to cover the unit during field sampling, to help protect the filters from rain and snow. We have typically used clear plastic cups to serve this purpose (see photos of field deployments in auxiliary material). After field exposure is complete, the HNO3 cassette samplers should be covered with the inlet cap, sealed in an airtight container, and transported on ice to the lab. Information regarding the filter extraction and analysis procedures is provided in the pilot study section below.
 We base the predicted utility of this sampler on our pilot study (see below) and results from several studies that evaluate the efficacy of passive samplers [Bytnerowicz et al., 2002; Gilbert et al., 2003; Roadman et al., 2003; Mukerjee et al., 2004; Bytnerowicz et al., 2005]. For example, Bytnerowicz et al.  present calibration results using passive samplers to monitor ambient HNO3 in the atmosphere, comparing their sampler to accuracy-tested honeycomb denuder systems in field trials in Riverside, California, with promising success rates (CV 9.2% in field calibrations). We propose that our sampler would retain HNO3 in a similar manner because (1) our sampler's structure is similar to Bytnerowicz et al. , though slightly smaller in diameter, (2) nylon is a proven collection medium for HNO3 gas [Parrish, 2000], (3) Teflon prefilters effectively capture particulate matter that would otherwise influence NO3− solvent analysis for ambient HNO3 levels [Bytnerowicz et al., 2005], and (4) tests suggest that the construction and design of passive samplers can be altered with similar, accurate results [Ferm and Svanberg, 1998]. An additional barrier that diffuses air at a consistent, measurable rate would, however, greatly enhance the open face HNO3 passive sampling technique [Bytnerowicz et al., 2002].
 The potential for use of our alternative HNO3 samplers in watershed research and management is promising; however, numerous validation experiments remain to assess their accuracy. First, one of the most important aspects of developing new technologies for measuring atmospheric HNO3 is the challenge of estimating deposition velocities (length per unit time), which are proportionality constants between ambient concentrations of an atmospheric gas and its depositional fluxes. Deposition velocities of HNO3 are heterogeneous across fine spatial and temporal scales, varying across gradients such as land cover and micrometeorological conditions [Dollard et al., 1987]. As a result, although passive samplers utilized in watershed studies can suggest the relative magnitude and spatial variability of HNO3 concentrations (i.e., where “hot spots” are located and whether spatial gradients in concentrations exist), estimates of absolute magnitudes of HNO3 deposition would require extensive measurements to calibrate deposition velocities at each individual study site. This is a significant task. Additional challenges remain, as HNO3 deposits readily onto most surfaces, reflecting its “sticky” properties as a compound. Thus, consideration must be given to the attachment of HNO3 to the sampler inlet or absorption onto the Teflon prefilter [Bytnerowicz et al., 2005]. Tests are needed to show how varying heat, pressure, and moisture conditions in ambient air and at the inlet influence this property, thus the ability of passive samplers to produce absolute, rather than relative, values of HNO3. Calibration of the HNO3 passive samplers alongside well-instrumented, standard methods, such as those employed at CASTNET sites, is also required. However, calibrating samplers against active (e.g., mechanized) samplers with a standardized precision range in a variety of settings (i.e., different land covers, elevations, etc.) might negate the location-based biases of national network comparisons [Bytnerowicz et al., 2002, 2005]. Finally, geochemical tests are needed (1) to estimate the influence of diffusion path length, effective sampling surface area, and sampler type on measured HNO3 concentrations and (2) to evaluate whether the composition of 37-mm cassette HNO3 filters (e.g., PTFE or acrylic) affect NO3− concentrations in solution eluates as a result of leaching processes. For example, Shooter et al.  found that solubility of NO2 in the plastic sampler materials might influence NO2− concentration measurements, but diffusion through plastic walls of sampler did not.
4. Pilot Study: Nitric Acid Deposition Across Watersheds of the South Korean Peninsula
 Rates of atmospheric N deposition to watersheds of the South Korean peninsula are a growing concern because of transboundary air pollution from industrializing areas to the north and west. In fact, recent research in the Republic of Korea (South Korea) [Bashkin et al., 2002] suggests a net surplus of nitrogen as a result, in part, of industrialization activities that lead to high rates of atmospheric N deposition. However, estimating spatial patterns of atmospheric N deposition to watersheds are required to more accurately assess the effect of atmospheric pollution on aquatic ecosystems.
 We estimated the spatial distribution of dry gaseous atmospheric nitrogen (N) deposition across the peninsula during July 2005 using passive sampling techniques. Here we focus on nitric acid (HNO3) measured at four sites in different watersheds throughout central South Korea, each with varying elevation and land uses, using our new passive atmospheric sampling technique (Table 3). The sites include Mount Changwang in Gangwan-Do, Chilbosan University Forest near Suwon, Taehwasan University Forest near Gwangju, and the Korea Forest Service office in Pyeongchang. The four sites are located at varying elevations ranging from 67 to 1176 m and include a diversity of watershed land uses (Table 3). Six or nine passive samplers were exposed at each site during one sampling round ranging from 14 to 18 days, depending on site-specific weather conditions. Included as auxiliary material to this article, we share photos to illustrate field deployments of passive samplers in various environmental settings.
Table 3. Sampling Sites Using Passive HNO3 Samplers Across the South Korean Peninsula
Adjacent Land Use
Chilbosan University Forest near Suwon
high-density industry, urban
Taehwasan University Forest near Gwanju
Pyeong Chang, Korea Forest Service Office
agriculture, low-density residential
Chungwant Mountain, Gangwan-Do
 After a period of deployment, samplers were retrieved and stored in clean, airtight containers and kept cold during travel. We froze the samplers (in the airtight containers) until analysis. Each filter was extracted from its sampler no more than 4 days after freezing. We recommend a clean lab setting and the use of gloves and forceps, aiming not to contaminate or disturb the HNO3 filter surface. Each HNO3 filter was eluted in 10 mL of high purity (18.2 Mohm cm) deionized water, and gently shaken in a small precleaned ehrlenmeyer flask on a shaking table for 30 min. We tested various shaking times for the elution, and deemed 30 min to be generously sufficient to ensure leaching of HNO3 from the filter. Concentrations of the HNO3 in the eluted solution were measured as NO3− using a Foss Tecator FIA Star 5000 N/P Flow Injection Analyzer (Sweden), whereas in other applications (E. Boyer, personal communication, 2007) we have used a Dionex ion chromatograph. In the spirit of quality assurance, we recommend work with blanks (e.g., fully assembled sampling units that are not deployed in the field but are treated the same way analytically); our work with such blanks identified a “clean” HNO3 filter.
 In the absence of calibrated deposition velocities, our first approximations of HNO3 fluxes were derived from the eluted solution concentration. We multiplied the solution concentration by the elution volume and normalized the product using the effective filter area (8.553 cm2) and the number of days the filters were exposed. On the basis of these estimates, dry HNO3 fluxes ranged from 0.016 to 0.032 kg ha−1 a−1 (4.3 to 8.9 μg m−2 d−1), well below the 10–11 kg ha−1 a−1 total atmospheric N input estimates for the region [see Bashkin et al., 2002]. Although fluxes of HNO3 are low, the relative magnitude of HNO3 suggests potential relationships between gaseous HNO3 deposition and elevation (Figure 2) and between gaseous HNO3 deposition and land use. For example, Chilbosan University forest, the lowest-elevation site surrounded by predominately urban land use, had the highest HNO3 fluxes (0.032 kg ha−1 a−1) compared to the high elevation, rural forested location at Chungwang (0.016 kg ha−1 a−1). In addition to the low flux values resulting from the lack of calibrated deposition velocities, these extremely low HNO3 flux values also potentially reflect two meteorological factors occurring within the Korean peninsula during July: (1) the monsoon season, during which deposition would most likely be delivered via precipitation and (2) a directional shift in wind patterns from the northwest, which is the predominate source of atmospheric emissions, to the southeast, an area with lower nitrogen emissions.
 As a result of this pilot study, we suggest that our passive sampling techniques for watershed research demonstrate potential; however, standardized methods for converting passively sampled atmospheric concentrations of HNO3 to deposition are limited, which is attributed to the high degree of spatial and temporal variability of deposition velocities (and consequent challenges estimating them). Therefore, our methods might have also contributed to underestimates of fluxes in the South Korean study. Further, because diffusion coefficients are difficult to estimate for our HNO3 sampler without tested calibration methods, translating filter solution concentrations to ambient air concentrations using Fick's first law of diffusion proves difficult. Caution should also be exercised when estimating ambient atmospheric concentrations of HNO3 in the absence of calibration, particularly because of the propensity of HNO3 to readily attach to surfaces. Passive sampling also provides only time-averaged concentrations [Shooter et al., 1997]; thus, this method overlooks diurnal, daily, and weekly fluxes in concentration levels, which are particularly important for watershed research in higher-elevation environments [Yamada et al., 1999]. Finally, practical concerns, such as the possible contact of rain, snow, or sleet entering from beneath sampler shelters and potential leaching during filter storage and transport also need to be considered. Given these limitations, however, passive atmospheric sampling techniques provide considerable insight to the relative magnitude of HNO3 deposition across heterogeneous watersheds throughout the South Korean peninsula. Application of these techniques using established commercial samplers and our new passive sampler for HNO3 should prove useful in future biogeochemical watershed research.
 We are grateful for financial support for this work through grants from the National Science Foundation East Asia and Pacific Summer Institutes Program, the Environmental Protection Agency Greater Research Opportunities program, and the New York Energy and Research Development Authority. We would like to thank the lab group of Don Koo Lee at Seoul National University for field work assistance. We appreciate thorough advice on this project from Russ Briggs and René Germain; interesting discussions on dry deposition associated with a related project from Doug Burns, Tom Butler, and Carol Kendall; and helpful information about passive sampling of HNO3 from Stuart Weiss. We would also like to thank three anonymous reviewers for helpful comments that substantially improved the manuscript. This paper has been reviewed in accordance with the U.S. Environmental Agency's peer and administrative review policies and has been approved for publication. Approval does not signify that the contents necessarily reflect the views and policies of the agency, nor does the mention of trade names or commercial products constitute endorsement or recommendation for use.