Environmental Pollution 135 (2005) 433–443 www.elsevier.com/locate/envpol Atmospheric nitrogen inputs to the Delaware Inland Bays: the role of ammonia Joseph R. Scudlarka,*, Jennifer A. Jenningsa,b, Megan J. Roadmana,c, Karen B. Savidgea, William J. Ullmana a College of Marine Studies, University of Delaware, 700 Pilottown Road, Lewes, DE 19958, USA b Watershed Assessment Branch, Department of Natural Resources and Environmental Control, Dover, DE 19901, USA c Bermuda Biological Station for Research, St. Georges GE 01, Bermuda Received 20 August 2004; accepted 15 November 2004 Local emissions from poultry production appear to significantly contribute to wet and dry atmospheric NHx loading to the Delaware Inland Bays. Abstract A previous assessment of nitrogen loading to the Delaware Inland Bays indicates that atmospheric deposition provides 15– 25% of the total, annual N input to these estuaries. A large and increasing fraction of the atmospheric wet flux is NHC 4 , which for most aquatic organisms represents the most readily assimilated form of this nutrient. Particularly noteworthy is a 60% increase in the precipitation NHC 4 concentration at Lewes, DE over the past 20 years, which parallels the increase in poultry production on the Delmarva Peninsula over this period (currently standing at nearly 585 million birds annually). To further examine the relationship between local NH3 emissions and deposition, biweekly-integrated gaseous NH3 concentrations were determined using Ogawa passive samplers deployed at 13 sampling sites throughout the Inland Bays watershed over a one-year period. Annual mean concentrations at the 13 sites ranged from !0.5 mg NH3 mÿ3 to O6 mg NH3 mÿ3, with a mean of 1.6 G 1.0 mg NH3 mÿ3. At most sites, highest NH3 concentrations were evident during spring and summer, when fertilizer application and poultry house ventilation rates are greatest, and seasonally elevated temperatures induce increased rates of microbial activity and volatilization from soils and animal wastes. The observed north-to-south concentration gradient across the watershed is consistent with the spatial distribution of poultry houses, as revealed by a GIS analysis of aerial photographs. Based on the average measured NH3 concentration and published NH3 deposition rates to water surfaces (5–8 mm sÿ1), the direct atmospheric deposition of gaseous NH3 to the Inland Bays is 3.0–4.8 kg haÿ1 yrÿ1. This input, not accounted for in previous assessments of atmospheric loading to the Inland Bays, would effectively double the estimated direct dry deposition rate, and is on par with the C ÿ C NOÿ 3 and NH4 wet fluxes. A second component of this study examined spatial differences in NO3 and NH4 wet deposition within the Inland Bays watershed. In a pilot study, precipitation composition at the Lewes NADP–AIRMoN site (DE 02) was compared with that at a satellite site established at Riverdale on the Indian River Estuary, approximately 21 km southwest. While C the volume-weighted mean precipitation NOÿ 3 concentrations did not differ significantly between sites, the NH4 concentration observed at Riverdale (26.3 mmoles Lÿ1) was 73% greater than at Lewes (15.2 mmoles Lÿ1). More recently, a NADP site was established at Trap Pond, DE (DE 99), which was intentionally located within the region of intense poultry production. A comparison of the initial two years (6/2001–5/2003) of precipitation chemistry data from Trap Pond with other nearby NADP– AIRMoN sites (Lewes and Smith Island) reveals fairly homogeneous NOÿ 3 wet deposition, but significant spatial differences * Corresponding author. Tel.: C1 302 645 4300; fax: C1 302 645 4007. E-mail address: scudlark@udel.edu (J.R. Scudlark). 0269-7491/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.envpol.2004.11.017 434 J.R. Scudlark et al. / Environmental Pollution 135 (2005) 433–443 (w60%) in the NHC 4 wet flux. Overall, these results suggest that local emissions and below-cloud scavenging provide a significant contribution to regional atmospheric N deposition. Ó 2004 Elsevier Ltd. All rights reserved. Keywords: Ammonia; Precipitation chemistry; Atmospheric deposition; Poultry; Estuaries 1. Introduction The Delaware Inland Bays are representative of a common but understudied class of estuarine ecosystem encountered along the Atlantic and Gulf Coasts of the United States. Such small, shallow embayments and lagoons possess several properties which distinguish them from large estuaries such as the Chesapeake Bay: (a) relatively poor flushing of large fractions of the system; (b) significant non-point sources of water and nutrients (atmospheric, groundwater and surface runoff) compared with point sources (municipal and industrial effluents); (c) shallow water depth, which enhances vertical mixing and sediment resuspension; and (d) an intensely agricultural watershed which lies entirely within the Coastal Plain physiographic province. Previous estimates of the atmospheric contribution to the total N loading in the Inland Bays range broadly from 8 to 38% (Ritter, 1986; Cerco et al., 1994; Valigura et al., 1996; Horsley and Witten, 1998; Scudlark and Church, 1999; Stacey et al., 2001). Much of the variability and uncertainty in these estimates relate to the use of convenient assumptions and generalizations which have dubious scientific justification, e.g., that dry and wet N deposition are equal in magnitude. Most of the cited studies also fail to consider the atmospheric deposition of organic nitrogen, which recent research indicates is an important component of the atmospheric N input (Cornell et al., 1995; Scudlark et al., 1998; Keene et al., 2002). In particular, due to a general lack of airborne concentration data, very few assessments of atmospheric inputs to coastal waters consider gaseous NH3 deposition. Compared with NOx, the atmospheric sources, reactivity, and deposition of NHx (NHx Z NH3 C NHC 4 ) are more poorly understood and regulated. National (Battye et al., 1994) and regional (Chimka et al., 1997) emission inventories identify agricultural operations as the dominant source of atmospheric NH3, in particular, release from animal wastes. The Delmarva Peninsula is one of the most intense poultry-production regions in the U.S., with annual production of about 585 million birds. For the past 50 years, Sussex County, Delaware, home to the Inland Bays, has been the largest broiler-producing county in the United States, with nearly 224 million birds/year produced on nearly 700 farms (DPI, 2004; USDA, 2002). Within the Inland Bays watershed alone, about 72 million birds/year are produced (DPI, 2004). Long-term precipitation chemistry records provided by the National Atmospheric Deposition Program (NADP) indicate that in several regions of the eastern concentrations have been U.S., precipitation NHC 4 increasing. Perhaps the most well-known example is at the Clinton NADP site (NC 35) located in southeastern North Carolina (Sampson County), where the doubling of the precipitation NHC 4 concentration over the past 10 years has been attributed to the rapid expansion in nearby commercial hog and poultry production (Walker et al., 2000, 2004). In the Inland Bays watershed, the precipitation NHC 4 concentration was shown to have increased nearly 60% over the past two decades, at a rate which parallels the growth in commercial poultry production on Delmarva (Scudlark and Church, 1999; Fig. 3). The steeper rate of increase at the North Carolina site is perhaps not surprising, in consideration of the larger NH3 emission factors associated with hog production (involving liquefied waste, lagoonal storage and spray irrigation) as compared with poultry production (Battye et al., 1994). It should be noted that over the same timeframe, the precipitation NOÿ 3 concentration in the Inland Bays watershed has remained relatively constant or perhaps decreased slightly (Fig. 3), which is consistent with most NADP sites in the northeast U.S. (Lynch et al., 1996). Consequently, atmospheric wet deposition of N to the Inland Bays is unique in that the input of reduced forms of N (NHC and DON) currently exceeds that of 4 oxidized forms (NOÿ 3 ) (Scudlark, 2002). The large and increasing role of NHx has several important consequences. As the dominant basic species in the atmosphere, NH3 is largely responsible for the neutralization of acid species, which influences their rate of deposition. Due to these neutralization reactions which involve rapid gas-to-particle conversion, NH3 has come under recent scrutiny with respect to fine particulate matter (PM 2.5) regulations, impacting both human health and visibility. Also due to its high reactivity, NH3 exhibits a relatively short atmospheric lifetime compared with NOÿ 3 , so that its ambient concentration is more greatly influenced by local/ regional sources. Of particular relevance to coastal waters, the increasing role of NHx deposition may also have significant ecological consequences, since the reduced forms of N are more readily assimilated by J.R. Scudlark et al. / Environmental Pollution 135 (2005) 433–443 most aquatic organisms, and thus may more rapidly or effectively induce a community response than oxidized forms (Willey and Cahoon, 1991; Paerl, 1995; Paerl and Whitall, 1999). While the parallel increase in precipitation NH3 concentration and local poultry production is highly suggestive, it does not establish a causal relationship. With this in mind, the specific objectives of this study are to: (a) examine spatial trends in airborne NH3 and precipitation NHC 4 concentrations across the Inland Bays watershed, in relation to known local sources; and (b) estimate wet and dry deposition of NHx relative to other N loading terms. 2. Experimental The Delaware Inland Bays consist of three interconnected estuaries located on the mid-Atlantic U.S. coast: Rehoboth Bay, Indian River Bay, and Little Assawoman Bay (Fig. 1). Hydrologically, Little Assawoman Bay is only weakly connected to the Rehoboth– 435 Indian River system, and is usually considered separately in water quality assessments. The Inland Bays cover a surface area of only 82 km2, of which 23 km2 is marsh. However, typical of coastal plain estuaries, the bays drain a fairly extensive and predominantly rural watershed of 750 km2 (Ritter, 1986). Freshwater inputs include surface waters (mean discharge 9 m3/s), direct groundwater discharge (0.6– 1.3 m3 sÿ1) and precipitation (averaging 115 cm yrÿ1) (Andres, 1992). Land use within the watershed is primarily agricultural (44%) and forest (44%) (Horsley and Witten, Inc., 1998). At present, less than 10% of the land use is classified as ‘‘built up,’’ but municipal and residential development is rapidly increasing, particularly in the eastern part of the watershed (Mackenzie and McCullough, 1999). Fig. 1 also depicts the location and relative size of all confined animal feeding operations within the watershed. These were initially identified using a GIS interpolation of aerial photographs, and the number of animal units and activity at each facility was subsequently verified by on-site inspection. Although there exist several small cattle, equine, dairy and swine Fig. 1. Location of the Delaware Inland Bays (Rehoboth and Indian River) and their watershed on the mid-Atlantic U.S. coast. Depicted are the confined animal feeding operations within the watershed, and the location of the 13 airborne NH3 monitoring sites employed in this study. 436 J.R. Scudlark et al. / Environmental Pollution 135 (2005) 433–443 operations, poultry production is clearly the dominant activity, which is most heavily concentrated in the western and southern part of the watershed. Most of the poultry production is for the commercial broiler and fryer market, and although many older, smaller (!30,000 bird capacity) poultry houses are still in operation, the historical trend has been toward larger, tunnel-ventilated houses (USDA, 2002). 2.1. Gaseous ammonia sampling Gaseous NH3 was measured using the Ogawa passive sampler (Ogawa & Company USA, Inc., Pompano Beach, FL). Passive sampling devices (PSDs) work on the principle of molecular diffusion through a fixed geometry, followed by chemisorption onto an acidimpregnated filter. Compared with traditional NHx measurement methods (e.g., denuders), PSDs typically lack the sensitivity to conduct measurements requiring high temporal resolution or precision, and they are not designed to sample particulate NHC 4 . However, PSDs offer the advantages of being simple, inexpensive, not requiring electric power or other ancillary equipment (pumps, mass flow controllers, etc.), which allows for unattended deployment over extended periods. Thus, with respect to the goals of this study, PSDs are ideally suited to continuously monitor the cumulative or average background NH3 concentrations at numerous remote locations. Gaseous NH3 sampling was conducted from 26 April 2000 through 10 April 2001. Measurements were conducted on a continual, bi-weekly basis at 13 sites distributed around the periphery of the bays, located at the extent of tidal influence at each of the major tributaries (Fig. 1). These sites were within regions exhibiting diverse land uses, soil types, and landcover. In consideration of the objective of gauging NH3 inputs to the surface waters of the bays, the sampling sites were thus representative of the NH3 concentrations at the estuarine boundary. To ensure a practical sample:blank ratio, the samplers were deployed for 14-day intervals, utilizing the PVC rain shelter and clips provided by the manufacturer. The shelters were deployed at the edge of vegetated riparian zones at a height of 1–1.5 m. Upon retrieval, the exposed samplers were sealed inside airtight vials, transported to the laboratory in an air-tight container and frozen until analysis. No effect of frozen storage, either prior to (%14 days) or after (%10 days) exposure, was observed. At the time of analysis, filters from both ends of the PSD were combined to double the effective sampling rate. One or more field blanks were run with each deployment, and were found to contribute 297 G 82 ng NH3 (n Z 44). For a typical 14-day deployment, the operational detection limit, defined as 2sfield blank translates to an effective concentration of 0.26 mg NH3 mÿ3. Overall precision, based on analysis of paired samplers deployed at one location for each 2-week interval, is estimated to be 18% at concentrations !1.5 mg NH3 mÿ3 and 7% at a mean concentration of 6.2 mg NH3 mÿ3. Analysis of the DI water/ammonium citrate extract was accomplished using automated, segmented flow colorimetry, based on the phenol–hypochlorite method (Solorzano, 1969) and employing an O/I Analytical Flow Solutions IV analyzer. Analytical precision was generally better than 10% except for extremely low extract concentrations (!0.5 mmoles/L). Additional details regarding the use, accuracy and precision of passive samplers for NH3 measurements can be found in Roadman et al. (2003). 2.2. Precipitation sampling Precipitation was collected in two successive sampling programs. A preliminary pilot study was conducted based on sampling at two locations: the Lewes, DE NADP–Atmospheric Integrated Research Monitoring Network (AIRMoN) site DE 02, and a satellite station established at Riverdale on the north shore of Indian River Bay (Fig. 2). Sampling at both sites was conducted on a daily basis (at 09:00 G1 h) using an automated, wet-only collector, employing AIRMoN collection and handling procedures (AIRMoN, 2002). At the Riverdale site, sampling was conducted during 1999 and 2000, from June through September. Sampling was carried out over the water at the end of a w100-ft private dock, approximately 1.5 km northeast of the 760 MW Indian River coal fired power plant, which during summer, is in the prevailing upwind direction. Winter sampling was not possible at this location because the dock is periodically submerged during winter storms. Sampling at the Riverdale site was carried out in parallel with the Lewes site, using identical collectors, materials and methods. Analyses of samples from both sites were carried out side-by-side in our laboratory using automated colorimetric methods. Independent analysis of paired sample splits from the Lewes site carried out at the NADP Central Analytical Laboratory (n Z 23) showed excellent agreement with analytical results obtained in our laboratory for both NOÿ 3 (m Z 1.02, 2 r2 Z 0.996) and NHC (m Z 1.09, r Z 0.989). 4 Due in part to the data from this pilot study, the Chesapeake Bay Program subsequently established a more permanent, year-round NADP–AIRMoN collection site at Trap Pond State Park (DE 99). This site was purposely situated within a region of intense poultry production on the Delmarva Peninsula, with a specific objective of establishing NHC wet deposition rates 4 within a high NH3 emission density region. The precipitation chemistry data from Trap Pond are compared with other nearby NADP–AIRMoN sites (Lewes and J.R. Scudlark et al. / Environmental Pollution 135 (2005) 433–443 437 Fig. 2. Location of the precipitation sampling locations at Lewes, Riverdale, Trap Pond and Smith Island. Smith Island), which reflect identical daily sampling protocols and side-by-side analyses at the NADP laboratory (AIRMoN, 2002). The data examined here span the initial two-year period of operation at Trap Pond (1 June 2001–31 May 2003) which has subsequently been converted to a weekly NADP–NTN site. 40 Precip [NH4+] 800 35 no. of chickens 700 Precip [NO3-] 30 600 25 500 20 400 15 300 10 200 5 80 85 90 95 100 2000 ÿ Fig. 3. Annual, volume-weighted mean precipitation NHC 4 and NO3 concentrations (mmoles Lÿ1) at Lewes, DE, in relation to the annual Delmarva poultry production (106 birds/year). The trend lines indicate the least mean squares fit of the data. The NHC 4 trend is statistically significant (r Z 0.64, p ! 0.005), and reflects a 58% increase over the past two decades. 3. Results and discussion 3.1. Airborne NH3 concentrations A summary of the bi-weekly average NH3 concentrations at the 13 sampling sites is presented in Fig. 4. The observed concentrations vary by two orders of magnitude, ranging from !0.1 mg NH3 mÿ3 to O15 mg NH3 mÿ3. Mean annual and seasonal NH3 concentrations across the watershed are summarized in Fig. 5. Concentrations in the northern part of the watershed (Sites 1–6) are generally consistent with regional background levels (%1 mg NH3 mÿ3). One notable exception is Site 3, which was situated near a large (w1000 head) dairy farm, about 1.8 km SE from the barns and manure holding lagoon, and only 0.4 km SE of pasture land that was occasionally spray irrigated with liquefied waste. Two other sites that exhibit somewhat elevated NH3 concentrations (1 and 4) were located near moderately traveled roads, and thus may be impacted by emissions from motor vehicle traffic, which recent studies suggest may provide an increasingly important source of atmospheric NH3 (Moeckli et al., 1996; Fraser and Cass, 1998; Kean et al., 2000; Perrino et al., 2002). Site 1 is perhaps the most noteworthy in this regard, being located only 0.4 km NNE (i.e., downwind under 438 J.R. Scudlark et al. / Environmental Pollution 135 (2005) 433–443 16 14 12 8 6 3) g/m [NH3] (µ 10 4 2 0 13 12 11 10 9 Sit 8 7 eN um 6 ber 5 4 3 2 1 6/2 5/18 5/4/00 4/10 3/27 3/12 2/26 2/12 1/23 1/8/01 12/27 12/13 11/29 11/10 10/26 10/11 9/28 9/12 8/29 8/15 7/27 7/13 6/29 6/15 le mp Sa d En te Da Fig. 4. Bi-weekly NH3 concentrations (mg NH3 mÿ3) at the 13 Inland Bays sampling sites, for the period 26 April 2000–10 April 2001. Note that sampling at Site 1 was terminated early due to repeated vandalism. prevailing winds) from a highway which serves an average of about 44,000 vehicles/day (Delaware Department of Transportation, personal communication). It is also located within the most urban/residential subwatershed (65% of land use), and is the only sub-basin containing no active poultry houses (Mackenzie et al., 1999). However, despite the high level of human activity, the NH3 concentration at this site does not appear to be significantly elevated compared with sites in the western and southern portion of the watershed [NH3] (µg NH3/m3] 12 10 8 6 4 2 0 1 2 3 4 5 6 7 8 9 10 11 12 13 Ave. Sp Su F W Site No. Fig. 5. Annual average NH3 concentrations (mg NH3 mÿ3) at each of the 13 Inland Bays sampling sites, for the period 26 April 2000–10 April 2001. Also shown is the annual arithmetic average concentration from all sites, and the seasonal averages. The error bars reflect the standard deviation based on all measurements. (Sites 7–13). In fact, the annual mean NH3 concentration at Site 1 may have a positive bias, since sampling was terminated during several winter months (see Fig. 4), when NH3 concentrations are typically lower (Fig. 5). Overall, mean annual NH3 concentrations increase to the south and west, reaching a maximum of O6 mg NH3 mÿ3 at Site 13 (Fig. 5). There does not appear to be a strong, direct correlation between land use and mean annual NH3 concentration. For example, agriculture is identified as the primary land use in the sub-watersheds surrounding Sites 2 and 5 (55% and 67%, respectively), which exhibit low NH3 concentrations. In contrast, Sites 11 and 13, which are in areas with comparable agricultural land use (53% and 70%, respectively), exhibit the largest NH3 concentrations. The major difference appears to be that in the northern part of the watershed (Sites 2 and 5), row crop and grain production (corn and soybeans) are the major agricultural activity and the number of chickens raised is relatively low (50–100 birds/acre of agricultural land), whereas agricultural lands in the more southern part of the watershed (Sites 11 and 13) support a greater poultry density (100–300 birds/acre of agricultural land; Mackenzie et al., 1999). J.R. Scudlark et al. / Environmental Pollution 135 (2005) 433–443 Previous researchers have reported inter-annual cycles in ambient NH3 concentrations at agricultural, forested and urban locations, with maximum levels observed during summer (Larsen et al., 2001; Pryor et al., 2001; Robarge et al., 2002; Walker et al., 2004; and others). This seasonality is evident at most sites in the Inland Bays watershed (Figs. 4 and 5), although there are several sites (5–7) where the NH3 concentrations remain consistently low throughout the year. The most apparent seasonality is at Site 13, where the maximum, bi-weekly summer concentration (15.6 mg NH3 mÿ3) is an order of magnitude greater than the winter minimum (!0.1 mg NH3 mÿ3). Averaged over all sampling sites, the mean summer:winter concentration ratio is 2.3 (Fig. 5), which falls between that reported for nonagricultural (2.06) and agricultural (3.13) sites in eastern North Carolina (Walker et al., 2004). The summer NH3 concentration maximum is in agreement with reported broiler house NH3 emission rates, where both NH3 production (via hydrolysis and ammonification) and ventilation rates increase as the ambient temperature rises (Walthes et al., 1997; Roadman et al., 2003; Siefert et al., 2004). Other seasonally variable factors, such as manure and fertilizer application to fields, and temperature-induced microbial activity and volatilization from soils and stored animal wastes, may also contribute to elevated summer emissions. Meteorological factors may also contribute to decreased ground-level concentrations during winter, including more frequent incursions of relatively ‘‘clean’’ oceanic air and increased atmospheric vertical mixing rates. The annual, arithmetic mean NH3 concentration measured at the 13 sites is 1.6 G 1.0 mg NH3 mÿ3 (N Z 337). This value is in good agreement with the median concentration of 1.30 mg NH3 mÿ3 measured nearby at Lewes, DE by Russell et al. (2003), based on 12-h sampling using a mist chamber technique, conducted over a 2-week period during August 2001. The mean Inland Bays NH3 concentration falls between the annual mean reported for Chesapeake Bay urban (3.3 G 2.1 mg NH3 mÿ3) and rural (1.2 G 1.0 mg NH3 mÿ3) locations (Larsen et al., 2001). 3.2. Precipitation A total of 34 paired samples were obtained over the two-year pilot study at the Riverdale and Lewes sites. These samples include only those events where precipitation occurred simultaneously at both sites and complete chemistry data were available. A comparison C of the NOÿ 3 and NH4 concentrations is plotted in Fig. 6. ÿ For NO3 , although there were occasionally large concentration differences between sites for a given event (usually related to differences in the rain amount), the volume-weighted mean concentrations at Riverdale 439 (32.1 mmoles Lÿ1) and Lewes (30.3 mmoles Lÿ1) were not significantly different. This observation is in accord with precipitation measurements conducted more than two decades ago on the grounds of the power plant, in conjunction with the EPRI/SURE study (Mueller and Hidy, 1983), which indicated that the volume-weighted mean NOÿ 3 concentration differed by less than 15% from simultaneous measurements at Lewes (Scudlark and Church, 1999). The apparent absence of a discernable power plant influence on the nearby precipitation NOÿ 3 concentration can be attributed to several factors. Firstly, the background NOÿ 3 burden inherited from more distant upwind sources may be sufficiently large to obscure the power plant emissions. Secondly, the NOx/NOÿ 3 reaction kinetics and vertical mixing rates are comparatively slow, so that very little of the local power plant NOx emissions would undergo terminal oxidation and incorporation into precipitating air masses at a sampling site in close proximity to the emission stacks. Thirdly, analysis of precipitation air mass trajectories at Lewes reveals that on average, the Riverdale sampling site is downwind from the power plant during only 37% of the events in a typical year, representing only 22% of the total incident precipitation (Scudlark et al., 1994). It should be emphasized that although these data suggest that the power plant NOx emissions do not appear to exert a large influence on the nearby precipitation NOÿ 3 concentration, they ultimately contribute to NOÿ 3 wet deposition at locations further downwind. Furthermore, the apparent absence of a significant impact on the nearby precipitation NOÿ 3 concentration should not be extrapolated to NOÿ 3 dry deposition, other power plant emissions (e.g., SO2, Hg) or air quality parameters for which NOx is a major reactant (e.g., ozone and photochemical smog). In contrast to NOÿ 3 , the volume-weighted mean precipitation concentration of NHC at Riverdale 4 (26.3 mmoles Lÿ1) was 73% greater than at Lewes (15.2 mmoles Lÿ1). This trend is consistent with the previously-cited EPRI–SURE study (Mueller and Hidy, 1983), in which the average NHC 4 concentration in precipitation sampled at the power plant was 37% greater than simultaneous measurements at Lewes (Scudlark and Church, 1999). This steep spatial gradient is also in accord with that observed for airborne NH3 at nearby sites (Fig. 5). It should be pointed out, however, that due to the absence of winter precipitation sampling, the annual differences may be somewhat less than these data indicate, since mean winter NHC 4 concentrations at this site are typically a factor of 4–5 lower than during summer (Scudlark et al., 1994). In addition to nearby agricultural emissions, the precipitation composition at Riverdale may be influenced by NHx emissions from the Indian River power plant, which utilizes Selective Non-Catalytic Reduction 440 J.R. Scudlark et al. / Environmental Pollution 135 (2005) 433–443 200 150 100 50 0 0 20 40 60 80 100 120 160 140 120 100 80 60 40 20 0 0 20 60 40 80 C Fig. 6. A comparison of NOÿ 3 and NH4 concentrations (mmoles/L) in paired precipitation samples collected at Lewes and Riverdale, DE. on their two largest units (3 and 4) during the ‘‘ozone season’’ (roughly May–September). With this technology, urea is directly injected into the furnace to chemically reduce NOx emissions to N2 and H2O. Depending on the operating conditions, excess NHx can be inadvertently released to the atmosphere, either directly or after reacting with other flue gas constituents (such as SO2). According to the 2002 Delaware Toxics Release Inventory, the Indian River power plant emitted 18,000 lbs of NH3 during the most recent year for which data are available (DNREC, 2004). However, compared with the estimated poultry emissions within the Inland Bays watershed (Roadman et al., 2003), the ‘‘NH3 slip’’ from the power plant appears to be quite small, representing !1% of the total, annual NH3 emissions. The initial two years of precipitation chemistry data from the regional NADP–AIRMoN sites (Table 1) corroborate the pilot study results: the volume-weighted mean NOÿ 3 concentrations and wet deposition do not vary significantly (G17%), while NHC 4 exhibits a steep spatial gradient across the region. If one assumes a general westerly atmospheric transport, the precipitation composition at Smith Island can be viewed as indicative of precipitating air masses transported into the region from upwind sources. For NOÿ 3 , the uniformity in NOÿ 3 concentration (and deposition) between the three sites implicates the dominance of upwind NOx emission sources. The somewhat greater precipitation NOÿ 3 concentration at Lewes may reflect the local impact of the power plant, nearby motor vehicle traffic, or possibly increased scavenging of vapor phase NOÿ 3 by alkaline sea salt aerosols along the coast and enhanced washout (Savoie and Prospero, 1982; Keene et al., 1998; Zhuang et al., 1999). In the case of NHC 4 , the significantly elevated precipitation concentration at Trap Pond appears to reflect the impact of nearby poultry operations which surround this site (Table 1). The intermediate NHC 4 Table 1 C Annual, volume-weighted mean concentrations (mmoles/L) of NOÿ 3 and NH4 in precipitation at three nearby NADP–AIRMoN sites, from 1 June 2001 to 31 May 2003 NHC 4 Site (NADP designation) Precipitation amount (cm) NOÿ 3 Concentration (mmoles Lÿ1) Deposition (kg N haÿ1) Concentration (mmoles Lÿ1) Deposition (kg N haÿ1) Lewes (DE 02) Trap Pond (DE 99) Smith Island (MD 15) 118.7 112.3 105.9 17.0 14.5 15.3 2.8 2.3 2.3 13.8 17.2 10.3 2.4 2.9 1.6 J.R. Scudlark et al. / Environmental Pollution 135 (2005) 433–443 concentration at Lewes may be the result of dispersion and deposition of upwind emissions over the Delmarva Peninsula, coupled with the greater influence of relatively ‘‘clean’’ air oceanic masses at the coast. In fact, the east–west gradient in precipitation amount (Table 1) likely reflects the increased influence of marine air masses and moisture as one approaches the Atlantic coast. It should be noted that the Lewes–Trap Pond NHC 4 concentration gradient (Table 1) is somewhat less steep than the Lewes–Riverdale gradient observed in the pilot study, which may be due to: (a) the closer proximity of the Riverdale Site to the agricultural emission sources (Figs. 1 and 2); (b) the lack of winter sampling in the pilot study, as previously discussed; and/ or (c) the somewhat lower average precipitation NHC 4 concentrations observed with the initial NADP data (Table 1) as compared with the long-term trend (Fig. 3). Limited summertime measurements at Lewes (Russell et al., 2003) indicate that gaseous NH3 comprises an average of 41% of the total atmospheric NHx, although the phase partitioning widely varies depending on the meteorological conditions. When boundary layer NH3 concentrations are low, the precipitation composition primarily reflects the in-cloud scavenging of particulate NHC and is associated with long range transport 4 (Shimshock and DePena, 1989; Asman, 1995). This appears to be the primary mechanism at Smith Island. However, in high emission areas within the Inland Bays region, the boundary layer NH3 concentrations can reach values that favor below-cloud scavenging, so that local emissions can have a direct impact on precipitation concentrations (Mészáros and Szentimrei, 1985; Walker et al., 2000). The observed enhancement of precipitation NHC 4 concentration at Trap Pond relative to Smith Island (73%) appears to be indicative of the local contribution. The gradient in precipitation concentrations observed in this study (Fig. 6 and Table 1) indicates that the previous estimates of the atmospheric N inputs to the Inland Bays (Ritter, 1986; Cerco et al., 1994; Valigura et al., 1996; Horsley and Witten, 1998; Scudlark and Church, 1999; Stacey et al., 2001), which were based entirely on data from Lewes, underestimate the spatiallyintegrated NHC 4 wet flux to the Inland Bays. Based on a simple linear interpolation of the precipitation data reported here, the actual NHC 4 wet flux to the Inland Bays appears to be about 20% greater than previously gauged. Like most estimates of atmospheric N loading to coastal waters, the above-cited studies also did not consider NH3 dry deposition. Based on the mean NH3 concentration determined here (1.6 G 1.0 mg NH3 mÿ3), and published over-water NH3 deposition velocities (5–8 mm/s; Lee et al., 1998; Larsen et al., 2001), the direct, atmospheric input of NH3 to the open waters of the bays is estimated to be 3.0–4.8 kg haÿ1. This flux is in 441 accord with the 4.2 kg N haÿ1 yrÿ1 estimate of Russell et al. (2003), which was based on independent airborne concentration measurements (cited above) and the application of a bulk exchange model (Valigura, 1995). Thus, even excluding indirect inputs via watershed transmission, the direct flux of NH3 to surface waters appears to provide a significant source of atmospheric N, C which is on par with NOÿ 3 and NH4 wet deposition. It should be pointed out, however, that during certain times of the year (especially summer) and at certain locations in the bays (particularly low salinity, highly productive/ high pH areas), NH3 can be volatilized from the water column (Larsen et al., 2001), so that the net air–water exchange in these regions may be somewhat less. 4. Conclusions The primary goal of this study was to evaluate the extent to which local emission sources contribute to the observed atmospheric N loading to the Inland Bays and their watershed, with a particular emphasis on the sources and input of NH3/NHC 4 . The wet deposition of NOÿ 3 appears to be relatively uniform, suggesting that local NOx emissions do not have a significant impact. In contrast, the gradient in airborne NH3 and precipitation NHC 4 concentrations both point to the existence of a large, regional emission source in the southwestern part of the watershed, which all evidence indicate is related to local poultry production. The direct flux of NH3 to the Inland Bays surface waters, not accounted for in previous estimates of atmospheric deposition, may provide an additional 3.0– 4.8 kg N haÿ1 yrÿ1. Similarly, results of this study suggest that earlier studies underestimated NHC 4 wet deposition by about 20%. Based on the NH3 dry deposition and revised NHC 4 wet deposition derived here, and the point and non-point loading estimates of Jennings (2003), the direct input of atmospheric N to Inland Bays surface waters is estimated to provide 15–19% of the total annual ‘‘new’’ N inputs. Acknowledgements This research was funded primarily through a grant from the U.S. EPA National Estuary Program (Contract 4D2618NANX), administered through the Center for the Inland Bays. Additional financial support was provided by the NOAA Air Resources Laboratory, under the auspices of the NADP/AIRMoN Program (Agreement 43-EA-NR-214069), and an EPA Star Grant (R826945). Appreciation is expressed to Annabella and Paul Larsen, for allowing the use of their dock and assistance with precipitation sampling. We are grateful to Jim 442 J.R. Scudlark et al. / Environmental Pollution 135 (2005) 433–443 Mulik of the EPA Methods and Testing Laboratory (retired), for his invaluable advice on the use of Ogawa passive samplers for ammonia measurements. We also wish to thank Karin Grosz of the Sussex Conservation District/DNREC Section 319, for conducting the animal census and producing the map showing active animal feeding operations (Fig. 1) that was so critical to the interpretations of this study. This report does not reflect the official views of the U.S. Environmental Protection Agency, and any mention of trade names or commercial products does not constitute and endorsement or recommendation for use by the EPA. References AIRMoN, 2002. !http://www.arl.noaa.gov/research/projects/airmon_wet. htmlO. Andres, A.S., 1992. Estimate of nitrate flux to Rehoboth and Indian River Bays, Delaware, through the direct discharge of ground water. Delaware Geological Survey Open File Report No. 35, 37 pp. Asman, W.A.H., 1995. Parameterization of below-cloud scavenging of highly soluble gases under convective conditions. Atmospheric Environment 29, 1359–1368. Battye, R., Battye, W., Overcash, C., Fudge, S., 1994. Development and selection of ammonia emission factors. U.S. EPA Report EPA/600/R-94/190, Research Triangle Park, NC. Cerco, C.F., Bunch, B., Cialone, M.A., Wand, H., 1994. Hydrodynamics and eutrophication model study of Indian River and Rehoboth Bay, Delaware. Report Prepared for the U.S. EPA. Delaware Department of Natural Resources and Environmental Control, and U.S. Army Engineering District, Philadelphia. Technical Report EL-94-5. Chimka, C.T., Galloway, J.N., Cosby, B.J., 1997. Ammonia and the Chesapeake Bay airshed. Chesapeake Bay Program. Scientific and Technical Advisory Committee Publication 97-1, 49 pp. Cornell, S., Rendell, A., Jickells, T.D., 1995. Atmospheric inputs of dissolved organic nitrogen to the oceans. Nature 376, 243–246. DNREC, 2004. Delaware department of natural resources, toxics release inventory for 2002. !http://www2.state.de.us/serc/ 02trirpt.htmO. DPI, 2004. Delmarva poultry industry. !http://www.dpichicken.orgO. Fraser, M.P., Cass, G.R., 1998. Detection of excess ammonia emissions from in-use vehicles and the implications for fine particle control. Environmental Science & Technology 32, 1053–1057. Horsley and Witten, Inc., 1998. Assessment of nitrogen and phosphorus loading to the Delaware Inland Bays. Report Prepared for the Center for the Inland Bays, Nassau, DE. Jennings, J.A., 2003. The role of land use and land cover in the delivery of nutrients to Delaware’s Inland Bays. M.S. thesis, College of Marine Studies. University of Delaware, Lewes, 163 pp. Kean, A., Harley, R., Littlejohn, D., Kendall, G., 2000. On-road measurement of ammonia and other motor vehicle exhaust emissions. Environmental Science & Technology 34, 3535–3539. Keene, W.C., Sader, R., Pszenny, A.A.P., Vogt, R., Crutzen, P.J., Galloway, J.N., 1998. Aerosol pH in the marine boundary layer: a review and model evaluation. Journal of Aerosol Science 29 (3), 339–356. Keene, W.C., Montag, J.A., Maben, J.R., Southwell, M., Leonard, J., Church, T.M., Moody, J.L., Galloway, J.N., 2002. Organic nitrogen in precipitation over Eastern North America. Atmospheric Environment 36 (28), 4529–4540. Larsen, R.K., Steinbacher, J.C., Baker, J.E., 2001. Ammonia exchange between the atmosphere and surface waters at two locations in the Chesapeake Bay. Environmental Science & Technology 35 (24), 4731–4736. Lee, D.S., Halliwell, C., Garland, J.A., Dollard, G.J., Kingdon, R.D., 1998. Exchange of ammonia at the sea surface – a preliminary study. Atmospheric Environment 32 (3), 431–439. Lynch, J.A., Bowersox, V.C., Grimm, J.W., 1996. Trends in Precipitation Chemistry in the United States, 1983–1994: An Analysis of the Effects in 1995 of Phase I of the Clean Air Act Amendments of 1990, Title IV. U.S. Geological Survey, Washington, DC (USGS 96-0346). Mackenzie, J., McCullough, K., 1999. Delaware Land-Use/Land Cover Transitions 1984–1992. !http://www.udel.edu/FREC/ spatlab/lulc/O. Mackenzie, J., Martin, J.H., Pintea, L., Boonmee, B., Gedamu, N., Thomas, T.C., 1999. Delaware Inland Bays Watershed Nutrient Management Project, Technical Report. College of Agriculture and Natural Resources, University of Delaware. !http://www.udel.edu/ FREC/spatlabO. Mészáros, E., Szentimrei, T., 1985. On the wet removal of gaseous and particulate sulfur and nitrogen species from the atmosphere. Journal of Atmospheric Chemistry 2, 405–413. Moeckli, M.A., Fierz, M., Sigrist, M.W., 1996. Emission factors for ethene and ammonia from a tunnel study with a photoacoustic trace gas detection system. Environmental Science & Technology 34, 3535–3539. Mueller, P.K., Hidy, G.M. 1983. The Sulfate Regional Experiment: Report of Findings. EPRI EA-1901. Electric Power Research Institute, Palo Alto, CA. Paerl, H.W., 1995. Coastal eutrophication in relation to atmospheric nitrogen deposition: current perspectives. Ophelia 41, 237–259. Paerl, H.W., Whitall, D.R., 1999. Anthropogenically-derived atmospheric nitrogen deposition, marine eutrophication and harmful algal bloom expansion: is there a link? Ambio 28 (4), 307–311. Perrino, C., Catrambone, M., DiMenno, A., Bucchianico, D., Allegrini, I., 2002. Gaseous ammonia in the urban area of Rome, Italy and its relationship with traffic emissions. Atmospheric Environment 36, 5385–5394. Pryor, S.C., Barthelmie, R.J., Sorenseon, L.L., Jensen, B., 2001. Ammonia concentrations and fluxes over a forest in the midwestern USA. Atmospheric Environment 35, 5645–5656. Ritter, W.F., 1986. Nutrient budgets for the Inland Bays. Report to the Delaware Department of Natural Resources, Dover, DE, 70 pp. Roadman, M.J., Scudlark, J.R., Meisinger, J.J., Ullman, W.J., 2003. Validation of Ogawa passive samplers for the determination of gaseous ammonia concentrations in agricultural settings. Atmospheric Environment 37, 2317–2325. Robarge, W.P., Walker, J.T., McCulloch, R.B., Murray, G., 2002. Atmospheric concentrations of ammonia and ammonium at an agricultural site in the southeastern United States. Atmospheric Environment 36, 1661–1674. Russell, K.M., Keene, W.C., Maben, J.R., Galloway, J.N., Moody, J.L., 2003. Phase partitioning and dry deposition of atmospheric nitrogen at the mid-Atlantic U.S. coast. Journal of Geophysical Research 108 (D21), 4656–4668. Savoie, D.L., Prospero, J.M., 1982. Particle size distributions of nitrate and sulfate in the marine atmosphere. Geophysical Research Letters 9, 1207–1210. Scudlark, J.R., Church, T.M., Conko, K.M., 1994. The wet deposition of trace elements on Delmarva and their utility as emission source indicators. Maryland Power Plant Research Program Report CBRM-AD-94-3, Annapolis, MD, 80 pp. Scudlark, J.R., Russell, K.M., Galloway, J.N., Church, T.M., Keene, W.C., 1998. Organic nitrogen in precipitation at the J.R. Scudlark et al. / Environmental Pollution 135 (2005) 433–443 mid-Atlantic U.S. coast – methods evaluation and preliminary measurements. Atmospheric Environment 32, 1719–1728. Scudlark, J.R., Church, T.M., 1999. A comprehensive reexamination of the input of atmospheric nitrogen to the Rehoboth and Indian River Estuaries. Final Report to the Center for the Inland Bays, Nassau, DE, 38 pp. Scudlark, J.R., 2002. Atmospheric nitrogen inputs to the Delaware Inland Bays: the impact of local sources. Final Report to the Center for the Inland Bays, 27 pp. Shimshock, J.M., DePena, R.G., 1989. Below-cloud scavenging of tropospheric ammonia. Tellus 41B, 296–304. Siefert, R.L., Scudlark, J.R., Potter, A.G., Simonsen, K.A., Savidge, K.B., 2004. Characterization of atmospheric ammonia emissions from a commercial chicken house on the Delmarva Peninsula. Environmental Science & Technology 38, 2769–2778. Solorzano, L., 1969. Determination of ammonia in natural waters by the phenol–hypochlorite method. Limnology and Oceanography 14, 799–801. Stacey, P.E., Greening, H.S., Kremer, J.N., Peterson, D., Tomasko, D.A., 2001. Contributions of atmospheric nitrogen deposition to U.S. estuaries: summary and conclusions. In: Valigura, R.A., Alexander, R.B., Castro, M.S., Meyers, T.P., Paerl, H.W., Stacey, P.E., Turner, R.E. (Eds.), Nitrogen Loading in Coastal Water Bodies: An Atmospheric Perspective. American Geophysical Union, Washington, DC. 443 USDA, 2002. US Department of Agriculture, Census of Agriculture, vol. 1. Geographic Area Series Part 8, AC-02-A-8. Valigura, R.A., 1995. Iterative bulk exchange model for estimating air–water transfer of HNO3. Journal of Geophysical Research 100 (D12), 26045–26050. Valigura, R.A., Luke, W.T., Artz, R.S., Hicks, B.B., 1996. Atmospheric nutrient input to coastal areas: reducing the uncertainty. NOAA Air Resources Laboratory Decision Analysis Series No. 9. Walker, J.T., Aneja, V.P., Dickey, D.A., 2000. Atmospheric transport and wet deposition of ammonium in North Carolina. Atmospheric Environment 34, 3407–3418. Walker, J.T., Whitall, D.R., Robarge, W., Paerl, H.W., 2004. Ambient ammonia and ammonium aerosols across a region of variable ammonia intensity. Atmospheric Environment 38, 1235–1246. Walthes, C.M., Holden, M.R., Sneath, R.W., White, R.P., Phillips, V.R., 1997. Concentrations and emission rates of aerial ammonia, nitrous oxide, methane, carbon dioxide, dust and endotoxin in UK broiler and layer houses. British Poultry Science 38, 14–28. Willey, J.D., Cahoon, L.B., 1991. Enhancement of chlorophyll a production in Gulf Stream surface seawater by rainwater nitrates. Marine Chemistry 34, 63–75. Zhuang, H., Chan, C.K., Fang, M., Wexler, A.S., 1999. Formation of nitrate and non-sea-salt sulfate on coarse particles. Atmospheric Environment 33, 4223–4233.