United States Environmental Protection Agency National Exposure Research Laboratory Research Triangle Park NC 27711 Research and Development EPA/600/SR-98/027 September 1998 Project Summary Clean Air Status and Trends Network Deposition Summary Report (1987-1995) The National Dry Deposition Network (NDDN) was established in 1986 to pro- vide long-term estimates of dry acidic deposition across the continental U.S. In 1990, NDDN was incorporated into the Clean Air Status and Trends Net- work (CASTNet) which was created to address the requirements of the Clean Air Act Amendments (CAAA). Approxi- mately 50 standard sites were opera- tional from 1989 through 1995 with the majority of the sites located in the rural eastern U.S. Each site is equipped with sensors for continuous measurements of ozone (O3) and meteorological vari- ables required for estimation of dry deposition rates. Weekly average at- mospheric concentrations of panicu- late sulfate (SO42), paniculate nitrate (NO3-), paniculate ammonium (NH4+), sulfur dioxide (SO2), and nitric acid (HNO3) were measured at all sites. Precipitation samples were collected at selected sites and analyzed for acidity and related species in order to esti- mate wet deposition. Under CASTNet, a visibility monitoring network and a Mountain Acid Deposition Program (MADPro) were established. A mi- crometeorological model has been ap- plied to calculate deposition velocities and dry deposition fluxes. Atmospheric concentration data showed species-dependent variability in space and time. In general, the high- est annual concentrations [>6.0 micro- grams per cubic meter (ug/m3) for SO42 ] were observed along the Ohio River valley, and these were a factor of 5 to 10 times higher than concentrations observed in the west. Significant con- centration gradients were observed be- tween the northeast and upper northeast, and the Midwest and upper Midwest. A comparison of 1989 and 1995 annual concentrations show 8O2 concentrations averaged over all east- ern sites decreased by 43%. Sulfate concentrations decreased by 23%. An- nual average concentrations of nitro- gen species showed little change between 1989 and 1995. These results do not account for the yearly variabil- ity in concentrations. Figure 1. A typical CASTNet monitoring station (Site 161, Gothic, CO). Dry deposition calculations for 1987 through 1995 showed that SO2 accounts for more than 70% of dry sulfur deposi- tion at eastern sites and more than 55% of dry sulfur deposition at west- ern sites. HNO3 accounts for more than 65% of dry nitrogen deposition at all sites. Data for all eastern sites showed a 29% reduction in SO2 deposition and a 6% reduction in SO42- deposition from 1989 to 1995. There is no apparent trend for western sites. The dry deposition calculations represent lower bound es- timates of actual fluxes and do not ac- count for quantified uncertainties of the model. Annual precipitation concentrations of SO/- from 1989 to 1995 declined significantly only in certain parts of the eastern region. There were no statisti- cally significant trends in precipitation concentrations of NO3-. Wet depositions Printed on Recycled Paper ------- from 1989 to 1995 showed statistically significant reductions of sulfur species for the eastern region. No statistically significant reductions were observed for nitrogen species. Total (wet plus dry) deposition esti- mates from 1989 to 1995 showed that dry deposition accounts for about 15 to 45% of total sulfur deposition, and 20 to 60% of total nitrogen deposition. The total deposition data showed a 32% reduction in atmospheric sulfur and no trend in nitrogen from 1987 through 1995. Ozone data indicated considerable geographic variability in annual aver- ages, but little yearly variability at indi- vidual stations. There was no discernible trend in annual averages. Hourly concentrations above the 1-hour National Ambient Air Quality Standard (NAAQS) were limited to sites in the Washington-New York corridor. Ozone measurements show that about 80% of the eastern sites would have recorded violations of the new 8-hour O3 stan- dard. Data from MADPro for 1994 to 1996 showed that cloudwater can be the pri- mary pathway for deposition of pollut- ants to high elevation ecosystems. An initial analysis of visual air qual- ity measurements taken in 1994 shows a strong relationship among atmo- spheric light scattering, fine particle concentrations, and visual quality. Fine particle concentrations peaked in the summer and were highly correlated with fine SO42% which is declining in the east- em U.S. Fine particle mass concentra- tions were below the proposed national standards of 15 ug/m3 annual average and 65 ug/m3 24-hour average. This Project Summary was developed by the U.S. Environmental Protection Agency's (EPA's) National Exposure Re- search Laboratory, Research Triangle Park, NC, to announce key findings of the research project that is fully docu- mented in a separate report of the same title (see Project Report ordering infor- mation at the back). Introduction Atmospheric deposition takes place via two pathways: wet deposition and dry deposition. Wet deposition is the result of precipitation events (rain, snow, etc.) which remove particles and gases from the at- mosphere. Dry deposition is the transfer of particles and gases to the landscape in the absence of precipitation. Wet deposi- tion rates of acidic species across the U.S. have been well documented over the last 10 to 15 years; however, comparable information is unavailable for dry deposi- tion rates. This lack of information in- creases the uncertainty in estimates of interregional, national, and international transport and confounds efforts to deter- mine the overall impact of atmospheric deposition. In 1986, the EPA contracted with Envi- ronmental Science and Engineering, Inc. (ESE) [now QST Environmental Inc. (QST)] to establish and operate the Na- tional Dry Deposition Network (NDDN). The objective of the NDDN was to obtain field data at approximately 50 sites throughout the U.S. to establish patterns and trends of dry deposition. The approach adopted by the NDDN was to estimate dry deposition using measured air pollut- ant concentrations and modeled deposi- tion velocities estimated from meteorological, land use, and site charac- teristic data. The model currently used for dry deposition calculations is a multi-layer version of the Big Leaf Model. The complete report summarizes results of NDDN and CASTNet monitoring activi- ties from 1987 through 1995. Concentra- tion and deposition data for atmospheric sulfur and nitrogen species are presented and analyzed. Annual and seasonal aver- ages for each year are discussed region- ally. Relative contribution of gases versus aerosols are evaluated. Wet deposition data for 21 CASTNet and 38 National Atmospheric Deposition Program (NADP) sites are presented and then used, along with dry deposition calculations, to esti- mate total depositions of sulfur and nitro- gen. The relative magnitude of wet and dry deposition is discussed. Ozone con- centrations and exposure statistics are analyzed and discussed in terms of exist- ing and proposed national air quality stan- dards. Data and results from the visibility and mountain acid deposition programs are briefly discussed. A mobile system for the direct measurement of dry deposition is described along with field studies in which the system was successfully utilized. Ini- tial results of the comparison between fil- ter packs and annular denuders are also given. Procedures Measurements of ambient O3, SO2, SO/-, NO?-, HNO3, NH/, and meteoro- logical variables required for dry deposi- tion calculations are taken at each CASTNet site. Meteorological variables and O3 concentrations are recorded con- tinuously and reported as hourly averages consisting of a minimum of nine valid 5- minute averages. Atmospheric sampling for sulfur and nitrogen species is inte- grated over weekly collection periods us- ing a 3-stage filter pack. In this approach, particles and selected gases are collected by passing air at a controlled flow rate through a sequence of Teflon®, nylon, and base-impregnated cellulose (Whatman) filters. The Teflon® filter re- moves particulate SO/-, NO3-, and NH4+; the nylon filter removes HNO3; and the cellulose filter, which is impregnated by potassium carbonate (K2CO3), is used for removal of SO2. In practice, a fraction (usually <20%) of ambient SO2 is cap- tured on the nylon filter. The nylon filter SO2 and Whatman filter SO2 are therefore summed to provide weekly average con- centrations. The nylon filter HNO3 is con- verted to NO3- and added to the Teflon® filter NO3- to provide weekly total NO3- concentrations. Filter packs are prepared and shipped to the field weekly and exchanged at each site every Tuesday. Blank filter packs are collected monthly to evaluate passive col- lection of particles and gases as well as contamination during shipment and han- dling. At 21 sites located more than 50 kilometers (km) from NADP sites, precipi- tation samples are collected weekly (ac- cording to NADP protocols) and shipped to QST for chemical analysis. Filter pack samples and O3 measure- ments are taken at 10 meters (m) using a tilt-down aluminum tower (Aluma, Inc.). Filter pack flow is maintained at 1.50 liters per minute (Lpm) at eastern sites and 3.00 Lpm at western sites, for standard conditions of 25 degrees Celsius (°C) and 760 millimeters of mercury (mmHg) with a mass flow controller (MFC). Precipitation samples are collected in precleaned poly- ethylene buckets using an Andersen Model APS precipitation sampler. Buckets are placed on the sampler on Tuesday and removed, whether or not rainfall has oc- curred, the following Tuesday. Buckets are weighed in the field, decanted to a 1-liter (L) polyethylene bottle, sealed, and shipped to QST for chemical analysis. Pre- cipitation amount (depth) is also moni- tored at wet deposition sites. O3 is measured via ultraviolet (UV) ab- sorbance with a Thermo-Environmental Model 49-103 analyzer operating on the 0- to 500-part per billion (ppb) range. Am- bient air is drawn from the 10-m air quality tower through a 3/8-inch TFE Teflon® sampling line. Teflon® filters housed at the tower inlet and the analyzer inlet pre- vent particle deposition within the system. Periodic checks indicate that online losses through the inlet system are consistently less than 3%. Zero, precision (60 ppb), ------- •142 Legend o Dry deposition only • Dry and wet deposition Figure 2. CASTNet monitoring sites (Source: QST). and span (400 ppb) checks of the O3 analyzer are performed every third day using an internal O3 generator. In addition, various observations are periodically made at the CASTNet sites to support model calculations of dry deposi- tion. Site operators record surface condi- tions (e.g., dew, frost, snow) and vegetation status weekly. Vegetation data are obtained to track evolution of the domi- nant plant canopy, from leaf emergence (or germination) to senescence (or har- vesting). Once a year, site operators also provide information on major plant spe- cies and land-use classifications within 1.0 km of the site. Additional land-use data were obtained by digitization and analysis of aerial photographs obtained from the U.S. Geological Survey (USGS) National Cartographic Information Center in Reston, VA. Leaf area index (LAI) measurements were taken at all CASTNet sites during the summers of 1991 and 1992. LAI was measured using an LAI-2000 Plant Canopy Analyzer manufactured by Li-Cor (Lincoln, NE). All field equipment is subjected to quar- terly inspections and multipoint calibra- tions, using standards traceable to the National Institute of Standards and Tech- nology (NIST). In addition, independent equipment audits were performed annu- ally by Ogden Environmental and Energy Services, Inc. (Ogden), and randomly by EPA or its designee. Results of field cali- brations are used to assess sensor accu- racy and flag, adjust, or invalidate field data. Following receipt from the field, exposed filters and blanks are extracted and then analyzed for SO42- and NO3- by micromembrane-suppressed ion chroma- tography (1C). Teflon® filter extracts are also analyzed for NH4+ by the automated indophenol method using a Technicon II or TRAACS-800 Autoanalyzer system. Wet deposition samples are filtered and then analyzed for pH, conductivity, acid- ity, sodium (Na*), potassium (K*), NH/, calcium (Ca2+), magnesium (Mg2+), chlo- ride (CI-), nitrite (NO2-), NO3-, and SO42-. A micrometeorological model called the multilayer model (MLM) was used to simu- late deposition velocities for the measured ambient species as the inverse sum of three separate resistances: atmospheric resistance (Ra), boundary layer resistance (Rb), and canopy resistance (Re). Dry depositions were then calculated as the product of concentrations and deposition velocities. An analysis of the uncertainties in simulated deposition velocities suggest that the MLM underestimates observed deposition velocities and consequently dry depositions. Since the uncertainties in the model have not been quantified, calcu- lated dry depositions represent lower bound estimates only. Results and Discussion The CASTNet measurements show sta- tistically significant reductions in annual ambient SO2, SO42-, and HNO3 concentra- tions averaged over all eastern sites. The data for the eastern sites combined show a 23% reduction in SO42- and a 43% re- duction in SO2 between 1989 and 1995 annual concentrations. Changes in sulfur species are more pronounced in the north- east and smaller, but still substantial, in the southeast. There is no apparent trend in the data from the western sites. The eastern data indicate about 70% of ambi- ent sulfur is in the form of SO2. These comparisons do not account for year-to- year variations in concentrations. Analy- ses that address these variations may yield slightly different results. ------- SOj concentrations show significant re- ductions in summer and winter, even though concentrations are much lower in summer when SO2 emissions are involved in photochemical reactions. A slight reduction in HNO3 concentra- tions results from the decline in summer averages. The downward trend is observed in most of the network, except for the southern peripheral sites and the western sites. No trends are observed in annual concentrations of NO3- aerosol and total N9r The geographic distributions of annual dry sulfur depositions (fluxes) are similar from year to year and correspond well with distributions of SO2 and SO42- con- centrations and reflect the locations of SOX sources. Annual fluxes of total nitrogen show a fairly uniform pattern of deposi- tions above 2.0 kg/ha (as N) from year to year. Lower depositions were calculated for the CASTNet sites in the Appalachian chain, similar to the results for HNO3 con- centrations. Time series of weekly fluxes show con- siderable site-by-site and weekly variabil- ity. The seasonal cycles of fluxes of individual species are similar to the sea- sonal behavior of concentrations. Scattergrams and correlation statistics show a strong relationship between con- centrations and depositions of SO2, SO/-, and HNO3. Results of linear regressions of dry depositions versus year for all eastern sites combined show downward trends for the sulfur species although the trend lines are not considered statistically significant. The downward slopes for SO2 and total sulfur are greater than for SO/-, consistent with the results for concentrations. The data show a 29% reduction in dry deposition of SO2 (as S) and only a 6% reduction in deposition of SO/- aerosol. These per- centage reductions in depositions are lower than the comparable reductions in con- centrations. The linear regressions show no significant trends in depositions of the nitrogen species. Depositions of sulfur and nitrogen species calculated for the west- ern sites exhibit no trends. Concentrations of anions and cations measured in precipitation samples col- lected at CASTNet sites and concentra- tions measured at those NADP sites approximately collocated with CASTNet sites that do not collect precipitation samples were combined to form one con- centration and wet deposition database. Annual SO/- concentrations exhibited sta- tistically significant downward trends for the northeastern and southern peripheral sites. Although the eastern region did not exhibit a statistically significant reduction, a downward trend is indicated by the data and analyses. Linear regressions of annual wet depo- sitions of SO42- (as S) show statistically significant downward trends for all the east- ern sites combined. The results show an overall decrease in wet SO/- deposition of approximately 35% over the period 1989 to 1995. No significant trends were indi- cated for wet NO3- depositions. However, the measurements still show a reduction of about 20% for the eastern data com- bined over the 7-year period. Dry and wet depositions were summed to obtain total depositions of sulfur and nitrogen species. A preliminary trends analysis of total deposition data, which does not correct for meteorological varia- tions or model uncertainties, shows statis- tically significant reductions in annual deposition of sulfur over the period 1989 through 1995. The downward trends are considered significant throughout the east- ern network. Figure 3 depicts the linear regression analysis of total sulfur deposi- tion versus year. The figure also shows nationwide annual SO2 emissions. No trend is apparent in the western data. Estimates of total deposition of nitrogen species ex- hibit no trend. The CASTNet O3 data provide estimates of exposure statistics and allow gauging compliance with the existing and proposed NAAQS for O3. After 1988, violations of the 1-hour standard were limited to subur- 14 12 10 8 o a. to Q Total S all eastern sites ban sites in the Washington-New York corridor. Concentrations above the pro- posed 8-hour standard of 85 ppb were measured throughout the midwest and northeast subregions. The measure SUM06 has been suggested as a second- ary standard for O3. From 1989 through 1995, many CASTNet sites show SUM06 values above 25 ppm-hr, the proposed numerical limit. Measurements from the CASTNet visibility network for 1994 show a strong relationship between fine particle mass (FPM) and fine SO/- concentrations. Fine SO/- contributes more than 85% of the mass of total particulate SO/-. The rural CASTNet measurements show com- pliance with the proposed 24-hour and annual FPM standards of 65 u,g/m3 24- hour average and 15 |ig/m3 annual aver- age. Furthermore, the CASTNet data show a downward trend in ambient SO/-, which is a major contributor to FPM in the east. MADPro is a component of CASTNet designed to study the deposition of air pollutants to high elevation forests. MADPro objectives are to measure cloud chemistry, determine total deposition and define source regions which impact high elevation ecosystems in the eastern U.S. The results to date show that cloudwater can be the primary pathway for deposition of air pollutants. A mobile system to measure direct dry deposition fluxes via eddy correlation has been in operation since 1994. The system has been deployed at several sites with 23 p OL Slope -0.629 p 0.01 r - squared 0.779 24 o CO oo -2 22 co CO "E 20 m 18 CO CD c o • 1989 1990 1991 1992 1993 Note: Emissions are indicated by a *. 1994 1995 Figure 3. Linear regressions of total sulfur depositions versus year for all eastern sites combined (Source: QST). ------- varying terrain settings and land uses to better understand deposition processes and to improve the MLM and other mod- els. Preliminary results of the CASTNet intercomparison study between annular denuders and filter packs show a high correlation between SO42- measurements using the two different methods, and simi- lar results for SO2. Filter pack HNO3 mea- surements were higher than denuder measurements. Conclusions and Recommendations The CASTNet database presents the opportunity for the first time to investigate trends in total (wet plus dry) deposition of sulfur and nitrogen species and contrast the results with trends in emissions. Total deposition of atmospheric sulfur (Figure 3) averaged over all CASTNet eastern sites has decreased by 32% over the pe- riod 1989 through 1995. Nationwide SO2 emissions have declined by 22% over the period 1985 through 1995. Electric utility SO2 emissions have dropped by 24% over that same period. A dramatic drop in SO2 emissions has been reported from 1994 to 1995. Nationwide SO2 emissions have dropped 13% and utility emissions 17% in one year. NOX emissions have been rela- tively flat since 1970. However, an 8% reduction in overall NOX emissions and 21% reduction in electric utility NOX emis- sions was reported between 1994 and 1995. Despite these reported recent NOX emission reductions, the CASTNet data show no change in total deposition of ni- trogen over the period 1989 through 1995. Advanced statistical analyses, beyond the scope of this report, are performed independently on the CASTNet data to further elucidate the apparent trends, and to decipher trends not detectable by simple linear regressions. Statistical analyses should investigate trends at individual sites as well as subregional averages to better understand the response to changes in emissions and meteorological fluctuations. The full report presents quantitative in- formation on dry deposition fluxes and atmospheric concentrations for the net- work from 1987 through 1995. A descrip- tion of the network's quality assurance/ quality control (QA/QC) program and key indicators of data quality such as preci- sion and accuracy statistics are presented as well. The results of the QA/QC pro- gram demonstrate conclusively that the observed changes in concentrations and depositions are real and not the result of network modifications or of data impreci- sion or inaccuracy. Concentration and deposition changes are the result of changes in emissions and of meteorological fluctuations. In short, the CASTNet data constitute an excep- tional database for the purpose of dis- cerning status and trends in air quality and of supporting other scientific activi- ties. To preserve the integrity of and to continue building upon the existing data- base, CASTNet should be operated with minimum disruptions. The number of sites and locations should be reviewed to opti- mize site locations for improvement in the detection of trends. Evaluation of the MLM should continue in order to reduce the uncertainties of the accuracy and preci- sion of deposition estimates. Model ac- ceptance criteria should be fine-tuned. Therefore, the mobile dry deposition field studies should continue until the MLM is fully evaluated and flux calculations are representative of a wide variety of land use and terrain settings. The visibility network should be contin- ued and perhaps expanded to better de- tect trends in visibility related air quality parameters and fine particle concentra- tions. The MADPro sites should also be continued so that investigation of the im- pact of cloud deposition to sensitive eco- systems may continue. CASTNet has produced an exceptional database that satisfies many of the re- quirements of the CAAA of 1990. The network will help assess compliance with the proposed NAAQS for O3 and fine par- ticles. It will continue to measure improve- ments in air quality and depositions associated with CAAA-mandated reduc- tions in SOX, NOX> and VOC emissions over the next 10 years. Ralph Baumgardner is the EPA Project Officer (see below). The complete report, entitled "Clean Air Status and Trends Network Deposition Summary Report (1987-1995)," (Order No. PB98-150097; Cost: $71.50, subject to change) will be available only from: National Technical Information Service 5285 Port Royal Road Springfield, VA 22161 Telephone: 703-605-6000 The EPA Project Officer can be contacted at: National Exposure Research Laboratory U.S. Environmental Protection Agency Research Triangle Park, NC27711 ------- ------- ------- United States Environmental Protection Agency Center for Environmental Research Information Cincinnati, OH 45268 Official Business Penally for Private Use $300 EPA/600/SR-98/027 BULK RATE POSTAGE & FEES PAID EPA PERMIT No. G-35 ------- |