|
Overview of the Program for Research on Oxidants: PHotochemistry, Emissions, and Transport (PROPHET) summer 1998 measurements intensive Mary Anne Carroll Steven B. Bertman Paul B. Shepson 1.
Introduction In this paper we introduce results obtained from the Program for Research on Oxidants: PHotochemistry, Emissions, and Transport (PROPHET) program that is being conducted at the University of Michigan Biological Station in northern Michigan. PROPHET is an independent consortium of individually funded scientists whose mutual interests and varied experiences have created a synergistic collaboration focused on studies of atmospheric chemical and meteorological processes linked to tropospheric ozone. Since 1997, the PROPHET science team has combined expertise to characterize the important atmospheric issues in this region and to begin to push the limits of our knowledge of the links between the biosphere and the atmosphere. The opportunity to conduct research in the physical context of the Biological Station enables this team to interact with a tremendous range of activities related to forest and ecosystem health and uniquely positions PROPHET to make contributions to the emerging field of biosphere-atmosphere interactions. Among other findings in this special section, Faloona et al. report on surprising measurements of the OH radical at night and discuss the possibility that biogenic emissions play a heretofore unappreciated role in nighttime chemistry and that our understanding of radical production in forests may be incomplete. Grossenbacher et al. report the first field measurements of isoprene nitrates and suggest that isoprene oxidation in this forest is slower, on average, than in areas with lower NOx burdens. Westberg et al. report on results of several approaches aimed at quantifying isoprene fluxes in this environment. It is our hope that the experiments described in this special section and future PROPHET experiments are paving the way toward a better understanding of natural forest systems and their interaction with anthropogenic inputs. Nitric oxide (NO) emitted into the atmosphere is rapidly oxidized to NO2. To the extent that the oxidation of NO to NO2 occurs through reaction with species other than ozone (e.g., peroxy radicals), the NO-NO2 cycle results in the catalytic production of ozone. It is believed that increased levels of ozone near the surface [Volz and Kley, 1988; Marenco et al., 1994] are closely tied to increases in and the distribution of anthropogenic NOx emissions [Watson et al., 1990; Levy et al., 1997, 1999; Wang and Jacob, 1998]. Furthermore, it is understood that photochemical ozone production and radical chemistry are strongly dependent on NOx levels in rural and remote environments [Sillman et al., 1990; National Research Council (NRC), 1991; Jacob et al., 1993; Levy et al., 1999; Lelieveld and Dentener, 2000]. Quantification of the physical and chemical processes that affect the availability of NOx in the atmosphere is thus critical to understanding the production of elevated levels of ozone and the photooxidant chemistry that determines oxidizing capacity and thus atmospheric residence times. This need for an improved understanding of the chemistry and climatology of oxidant formation led to the development of PROPHET. With an initial focus on the photochemical and transport processes responsible for the formation of ozone at a rural site in the upper Midwest, PROPHET seeks to quantify the impact of midwestern industrial and biogenic emissions on oxidant production and distribution and to improve our understanding of the chemistry, partitioning, and fate of atmospheric nitrogen in this temperate forest region. Because of the directional dependence of NO emission sources and the variability of the flow regimes between anthopogenic emission impacted and clean northwesterly flow, this is an ideal site at which to investigate both the NOx dependence of oxidant chemistry and the impact of biogenic volatile organic compounds (VOCs). Thus far, PROPHET activities have consisted of facility development, a small suite of continuous measurements, seasonal measurement intensives during which a greater number of species ambient concentration and/or flux measurements are made, transport and chemical modeling studies, and the training of undergraduates, graduate students, and postdoctoral fellows. In this special section we focus on the measurement and modeling activities associated with the summer 1998 measurements intensive and the seasonal behavior of PAN, PPN, and O3. In this paper we provide a detailed description of the research site and highlight key results while introducing the manuscripts that make up this special section. The University of Michigan owns approximately 10,000 acres of forested land in the Onaway Subdistrict of the Presque Isle District near the tip of Michigan's lower peninsula (see Plate 1). The University of Michigan Biological Station (UMBS) is located along the boundary of Cheboygan and Emmet Counties (45°30’N, 84°42’W, elevation 238 m). The northern portion of the lower peninsula is characterized as “mixed” or “transition" forest, with northern hardwood (maple, beech, birch, basswood), mixed aspen (native, bigtooth, trembling, representing new or re-growth), bog conifers in lower, wet areas (white and black spruce and balsam fir), pine and/or red oak in the drier upland regions, and grass-covered sand dunes along the coast. This transition forest region lies south of boreal forests, with its northern border abutting primarily temperate needle leaf forests and woodlands and north of the eastern deciduous forests. The height of the canopy varies, with an average height near 20 m. The overstory age of the hardwood forest is approximately 75 years with overstory biomass equal to 263 Mg/ha [e.g., Pregitzer et al., 1992]. This region and that to the west, including Wisconsin and Minnesota, is largely characterized by forests and woodlands, lakes, small homesteads, villages or small towns, and farms. Towns near the UMBS include the town of Cheboygan, on the northern shore of Lake Huron, ~ 24 km to the east-northeast (population ~ 5500), Petosky ~ 40 km to the southwest on Lake Michigan (population ~ 7800), Traverse City ~ 180 km to the southwest (the shoreline in this entire region is populated with summer homes with little inland; population ~ 15,000), Alpena ~ 120 km to the southeast on Lake Huron (population ~ 11,500, where there is a cement plant with waste combustion), and Gaylord and Grayling ~ 80 and 120 km to the south (population ~ 4000 and ~ 2000, respectively, where plywood factories are located). Larger urban centers include Detroit (~ 350 km to the southeast, metropolitan area population ~ 3,800,000), Chicago (over 400 km to the southwest, metropolitan area population ~ 7,000,000), Toronto, Ontario (over 400 km to the east-southeast, population ~ 4,700,000), and Sault St. Marie, Ontario (~ 130 km to the north, population ~ 81,000). Still farther are Toledo, Ohio (population ~ 610,000), and Madison and Milwaukee, Wisconsin (population ~430,000 and ~1,500,000, respectively). With the predominant flow regimes northwesterly, westerly, or southwesterly with occasional northerly flow [Moody and Sampson, 1989], potential sources of fresh pollution that lie to the east, south, and southeast of the UMBS are less frequently upwind. The PROPHET research site is located in a forested area located on University of Michigan property ~ 3.5 km west of the UMBS (Plate 2). The nearest town is Pellston, Michigan (population < 600), located ~5.5 km to the west. Rural Highway 31 runs north-south and intersects the town of Pellston. The Pellston Airport is located ~1 km north of Pellston and is open year-round, although 50% of its traffic occurs between mid-June and mid-September, representative of the seasonal nature of the traffic in this area in general. During summer there are approximately 12 flights into or out of this airport daily. There is a local road ~2 km south of the site and Interstate Highway 75 runs north-south ~7.5 km to the east. PROPHET has developed significant infrastructure for atmospheric chemistry research with support from the National Science Foundation, the University of Michigan, Western Michigan University, and Purdue University. Facilities include a 31 m measurement tower and a laboratory. The laboratory, an insulated 6.1 m by 9.1 m pole barn, contains a 15 ton water-cooled air conditioning system and high-quality and abundant power distributed in a manner that allows interference-free operation of numerous measurement systems. The tower is a scaffolding tower with an internal stair system, equipped with a fall protection system. A small triangular tower, attached to the main tower, supports a 5 cm ID Pyrex sampling manifold, which brings air from the top of the tower (34 m) to near ground level and into the laboratory. A large blower controls the flow through the manifold at ~3300 liters per minute (LPM) resulting in a residence time of < 2 s. Sample ports are located along the length of the manifold that runs horizontally inside the laboratory ceiling as well as on the lowest vertical section just upstream of a glass elbow, prior to its entering the laboratory. A Teflon screen was added to the manifold inlet prior to the summer 1998 intensive to reduce the intake of insects. Instruments and/or inlet systems can be mounted on the tower itself (e.g., for species such as OH, HO2, HNO3, and NOy, which are too reactive to be sampled with integrity from the manifold) or are located inside the laboratory. The laboratory contains a separate room where vacuum pumps and the sampling manifold blower and laboratory exhaust system components are located. Calibration gases, pump exhaust, manifold exhaust, and heat are transported underground several hundred meters and exhausted near the site access road, which is located to the east (typically downwind) of the measurements site. Additional UMBS research facilities include a DOE AmeriFlux Network Site, a USDA Surface UVB Network Site, and a National Acid Deposition Program/National Trends Network Site. The AmeriFlux site, located 132 m north-northeast of the PROPHET Tower, includes a 50 m tower from which PROPHET flux measurements are made. Additional radiation and particle measurements were made in summer 1998 at the UV-B monitoring site, located ~ 4.5 km east of the PROPHET tower. The region around the PROPHET and AmeriFlux towers is heavily forested with minimal local sources of pollution. A biomass survey conducted in 1998, which included all trees within a 60 m radius of the AmeriFlux tower and leaf trap analyses, yielded an average isoprene (2-methyl-1,3-butadiene) emitting biomass of 156 ± 11 g m-2. Aspen is the dominant species accounting for about 90% with the remaining 10% of the isoprene emitting biomass attributable to Northern Red Oak [Westberg et al., this issue]. With high emissions of isoprene and relatively low but highly variable NOx concentrations, this is an ideal site for studies of isoprene chemistry. The suite of chemical species and atmospheric parameters measured continuously at the PROPHET Tower Laboratory is shown in Table 1. Continuous measurements of O3 and CO began in December 1996, using a Thermo Environmental Instruments, Inc. model 49C ozone analyzer and a Thermo Environmental Instruments, Inc. model 48C CO analyzer. Ambient air was sampled via ~30 m Teflon tubing prior to the construction of the sampling tower and installation of the glass-sampling manifold in June 1997. Standard addition calibrations of the TECO 48C CO analyzer were performed twice daily during intensives and bimonthly between intensives. The system was fully automated during summer 2000 to allow more frequent calibrations. Measurement capabilities for CO and O3 are shown in Table 2. Continuous measurements of peroxyacetyl nitrate (PAN) and peroxypropionyl nitrate (PPN) began in June 1997. These measurements are made using a custom gas chromatograph equipped with an electron capture detector. Measurement frequency varies from every 20 min to 2 hours, with more frequent measurements occurring during intensives. Measurement capabilities for PAN and PPN are shown in Table 2, and a complete discription of the operation of this instrument is given by Pippin et al. [this issue]. Continuous measurements of net radiation, temperature, relative humidity, wind speed, and wind direction were added in July 1997, following the tower-top installation of an Eppley Laboratory, Inc. TUVR ultraviolet radiometer (295-385 nm), a Rotronics Instrument Corp. MP100 temperature and relative humidity probe, and a R. M. Young Company wind monitor-RE. Measurement capabilities for radiation and meteorological parameters are shown in Table 3. 4. Seasonal Measurement Intensives In addition to these continuous measurements, six intensive measurement campaigns have been carried out as part of PROPHET: early and late summer 1997, fall 1997, winter 1998, summer 1998, and summer 2000. The papers in this special section focus on results obtained during the summer 1998 intensive and on the seasonal variability observed in PAN, PPN, and O3 over the period June 1997 through August 1999.The summer 1998 measurements intensive was carried out at the UMBS over the period July 11 through August 20. Measurements of NO, NO2, NOy, PAN, PPN, C3-C5 alkyl nitrates, isoprene nitrates, HCHO, OH, HO2, H2O2, total organic peroxides, isoprene, methyl vinyl ketone methacrolein, 3-methyl furan, C2 – C10 alkanes, alkenes, and alkynes, benzene, toluene, xylenes, SO2, particle size and number, and isoprene, CO2, and H2O fluxes were made throughout the intensive. In addition, measurements of HNO3, HONO, particle phase nitrate and sulfate, and an expanded suite of VOCs and OVOCs were made over the period August 5-15. Research group affiliations, measurement and modeling methods, measurement frequencies, limits of detection, and uncertainties are shown in Table 4. 6. Meteorological Conditions and Flow Regimes During Summer 1998 A discussion of the synoptic meteorology and air mass transport during the summer 1998 intensive is found in the work of Cooper et al. [this issue]. Although the predominant flow regimes in the north regions of the Michigan lower peninsula during summer are northwesterly and southwesterly [Moody and Samson, 1989], the regional flow was changed significantly by La Niña in 1998. Amplification of a ridge over western Canada and the Hudson Bay low resulted in prolonged periods of dry northerly, surface-level flow over the Great Lakes region, reducing the frequency and intensity of precipitation events [Cooper et al., this issue]. Air mass origin over the PROPHET site oscillated between relatively clean regions in the north and regions of greater anthropogenic emissions to the south, and higher mixing ratios of ozone, CO, NOx, PAN, PPN, alkyl nitrates, and HCHO were generally associated with southerly transport [Cooper et al., this issue; Thornberry et al., this issue; Pippin et al., this issue; Ostling et al., this issue; Sumner et al., this issue]. However, the unusually short duration of southwesterly flow episodes prevented the occurrence of significant stagnation, significantly reducing in situ ozone production [Cooper et al., this issue; Pippin et al., this issue]. For example, as shown in Plate 3, significantly higher values of O3 were observed during the early summer intensive in 1997 (A), and NOx mixing ratios were also observed to be considerably higher during extended periods of southwesterly flow [Emmons et al., 1998; Thornberry et al., 1998]. Back trajectory analyses indicate that four types of continental air masses influenced the PROPHET site during the study period: northwesterly, southwesterly, southeasterly, and center of high. Northerly air masses originated over Canada and influenced the UMBS in the wake of passing cold fronts. Southwesterly air masses approached the UMBS on the western side of anticyclones, east of approaching cold fronts. Southeasterly flow occurred during one extended period when a stationary surface low was positioned to the SW of Michigan. Finally, center of high air masses occurred when the site was under the center of a surface anticyclone, characterized by light variable winds and indeterminate origin, associated with the transition between northerly and southwesterly flow [Cooper et al., this issue]. As shown in Plate 4, temperature and water vapor and ozone mixing ratios are lower during northwesterly-northerly flow, radiation measurements indicate that clear-sky conditions occurred more frequently during northwesterly-northerly flow, and ozone levels were highest during southeasterly-southerly flow.These transport category definitions were employed in comparisons of the behavior of additional measured species in polluted southwesterly and southeasterly flow versus that observed during periods of cleaner flow from the north and northwest. Cooper et al. [this issue] examined the influence of air mass transport on the slope of the O3/CO relationship. They find, for this particular site, that the slope of 0.3 is primarily driven by the mixture of polluted (high CO and high O3) air masses from the south and clean air masses (low CO and low O3) from the north. Thornberry et al. [this issue] report that NOy mixing ratios ranged from 180 parts per trillion by volume (pptv) to more than 12 parts per billion by volume (ppbv) and exhibited a strong dependence on the direction of synoptic flow. They report similar a behavior for NOx, with levels ranging from 50 pptv to 9.6 ppbv and with mixing ratios typically between 400 and 700 pptv in southerly flow and 100 to 400 pptv in northerly flow. PAN and PPN mixing ratios ranged from 18 to 943 pptv and from below the 3 pptv detection limit to a high of 109 pptv, respectively, and both species were strongly influenced by the direction of transport [Thornberry et al., this issue]. The sum of C3-C5 alkyl nitrates ranged from 3.45 to 65.8 pptv and contributed between 0.5% and 3% to total NOy. The levels of alkyl nitrates and their relative contribution to NOy were also observed to vary significantly with air mass origin [Ostling et al., this issue; Thornberry et al., this issue]. Mixing ratios of total isoprene nitrates ranged from below the instrument detection limit of 0.5 pptv to 31.0 pptv, and total isoprene nitrate contribution to NOy ranged from < 0.1% to 4.7% [Grossenbacher et al., this issue; Thornberry et al., this issue]. Formaldehyde mixing ratios ranged from 0.5 to 12 ppb, with the highest concentrations observed in southeasterly air masses [Sumner et al., this issue]. Interestingly, as discussed by Thornberry et al. [this issue], higher values of NOx/NOy appear to be associated with northwesterly flow. This behavior is shown in Plate 5, where a polar plot of ozone versus local wind direction is color coded by NOx/NOy. Wind direction measurements made at the PROPHET tower were also used in analyses of chemical species behavior. The Pippin et al. [this issue] treatment of seasonal PAN, PPN, and O3 data used flow categories defined by the local wind direction. For comparisons of species behavior in northerly and southerly flow, they define the “south” category when local winds are within 105°-255° and the “north” category when local winds are within 285°-75°. Pippin et al. conducted a comparison of the average of the summer 1998 PAN, PPN, and O3 data determined by categorization by trajectory with data categorized by local wind direction. They found that the average and median in southerly flow defined by local wind direction are ~30% less and the average and median in northerly flow is 20% and 10% higher, respectively. The polar plot in Plate 6 shows the distribution of observed local winds during periods of northwesterly, southwesterly, and southeasterly flow as defined using trajectory analyses. It is clear that some degree of overlap exists for all flow regimes so defined. Pippin et al. [this issue] examine the seasonal trends of PAN, PPN, and O3 mixing ratios and characterize the primary flow regimes and source regions by contrasting observations in polluted air from the south with cleaner air from the north. Although the synoptic flow pattern was atypical due to El Niño in 1997 and La Niña in 1998, the seasonal trends in PAN, PPN, and O3 observed at the PROPHET site indicate a springtime maximum and a winter minimum for PAN and O3, and a winter maximum with a summer minimum for PPN. Pippin et al. report that the influence of polluted air transported from regions to the south of the site is evident in the diurnal patterns of PAN, PPN, and O3 during summer. Many of the chemical species measured during the summer 1998 intensive varied diurnally at tower height, and this variation was frequently more intense under polluted conditions. For example, while NOy mixing ratios exhibited a diurnal pattern with a systematic increase in NOy mixing ratios after sunrise, peaking between 0800 and 1000 (EDT) before decreasing to a midafternoon minimum, the median increase in NOy during this morning peak was ~ 300 pptv under northerly flow and ~ 1 ppbv under southerly flow. [Thornberry et al., this issue]. NOx values were typically higher during the night and at a minimum in the afternoon in both flow regimes. However, afternoon NOx mixing ratios were typically between 400 and 700 pptv in southerly flow and 100 to 400 pptv in northerly flow. The NOx and NOx/NOy diurnal patterns also displayed a significant peak between 0800 and 1000 during periods of polluted flow [Thornberry et al., this issue]. PAN mixing ratios reached a minimum at night and exhibited a broad afternoon maximum; however this behavior was only clearly seen during northerly flow. In contrast, PAN/NOy was observed to have a significant diurnal variation under both clean northerly and polluted southerly flow [Thornberry et al., this issue]. Isoprene nitrate mixing ratios reach a maximum near 1400 EDT and decrease rapidly in the late afternoon/evening [Grossenbacher et al., this issue]. HCHO mixing ratios also exhibited a midafternoon maximum and a late night minimum [Sumner et al., this issue]. Isoprene exhibits a rapid nighttime decay, falling from several ppbv to levels below 100 pptv. While for most cases the decay was essentially a single first-order exponential, the data appear to exhibit a second decay period occurring 2-3 hours later in ~ 20% of the cases analyzed [Hurst et al., this issue; Faloona et al., this issue]; and while OH and HO2 exhibited daytime maxima as expected, the amplitude modulation factor of the OH diurnal profile was only ~ 4. Furthermore, the nighttime decay of HOx lagged that of radiation, and both species were observed to be sustained in significant amounts throughout the night [Faloona et al., this issue]. As mentioned above, continuous measurements of PAN, PPN, O3, and meteorological parameters are made at the PROPHET site, and Pippin et al. [this issue] examine monthly averages of data obtained from June 1997 to August 1999. They report a minimum in monthly averages of PAN in late autumn/early winter (November-January), with mixing ratios ranging from 100 to 290 pptv and a maximum in February through June, with mixing ratios of 212-536 pptv. In contrast, average PPN mixing ratios were 25% lower in summer than in winter. During this period, monthly average ozone mixing ratios ranged from 32 to 56 ppbv during summer months and 22-32 ppbv during winter months. Pippin et al. also examined correlations among these species. They found a dramatic variation in the PPN/PAN slope from summer when values ranged from 0.09 to .0126, to winter, when slopes ranged from 0.20 to 0.23, with a high degree of association (r2 values of 0.76 - 0.96) during all months. Ozone and PAN were also observed to exhibit a strong positive correlation during summer. However during winter the O3/PAN slope was negative with a fairly weak correlation [Pippin et al., this issue]. Pippin et al. compare these seasonal trends with long-term records at other sites and explore the utility of the difference between winter and summer PPN to PAN correlation as an indicator of summertime biogenic contribution to PAN production. The suite of simultaneous measurements made during the summer 1998 Intensive provided a unique opportunity to study local isoprene chemistry and to evaluate the budgets of HCHO and HOx under relatively low but variable NOx conditions. 9.1 NOx, NOz, and O3 During the summer 1998 intensive, NOx levels at the PROPHET site during the afternoon were typically ~300 pptv in northerly flow and ~600 pptv in southerly flow. Significantly, NOx limited chemistry is expected for such NOx levels and isoprene and formaldehyde mixing ratios in the several ppbv range [Thornberry et al., this issue; Hurst et al., this issue; Sumner et al., this issue]. Regression analyses of O3 - NOz (NOz = NOy - NOx ) data obtained during northerly flow are highly correlated (r2 = 0.783) and yielded a background ozone level of ~ 25 ppbv and an average slope of 22. In contrast, regression analysis of O3 - NOz data obtained during flow from the south are moderately correlated (r2 = 0.554) and yielded a background ozone level of ~ 40 ppbv and an average slope of 14, and regression analyses of southerly flow data filtered for NOx / NOy < 0.25 (r2 = 0.608) results in a background ozone level of ~ 55 ppbv and a slope of 8.7. 9.2 Isoprene Emissions While the biogenic emission inventory system (BEIS) is employed to inventory isoprene emissions for use in regional oxidant models [Geron et al., 1994], many of the basal emission factors suggested for use need further validation. As well, observations are needed to establish basil emission rate change over the course of the growing season [Westberg et al., this issue]. Thus quantification of isoprene fluxes is needed for studies of photooxidant chemistry in this region. Isoprene flux measurements made in the surface layer immediately above the forest canopy provide new isoprene emission estimates for northern hardwood forests and a quantitative estimate of uncertainty [Westberg et al., this issue]. Fluxes were measured from the PROPHET tower in 1997 by relaxed eddy accumulation and from the AmeriFlux tower in 1998 by relaxed eddy accumulation and eddy covariance, which provided a much improved temporal record of isoprene emissions. The towers are only 132 m apart, and total oak and aspen isoprene-emitting biomass density was found to be ~ 150 gm-2 at both towers [Westberg et al., this issue]. In addition, isoprene profiles, obtained via canister sampling from aircraft [Hurst et al., this issue], and a mixed layer gradient modeling technique were used to estimate larger scale isoprene fluxes. Westberg et al. report that the mixed layer modeling approach gave isoprene fluxes that were consistent with those made at the same time at the canopy scale. They recommend that a standard flux value of 11.4 mg m-2 hr-1 and a standard emission rate of 76 mg g-1 hr-1 be used for estimating regional isoprene emissions for this northern hardwood forest regime. 9.3 Ambient Isoprene Measurements and Photochemistry In addition to the above mentioned grab samples, ambient isoprene mixing ratios were determined throughout the summer 1998 intensive using the Purdue gas chromatograph/mass spectrometer (GC/MS) and quadrupole ion trap MS and the Washington State chemiluminescence-based fast isoprene sensor. Isoprene measurements were also made using two additional GC/MS instruments from the National Center for Atmospheric Research (NCAR) and the University of Miami during August 5-15 [Hurst et al., this issue; Barket et al., this issue; Westberg et al., this issue; Apel et al., 2001]. Isoprene data are presented by Hurst et al. [this issue] and Barket et al. [this issue], wherein the results of an informal intercomparison of the data obtained with these five instruments are presented. Daytime isoprene chemistry is discussed by Apel et al. [2001] and Sumner et al. [this issue], Grossenbacher et al. [this issue], and Tan et al. [this issue], and isoprene nighttime decay is discussed by Hurst et al. [this issue], Faloona et al. [this issue], and Sillman et al. [2001]. Barket et al. [this issue] report that regression analysis of the GC/MS methods show excellent agreement, and all five methods demonstrate relatively good agreement in capturing diel profiles, decays, and onsets. However, they found significant relative differences among the five techniques for isoprene mixing ratios below 1 ppbv and identified an apparent bias in the non-GC/MS methods when isoprene mixing ratios exceeded 5 ppbv. Apel et al. [2001] focus on isoprene, methyl vinyl ketone (MVK), and methacrolein (MACR) measurements and comparisons of observations and modeled ratios of the isoprene oxidation products to isoprene. They report that model results indicate that the air masses studied represented relatively fresh emissions with a photochmeical age of measured isoprene between 3 and 18 min, which is significantly less than the expected photochemical lifetime of isoprene (t = 45 min at [OH] = 3.4 x 106 molecules cm-3). Furthermore, Apel et al. find the average daytime (MVK + MACR)/isoprene ratio to be 0.12, which is somewhat lower than previous studies and lower than is predicted by the Sillman et al. [2001] model. They conclude that a large portion of the isoprene that reaches the sampling manifold has not had time to react completely with OH yielding lower than expected ratios when compared with model calculations that do not explicitly take this into account. While isoprene dominates as an OH sink in this forested environment, it can also effectively limit ozone production through sequestration of NOx during the formation of isoprene nitrates, which can significantly impact the conversion of NOx to NOz as well as the long-range transport and deposition of nitrogen. Grossenbacher et al. [this issue] report the results of the first quantitative ambient measurements of organic nitrates derived from isoprene. They found that isoprene nitrate and alkyl nitrate mixing ratios are comparable and report typical contributions of the isoprene nitrates to NOy of 0.5 - 1.5% and up to ~4% in well-aged air, which are lower than expected considering known yields of isoprene nitrates and MACR from the OH initiated reaction of isoprene in the presence of NO. However, Grossenbacher et al. note that reactions involving the isoprene peroxy radicals and other peroxy radicals can lead to a different relative production of MACR and the isoprene nitrates under NOx-limited conditions. 9.4 HCHO and HOx Budgets Sumner et al. [this issue] explore the rates of HCHO production and OH reactivity associated with the different flow regimes experienced during the summer 1998 intensive. They found isoprene oxidation to be the most important source of HCHO and dry deposition to be an important sink during daytime and the dominant sink at night. Calculation of HCHO production and loss terms indicates that meteorological effects are also important. For example, in southerly flow, NOx and OH levels increase and the rates of HCHO production from isoprene oxidation also increase [Sumner et al., this issue]. Determination of species contributions to OH reactivity in southeasterly flow indicates that isoprene, HCHO, CO, and acetaldehyde contribute 75, 10, 4, and 4%, respectively. Measurements of OH, HO2, and a full range of species important to HOx also allow assessment of the HOx budget. Thirty-minute average values of O3, NO, NO2, CO, H2O, iosprene, MACR, MVK, 3-methyl furan, HCHO, acetaldehyde, acetone, a-pinene, b-pinene, several terpenes, toluene, PAN, MPAN, SO2, calculated photolysis frequencies, and aerosol surface area were used in a photochemical point model for comparisons of modeled OH and HO2 with OH and HO2 measurements made during August 5-15 [Tan et al., this issue]. Because measured photolysis frequencies were not available, J values were calculated using data from the PROPHET tower Eppley radiometer and the UV-B network site Yankee UV-B radiometer, along with the TOMS– and GOME-derived ozone columns for the PROPHET site and the Madronich algorithms [Madronich, 1987a, 1987b]. Tan et al. [this issue] found modeled HO2 to be in good agreement with measurements, and modeled OH to be, on average, 2.7 times lower than observations. Even when the model was run with an additional postulated OH source from the ozonolysis of unmeasured terpenes, measured OH is 1.5 times greater than the model. Tan et al. conclude that model mechanisms currently used to model HOx in a high biogenic hydrocarbon/low NOx environment do not correctly capture HO2 to OH cycling and appear to be missing an OH source. As mentioned above, isoprene was observed to undergo a rapid nighttime decay. Calculations of decay rate constants from the average slope of ln[isoprene] versus time yield isoprene lifetimes ranging from 2 to 5 hours [Hurst et al., this issue]. They report that reaction with O3 will not occur at a rate consistent with the observed decay, and with nighttime NO levels of only 2 pptv [Thornberry et al., this issue], the NO3 radical only becomes an important sink for isoprene after the majority of the isoprene decay has occurred. The isoprene flux data were not consistent with dry deposition playing a significant role in nighttime forest loss [Westberg et al., this issue]. Instead, and possibly corroborating observed levels of OH [Faloona et al., this issue], observed decay rates show good overall agreement with the second-order isoprene + OH rate constant, and the isoprene decay can, for several evenings, be simulated using measured OH [Hurst et al., this issue; Faloona et al., this issue]. However, the reported OH data were found to overpredict the isoprene loss rate on most nights, and Hurst et al. estimate that vertical mixing with isoprene-depleted air probably contributes to the observed isoprene decay. Faloona et al. [this issue] report that OH and HO2 are sustained in significant amounts throughout the night in this northern forested region, with typical overnight OH levels of 0.04 pptv (1.1x106 molecules/cm3) and HO2 levels of 1 to 4 pptv. They employ a steady state model of nitrate radical concentrations and a zero-dimensional, steady state model, which iterates the main production and loss reactions in the HOx budget until self-consistent solutions for the OH and HO2 concentrations are obtained, in their investigation of nocturnal chemistry during the summer 1998 intensive. In assessing nighttime isoprene chemistry, Faloona et al. also conclude that reaction with O3 and NO3 are probably not important on timescales of interest. They calculate an isoprene lifetime with respect to the median hydroxyl concentration of 106 molecules cm-3 of only 2.7 hours, in agreement with the Hurst observations. Furthermore, they report that observed aerosol dynamics probably indicate the presence of an ample pool of lower volatility precursor gases at late hours, perhaps providing additional support for the occurrence of prodigious oxidation in the nocturnal boundary layer. Proposing that the elevated OH concentrations observed at night involve an unmeasured, extremely reactive olefinic compound, Faloona et al. examine four model scenarios. These include a baseline run, which only includes the OH production terms measured and expected by conventional atmospheric chemistry; a run with additional terpene ozonolysis, sufficient to bring the HO2 estimates within 5% of the observed levels on average; a run with a greater terpene source, sufficient to match the observed OH levels; and a run with isoprene peroxy radical levels arbitrarily inflated to greater than 90 pptv and an increased uptake coefficient for HO2 onto atmospheric aerosols. However, such theoretical overnight OH production mechanisms are also substantial HO2 sources and, consequently, pull steady-state estimates of HO2 well above measured levels [Faloona et al., this issue]. Sillman et al. [2001] employ a one-dimensional Lagrangian model for atmospheric transport and photochemistry to investigate the observed nighttime decay in isoprene and high nighttime OH. They report that model results compare well with measured isoprene, methacrolein, NOx, and isoprene vertical profiles but exhibit significant discrepancies for terpenes, MVK, HCHO, and the ratio OH/HO2. Results indicate that while significant amounts of OH can be generated at night in the model through terpene chemistry, this nighttime OH is confined to a shallow vertical layer and thus has a limited impact on isoprene. Instead, Sillman et al. [2001] suggest that the observed decrease in isoprene at night can be reproduced in models with low OH, primarily through vertical mixing with isoprene-poor air aloft. 11. Special Section Organization The organization for this special section is shown in
Table 5. A discussion of the meteorological conditions is presented
first, followed by manuscripts focusing on daytime nitrogen and isoprene
chemistry and nighttime isoprene decay and HOx. Acknowledgments. The PROPHET Science Team thanks J. A. Teeri, C. A. Sutterly, L. Readmond, P. J. Cunningham, R. Spray, T. F. Crandell, K. Gasper, and R. J. Vande Kopple for infrastructural and operational support at the University of Michigan Biological Station. We also thank P. Curtis, H. P. Schmid, and C. Vogel for access to and collaborations involving the AmeriFlux tower and S. Thomas for superb program management. The authors gratefully acknowledge the efforts of the other members of the science team, the loan of tower components by the Atmospheric Chemistry Division of the National Center for Atmospheric Chemistry, and funding from the National Science Foundation (MAC, PBS, and SBB), the Environmental Protection Agency (PBS), the University of Michigan, Purdue University, and Western Michigan University. Apel, E. C., et al., Measurement and interpretation of isoprene fluxes and isoprene, methacrolein, and methyl vinyl ketone mixing ratios at the PROPHET site during the 1998 intensive, J. Geophys. Res., in press, 2001. Barket, D., Jr., et al., Intercomparison of automated methodologies for determination of ambient isoprene during the PROPHET 1998 summer campaign, J. Geophys. Res., this issue. Cooper, O. R., J. L. Moody, T. Thornberry, M. Town, and M. A. Carroll, PROPHET 1998 meteorological overview and air mass classification, J. Geophys. Res., this issue. Emmons, L. K., D. Hauglustaine, M. A. Carroll, and G. Brasseur, Data-based climatologies of tropospheric carbon monoxide and ozone, Eos Trans. AGU, 79(17), Spring Meet. Suppl., S21, 1998. Faloona, I., et al., Nighttime observations of anomalously high levels of hydroxyl radicals above a deciduous forest canopy, J. Geophys. Res., this issue. Geron, C. D., A. B. Guenther, and T. E. Pierce, An improved model for estimating emissions of volatile organic compounds for forests in the eastern United States, J. Geophys. Res., 99, 12,773-12,791, 1994. Grossenbacher, J. W., et al., Measurements of isoprene nitrates above a forest canopy, J. Geophys. Res., this issue. Hurst, J. M., et al., Investigation of the nighttime decay of isoprene, J. Geophys. Res., this issue. Jacob, D. J., J. A. Logan, G. M. Gardner, R. M. Yevich, C. M. Spivakovsky, and S. C. Wofsy, Factors regulating ozone over the United States and its export to the global atmosphere, J. Geophys. Res., 98, 14,817-14,826, 1993. Lelieveld, J., and F. J. Dentener, What controls tropospheric ozone?, J. Geophys. Res., 105, 3531-3551, 2000. Levy H., II, P. S. Kasibhatla, W. J. Moxim, A. A. Klonecki, A. I. Hirsch, S. J. Oltmans, and W. L. Chameides, The global impact of human activity on tropospheric ozone, Geophys. Res. Lett., 24, 791-794, 1997. Levy H., II, W. J. Moxim, A. A. Klonecki, and P. S. Kasibhatla, Simulated tropospheric NOx: Its evaluation, global distribution and individual source contributions, J. Geophys. Res., 104, 26,279-26,306, 1999. Madronich, S., Intercomparisojn of NO2 photodissociation and U.V. radiometer measurements, Atmos. Environ., 21(3), 569-578, 1987a. Madronich, S., Photodissociation in the atmosphere, 1, Actinic flux and the effects of ground reflection and clouds, J. Geophys. Res., 92, 9740-9752, 1987b. Marenco, A., H. Gouget, P. Nédélec, J.-P. Pagés, and F. Karcher, Evidence of a long-term increase in tropospheric ozone from Pic du Midi data series: Consequences: Positive radiative forcing, J. Geophys. Res., 99, 16,617-16,632, 1994. Moody, J. L., and P. J. Samson, The influence of atmospheric transport on the composition of precipitation at two sites in the Midwestern United States, Atmos. Environ, 23, 2117-2132, 1989. National Research Council (NRC), Committee on Tropospheric Ozone Formation and Measurement, in Rethinking the Ozone Problem in Urban and Regional Air Pollution, Nat. Acad. Press, Washington, D. C., 1991. Ostling, K., B. Kelly, S. Bird, S. B. Bertman, M. Pippin, T. Thornberry, and M. A. Carroll, Fast-turnaround alkyl nitrate measurements during the PROPHET summer 1998 intensive, J. Geophys. Res., this issue. Pippin, M., S. Bertman, T. Thornberry, M. Town, M. A. Carroll, and S. Sillman, Seasonal variations of PAN, PPN, and O3 at the upper Midwest PROPHET site, J. Geophys. Res., this issue. Pregitzer, K. S., A. J. Burton, G. D. Mroz, H. O. Liechty, and N. W. MacDonald, Foliar sulfur and nitrogen along an 800-km pollution gradient, Can. J. For. Res., 22(11), 1761-1769, 1992. Sillman, S., J. A. Logan, and S. C. Wofsy, The sensitivity of ozone to nitrogen oxides and hydrocarbons in regional ozone episodes, J. Geophys. Res., 95, 1837-1851, 1990. Sillman, S., et al., Loss of isoprene and sources of nighttime OH radicals at a rural site in the U.S.: results from photochemical models, J. Geophys. Res., in press, 2001. Sumner, A. L., et al., A study of formaldehyde chemistry above a forest canopy, J. Geophys. Res., this issue. Tan, D., et al., HOx budgets in a deciduous forest: Results from the PROPHET summer 1998 campaign, J. Geophys. Res., this issue. Thornberry, T., et al., Simultaneous measurements of NO, NO2, and NOy during the PROPHET fall 1997 and winter 1998 seasonal intensives, Eos Trans. AGU, 79(17), Spring Meet. Suppl., S22, 1998. Thornberry, T. D., et al., Observations of reactive oxidized nitrogen and speciation of NOy during the PROPHET 1998 summer intensive, J. Geophys. Res., this issue. Volz, A., and D. Kley, Evaluation of the Montsouris series of ozone measurements made in the nineteenth century, Nature, 332, 240-242, 1988. Wang, Y. H., and D. J. Jacob, Anthropogenic forcing on tropospheric ozone and OH since pre-industrial times, J. Geophys. Res., 103, 31,123-31,135, 1998. Watson, R. T., H. Rodhe, H. Oeschger, and U. Siegenthaler, Greenhouse gases and aerosols, in Climate Change: The IPCC Scientific Assessment, pp. 1-40, Cambridge Univ. Press, New York, 1990. Westberg, H., B. Lamb, R. Hafer, A. Hills, P. Shepson, and C. Vogel, Measurement of isoprene fluxes at the PROPHET site, J. Geophys. Res., this issue.
Plate Captions:Plate 1. Regional map showing the location of Pellston, Michigan (star) near northern tip of the Michigan lower peninsula (source: www.mapquest.com). Plate 2. Map showing location of the University of Michigan Biological Station (cross) and PROPHET Research Site (asterix) with town of Pellston and Pellston Airport to the west and Interstate Highway 75 to the east (source: www.topozone.com). Plate 3. Ozone mixing ratios measured during three PROPHET intensives are shown above the wind rose showing the distribution of 1 hour average wind directions as measured at the Pellston Airport during the summer 1997 A intensive. Plate 4. Temperature, water vapor, radiation, and ozone mixing ratios measured at the PROPHET tower during the summer 1998 intensive are shown, color coded by trajectory-defined flow categories (blue- NW/N, red – SW, green – SE/S). Plate 5. One minute average ozone mixing ratios measured during the summer 1998 intensive are shown in a polar plot, color coded for NOx/NOy. Plate 6. Distribution of observed local winds measured at the PROPEHT tower during periods categorized as NW, SW, and SE flow per back trajectory analyses. |