Long-range transport of anthropogenic emissions appears to be transforming the composition of much of the remote troposphere. For example, North America exports a large amount of ozone to the North Atlantic [Parrish et al., 1993; Chin et al., 1994]. However, this flux is not the total impact on the remote troposphere, since ozone formation continues in airflows with levels of NO_X (=NO + NO₂) above about 80 pptv [Moxim et al., 1996]. Determining the influence of long-range transport of anthropogenic pollution on the composition of the North Atlantic troposphere, especially with regard to the ozone budget, is one of the primary goals of NARE. In this paper we summarize the results from three ground-based measurement campaigns organized by Michigan Technological University. We also present results from modeling studies of the fate of reactive nitrogen oxides and ozone production within the North Atlantic marine boundary layer (MBL).

We conducted campaigns during the NARE intensives in 1993, 1996, and 1997 to determine reactive nitrogen oxide levels and investigate photo- and physicochemical processes in remote regions of the North Atlantic MBL. The 1993 study was in the summer at the 1 km peak of Santa Barbara volcano (27.322^oW, 38.732^oN) on the Azores island of Terceira. The Azores provide the only location for making extended observations of background trace gas levels and long-range transport events over the central North Atlantic Ocean [see Honrath and Fialho, this issue]. During the winter/spring of 1996, measurements were conducted at the Canadian Coast Guard's Cape Norman lighthouse (55.90^oW, 51.60^oN, 30 m elevation) at the northern tip of Newfoundland. This site maximized exposure to airflows from the Arctic, provided direct exposure to the unmodified marine troposphere, minimized impacts from local emissions, and provided opportunities to sample long-range transport events from the heavily populated regions of North America. The fall 1997 NARE intensive was located at the Canadian Coast Guard's Cape Pine lighthouse (53.31^oW, 46.37^oN, 100 m elevation) at the southern tip of Newfoundland. This location was chosen so that local and regional emissions from Newfoundland would not interfere with analysis of long-range transport events from more southerly portions of North America. The species measured in all three studies included NO and NO_Y (the sum of all reactive nitrogen oxides), O₃, CO and non-methane hydrocarbons. The 1996 and 1997 measurements also included NO_X, nitric acid, nitrate aerosol, peroxyacyl nitrates (PANs) and alkyl nitrates.

Measurements and modeling of reactive nitrogen oxide levels at Terceira in the Azores demonstrate the importance of downward transport of NO_Y-rich air from the free troposphere (FT) to the MBL, and also demonstrate the importance of considering the MBL lifetimes of trace gas species. Observations indicate that NO_Y levels in the overlying FT were 300 to 500 pptv at Terceira during NARE 1993 [Peterson et al., 1998]. When combined with transport of this NO_Y-rich air, relatively simple modeling shows that the lifetime of NO_Y in the MBL must be 0.6 to 1.8 days in order to explain background MBL NO_Y levels of 59 to 93 pptv [Peterson et al., 1998; see also Table 1 in Cooper et al., this issue]. These low MBL NO_Y levels, and lack of CO and O₃ correlations, indicate an absence of long-range transport events in this region during the period of the NO_Y measurements [Peterson et al., 1998]. This is not always the case. Parrish et al. [1998] show that CO and O₃ are correlated at Terceira during short periods in other seasons. However, it is important to note that the simple modeling results show that caution is needed when interpreting correlations between CO and O₃ levels observed in the MBL, as these correlations will be degraded relative to those in the FT due to the different lifetimes of these species in the MBL.

Figure 1. Transformations in NOY speciation during subsidence from the high latitude free troposphere in May.

A chemically explicit model, free of the simplifications required in models that use lifetimes to compute mixing ratios, and not dependent on interpretations of correlations, was also used to study trace gas levels in subsiding air. The cycling of reactive nitrogen oxides and their effect on ozone during long-range transport from the high latitude FT to the North Atlantic MBL was followed with the NCAR Master Mechanism employed as a Lagrangian box model advecting on the path of typical (subsiding) isentropic trajectories [Hamlin and Honrath, 2001]. In these simulations, subsidence-induced warming releases NO_X from air rich in the NO_X-reservoir species peroxyacetyl nitrate (PAN) as it enters the North Atlantic MBL. The situation representing our best estimate of the upper-limit of PAN decomposition is depicted in Figure 1. When generalized to all airflows, this process is responsible for more than 80% of the NO_X found in portions of the North Atlantic MBL during winter [Moxim et al., 1996]. Thus, anthropogenic emissions that form PAN enhance NO_X levels in the remote MBL. Although the NO_X released from PAN is oxidized to nitric acid in a few days (Figure 1) and deposited, elevations in NO_X levels are responsible for increasing the ozone tendency by as much as 1 ppbv/day, relative to the ozone tendency without NO_X release from PAN [Hamlin and Honrath, 2001].

While current models and measurement techniques are becoming increasingly sophisticated and accurate, discrepancies between model results and observations are significant in some cases. One such example involves our NO_X measurements from Cape Norman in northern Newfoundland and an explicit set of global chemical transport model (GCTM) simulations of the influence of long-range transport on NO_X levels over the North Atlantic [Moxim et al., 1996]. The observed median NO_X level of 24 pptv in the MBL of this region is similar to anthropogenically-impacted levels observed in the western Pacific MBL, and is 35%-48% higher than previous observations in the clean midlatitude MBL [Peterson and Honrath, 1999]. Despite this, observed monthly average NO_X levels at this site are about one-half of those predicted for the northwestern North Atlantic MBL by this GCTM [Peterson and Honrath, 1999]. Reactive nitrogen oxide levels observed within the MBL at Terceira in the Azores are also much smaller than predicted by this and another GCTM [Kasibhatla et al., 2000; Lawrence and Crutzen, 1999], both including ship emissions [Corbett and Fischbeck, 1997]. Analysis of NO_X and NO_Y observations made during background MBL airflow periods at Cape Norman in Newfoundland supports this result. Although NO_X/ NO_Y ratios were greater than 0.25 during half of these periods, elevations in NO_X of just 20 pptv above background levels are sufficient to explain this high NO_X/ NO_Y ratio. While we are not certain of the origin of this NO_X enhancement, it is much smaller than expected from ship emission impacts, based on the results of the two GCTMs [Kasibhatla et al., 2000]. Taken together, these results indicate that current GCTMs overestimate reactive nitrogen oxide levels in the North Atlantic MBL. Kasibhatla et al. [2000] suggest that this may be due to, "the lack of parameterized representations of plume dynamics and chemistry in these models". Peterson and Honrath [1999] state that, "the durations of simulated events of elevated NO_X are much longer than are observed". In both cases the authors conclude that simulations of relatively short events with high levels of reactive nitrogen oxides (i.e., long-range transport events, ship emissions) are currently inadequate.

The magnitude of nitrogen oxide export from source regions is a key determinant of global ozone production. Measurements at Cape Pine, southern Newfoundland during fall 1997 indicate significant NO_Y export, but less than calculated by a regional model [Liang et al., 1998]. Nevertheless, the influence of the exported NO_Y upon the O₃ budget is likely greater than the influence of direct O₃ export. To fully assess the ultimate impacts of anthropogenic emissions on ozone levels and production rates over the North Atlantic, additional measurements are needed. One goal of our program at Michigan Technological University is to extend the available data. This is the goal of the PICO-NARE study [Honrath and Fialho, this issue] at a FT site in the Azores.

References