Introduction

Summit, Greenland (72˚34’N, 37˚38’W; 3420ma.s.l.), is a widely studied location for ice-core data (e.g. Reference Meeker, Mayewski, Twickler, Whitlow and MeeseMeeker and others, 1997) and air/snow transfer processes (Reference Dibb and JaffrezoDibb and Jaffrezo, 1997). the sources of airborne material reaching the Greenland ice sheet have also been studied (Reference Davidson, Jaffrezo, Small, Summers, Olson and Borys.Davidson and others, 1993; Reference Kahl, Martinez, Kuhns, Davidson, Jaffrezo and HarrisKahl and others, 1997, Reference Kahl, Martinez, Kuhns, Davidson, Jaffrezo and Harris1999; Reference Slater, Dibb, Keim and KahlSlater and others, 2001). the general conclusion of studies like these is that long-range transport is an important determinant of the content of the air and surface snow at the receptor, and thus is a significant air/snow transfer process.

In this work we also study long-range air movement to Summit, but here we emphasize the vertical component of transport. As air parcels travel long distances, they vary in height as they respond to atmospheric motions on a variety of scales. By using a three-dimensional analysis of air-parcel trajectories, we can determine the altitude at which the trajectories pass over particular upwind areas on the way to Summit. If the trajectories arriving at Summit follow a path that passes near the surface at certain upwind locations, these locations will have a higher probability of contributing to the composition of the air at Summit. By tracking the number of hours traveling air parcels spend over a location at a particular altitude, we determine which areas are frequent passageways for the air trajectories before arriving at the receptor at Summit.

We recognize that the determination of source regions contributing material deposited to the snow surface at Summit requires the consideration of several atmospheric processes. These include background concentrations of chemical species of interest (i.e. atmospheric loading beyond the reach of the back trajectories), atmospheric injection, transformation and removal processes, as well as transport and diffusion over continental and hemispheric-scale distances. As discussedbyReference Kahl, Martinez, Kuhns, Davidson, Jaffrezo and HarrisKahl and others (1997), the uncertainties introduced by considering atmospheric transport alone (as revealed by modeled trajectories) are often constrained during long-range airflow to the Greenland ice sheet. We therefore consider our analysis to represent the potential for upwind source contributions to the air and snow chemistry at Summit.

In a few previous studies addressing atmospheric transport to Summit, a two-dimensional approach has been used. Reference Kahl, Martinez, Kuhns, Davidson, Jaffrezo and HarrisKahl and others (1997) found that for the two pressure altitudes examined (700 and 500 hPa, or approximately 2.8 and 5.5 km m.s.l., respectively) the summertime transport is shorter than the transport in other seasons due to the weaker atmospheric circulation. the 700 hPa analysis revealed that the North American continent is the most important potential source region for continental impurities, while at 500 hPa, transport occurs from North America, as well as regions further west, including the North Pacific Ocean and Asia. In a later study, Reference Kahl, Galbraith and Martinez.Kahl and others (1999) applied a two-dimensional residence-time approach, a precursor to the methodology used here, to diagnose the interannual variability of large-scale transport features.

Three-dimensional trajectories have also been utilized for studying the transport to Greenland. Using three-dimensional trajectory analysis to depict transport to Dye 3, along the southeast coast of Greenland, Reference Davidson, Jaffrezo, Small, Summers, Olson and Borys.Davidson and others (1993) found that there is less long-range transport in the summer, due to weaker wind speeds. These authors emphasized the need for longer trajectories in order to identify potential contributions from more distant source regions. Reference Dibb, Talbot, Whitlow, Shipham and Bales.Dibb and others (1996) used three-dimensional trajectories to investigate chemical signatures associated with the transport of biomass-burning emissions from North America. More recently, Reference Slater, Dibb, Keim and KahlSlater and others’ (2001) analysis of three-dimensional back trajectories again revealed the importance of transport from North America. Although Slater and others’ work considered three-dimensional transport to Summit, the vertical component of transport was not considered in a comprehensive or systematic manner.

The present work expands upon previous studies by emphasizing the vertical component of transport through three-dimensional residence-time analysis of the paths of 10 day air trajectories to Summit.

Data and Methodology

Results and Discussion

Examination of the 10 year average total-column residence time, which does not involve any type of vertical dependency, reveals that the residence time is nearly equally distributed in the area surrounding the receptor (Fig.1a and b). This is due in part to a natural geometric tendency (Reference Poirot and WishinskiPoirot andWishinski, 1986) requiring all trajectories to pass over nearby gridcells en route to Summit. However, there is a slight extension toward the west that would suggest that the main summer pathway to the receptor at Summit is to the west of Greenland, in north-eastern North America. Because of this distribution, any non-marine material contributions that would affect the air composition at Summit are likely to be from North America. This finding is consistent with previous studies (Reference Kahl, Martinez, Kuhns, Davidson, Jaffrezo and HarrisKahl and other, 1997; Reference Slater, Dibb, Keim and KahlSlater and others, 2001).

Fig. 1 Ten-year averaged, summertime residence-time values for 10 day back trajectories arriving at Summit, Greenland: (a) total-column residence time (hours), (b) lowest-bin residence time (hours), and (c) mean pressure (hPa). the residence time shows the amount of time trajectories spend over an area before their arrival at Summit; total column includes trajectories at all heights, while lowest bin includes the lowest 1500m. Pressure values of the trajectories en route to Summit are analogous to altitude, where high-pressure values indicate low altitudes (black indicates areas devoid of trajectory passage).

The mean summer residence times are 45 hours in most of Canada, and 450 hours in extreme northeastern Canada. the extreme western portion of North America is not frequented by summertime trajectories en route to Greenland. the east coast of the United States is also noted as a potential pathway, since residence times in this region exceed 5 hours.

There is very little influence from Europe, although in northern Europe, spanning from the western edge of Russia to the United Kingdom, there is an area where the 10 year average residence times are 5–25hours. In a northern section of Scandinavia, there is a small area with average residence-time values of 25–50 hours. There is little influence from the interior of Asia, or from locations south of 40˚N latitude.

We now turn to the vertical component of transport by presenting the mean pressure of the trajectories passing over each gridcell, as well as the residence time of the lowest 1500m (``lowest bin’’).

The mean pressure (Fig. 1c) shows the average vertical position of the trajectories when they pass over a given location; large values imply trajectory passage at lower altitudes.The highest values are in the lower latitudes, south of Greenland, with mean pressures of 1000–800hPa (approximately 99 and 1930mm.s.l., respectively). In the areas closer to Summit, the trajectories pass through the area between 800 and 600 hPa (1930 and 4229mm.s.l.), which is consistent with the high altitude of the receptor. the relatively few trajectories that pass through Europe (see Fig. 1a) tend to do so at low altitudes (1000–800hPa).

The residence time of the lowest vertical bin (0–1500m) is crucial in determining the areas where surface-level sources could contribute material to the air that arrives at Summit (Fig. 1b). the lowest-bin results show that the predominant areas where trajectories pass near the surface during travel to Summit are to the west of Greenland. There is a maximum of 25 hours in northeastern Canada, areas known to contribute organic material related to forest fires (Reference Dibb, Talbot, Whitlow, Shipham and Bales.Dibb and others, 1996). There are also areas of 45 hours in northern Europe. Because the total residence time for the areas in North America is in the range 5–50 hours (Fig.1a), it can be concluded that most trajectories that pass over the eastern region of North America do so at a low level.

In order to examine deviations from the 10 year average, we now examine the individual anomalous summers from 1989 through 1998. the summer of 1990 was a particularly anomalous period with respect to transport. Note that the total-column residence time for this period (Fig. 2a), as for the other individual summers, is rather noisy as compared to the 10 year mean (Fig. 1a). the 1990 total-column residence time shows transport extending further in all directions, reflecting a particularly vigorous atmospheric circulation. There appears to be a path extending westward from north-western Russia, with total-column residence times exceeding 100 hours. Examination of individual trajectories for a 2 day period in June 1990 (Fig. 3) reveals a pattern of continuous transport from this area. Synoptic meteorological charts for this period (not shown) indicate an anticyclone in the Barents Sea as the cause of this easterly flow to Summit.

Fig. 2 Residence time and pressure for trajectories of summer 1990, as in Figure 1: (a) total-column residence time (hours), (b) lowest-bin residence time (hours), and (c) mean pressure (hPa).

Fig. 3 Trajectory plot for 20–22 June 1990. This plot shows a 3 day span of 10 day isentropic trajectories to Summit that originate primarily from northern Europe.

There are also values as high as 50 hours in central Europe, which is appreciably south and east of where similar values are located in the average plot. Significant lowest-bin residence-time values are also found in this area (Fig.2b). the mean pressure further elucidates the significant low-level transport from central Europe for the summer of 1990 (Fig. 2c). the pressures at which trajectories passed through this area are 1000–800hPa (99 and 1930mm.s.l.), with a significantly large area of 1000–900hPa (99–980mm.s.l.) pressure spanning much of Europe.

There is also a small maximum in lowest-bin residence time over the Great Lakes region of the United States. Inspection of the plots of lowest-bin residence time, as well as the mean pressure, reveals that these trajectories followed a pathway to Summit at low altitudes.

The relatively high number of residence-time hours in northeastern Canada (Fig. 2a) appear to be mostly mid-level trajectories (approximately 700–600hPa, or 2954 and 4229mm.s.l., respectively; Fig. 2c). Residence time of the lowest bin of >50 hours is also shown near the Alaska/ Canada border. Thus, the summer of 1990 exhibited an extremely varied flow, with longer than usual trajectories following several different pathways.

The summer of 1998 was another unusual year, in terms of atmospheric transport to Summit. There is a maximum of total-column residence-time hours centered over the eastern United States, with values exceeding 50 hours in a larger area (Fig. 4a). There is also an area exceeding 50 hours in south-eastern Canada. the lowest-bin results (Fig. 4b) demonstrate that these areas are dominated by low-altitude trajectories. the mean pressure plot from this season verifies this claim (Fig. 4c), with pressures of 1000–900hPa (99–980mm.s.l.), and 900–800hPa (980–1930mm.s.l.) for areas in Canada and the eastern United States, respectively. These plots also reveal anomalous, low-altitude transport from central and southern Europe.

Fig. 4 Residence time and pressure for trajectories of summer, 1998, as in Figure 1: (a) total-column residence time (hours), (b) lowest-bin residence time (hours), and (c) mean pressure (hPa).

The summer of 1996 was anomalous in a different way, with <5 hours total-column residence time for all of Europe (Fig.5a). Nearly all of the transport came from the west, with a reduced southern component. Nearly all of the residence-time hours are concentrated in Canada, with many of them being low-altitude trajectories (Fig. 5b). There are areas dispersed throughout Canada where the lowest-bin residence times exceed 50 hours, with mean pressures of 1000– 800hPa (99–1930mm.s.l.; Fig. 5c). for 1996, Reference Slater, Dibb, Keim and KahlSlater and others (2001) found that >20% of 10 day trajectory origins were located in theYukon and Siberia.

Fig. 5 Residence time and pressure for trajectories of summer 1996, as in Figure 1: (a) total-column residence time (hours), (b) lowest-bin residence time (hours), and (c) mean pressure (hPa).

Of the 10 summers, the origins of the 10 day transport to Summit were furthest afield in 1996. the trajectories extended as far westward as eastern Russia. Furthermore, the residence-time analysis areas show that these were low-altitude trajectories. This pattern is indicative of a summer with a much more vigorous circulation. There is also a highly anomalous feature of five residence-time hours, in the lowest bin, over a large area within central Siberia. This is likely a result of rapid transport over the North Pole.

The remainder of the individual summer seasons, excluding 1990,1996 and 1998, exhibited transport patterns quite similar to the 10 year average (Fig.1).

Conclusions

The three-dimensional residence-time approach is a particularly valuable tool for diagnosing details of both the horizontal and vertical components of long-range atmospheric transport. Knowledge of the vertical component of transport, rather than just the horizontal projection of long-range air trajectories, allows an assessment of the potential for particular upwind source regions to contribute material, whether natural or anthropogenic, to a remote location. In addition to revealing features of the vertical component of transport, our analysis has confirmed many previous conclusions drawn by studies utilizing two-dimensional trajectories.

Application of the three-dimensional residence time technique to summertime, 10 day atmospheric transport to Summit, Greenland, has revealed the following:

North America has the greatest potential to contribute non-marine material to be transported to Summit.

Most transport from North America tends to occur in the lowest 1500m of the atmosphere, thus enhancing the potential for both natural and anthropogenic source contributions.

Considerable interannual variability is found for transport from Europe, with some years exhibiting transport at low altitudes.

The summertime Northern Hemispheric circulation was particularly vigorous in 1990, with potential source contributions further afield in all directions. Significant low-altitude transport from central and northern Europe, as well as from the Great Lakes region of North America, was observed during this season.

Unusually persistent, low-altitude transport from eastern North America was observed during summer 1998.

No transport from Europe was observed during summer 1996. Very strong low-altitude winds during this season created trajectories extending as far upwind as eastern Siberia, as well as over-the-pole transport from central Siberia.

Future work will include comparing these transport results to corresponding chemical measurements in order to further assess the role of meteorology in the air/snow transfer process. the methodology described here is also being applied to atmospheric transport issues at other locations.