RADON-222 AS A TRACER OF ATMOSPHERIC TRANSPORT PHENOMENA ON DIFFERENT SPATIAL AND TEMPORAL SCALES

Transcription

1 RADON-222 AS A TRACER OF ATMOSPHERIC TRANSPORT PHENOMENA ON DIFFERENT SPATIAL AND TEMPORAL SCALES Zahorowski W., Chambers S. and Williams A.G. Australian Nuclear Science and Technology Organisation (ANSTO), Menai, New South Wales, Australia 1. Introduction Radon-222 (radon) is the most frequently used naturally occurring radioactive tracer of atmospheric dynamics on local, regional, and hemispheric scales [1]. This is due to a number of useful radon properties, which include: (a) a terrestrial flux that can be assumed to be horizontally homogenous on regional scales and approximated by a global average of 22 mbq m -2 s -1 ; (b) a very low flux from oceans (about three orders of magnitude lower than from land); (c) its non-reactive nature, with a simple, well defined sink by radioactive decay alone; and (d) a half-life (3.8 days) that is long enough compared to the turbulent mixing time in the Atmospheric Boundary Layer (ABL) but short enough to ensure a significant gradient across the ABL-troposphere interface and to minimize horizontal variations in concentration above the ABL. Radon s tracing capacity has been applied to (a) selecting the least perturbed marine air masses at baseline stations [2]; (b) defining footprints of ground-based sites [3]; (c) calibrating, on regional scales, terrestrial emissions of greenhouse gases [4]; (d) evaluating the performance of regional and global atmospheric transport models [1]; and (e) quantifying mixing within the ABL [5]. ANSTO has developed advanced instruments and methods for radon observations on the ground, at sea, and from airborne platforms. Such measurements are demanding as radon concentrations in air may vary from 1 to >1, mbq m -3, with the lower limit corresponding to 5 radon atoms per litre of air, a very low concentration by any standard. The time resolution of such instruments, necessitated by the processes of interest, is also demanding. While it is sufficient for ground-based measurements to resolve diurnal and synoptic time scales for horizontal transport, sampling times shorter than 1-15 min are essential to conduct vertical profiles. Also, as a rule, ground- and sea-based instruments have to run unattended for prolonged periods, and airborne samplers are tightly constrained by weight and size, have to meet strict aviation standards and be simple enough to operate easily in flight. This paper illustrates how radon has been applied as a tracer on regional scales in the East Asia/Northern Pacific region. It also reports on the most recent ANSTO developments in the measurement of vertical radon profiles for the characterisation of mixing within the ABL and exchange with the troposphere, from groundbased and airborne observations. Together with supporting meteorological measurements, these systems provide new information regarding mixing and exchange processes across the atmospheric surface layer, the nocturnal stable layer, and the convective mixed layer, and can be used in the evaluation of mixing schemes for numerical models. 2. Tracing atmospheric transport on regional scales Examples presented here are derived from a five year experiment in which hourly measurements of atmospheric radon concentration were made at three ground-based coastal stations in East Asia (Hok Tsui, Hong Kong Island, China; Gosan, Jeju Island, Korea; Sado, Sado Island, Japan). Results are compared with data from Mauna Loa Observatory (MLO), Hawaii, gathered during the same period. The experiment was part of the most recent Aerosol Characterization Experiment in East Asia (ACE-Asia), and aimed to comprehensively characterise air masses associated with East Asian continental outflow events to the Pacific. Site locations were chosen to span the latitudinal band within which most of the low level Asian continental outflow events to the Pacific occur (2-4 N), and coincided with ACE-Asia network sites and locations where campaign-style observations were made. The data set is unique with respect to its spatial and temporal coverage, and well suited to air mass characterisation of a region that is a globally significant source of natural and anthropogenic pollution.

2 Figure 1a Principle of operation of the dual flow loop, two filter radon detector Figure 1b Schematic view of an air parcel s movement en route to a ground station Dual-flow loop, two-filter detectors, developed at ANSTO [6], were used at each of the sites. The principle of operation of these detectors is illustrated in Figure 1a. The external flow loop draws sample air through the detector at a low flow rate. Between the inlet and the first filter the sample is delayed to reduce radon- 22 (thoron) content to negligible levels. The first filter removes aerosols and ambient radon progeny. The air then enters the radon delay volume where it is circulated at a high flow rate by the internal flow loop, passing repeatedly through the second filter. A fraction of the new radon progeny generated in the radon delay volume is captured on the second filter. These short-lived, alpha emitting radon progeny are counted using an assembly consisting of a ZnS scintillator, photomultiplier tube and counting electronics. Each of the three detectors deployed in East Asia had a volume of 75L and sampled at a flow rate of ~4L min -1 ; the volume and the flow rate of the MLO detector were 1,5L and 8L min -1, respectively. The response time of the detectors, defined as the time required to reach 5% of the maximum count rate after a step change in radon, was about 45 minutes. Figure 2 Areas of high (left panel) and low (right panel) contribution of continental emissions to air masses approaching Hok Tsui. The blue square shows the location of the station and dotted lines approximate the main direction in which the air parcels approach the station. Trajectory selection was based on hourly radon concentrations recorded at Hok Tsui in winter 23. Atmospheric species of terrestrial origin, including aerosols and climatically sensitive gases, are usually monitored at ground stations where air masses are sampled and their composition determined. That composition will depend upon the stations footprint. A station s footprint is the region most likely to contribute to the atmospheric composition of air masses arriving at the station. Back trajectories, derived from assimilated meteorological data collected around the world, approximate paths that air parcels follow on their way to a station (Figure 1b). However, back trajectories alone can not quantify the degree of interaction between an air parcel and the Earth s surface, which largely determines constituent concentrations. Interaction with the Earth s surface can only be estimated by concurrent measurements of a suitable tracer. Because of its unique properties, radon is the most suitable tracer in this context.

3 1 Radon (mbq m-3) Hok Tsui Sado Is. 1 1 Figure 3a Outline of seasonal footprints and mean trajectories in East Asia at HokTsui (red lines) and Gosan (black lines). Solid lines represent winter, dotted lines spring. Shaded areas depict footprints of Hok Tsui and Gosan (red) and MLO (blue). Gosan Week of year 4 5 Figure 3b Seasonal change in background radon concentration at the three sites. Curves represent weekly 1th percentile values o Latitude ( ) Latitude (o) Since many factors can affect species concentrations, the best way to analyse the variability in a stations footprint is by determining footprints corresponding to extremes of land contact. Previously this has been achieved by analyzing back trajectories corresponding to high and low hourly radon concentrations over a period of interest [3]. A separate consideration is the length of back trajectories used. Under conditions of a spatially homogenous radon flux and entrainment velocity, approximately 8% of the radon observed in an air mass is contributed over the last 1 days of its journey. The following examples of back trajectories were calculated using the NOAA Air Resources Laboratory (ARL) HYSPLIT 4 package [7] Longitute (o) Longitute (o) 21 Figure 4 An example of the latitudinal separation of the MLO footprint for Spring 23. The left and right panels show the 2-3 and 3-4 latitudinal bands. The first example of radon-derived footprints from Hok Tsui in winter (Figure 2) shows that the technique can identify subtle changes in fetch. The left and right panels show trajectories corresponding to hourly radon concentrations higher (lower) than the seasonal 9th (1th) percentile values, respectively. Although the corresponding air masses approach Hok Tsui from a similar direction (dotted lines) their composition is very different. A high radon concentration indicates recent land contact whereas a low concentration most likely indicates recent marine fetch. In each case the last 3-5 days of the Hok Tsui footprint is well defined, indicating an overall coastal fetch in winter, but with distinct on-shore and off-shore components. The second example illustrates the capability of the above method to obtain footprints characterised by strong air mass interaction with continental emissions. In Figure 3a schematic footprints are superimposed on the corresponding mean trajectories for winter (solid lines) and spring (dotted lines). The result clearly indicates that, although Hok Tsui and Gosan are located within the 2-4o latitudinal band, their recent fetch

4 is limited coastal regions. Furthermore, their footprints extend to the interior of the continent above 5 o north. As such, they offer limited information about inland continental emissions within the ACE-Asia study domain (2-4 N). As indicated in blue in Figure 3a, however, such information can be derived from the MLO observations. Trajectories of high radon events in spring corresponding to the MLO footprints (Figure 4), demonstrate that MLO high radon events can be separated into those originating from the 2-3 o and 3-4 o latitudinal bands, thus enabling separate probing of these two important continental bands. An interesting aspect of the East Asian network measurements is that their radon background varies strongly with both season and latitude (Figure 3b). Here the background is defined by the weekly 1 th percentile concentration to reduce the influence of synoptic scale events. Only in summer (weeks 25-3), when regional flow is onshore, does the background approach concentrations associated with oceanic radon emissions. In other seasons, when offshore flow dominates, the background reflects seasonal and latitudinal dependence. 3. Tracing local air movement on diurnal time scales Vertical radon profiles in the lower atmosphere can be used to determine the degree of mixing in the ABL and exchanges with the surface and the free atmosphere. The use of radon in ABL process and model evaluation studies therefore carries the potential to significantly reduce systematic errors in the reproduction of diurnal and seasonal cycles in weather and climate prediction models on a range of scales. In the last two years ANSTO has developed and implemented a measurement system for obtaining accurate radon concentration profiles through the ABL and extending into the free troposphere. The adopted approach involves a combination of ground-based continuous hourly observations of radon concentration gradients and aircraft-based measurements of radon profiles day subset Radon (mbq m -3 ) Disconnected layers Day of 25 Figure 4 Hourly radon concentrations recorded at two levels (2m and 5m) on the 5m ANSTO tower. The top panel shows all data for September 25; the bottom panel shows a 1 day subset of that period. The ground-based observations rely on taking air samples using inlet lines mounted at two heights, on two towers: one tower covering the first 5m above the ground; the other extending to 2m. Each of the towers is equipped with two radon detectors of the type discussed in the previous section. All four instruments have identical technical parameters (1,5 L delay volume; 8 L min -1 inlet flow rate; 3-4 mbq m -3 lower limit of detection). At present, inlets are mounted and 2 and 5m, and 2 and 2m, on the 5m and 2m towers, respectively. The shorter tower is located at ANSTO and became operational in mid 25. The taller tower is located at the Cabauw Experimental Site for Atmospheric Research (CESAR) in the Netherlands and became operational in March 26. The two-tower system offers a unique opportunity for high precision measurements of diurnal and seasonal cycles of vertical mixing using naturally occurring radioactivity. Radon time series from 2 and 5m on the ANSTO 5m tower at are compared in Figure 4 for the whole month of September 25 (top panel) and for a 1-day subset (bottom panel). Close to the surface, radon concentrations build up at night with the formation of the nocturnal stable boundary layer (SBL). On some 5mRn 2mRn

5 nights, indicated by arrows, the radon data indicates that the local SBL height is below 5m and the surface is thermodynamically decoupled from the air aloft. Radon (mbq m -3 ) Temperature (C) Wind speed (m s -1 ) Radiation (W m -2 ) m 5m Incoming shortwave Net radiation 5m 2m 5m AGL Day of 25 Figure 5 Hourly radon recorded at 2 and 5m on the 5m ANSTO tower for 4 days in September 25 (top panel) as well as temperature, wind speed and solar radiation. Figure 5 shows details of the 4 days from 4-7 September 25. Mean radon concentrations at 2 and 5m differed only slightly during the 24-hour period containing the night of day 248. In contrast, the corresponding period for day 25 exhibited radon differences up to 4 Bq m -3. The meteorological data reveals the reasons for this contrasting behaviour. On day 248, conditions were overcast and wind speeds moderate (at times >6 ms -1 ). This resulted in near neutral conditions and well mixed air to depths in excess of 5m for the entire period, as can be seen by the small temperature gradients and diurnal range in near-surface air temperature. On day 25, however, conditions were clear and the nocturnal wind speeds lower. As can be seen in the radon time series and the decoupling of temperature aloft from the surface, a strong nocturnal inversion developed with a depth below 5m. The two-tower system is complemented by a sampling system for measurement of radon concentrations from airborne platforms. A radon sampler was developed at ANSTO after completion of successful tests of two prototypes. The new radon sampler consists of 6 charcoal traps (Figure 6a). Each trap has a volume of approximately 3 cm -3, holding 11g of charcoal. During a sampling event a known volume of air passes through the traps using a 24V dual membrane pump. Samples are collected for periods ranging from 5 to 15min at a flow rate of about 65 L min -1 STP. Instrument dimensions, weight and power consumption (76 x 3 x 2 mm, 2kg, and 24W, respectively) make it possible to fit in the wing-pod of a motor-glider (Figure 6a). Radon collected in the traps is transferred to evacuated Lucas cells in a two step process in a laboratory-based radon rig. Estimated counting errors as a function of radon concentration, calculated for three counting periods, are shown in Figure 6b. The estimate points to the possibility of getting the lower limit of detection down to 1 mbq m -3. The actual lower limit of detection depends critically on the concentration of radium-226 in charcoal: a typical commercial charcoal batch results in a peak-to-background ratio of.2 at the 1 mbq m -3 radon concentration level the same ratio for the counting system alone is about 6. Together with the aforementioned two-tower system, the capability of the new system to record vertical radon concentration profiles covers the boundary layer and lower troposphere up to 3m above the ground.

6 3 Counting error (% of net count) hour count 9 hour count 12 hour count Radon concentration (Bq m -3 ) Figure 6 The left panel shows the radon sampler mounted in the wing-pod of Airborne Research Australia s motor-glider VH-EOS. The central section of the sampler contains solenoid valves and an air pump; it provides the mouting points for the six charcoal-filled stainless steel tubes (three on each side). The right panel shows the counting efficiency of the lab-based system for processing the samples for three counting periods. 4. Conclusions A comprehensive system of measurements, including unique instrumentation and methods, has been established at ANSTO to apply radon tracing capacities for the current problems in atmospheric research. On a regional scale, these measurements have provided accurate footprint description for important coastal ground stations in East Asia and the Northern Pacific. On a local scale, the newly developed airborne measurements offer potential for new insights into mixing and exchange processes across the Atmospheric Boundary Layer. 5. References [1] Zahorowski W., Chambers S. and A. Henderson-Sellers, Ground based radon-222 observations and their application to atmospheric studies, Journal of Environmental Radioactivity, Vol. 76, 24, pp.3-33 [2] Zahorowski, W. and S. Whittlestone, Radon database : A review, In Baseline Atmospheric Program (Australia) 1996, 1999, pp.71-8 [3] Zahorowski W., S. Chambers, T. Wang, C.-H. Kang, I. Uno, S. Poon, S.-N. Oh, S. Werczynski, J. Kim and A. Henderson-Sellers, Radon-222 in boundary layer and free tropospheric continental outflow events at three ACE-Asia sites, Tellus, Vol. 57B, 25, pp [4] Biraud S., Ciais P., Ramonet M., Simmonds P., Kazan V., Monfray P., et al.,. European greenhouse gas emissions estimated from continuous atmospheric measurements and radon 222 at Mace Head, Ireland, Journal of Geophysical Research, Vol. 15(D1), 2, pp [5] Lee H.N. and R.K. Larsen, Vertical Diffusion in the Lower Atmosphere Using Aircraft Measurements of 222 Rn, Journal of Applied Meteorology, Vol. 36, 1997, pp [6] Whittlestone S. and W. Zahorowski, Baseline radon detectors for shipboard use: Development and deployment in the First Aerosol Characterisation experiment (ACE 1), Journal of Geophysical Research, Vol. 23, 1998, pp.16, ,751 [7] Draxler R.R and G.D. Hess, An overview of the HYSPLIT 4 modelling system for trajectories, dispersion and deposition, Australian Meteorological Magazine, Vol.47, 1998, pp