1. Introduction
Giant (2–20-μm diameter) and ultragiant (>20-μm diameter) aerosol particles are of particular interest in the warm rain process. Deliquesced, soluble giant particles and any ultragiant particles1 can enter a cloud and initiate coalescence quite quickly; if they are larger than 50-μm diameter, their collection efficiencies are greater than 20%. Although their importance to warm rain production has not been firmly established, numerous modeling studies have shown that they can produce radar echoes ≥10 dBZ in less than twenty minutes (e.g., Ochs and Semonin 1979; Johnson 1982), and more recent studies that incorporated radar observations have discussed their potential for dominating the early radar echoes (Lasher-Trapp et al. 2001; Knight et al. 2002).
Woodcock (1952, 1953) first presented size distributions of giant salt particles (deliquesced at 99% relative humidity) collected off the coasts of Hawaii and Florida. Particles were collected by impaction on coated glass slides extended from an aircraft and were later sized under a microscope in a temperature- and humidity-controlled environment. Other studies have documented giant and ultragiant aerosol particles, both soluble and insoluble, over continents and oceans (e.g., Reitan and Braham 1954; Okita 1955; Nelson and Gokhale 1968; Noll and Pilat 1971; Mészáros and Vissy 1974; Johnson 1976), with maximum concentrations ranging from over 10−2 cm−3 at the smaller sizes to less than 10−8 cm−3 at the larger sizes. These studies have shown that concentrations of these particles in the atmosphere may vary by several orders of magnitude, depending on meteorological processes, local sources, altitude, and numerous other factors. Thus, it appears critical to take measurements of the sizes and concentrations of ultragiant aerosol particles (UGA) during field campaigns where their influence on radar echoes and precipitation needs to be quantified.
2. Some estimates of ultragiant aerosol particles from the SCMS
a. The SCMS
The Small Cumulus Microphysics Study (SCMS) was a field campaign conducted near Cape Canaveral, Florida, to collect observations on small warm cumuli and the initial stages of the precipitation process. A dual-wavelength radar and three different instrumented aircraft were deployed for this purpose. The Florida coastal region can be influenced by air over the continent and ocean, depending on the prevailing wind direction. Thus, special patterns (long, overlapping ovals at low levels) were flown by the National Center for Atmospheric Research (NCAR) C130 aircraft each day of the project to characterize the aerosol in and above the boundary layer.
b. Estimates from the 260X probe
The 260X is a one-dimensional optical array probe (manufactured by Particle Measuring Systems, Inc.) that records the cross-flight dimension of shadows of particles passing through its laser beam. For the SCMS, the probe was modified to measure particle sizes 51–1054-μm diameter at 17-μm resolution. The size-dependent sample volumes range from ∼30 to 4000 cm3 (100 m)−1 of flight. (These values were determined by calibrations conducted after the SCMS.) This probe is thus capable of providing estimates of the number of UGA from the long periods of clear-air flight described above. For example, if the UGA concentration is ∼10−4 cm−3 for 51-μm particles (∼10−6 cm−3 for 204-μm particles), an averaging time of 4000 s is needed to measure this concentration with an uncertainty of ∼30%, for a flight speed of 100 m s−1. Some examples of size distributions computed from the SCMS 260X data are shown in Fig. 1. The size distributions appear quite realistic and are not a result of a few anomalously high values over short time intervals.
Laird et al. (2000) compiled composite size distributions of UGA over the entire duration of the SCMS using clear-air data from the 260X probe. Their analysis showed that UGA were present in number concentrations similar to the giant particles (deliquesced at 99% relative humidity) measured by Woodcock (1953) in weak-to-moderate winds off the coasts of Hawaii and Florida. Laird et al. (2000) concluded that the data were not contaminated by noise because the resulting size distribution from the particle counts was exponential, not flat, and not limited to one channel.
However, comparisons of the 260X data with a 2DC probe mounted on the aircraft suggest that the particles recorded by the 260X were not always real. The 2DC is a two-dimensional optical array probe (manufactured by Particle Measuring Systems, Inc.) that records two-dimensional images of particles passing through its laser. During the clear-air periods when the 260X recorded particles, very few or no images appeared on the 2DC, nor were there many instances that the probe triggered from particles too small to record images. The usable size ranges of the 260X (51–1054 μm) and 2DC (50–800 μm) overlap well, and the sample volume of the 2DC (5 L maximum for 100-m flight) is even greater than that of the 260X (3–4 L maximum for 100-m flight). In addition, the 2DC diodes are only required to be shadowed 40% when particles pass through the laser, versus the 70% shadowing required by the 260X to record a particle. Thus, when the 260X records particles, the 2DC should as well. The 2DC probe appears to be working correctly during the SCMS, because the 260X and 2DC size distributions from in-cloud penetrations agree well.
Additional evidence indicates that the 260X data are contaminated by noise. When the aerosol legs and aircraft soundings are partitioned into periods during which the aircraft was below 1500 m MSL (which included the aerosol legs) and periods during which it was above 1500 m MSL (during the sounding, with maximum height over 3 km MSL), the concentrations of UGA measured by the 260X fail to decrease with height, as expected of these large particles due to their relatively large settling velocities. Figure 2 shows size distributions measured by the 260X in clear air below and above 1500 m on three different days; the concentrations are nearly identical. Other days investigated in the same way show similar results.
Perhaps the most convincing evidence that a problem is occurring with the 260X comes from data collected toward the end of the SCMS. At this time, two short flights were conducted to compare data from the three different one-dimensional optical array probes flown on the three different aircraft during the project. For the comparison, all three probes (NCAR's 260X, the University of Wyoming's 1DC, and Météo-France's 200X) were flown on the NCAR C130. Results show that the 260X often recorded counts in clear air when both of the other 1D probes had none (Fig. 3). The 2DC, also present on the C130, also showed no images or counts from particles too small to record images.
We thus conclude that the distributions of UGA given by Laird et al. (2000) are contaminated by noise. The spurious counts recorded by the 260X are most likely not a problem in a cumulus cloud penetration because the time interval of interest is so short, but when averaged over long intervals in clear air, they dominate the measurements.
c. Estimates from the 2DC probe
Particle images from the 2DC, and lack thereof, can be used to provide some insight into potential concentrations of UGA from the boundary layer in the SCMS area. Six flights having the longest clear-air periods flown by the NCAR C130 at low altitudes are used; the time periods, altitude, relative humidity and 2DC images are listed in Table 1. Two of the legs (one being as long as 3000 s) had no 2DC images; the others had very few images. The occurrence of images does not appear to be related to the ambient relative humidity.
Because of the scarcity of images, concentrations derived directly from those images are misleading because of the poor counting statistics. In addition, it is impossible to determine if the noise affecting the 260X also may have been affecting the 2DC, creating false particle images. As a result of these uncertainties, the 2DC data listed in Table 1 are used to derive upper limits on the UGA size distributions, rather than actual measured size distributions, for UGA 50 < D < 200 μm, at a 95% confidence level.
The upper limits calculated in this manner for the six flights in Table 1 are shown in Fig. 4. The size distributions all fall within a relatively narrow region, regardless of the number of particles and the size of the accumulated sample volume. The maximum concentration of particles over the 100–200-μm size range shown is ∼6 × 10−6 cm−3.
Overlaid on Fig. 4 for comparison are the size distributions from Woodcock (1953), evaluated at 60% and 90% relative humidity, for data collected in weak wind (W) and tropical storm–force wind (S). The 60% and 90% relative humidity cases are meant to bracket the relative humidities observed for the SCMS cases listed in Table 1. It is important to note that the sizes of particles measured by Woodcock off the coast of Florida are only greater than 20-μm diameter when their deliquesced sizes at very high relative humidities are considered; otherwise they are all smaller, and undetectable by the optical array probes used for the SCMS.
The UGA project-composite distribution for 85%–90% relative humidity based on the 260X data from Laird et al. (2000) is shown in Fig. 4 as well, corrected for an error that made the concentrations an order of magnitude too low as reported in the original paper (Laird et al. 2000, corrigendum). It is quite clear that the scarcity of 2DC images implies that the concentrations of UGA present in the boundary layer are much less than the composite distribution derived from the 260X data, by at least three orders of magnitude. Figure 5 compares the upper limits derived from the 2DC data with the 260X distributions for some of the individual cases listed in Table 1 (the 260X was not operating on 24 July and 6 August, so those cases are not plotted). The individual 260X distributions are again at least three orders of magnitude greater than the upper limits derived from the 2DC data, and much more at the larger sizes.
3. A possible source of the 260X noise
The source of noise to the NCAR 260X data from the SCMS is at this time unknown, although it does not appear to be attributable to radio frequency interference. Two temperature sensors mounted near the nose of the C130 were known to be affected by use of the aircraft radio, and when these data are compared to those from the 260X it is clear that the noise affecting the 260X occurred equally as often when the radio was and was not being used.
The source of the noise does appear to be consistent with the effects of electromagnetic interference on one-dimensional and optical array probes studied by a group at the GKSS Research Center at the Institute for Atmospheric Physics in Hamburg, Germany (D. Nagel 2000, personal communication). In an unpublished study, they found that electromagnetic radiation from the 28-V aircraft generators or the aircraft inverters (which convert the raw 28-V dc from the aircraft generators or alternators to 110-V ac for the probe electronics) could trigger the forward-scattering spectrometer probe FSSP-100 and cause narrow streaks on the 2D probe images. They also found that 400-Hz inverters produced much more noise than 50-Hz inverters. This noise occurred as counts in any channels of an FSSP-100 or 2DC and could produce an exponential size distribution. It is impossible to distinguish this kind of noise from real particle counts. More than one frequency of interfering radiation was required to replicate the variety of noise observed on the probes, and certain probes were sensitive to certain frequencies, while other probes were completely immune to those frequencies. This could explain why the 260X during the SCMS was subject to noise while the 2DC was not.
The GKSS group solved this noise problem by adding filters at the output of the aircraft inverters, as well as shielding and filtering all the data and power lines from within the probe to the data-acquisition system. Further study is needed to determine if this problem, and this solution, are relevant to the 260X noise on the NCAR C130.
4. Summary and discussion
Previous estimates of UGA derived from the SCMS 260X data have been shown to be most likely contaminated by noise. The 260X data are inconsistent with data from other 1D probes flown during the project, as well as the 2DC, when flown in clear air. The possibility of electromagnetic noise producing artificial particle counts has been discussed. New upper limits of UGA have been presented here, based on data from a 2DC probe. These upper limits are three orders of magnitude lower than the estimates from the 260X data.
Much higher concentrations of UGA than those estimated from the 2DC data have been used to explain the appearance of early radar echoes in Florida cumuli from the SCMS, with reasonable results (Lasher-Trapp et al. 2001). The reason for this apparent discrepancy is at this time unknown. One possibility is that the UGA concentrations in the boundary layer fluctuate over much smaller spatial and temporal scales than those represented here. (This cannot be investigated with the SCMS dataset because such long time-averaging is required to attain adequate sample volumes.) Of course, another equally likely possibility is that UGA are not the source of the large drops responsible for the early radar echo in these clouds. Both possibilities are topics for future work.
The noise contamination of the NCAR 260X data in clear air may not be limited to the SCMS. Investigation of a few flights from the Indian Ocean Experiment (INDOEX) showed similarities to the SCMS data: the 260X recorded very large particles in clear air, with no corresponding images on the 2DC.
Finally, based on the limitations experienced during this field campaign, it has become clear that better instrumentation is urgently required to measure ultragiant particles in the atmosphere. Upper limits of UGA concentrations in clear air calculated here with the SCMS 2DC data still vary by an order of magnitude, limiting their usefulness for any future studies quantifying their effect in clouds. Because of the large sample volumes required (several to hundreds of cubic meters at a minimum, for UGA as large as 200-μm diameter), the present electronic optical array probes are inadequate, especially if quantification of temporal and spatial variability is required. An additional desirable aspect in any new instrumentation would also be the ability to measure the solubility of the UGA, in order to include it in microphysical calculations of their growth.
Acknowledgments
Dr. Charles Knight provided considerable comments on this work as it evolved, as well as on the manuscript. Dr. Walter Strapp suggested the connection between the inconsistencies in the probe measurements with the work performed at the GKSS and helped make contact with Dr. Dagmar Nagel, who shared her experience with electromagnetic interference and the microphysical probes. Dr. Greg McFarquhar provided the data from INDOEX. This work was supported in part by NSF Grants ATM-9707301 and ATM-9981937.
REFERENCES
Johnson, D. B., 1976: Ultragiant urban aerosol particles. Science, 194 , 941–942.
Johnson, D. B., . 1982: The role of giant and ultragiant aerosol particles in warm rain initiation. J. Atmos. Sci., 39 , 448–460.
Knight, C. A., Vivekanandan J. , and Lasher-Trapp S. G. , 2002: First radar echoes and the early ZDR history of a Florida cumulus. J. Atmos. Sci., in press.
Laird, N. F., Ochs III H. T. , Rauber R. M. , and Miller L. J. , 2000: Initial precipitation formation in warm Florida cumulus. J. Atmos. Sci, 57 , 3740–3751. Corrigendum,. 58 , 2668–2669.
Lasher-Trapp, S. G., Knight C. A. , and Straka J. M. , 2001: Early radar echoes from ultragiant aerosol in a cumulus congestus: Modeling and observations. J. Atmos. Sci., 58 , 3545–3562.
Mészáros, A., and Vissy K. , 1974: Concentration, size distribution and chemical nature of atmospheric aerosol particles in remote oceanic areas. J. Aerosol Sci., 5 , 101–109.
Nelson, R. T., and Gokhale N. R. , 1968: Concentrations of giant particles below cloud bases. Preprints, First National Conf. on Weather Modification, Albany, NY, Amer. Meteor. Soc., 89–98.
Noll, K. E., and Pilat M. J. , 1971: Size distribution of atmospheric giant particles. Atmos. Environ., 5 , 527–540.
Ochs, H. T. I. I. I,, and Semonin R. G. , 1979: Sensitivity of a cloud microphysical model to an urban environment. J. Appl. Meteor., 18 , 1118–1129.
Okita, T., 1955: Interaction between atmospheric aerosols and raindrops. J. Meteor. Soc. Japan, 33 , 220–226.
Reitan, C. H., and Braham R. R. , 1954: Observations of salt nuclei over the midwestern United States. J. Meteor., 11 , 503–506.
Taylor, J. R., 1982: An Introduction to Error Analysis. University Science Books, 270 pp.
Woodcock, A. H., 1952: Atmospheric salt particles and raindrops. J. Meteor., 9 , 200–212.
Woodcock, A. H., . 1953: Salt nuclei in marine air as a function of altitude and wind force. J. Meteor., 10 , 362–371.
Size distributions from 260X data collected in clear air for (a) 1442–1450 UTC 22 Jul 1995 and (b) 1403–1440 UTC 24 Jul 1995
Citation: Journal of Atmospheric and Oceanic Technology 19, 3; 10.1175/1520-0426(2002)019<0402:MOUAPI>2.0.CO;2
Size distributions from the 260X in clear air (all times UTC): (a) 1513–1527 19 Jul, (b) 1530–1540 19 Jul, (c) 1415–1430 22 Jul, (d) 1430–1440 22 Jul, (e) 1323–1332 11 Aug, and (f) 1342–1359 11 Aug 1995. Total concentration and aircraft altitude are shown for each plot in upper right-hand corner
Citation: Journal of Atmospheric and Oceanic Technology 19, 3; 10.1175/1520-0426(2002)019<0402:MOUAPI>2.0.CO;2
Time series of total concentrations (L−1) of particles recorded by four different optical array probes mounted on the NCAR C130 during flight 20 of the SCMS. No particles were recorded by the 2DC, the 1DC, or the 200X
Citation: Journal of Atmospheric and Oceanic Technology 19, 3; 10.1175/1520-0426(2002)019<0402:MOUAPI>2.0.CO;2
Comparison of size distributions of UGA from different studies. Squares denote data from Woodcock (1953) for strong (S) and weak (W) wind cases at relative humidities of 60% or 90%. Open circles denote composite distribution from 85% to 90% relative humidity during the SCMS from Laird et al. (2000) using 260X data. Solid and dashed lines denote upper limits on UGA size distribution based on SCMS 2DC data for the six cases listed in Table 1
Citation: Journal of Atmospheric and Oceanic Technology 19, 3; 10.1175/1520-0426(2002)019<0402:MOUAPI>2.0.CO;2
As in Fig. 4, except size distributions of UGA from 260X data for four individual days are compared to the corresponding upper limits derived from 2DC data. The plot is truncated at a diameter of 500 μm, but counts at sizes as large as 884 μm were recorded by the 260X on 21 Jul
Citation: Journal of Atmospheric and Oceanic Technology 19, 3; 10.1175/1520-0426(2002)019<0402:MOUAPI>2.0.CO;2
Description of six clear-air flight segments by the NCAR C130 during the SCMS in 1995
The National Center for Atmospheric Research is sponsored by the National Science Foundation.
For the duration of this note, particles with D > 50 μm will be called ultragiant, since the probes used here cannot distinguish the size of the original particle before it was deliquesced, or even if it has deliquesced.
