TITLE: Cloud Condensation Nuclei Spectra on C-130 AUTHORS: James G. Hudson and Seong Soo Yum DAS-DRI 2215 Raggio Pkwy. Reno, NV 89512-1095 775-674-7020 fax -7007 hudson@dri.edu 1.0 DATA SET OVERVIEW: These are measurements of the cloud condensation nuclei (CCN) spectra with the original Desert Research Institute (DRI) airborne CCN spectrometer (Hudson 1989), which was on board the NCAR C-130 during INDOEX. Data reported at this time are only for the three best gradient flights Feb. 20, 24, and March 4. These were the most interesting INDOEX flights as they covered both the polluted northern Hemisphere and clean maritime southern Hemisphere. Throughout the project concentrations were rather similar throughout the northern and southern Hemispheres. Therefore the measurements reported here are typical of these two distinct regions throughout INDOEX. Data reported here are only for the 3 to 4 hour period of the north-to- south transect from Male (4 degrees north) to about 9 degree south latitude along a constant longitude of about 73 degrees east. These low level transects took place during the early parts of these flights at altitudes below about 1 km (900 mb). Some of these data are presented by Hudson and Yum (2001). The same data files can also be found at the DRI anonymous ftp site (ftp.dri. edu at the subdirectory ccndata/indoex) 2.0 INSTRUMENT DESCRIPTION: This instrument is a parallel plate thermal gradient diffusion cloud chamber (Hudson 1989). This instrument also has a streamwise supersaturation gradient, as each plate is divided into eight temperature controlled zones. The temperature gradient between the plates then increases in eight steps as the sample passes from the entrance to the optical particle counter (OPC); this is a distance of about 30 cm (chamber length), which generally took about 30 s. The temperature gradient between the two parallel plates increases from 0 to about 8 degrees Celsius at mean temperatures that approximate cabin temperature. The plates are oriented vertically so that the 1.5 cm separation between the two plates is horizontal and transverse to the flow of sample, which is also horizontal along the length of the cloud chamber (30 cm). Sample enters through a vertical tube transverse to the length. This tube is placed between the plates at the end away from the OPC. This sample tube has 21 holes spread along the middle 15 cm of the entire 30 cm width of the cloud chamber, which is actually vertical, transverse to the plate separation and length of the cloud chamber. Sample is then spread vertically across 15 cm of the 30 cm width and is confined to about the central mm of the 1.5 cm plate separation (the transverse dimension) by a clean air sheath flow, which is set up upstream of the sample entrance. This confines the sample to the maxima supersaturations near the central plane between the plates. The sample is also kept away from the uncontrolled plastic sidewalls (actually at the top and bottom) of the cloud chamber as the holes only extend the middle 15 cm of the 30 cm "width." All of the sample and the clean air then converges through a small opening at the end of the chamber and passes through the Royco OPC, which counts and sizes the droplets individually. The increasing supersaturation field within the chamber causes the better nuclei (those with lower critical supersaturations Sc) to begin condensing water earlier and to grow faster than higher Sc nuclei. Thus the better nuclei (generally the larger particles, though this really depends on the number of soluble ions, which is characterized by Sc i.e., Hudson and Clarke 1992) make larger droplets. The division according to droplet size (actually channel number 128 channels) is then related to the Sc of the nuclei. This relationship is established by calibrations with monodisperse soluble electrolyte particles (i.e., salt usually ammonium sulfate) of known composition and size, which determines the Sc (Gerber et al. 1977). Several sizes are used to produce a calibration curve of Sc versus channel number that is then applied to all sample spectra (concentration versus channel number) to produce a CCN spectrum of concentration versus Sc. This is usually expressed cumulatively since this is the way CCN act in clouds. All nuclei with Sc below specific values generally turn into cloud droplets in an adiabatic (unmixed) cloud parcel. It is assumed that particles with the same Sc regardless of differences in size and composition (which are what determines Sc) will grow to equally sized droplets if exposed to the same supersaturation history either in this or any cloud chamber or in the atmosphere. This has been demonstrated for inorganic soluble electrolytes (Gerber et al. 1977) but it has been questioned (e.g., Facchini et al. 1999) for some organic compounds that may compose some CCN. If this is true it would not only undermine the technique used in the DRI CCN spectrometer but it would undermine any CCN measurements and the comparisons of CCN with real clouds, which is their fundamental value. This challenge essentially says that for some particles Sc does not completely characterize the growth of some particles in supersaturated environments. We are currently carrying out laboratory tests of various organic particles to determine the validity of this speculation. Preliminary results so far are negative with respect to this speculation. That is, we have not found discernable differences in the relative growth rates of water on the organic particles that have been tested with respect to inorganic particles. In other words we deduce similar Sc values for these organic aerosols even with different supersaturation histories in the cloud chamber. In other words when we operate the chamber with different delta temperatures between the plates (different supersaturations) or with different flow rates (this changes the time of exposure to the supersaturations) we still deduce the same Sc for the same sized organic particles. This suggests that Sc is the only, or at least the major, determinant of droplet growth. In the end this may or may not be true for every potential organic CCN but at least the DRI CCN spectrometer produces some sort of hierarchy of the likelihood that particles will turn into cloud droplets. This is what is needed in order to make useful comparisons with cloud droplet spectra. It must be borne in mind that organic particles were probably in greater abundance in northern Hemisphere INDOEX than in the southern Hemisphere or most previous projects. Nonetheless, even in this cloud chamber Sc of the sample particles is not the only determinant of droplet growth and final channel number. The sample has a certain spread in two dimensions (with respect to the distance from the central plane between the temperature controlled plates and with respect to the uncontrolled plastic sidewalls), which causes different portions to be exposed to slightly different supersaturations. The convergence at the end of the chamber also causes differences in the supersaturation history of the particles. Furthermore the optics and electronics cause a further spread in the final distribution that causes inherent broadening of even the most monodisperse input signals. This artificial spreading of the signals means that channels that should have relatively lower concentrations can inadvertently receive signals from adjacent channels and that channels with higher concentrations can appear to lose counts. This tendency can be inhibited by inverting the data as described by Yum and Hudson (2000). This requires knowledge of the artificial spread of the calibration data, which requires knowledge of the actual spread of the monodisperse calibration particles. We feel that we have done a reasonable job of this and note that for supersaturations above 0.1% there are negligible inversion corrections. It would be too tedious and inefficient to apply the inversion algorithm to each data record. Thus, the inversion correction was applied to averages over several data records that appeared to be consistent. Therefore a low time resolution data set is supplied with the inversion corrections and a higher time resolved data set is presented without the inversion correction. These are generally similar for Sc above 0.1% except for the different time resolution. The average over the same period of the high time resolved data should match the inverted data for Sc above 0.1%. The error in concentration generally increases for lower Sc. This is mainly because the shape of the spectrum generally becomes steeper for lower S. A steeper spectrum causes more error as the concentration changes more readily with Sc. The error is concentration is probably about 10% and the error in supersaturation is about 10% for Sc above 0.1% and higher for lower Sc. The precision is probably much better. 3.0 DATA COLLECTION AND PROCESSING: The instrument was operated throughout the flights. The sample flow rate was directly measured with a mass flow meter and this was combined with the measured sample integration time period to present concentrations in number per cubic cm. The sample flow rate is measured with normalization to standard temperature and pressure. Therefore the concentration is normalized to the volume at surface pressure. For this low altitude data set the difference from concentration at the ambient pressure is less than 10% anyway. The sample flow meter was calibrated before and after the experiment. Accuracy of the flow rate is 10% but precision is better. Calibrations to designate the Sc of the droplet size channels were carried out at least once per flight. 4.0 DATA FORMAT: Data are presented in space delimited ASCII code. There are three files of about 1000 records (rows) for the original uninverted data. These files are designated feb2099o.txt, feb2499o.txt, and mar0499o.txt. These have much greater time resolution than the inverted files, which are feb2099i.txt, feb2499i.txt, and mar0499i.txt. For the original uninverted data the first column is the universal date and time at the beginning of the counting interval. This is in accord with the directed convention. The 2nd column is the ambient pressure according to the gage on the DRI spectrometer. This is not the same as the real ambient pressure measured by the C-130 but it is roughly proportional. The 3rd column is the total CCN concentration at some unspecified supersaturation above 1%. The next 10 columns are the cumulative concentrations at the various Sc. For the inverted data the 1st column is the midpoint time of the count duration shown in the 2nd column. The third column is the ambient pressure and the 4th column is the total CCN concentration. The next 14 columns are the cumulative concentrations at the specified Sc. Note that there are 4 additional Sc values for the inverted data, 0.7, 0.5, 0.3, and 0.15%. 5.0 DATA REMARKS: Periodically the instrument did not monitor the ambient aerosol as it was often used to measure processed sample, which was done by heating the sample, passing it through a size discrimination (DMA) or through the counterflow virtual impactor (CVI) (Twohy et al. 2000). The latter was usually done only during in-cloud measurements, which presented artifacts for the ambient data collection anyway. At any rate these sample processing and measurements done within clouds were edited from this data set. This accounts for most of the interruptions of the continuous data. Other interruptions were accounted for by calibrations. 6.0 REFERENCES: Facchini, M.C., M. Mircea, S. Fuzzi, and R.J. Charlson, 1999: Cloud albedo enhancement by surface-active organic solutes in growing droplets. Nature, 401, 257-259. Gerber, H.E., W.A. Hoppel, and T.A. Wojciechowski, 1977: Experimental verifications of the theoretical relationship between size and critical supersaturation of salt nuclei. J. Atmos. Sci., 34, 1410-1420. Hudson, J.G., 1989: An instantaneous CCN spectrometer. J. Atmos. & Ocean. Techn., 6, 1055- 1065. Hudson, J.G., and A.D. Clarke, 1992: Aerosol and cloud condensation nuclei measurements in the Kuwait plume. J. Geophys. Res. 97, 14533-14536. Hudson, J. G. and Yum, S. S., 2001: Spatial distribution of cloud condensation nuclei spectra over the Indian Ocean. J. Geophys. Res. Submitted Oct. Twohy, C. H., J.G. Hudson, J.R. Anderson, S.K. Durlak, S.S. Yum, and D. Baumgardner, 2000: Characteristics of cloud nucleating aerosols in the Indian Ocean region. J. Geophys. Res., in press. Yum, S.S., and J.G. Hudson, 2000: Vertical distributions of cloud condensation nuclei spectra over the springtime Arctic Ocean. J. Geophys. Res., in press.