1. OBSERVATION ON ION DYNAMICS Urmas Hõrrak, Hannes Tammet Institute of Environmental Physics, University of Tartu, 18 Ülikooli St., 50090 Tartu, Estonia. E-mail: Urmas.Horrak@ut.ee

3. ESTONIA Location of Tahkuse Observatory

4. Measurements Location: Tahkuse Observatory (58°31'N 24°56'E), a sparsely populated rural region. Instrumentation: three original multichannel aspiration spectrometers (second-order differential mobility analyzers). Method: particle classification by electrical mobilities. Mobility range: 0.00041–3.14 cm 2 V –1 s –1, logarithmically divided into 20 intervals. Mobility spectra of positive and negative ions are measured in every 5 minute. Height of measurements: 5 m above the ground. The database consists of 8900 hourly average spectra measured during the period Sept. 1993 – Oct. 1994.

5. Diagram of the measuring system and the small air ion spectrometer (cross-section). E is electrometer, HVS is high voltage supply VS is voltage supply. External dimensions of the cylindrical aspiration spectrometer: height 695 mm, diameter 122 mm.

6. Classification of air ions Average spectra of air ions at Tahkuse. Sept.1993 - Oct.1994.

7. Classification of air ions and aerosol particles Class of air ionsMobility range cm 2 V –1 s –1 Diameter range * Class of particles Small cluster ions1.3 – 3.20.36 – 0.85Clusters Big cluster ions0.5 – 1.30.85 – 1.6Clusters Intermediate ions0.034 – 0.51.6 – 7.4Nanometer particles Light large ions0.0042 – 0.0347.4 – 22Ultrafine particles Heavy large ions0.00087 – 0.004222 – 79Aitken particles * Estimates of equivalent diameter ranges assume single charged particles. Hõrrak, U., Salm, J. and Tammet, H. (2000) Statistical characterization of air ion mobility spectra at Tahkuse Observatory: Classification of air ions. J. Geophys. Res. Atmospheres 105, 9291–9302.

8. Small or cluster ions Average spectra of air ions. Sept. 1993 - Oct. 1994. Tahkuse Observatory

10. Average characteristics of small ions (means and standard deviations) ParameterSmall ions, cm –3 Mean mobility, cm 2 V  1 s  1 Conductivity of small ions, fS m  1 Polarity (+) 274  961.36  0.066.0  2.1 Polarity (–) 245  881.53  0.106.0  2.1 Ratio (+/–) 1.13  0.070.89  0.041.00  0.05 Average mobility spectra of small ions. Sept. 1993 – Oct. 1994. Tahkuse.

11. The frequency distributions (histograms) of the mean mobility of positive and negative small ions

12. Diurnal variation in the concentration of small ions (0.5– 3.14 cm 2 V  1 s  1 ) and wind speed. August 21 – August 30, 1994.

13. Average diurnal variation in small (or cluster) ion concentration Average diurnal variation of positive cluster ion concentration in the warm season (Sept.1993, May – Sept. 1994) and in the cold season (Nov. 1993 – April 1994). Statistics: median, box (25% and 75%) and whiskers (10% and 90% quantiles).

14. Variation in the activity concentration of radon at Tahkuse. August 19 - September 18, 1998.

15. Correlation coefficients (in percent) between positive air ion mobility fractions. November 1, 1993 – Aprill 30, 1994.

16. Scatterplot: the concentrations of small positive ions (0.5–3.2 cm 2 V  1 s  1 ) versus heavy large ions (52 –79 nm). Cold season (Nov. 1993 – April 1994), Warm season (May–Sept. 1994).

17. Evolution of cluster ion mobility spectra in wintertime The effect of aerosol particle concentration on the mobility distribution of cluster ions

18. Evolution of cluster ion mobility spectra in summertime The combined effect of the concentration aerosol particles and radon (Rn 222 ) on the mobility distribution of cluster ions

19. Small and big cluster ions It is rational to subdivide cluster ions (0.5–0.14 cm 2 V  1 s  1 ) into two classes of small and big cluster ions (with the boundary at 1.3 cm 2 V –1 s –1 for negative ions and 1.0 cm 2 V  1 s  1 for positive ions). Subdividing leads to more distinct shape of the diurnal variation in the concentration of these ion classes. The ratio of the concentrations of small and big cluster ions is closely correlated with the mean mobility of cluster ions (the correlation coefficients are 97% and 95% for positive and negative small ions, respectively).

20. Diurnal variation in the median concentration of positive and negative small cluster ions and big cluster ions in the warm season (Sept. 1993, May – Sept. 1994).

21. Average diurnal variation of the natural mean mobility of small positive and negative ions in the warm season (Sept. 1993, May – Sept. 1994). Statistics: median, box (25% and 75%) and whiskers (10% and 90% quantiles).

22. Variations in the mean mobility of negative and positive small ions and air temperature at Tahkuse, September 1–30, 1993.

23. Variations in the mean mobility of negative and positive small ions and air temperature at Tahkuse. December 23, 1993 - January 31, 1994

24. Variations in the mean mobility of negative and positive small ions and accumulation mode aerosol particle (100–500 nm) concentration at Tahkuse from April 14 to May 16, 1994.

25. Conclusions The average diurnal variation of cluster ions at Tahkuse had the ordinary shape for the continental stations of high latitude. The results are in accordance with the hypothesis that radon is the main ionizing agent, which causes the variation in the ionization rate, and therefore in the concentration of cluster ions near the ground. The different behavior of small and big cluster ions during diurnal cycle causes the diurnal variation in the natural mean mobility of small ions. The diurnal variation of the mean mobility was considerable only in the warm season while in the cold season the mean mobility was mainly correlated with the changes in air masses.

26. The factors of the mobility distribution of cluster ions We suppose that the changes in the mobility distribution of cluster ions are due to the changes in the chemical composition and the concentration of some trace gases or vapors in the air, probably generated by photochemical reactions; the changes in the ionization rate and the concentration of aerosol particles. The last two factors have an effect on the lifetime of cluster ions.

27. Burst events of intermediate ions (charged nanometer particles ) in the atmospheric air Average spectra of air ions. Sept. 1993 - Oct. 1994. Tahkuse Observatory

28. Variation in the concentration of positive intermediate ions, air temperature, and relative humidity at Tahkuse, Sept. 11–27, 1993.

29. Time series of the concentration of positive intermediate ions, air temperature, relative and absolute humidity. Tahkuse, Oct. 10 - 28, 1994.

30. Variation in the concentration of positive intermediate ions, air temperature, relative and absolute humidity. Tahkuse, November 5 - 18, 1993 relative and

31. Time variation of the concentration of aerosol particles (100 - 560 nm) and intermediate ions (1.6 - 7.4 nm). Tahkuse 14.04 - 16. 05. 1994. and

32. Variation in the concentration of positive intermediate ions, the mean mobility of negative and positive small ions and relative humidity. Sept. 11–27, 1993.

33. Scatterplot of the mean mobility of small ions versus intermediate ion concentration. Sept. 1993 – Oct.1994.

34. Contour plot of the evolution of air ion spectrum at Tahkuse. October 20, 1994

35. Evolution of positive air ion mobility spectra. Oct. 20, 1994.

36. Evolution of positive ion mobility spectra at Tahkuse. March 22, 1994.

37. Contour plot of the evolution of air ion spectrum at Tahkuse. May 22, 1996.

38. Contour plot of positive air ion spectrum. 19.11 - 20.11.1995, Tahkuse

39. Average diurnal variation of positive air ion concentration during the burst events. Tahkuse 1.09.1993 - 27.10.1994.

40. Diurnal variation in the median concentration of intermediate ions (diameter 2.1–7.4 nm), light large ions (7.4–22 nm) and heavy large ions (22–79 nm) in the warm season. Sept. 1993, May – Sept. 1994.

41. Evolution of positive and negative air ion mobility spectra. December 5, 1993, Tahkuse.

42. Time variations in the fraction number concentrations of aerosol particles (PNC), positive air ions (INC) and global radiation. Nov.19, 1995.

43. Location of the measurement points (Tahkuse, Tõravere and Kellasaare)

44. Monthly concentrations of intermediate ions (1.6 - 7.4 nm) Sept. 1993 – Oct. 1994

45. Statistics of intermediate ion burst events. Sept.1993- Oct.1994. The number of days in the month, when the concentration of positive intermediate ions exceeds a certain value. * burst of short duration: October, December within 2 hours, January within 3 hours more than 100 cm  3.

46. CONCLUSIONS The size distribution of aerosol ions (1.6–22 nm) is strongly affected by the nucleation bursts. The photochemical nucleation can initiate a burst of intermediate ion concentration (charged aerosol particles in the size range of 1.6–7.4 nm) and subsequent evolution of aerosol ion size spectra below 80 nm. The generated new aerosol particles grow toward sizes 10–15 nm during 2–3 hours. In general, the disturbed region of air ion size spectra affected by the bursts is 1.1–34 nm (0.002– 1.0 cm 2 V  1 s  1 ) including the groups of big cluster ions, nanometer particles and a fraction of Aitken particles. The frequency of occurrence of the bursts (about 80 per year) has peaks in spring and autumn.

47. Diurnal variation of aerosol ion concentration during the non–burst days During the non–burst days, the average diurnal variation in the concentration of aerosol ions was weak, in general, and displayed different behavior considering the cold and warm season.

48. Diurnal variation in the median concentration of positive cluster ions (0.5–3.14 cm 2 V  1 s  1 ), intermediate ions (2.1–7.4 nm), light large ions (7.4–22 nm) and heavy large ions (22–79 nm) in the cold season (November – March).

49. Diurnal variation in the median concentration of positive cluster ions (0.5–3.14 cm 2 V  1 s  1 ), intermediate ions (2.1–7.4 nm), light large ions (7.4–22 nm) and heavy large ions (22–79 nm) in the warm season (May – September).

50. Diurnal variation in the median concentration of heavy large ions (22-79 nm) and small ions in the warm season (Sept. 1993, May - Sept. 1994) and in the cold season (Nov. 1993 - March 1994).

51. The diurnal variation of heavy large ion concentration in the warm season is generally opposite to that in the cold season This contrast can be explained by the differences in the mixing of the boundary layer air and by the different processes of aerosol particle generation (combustion of fuel versus gas-to-particle conversion and radiolytic processes initiated by 222 Rn, 220 Rn decay).