ABSTRACT

The basic findings following from evaluation of freezing nucleus spectra in 351 daily precipitation samples are: 1. The freezing nucleus spectrum may be expressed as a superposition of background and supplementary spectra. The background spectrum is quite conservative and smooth, the supplementary spectrum is expressed by steps superposed on the background spectrum. 2. Concentration of highly active supplementary nuclei increases with increasing instability of stratification of the atmosphere. Supported by similar findings referred in literature, it is hypothesized that the highly active nuclei have their origin from local sources. Regarding their supposed crucial role in cloud glaciation, author suggests to take them into consideration in prediction and modelling the cloud development.

1 Introduction

The drop freezing technique provides the freezing nucleus spectrum of hydrosol particles contained in the aqueous sample. The very laboratory procedure gives very accurate picture of the freezing nucleus content in the aqueous sample. The great problems however arise when one tries to relate the results of the drop freezing experiments applied on precipitation samples to cloud conditions (Rosinski et al., 1974; Rosinski et al., 1973). Firstly, ice nuclei (IN) detected in the precipitation sample are generally not identical with those playing active role in cloud glaciation, and secondly, four heterogeneous ice nucleation modes may function in the cloud, but only immersion freezing is detected by drop freezing experiments. Consequently, one can hardly reconstruct the origin of hydrosolized particles and their possible role in cloud process based on the results of the drop freezing experiment.
This contribution presents findings following from the evaluation of freezing nucleus spectra measured in 351 cumulative precipitation samples collected at the ground level. Supporting the findings with those referred in the literature, and considering these findings rather qualitatively instead of quantitatively, author suggests application on cloud conditions.

2 Experimental

Sampling site is situated on the low elevation (about 50 m above the nearly completely flat terrain), relatively far from the main local sources of air pollutants, at the southern margin of Hradec Kralove. The freezing nucleus content in the precipitation sample is determined by a computer-driven drop freezing technique (Dubrovsky and Skoloud, 1989). The single drop freezing experiment consists in freezing the set (N=100) of 0.001 ml drops resting on the cold stage the surface temperature of which linearly decreases at a rate of 0.05 Ks^-1. The primary outcome of the experiment is a distribution function P(T) being a fraction of drops frozen at temperature above T. Cumulative freezing nucleus spectrum n_f(T) (number of freezing nuclei active above T) may be derived from P(T) according to the relation (Vali, 1971): n_f(T) = - (1/v) ln[1 - P(T)], where v is a volume of a single drop.

3 Results and discussion

Evaluation of freezing nucleus spectra determined on 351 daily precipitation samples¹

¹ The measurements were performed only on precipitation samples of volume greater than 40 ml due to the technique of forming the drops. Two separate measurements were performed on each sample with subsequent cumulation of drop freezing temperatures.

collected during 1983-1987 indicates following, worth mentioning findings: 1. The freezing nucleus spectrum may be expressed as a superposition of background and supplementary spectra. Background spectrum is quite conservative and smooth (possible to be parametrized as n_f0(T) = a[exp(-bT) - 1], where a = 10^-16mm^-3, b = 1.8 K^-1), the supplementary population of freezing nuclei is often expressed by steps superposed on background spectrum. 2. Concentration of highly active supplementary nuclei increases with increasing instability of stratification of the atmosphere.
Some illustrative examples are given in Fig.1. Sample belonging to 84/01/09 is a typical sample containing only background spectrum, while the other samples contain some amount of the supplementary IN effective at warmer temperatures (referred to as super IN in next). Worth to mention is the step-wise character of some distribution functions, with frequent occurrence of -9 C step. The dependence of super IN concentration on stability of atmospheric stratification manifests itself also in a generally lower content of super IN in the winter precipitation samples, while the samples belonging to the stormy summer days with high-intensity showers often contain great concentrations of IN active at temperatures -9 C and warmer.
These findings well correspond with those of other authors: Rosinski and co-authors (Rosinski et al, 1971; Rosinski and Nagamoto, 1976) distinguish two populations of aerosol particles in severe convective storms: the normal background aerosol and aerosolized soil particles. Large aerosol particles of the latter population entrained into the cloud due to the convective activity then give an origin to the warm-temperature peaks on the freezing nucleus spectra. The presence of highly active freezing nuclei in the convective precipitation samples, including the hail samples, is also expressed in measurements by Vali [Vali and Stansbury (1966), Vali (1968, 1969, 1971)]. Some measurements of both authors exhibit the significant nucleation activity maximum belonging to nuclei active at temperatures close to -9 C. Freezing nuclei with activity close to -9 C were also detected in soil samples by Rosinski et al. (1973).
The curve in Fig.1 belonging to the bacterial suspension (Dubrovskþ et al., 1989) evokes idea on the biogenic origin of the atmospheric IN (cf. Schnell and Vali, 1976; Vali et al., 1976). Unfortunately, in time of our measurements no tests were made to learn an origin of the super IN.

4 Suggestions and conclusion

It follows from above that the IN population in the atmosphere consists of background population which may be the rather large-scale property of the air mass, and the supplementary population having origin from local sources. This "local population" may be ingested into the cloud due to the convective activity and may involve giant aerosol particles functioning as highly active IN. Although our measurements do not allow to determine an origin of highly active freezing nuclei in the precipitation samples and their possible role in cloud development, based on Vali (1974)¹ and Rosinski et al (1971)² it appears that at least the part of the super IN found in the

¹ who compared the nucleus spectra for near-simultaneous samples at cloud base and at ground level and concludes, that the majority of freezing nuclei found in the rain have entered the precipitation elements during the formative stages of the precipitation.
² who found highly active IN in the centres of hailstones, indicating that IN active at higher temperatures nucleate some of the hail embryos in severe storms.

precipitation samples played an active role in the cloud glaciation. In case the super IN are ingested into the totally water cloud, the effect of the super IN would resemble the cloud seeding: the type of super IN determines the onset temperature of cloud glaciation and the concentration of super IN determines the initial concentration of ice particles. Due to the scarcity of the super IN and sufficient time delay between the activation of the super IN and background IN (as indicated by length of step on some distribution functions in Fig.1) the ice particles formed on super IN have enough time and enough water to grow to give rise to hail embryos. Consequently, presence of even limited concentration of highly active IN may be of great importance for glaciation process and should be paid corresponding attention.
Prediction of rain. In fact, information on the probability (or onset, course) of cloud glaciation may help while giving prediction of probability of occurrence (or onset, character) of precipitation event. In view of the above findings, the process of the cloud glaciation may be under suitable conditions incited (and in fact governed) by supplementary ingredient of the IN spectrum. The task is then to estimate the shape of the IN spectrum probable to occur in the cloud. The type of IN could be deduced on the basis of knowledge of possible sources of atmospheric aerosol (local sources may predominate in case of the local convective clouds) and the amount of particulate matter ingested into the cloud could be estimated as a function of the actual meteorological conditions with a stress upon the stability of the stratification and wind speed.
Regarding the frequent usage of predictors based on level of intensive crystallization (LIC, often taken as -12 C level) in prediction of convective phenomena, one can consider the LIC temperature as variable dependent upon local weather and local sources of IN.
Parametrization of ice formation. Ice formation rate is usually parametrized to monotonously increase with decreasing temperature. In view of the possible role of super IN, the supplementary population of highly effective ice nuclei of limited concentration is suggested to be considered. The total cumulative IN spectrum n(T) may take the form n(T) = n₀(T) + n'(T), where n₀(T) is a background spectrum and n'(T) is a supplementary spectrum. Background spectrum may be parametrized by some exponential or power-law function, the supplementary spectrum can be parametrized by simple step-wise function, e.g., n'(T) = SUM_i=1,..,r c_i{1 - exp[-(T/T_i)^ai]}, where r is number of types of supplementary IN, c_i is the concentration of the IN of the i-th kind; T_i, a_i are parameters of the activity spectrum of the IN of the i-th kind (T_i is a characteristic activation temperature and a_i characterizes dispersion of the actual activation temperatures around T_i). It is not obvious how to turn the results obtained with a drop freezing technique referred in the previous chapter into the quantitative expression of the IN spectra associated with the individual nucleation modes. In a first approach one may take r = 1; T₁ = -9 C (for immersion freezing; should be suitably modified for other nucleation modes); a₁ = infinity. Concentration c₁ should be given some realistic value. The test of sensitivity upon varying parameters T1 and c1 could be of great value for proper evaluation of possible role of super IN in cloud development.
The presence of supplementary IN population may be also reflected in adequate form in the bulk parametrization schemes by adding some steps of variable hight and characteristic temperature (being possibly dependent upon atmospheric conditions and locality) to ice formation governing functions.

References

Figure

Fig. 1 Distribution functions of the drop freezing temperatures. Triangles, asterices, filled squares, empty squares, x: daily precipitation samples; line : bacterial suspension (Pseudomonas syringae); Thunderstorms occurred on 84/04/16, 84/05/21 and 84/05/23.