Many measurements of stratiform precipitation have been performed at different locations (Mt. Uetli, SAMM I, SAMM II). The perfect case is characterised by long-lasting and uniform precipitation with a descending or ascending melting layer passing a station equipped with our instruments. During the years, several such cases were measured. Below, one of this cases is shown in detail, then general results of all cases together are shown.

A case study

On 17/18 February 1995, an unusual precipitation event took place in the region of Zurich. The most striking feature of this event was the uniformity of the precipitation over a wide range (> 30 km) and a long time (> 6 h). Precipitation was triggered by a very weak cold front. No strong winds were observed nor a rapid change of any meteorological quantity. All changes occured slowly and gradually. Due to these conditions, the melting layer was descending slowly and steadily by 400 m during the 5.5 hours of the experiment.

Microphysical measurements were performed along the steep slope of Mt. Uetli with a vertically pointing X-band Doppler radar and a disdrometer at the base of the mountain and an optical spectrometer on top of it. At the beginning of the measurements, the melting layer was above the optical spectrometer, at the end it was located between the optical spectrometer and the X-band radar/disdrometer. The optical spectrometer at the top was first measuring rain, then melting snow and at last dry snow. The disdrometer at the bottom was measuring rain all the time. The radar picture below shows the evolution of the precipitation during 5.5 hours.

The position parameter H

The definition of the position parameter H: Every radar profile with a pronounced bright band and a sufficient vertical resolution shows three characteristic features: the top, the bottom and the maximum of the bright band. The top of the bright band marks the beginning of the melting of ice particles. The bottom of the bright band is generally believed to mark the end of melting, i.e. to be at the height where all the ice particles have melted to raindrops. However, the maximum of the bright band is not so obviously connected to melting alone, but is a consequence of the sum of several microphysical properties. Nevertheless, the maximum of radar reflectivity is the most prominent feature of a stratiform radar profile and will be included in the definition of the position parameter H.

First, the top of the bright band is determined. The new position parameter H is set to 1 at the top of the bright band and to zero at the maximum of radar reflectivity. The figure shows a profile of radar reflectivity. The solid horizontal line is the position of the maximum radar reflectivity (H=0) and the dashed line indicates the top of the bright band (H=1). With this, the part above the maximum of radar reflectivity has positive values of H. Since radar profiles are usually not symmetric, in a second step H is set to -1 at the bottom of the bright band. With this, the part below the maximum of radar reflectivity has negative values of H.

This parametrisation has many advantages. First of all, measurements made in different precipitation events with the 0 deg C isotherm at different heights can be compared easily. A position parameter of, e.g., H=0.5 indicates that measurements were made in the middle of the upper part of the melting layer, independent of its actual height above ground level. Another advantage is the elimination of the effect of different rainrates. The vertical extension of the melting layer and hence of the bright band depends on the rainrate. With the conception of the position parameter H, cases with different rainrates may easily be compared.

It is proposed that the position parameter H is a much better reference to describe microphysical properties within the melting layer than the vertical distance to the top of the bright band. Microphysical observations at the same position H of different cases are assumed to be more consistent than at the same vertical distance below the top of the bright band. Two examples, proving the validity of this assumption, can be found further below.

Particle size and velocity distribution

Particle size and velocity distribution is measured with the optical spectrometer. The two time series below illustrate the change of hydrometeor size distribution and fall velocity as the melting layer passes the station with the optical instrument.

Particle size distributions are depicted on the left. Average distributions have been calculated for 15-minute intervals. Three intervals are missing (20:30, 20:45, 21:15) due to a system breakdown. Maximum hydrometeor sizes depicted are limited to 15 mm in order to enhance the clarity at small particle sizes. However, between 21:45 and 22:45 large snowflakes with sizes over 20 mm have been sampled.

At the beginning, particle fall velocities (depicted on the right) are equal to the fall velocity of raindrops (solid line in the figure). Suddenly, large hydrometeors appeare, the fall velocities of which differ considerably from the Gunn and Kinzer curve. At 21:30, two regimes are clearly visible. Particles with sizes up to approx. 1.5 mm have fall velocities equal to the velocity of raindrops. These particles are completely melted. Particles larger than 1.5 mm have, with growing size decreasing fall velocities. The largest particles have a fall velocity of below 2 m/s. The decreasing fall velocity with increasing size illustrates the different stage of melting. While particles around 2 to 3 mm are almost completely melted, large particles with sizes over 10 mm have just begun to melt, if at all. The next few intervals show a similar behaviour with a clearly visible upper limit for the size of the completely melted drops. This limit decreases from 1.5 mm at 21:30 to about 0.5 mm at 22:15.

No upper limit for the size of the completely melted drops can be found in the spectra after 22:15. The fall velocities of the particles are similar to the fall velocity given by Locatelli and Hobbs (1974, dotted line) for aggregates of side planes and by Zikmunda (1972, dashed line) for general aggregates.

The particle number flux

To compare the measurements made with the optical spectrometer with the data obtained with the disdrometer, a quantity called the number flux is introduced. The number flux is defined as

where n_i is the number of particles sampled during one minute in the size class i and A_i is the size of the corresponding measuring area. The number flux can be defined for the optical spectrometer as well as for the disdrometer. In the figure left, the number flux at the disdrometer (open circles) and at the optical spectrometer (solid boxes) is shown in dependance of the position parameter H. Here it has to be noted that the position parameter H is a reference to the height only of the optical spectrometer within the melting layer. It does not contain any information about the position of the disdrometer. The disdrometer stands throughout the whole experiment at the bottom station and measures rain from the begin to the end. If the NF at one of the instruments is given e.g. at H=-0.5 this means that the optical spectrometer measures melting particles at H=-0.5 within the melting layer while the disdrometer measures rain at some distance below the melting layer. However, the NF at the disdrometer can also be depicted in dependence of the position parameter H, simply by using the same transformation from time to H as was used to calculate the position of the optical spectrometer.

The number flux ratio

In order to compare the number fluxes at the optical spectrometer and the disdrometer, a second quantity is introduced which is called the number flux ratio (NFR). The number flux ratio is defined as

The number flux ratio is equal to one if both instruments measure the same number of particles and smaller than one if the optical instrument measures less particles than the disdrometer.

General results

Aggregation and breakup of hydrometeors

Using the number flux at two instruments, one (disdrometer) always below the melting layer and measuring rain and one (optical instrument) somewhere within, below or above the melting layer measuring all types of hydrometeors, the quantity number flux ratio is calculated By observing the behaviour of the number flux ratio in dependance of the position parameter H several assumptions can be made:

The largest snowflakes are observed within 100 m of the height of the maximum of radar reflectivity. This height is found to correspond to temperatures between 0.5 and 1.5 deg C. At the bottom of the bright band, the temperature is approx. 2.5 deg C.

By comparing number fluxes (number of hydrometeors falling through one square meter within one minute) at the top and at the bottom of the melting layer (NFR=1 at H=1), it seems that, on the average, one snowflake yields one raindrop.

By comparing number fluxes in the upper part of the melting layer and in rain below the melting layer (NFR decreasing at -0.3<H<0.5), aggregation seems to be the cause of a rapid decrease of the number of particles within the upper part of the melting layer. This strong aggregation leads to the large snowflakes observed in the middle of the melting layer.

By comparing the number flux in the lower part of the melting layer and in rain below the melting layer (NFR increasing at -0.3>H>-1), breakup seems to be responsible of a rapid increase of the number of particles within the lower part of the melting layer. Spontaneous and collisional induced breakup seems to occur frequently within the lower part of the melting layer where the range of fallspeeds of the hydrometeors is rather large.

Melt distances

By comparing particle fall velocities within the melting layer with the fall velocity of raindrops, snowflakes and partly melted particles can be separated from raindrops. Thus, the largest melted hydrometeor can be determined as a function of the fall distance from the top of the bright band. Therefore, melt distances of snowflakes of given mass can be measured in natural rain.

Left picture: Melt distances measured on four different days (dots, squares, triangles and plus signs). Right picture: Comparison to model calculations of melt distances of different authors. The tiny black dots represent the experimentally measured melt distances (same as in the left picture). The two black curves labeled A and B are melt distances calculated by Wexler (1955) with a lapse rate of 6 deg/km. The large dots (red, green, yellow) are data taken from Matsuo and Sasyo (1981). Melt distances are calculated for different ambient humidity and a constant lapse rate of 6 deg/km. The blue cross is the melt distance of snowflakes with melted diameter of 2.7 mm obtained by Mitra et al (1990) in a wind tunnel. The blue dots are calculated values of Mitra et al. The symbols number 1 to 4 are all calculated with 100% humidity, number 5 with 95% humidity and number 6 with 90% humidity. Number 1 and 2 use a lapse rate of 10 deg/km, the rest of 6 deg/km

The largest melted hydrometer can be depicted in dependence of the position parameter H instead of the distance to the top of the melting layer. The two pictures above show this. The left picture shows again the four different cases. The right picture indicates the according precipitation rates. For small melted diameters, the concept of the position parameter H is much better suited to describe the microphysics than the distance to the top of the melting layer.

Temperatures within the melting layer

Temperatures were measured while the melting layer passed a station. It seems that measurements at the top station of Mt. Rigi are not very good (blue and orange dots).