Published in Proceedings of the Workshop on Synergy of Active Instruments in the Earth Radiation Mission, 12-14 November 1997, GKSS Research Centre, Geesthacht, Germany.

Radar-lidar observations of clouds during the CLARA-campaigns

H.W.J. Russchenberg, V.K.C Venema, Delft University of Technology
International Research Centre for Telecommunication-Transmission and Radar,

A.C.A.P. van Lammeren, Royal Netherlands Meteorological Institute

A. Apituley, National Institute of Public Health and the Environment

Correspondence: H..W.J. Russchenberg, Delft University of Technology-IRCTR
P.O. Box 5031, 2600 GA Delft, The Netherlands, E-mail:

download article
You can download this article. We have a two versions:

1. Introduction

The vertical cloud structure is a key element in the energy budget of the atmosphere. On a global scale, however, too little experimental data is available. To meet this deficit, the European Space Agency is considering to operate the Earth Radiation Mission: a space-based combination of radar and lidar to monitor the clouds. To study the benefit of this combination, ground-based experiments were done in The Netherlands during the Clouds and Radiation-campaigns in 1996 [Van Lammeren et al, 1997].

In this paper, two different examples of cloud observations will be discussed: stratocumulus and altostratus. Based on the observations, some preliminary remarks concerning space-based observations will be made.

2. The instruments

The Delft Atmospheric Research Radar DARR is an S-band FM-CW Doppler radar. It operates at 3.315 GHz, whereas other cloud radars operate at frequencies above 35 GHz. The radar can observe scattering due to cloud particles and clear-air phenomena. Gaseous absorption as well attenuation due to hydrometeors is negligible. The cloud measurements were done with a range resolution of 15 or 30 meter, depending on the meteorological situation. The temporal resolution is 5.12 seconds. The sensitivity of the radar is -30 dBZ at 2 km.

The lidar system for the measurements in this paper is a Vaisala CT75. It has a range resolution of 30 meter, and a temporal resolution of 30 seconds. To cross-check the range accuracy of the Vaisala-lidar and the radar an experiment was performed by pointing them at a chimney at a distance of approximately 1100 meter from the radar site.

For cloud observations, the radar and lidar were pointed towards the zenith. The systems were collocated at the Faculty of Electrical Engineering in Delft, near the coast of the North Sea.

3. Particle scattering

The radar wavelength is much larger than the average cloud droplet size, which implies that the radar signal is due to Rayleigh-scattering: the received power is proportional to the 6th order moment of the dropsize distribution. The lidar wavelength is, however, much smaller than the cloud droplets, which implies that scattering is happening in the resonance-region. The received power is approximately proportional to the 2nd order moment of the dropsize distribution.

Table 1 gives the drop concentration that is needed to cause the reflectivity level given in the first column, assuming a uniform dropsize distribution. It shows that the radar reflectivity is very sensitive to the dropsize. A reflectivity of -30 dBZ can be caused by a billion drops of 10 m m, which is of the order occurring in stratus clouds, or by one droplet of 316 m m. This means that in case of precipitating or nearly precipitating clouds, the reflectivity may be dominated by a few large drops. Table 1 also gives the optical thickness that corresponds to the radar reflectivity, for the given diameter and concentration, and assuming a cloud thickness of 500 m. The shaded part of the table gives the optically very thin clouds. Note that they are still very well distinguishable with the radar.

Table of typical reflections and optical densities.
Table 1 Necessary drop concentration to cause typical reflectivity levels as function of dropsize. For each drop size and corresponding concentration the optical thickness is given. The last column gives the diameter of a single drop that gives the reflectivity given in the first column. The grey part of the table indicates the optically very thin clouds.

4. Impact of the vertical cloud structure of stratocumulus on the radar-lidar observation

In stratocumulus clouds the liquid water content as well as the mean dropsize is often seen to increase with height [Martin et al, 1994] [Gerber, 1996] [Slingo et al, 1982]. This will have an impact on the vertical profiles of the lidar and radar signals, due to differences in the scattering mechanisms in the microwave and optical range. In the absence of attenuation, the radar will receive its maximum power from the top of the cloud. The lidar signal, however, will have a peak in the lower part of the cloud: the extinction prevents the lidar from penetrating into the cloud towards the top. From space, however, the situation will be different: both the radar and the lidar will see a maximum near the cloud top.

The actual situation is of course more complex than these idealized clouds. The concentration is not constant and the height dependence of the liquid water content is never truly linear, but the idealization clearly shows a trend.

5. Measurements

5.1 Example of stratocumulus

Figure 1 shows the radar reflectivity and the lidar backscatter as function of time. Throughout the measurement, a thin cloud layer is observed at approximately 1700 meter. In the second half of the measurement, the convective activity in the boundary layer is increasing as can be recognized in the enhanced speckle patterns in the radar plot. The cloud layer breaks up in separate cumuli. The lidar backscatter varies has a maximum backscatter of 3000 [1000 srad km]^-1; the radar reflectivity varies between -15 and -30 dBZ.

The differences between the radar and lidar are intriguing. They give different cloud heights. The radar sees a cloud that is approximately 75 meter higher than that the lidar sees. These measurements seem to justify the considerations of section 4: the radar and lidar receive the maximum return from different heights; the sensitivity of the radar is probably not good enough to observe the cloud base.

It is not shown in this paper, but further analysis of this measurement revealed that the temporal correlation between the maximum received power from the lidar and the radar was quite significant, even though the two instruments did not observe the peak reflections at the same heights. Details can be found in [Russchenberg et al, 1997].

measurement of stratocumulus.
Figure 1 Height-time diagram of the radar reflectivity with contours of the lidar backscatter on April 19, 1996; Stratocumulus breaking in to cumuli after 10 utc. The lidar backscatter varies has a maximum backscatter of 3000 [1000 srad km]^-1. Resolution: 15 meter, 5.12 seconds.

5.2 Example of altostratus

Figure 2 shows the radar reflectivity and the lidar backscatter as function of height and time. A thick layer occurs between 2500 and 5000 meter. The isolated radar reflections underneath the cloud layer are still of unknown origin, but are often seen in case of wind coming from the North Sea. The lidar backscatter has a maximum backscatter of 300 [1000 srad km]^-1; the radar reflectivity varies between 0 and -20 dBZ.

Two hours before the cloud passage a radiosonde was launched: between 2500 and 5000 meter the temperature varied from 0 ° C to -16 ° C. However, measurements with an infra-red radiometer during the cloud passage showed a sky temperature varying between -10 and -30 ° C. In case of optically thick clouds, the infra-red sky temperature would have been equal to the temperature at the cloud base. The -10 ° C that was observed therefore implies an optically thin cloud (with an emissivity less than 1).

The large radar reflectivity indicates the presence of ice crystals: it can not be caused by small water droplets, unless it is a optically thick cloud, but that would have resulted in a large lidar signal as well: the lidar backscatter is, however, a factor 10 smaller than in case of the measurement of stratocumulus (Figure 1).

Some portions of the cloud are, although giving a lidar-reflection, not observed by the radar. This seems to indicate the lidar reflection is mainly due to the supercooled water droplets in the cloud, or by ice crystals that are too small to resolve with the radar.

The heights where both instruments received the maximum power from varies: sometimes the lidar gets its maximum from a larger heights than the radar and sometimes it is the other way around [Russchenberg et al, 1997].

Measurement of altostratus.
Figure 2 Height-time diagram of the radar reflectivity and lidar backscatter on April 18, 1996; Altostratus. The lidar backscatter varies has a maximum backscatter of 300 [1000 srad km]^-1. The horizontal contour line at 1500 meter is due to the top of a boundary layer. Resolution: 30 meter, 5.12 seconds

6. Conclusions

The lidar signal from clouds is due to scattering in the resonance-region and the radar signal is due to Rayleigh-scattering, which makes the two instruments differently sensitive to the size of the cloud droplets, and consequently to the vertical distribution of water or ice inside the cloud. In case of clouds in which the water content increases with height, the radar receives its maximum from the upper half of the cloud. The extinction of the lidar signal will in these cases result in a peak somewhere at the bottom of the cloud. This has been illustrated with an example of observations of stratocumulus. From space the radar and lidar will both see a maximum at the cloud top, in that case.

In altostratus, ice crystals may dominate in the radar return, whereas the lidar signal is still representative for the liquid water content. This may possibly be used to derive the vertical distribution of water and ice from synergy of radar and lidar.


The work described in this paper is supported by the National Research Program on Global Change NRP, the European Space Agency ESA and the Netherlands Technology Foundation STW.


Gerber, H., 1996, 'Microphysics of marine stratocumulus clouds with two drizzle modes', J. Atmos. Sci., Vol 53, No 12, pp1649-1662

Lammeren. A.C.A.P., Russchenberg, H.W.J., Apituley, A., Ten Brink, H. 'CLARA: a data set to study sensor synergy', these proceedings

Martin, G.M., Johnson, D.W, Spice, A., 1994, 'The measurement and parametrization of effective radius droplets in warm stratocumulus clouds', J. Atmos. Sci., Vol 51, No 13, pp1823-1842

Russchenberg, H.W.J., V.K.C Venema, A.C.A.P. van Lammeren, A. Apituley, 1997, 'Cloud measurements with lidar and a 3 Ghz radar, ESTEC Final report PO 151912', Noordwijk, The Netherlands

Slingo, A., Nicholls, S., Schmeitz, J., 1982, 'Aircraft observations of marine stratocumulus during JASIN', Quart. J. R. Met. Soc., 108, pp833-856