Improving model representation of cloud ice using cloud radar and aircraft observations

By: Peggy Achtert

Understanding the evolution of the ice phase in clouds is of great importance for understanding the development of thunderstorms and the formation of heavy rain. However, cloud ice poses an enormous challenge for both measurements and modelling. While we can probe ice particles in the atmosphere with remote-sensing instruments that make use of electromagnetic radiation over a range of wavelengths that spans from many centimeters down to hundreds of microns, we cannot get a direct measure of the parameters we need to know to understand the role of cloud ice in atmospheric processes. For this, we need to have a model that describes the scattering of electromagnetic radiation by ice particles of different sizes, a model that describes the size distribution of the ice particles, and a relationship between particle mass and particle size.

We have designed an atmospheric experiment to evaluate the three assumptions implicit in remote-sensing techniques by performing collocated measurements with multiple ground-based radars and the Facility for Airborne Atmospheric Measurements (FAAM, https://www.faam.ac.uk/) research aircraft. With this new knowledge, we want to test and develop microphysical parameterizations used in atmospheric models.

Within the Parameterizing Ice Clouds using Airborne Observations and Triple-frequency Doppler Radar Data (PICASSO) project, we want to look at how the scattering behaviour of snowflakes changes with wavelength. For this, we are operating three scanning radars at Chilbolton Observatory (https://www.chilbolton.stfc.ac.uk/Pages/home.aspx) that are probing the same cloud volumes: the 3 GHz Chilbolton Advanced Meteorological Radar (CAMRa) radar (25m antenna), the 94 GHz Galileo radar installed on side of the CAMRa antenna, and the 35 GHz Kepler cloud radar, with its antenna slaved to the CAMRa radar. The combination of the different radars allows for detecting ice particles over a wide range of sizes. More importantly, the difference in radar reflectivity between two wavelengths – say 10 cm and 9 mm [3 and 35GHz] – is related to both the size and the shape of the ice particles. For any given assumption about the particle shape (or scattering model) we can directly predict what we should get for a different wavelength pair from the measurements. This takes one degree of freedom from the problem and enables a further consistency check. A match of observations and scattering model is evidence that the latter is an appropriate choice. A mismatch tells us that something is wrong with the scattering model. While a few studies of that type, i.e. using multi-wavelengths radar observations, have been conducted, there has so far been almost none with collocated in-situ sampling to support the interpretation of the remote-sensing data. Within PICASSO, the ground-based remote sensing is therefore complemented by airborne measurements of cloud droplets and ice particles up to 2 cm in size together with the concentration of ice and liquid water. These independent measurements provide the “truth” that we want to be able to retrieve from the remote-sensing measurements.

Figure 1: Picture of the CAMRa antenna dish at Chilbolton Observatory with the FAAM research aircraft in the background (red circle).

Normally in a campaign like this we might scan our radars up and down a prearranged flight radial and attempt to match up the radar and in-situ data after the fact. This approach generally leads to substantial scatter in the comparison between radar and aircraft. We therefore used a real-time position feed from the aircraft to drive the antenna automatically – a technique that builds on other work at Chilbolton to track satellites and other objects in space. In a nutshell, the radar antennas track the aircraft’s movement as it flies towards Chilbolton. The aircraft can be identified in the display of radar reflectivity in Figure 2 as a thin line of strong signal at 1 km height that reaches as far as 60 km from Chilbolton. In the closest few km the antenna runs ahead to reach vertical as the aircraft performs an overpass.

Figure 2: Distance-height display of radar reflectivity measured with CAMRa at Chilbolton Observatory during PICASSO. Warm colours refer to strong signal while cold colours refer to weak signals. The thin red line at 1 km height marks the flight path of the FAAM research aircraft.

In the PICASSO data set, we can now select the radar measurements as close as possible to the aircraft echo, and plot a time series of reflectivity along the aircraft track. We can then calculate the same thing from the in-situ particle size distributions. In a preliminary analysis of an ice cloud sampled during PICASSO, we find that the reflectivity calculated from the in-situ measurements is highly correlated with the radar observations. But if we select a specific mass-size relationship commonly used in the analysis of radar observations, we find a significant (factor of 4) difference in magnitude. This suggests that the ice particles in this particular cloud were around twice as dense as predicted by today’s parameterizations. In the next steps of this work, we will use the full set of data collected during PICASSO to evaluate the available parameterizations and models used in the analysis of radar observations and, if necessary, propose new relationships for an improved retrieval of cloud ice from radar data.

This entry was posted in Clouds, Microphysics and tagged . Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *