By: Karina McCusker
How do we measure cloud ice, and why do we need to?
Ice particles in clouds have complex geometries, making them more difficult to understand than droplets. As a result, ice clouds are a source of uncertainty in weather and climate simulations. To improve this, high-quality global observations are required. Microwave remote sensing instruments, such as radars and radiometers, allow observations of cloud over large areas on a continuous basis. This data is useful for improving microphysical schemes and evaluating numerical weather prediction and climate models.
Measurements of atmospheric cloud ice may also be assimilated into forecasts. For many years a major limitation to forecasts was that only clear-sky radiances were assimilated into numerical weather prediction models, and cloudy cells were discarded. This means a great deal of useful information was lost, thus in recent years a lot of focus was put into enabling all-sky assimilation of satellite radiances. At the ECMWF, all-sky assimilation of microwave data has been shown to have the largest relative impact on the quality of the 24-hour operational weather forecasts of all observations. Currently microwave data is assimilated using a combination of clear-sky and all-sky techniques, but by October an exclusively all-sky assimilation framework will be used for microwave observations.
To address the above points it is necessary to understand the relationship between the size and shape of an ice particle and its microwave scattering properties. To obtain information on particle properties, such as size, shape, and mass, direct measurements are also required. Thus, a novel dataset has been obtained as part of the PICASSO campaign, involving co-located in-situ aircraft observations with remote sensing measurements from ground-based radar. Unique tracking was used to ensure the same cloud was sampled by the aircraft instruments and the radars. Data from synchronized 3, 35, and 94 GHz radars was collected, allowing studies of how scattering by snowflakes changes with wavelength. Further details can be found in this blog post. Here we use the aircraft probes and multi-frequency radar information from the PICASSO dataset to begin to evaluate different ice particle shape models.
What are the benefits of using multiple radar frequencies?
Ice particles in clouds have a wide range of sizes, from frozen cloud droplets of about 10 μm to large aggregates of crystals that can reach 4-5 cm. At 3 GHz (i.e. 10 cm wavelength), the particle size is always less than the wavelength. This means the particles scatter in the Rayleigh regime. If we were to consider the case of homogeneous spheres in the Rayleigh regime, the amount of scattering would increase as the sixth power of particle size (D6). This is a bit more subtle for ice, where generally scattering is proportional to mass2. Either way, larger particles tend to scatter much more than smaller particles. For higher frequencies such as 94 GHz (shorter wavelengths), small particles scatter in the Rayleigh regime, but larger particles (which are comparable in size to the wavelength) will scatter in the Mie regime. This means that by using a combination of measurements at two different frequencies (i.e. the dual-frequency ratio; DFR), we can get a better indication of the size of the particles, consequently improving estimations of ice water content (IWC). Triple-frequency measurements have shown potential for providing information on particle shape/structure and density (e.g Kneifel et al. (2011; 2015)), along with potential to identify regions of aggregation, melting, and riming (Dias Neto et al., 2019). Stein et al. (2015) used triple-frequency observations to evaluate particle models, and we can perform similar experiments using the PICASSO dataset.
Why do we need to evaluate particle models?
RTTOV-SCATT is a fast multiple-scattering radiative transfer model designed to assimilate all-sky MW radiances in numerical weather prediction (Bauer et al., 2006; Saunders et al., 2020; Geer et al., 2021). Assimilation of observations requires accurate hydrometeor scattering models, and optimisation of particle representation is necessary in order to extend all-sky capabilities to include higher frequencies and observations from new sensors, e.g. the Ice Cloud Imager. The default in version 13 of RTTOV-SCATT is to use a range of realistic, non-spherical particles to represent frozen hydrometeors (i.e. snow, graupel, and cloud ice). These are obtained from the ARTS scattering database (Eriksson et al., 2018), as outlined in table 1 of Geer et al. (2021). Here we examine 4 particle mixtures from the ARTS scattering database – plates, columns, block columns, and ICON snow.
Examples of experiments performed
Results of the simulated IWC and radar reflectivity (Z) are shown in Fig. 1 for one of the aircraft runs on 13th February 2018. The in-situ measured particle-size distributions (PSDs) were used to perform the simulations. The red lines show the measurements obtained from the Nevzorov probe and the CAMRa 3 GHz radar, respectively. We find that none of the 4 particle mixtures simultaneously provide a good fit to IWC and Z. However, the block mixture tends to overestimate measurements of both quantities, and the column mixture underestimates measurements.
Figure 1: – (a) Simulated and measured IWC. The IWC measured using the Nevzorov probe is shown in red, with the IWC simulated using the in-situ measured PSDs and the 4 particle mixtures shown by the other colours, as outlined in the figure legend. (b) Same as panel (a) but for the 3 GHz radar reflectivity.
We also looked at the dual-frequency ratio, and found that columns predict a larger value of DFR(3,35) than the other mixtures (Fig. 2a), while ICON snow predicts a lower value of DFR(35,94) than the other shapes (Fig. 2b). Fig. 2c shows the DFRs calculated for all the runs during this case study, plotted in triple-frequency space with DFR(35,94) on the x-axis and DFR(3,35) on the y-axis. The dots are the values simulated using the in-situ measured PSDs, and the lines are simulated using exponential PSDs. The large variation of the dots from the lines shows that even if the chosen shape model is realistic, commonly-used parameterisations of the PSD (such as assumptions of exponential and gamma distributions) may still introduce a large error to the calculations.
We are currently in the process of comparing the DFR simulations to measurements in order to evaluate the particle models and determine whether any of them are realistic. This work will be useful to guide microphysical schemes and assumptions that are used in weather and climate models, and in data assimilation.
Figure 2: (a) Simulated DFR calculated at 3 and 35 GHz for one of the aircraft runs. (b) Same as panel (a) but for 35 and 94 GHz. (c) The two DFRs calculated for all the runs during this case study, plotted in triple-frequency space with DFR(35,94) on the x-axis and DFR(3,35) on the y-axis. The dots are the values simulated using the in-situ measured PSDs, and the lines show the results calculated using exponential PSDs.
Kneifel, S., M. S. Kulie, and R. Bennartz, 2011: A triple‐frequency approach to retrieve microphysical snowfall parameters, J. Geophys. Res., 116, D11203, https://doi.org/10.1029/2010JD015430.
Kneifel, S., A. von Lerber, J. Tiira, D. Moisseev, P. Kollias, and J. Leinonen, 2015: Observed relations between snowfall microphysics and triple-frequency radar measurements. J. Geophys. Res. Atmos., 120, 6034– 6055, https://doi.org/10.1002/2015JD023156.
Dias Neto, J., and Coauthors, 2019: The TRIple-frequency and Polarimetric radar Experiment for improving process observations of winter precipitation, Earth Syst. Sci. Data, 11, 845–863, https://doi.org/10.5194/essd-11-845-2019.
Stein, T. H. M., C. D. Westbrook, and J. C. Nicol, 2015: Fractal geometry of aggregate snowflakes revealed by triple-wavelength radar measurements, Geophys. Res. Lett., 42, 176–183, https://doi.org/10.1002/2014GL062170.
Bauer, P., E. Moreau, F. Chevallier, and U. O’Keeffe, 2006: Multiple-scattering microwave radiative transfer for data assimilation applications, Quarterly Journal of the Royal Meteorological Society, Wiley, 132 (617), pp.1259-1281, https://doi.org/10.1256/QJ.05.153.
Saunders, R., J., and Coauthors, 2020: RTTOV-13 science and validation report, EUMETSAT NWP-SAF.
Geer, A. J., and Coauthors, 2021: Bulk hydrometeor optical properties for microwave and sub-mm radiative transfer in RTTOV-SCATT v13.0, Geosci. Model Dev. Discuss. [preprint], https://doi.org/10.5194/gmd-2021-73, in review.
Eriksson, P., R. Ekelund, J. Mendrok, M. Brath, O. Lemke, and S. A. Buehler, 2018: A general database of hydrometeor single scattering properties at microwave and sub-millimetre wavelengths, Earth Sys. Sci. Data, 10, 1301–1326, https://doi.org/10.5194/essd-10-1301-2018.