Abstract

This project addresses the need to enhance knowledge about the distribution of hydrometeors in clouds during the onset phase of precipitation and its representation in numerical models. SPOCC proposes to derive estimates of the hydrometeor ratio of the above-mentioned particle types from spectrally resolved polarimetric Ka-band cloud radar observations as well as from a spectral-bin microphysical model. Basis for the work is the quality-assured dataset obtained during the Analysis of the Composition of Clouds with Extended Polarization Techniques (ACCEPT) field experiment (Myagkov et al., 2016b), during which co-located observations of a scanning hybrid-mode Ka-band and a vertical-stare LDR-mode cloud radar, scanning polarimetric S-Band radar, multiwavelength polarization lidar, Doppler lidar, and atmospheric radio-soundings were performed for the first time. ACCEPT was conducted in Cabauw, the Netherlands in fall 2014.

Status 2023

Contribution of the Observations Work Package: Spectrally resolved retrieval of hydrometeor shape and orientation

With respect to the previous update, the retrieval technique including correction for wind effects and spectral separation technique was finalized. The operational retrieval is in the following described by means of a case study from Cabauw, NL observed on 3 November 2014. On that day around 2000 UTC, LDR (linear depolarization ratio) values from cloud top (around 6 km height) to cloud base (around 2 km height) decreased from -20 dB to -30 dB (Fig. 1).

The decrease of LDR between cloud top and base indicates a transition in particle shape, which is in the following evaluated based on the spectrally resolved retrieval technique. After correction for aliasing (Doppler folding) effects we obtain for each height level of 30 m the Doppler spectra of differential reflectivity ZDR and correlation coefficient RHV as a function of the elevation angle from 30° to 90° (zenith-pointing). In Figure 2, an example is depicted for the RHI scan obtained at 20:00 UTC on 3 Nov 2014 at a height level of 2900 m.

It can be seen that ZDR and RHV behave differently for different elevation angles and velocities. For velocities close to 0 m/s, the ZDR spectrogram shows increased values of up to 3 dB for low elevation angles. Also the RHV values are generally lower at the slow falling velocities. Therefore, a first conclusion is that particles with slowest fall speed have different shapes compared to the particles with faster fall speeds. Using the spectrally resolved retrieval approach, the spectrogram is therefore split into 5 parts (see Fig. 4) in order to quantify the particle shape and orientation as a function of Doppler velocity.
Figure 3 shows the relationships of ZDR and RHV as a function of elevation angle, linearly distributed into 5 parts between minimum and maximum detected velocity at each elevation angle.

All Doppler parts show individual relationships between ZDR, RHV and elevation angles. They can be interpreted as follows:

1. In part 1, which related to fastest falling particles, ZDR values fluctuate strongly around zero and values of RHV are quite chaotic. In this part, likely effects of low-SNR (not shown) and potentially also Mie scattering effects cause a strong fluctuation of the ZDR and RHV values. Therefore, the Doppler part of the fastest falling hydrometeors should be handled with care. It is required to define an increased threshold value for SNR in the spectrally resolved approach to minimize statistical uncertainties.
   
2. In part 2, values of ZDR fluctuate smoothly around zero and values of RHV is close to 1 (a bit far from 1 for low elevation angles) which indicates spherical particles.
   
3. In part 3, ZDR for zenith angle is around zero, while it gradually increases with decreasing elevation angles. RHV is also close to unity for all elevation angles, which is indicative for light oblate particles.
   
4. The behavior of part 4 is similar to part 3. But the ZDR values increase more strongly with decreasing elevation angles compared to part 2. Also RHV is still close to one for all elevation angles. These relationships are indicative for the presence of oblate particles.
   
5. In part 5 the ZDR and RHV signatures are quite different. The gradient of ZDR is greater than for part 3 and 4. RHV values at 90° are lowest and incease with decreasing elevation angles. This behaviour of ZDR and RHV indicates the presence of prolate particles. The observed fluctuation of ZDR and RHV is again due to the presence of rather low SNR values.
   

Figure 4 shows the retrieval result using the spectrally resolved approach. Parts 2, 3 and 4 yield oblate particles as expected form the spectrum. The shape of the particles for part 4 is more oblate compared to parts 2 and 3. Parts 1 and 5 return prolate particles. As discussed already, result of part 1 can be considered unreliable, but part 5 can return prolate particle as expected.

In conclusion, the spectrally resolved approach revealed that a mix of different particle types was present in the discussed case study of 03 Nov 2014, 20:00-20:15 UTC. In contrast, the original main-peak approach, that provides the shape retrieval only for the strongest SNR peak, would show only the presence of isometric particles.

Contribution of the Modelling Work Package: Spectrally resolved retrieval of hydrometeor shape and orientation

During the Analysis of the Composition of Clouds with Enhanced Polarization Techniques (ACCEPT) campaign led by TROPOS, which occurred at the Cabauw Experimental Site for Atmospheric Research in the Netherlands from 7 October to 17 November 2014, we demonstrated the effectiveness of our measurement technique.
The ACCEPT campaign was meanwhile simulated using the COSMO model (Consortium for Small-scale Modeling, Baldauf et al., 2011) of the German Weather Service coupled with the two-moment microphysical scheme (Seifert and Beheng, 2006). To enable comparison with the observational data, the simulation outputs were converted into radar variables using one of the radar foward simulators, the Passive and Active Microwave TRopospheric RAinfall Retrieval (PAMTRA) (Mech et al., 2020).
Figure 5 shows that the (a) radar reflectivity derived from the COSMO-2M simulation results using PAMTRA closely matched the (b) radar reflectivity observed during the campaign and provided by CLOUDNET (Illingworth et al., 2007) for the 02 November 2014. The simulation outputs capture well for the mixed-phase cloud at 3-6 km altitude, despite minor timing discrepancies. Our future studies will employ the spectral-bin model.

References

• Baldauf, M. et al., 2011: Operational Convective-Scale Numerical Weather Prediction with the COSMO Model: Description and Sensitivities, https://doi.org/10.1175/MWR-D-10-05013.1.
• Illingworth, A. J. et al., 2007: Cloudnet. Continuous Evaluation of Cloud Profiles in Seven Operational Models Using Ground-Based Observations, https://doi.org/10.1175/BAMS-88-6-883.
• Mech, M. et al., 2020: PAMTRA 1.0: the Passive and Active Microwave radiative TRAnsfer tool for simulating radiometer and radar measurements of the cloudy atmosphere. Geoscientific Model Development, 13(9):4229–4251, doi:10.5194/gmd-13-4229-2020.
• Seifert, A., and K. D. Beheng, 2006: A two-moment cloud microphysics parameterization for mixed-phase clouds. Part I: Model description, in: Meteorology and Atmospheric Physics 92, pp. 45–66.

Status summer 2021

Contribution of the Observations Work Package: Spectrally resolved retrieval of hydrometeor shape and orientation

Recent developments within SPOCC focused on the derivation of multiple hydrometeor species from the range-height-indicator (RHI) scans of the polarimetric hybrid-mode Ka-band cloud radar. This is a delicate task, since the Doppler spectra from the lowest elevation angle of 30° and the highest elevation angle of 90° cannot be compared directly. Horizontal wind effects and differential contributions of the fall velocity of the hydrometeors to the Doppler spectra at the different elevation angles need to be considered. After correction for these effects based on observed profiles of the horizontal wind field, the Doppler spectra at all elevation angels were split into a series of 5 parts. The polarimetric shape and orientation retrieval can then be applied to each of the parts of the Doppler spectrum, separately. An example of the extended shape retrieval is shown for a measurement from Cabauw, NL, taken on 3 November 2014. The 24-hour overview of that day is shown in Fig. 1. A warm-frontal system with embedded precipitation passed the measurement site on that day. The melting layer was located at approximately 1.5 km height, as can be seen from the radar bright band (dark red color). Vertical profiles of the derived polarizability ratio (density-weighted geometric axis ratio) and degree of orientation (deviation of particle orientation from horizontal alignment) are shown in Figure 2. Therein, Figure 2(a) illustrates the result of the plain main-peak retrieval (Myagkov et al., 2016) when applied to a series of 4 RHI scans which were performed between 20:00 and 20:15 UTC. If only the main peak of the Doppler spectrum is considered, a polarizability ratio of approximately 1 (isometric) is derived at almost all heights (except for the invalid region of the melting layer). When the spectrally resolved retrieval is applied to the RHI scan, a different picture is obtained, as Figure 2(b) shows. The slowest falling Doppler spectral part 5 (pink curve) shows particles which are more oblate then the other 4 parts. Also the degree of orientation of part 5 is closer to perfect horizontal alignment (orientation of +1). This illustrates that besides the rather fast falling isometric particles (dominating in parts 1-4 of the Doppler spectrum), also smaller (slower falling) but strongly dendritic (oblate), horizontal aligned particles were present in the observed cloud system. In conclusion, the case study presents the value of the extended shape and orientation retrieval. It can provide valuable insights into the shape partitioning in complex mixed-phase clouds systems, including the detection of regions of secondary ice formation, aggregation or riming. It also enables one to track how ice particle shapes change during precipitation between cloud top to melting layer. The next steps of Work Package 1 comprise the publication of the extended spectrally resolved shape retrieval and the application of the retrieval to prominent case studies of precipitation formation and mixed-phase clouds.

Contribution of the Modelling Work Package: Sensitivity of mixed-phase cloud microphysics to aerosol perturbations

The simulations within Work Package 2 were dedicated to the investigation of the response of mixed-phase cloud microphysical properties to perturbations in the number concentration of cloud condensation nuclei (CCN) and ice nucleating particles (INP). The main method used is the spectral-bin microphysical methodology called AMPS (Advanced Microphysics Prediction System; Hashino et al., 2020). The AMPS was coupled with a simple 1-D dynamic core KiD (Kinematic Driver for microphysics intercomparison). Different microphysical schemes, such as the Morrison 2-moment, TAU spectral bin model, Thomson 2009, and Thomson 2007 were applied to the KiD-AMPS model setup. The impact of the CCN and INP perturbations on mixed-phase cloud properties were then investigated for all of the investigated microphysical schemes. As an example, Figure 3 shows the response of the ice water mass simulated for contrasting, but overall realistic, CCN and INP concentrations with the AMPS microphysics scheme. A strong susceptibility of the microphysical evolution to the CCN and INP conditions was found. A more detailed investigation revealed that immersion and contact freezing were the main drivers for primary ice formation and that the increasing CCN and INP concentrations were correlated with an increase in the number ice aggregates.

Next steps will of Work Package 2 will be to couple the output of the simulations to radar forward simulators in order to evaluate the detectability of the aerosol effects on ice microphysics with radar remote sensing techniques.

References:

• Hashino, T., de Boer, G., Okamoto, H., & Tripoli, G. J., 2020: Relationships between Immersion Freezing and Crystal Habit for Arctic Mixed-Phase Clouds—A Numerical Study, Journal of the Atmospheric Sciences, 77(7), 2411-2438. Retrieved 13 Sep 2021 from https://journals.ametsoc.org/view/journals/atsc/77/7/jasD200078.xml.
• Myagkov, A., Seifert, P., Bauer-Pfundstein, M., and Wandinger, U., 2016: Cloud radar with hybrid mode towards estimation of shape and orientation of ice crystals, Atmos. Meas. Tech., 9, 469–489, https://doi.org/10.5194/amt-9-469-2016.

Status 2020

The work in the first year addresses the implementation of the analysis of bulk observations of the polarimetric Doppler radar observations (WP 1, observation) as well as the improvements of the ice part of the SPECS model (WP 2, modeling).

Contribution of WP1-Observations

Based on the ACCEPT dataset, Myagkov et al. (2016a) developed a technique to derive the shape and orientation of cloud hydrometeors. Application of this technique to the peak signal in the observed cloud-radar Doppler spectra yield a statistic about the relationship between shape of ice crystals and temperature in stratiform mixed-phase clouds (Myagkov et al., 2016b). Due to the limitation to the main peak in the Doppler spectrum, the current state of the shape and orientation retrieval does not allow for the retrieval of hydrometor ratios in the observed cloud volume. It is thus goal of the observations part of SPOCC to extend the bulk technique towards the full Doppler spectrum in order to enable the classification of several types of hydrometeors present in the same observation volume. In a first step, the shape and orientation retrieval of Myagkov et al. (2016a) was reproduced by means of an independent implementation (see Fig. 1), which will be the basis for the future incorporation of the full Doppler spectrum into the shape and orientation retrieval.

Another step of the first project year was to familiarize with the observational datasets of the project which are the prerequisite for the shape and orientation retrieval. Figure 2 shows range-height-indicator (RHI) scans of differential reflectivity and co-cross correlation coefficient as observed at Ka-band during ACCEPT (Fig. 2a) and with the scanning experimental C-Band precipitation radar of the German Meteorological Service (DWD) in Hohenpeißenberg (see Fig. 2b) from which shape and orientation of hydrometeors can be retrieved. The gained experience in application of the shape- and orientation retrieval will in a sub-project be applied to observations of the weather radar network of DWD.
In the course of the project, the obtained results will be the basis for next step to evaluate parameterizations of cloud microphysical processes in collaboration with the modeling part of SPOCC.

Contribution of WP2-Modeling

The first goal of the contribution to SPOCC was an extension of the ice part of the Spectral Microphysics (SPECS) model (Simmel et al., 2017). So far, we have tested the spectral-bin microphysical methodology called AMPS (Advanced Microphysics Prediction System; Hashino et al., 2006) to implement the habit and shape prediction of ice particles in SPECS. Until now, significant efforts have been underway to improve microphysical processes in AMPS, such as the schemes for immersion freezing and habit prediction. Despite these efforts, it is still challenging using modeling alone to resolve such complexity of microphysical processes due the large number of associated parameterizations and assumptions. In particular, the ice habit prediction system in AMPS is sensitive to the 3-D Eulerian advection scheme, such as COSMO.

The steps are as follows. First of all, AMPS was coupled with a simple 1-D dynamic core KiD (Kinematic Driver for microphysics Intercomparison; Shipway and Hill, 2012). By doing so, we were enabled to compare the scheme with other microphysics schemes (i.e., Morrison 2-moment) (see Fig. 3). In the next step, we will evaluate the simulation of a case study of a mixed-phase cloud system against co-located observational data from the ACCEPT campaign. In the course of the work, AMPS will be coupled with the German weather prediction system COSMO (Consortium for Small-scale Modeling; Baldauf et al., 2011) model. Also, we will use the radar forward operator CR-SIM (Cloud Resolving Model Radar Simulator) to translate the dataset of simulation output into radar variables. Therefore, we will directly compare the hydrometeor properties as obtained from the ground-based observations and from the modeling datasets.

References

• Baldauf, M., Seifert, A., Förstner, J., Majewski, D., Raschendorfer, M., & Reinhardt, T., 2011: Operational Convective-Scale Numerical Weather Prediction with the COSMO Model: Description and Sensitivities. Monthly Weather Review, 139(12), 3887–3905.
  
• Hashino, T., & Tripoli, G. J., 2006: The Spectral Ice Habit Prediction System (SHIPS). Part I: Model Description and Simulation of the Vapor Deposition Process. Journal of the Atmospheric Sciences, 64(7), 2210–2237.
  
• Myagkov, P. Seifert, M. Bauer-Pfundstein, and U. Wandinger, 2016a: Cloud radar with hybrid mode towards estimation of shape and orientation of ice crystals, Atmos. Meas. Tech., 9, 469–489.
  
• Myagkov, A., Seifert, P., Wandinger, U., Bühl, J., and Engelmann, R., 2016b: Relationship between temperature and apparent shape of pristine ice crystals derived from polarimetric cloud radar observations during the ACCEPT campaign]], Atmos. Meas. Tech., 9, 3739–3754,https://amt.copernicus.org/articles/9/3739/2016/.
• Shipway, B. J., & Hill, A. A., 2012: Diagnosis of systematic differences between multiple parametrizations of warm rain microphysics using a kinematic framework. Quarterly Journal of the Royal Meteorological Society, 138(669), 2196–2211.
  
• Simmel, M., Bühl, J., Ansmann, A., and Tegen, I., 2015: Ice phase in altocumulus clouds over Leipzig: remote sensing observations and detailed modeling]], Atmos. Chem. Phys., 15, 10453–10470, https://acp.copernicus.org/articles/15/10453/2015/.