Research on nowcasting through dual-polarization weather radar data using deep learning approach is rare but worth exploring. This paper lightens a previous work, the MCT (Multivariate Channel Transformer) model, which leads to the design of the MSF (Multivariate Swin Fusion) model. The commonalities between the two are as follows: on one hand, both fuses several dual-polarization observables including reflectivity (Z), specific differential phase (Kdp ), and differential reflectivity (Zdr ) to more comprehensively consider meteorological particle features; on the other hand, they introduces the attention mechanism to more fully fuse multi-frame, multi-variate, and multi-scale features. In the experimental evaluation, this study first selects observation data from KMLB radar in FL, USA, and uses traditional optical flow method, deep learning TrajGRU method, etc. as controls. The results show that both MCT and MSF perform better than the control, and the 60min forecast scores of both are 8.78/9.31 for RMSE and 0.46/0.18/0.07 for CSI (20/35/45dBZ), and this conclusion is verified by case study. Further, the role of the attention mechanism is verified by ablation experiments.
more »
« less
Physical Analysis of the Impact of Polarization Parameters on Deep-learning Networks for Nowcasting
The task of nowcasting by deep learning using multivariate, rather than just reflectivity, is limited by poor interpretability. The previous experiment designed MCT (Multivariate Channel Transformer), a deep learning model capable of nowcasting with dual-polarization radar data. Four analytical methods are designed to further explore the contribution of polarization parameters: (i) Case studies of different meteorological processes. (ii) A permutation test ranking the significance of each variable. (iii) Visualization of the feature maps obtained by forward propagation of the input data. (iv) Data downscaling of polarimetric radar data. The results show that the polarization parameters serve as a guide to predict the location and shape of strong reflectivity, as well as the energy retention of strong echoes at 40-50 dBZ. The contributions of Zdr and Kdp are more evident in the prediction results after 30 min, and the importance of Kdp exceeds that of Zdr in case of strong convective weather.
more »
« less
- Award ID(s):
- 2239880
- NSF-PAR ID:
- 10515826
- Publisher / Repository:
- IEEE
- Date Published:
- ISBN:
- 979-8-3503-2010-7
- Page Range / eLocation ID:
- 6799 to 6802
- Format(s):
- Medium: X
- Location:
- Pasadena, CA, USA
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
In this study, a polarimetric radar forward model operator was developed for the Weather Research and Forecasting (WRF) model that was based on a scattering algorithm using the T-matrix methodology. Three microphysics schemes—Thompson, Morrison 2-moment, and Milbrandt-Yau 2-moment—were supported in the operator. This radar forward operator used the microphysics, thermodynamic, and wind fields from WRF model forecasts to compute horizontal reflectivity, radial velocity, and polarimetric variables including differential reflectivity (ZDR) and specific differential phase (KDP) for S-band radar. A case study with severe convective storms was used to examine the accuracy of the radar operator. Output from the radar operator was compared to real radar observations from the Weather Surveillance Radar–1988 Doppler (WSR-88D) radar. The results showed that the radar forward operator generated realistic polarimetric signatures. The distribution of polarimetric variables agreed well with the hydrometer properties produced by different microphysics schemes. Similar to the observed polarimetric signatures, radar operator output showed ZDR and KDP columns from low-to-mid troposphere, reflecting the large amount of rain within strong updrafts. The Thompson scheme produced a better simulation for the hail storm with a ZDR hole to indicate the existence of graupel in the low troposphere.more » « less
-
In this study, a polarimetric radar forward model operator was developed for the Weather Research and Forecasting (WRF) model that was based on a scattering algorithm using the T-matrix methodology. Three microphysics schemes—Thompson, Morrison 2-moment, and Milbrandt-Yau 2-moment—were supported in the operator. This radar forward operator used the microphysics, thermodynamic, and wind fields from WRF model forecasts to compute horizontal reflectivity, radial velocity, and polarimetric variables including differential reflectivity (ZDR) and specific differential phase (KDP) for S-band radar. A case study with severe convective storms was used to examine the accuracy of the radar operator. Output from the radar operator was compared to real radar observations from the Weather Surveillance Radar–1988 Doppler (WSR-88D) radar. The results showed that the radar forward operator generated realistic polarimetric signatures. The distribution of polarimetric variables agreed well with the hydrometer properties produced by different microphysics schemes. Similar to the observed polarimetric signatures, radar operator output showed ZDR and KDP columns from low-to-mid troposphere, reflecting the large amount of rain within strong updrafts. The Thompson scheme produced a better simulation for the hail storm with a ZDR hole to indicate the existence of graupel in the low troposphere.more » « less
-
null (Ed.)A multi-radar analysis of the 20 May 2013 Moore, Oklahoma, U.S. supercell is presented using three Weather Surveillance Radars 1988 Doppler (WSR-88Ds) and PX-1000, a rapid-scan, polarimetric, X-band radar, with a focus on the period between 1930 and 2008 UTC, encompassing supercell maturation through rapid tornado intensification. Owing to the 20-s temporal resolution of PX-1000, a detailed radar analysis of the hook echo is performed on (1) the microphysical characteristics through a hydrometeor classification algorithm (HCA)—inter-compared between X- and S-band for performance evaluation—including a hail and debris class and (2) kinematic properties of the low-level mesocyclone (LLM) assessed through ΔVr analyses. Four transient intensifications in ΔVr prior to tornadogenesis are documented and found to be associated with two prevalent internal rear-flank downdraft (RFD) momentum surges, the latter surge coincident with tornadogenesis. The momentum surges are marked by a rapidly advancing reflectivity (ZH) gradient traversing around the LLM, descending reflectivity cores (DRCs), a drop in differential reflectivity (ZDR) due to the advection of smaller drops into the hook echo, a decrease in correlation coefficient (ρhv), and the detection of debris from the HCA. Additionally, volumetric analyses of ZDR and specific differential phase (KDP) signatures show general diffusivity of the ZDR arc even after tornadogenesis in contrast with explosive deepening of the KDP foot downshear of the updraft. Similarly, while the vertical extent of the ZDR and KDP columns decrease leading up to tornadogenesis, the phasing of these signatures are offset after tornadogenesis, with the ZDR column deepening the lagging of KDP.more » « less
-
Abstract. Radar dual-wavelength ratio (DWR) measurements from the Stony Brook RadarObservatory Ka-band scanning polarimetric radar (KASPR, 35 GHz), a W-bandprofiling radar (94 GHz), and a next-generation K-band (24 GHz) micro rainradar (MRRPro) were exploited for ice particle identification using triple-frequency approaches. The results indicated that two of the radarfrequencies (K and Ka band) are not sufficiently separated; thus, thetriple-frequency radar approaches had limited success. On the other hand, ajoint analysis of DWR, mean Doppler velocity (MDV), andpolarimetric radar variables indicated potential in identifying ice particletypes and distinguishing among different ice growth processes and even inrevealing additional microphysical details. We investigated all DWR pairs in conjunction with MDV from the KASPRprofiling measurements and differential reflectivity (ZDR) and specificdifferential phase (KDP) from the KASPR quasi-vertical profiles. TheDWR-versus-MDV diagrams coupled with the polarimetric observables exhibiteddistinct separations of particle populations attributed to different rimedegrees and particle growth processes. In fallstreaks, the 35–94 GHz DWRpair increased with the magnitude of MDV corresponding to the scatteringcalculations for aggregates with lower degrees of riming. The DWR valuesfurther increased at lower altitudes while ZDR slightly decreased,indicating further aggregation. Particle populations with higher rimedegrees had a similar increase in DWR but a 1–1.5 m s−1 largermagnitude of MDV and rapid decreases in KDP and ZDR. The analysisalso depicted the early stage of riming where ZDR increased with theMDV magnitude collocated with small increases in DWR. This approach willimprove quantitative estimations of snow amount and microphysical quantitiessuch as rime mass fraction. The study suggests that triple-frequencymeasurements are not always necessary for in-depth ice microphysical studiesand that dual-frequency polarimetric and Doppler measurements cansuccessfully be used to gain insights into ice hydrometeor microphysics.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1109/IGARSS52108.2023.10282636
