Short Courses
We are pleased to offer all of the short courses completely free of charge to all conference participants
Short courses
Short course venue details will be confirmed and shared with all registered participants as soon as final attendance numbers are tallied.
Lecture topics
- Basic definitions of polarimetric radar variables and their physical meaning. Comparison between S, C, and X bands.
- Data quality and measurement errors. Calibration, attenuation correction (S, C, and X bands).
- Polarimetric “fingerprints” of different microphysical processes in clouds and precipitation.
- Radar echo classification. Detection of hail and determination of its size. Tornado detection. Classification of winter precipitation. Comparison between S, C, and X bands.
- Quantitative precipitation estimation. Advanced polarimetric algorithms for rain and snow measurements. The difference between QPE algorithms at S, C, and X bands.
- Polarimetric microphysical retrievals in rain and ice. Radar polarimetry and cloud modeling.
Ground clutter filtering is an important and necessary step for quality control of weather radar data. Clutter filtering typically has been done in the Doppler spectral domain via use of the Discrete Fourier Transform (DFT) to obtain the power spectrum. However, before applying the DFT, the I (in-phase) and Q (quadrature) time series are multiplied by a window function, such as the von Hann or Blackman, which suppresses clutter leakage but also significantly attenuates the time series power. The clutter signal is eliminated from the Doppler spectrum by setting spectral coefficients to zero where there is clutter signal. Recently, it has been shown that a regression clutter filter is a viable alternative (Hubbert et al. 2021, 2025, JTECH). The regression filter is based on the observation that in the time domain, the clutter signal varies slowly, whereas a typical weather signal varies more rapidly. Thus, a least-squares polynomial fit, of appropriate order, is used that follows the trend of the clutter signal. The polynomial fit is then subtracted from the I and Q time series thereby suppressing the clutter signal. The regression filter rejects clutter as effectively as the spectral technique, or better, but has the distinct advantage that no attenuating window function is necessary. As a result, the standard error of estimates of the radar variables are, in general, improved by about 25% to 50% over a comparable spectral domain filter. The regression filter is seemingly straight forward in concept, but the Devil is in the details!
An operational regression based clutter filter includes 1) a robust, low error polynomial fit routine, 2) an algorithm to specify the required polynomial order, 3) an interpolation routine, across the zero velocity gap created by the filter, that reduces the bias of the reflectivity and velocity estimates and 3) real time clutter identification. The theory for the regression filter is developed in context of the more traditional filters such as FIR, IIR and spectral filters. The regression filter is then compared to a spectral based filter via simulations and application to experimental radar data. All Devils will be revealed!
More recently, it has been shown how a high-pass regression clutter filter can be transformed into a band-stop filter centered on an arbitrary frequency (velocity) (Hubbert et al. 2025, 2026, JTECH). This novel filter is developed and then applied to the suppression of the strong trip signal in SZ phase coding used in NEXRADs.
A critical aspect of regression filtering is to use a set of discrete orthogonal polynomials, which by their nature, greatly reduce the round-off error, especially evident in higher order fits. Participants of this class will be given a subroutine that executes the discrete polynomial fit in MATLAB, C++, and FORTRAN.
An operational regression based clutter filter includes 1) a robust, low error polynomial fit routine, 2) an algorithm to specify the required polynomial order, 3) an interpolation routine, across the zero velocity gap created by the filter, that reduces the bias of the reflectivity and velocity estimates and 3) real time clutter identification. The theory for the regression filter is developed in context of the more traditional filters such as FIR, IIR and spectral filters. The regression filter is then compared to a spectral based filter via simulations and application to experimental radar data. All Devils will be revealed!
More recently, it has been shown how a high-pass regression clutter filter can be transformed into a band-stop filter centered on an arbitrary frequency (velocity) (Hubbert et al. 2025, 2026, JTECH). This novel filter is developed and then applied to the suppression of the strong trip signal in SZ phase coding used in NEXRADs.
A critical aspect of regression filtering is to use a set of discrete orthogonal polynomials, which by their nature, greatly reduce the round-off error, especially evident in higher order fits. Participants of this class will be given a subroutine that executes the discrete polynomial fit in MATLAB, C++, and FORTRAN.
This short course provides a hands-on introduction to open-source weather radar processing, guiding participants through the complete
workflow from raw radar data to analysis-ready precipitation products. Building on recent developments in the open radar science ecosystem, the course emphasises interoperable and reproducible workflows based on tools such as xradar, Py-ART, and wradlib. Participants will work with real-world datasets to perform quality control, apply essential corrections, and generate gridded radar products within a modern, xarray-based framework. In addition, the course highlights how established systems such as BALTRAD and LROSE can be integrated into these workflows, bridging research and operational environments. Particular emphasis is placed on transparency and adaptability of processing methods, enabling participants not only to apply existing tools but also to understand and modify key algorithmic components. The course forms the first part of a two-day training sequence and provides the processed datasets required for subsequent precipitation nowcasting applications.
workflow from raw radar data to analysis-ready precipitation products. Building on recent developments in the open radar science ecosystem, the course emphasises interoperable and reproducible workflows based on tools such as xradar, Py-ART, and wradlib. Participants will work with real-world datasets to perform quality control, apply essential corrections, and generate gridded radar products within a modern, xarray-based framework. In addition, the course highlights how established systems such as BALTRAD and LROSE can be integrated into these workflows, bridging research and operational environments. Particular emphasis is placed on transparency and adaptability of processing methods, enabling participants not only to apply existing tools but also to understand and modify key algorithmic components. The course forms the first part of a two-day training sequence and provides the processed datasets required for subsequent precipitation nowcasting applications.
TBD
TBD
These short courses are organized with the help of our institutional partners, the Hydrometeorological Service of Serbia (RHSS) and the University of Belgrade Faculty of Physics Department of Meteorology
9:00 AM
RHSS
Kneza Višeslava 66
Kneza Višeslava 66
Dept. of Meteorology
Dobračina 16
Dobračina 16
Online short courses
Access to online short courses is through your Virtual conference section for registered participants.
Take me there All text, photo, and video materials belong to their owners and are for demonstration purposes only. Please do not use them in commercial projects.
