Contenu

Collaboration in radar meteorology with the EPFL

Le contenu de cette page est actuellement disponible uniquement en anglais.
Début du projet 01.01.2015
Fin du projet 31.12.2024
Thèmes Système de mesure et de prévision
Région Suisse
Statut Projects actuels

Collaboration in radar meteorology between EPFL (Environmental Remote Sensing Lab) and MeteoSwiss (Radar, Satellite and Nowcasting). The renewed weather radar network is a fundamental pillar of the national meteorological observing system. It is an indispensable basis for the generation of alerts and the management of natural hazards related to heavy rainfall, severe thunderstorms and hail, which are top priority tasks of MeteoSwiss. The collaboration in radar meteorology with the École Polytechnique Fédérale de Lausanne (EPFL) aims at a thorough exploitation of the large potential of the novel radar technology of the renewed weather radar network. Key topics are hydrometeors, microphysics, hail, winter precipitation, quantitative precipitation estimation, dual-polarisation Doppler signatures of severe convection. This is combined with data from other observing systems, advanced integration in the forecasting system “COSMO” and novel data processing technology such as machine learning. At the national level, this collaboration is an example of a fruitful mutual interaction between academy (under EAER - The Federal Department of Economic Affairs, Education and Research) and government (under FDHA - The Federal Department of Home Affairs) and represents a modern set-up for capacity building and professional training with young researchers and senior scientists who closely interact on a daily basis remotely and face-to-face. It unifies the academic context and strive for scientific excellence of EPFL with the user-oriented context of MeteoSwiss and its long tradition in applied research and the supervision of PhD students. EPFL has many students eager to learn and work hard, while MeteoSwiss can offer a unique set of observational data, large libraries of algorithms and many research topics with high practical significance. A key element is also the scientific and professional training of PhD students and young Post-Doc scientists in a field that is relevant for MeteoSwiss.

People, activities and temporal evolution of the project

In July 2014, MeteoSwiss has signed a contract with EPFL for the collaboration in applied research and innovation in the field of radar meteorology. It joins the excellence of the Environmental Remote Sensing Lab (LTE) of EPFL and MeteoSwiss in the field of radar meteorology and aims to challenge state-of-the-art radar algorithms, to exploit new radar technologies and to develop new remote sensing applications to meet future user needs in the Alpine region.

Researchers from EPFL and MeteoSwiss and, depending on the topic, by other MeteoSwiss partners have been concentrated on different research activities, to mention just a few of them:
   

Hydrometeor classification and winter precipitation

A focus was given to a hydrometeor classification and solid precipitation (see [2], [5], [11], [12] as well as on winter precipitation. Another focus was given on the vertical structure of dual-polarisation moments in the Alps in view of an improved precipitation estimation in complex orography regions. (Ph.D. Thesis N. 7203, 2019). Furthermore, at the end of 2016, some researchers have been concentrated on combining different research results in order to improve winter precipitation estimation.
   

Automatic alert system

In July 2015 an urgent need came up: MeteoSwiss had to replace the precipitation alert system dated back to 2003 with a novel operational heavy precipitation alert system that integrate radar-gauge precipitation fields with forecast systems. The collaboration with EPFL was an ideally suited context for this task. The new automatic alert system has been successfully implemented (NowPAL, see [1], [14], [16]).
    

Novel radar quantitative precipitation estimation

Another research activity has been started with focus on the link between weather radar and natural hazards, in particular the development and evaluation of novel radar quantitative precipitation estimation. Specifically, the potential of using retrieved fields of specific differential propagation phase delay (KDP) for rain rate estimation in Switzerland has been thoroughly investigated. The study was part of a broader project aiming to improve the Quantitative Precipitation Estimation retrieval by complementing it with polarimetric data, instead of relying solely on the reflectivity. A huge dataset of Drop Size Distribution measurements was used to derive a Rain Rate-to-KDP model specifically valid for Switzerland, which takes into account the type of air mass and the incidence angle and provides error estimates.

During the second part of the project the very difficult task of QPE in complex orography has been tackled with a new twist, by training a random forest (RF) regression to learn a QPE model directly from a large database comprising four years of combined gauge and polarimetric radar observations. The algorithm is carefully fine-tuned by optimizing its hyper-parameters and then compared with MeteoSwiss' current operational non-polarimetric QPE method. The evaluation shows that the RF algorithm is able to significantly reduce the error and the bias of the predicted precipitation intensities, especially for large and solid/mixed precipitation (see the recently submitted paper [C]).

Impacts and benefits

The collaboration in applied research and innovation in radar meteorology with EPFL initiated in 2014 and set-up in 2015 has generated a number of important results and benefits with practical significance for MeteoSwiss during this first almost-6-year period (February 2015- October 2020).

Both partners, EPFL and MeteoSwiss, to fulfil their mission and legal obligations, are very active at an international level. Both have many key partners for specific topics. Networking includes collaboration with many research institutes around the world, within weather services, academia and industry. Worth mentioning is also active participation in OPERA (Eumetnet), in the previous WMO Inter-Program Expert Team on Operational Weather Radars and in the newly founded Joint Expert Team on Operational Weather Radars under the WMO Infrastructure Commission as well as the international ISO-WMO initiative to establish a standard for weather radars. The collaboration with EPFL in radar meteorology is a win-win solution that combines the excellence of both partners, exploits the potential of the new radar network and makes sure we are prepared to respond to the user demands of tomorrow.

Quality assurance of new algorithms and data is certified by the publication of research in scientific journals, the corresponding peer-review process, and presentation of the results to the international community at conferences and workshops. The review by independent and anonymous reviewers is an effective way to be exposed to open criticism by experts in the field, and, once approved by the reviewers and editors of the journal; the paper earns credibility and builds the basis for follow-up research in the community.

The collaboration brings significant benefit at reasonable costs both at the technical and human level. Since it provides knowledge and scientific basis for continuous improvement and extension of radar data, products and services, it represents a unique opportunity in terms of knowhow. From the point of view of innovative operational radar products and services, particularly impressive are the first three items in the following list:

  • development of a novel semi-supervised technique based on dual-polarisation radar observations for hydrometeor classification (including melting hail, ice hail, high density graupel), which runs operationally on Rad4Alp central computing server (see ref. [2] and [12] in the Section “Publications and quality assurance” below);
  • development of a CombiPrecip-INCA-COSMO based heavy precipitation alert system (NowPAL), also running operationally (see [1] regarding the algorithm NowPAL and [14] regarding a preliminary study that shows the way for setting optimized alert thresholds);
  • access to data and observations from field experiments: huge dataset of Drop Size Distribution measurements in the Swiss and French Alps and on the Swiss plateau;, unique observational set of pictures of snowflakes and ice crystals (see [6]), data from cloud radar;
  • a detailed comparison of space-borne (GPM) and ground-based (Rad4Alp) radar precipitation estimates over Switzerland (see [7] and [8]) and a machine learning comparison between precipitation fields at a European scale between the OPERA (Eumetnet) weather radar composite and the Eumetsat geostationary satellite, (see [14]);
  • access to instruments for special purposes field experiments: high-resolution radar scans of the Payerne (PARADISO) and Valais field campaigns with a mobile X-band radar for improved low-altitude coverage at short range (see [16]); exploring potential of innovative calibration techniques (see [4]) and observations (see [9]); an innovative and adaptive thunderstorm measurement concept based on high resolution scans of a dedicated X-band radar optimized for low-altitude coverage; the X-band radar performs adaptive three-dimensional scans of one automatically selected cell within its surveillance range, namely the most intense storm cell previously identified by the operational C-band weather radar network [20];
  • first studies of the four-dimensional structure of dual-polarisation moments over the Alps (see [16] and [21]);
  • development of a COSMO forward operator for dual-polarisation radar data, ready to explore the value of dual-polarisation radar data in COSMO (see [11]);
  • experiments with new radar snow algorithms and a comparison of radar snowfall estimates with ground measurements in the region of Davos (see [13], together with SLF in addition to EPFL);
  • unprecedented insight into the microphysical structure of falling snowflakes, which forms the basis for proper interpretation of dual-polarisation radar signatures of solid precipitation (see [6]);
  • Abstract book of the 11th European Conference on Radar in Meteorology and Hydrology (ERAD2020), can be freely download at www.erad2020.ch
       

Publications related to Project 4 on Natural Hazards and QPE

[A] Wolfensberger D., Gabella, M., Boscacci M., Berne A., Germann U., 2018: Potential use of specific differential propagation phase delay for the retrieval of rain rates in strong convection over Switzerland, 11th International Workshop on Precipitation in Urban Areas, December 5-7, Pontresina, Switzerland.

[B] Gabella M., Panziera L., Sideris I., Boscacci M., Wolfensberger D., Clementi L., Germann U., 2018: Twelve years of operational real-time precipitation estimation in the Alps and an example of QPE during an extreme event : Lausanne, 11.6.2018., 11th International Workshop on Precipitation in Urban Areas, December 5-7, Pontresina, Switzerland.

[C] Wolfensberger D., Gabella, M., Boscacci M., Berne A., Germann U., 2020: RainForest: A random forest algorithm for quantitative precipitation estimation over Swizerland, submitted to Atmos. Meas. Tech., https://amt.copernicus.org/preprints/amt-2020-284/
    

European Conference on Radar in Meteorology and Hydrology -- ERAD2020 Book of Abstracts

www.erad2020.ch or www.erad2022.ch

    

Peer-reviewed papers on international journals and quality assurance

[22] Feldmann M., Curtis J., Boscacci M., Leuenberger D., Gabella M., Wolfensberger D., Germann U., and A. Berne, 2020: R2D2 – A Region-based Recursive Doppler Dealiasing algorithm for operational weather radar, J. Atmos. Ocean. Tech., accepted for publication.

[21] Van den Heuvel F., Foresti L., Gabella M., Germann U., and A. Berne, 2019: Learning about the vertical structure of radar reflectivity using hydrometeor classes and neural networks in the Swiss Alps, Atmos. Meas. Tech., 13, 2481-2500. https://doi.org/10.5194/amt-13-2481-2020

[20] Grazioli J., Leuenberger A., Peyraud L., Figueras J., Gabella M., Hering A., Germann U., 2019: An adaptive Thunderstorm Measurement Concept using C-band and X-band Radar Data, IEEE Geoscience and Remote Sensing Letters, 16, 1673-1677.  https://doi.org/10.1109/LGRS.2019.2909970

[19] Nerini D., Foresti L., Leuenberger D., Robert S., Germann U., 2019: A Reduced-Space Ensemble Kalman Filter Approach for Flow-Dependent Integration of Radar Extrapolation Nowcasts and NWP Precipitation Ensembles, Mon. Wea. Rev., 147, 987-1006. https://doi.org/10.1175/MWR-D-18-0258.1

[18] Gabella M., 2018: On the Use of Bright Scatterers for Monitoring Doppler, Dual-Polarization Weather Radars, Remote Sensing, 10, 14 pages. https://doi.org/10.3390/rs10071007

[17] Foresti L., Sideris I., Panziera L., Nerini D., and U. Germann: 2018: A 10-year radar-based analysis of orographic precipitation growth and decay patterns over the Swiss Alpine region, Q. J. Royal Met. Soc., 144, 2277-2301. https://doi.org/10.1002/qj.3364

[16] Van den Heuvel F., Gabella M., Germann U., and A. Berne, 2018: Characterization of the melting layer variability in an Alpine valley based on polarimetric X-band radar scans, Atmos. Meas. Tech., 11, 5181-5198. https://doi.org/10.5194/amt-11-5181-2018

[15] Panziera L., Gabella M., Germann U., Martius O., 2018: A 12-year radar-based climatology of daily/sub-daily extreme precipitation over the Swiss Alps, Int. J. Climatol., 7-21. https://doi.org/10.1002/joc.5528

[14] Beusch L., Foresti L., Gabella M., Hamann U., 2018: Satellite-Based Rainfall Retrieval: from Generalized Linear Models to Artificial Neural Networks, Remote Sensing, 10, 24 pages. https://doi.org/10.3390/rs10060939

[13] Gerber F., N. Besic, V. Sharma, M. Daniels, M. Gabella, R. Mott, U. Germann, A. Berne, and M. Lehning, 2018: Spatial variability in snow precipitation and accumulation in COSMO–WRF simulations and radar estimations over complex terrain, The Cryosphere, 12, pp. 3137-3160. https://doi.org/10.5194/tc-12-3137-2018

[12] Besic N., Gehring J., Praz C., Figueras J., Grazioli J., Gabella M., Germann U., and A. Berne, 2018: Unraveling hydrometeor mixtures in polarimetric radar measurements, Atmos. Meas. Tech., 11, 4847-4866. https://doi.org/10.5194/amt-11-4847-2018

[11] Wolfensberger D. and A. Berne, 2018: From model to radar variables: a new forward polarimetric radar operator for COSMO, Atmos. Meas. Tech., 11, 3883-3916. https://doi.org/10.5194/amt-11-3883-2018

[10] Gabella M. and A. Leuenberger, 2017: Dual-Polarization Observations of Slowly Varying Solar Emissions from a Mobile X-Band Radar, Sensors, 17, 1-15. https://doi.org/10.3390/s17051185

[9] Gabella M., Huuskonen A., Sartori M., Holleman I., Boscacci M., Germann U., 2017: Evaluating the Solar Slowly Varying Component at C-Band Using Dual- and Single-Polarization Weather Radars in Europe, Advances in Meteorology, vol. 2017, 8 pages. https://doi.org/10.1155/2017/4971765

[8] Gabella M., Speirs P., Hamann U., Berne A., and U. Germann, 2017: Measurement of precipitation in the Alps using dual-polarization C-band ground-based radars, the GPM spaceborne Ku-band radar and rain gauges, Remote Sensing, 9, 19 pages. https://doi.org/10.3390/rs9111147

[7] Speirs P., Gabella M., and A. Berne, 2017: A comparison between the GPM dual-frequency precipitation radar and ground-based radar precipitation rate estimates in the Swiss Alps and Plateau, J. of Hydrometeorol., 18, 1247-1269. https://doi.org/10.1175/JHM-D-16-0085.1

[6] Praz C., Roulet Y.A., and A. Berne, 2017: Solid hydrometeor classification and riming degree estimation from pictures collected with a Multi-Angle Snowflake Camera, Atmos. Meas. Tech., 10, 1335-1357. https://doi.org/10.5194/amt-10-1335-2017

[5] Nerini D., Besic N., Sideris I., Germann U., and L. Foresti: 2017: A non-stationary stochastic ensemble generator for radar rainfall fields based on the short-space Fourier transform, Hydrol. Earth Syst. Sci., 21, 2777-2797. https://doi.org/10.5194/hess-21-2777-2017

[4] Gabella M., Sartori M., Boscacci M., Germann U., 2016: Calibration accuracy of the dual-polarization receivers of the C-band Swiss weather radar network, Atmosphere, 7, 10 pages. https://doi.org/10.3390/atmos7060076

[3] Wolfensberger D., Scipion D. and A. Berne, 2016: Detection and characterization of the melting layer based on polarimetric radar scans. Quarterly Journal of the Royal Meteorological Society, 142, 108-124. https://doi.org/10.1002/qj.2672

[2] Besic N., Figueras J., Grazioli J., Gabella M., Germann U., and A. Berne, 2016: Hydrometeor classification through statistical clustering of polarimetric radar measurements: a semi-supervised approach, Atmos. Meas. Tech., 9, 4425-4445. https://doi.org/10.5194/amt-9-4425-2016

[1] Panziera L., Gabella M., Zanini S., Hering A., Germann U., and A. Berne, A., 2016: A radar-based regional extreme rainfall analysis to derive the thresholds for a novel automatic alert system in Switzerland, Hydrol. Earth Syst. Sci., 20, 2317-2332. https://doi.org/10.5194/hess-20-2317-2016

Informations complémentaires