{"id":{"repo_id":"vilnius","oai_identifier":"oai:vu.lt:elaba:308335823"},"canonical_url":"https://search.ndltd.org/etd/vilnius/oai:vu.lt:elaba:308335823","repository":{"repo_id":"vilnius","name":"Vilnius University","base_url":"https://epublications.vu.lt/oai"},"display":{"title":"Labai mažos trukmės kritulių intensyvumo prognozių modelio kūrimas pagal orų radarų duomenis Lietuvoje /","abstract":"There are currently two operational weather radars in Lithuania: Laukuva (since 2010) and Trakų Vokė (since 2014). Thanks to their high spatial (1°x300 m) and temporal (5 min) resolutions, their observations may be used for precipitation nowcasting (up to 90 minutes). Standard numerical weather models often struggle within this timeframe due to their lower spatial resolution, sparse forecast time steps, and extensive computation time. Various nowcasting methods are currently being developed and deployed by national weather service organizations worldwide, providing vital information for businesses and society. In this paper we developed a stochastic model based on the S-PROG nowcasting model, replacing Fourier decomposition with Dual-Tree Complex Wavelet Transform. We calibrated the Z-R relationship, and optimized the model parameters. The developed model lags behind S-PROG in terms of categorical metrics but outperforms it in continuous metrics. According to the FSS score, the useful forecast duration for rain detection is 60 minutes; for light rain, 30 minutes; for medium-intensity rain, 15 minutes. Because our model incorporates a stochastic component, the output frames are less blurred and better maintain the fragmented rain cell structures. Moreover, it is less susceptible to radar noise and composition artifacts. However, under complex synoptic conditions – when multiple rain cells develop at different rates and are in different phases of their life cycles – our model fails to capture these dynamics; S-PROG suffers from the same limitation. We developed a method for determining the optimal number of input frames for optical flow estimation and autoregressive model parametrization. Since the majority of non-machine learning models operate using optical flow and/or autoregression, this method may also be applied to other models.","abstract_html":"There are currently two operational weather radars in Lithuania: Laukuva (since 2010) and Trakų Vokė (since 2014). Thanks to their high spatial (1°x300 m) and temporal (5 min) resolutions, their observations may be used for precipitation nowcasting (up to 90 minutes). Standard numerical weather models often struggle within this timeframe due to their lower spatial resolution, sparse forecast time steps, and extensive computation time. Various nowcasting methods are currently being developed and deployed by national weather service organizations worldwide, providing vital information for businesses and society. In this paper we developed a stochastic model based on the S-PROG nowcasting model, replacing Fourier decomposition with Dual-Tree Complex Wavelet Transform. We calibrated the Z-R relationship, and optimized the model parameters. The developed model lags behind S-PROG in terms of categorical metrics but outperforms it in continuous metrics. According to the FSS score, the useful forecast duration for rain detection is 60 minutes; for light rain, 30 minutes; for medium-intensity rain, 15 minutes. Because our model incorporates a stochastic component, the output frames are less blurred and better maintain the fragmented rain cell structures. Moreover, it is less susceptible to radar noise and composition artifacts. However, under complex synoptic conditions – when multiple rain cells develop at different rates and are in different phases of their life cycles – our model fails to capture these dynamics; S-PROG suffers from the same limitation. We developed a method for determining the optimal number of input frames for optical flow estimation and autoregressive model parametrization. Since the majority of non-machine learning models operate using optical flow and/or autoregression, this method may also be applied to other models.","abstract_has_math":false,"creators":["Stanaitis, Rimantas,"],"institution":"Institutional Repository of Vilnius University","degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2026,"date_issued":"2026","date_published":"2026","updated_at":"2026-11-19T10:12:55Z","subjects":[],"languages":["lit"],"rights":["info:eu-repo/semantics/openAccess"],"rights_urls":[],"identifier_entries":[],"person_identifier_entries":[]},"links":{"outbound_url":"https://repository.vu.lt/VU:ELABAETD308335823&prefLang=en_US","outbound_label":"Repository record","outbound_source":"dc:identifier","record_url":"https://repository.vu.lt/VU:ELABAETD308335823&prefLang=en_US","record_label":"Repository record","record_kind":"record","record_source":"dc:identifier","download_url":null,"download_label":null,"download_source":null,"access_status":"open","embargo_until":null,"access_note":null},"source_record":{"url":"https://epublications.vu.lt/oai?verb=GetRecord&metadataPrefix=oai_dc&identifier=oai%3Avu.lt%3Aelaba%3A308335823","prefix":"oai_dc"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:creator","label":"Author","values":["Stanaitis, Rimantas,"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2026"]},{"key":"dc:publisher","label":"Institution","values":["Institutional Repository of Vilnius University"]},{"key":"dc:relation","label":"Dc Relation","values":["https://epublications.vu.lt/object/elaba:308335823/308335823.pdf"]},{"key":"dc:type","label":"Dc Type","values":["info:eu-repo/semantics/bachelorThesis"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["lit"]},{"key":"dc:rights","label":"Dc Rights","values":["info:eu-repo/semantics/openAccess"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["https://repository.vu.lt/VU:ELABAETD308335823&prefLang=en_US"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["There are currently two operational weather radars in Lithuania: Laukuva (since 2010) and Trakų Vokė (since 2014). Thanks to their high spatial (1°x300 m) and temporal (5 min) resolutions, their observations may be used for precipitation nowcasting (up to 90 minutes). Standard numerical weather models often struggle within this timeframe due to their lower spatial resolution, sparse forecast time steps, and extensive computation time. Various nowcasting methods are currently being developed and deployed by national weather service organizations worldwide, providing vital information for businesses and society. In this paper we developed a stochastic model based on the S-PROG nowcasting model, replacing Fourier decomposition with Dual-Tree Complex Wavelet Transform. We calibrated the Z-R relationship, and optimized the model parameters. The developed model lags behind S-PROG in terms of categorical metrics but outperforms it in continuous metrics. According to the FSS score, the useful forecast duration for rain detection is 60 minutes; for light rain, 30 minutes; for medium-intensity rain, 15 minutes. Because our model incorporates a stochastic component, the output frames are less blurred and better maintain the fragmented rain cell structures. Moreover, it is less susceptible to radar noise and composition artifacts. However, under complex synoptic conditions – when multiple rain cells develop at different rates and are in different phases of their life cycles – our model fails to capture these dynamics; S-PROG suffers from the same limitation. We developed a method for determining the optimal number of input frames for optical flow estimation and autoregressive model parametrization. Since the majority of non-machine learning models operate using optical flow and/or autoregression, this method may also be applied to other models."]},{"key":"dc:format","label":"Dc Format","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Labai mažos trukmės kritulių intensyvumo prognozių modelio kūrimas pagal orų radarų duomenis Lietuvoje /","Development of a weather radar-based precipitation nowcasting model in lithuania."]}]}],"canonical_facts":{"dc:creator":["Stanaitis, Rimantas,"],"dc:date":["2026"],"dc:description":["There are currently two operational weather radars in Lithuania: Laukuva (since 2010) and Trakų Vokė (since 2014). Thanks to their high spatial (1°x300 m) and temporal (5 min) resolutions, their observations may be used for precipitation nowcasting (up to 90 minutes). Standard numerical weather models often struggle within this timeframe due to their lower spatial resolution, sparse forecast time steps, and extensive computation time. Various nowcasting methods are currently being developed and deployed by national weather service organizations worldwide, providing vital information for businesses and society. In this paper we developed a stochastic model based on the S-PROG nowcasting model, replacing Fourier decomposition with Dual-Tree Complex Wavelet Transform. We calibrated the Z-R relationship, and optimized the model parameters. The developed model lags behind S-PROG in terms of categorical metrics but outperforms it in continuous metrics. According to the FSS score, the useful forecast duration for rain detection is 60 minutes; for light rain, 30 minutes; for medium-intensity rain, 15 minutes. Because our model incorporates a stochastic component, the output frames are less blurred and better maintain the fragmented rain cell structures. Moreover, it is less susceptible to radar noise and composition artifacts. However, under complex synoptic conditions – when multiple rain cells develop at different rates and are in different phases of their life cycles – our model fails to capture these dynamics; S-PROG suffers from the same limitation. We developed a method for determining the optimal number of input frames for optical flow estimation and autoregressive model parametrization. Since the majority of non-machine learning models operate using optical flow and/or autoregression, this method may also be applied to other models."],"dc:format":["application/pdf"],"dc:identifier":["https://repository.vu.lt/VU:ELABAETD308335823&prefLang=en_US"],"dc:language":["lit"],"dc:publisher":["Institutional Repository of Vilnius University"],"dc:relation":["https://epublications.vu.lt/object/elaba:308335823/308335823.pdf"],"dc:rights":["info:eu-repo/semantics/openAccess"],"dc:title":["Labai mažos trukmės kritulių intensyvumo prognozių modelio kūrimas pagal orų radarų duomenis Lietuvoje /","Development of a weather radar-based precipitation nowcasting model in lithuania."],"dc:type":["info:eu-repo/semantics/bachelorThesis"]},"updated_at":"2026-11-19T10:12:55Z"}