View Item 
        •   Utrecht University Student Theses Repository Home
        • UU Theses Repository
        • Theses
        • View Item
        •   Utrecht University Student Theses Repository Home
        • UU Theses Repository
        • Theses
        • View Item
        JavaScript is disabled for your browser. Some features of this site may not work without it.

        Browse

        All of UU Student Theses RepositoryBy Issue DateAuthorsTitlesSubjectsThis CollectionBy Issue DateAuthorsTitlesSubjects

        Nowcasting rainfall in the Netherlands with the focus on extreme summer precipitation events

        Thumbnail
        View/Open
        Karolina Mazurek 7789831.pdf (1.492Mb)
        Publication date
        2022
        Author
        Mazurek, Karolina
        Metadata
        Show full item record
        Summary
        Extreme summer precipitation often caused by convection is a phenomenon that can lead to flooding, but at the same time, it is challenging for forecasting methods. Among others, deep learning tools are being used to tackle the problem. The study concerns implementation of DeepMind’s Deep Generative Model of Rainfall (DGMR) to data about rainfall in the Netherlands gathered in Nationale Regenradar. DGMR is a nowcasting method which allows to forecast precipitation within the lead time of 90 minutes, based on an input of data referring to 20 minutes. In the research, data for summer, extreme precipitation events was used as an input of the model. Thirteen such events were chosen in consultation with a meteorologist from Koninklijk Nederlands Meteorologisch Instituut (KNMI). The study’s results show that the DGMR proves applicable to the NRR data and this analysis can be used as a proof of concept for further research. Two research questions were addressed: model’s performance in nowcasting precipitation of convective or partly convective type, and, change of nowcast’s accuracy with increasing lead time. The results of the study approve that convective rainfall is more difficult for the model to nowcast than a mixed type one. Additionally, as the lead time was increasing, a drop in nowcast accuracy was observed.
        URI
        https://studenttheses.uu.nl/handle/20.500.12932/42826
        Collections
        • Theses
        Utrecht university logo