Disputas: Valeriya Filipova

Valeriya Filipova disputerer 11. desember for doktorgraden med avhandlingen “Modeling of extreme floods: a case study of Norway”.


11 Dec

Praktisk informasjon

  • Dato: 11. desember 2018
  • Tid: kl. 10.00 - 16.00
  • Sted: Campus Bø
  • Bedømmelseskomité

    • Første opponent: Professor Renata J. Romanowicz, Institute of Geophysics, Polish Academy of Sciences
    • Andre opponent: Professor Oddbjørn Bruland, NTNU
    • Administrator av komiteen: Førsteamanuensis Mona Sæbø, USN

    Veiledere

    • Professor Harald Klempe, USN (hovedveileder)
    • Seniorforsker Deborah Lawrence 
    • Dr. Inger Hansen-Bauer
    • Dr. Eirik Førland

    Program

    Kl. 11:00 Prøveforelesning med følgende tema:

    “Influence of floods on riverine ecosystems - their positive and negative consequences”

    Kl. 12:15 Disputas

    Prøveforelesning og disputas er åpen for alle interesserte

Graden avlegges ved  Universitetet i Sørøst-Norge (USN), Fakultet for teknologi, naturvitenskap og maritime fag.

Om avhandlingen

With the fall of 2018 and two heavy floods in mind - one from combined snowmelt and rain, and one from intense rain - we see the importance of estimating such floods to avoid damage to infrastructure and houses.  In addition, the estimation of extreme floods is needed in the design and evaluation of structures such as dams, levees and bridges. 

Valeriya Filipova disputas usn. foto.

Although several methods exist and are currently applied to model extreme floods in Norway, they have high uncertainty. In addition, as new data and data analysis techniques are now available, these existing methods need to be updated. 

For example, one of the methods to estimate flood events is to use extreme precipitation sequence as input into the hydrological model PQRUT. The regional equations to estimate the parameters of the model have been developed in the 1980s and are still in use. As more data is now available, this thesis presents a new set of equations which result in a better overall performance of the PQRUT model. 

Another shortcoming of this approach is the need to assign snowmelt and initial conditions for the simulation. These are both important inputs as the flood risk is much higher if the storm event is combined with snowmelt and fully saturated soil. In order to consider a large number of possible combinations of these factors, we consider a method based on Monte Carlo simulations.

In addition to using a hydrological model, the streamflow series can be analysed directly to obtain the return period of interest.  Typically, the flood peak and the flood volume are modelled separately. However, as there is significant correlation between these two values, this can result in inaccurate estimation of the risk. A way to model this correlation is to use copulas but this requires additional parameters. In this case, it is beneficial to use regional information. Our study evaluates whether catchment properties have an effect on the correlation between flood peak and volume and also on the selection of the copulas. 

Our study has now focused on catchments that do not show any non-stationarity. However, due to river regulation, changes in land use and climate change, in many places the flood risk might change in the future. Therefore, more research is needed in this area.