River flow hydrograph simulation in the Western Siberia lowland north for the extreme flood flow prediction based on the HBV-Light model

The purpose of this work is to study the possibility of predicting the flow hydrographs of the Arctic rivers (including maximum discharges of snowmelt and rain floods) and estimating the maximum water discharges of a given probability of exceedance on a basis of the conceptual model HBV-light. The calculations were performed for the rivers Poluy, Nadym, Pur, Taz, etc. In total, the processes of runoff formation from nine basins were modeled. The area of the smallest river basin of the Pravaya Hetta River at Pangody is 1200, and the largest basin of the Taz River at the settlement of Sidorovsk is 100,000 km² in area. Prediction of runoff hydrographs in the HBV-light model consists of the preparation of initial data and the selection of model parameters that allow calculating the daily runoff depth based on daily air temperature and precipitation at the nearest weather stations. In this work, the division of catchments into altitude and landscape zones was not carried out. A standard model structure was chosen without taking into account the specifics of land use in the basin. The parameters were obtained for a whole area of each model catchment, without its detailing. Calibration of model parameters for rivers shows that the relatively simple conceptual model HBV-light reproduces well the intra-annual dynamics of runoff, including water discharge during the flood period. The accuracy of reproducing the hydrographs of the northern rivers runoff is not inferior to the accuracy of more complex deterministic models due to the lack of a sufficiently dense network of hydrometeorological observations and the need to determine the numerous parameters of more complex deterministic models. Manual calibration of the model parameters makes it possible to meaningfully use the HBV-light model for regional modeling of the elements of the water cycle and flow hydrographs. In this case, six individual parameters for each basin and nine general regional parameters are determined, reflecting the similarity of landscape conditions throughout the study area. The individual parameters include: TT, FC, LP, K1, K2, MAXBAS. It is shown that the values of each of the other nine parameters of the model can be set the same for the entire study area, that is, they have a regional character, which makes it possible to calculate runoff hydrographs, including the maximum water discharge of ungauged and poorly studied rivers. In general, the obtained model quality criteria for the period of calibration and validation of the models are quite high, and some models could, under certain conditions, be used to assess the change in river runoff induced by the climate change, as well as in the longer time period. Peak flood due to snowmelt is modeled better than peaks due to rainfall floods. Therefore, for basins with high model performance, a comparison of empirical probability distribution functions shows a good agreement between the observed and modeled values of the maximum flood flow rates over the entire amplitude of their fluctuations.

Keywords

Authors

References