Scholarly record

ENHANCING SIMULATION OF LOW AND HIGH FLOWS USING A TWO-REGIME SWITCHING RAINFALL-RUNOFF MODEL

Ing. Patrik Sleziak, Univ. Doz. Dr. Juraj Parajka

Abstract

Traditional one-regime rainfall-runoff (r-r) models are usually able to adequate simulate only flows in some interval (low or high flows). Proper simulation of both type of flows has a great importance in many practical applications, for example flood protection, risk analysis, etc. For this purpose more sophisticated modeling tools are needed. The aim of the paper is the improvement of simulations of low and high flows using a two-regime switching r-r model. This model uses a unique concept of separate simulation of both type of flows. The structure of the model is based on fixing all model parameters except of three parameters (k0, k1, k2) which are responsible for runoff. These parameters are calibrated separately by two optimization functions. The first optimization function (for low flows) compares only values below a 5th percentile of the observed flows. The second optimization function (for high flows) compares only values above a 95th percentile of the observed flows. By simulating flows separately a satisfactorily results have been achieved. In the first part of the study, the two-regime switching r-r model is described. Subsequently, this model is calibrated in period 2001 пїЅ 2010 and validated in period 1981 пїЅ 1990 for 213 catchments in Austria. The model performance is evaluated by the Nash-Sutcliffe coefficient (NSE) and the volume error (VE). The results showed that the model performed better in calibration period (the values of NSE was higher). In validation period the model also achieved satisfactorily results. Next, the results are compared with the results obtained by traditional one-regime r-r model. It showed that the two regime switching r-r model allow as to adequate describe the representation of low flows (some values are higher, some similar in comparison with one-regime model). To the contrary, the two-regime model showed a slightly poorer ability to simulate high flows (the lower values of NSE/the values of VE are comparable). Some of the comparisons are discussed in conclusion.

Publication Impact Profile

• Citations
• Scopus - Citation Indexes: 2

• Captures
• Mendeley - Readers: 5

PLUMX see details

DOI resolver: doi.org/10.5593/sgem2017/31/s12.029

Publication details

Back to publication list