Application of Conceptual Rainfall-Runoff Model METQ for Simulation of Daily Runoff and Water Level: The case of the Lake Burtnieks Watershed

Abstract

In this study a conceptual rainfall-runoff METQ model-the latest version METQ2007B-DOPT-was applied to simulate the daily runoff and water level of the Lake Burtnieks watershed from 1990 to 1999. The model structure and parameters were basically the same as in the METQ98, with some additional improvements and semi-automatical calibration performance. Model calibration was done for four rivers and one lake gauging station. The results of calibration showed a good correlation between the measured and simulated daily discharges. The Nash-Sutcliffe efficiency R 2 varied from 0.90 to 0.58 and correlation coefficient r from 0.95 to 0.83. The highest values of R 2 = 0.90 and r = 0.95 were obtained for the River Salaca and the lowest R 2 = 0.53 and r = 0.83 for Lake Burtnieks. We observed some relationships between the model parameter values and physiographic characteristic of the sub-catchments.

Climate change impacts on river runoff in Latvia

“…These models predict that impacts of climate change on river runoff can be quite pronounced in areas where snow accumulation and melt presently dominate in the hydrological regime (Berg ström et al 2001), as is the case in Latvia. Further, the semi-automatic or automatic calibration method is aimed at successfully finding the single bestfitting para meter values and reducing parameter uncertainty in the hydrological model, which could arise by manual calibration (Apsīte et al 2008, Moradkhani & Sorooshian 2009). However, the observations in rainfall-runoff modelling consist of the measurement of the input and output fluxes of the hydrological system and even the storage in the system.…”

“…Another explanation of the above-mentioned differences might be a broad flood plain and a high percentage of wetlands in the Seda River basin. These characteristics determine the specific hydrological regime, which differs from that of the other studied rivers; therefore, it is difficult to simulate the rainfall-runoff processes without additional riverbed measurements (Apsīte et al 2008). Usually, a shorter calibration period is selected in studies where the conceptual rainfall-runoff model is applied.…”

“…During the last 20 yr, several versions of mathematical models of hydrological processes have been developed in Latvia: METUL (Krams & Zīverts 1993), METQ96 (Zīverts & Jauja 1996), METQ98 (Zīverts & Jauja 1999), METQ2005 and METQ2006, and the latest version of the METQ2007BDOPT with semi-automatic calibration performance (Apsīte et al 2008). The METQ model is successfully applied to both small and relatively large catchments in Latvia, namely Vienziemīte Brook (drainage area 5.92 km 2 ) and the Daugava River (drainage area 81 000 km 2 ), respectively (Zīverts & Jauja 1999).…”

“…modifications of the unit hydrograph approach (Zīverts & Jauja 1999). If there is a lake or a reservoir in the catchment and it considerably influences the hydrological regime of the river, the hydraulic runoff routing is required (Apsīte et al 2008). It is assumed by Zīverts & Jauja (1999) that water evaporates only from the upper layer of land, i.e.…”

“…The influence of the radiation part to the total value of the degree-day ratio is higher in open areas than in the forest. A more detailed description of the model METQ is presented elsewhere (Zīverts & Jauja 1999, Apsīte et al 2008). …”