Articles | Volume 17, issue 6
Hydrol. Earth Syst. Sci., 17, 2297–2303, 2013
https://doi.org/10.5194/hess-17-2297-2013

Special issue: Statistical methods for hydrological applications

Hydrol. Earth Syst. Sci., 17, 2297–2303, 2013
https://doi.org/10.5194/hess-17-2297-2013

Research article 25 Jun 2013

Research article | 25 Jun 2013

Stochastic modeling of Lake Van water level time series with jumps and multiple trends

H. Aksoy1, N. E. Unal1, E. Eris2, and M. I. Yuce3 H. Aksoy et al.
  • 1Istanbul Technical University, Istanbul, Turkey
  • 2Ege University, Izmir, Turkey
  • 3University of Gaziantep, Gaziantep, Turkey

Abstract. In the 1990s, water level in the closed-basin Lake Van located in the Eastern Anatolia, Turkey, has risen up about 2 m. Analysis of the hydrometeorological data shows that change in the water level is related to the water budget of the lake. In this study, stochastic models are proposed for simulating monthly water level data. Two models considering mono- and multiple-trend time series are developed. The models are derived after removal of trend and periodicity in the dataset. Trend observed in the lake water level time series is fitted by mono- and multiple-trend lines. In the so-called mono-trend model, the time series is treated as a whole under the hypothesis that the lake water level has an increasing trend. In the second model (so-called multiple-trend), the time series is divided into a number of segments to each a linear trend can be fitted separately. Application on the lake water level data shows that four segments, each fitted with a trend line, are meaningful. Both the mono- and multiple-trend models are used for simulation of synthetic lake water level time series under the hypothesis that the observed mono- and multiple-trend structure of the lake water level persist during the simulation period. The multiple-trend model is found better for planning the future infrastructural projects in surrounding areas of the lake as it generates higher maxima for the simulated lake water level.