Comparison of forecasting performance of nonlinear models of hydrological time series

Abstract Mean monthly flows of the Tatry alpine mountain region in Slovakia are predominantly fed by snowmelt in the spring and convective precipitation in the summer. Therefore their regime properties exhibit clear seasonal patterns. Positive deviations from these trends have substantially different features than the negative ones. This provides intuitive justification for the application of nonlinear two-regime models for modelling and forecasting of these time series. Nonlinear time series structures often have lead to good fitting performances, however these do not guarantee an equally good forecasting performance. In this paper therefore the forecasting performance of several nonlinear time series models is compared with respect to their capabilities of forecasting monthly and seasonal flows in the Tatry region. A new type of regime-switching models is also proposed and tested. The best predictive performance was achieved for a new model subclass involving aggregation operators.

[1]  K. Chan,et al.  Percentage Points of Likelihood Ratio Tests for Threshold Autoregression , 1991 .

[2]  Michael P. Clements,et al.  A Monte Carlo Study of the Forecasting Performance of Empirical Setar Models , 1999 .

[3]  Ruey S. Tsay,et al.  On the Ergodicity of Tar(1) Processes , 1991 .

[4]  C. Gouriéroux.  ARCH Models and Financial Applications , 1997 .

[5]  T. Teräsvirta.  Specification, Estimation, and Evaluation of Smooth Transition Autoregressive Models , 1994 .

[6]  Qiwei Yao,et al.  Quantifying the influence of initial values on nonlinear prediction , 1994 .

[7]  C. Granger,et al.  Forecasting from non-linear models in practice , 1994 .

[8]  B. Hansen,et al.  Inference in TAR Models , 1997 .

[9]  Luca Ridolfi,et al.  Multivariate nonlinear prediction of river flows , 2001 .

[10]  Philip Hans Franses,et al.  Time Series Models for Business and Economic Forecasting , 1998 .

[11]  A. Ullah,et al.  Handbook of Applied Economic Statistics , 2000 .

[12]  Philip Hans Franses,et al.  Non-Linear Time Series Models in Empirical Finance , 2000 .

[13]  Roberto Revelli,et al.  A comparison of nonlinear flood forecasting methods , 2003 .

[14]  R. Engle.  Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation , 1982 .

[15]  Ruey S. Tsay,et al.  Testing and Modeling Threshold Autoregressive Processes , 1989 .

[16]  Ezio Todini,et al.  Role and treatment of uncertainty in real‐time flood forecasting , 2004 .

[17]  Michael P. Clements,et al.  The Performance of Alternative Forecasting Methods for SETAR Models , 1997 .

[18]  Alessandra Amendola,et al.  Modelli non lineari e previsioni in tempo reale , 2000 .

[19]  B. Hansen.  Sample Splitting and Threshold Estimation , 2000 .

[20]  N. T. Kottegoda,et al.  Stochastic Modelling of Riverflow Time Series , 1977 .

[21]  Jose D. Salas,et al.  Retrospective analysis and forecasting of streamflows using a shifting level model , 2004 .

[22]  Paulin Coulibaly,et al.  Nonstationary hydrological time series forecasting using nonlinear dynamic methods , 2005 .

[23]  Mark M. Meerschaert,et al.  Identification of periodic autoregressive moving average models and their application to the modeling of river flows , 2006 .

[24]  James D. Hamilton.  Time Series Analysis , 1994 .

[25]  T. McMahon,et al.  Stochastic generation of annual, monthly and daily climate data: A review , 2001 .

[26]  H. Tong,et al.  ON ESTIMATING THRESHOLDS IN AUTOREGRESSIVE MODELS , 1986 .

[27]  Shlomo Havlin,et al.  A stochastic model of river discharge fluctuations , 2003 .

[28]  J. Cunderlik,et al.  Impact of climate change on the seasonal distribution of runoff in mountainous basins in Slovakia , 1998 .

[29]  David W. Bacon,et al.  Estimating the transition between two intersecting straight lines , 1971 .

[30]  V. Singh,et al.  Computer Models of Watershed Hydrology , 1995 .

[31]  J. Scott Armstrong,et al.  Principles of forecasting : a handbook for researchers and practitioners , 2001 .

[32]  D. Tjøstheim.  Non-linear Time Series: A Selective Review* , 1994 .

[33]  Jan G. De Gooijer,et al.  On forecasting SETAR processes , 1998 .

[34]  Jozef Komorník,et al.  Application of aggregation operators in regime-switching models for exchange rates , 2005, EUSFLAT Conf..

[35]  C C Reyes-Aldasoro,et al.  A hybrid model based on dynamic programming, neural networks, and surrogate value for inventory optimisation applications , 1999, J. Oper. Res. Soc..

[36]  C. Granger,et al.  Modelling Nonlinear Economic Relationships , 1995 .

[37]  A. Turgeon,et al.  Fuzzy Learning Decomposition for the Scheduling of Hydroelectric Power Systems , 1996 .

[38]  Timo Teräsvirta,et al.  Testing linearity against smooth transition autoregressive models , 1988 .

[39]  Gwilym M. Jenkins,et al.  Time series analysis, forecasting and control , 1972 .

[40]  Kung-Sik Chan,et al.  On Likelihood Ratio Tests for Threshold Autoregression , 1990 .

[41]  Stefania Tamea,et al.  Probabilistic nonlinear prediction of river flows , 2005 .

[42]  Michael P. Clements,et al.  On SETAR non- linearity and forecasting , 1999 .

[43]  Timo Teräsvirta,et al.  Modelling Economic Relationships with Smooth Transition Regressions , 1996 .

[44]  Pavla Pekarova,et al.  Long‐term discharge prediction for the Turnu Severin station (the Danube) using a linear autoregressive model , 2006 .

[45]  W. Enders.  Applied Econometric Time Series , 1994 .

[46]  Bryan W. Brown,et al.  Predictors in Dynamic Nonlinear Models: Large-Sample Behavior , 1989, Econometric Theory.

[47]  K. Chan,et al.  Testing for threshold autoregression , 1990 .

[48]  H. Tong,et al.  Threshold Autoregression, Limit Cycles and Cyclical Data , 1980 .

[49]  B. Akintug,et al.  A Markov switching model for annual hydrologic time series , 2005 .

[50]  Wen Wang,et al.  Testing and modelling autoregressive conditional heteroskedasticity of streamflow processes , 2005 .

[51]  T. Bollerslev,et al.  Generalized autoregressive conditional heteroskedasticity , 1986 .

[52]  Roman Krzysztofowicz,et al.  The case for probabilistic forecasting in hydrology , 2001 .

[53]  Hung Man Tong,et al.  Threshold models in non-linear time series analysis. Lecture notes in statistics, No.21 , 1983 .

[54]  Alessandra Amendola,et al.  Predictor distribution and forecast accuracy of threshold models , 2004 .

[55]  Richard S.J. Tol.  Autoregressive conditional heteroscedasticity in daily temperature measurements , 1996 .

[56]  J. Petruccelli,et al.  A threshold AR(1) model , 1984, Journal of Applied Probability.

[57]  Heung Wong,et al.  Change-point analysis of hydrological time series using grey relational method , 2006 .

[58]  F. Diebold,et al.  Comparing Predictive Accuracy , 1994, Business Cycles.

[59]  J. Salas,et al.  Conceptual Basis of Seasonal Streamflow Time Series Models , 1992 .

[60]  A. Ramachandra Rao,et al.  Linearity analysis on stationary segments of hydrologic time series , 2003 .

[61]  D. K. Srivastava,et al.  Application of ANN for Reservoir Inflow Prediction and Operation , 1999 .

[62]  Henghsiu Tsai,et al.  Testing for nonlinearity with partially observed time series , 2000 .