A new evolutionary time series model for streamflow forecasting in boreal lake-river systems

Abstract Genetic programming (GP) is an evolutionary regression method that has received considerable interest to model hydro-environmental phenomena recently. Considering the sparseness of hydro-meteorological stations on northern areas, this study investigates benefits and downfalls univariate streamflow modeling at high latitudes using GP seasonal autoregressive integrated moving average (SARIMA). Furthermore, a new time series model, called GP-SARIMA, introduced enhance forecasting accuracy long-term horizons in lake-river system. The paper includes testing for one-step-ahead forecasts daily mean, weekly monthly mean headwaters Oulujoki River, Finland. results showed combination correlogram mutual information (AMI) analysis might yield selection optimum lags are needed be used as predictors models. With Nash-Sutcliffe efficiency values more than 99%, both SARIMA models exhibited good performance prediction. However, they were not able precisely intramonthly snow water equivalent forecast. proposed ensemble which integrates best with most efficient predictor, may eliminate one-fourth root squared errors standalone GP-SARIMA also up three times improvement based Nash-Sutcliff measure.