TY - JOUR
T1 - A functional autoregressive model based on exogenous hydrometeorological variables for river flow prediction
AU - Beyaztas, Ufuk
AU - Shang, Hanlin
AU - Yaseen, Zaher
PY - 2021/7
Y1 - 2021/7
N2 - In this research, a functional time series model was introduced to predict future realizations of river flow time series. The proposed model was constructed based on a functional time series’s correlated lags and the essential exogenous climate variables. Rainfall, temperature, and evaporation variables were hypothesized to have substantial functionality in river flow simulation. Because an actual time series model is unspecified and the input variables’ significance for the learning process is unknown in practice, it was employed a variable selection procedure to determine only the significant variables for the model. A nonparametric bootstrap model was also proposed to investigate predictions’ uncertainty and construct pointwise prediction intervals for the river flow curve time series. Historical datasets at three meteorological stations (Mosul, Baghdad, and Kut) located in the semi-arid region, Iraq, were used for model development. The prediction performance of the proposed model was validated against existing functional and traditional time series models. The numerical analyses revealed that the proposed model provides competitive or even better performance than the benchmark models. Also, the incorporated exogenous climate variables have substantially improved the modeling predictability performance. Overall, the proposed model indicated a reliable methodology for modeling river flow within the semi-arid region.
AB - In this research, a functional time series model was introduced to predict future realizations of river flow time series. The proposed model was constructed based on a functional time series’s correlated lags and the essential exogenous climate variables. Rainfall, temperature, and evaporation variables were hypothesized to have substantial functionality in river flow simulation. Because an actual time series model is unspecified and the input variables’ significance for the learning process is unknown in practice, it was employed a variable selection procedure to determine only the significant variables for the model. A nonparametric bootstrap model was also proposed to investigate predictions’ uncertainty and construct pointwise prediction intervals for the river flow curve time series. Historical datasets at three meteorological stations (Mosul, Baghdad, and Kut) located in the semi-arid region, Iraq, were used for model development. The prediction performance of the proposed model was validated against existing functional and traditional time series models. The numerical analyses revealed that the proposed model provides competitive or even better performance than the benchmark models. Also, the incorporated exogenous climate variables have substantially improved the modeling predictability performance. Overall, the proposed model indicated a reliable methodology for modeling river flow within the semi-arid region.
KW - Functional autoregressive
KW - Hydrometeorological variables
KW - River flow prediction
KW - Semi-arid environment
UR - http://www.scopus.com/inward/record.url?scp=85106594383&partnerID=8YFLogxK
U2 - 10.1016/j.jhydrol.2021.126380
DO - 10.1016/j.jhydrol.2021.126380
M3 - Article
VL - 598
SP - 1
EP - 19
JO - Journal of Hydrology
JF - Journal of Hydrology
SN - 0022-1694
M1 - 126380
ER -