One of the water resources modeling requirements is sufficient knowledge of long-term series of meteorological and hydrological parameters. In this study the nearest neighbor resampling method presented by Lall and Sharma was developed. In the developed model, the KNN regression was used for time series forecasting instead of local polynomial used in the developed algorithm by Prairie. In this case problems caused by polynomial degree estimation were solved. This caused that multivariate time series simulation became feasible. To evaluate the performance of the developed model recorded data in the Kawtar hydrometer station located on one of the main branch of Mahabad River and Mahabad synoptic station were used. The results from time series simulation showed that the developed model is able to keep important statistical characteristics of historic series. In addition, it could solve the classic nearest neighbor resampling method problem in order to produce values not seen in the historical record. Also, this model showed that it could simulate the multivariate hydrologic series.
Sharifazari, S., & Araghinejad, S. (2013). Develop a non-parametric model to simulate monthly hydrological data. Water and Irrigation Management, 3(1), 83-95. doi: 10.22059/jwim.2013.35739
MLA
Salman Sharifazari; Shahab Araghinejad. "Develop a non-parametric model to simulate monthly hydrological data", Water and Irrigation Management, 3, 1, 2013, 83-95. doi: 10.22059/jwim.2013.35739
HARVARD
Sharifazari, S., Araghinejad, S. (2013). 'Develop a non-parametric model to simulate monthly hydrological data', Water and Irrigation Management, 3(1), pp. 83-95. doi: 10.22059/jwim.2013.35739
VANCOUVER
Sharifazari, S., Araghinejad, S. Develop a non-parametric model to simulate monthly hydrological data. Water and Irrigation Management, 2013; 3(1): 83-95. doi: 10.22059/jwim.2013.35739
)