Integrating IHACRES with a data-driven model to investigate the possibility of improving monthly flow estimates
Estimating the outflow of basins is a critical step in surface water resources planning and management, especially in basins that lack reliable long-term observed data of streamflow. Hydrological models, which can simulate the process of rainfall-runoff, can be used to obtain reliable estimates of s...
Gespeichert in:
Veröffentlicht in: | Water science & technology. Water supply 2022-01, Vol.22 (1), p.360-371 |
---|---|
Hauptverfasser: | , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | Estimating the outflow of basins is a critical step in surface water resources planning and management, especially in basins that lack reliable long-term observed data of streamflow. Hydrological models, which can simulate the process of rainfall-runoff, can be used to obtain reliable estimates of streamflow from precipitation data and the physical characteristics of basins. The focus of the present study was to estimate the outflow of 19 sub-basins located in Guilan Province, northern Iran. To achieve this, hybrid models were developed by integrating the IHACRES (identification of unit hydrograph and component flows from rainfall, evapotranspiration, and streamflow) hydrological model with the intelligent-based GMDH (group method of data handling) model. The IHACRES model was calibrated using monthly ground-based precipitation and temperature data as well as satellite-based precipitation data. The lowest and highest Nash-Sutcliffe coefficient (NS) for the IHACRES models were, respectively, 0.14 and 0.68 in the calibration phase and 0.11 and 0.73 in the validation phase. It was also observed that using satellite-based precipitation data reduces NS by 10–75% in the 19 sub-basins under study. After calibrating and validating the IHACRES models, the hybrid models were developed by integrating IHACRES and GMDH models. The lowest and highest NS for the hybrid models were, respectively, 0.23 and 0.81 in the calibration phase and 0.11 and 0.81 in the validation phase. It was observed that, on average, integrating IHACRES and GMDH increases the NS by 44.1% in the calibration phase and 37.0% in the validation phase in comparison with the IHACRES model. According to the NS, the hybrid model had ‘acceptable’ performance in six sub-basins in which the IHCRES model had ‘unacceptable’ performance. It was observed that integrating the IHACRES model with a data-driven model (the GMDH model) can generally improve the simulation results in all sub-basins under study. |
---|---|
ISSN: | 1606-9749 1607-0798 |
DOI: | 10.2166/ws.2021.267 |