Hydrological stream flow modelling using soil and water assessment tool (SWAT) and neural networks (NNs) for the Limkheda watershed, Gujarat, India

Autor: Jaydip J. Makwana, Mukesh K. Tiwari
Rok vydání: 2017
Předmět:
Zdroj: Modeling Earth Systems and Environment. 3:635-645
ISSN: 2363-6211
2363-6203
DOI: 10.1007/s40808-017-0323-y
Popis: Investigation of continuous daily streamflow based on rainfall in arid and semi-arid region is challenging, particularly when climate records are limited, time consuming or unavailable. A calibrated and validated model to simulate hydrological processes will be a great help to the concerned watershed management. In this study the accuracy of the Soil and Water Assessment Tools (SWAT) and Neural Networks (NNs) are compared to perform continuous simulation of runoff in a hilly and agricultural watershed, named Limkheda watershed of Gujarat, India. We used the remote sensing data (SRTM-DEM imagery, soil maps and land use/cover classification from LISS-III imagery, etc), climatic and discharge data are used as primary inputs for SWAT models, whereas only climatic data and discharge data were used for NN model setup. The climatic and observed streamflow data from 2 years (2009–2010) were used for calibration and another 2 years (2011–2012) data were used for model validation. To examine the efficiency of both models five performance indices were applied. In the present study, performance of the NNs model was found better than SWAT model for simulating surface runoff from the watershed based on calibration and validation results. It is found in this study that SWAT model provides a better description of water balance of the watershed, whereas NN models present the surface runoff at the outlet without any explicit consideration of different components of the hydrologic cycle.
Databáze: OpenAIRE
Externí odkaz: