Runoff Prediction Using Artificial Neural Network …
41
5 Conclusions
Rainfall-runoff modeling has been carried out using the conceptual and practical
method, the SCS-CN method accounting the spatial variation of land use, soil texture
throughout the region, and heterogeneous rainfall distribution in the GIS environment. The computed runoff for the year 2000 is highest, which is also found from
the available literature discussed in the result section. The summed runoff of individual runoff depth for each Thiessen cell shows a higher magnitude compared to
the weighted curve number approach applied on all the Thiessen cells. This insight
needs further more study in terms of applicability of curve number with observed
data in a distributed scenario rather than applying as a cumulative approach.
The results from the SCS-CN method have been used in modeling the ANN as
there is no observed data in the gauging site. In the ANN method, both the FFBP
and CFBP model reflect a good correlation between the rainfall-runoff data set.
Among all three models (SCS-CN model, FFBP model, and CFBP model), CFBP
model shows a very satisfying performance in terms of rainfall pattern in its peaks.
Though with the help of a smaller number of variables, the ANN model can predict
runoff in a shorter time, and it is not always the right decision to say the ANN is
better than an existing conceptual model in the field of hydrology.
Comparison of the present results with the other popular models viz HEC-HMS
model, SWAT model, and VIC model will lead the study more pronounced and
significant in this field of research.
References
1. Duan Q, Gupta V (1992) Effective and efficient global optimization. Water Resour 28:1015–
1031
2. Bronstert A (2004) Rainfall-runoff modelling for assessing impacts of climate and land-use
change. Hydrol Process 18:567–570
3. Arnell NW, Gosling SN (2013) The impacts of climate change on river flow regimes at the
global scale. J Hydrol 486:351–364
4. Pisarenko VF, Lyubushin AA, Bolgov MV, Rukavishnikova TA, Kanyu S, Kanevskii MF,
Savel’eva EA, Dem’yanov VV, Zalyapin IV (2005) Statistical methods for river runoff
prediction. Water Resour 32:115–26
5. Patil S, Patil S, Valunjkar S (2012) Study of different rainfall-runoff forecasting algorithms for
better water consumption, vol 051
6. Adamowski J, Prasher SO (2012) Comparison of machine learning methods for runoff
forecasting in mountainous watersheds with limited data. J Water L Dev 17:89–97
7. Ki¸ si Ö (2007) Streamflow forecasting using different artificial neural network algorithms. J
Hydrol Eng 12:532–539
8. Phukoetphim P, Shamseldin AY, Melville BW (2014) Knowledge extraction from artificial
neural networks for rainfall-runoff model combination systems. J Hydrol Eng 19:1422–1429
9. Noori N, Kalin L (2016) Coupling SWAT and ANN models for enhanced daily streamflow
prediction. J Hydrol 533:141–151
10. Elsafi SH (2014) Artificial neural networks (ANNs) for flood forecasting at Dongola Station
in the River Nile. Sudan Alexandria Eng. J 53:655–662
41
5 Conclusions
Rainfall-runoff modeling has been carried out using the conceptual and practical
method, the SCS-CN method accounting the spatial variation of land use, soil texture
throughout the region, and heterogeneous rainfall distribution in the GIS environment. The computed runoff for the year 2000 is highest, which is also found from
the available literature discussed in the result section. The summed runoff of individual runoff depth for each Thiessen cell shows a higher magnitude compared to
the weighted curve number approach applied on all the Thiessen cells. This insight
needs further more study in terms of applicability of curve number with observed
data in a distributed scenario rather than applying as a cumulative approach.
The results from the SCS-CN method have been used in modeling the ANN as
there is no observed data in the gauging site. In the ANN method, both the FFBP
and CFBP model reflect a good correlation between the rainfall-runoff data set.
Among all three models (SCS-CN model, FFBP model, and CFBP model), CFBP
model shows a very satisfying performance in terms of rainfall pattern in its peaks.
Though with the help of a smaller number of variables, the ANN model can predict
runoff in a shorter time, and it is not always the right decision to say the ANN is
better than an existing conceptual model in the field of hydrology.
Comparison of the present results with the other popular models viz HEC-HMS
model, SWAT model, and VIC model will lead the study more pronounced and
significant in this field of research.
References
1. Duan Q, Gupta V (1992) Effective and efficient global optimization. Water Resour 28:1015–
1031
2. Bronstert A (2004) Rainfall-runoff modelling for assessing impacts of climate and land-use
change. Hydrol Process 18:567–570
3. Arnell NW, Gosling SN (2013) The impacts of climate change on river flow regimes at the
global scale. J Hydrol 486:351–364
4. Pisarenko VF, Lyubushin AA, Bolgov MV, Rukavishnikova TA, Kanyu S, Kanevskii MF,
Savel’eva EA, Dem’yanov VV, Zalyapin IV (2005) Statistical methods for river runoff
prediction. Water Resour 32:115–26
5. Patil S, Patil S, Valunjkar S (2012) Study of different rainfall-runoff forecasting algorithms for
better water consumption, vol 051
6. Adamowski J, Prasher SO (2012) Comparison of machine learning methods for runoff
forecasting in mountainous watersheds with limited data. J Water L Dev 17:89–97
7. Ki¸ si Ö (2007) Streamflow forecasting using different artificial neural network algorithms. J
Hydrol Eng 12:532–539
8. Phukoetphim P, Shamseldin AY, Melville BW (2014) Knowledge extraction from artificial
neural networks for rainfall-runoff model combination systems. J Hydrol Eng 19:1422–1429
9. Noori N, Kalin L (2016) Coupling SWAT and ANN models for enhanced daily streamflow
prediction. J Hydrol 533:141–151
10. Elsafi SH (2014) Artificial neural networks (ANNs) for flood forecasting at Dongola Station
in the River Nile. Sudan Alexandria Eng. J 53:655–662