Evapotranspiration Prediction Using ARIMA, ANN and Hybrid Models for Optimum Water Use in Agriculture: A Case Study of Keiskammahoek Irrigation Scheme, Eastern Cape, South Africa

Mbulelo Phesa, Yali Woyessa, Nkanyiso Mbatha

2022

Abstract

Evapotranspiration is the main limitation for irrigation development in developing countries and semi-arid regions. Proper prediction of this variable is key for proper planning and positively contribute to daily management of irrigation schemes. This study used 18 years (2001-2018) of remotely sensed data extracted at Keiskammahoek Irrigation Scheme, Eastern Cape province of South Africa, a province that has been declared drought disaster region forcing many irrigation schemes in this region to close some irrigated sections in order to deal with reduced dam levels. This study used three prediction models, namely Auto-Regressive Integrated Moving Average (ARIMA), Artificial Neural Networks (ANN), and Hybrid (ARIMA-ANN) to predict ET for optimal water use in this irrigation scheme. The prediction models were evaluated using four model performance statistics, namely Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE), Mean Absolute Error (MAE), and the Pearson’s correlation of coefficient (R). The results show that the hybrid (ARIMA-ANN) model outperformed both the ARIMA and ANN consecutively with less values of the statistical performance evaluation showing RMSE = 33.80, MAE = 27.02, MAPE = 17.31, and R = 0.94 compared to higher values of ARIMA and ANN. In general, these prediction results show the dominance of the Hybrid (ARIMA-ANN) model over ARIMA and ANN. These results will assist water managers at Keiskammahoek Irrigation Scheme to plan effectively.

Paper Citation

in Harvard Style

Phesa M., Woyessa Y. and Mbatha N. (2022). Evapotranspiration Prediction Using ARIMA, ANN and Hybrid Models for Optimum Water Use in Agriculture: A Case Study of Keiskammahoek Irrigation Scheme, Eastern Cape, South Africa. In Proceedings of the 3rd International Symposium on Water, Ecology and Environment - Volume 1: ISWEE; ISBN 978-989-758-639-2, SciTePress, pages 276-286. DOI: 10.5220/0012009000003536

in Bibtex Style

@conference{iswee22,
author={Mbulelo Phesa and Yali Woyessa and Nkanyiso Mbatha},
title={Evapotranspiration Prediction Using ARIMA, ANN and Hybrid Models for Optimum Water Use in Agriculture: A Case Study of Keiskammahoek Irrigation Scheme, Eastern Cape, South Africa},
booktitle={Proceedings of the 3rd International Symposium on Water, Ecology and Environment - Volume 1: ISWEE},
year={2022},
pages={276-286},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012009000003536},
isbn={978-989-758-639-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 3rd International Symposium on Water, Ecology and Environment - Volume 1: ISWEE
TI - Evapotranspiration Prediction Using ARIMA, ANN and Hybrid Models for Optimum Water Use in Agriculture: A Case Study of Keiskammahoek Irrigation Scheme, Eastern Cape, South Africa
SN - 978-989-758-639-2
AU - Phesa M.
AU - Woyessa Y.
AU - Mbatha N.
PY - 2022
SP - 276
EP - 286
DO - 10.5220/0012009000003536
PB - SciTePress