Application of Artificial Neural Network and Climate Indices to Drought Forecasting in South-Central Vietnam
More details
Hide details
Faculty of Water Resource Engineering, Thuyloi University, Hanoi, Vietnam
Department of Engineering Systems and Environment, University of Virginia, Charlottesville, VA USA
Luong Bang Nguyen   

Faculty of Water Resource Engineering, Thuyloi University, Viet Nam
Submission date: 2019-01-16
Final revision date: 2019-02-27
Acceptance date: 2019-03-26
Online publication date: 2019-09-18
Publication date: 2020-01-16
Pol. J. Environ. Stud. 2020;29(2):1293–1303
Widespread negative consequences of droughts related to climate indices in Vietnam have motivated many studies integrating those indices to predict the onset of drought in the region. This study aims to examine the capacity of eight climate Pacific Ocean indices as input variables for forecasting the drought index at 30 stations of south-central Vietnam during the period 1977 to 2014. The standardized precipitation evapotranspiration index (SPEI) was selected as a predicted target drought index at four multiple time scales (3, 6, 9, and 12 months). Input variable selection filters (partial correlation input selection and partial mutual information selection) were used to select the suitable climate indices as input parameters, and an artificial neural network was applied for the drought model. The results showed that partial correlation input selection selected a better optimal input set for the drought model. The west tropical Pacific index (NINOW), east central tropical Pacific index (NINO34), and south oscillation index (SOI) were climate indices that could improve the drought forecasting performances at the given study.