Drought Forecasting using Wavelet-GMDH Model with Standardized Precipitation Index

Abstract

This paper proposes Wavelet-Group Methods of Data Handling (W-GMDH) model to explore its ability of drought forecasting. The W-GMDH model was developed by combining Discrete Wavelet Transform (DWT) and GMDH model using the Standardized Precipitation Index (SPI) drought data for forecasting to assess the effectiveness of the new (W-GMDH) model. These methods were used on four SPI data sets (SPI3, SPI6, SPI9 and SPI12). To achieve this, a 624 month of SPI data from January 1956 to December 2008 was used and divided into two parts (80% for training and 20% for testing). The results of the W-GMDH model were then compared with the conventional GMDH model using Root Mean Square Error (RMSE), Mean Average Error (MAE) and coefficient of correlation as the performance evaluation measures. Both results of the proposed W-GMDH model and the GMDH showed very clearly that the propose method can achieve the best forecasting performance in terms of accuracy for each of the SPI data series. The key role played by the DWT is to smooth the analysis of SPI data obtained after the wavelet decomposition which is also used to decompose the SPI data into different number of component series to minimize the forecasting error. In all the results computed, the proposed model has a minimum error indicating its superiority over the GMDH model. This indicates that W-GMDH model’s performance has outweighed that of the conventional GMDH model in SPI drought forecasting. The research contributes to the discovering of viable forecasting of drought and demonstrates that the established model is good and appropriate for drought. In all the analysis W-GMDH model has the minimum error. The overall results showed that SPI12 has the minimum error among all the SPI data considered.

Review of the limitations and potential empirical improvements of the parametric group method of data handling for rainfall modelling

“…The fundamental idea behind WT is their ability to perform signal analysis at different resolutions, which precipitates their most important feature, that of multiresolution decomposition (Lubis et al 2017 ). Özger et al ( 2012 ) employed the continuous WT (CWT) in their study of drought forecasting, while Alfa et al ( 2019 ) utilised the discrete WT (DWT) in a hybrid with GMDH for their drought forecasting study. Moosavi et al ( 2021 ) employed the DWT as a hybrid with GMDH in their research on groundwater level modelling.…”

Limitations of Parametric Group Method of Data Handling and Empirical Improvements for the Application of Rainfall Modelling

“…employed the continuous WT (CWT) in their study of drought forecasting whilstAlfa et al (2019) utilised the discrete WT (DWT) in a hybrid with GMDH for their drought forecasting study Moosavi et al (2021). …”