Hybrid COOT–ANN: a novel optimization algorithm for prediction of daily reference evapotranspiration in Australia

Abstract

The present study evaluates the capability of a novel optimization method in modeling daily reference evapotranspiration (ET0), a critical issue in water resource management. A hybrid predictive model based on the ANN Algorithm that is embedded within the COOT method (COOT bird natural life model- Artificial Neural Network (COOT-ANN)) is developed and evaluated for its suitability for the prediction of daily ET0 at seven meteorological stations in different states of Australia. Accordingly, a daily statistical period of 12 years (01-01-2010 to 31-12-2021) for climatic data of maximum temperature, minimum temperature, and ET0 were collected. The results are evaluated using six performance criteria metrics: correlation coefficient (R), Root Mean Square Error (RMSE), Nash-Sutcliffe efficiency (NSE), RMSE-observation standard deviation ratio (RSR), Scatter Index (SI), and mean absolute error (MAE) along with the Taylor diagrams. The performance of the COOT-ANN model was compared with those of the conventional ANN model. The results showed that the COOT-ANN hybrid model outperforms the ANN model at all seven stations; and so this study provides an innovative method for prediction in agricultural and water resources studies.

Shannon entropy of performance metrics to choose the best novel hybrid algorithm to predict groundwater level (case study: Tabriz plain, Iran)

“…The COOT-ANN hybrid approach stops when an expression of mathematical functions obtains a certain or satisfactory fit between COOT and ANN or when the number of iterations reaches its maximum. COOT optimization enables the model to reach its maximum capabilities (Mirzania et al 2022;Mohammadi et al 2020). Parameters of ANN, COOT-ANN, and HBA-ANN models are presented in Table 4.…”