Abstract:Evaporation prediction is essential for efficient water resource management and environmental monitoring. Our study introduces a new model named recurrent attention unit (RAU)–gated recurrent unit (GRU)–extreme learning machine model (ELM) to extract temporal features and predict evaporation data. While the ELM model may not extract temporal features from data, the GRU model can extract temporal features from time series data. The GRU–ELM model may not pay attention to important features of time series data. Our study introduces a RAU to address this limitation of the GRU–ELM model. RAU can extract and store important features. Our study also introduces a new optimization algorithm called Gaussian mutation (GM)–Alpine skiing optimization algorithm (ASOA) to improve the exploitation ability of the ASOA for feature selection. The novelties of the current research include the development of the RAU–GRU–ELM (RGE) model for predicting evaporation data and the introduction of a new optimizer for selecting inputs. The GM–ASOA model chose the maximum temperature, minimum temperature, and number of sunny hours as the optimal input scenario. Determining the best input scenario was difficult and complex as there were 2 6 -1 input combinations. GM–ASOA automatically selects the most appropriate input scenario for predicting evaporation. While ELM was unable to extract spatiotemporal patterns, GM–ASOA–RGE was able to extract spatiotemporal patterns and make accurate evaporation predictions. The study results indicated that the GM–ASOA RAU–GRU–ELM (GM–ASOA–RGE) had the best accuracy among other models. The GM–ASOA–RAU–GRU–ELM (GM–ASOA–RGE) decreased training MAE of the ASOA–RGE, RGE, GRU–ELM, RAU–GRU, GRU, and ELM by 4.6%, 8.03%, 20.1%, 54.01%, 64%%, respectively. The testing MAE of the GM–ASOA–RGE and ELM was 0.126 mm and 0.300 mm, respectively. The training NSE of the GM–ASOA–RGE, ASOA–RGE, RGE, GRU–ELM, RAU–GRU, GRU, and ELM model was 0.98, 0.97, 0.95, 0.94, 0.86, 0.82, and 0.78, respectively. The GM–ASOA–RGE model also outperformed benchmark models including the recurrent neural network model (RNN), the multilayer perceptron (MLP) and the radial basis function neural network model (RBFNN). The GM–ASOA–RGE reduced the testing MAE of RNN, MLP, and RFBNN by 36%, 48%, and 51%, respectively. The study results indicate that the RAU–GRU–ELM model is a reliable model for monitoring hydrological parameters.

Accuracy enhancement for monthly evaporation predicting model utilizing evolutionary machine learning methods

Enhancing the performance of data-driven models for monthly reservoir evaporation prediction

Coupling a firefly algorithm with support vector regression to predict evaporation in northern Iran

Prediction of reservoir evaporation considering water temperature and using ANFIS hybridized with metaheuristic algorithms

Pan evaporation prediction using a hybrid multilayer perceptron-firefly algorithm (MLP-FFA) model: case study in North Iran

Advancements in evaporation prediction: introducing the Gated Recurrent Unit–Multi-Kernel Extreme Learning Machine (MKELM)–Gaussian Process Regression (GPR) model

Modeling monthly pan evaporation using wavelet support vector regression and wavelet artificial neural networks in arid and humid climates

SVR-RSM: a Hybrid Heuristic Method for Modeling Monthly Pan Evaporation

Modeling monthly pan evaporation process over the Indian central Himalayas: application of multiple learning artificial intelligence model

Forecasting monthly pan evaporation using hybrid additive regression and data-driven models in a semi-arid environment

Predicting evaporation with optimized artificial neural network using multi-objective salp swarm algorithm

Forecasting pan evaporation with an integrated artificial neural network quantum-behaved particle swarm optimization model: a case study in Talesh, Northern Iran

Utilizing Machine Learning Models with Limited Meteorological Data as Alternatives for the FAO-56PM Model in Estimating Reference Evapotranspiration

Integration of fruit fly and firefly optimization algorithm with support vector regression in estimating daily pan evaporation

Reference Evapotranspiration Prediction Using Hybridized Fuzzy Model with Firefly Algorithm: Regional Case Study in Burkina Faso

Gaussian mutation-alpine skiing optimization algorithm-recurrent attention unit-gated recurrent unit-extreme learning machine model: an advanced predictive model for predicting evaporation

Random Forest and Multilayer Perceptron hybrid models integrated with the genetic algorithm for predicting pan evaporation of target site using a limited set of neighboring reference station data

An approach for flood flow prediction utilizing new hybrids of ANFIS with several optimization techniques: a case study

Solar radiation prediction using improved soft computing models for semi-arid, slightly-arid and humid climates

Enhancing the accuracy of metaheuristic neural networks in predicting underground water levels using meteorological data and remote sensing: A case study of Ardabil Plain, Iran

Monthly evapotranspiration estimation using optimal climatic parameters: efficacy of hybrid support vector regression integrated with whale optimization algorithm