The climate crisis has resulted in an increase in severe surface air temperature and precipitation events over the past few decades, leading to heatwaves, floods, droughts, and adverse effects on irrigation systems. To prevent the severe effects of catastrophic events, reliable, up-to-date, and thorough surface air temperature and precipitation data are imperative. With the proposed hybrid prediction mechanism, the study combines artificial neural networks (ANNs) with starfish optimisation algorithms (SFOAs) to improve surface air temperature and precipitation occurrences. The developed ANN-SFOA model is evaluated over datasets from two German locations: Düsseldorf and München. The analysis addressed surface air temperature and precipitation by using seven input features: cloud cover, wind speed, wind gust, humidity, pressure, global radiation, and sunshine. A comparative analysis is performed on five models: ANN, ANN-GWO, ANN-RSA, ANN-COA, and ANN-PO, utilising statistical benchmarks including root mean squared error (RMSE), scatter index (SI), mean absolute error (MAE), and determination coefficient (R2) to emphasize the effectiveness of ANN-SFOA. Further, the experimental findings indicate that ANN-SFOA surpasses other models in accurately predicting surface air temperature and precipitation for the locations of Düsseldorf and München. The ANN-SFOA demonstrated exceptional performance, with high accuracy in predicting surface air temperature and precipitation for Düsseldorf (R2 = 0.9797 ± 0.0102, R2 = 0.9626 ± 0.0385, MBE = 0.3465 ± 0.1454, MBE = 0.0178 ± 0.0126) and Müenchen (R2 = 0.9636 ± 0.0226, R2 = 0.9688 ± 0.0248, MBE = 0.2814 ± 0.1896, MBE = −0.0142 ± 0.0129), respectively. Eventually, this research could potentially assist ecological management and address the climate problem, particularly in understanding and regulating extreme surface air temperatures and precipitation.
Rani et al. (Wed,) studied this question.