Predictive Modeling of Hourly Air Temperature Based on Atmospheric Conditions of Karak in Jordan

In this paper, the focus is mainly on building machine learning (ML) models for AAT forecasting every hour over Karak City in Jordan.The dataset consisted of comprehensive meteorological readings, which were subject to heavy preprocessing in order to establish data integrity essential for building strong ML models.The investigation involved a number of ML models Support Vector Regression (SVR) with RBF Kernel, Decision Tree Regressor (DTR), Ridge Regressor (RR) & Lasso Regressors (LSR) and Linear Regression (LR) because each was found to have unique strengths in capturing the intricate dynamics of temperature behavior.Excellent accuracy of the models, mainly SVR with RBF Kernel and relevance for better forecasting of weather in a region with peculiar difficulties to data-based modeling were shown by it.Our research not only confirmed the capability of different ML methodologies in regional temperature forecasting but also provided an important reference for planners and stakeholders concerned with environmental planning and management.The study provides a better understanding of the regional climate adaptation approaches, vide its case location in Karak City only whereas support local data analysis is necessary to address global climate variability.The results have important implications for the improvement of decision-making in agriculture, disaster management and sustainability schemes especially under changing climatic conditions.