Empowering Precision Forecasting: Self-Supervised LSTM for Hourly Pressure and Temperature Prediction

The most important parts of any flight are landing and takeoff, and an aircraft's takeoff configuration must balance the regulated takeoff weight, runway length, and weather conditions to ensure a safe departure and arrival. In addition to runway length, wind, temperature, pressure, and visibility determine the total allowed takeoff weight and the economic viability of any trip. Thus, any meteorological office involved in flight planning and operation at any airport must accurately assess these factors, known as takeoff data. This research paper suggests multivariate self-supervised LSTM-based models to accurately predict the temperature and pressure (MSLP) of the takeoff data. The suggested prediction algorithms are based on deep neural networks (LSTM), making them straightforward to design and resource-efficient. Based on this approach, a Nowcast of temperature and pressure for the next one to six hours could be generated using the time series multivariate dataset for airport temperature and pressure (representative station: Patna Airport) with input features including date and time, temperature, atmospheric pressure, humidity, dew point temperature, wind direction, wind speed, cloud amount, and present and past weather. Pearson correlation coefficients determined input features. The model's efficacy and accuracy are measured by comparing anticipated temperatures and pressure (MSLP) to observed temperatures and using several performance indicators. This technique aims to construct and implement robust and dynamic self-supervised LSTM models to automate takeoff data, which is essential for flight safety and efficiency.