Estimation and prediction of weather variables from surveillance data using spatio-temporal Kriging

Abstract

State-of-the-art weather data obtained from numerical weather predictions are unlikely to satisfy the requirements of the future air traffic management system. A potential approach to improve the resolution and accuracy of the weather predictions could consist on using airborne aircraft as meteorological sensors, which would provide up-to-date weather observations to the surrounding aircraft and ground systems. This paper proposes to use Kriging, a geostatistical interpolation technique, to create short-term weather predictions from scattered weather observations derived from surveillance data. Results show that this method can accurately capture the spatio-temporal distribution of the temperature and wind fields, allowing to obtain high-quality local, short-term weather predictions and providing at the same time a measure of the uncertainty associated with the prediction.