Machine learning methods for Precipitable Water Vapor estimation by radiometric data in millimetre wavelength

Abstract

This work deals with the first try to calculate the amount of Precipitable Water Vapor (PWV) in atmosphere by using machine learning and AI methods. We use the detector voltages series measured by radiometric system “MIAP-2” as the initial data for machine learning. The radiometer MIAP-2 works by “atmospheric dip method” in 2mm and 3mm atmospheric transparency windows. We also have PWV data series collected by Water Vapor Radiometer and GNSS receiver for data validation. The best convergence results were demonstrated by the independent component analysis (ICA) method with coefficient of determination R2= 0.53 and artificial neural network method (ANN) with R2= 0.8. These methods allow to reduce the systematic errors due to direct PWV calculation from raw radiometric data avoiding unnecessary steps opacity calculation.

Measurements and Evaluations of the Atmospheric Transparency at Short Millimeter Wavelengths at the Candidate Sites for Millimeter and Submillimeter Wave Telescopes

“…The bias introduced by the flat layered atmospheric model was also excluded. The coefficient of determination R 2 on the validation part of the dataset was 0.8 for kNN and 0.86 for ANN with symmetrical deviations from the trend line [58]. The result of this findings was that the PWV may be consistently determinated from the sky dip data using the statistical approach.…”

“…Recently we proposed the statistical approach for the determination of the PWV values based on machine learning algorithms and MIAP-2 data [58]. The idea was to use non-linear regression to calculate PWV values from the MIAP data using the receiver responses on different angles in the 2 mm and 3 mm bands as the input data, without addressing to the physical model.…”