The Use of Linear Regression Relations Derived from Model and Experimental Data for Retrieval of the Water Content of Clouds from Ground-Based Microwave Measurements

Research output

Abstract

Estimates of the error in determining the cloud liquid water path by the multiple linear regression (MLR) technique using different regression relations obtained both by model calculations and by experimental data (for reference, results of the method based on the inversion of the radiative transfer equation were taken) are presented. It is shown that if the MLR method is trained by experimental data and measurements in seven spectral channels of the radiometer, the random component of the liquid water path error in the cloud is 0.015–0.017 kg/m2, which is half that obtained when trained by the results of model calculations. The cloud liquid water path bias does not exceed 0.005 kg/m2. The MLR results allow one to reliably identify periods of clear sky by the criterion of the minimum variance of the water content.

Original languageEnglish
Pages (from-to)569-577
JournalAtmospheric and Oceanic Optics
Volume32
Issue number5
Early online date17 Oct 2019
DOIs
Publication statusPublished - 2019

    Fingerprint

Scopus subject areas

  • Earth-Surface Processes
  • Atomic and Molecular Physics, and Optics
  • Oceanography
  • Atmospheric Science

Cite this