Humidity sensor failure: a problem that should not be neglected

The problem of abnormally dry bias induced by radiosonde humidity sensor failure in the low and mid-troposphere is studied based on the global operational radiosonde relative humidity observations from December 2008 to November 2009. The concurrent humidity retrievals from the FORMOSAT-3/COSMIC...

Full description

Bibliographic Details
Main Authors: Y. Liu, N. Tang
Format: Article
Language:English
Published: Copernicus Publications 2014-11-01
Series:Atmospheric Measurement Techniques
Online Access:https://www.atmos-meas-tech.net/7/3909/2014/amt-7-3909-2014.pdf
Description
Summary:The problem of abnormally dry bias induced by radiosonde humidity sensor failure in the low and mid-troposphere is studied based on the global operational radiosonde relative humidity observations from December 2008 to November 2009. The concurrent humidity retrievals from the FORMOSAT-3/COSMIC radio occultation mission are also used to assess the quality of the radiosonde humidity observations. It is found that extremely dry relative humidity are common in the low and mid-troposphere, with an annual globally averaged occurrence of 4.2%. These low-humidity observations usually exist between 20 and 40° latitude in both the Northern Hemisphere and Southern Hemisphere, and from heights of 700 to 450 hPa. Winter and spring are the favored seasons for their occurrence, with a maximum fraction of 9.53 % in the Northern Hemisphere and 16.82% in the Southern Hemisphere. The phenomenon does not result from natural atmospheric variability, but rather humidity sensor failure. If the performance of humidity sensors is not good, low-humidity observations occur easily, particularly when the radiosonde ascends through stratiform clouds with high moisture content. The humidity sensor cannot adapt to the huge change of the atmospheric environment inside and outside stratiform clouds, resulting in sensor failure and no response to atmospheric change. These extremely dry relative humidity observations are erroneous. However, they have been archived as formal data and applied in many research studies. This may seriously undermine the reliability of numerical weather prediction and the analysis of weather and climate if quality control is not applied before using these data.
ISSN:1867-1381
1867-8548