Measurement Noise and Stratospheric Gravity Wave Characteristics Obtained from GPS Occultation Data

Abstract

Tsuda and Hocke (2002) reduce the amount of filtering applied to raw radio occultation data in order to increase the vertical resolution of GPS based temperature retrievals in the stratosphere. The authors discuss gravitywave characteristics obtained from these optimised retrievals, including characteristics of vertical wave number spectra for vertical wavelengths below 2 km, and the meridional distribution of potential energies for such short wavelengths. The filtering of raw data is required because real measurements contain noise. In this note, we discuss the role of measurement noise for gravity wave characteristics obtained from dry temperature retrievals of GPS based radio occultation soundings. We show that for realistic noise levels, temperature fluctuations can be safely interpreted as originating from small scale atmospheric waves in the altitude region below 30 km and for wavelengths greater than 2 km, with the detailed limits depending on noise and retrieval characteristics. An optimising procedure as the one applied by Tsuda and Hocke, however, may raise noise related effects to the level of atmospheric waves for short vertical wavelengths. We further show that conclusions for vertical wavelengths less than 2 km, if at all, can only be drawn after taking the actual noise characteristics of the observations and the details of the filter procedures applied in the retrieval into account.