A Matern class covariance function for modeling atmospheric delays in SAR interferometry
Abstract
The fine spatial resolution of InSAR data allows atmospheric delays to be spatially modeled by least-squares collocation (LSC). LSC requires a covariance model consistent with the data. The spectral behavior of atmospheric delays derived from InSAR data allows the parameters of the power spectral density function (variance, power-law exponent and frequency shift) to be estimated and used for constructing the corresponding variance-covariance function in the spatial domain. Introducing an additional parameter (shift in frequency domain) provides more flexibility in modeling different weather conditions. A new covariance function for atmospheric delays has been derived based on the Matern class of covariance functions. The constructed analytical covariance model is used to model the empirical covariance function of repeat-pass InSAR data. The new covariance model is tested for the prediction of atmospheric delays under varying weather conditions. The quality of the predicted atmospheric delays is discussed and conclusions about the applicability of the Matern class of covariance functions are drawn.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2005
- Bibcode:
- 2005AGUFM.G51C0858G
- Keywords:
-
- 1220 Atmosphere monitoring with geodetic techniques (6952)