Estimation of precipitable water vapour from NOAA-AVHRR data during the Hapex Sahel experiment

Publisher Website
Google Scholar
Abstract
The simple split-window and the transmittance ratio method for estimation of atmospheric precipitable water content were analysed using data gathered through the Hapex Sahel Experiment conducted in Niger from August to October 1992. Results are presented for the Hapex Sahel square degree. Both methods are designed to work with thermal infrared data from the Advanced Very High Resolution Radiometer flown on the NOAA polar orbiting satellites. The performance of the methods were evaluated with radiosonde measurements and prognosticated water vapour fields obtained from the global circulation model used at the European Centre for Medium Weather Forecast. The implications of extending the simple split-window method to work over land surfaces were investigated by use of both simulated and observed data. It was shown that changing surface conditions, i.e., emissivity and surface temperature, could significantly influence the split-window estimate of water vapour content and therefore inhibit the implementation of a global model. An optimized split-window algorithm with inclusion of the land surface temperature could increase the explained variance of the radiosonde measurements and it could reproduce the variation of the water vapour through the growing season very well. The performance of the transmittance ratio method was shown to be more unstable but in some cases it compared well with observed data. This result might be attributed both to the non-optimal characteristic of the satellite sensor as well as the problems with thermal homogeneity of the land surface.