Spatial downscaling of AIRS-derived column water vapor using ratio model to improve LST retrieval

Atmospheric column water vapor, which is the total atmospheric precipitable water vapor contained in a vertical air column, is one of the most important factors in all surface-atmosphere interactions (such as energy fluxes between the earth and the atmosphere) and plays a key role in wide variety of environmental studies, ecological and agricultural applications. However, measuring this parameter at meteorological stations requires the use of radiosonde instruments, which being pointwise and costly are limitations of these observations. Therefore, remote sensing is used as an alternative to estimate this important atmospheric parameter. Compared to other atmospheric parameters, atmospheric water vapor which attenuates remotely sensed radiance is of great importance. Although this atmospheric parameter is measured by AIRS (Atmospheric Infrared Sounder) sensor, its low resolution (about 40 km) is not acceptable for many applications. Therefore, developing an algorithm to downscale the AIRS-derived column water vapor is the main goal of this study, so that its spatial resolution can be improved. To do this, using the ratio method, the AIRS-derived column water vapor is fused with the MODIS (Moderate Resolution Imaging spectroradiometer) data. Then, due to the major influence of this parameter on Land Surface Temperature (LST) estimation, the role of improved resolution atmospheric column water vapor in the estimation of LST is investigated as a secondary goal. In order to validate the estimated parameters and evaluate their accuracy, independent datasets were used. Results of the implementation indicate that proposed downscaling method has high potential to enhance the spatial resolution of AIRS-derived atmospheric column water vapor, without significant degradation of the RMSE. It was also found that the atmospheric column water vapor when moving into higher spatial resolution can dramatically increase the accuracy of the LST estimation.