Effects of ATMS Spectral Response Function Imbalance at Double-Side Band Channel on Brightness Temperature Simulations

Eleven of the Advanced Technology Microwave Sounder (ATMS) channels are designed with double side bands, which are required for reducing signal noise of the ATMS data. However, the current approach of using the symmetric boxcar spectral response functions (SRFs) for training the fast forward radiative transfer models (RTMs) could become problematic if the instruments are designed with large unbalanced side-band SRFs. For the eleven Joint Polar Satellite System-1 (JPSS-1) ATMS channels designed with side-band receivers, the SRFs are measured at each sideband, and the requirements for the side-band SRFs are specified to be less than 2 dB in terms of the mean gain ratio, which is defined as the averaged difference between the lower band and the high band parts. It is found that the mean gain ratio of some ATMS channels could reach 4 dB. This magnitude of the unbalanced side-bands can result in large systematic or random errors in model simulations of brightness temperatures at this channel. This study investigates the effects of unbalanced ATMS SRFs between the side bands on the simulations of brightness temperatures by using a line-by-line RTM. Brightness temperature simulations with the boxcar approximated SFRs are compared with those using the exact SRF at each side-band. It is shown that for an unbalanced gain of 2 to 4 dB at ATMS channel 18 to 22, the bias can be as large as 0.5 K. Variations of the SRF-imbalance-induced biases with respect to scan angle and scene temperature are also investigated for the simulations of the actual ATMS observations. It is suggested that the actual SRFs from all the sidebands must be carefully measured by the instrument vendor and provided to Numerical Weather Prediction (NWP) users for building an accurate fast RTM for satellite data assimilation in NWP models.