The next generation of GOES satellites (GOES-I to M, sometimes referred to as GOES-NEXT), the first of which is scheduled for launch in 1992 (Smith 1989), will not carry a VAS. Instead, it will carry two new visible and infrared radiometers, one for imaging and the other for sounding. The spacecraft itself will be of a new design and will be three-axis stabilized so that the radiometers will always face the Earth.
The GOES-I/M Imager will be a calibrated 5-channel radiometer (Table 3), with one visible channel with 1-km resolution, three infrared channels with 4-km resolution, and one infrared channel with 8-km resolution; these data will be digitized with 10-bit resolution (Rao et al. 1990). The predicted radiometric performance is given by Koenig (1990).
The GOES-I/M Sounder will be a 19-channel radiometer. All but one will be in the infrared (Table 4), with a spatial resolution of 10 x 40 km at nadir. As with the VAS, the channel selection will be by a filter-wheel. These data will be digitized to 12-bit resolution (Rao et al. 1990). The predicted radiometric performance of each channel was given by Koenig (1990), but recent bench tests of the GOES-I Sounder show that the actual noise levels are likely to be much higher, at least for the first instrument of the series and may result in little improvement over current VAS sounding capabilities (Hayden and Schmit 1991).
Both the Imager and Sounder are design derivatives of the AVHRR and HIRS radiometers on the NOAA polar-orbiters. Thus, they will have external blackbody calibration targets for in-flight calibration. This will avoid the need to model the contributions of the telescope foreoptics and permit frequent in-flight calibration measurements. The temperature of the calibration target will be monitored by thermistors at four locations. Temperature changes across the face of the target are expected to be less than 0.3K in most cases, but could be as large as 1.0K at the worst sun angles. Both the Imager and Sounder will view cold space at 2-minute intervals. Based on worst-case thermal changes, the largest change in calibration offset of the most sensitive channels in this interval is expected to be ~0.15K for Channel 5 of the Imager, which has a specified NEDT of 0.35K, and ~0.037K in Channels 7 and 13 of the Sounder, which matches the specified NEDT. The Imager will view the calibration blackbody target at intervals of between 10 and 25 minutes, the Sounder will view every 10 minutes. The worst-case uncertainties in the infrared measurements caused by gain changes in these intervals are expected to be less than the NEDT in all channels. The Imager will use at least 100 samples of the space and calibration target views for each calibration cycle, while the Sounder will use 40 samples per channel (Koenig 1990).
The non-linearities of the responses of the mercury cadmium telluride infrared detectors will be accommodated by a quadratic expression relating the detector output to radiance, in which the linear part is given by the in-flight calibration and the quadratic component by the pre-launch testing (Weinreb 1990). As a result of the three-axis stabilization, the spacecraft will be subjected to larger diurnal temperature changes than the current GOES platforms. Sunlight will not be totally excluded from entering the radiometer apertures during the four hours or so around local midnight (Koenig 1990), and this will have consequences on the thermal conditions inside the radiometer. Thermistors inside the radiometers will monitor internal temperatures, and a simple linear correction, with a diurnal cycle, will be applied during the data processing (Weinreb 1990).
Data from both the Imager and the Sounder will be processed at the Command and Data Acquisition Station at Wallops and immediately rebroadcast to users via the satellite, as the GVAR (GOES variable formatted data stream). For the Sounder, both raw and processed data will be rebroadcast; for the Imager, only the processed data can be accommodated. A limitation imposed by the bandwidth (Weinreb 1990).
There is no provision for in-flight calibration of the visible channels on either instrument, but an improved normalization scheme will be used to avoid stripes across the visible imaged data caused by differences in detector responses (Weinreb 1990).
The vertical resolution of atmospheric profiles derived from sounder data is limited by the number and spectral resolution of the channels. As a consequence,
radiometers with channel definition determined by filter wheels have limited scope for development for improved meteorological products, especially for mesoscale
studies (Smith et al. 1990). In the mid- to late '90s a new type of sounding radiometer is planned for both geosynchronous and polar orbiting satellites. This
will be either an interferometer or grating spectrometer (AIRS - Advanced Infrared Radiometer Sounder) and will have up to 4000 channels in the infrared spectral
range of 3.7 to 17 mm, with nominal spectral resolution of 0.4-1.0 
, and a ground resolution of ~10 km (Smith 1989). This great increase in the spectral
resolution is anticipated to significantly reduce the errors in the retrievals of the atmospheric temperature and humidity profiles. An airborne prototype
(HIS - High spectral resolution Interferometer Sounder) with 2500 channels has already flown on the NASA ER-2 aircraft and produced results which compare well
with radiosonde profiles (Smith 1989).
A possible method of quickly implementing an interferometric spectrometer on a geosynchronous satellite would be to modify the GOES-I/M Sounder by replacing
the filter-wheel assembly with a HIS type interferometer module (Smith et al. 1990). This proposed instrument would be called the GOES High Resolution Interferometer
Sounder (GHIS). A more flexible scan system, which would increase the data rate from 40 kbs-1 to 500 kbs-1, could also be implemented. The improvement in the
noise performance for GHIS would be as much as a factor of 3 to 8 over that specified for the long-wave channels of the GOES-I/M Sounder. The expected accuracies
of the derived temperature and water vapor profiles would be 
1K for temperature and 2-3K for dew-point. However, the major advance would be in the vertical
resolution of the profiles, especially for water vapor (Smith et al. 1990).
