Anomalies have been discovered in the ratio NIR/VIS of near-infrared to visible albedo of clouds measured from aircraft (Stephens and Tsay, 1990; Francis et al., 1997; Valero et al., 1997, 1999) and from the Nimbus7 satellite (Collins, 1998). Current radiative transfer models cannot explain these anomalies. This was part of the motivation for adding a near-infrared channel to NISTAR.
Most studies of this anomaly were based upon data sets of a few hours duration from aircraft campaigns, and thus could be dismissed as ephemeral or unrepresentative. Therefore, it came as somewhat of a surprise when a similar anomaly was discovered in the global record from the Nimbus-7 satellite, recurring continuously from 1979 to 1987 (Collins, 1998). Because the NIR channel on Nimbus7 was wide-field-of-view, the anomaly could not be attached to individual Earth pixels, but it was revealed by Collins using the method shown in Fig. 29. The ratio NIR/VIS has long been used as a measure of differences in cloud radiative properties between actual and modeled cloud systems (Stephens and Tsay, 1990). Anomalies in the measured spectral albedo ratio can also be linked directly to anomalies in the magnitude of atmospheric absorption (Collins, 1998).
Figure 29 plots the NIR/VIS albedo ratio versus the total shortwave albedo as a dashed line, while a solid line shows simulations from the NCAR CCM3. The Nimbus7 data and model simulation are global and extend from 1979 to 1987. Farther to the left in the plot (relatively cloud-free conditions), the observed and modeled albedo ratios are close. Farther to the right (increasing cloudiness and/or cloud albedo), the observations and simulations increasingly diverge, regardless of cloud type or cloud phase. The anomaly occurs for all tropical and mid-latitude ocean regions and exhibits minimal seasonal and inter-annual variability. Almost identical anomalies are obtained when the Nimbus7 data are compared with radiative transfer calculations based upon International Satellite Cloud Climatology Project (ISCCP) cloud data; therefore the anomalies are not an artifact of the CCM3 climate model.
Since Nimbus-7, no broadband near-infrared channel has been flown on a radiation space instrument. NISTAR not only restarts this measurement, but also provides several significant improvements over Nimbus-7. Perhaps the most important improvement will be the much better calibration and stability of the NISTAR radiometers. The Nimbus shortwave and near-infrared radiometers gradually degraded in the space environment (Kyle et al, 1995; Tashima and Hartmann, 1999). Calculations of the spectral albedos which include and exclude the degradation, and analysis of the time series of the spectral anomalies in the Nimbus data, both suggest that the long-term changes in the Nimbus instruments do not affect the detection of a cloud-albedo anomaly (Collins, 1998). Nonetheless, the conclusions from the Nimbus analysis should be evaluated with an independent set of observations.
The near-infrared radiometer will serve several purposes: first, it will help to test the Nimbus-7 result of Collins (1998) with accuracy and signal to noise ratio far superior to the Nimbus-7 ERB instruments. Second, it will provide a globally integrated test of the episodic but highly time- and space-localized findings of discrepant near infrared to total cloud albedo ratios (Stephens and Tsay, 1990; Francis et al., 1997; Valero et al., 1997). Third, because the near-infrared channel is sensitive to vegetation and snow/ice cover in addition to clouds, the near-infrared to total ratio is an attractively simple and fundamental analysis tool for studying global change, and Triana is the perfect vantage point to begin using that tool. (No current or planned LEO or GEO Earth radiation budget satellites have a broadband near-infrared channel, although CERES is apparently planning to add one in the post-2003 timeframe, which should serve as a nice complement to that on Triana.)
Simulated time series of the shortwave albedos and spectral albedo ratios for the Pacific Ocean are shown in Figures 30 and 31, respectively. The values are calculated for a subsatellite point of 0°N, 160°W for which 88% of the planetary surface viewed from Triana is covered by ocean. For simplicity, we have assumed that Triana will view this point once per day. The modeled time series for 1984 is derived by averaging fluxes computed from the NCAR Column Radiation Model applied to 3-hourly atmospheric state information from ISCCP (Collins, 1998). The observational time series is constructed by averaging fluxes measured by the Nimbus-7 wide-field-of-view radiometers. Although the correct procedure for computing the radiation at L-1 involves
integrating over the radiance field, geometrical averaging of fluxes provides reasonably accurate estimates. The ISCCP calculations are sampled at local noon, corresponding to the local noon equator crossing time of Nimbus-7.
Figures 30 and 31 illustrate that while the model is able to reproduce the observed albedo time series, the modeled albedo ratio is consistently larger than the observed albedo ratio. The results also show that there is considerable monthly and seasonal temporal variability in the albedo and albedo ratios, showing the necessity of long time series to acquire the needed information. Variability on seasonal and interannual time scales will be used to track the anomalies in cloud spectral radiation associated with synoptic weather patterns (e.g., the Madden Julian oscillation) and with ENSO.
If the measured, not the modeled lines in Figs. 29 and 31 are true, it would signify a major uncertainty in the Earth's energy balance. Other lines of evidence point to a similar conclusion, notably the globally averaged record of shortwave insolation at the surface, which shows a considerable deficit compared to global models. But this remains controversial.
While NISTAR, like Nimbus7's NIR channel, cannot resolve individual cloud masses, it is clearly of interest to begin to look at the NIR/VIS ratio on a global scale with a modern radiometer. We believe that because of its intrinsic interest to the vegetation, cloud, and snow/ice communities, this ratio also has the potential to become, in future, an important global change variable.
