A Consistency Analysis of ARESE Aircraft Measurements

Li, Z., University of Maryland

Radiation Processes

Radiative Processes

Li, Z., A.P. Trishchenko, H.W. Barker, G.L. Stephens, and P. Partain, 1999: "Analyses of Atmospheric Radiation Measurement (ARM) program's Enhanced Shortwave Experiment (ARESE) multiple data sets for studying cloud absorption," J. of Geophys. Res. 104(D16):19127-19134


Figure 1. Comparisons of TOA albedos inferred from measurements made by TSBR, GOES-8, and SSP. Two sets of GOES-based estimates are shown for an aircraft along the Egrett's flight path: one for an aircraft skimming the cloud tops (dotted lines); and another for one at 14 km (thin solid lines). Squares represent SSP measurements made near SGP/CF.


Figure 2. The relations between TOA albedo and amospheric transmittance. TOA albedos were derived from GOES-8 for ARESE and April 1996, from GOES-7 for April 1994, from ScaRaB for 1994, and from SSP for October 1995. Atmospheric transmittance was computed from surface irradiance measurements from BSRN. Regression equations are listed on the plot.


Figure 3. Atmospheric absorptance computed from satellite (ScaRaB) and surface (BSRN) measurements from March 1994 to February 1995 at the SGP/CF. Data for inhomogeneous clouds were screened out by restricting the standard deviation of 30 minute surface observations to less than 20 Wm-2.


Figure 4. A comparison of TOA visible albedos inferred from both GOES-8 (thin line) and SSP (solid line) data in which the SSP used GOES-8's response function.


Figure 1. Comparisons of TOA albedos inferred from measurements made by TSBR, GOES-8, and SSP. Two sets of GOES-based estimates are shown for an aircraft along the Egrett's flight path: one for an aircraft skimming the cloud tops (dotted lines); and another for one at 14 km (thin solid lines). Squares represent SSP measurements made near SGP/CF.

Figure 2. The relations between TOA albedo and amospheric transmittance. TOA albedos were derived from GOES-8 for ARESE and April 1996, from GOES-7 for April 1994, from ScaRaB for 1994, and from SSP for October 1995. Atmospheric transmittance was computed from surface irradiance measurements from BSRN. Regression equations are listed on the plot.

Figure 3. Atmospheric absorptance computed from satellite (ScaRaB) and surface (BSRN) measurements from March 1994 to February 1995 at the SGP/CF. Data for inhomogeneous clouds were screened out by restricting the standard deviation of 30 minute surface observations to less than 20 Wm-2.

Figure 4. A comparison of TOA visible albedos inferred from both GOES-8 (thin line) and SSP (solid line) data in which the SSP used GOES-8's response function.

In an attempt to resolve the recent debate over the cloud absorption anomaly, the U.S. Department of Energy sponsored a field experiment in the fall of 1995 under the auspices of its Atmospheric Radiation Measurement (ARM) Program. The experiment named ARM Enhanced Shortwave Experiment (ARESE) took place around the southern great plains (SGP) central facility (CF) in Oklahoma. Following ARESE, a cloud absorption anomaly of unprecedented magnitude and unknown origin was presented. Valero et al. (1997) employed coeval measurements of upwelling and downwelling radiative fluxes made on two stacked aircraft (above and below cloud) and reported that cloud absorption increases dramatically with cloud fraction. For a heavy overcast (October 30, 1995), they claimed that the layer between the aircraft (mostly cloud) absorbed 37% of the incoming solar irradiance. This contrasts sharply with model estimates of total atmospheric absorptance which are usually around, or less than, 24% regardless of sky condition (Li et al. 1997). If clouds on October 30 were indeed so absorptive and representative, other relevant radiometric measurements should be able to detect such a strong signal. The purpose of this study is to examine if measurements made by other instruments support the finding of Valero et al. (1997).

The study employed measurements of solar radiation made by space-borne, air-borne, and ground-based radiometers over the SGP/CF site in Oklahoma (36.60ºN, 97.485ºW).

Satellite data include measurements from the Scanner for Radiation Budget (ScaRaB) onboard Meteor 3 satellite and the Visible and Infrared Spin-Scan Radiometer (VISSR) onboard GOES-7 and GOES-8. ScaRaB provided calibrated shortwave (SW) (0.2 - 4.8 mm) and visible (~0.6 mm) reflected flux/albedo measurements. The calibration accuracy is estimated to be 1-2% (Kandel et al. 1998; Trishchenko and Li 1998). While GOES provides data at high temporal (every 15-30 minutes) and spatial (about 1 km at nadir viewing) resolutions with no on-board calibration. The indirect post-launch calibration was used to derive SW albedos by means of narrowband-to-broadband conversion. Both the calibration and narrowband-to-broadband conversion used in processing GOES-7 data were validated against ScaRaB measurements (Trishchenko and Li 1998). Calibration for GOES-8 was based on a comparison with NOAA-14 AVHRR measurements.

Aircraft data used in this study were restricted to the overcast day (October 30) and included observations from zenith- and nadir-pointing Total Solar Broadband Radiometers (TSBRs) that measured total solar irradiance between 0.224 mm and 3.91 mm, Total, Direct, Diffuse Radiometers (TDDRs) that measured 10 nm wide spectral bands centered at seven wavelengths between 0.5 mm and 1.75 mm (Valero et al. 1997), Scanning Spectral Polarimeter (SSP) (Stephens et al. 1998). We used data from the Egrett aircraft flown at the 14 km above the surface level.

To obtain SW albedos at the TOA from these spectral fluxes for comparison with satellite observations, narrowband visible albedos are first computed at the Egrett altitude and then converted to TOA values. The conversions were made by modeling with a 2-stream radiative transfer model. The narrowband spectral albedos were then integrated over the ScaRaB visible bandpass and weighted by its spectral response function. The resulting ScaRaB equivalent visible albedos at the aircraft's altitude are further corrected to TOA visible albedo by means of radiative transfer modeling. From these equivalent ScaRaB visible albedos at the TOA, SW albedos were derived using a observational narrow-to-broadband conversion relation derived from ScaRaB over the SGP region (Li and Trishchenko 1997). The conversion is accurate to within 2%. Surface irradiance measurements made at the SGP site were employed from 1994 to 1996. They were made with an observing system known as the BSRN.

Figure 1 shows TOA albedos obtained from TSBR, GOES-8, and SSP on October 30, 1995. The curve denoting albedos derived from GOES-8 are discontinuous as the GOES images were separated by approximately 15 minutes. Cloud reflectances were assumed to remain invariant during this interval. It is seen that fluctuations in albedo as measured by the aircraft are similar to those inferred from satellite radi-ances, but the magnitudes of albedos inferred from GOES, TSBR, and SSP differ drastically. The mean-bias difference is 6% between GOES and TSBR and about 14.4% between SSP and TSBR! The disparity in albedo is comparable to the magnitude of the CAA reported by Valero et al. (1997).

To understand the discrepancies, comparisons were made between ScaRaB, GOES-7, GOES-8, and SSP. For the ARM experiment, GOES-8 data were made available in 1995 and 1996 (Minnis and Smith 1998), GOES-7 and ScaRaB in 1994. The difference in time and strong cloud variability make the direct comparison of TOA albedos meaningless. However, it is revealing to compare the relationship between TOA albedo and surface transmittance. The slope of the relation was proposed to assess cloud absorption by Cess et al. (1995). Although the slope can be an ambiguous indicator of cloud absorptance (Li and Moreau 1996; Barker and Li 1997), it is much less variable than TOA albedo and atmospheric transmittance, especially for overcast scenes. For broken clouds, the approach suffers from considerable uncertainties due to large errors in matching TOA and surface measurements (Arking et al. 1996) and to horizontal transport of photons (Barker and Li 1997). Given these limitations, all matched pairs of TOA and surface measurements were screened based on the standard deviation (SD) of surface irradiance. Since partly cloudy scenes evolve more rapidly than do clear and overcast scenes, data were retained if SD were less than 20 Wm-2.

Figure 2 presents an albedo-transmittance plot for screened data. TOA albedos were derived from GOES-7, GOES-8, ScaRaB, and SSP. GOES data represent averages over grid cells of 0.3o�0.3o centered at the ARM CF, and ScaRaB data are for individual pixels closest to the CF. SSP data were taken within 1 km around the CF for homogeneous cloud scenes. Atmospheric transmittances were computed from broadband surface irradiances observed with BSRN at the CF. Due to the data screening, the majority of data points correspond to either clear or overcast scenes. The most striking feature of Figure 2 is the tight cluster of clear-sky points on the right, and the presence of two distinct clusters on the left (overcast). One cluster consists of data from GOES-7, ScaRaB, and SSP while the other consists of GOES-8 data. The slopes for the least-square linear regression lines for GOES-7 (-0.78) and ScaRaB (-0.82) data are indistinguishable from model values (~0.8) (Cess et al. 1995; Li and Moreau 1996). The SSP data points distribute closely around the regression lines of ScaRaB and GOES-7. This finding contradicts the existence of a significant cloud absorption anomaly. Indeed, atmospheric absorptances computed from ScaRaB TOA and surface measurements show no systematic difference between clear and overcast skies (see Figure 3) for the data collected in 1994. The regression slope for GOES-8 data, however, is significantly smaller than the others.

If all measurements are correct, one has to conclude that clouds in 1995 and 1996 were anomalously absorptive relative to those in 1994. This conclusion is difficult to accept, barring protracted, yet intermittent, environmental changes that produced dirty clouds over the period and location in question. To our knowledge, this did not occur. A more sound explanation is inconsistent calibration in processing GOES-7 and GOES-8 data. The calibration inconsistency is more evident from a comparison of visible albedos observed by GOES-8 and computed from SSP measurements weighted by the GOES-8 response function in the visible band, as is shown in Figure 4.

Similar inconsistent results were obtained by Dong et al. (1998). Excellent agreement was achieved for GOES-7 during April 1994, but GOES-8 TOA albedos were generally less than those deduced from ground-based measurements by about 15%.

Other potential contributing factors to the discrepancy may be differences in the calibrations of the BSRN radiometers used in April 1994 and October 1995 and an uncertainties in the SSP calibrations. This was found not to be a significant problem when compared to other instruments for a variety of scenes.

Since the albedos observed by TSBR are the lowest (6% less than those from GOES-8), an even larger calibration uncertainty for TSBR certainly cannot be ruled out. It is thus imperative to solve the calibration conundrum before accepting the conclusion of the existence of an enormous CAA, as found by Valero et al. (1997).

Following the ARM/ARESE experiment, Valero et al. (1997) showed a cloud absorption anomaly (CAA) of unprecedented magnitude. An analysis is presented here to examine if their finding is consistent with observations from multiple sensors on various platforms including those onboard satellite (ScaRaB, GOES) and aircraft (TSBR, SSP) and on the ground (BSRN). It was found that albedos measured with the TSBR radiometer as used by Valero et al. (1997) systematically less by 0.06 and 0.144, respectively, than values estimated from GOES-8 imagery and SSP. The discrepancy appears to stem from inconsistent calibrations among the radiometers. An analysis of the regression between TOA albedo and atmospheric transmittance revealed a nearly identical slopes derived from SSP, ScaRaB and GOES-7 which are in excellent agreement with model values and at variance with the finding of Valero et al. (1997).

Arking, A., M.-D. Chou, and W.L. Ridgway, 1996: On estimating the effects of clouds on atmospheric absorption based on flux observations above and below cloud level, Geophy. Res. Let., 23, 829-832.

Barker, H.W., and Z. Li, 1997: Interpreting shortwave albedo-transmittance plots: True or apparent anomalous absorption, Geophy. Res. Let. 24, 2023-2026.

Cess.R.D., M.H. Zhang, P. Minnis, L. Corsetti, E.G. Dutton, B.W. Forgan, D.P. Garber, W.L. Gates, J.J. Hack, E.F. Harrison, X. Jing, J.T. Kiehl, C.N. Long, J.-J. Morcrette, G.L. Potter, V. Ramanathan, B. Subasilar, C.H. Whitlock, D.F. Young, Y. Zhou, 1995: Absorption of solar radiation by clouds: Observations versus models. Science, 267, 496-499.

Dong, X., W.L. Smith, Jr., P. Minnis, 1998: Comparison of stratus cloud optical depths retrieved from surface and GOES measurements over the SGP ARM central facility, ARM Science Team Meeting, Tucson, Arizona, March 23-27, 1998.

Kandel, R., M. Viollier, P. Raberanto, J.P. Duvel, L.A. Pakhomov, V.A. Golovko, A.P. Trishchenko, J. Mueller, E. Raschke, and R. Stuhlmann, 1998: The ScaRaB earth radiation budget dataset. Bull. Amer.Meteor.Soc., in press.

Li, Z., and L. Moreau, 1996: Alteration of atmospheric solar absorption by clouds: Simulation and observation, J. Appl. Meteor., 35, 653-670.

Li, Z., L. Moreau, A. Arking, 1997: On solar energy disposition, A perspective from observation and modeling, Bull. Amer. Meteor. Soc., 78, 53-70.

Li, Z., and A. Trishchenko, 1997: A study towards an improved understanding of the relationship between visible and SW albedo measurements, J. Atmos. & Ocean. Tech. Submitted.

Minnis, P. and W. L. Smith, Jr., 1998: Cloud and radiative fields derived from GOES-8 during SUCCESS and the ARM-UAV spring 1996 flight series, Geophys. Res. Lett., In press.

Stephens, G.L., R. McCoy, P. Partain and P. Gabriel, 1998: A multipurpose spectrally radiometer - instrument characteristics and example measurements, submitted to J. Oceanic Atmos Tech.

Trishchenko A., Li. Z, 1998, Use of ScaRaB measurements for validating a GOES-based TOA radiation product, J. Appl. Meteor., in press.

Valero, F. P. J., R.D. Cess, M. Zhang, S.K. Pope, A. Bucholtz, B. Bush, J. Vitko, Jr., 1997: Absorption of solar radiation by the cloudy atmosphere: Interpretations of collocated aircraft measurements, J. Geophy. Res., 102, 29,917-29,928.