Investigating anomalous absorption using surface measurements

Transcription

1 JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 108, NO. D24, 4761, doi: /2003jd003411, 2003 Investigating anomalous absorption using surface measurements M. Sengupta 1 and T. P. Ackerman Pacific Northwest National Laboratory, Richland, Washington, USA Received 14 January 2003; revised 4 June 2003; accepted 13 June 2003; published 17 December [1] Flux measurements from the 415 nm band of the multifilter rotating shadowband radiometer and liquid water path from the microwave radiometer were used to derive effective radii in warm boundary layer clouds at the U.S. Department of Energy Atmospheric Measurement Program Southern Great Plains site. Surface fluxes computed using the effective radii retrieved using 415 nm measurements showed no bias when compared with observed broadband fluxes. It is therefore inferred that there is no excess absorption at solar and near-infrared wavelengths in the presence of warm boundary layer clouds. INDEX TERMS: 0305 Atmospheric Composition and Structure: Aerosols and particles (0345, 4801); 3359 Meteorology and Atmospheric Dynamics: Radiative processes; 3367 Meteorology and Atmospheric Dynamics: Theoretical modeling; KEYWORDS: anomalous absorption Citation: Sengupta, M., and T. P. Ackerman, Investigating anomalous absorption using surface measurements, J. Geophys. Res., 108(D24), 4761, doi: /2003jd003411, Introduction [2] Using data from the Atmospheric Radiation Measurement Program (ARM), Southern Great Plains (SGP), Oklahoma site, several scientists [Dong et al., 2000; Min and Harrison, 1996] have demonstrated that, for single layer water clouds, cloud optical depth and effective radius can be deduced from a combination of liquid water path and solar transmission measurements. In their technique, Min and Harrison use solar transmission measured in a 10 nm wide passband around 415 nm to derive cloud optical depth and concurrent liquid water path measurements to calculate, subsequently, cloud droplet effective radius. We hypothesize that this value of the optical depth can be used to compute accurately the broad-band solar transmission, when appropriately scaled across the solar spectrum using a lognormal size distribution with the retrieved effective radius and known refractive indices for water. We can check this hypothesis easily because the ARM sites provide us with simultaneous measurements of narrowband and broad-band solar transmission. Given a sufficiently large number of events with varying values of liquid water path, we think that this comparison of measured and computed values of broad-band solar transmission can also be used to prove or disprove the existence of significant excess or anomalous absorption [Ramanathan et al., 1995; Pilewskie and Valero, 1995; Cess et al., 1995; Zender et al., 1997]. If anomalous absorption actually occurs, we would expect a persistent bias between calculated and observed surface broad-band flux because our model by definition only accounts for known sources of absorption. Agreement between any single comparison of calculated and 1 Now at Cooperative Institute for Research in the Atmosphere (CIRA), Colorado State University, Fort Collins, Colorado, USA. This paper is not subject to U.S. copyright. Published in 2003 by the American Geophysical Union. observed surface broad-band flux cannot be used to disprove anomalous absorption because some fortuitous choice of values could still produce agreement in transmission. However, as we demonstrate in the next section, it is extremely unlikely that any plausible mechanism for anomalous absorption could produce agreement between calculated and observed broad-band flux across a broad range of liquid water path and, by association, cloud optical depth. Such agreement could only occur if anomalous absorption were both perfectly gray, i.e., had no spectral dependence across the entire solar spectrum, and perfectly correlated with cloud optical depth. 2. Effects of Simulated Anomalous Absorption on Solar Transmission [3] In order to illustrate the relationship between atmospheric absorption and downwelling surface flux in the presence of anomalous absorption, we postulate three possible mechanisms for this absorption: externally mixed aerosol, some sort of internally mixed spectrally varying absorber (which we simulate by artificially increasing the imaginary part of the index of refraction for water), and a spectrally gray absorption that is proportional to the amount of cloud liquid water path. In addition to these three simulation sets, we include a simulation of the effect of changes in surface albedo. We shall see that surface albedo variations do not affect absorption but have an impact on solar transmission measurements. The studies supporting the existence of anomalous absorption have argued that measurements show significant enhancements as large as 40 to 50% in atmospheric column absorption compared to model calculations. We present our results as ratios, so a value of 1 in the absorption ratio implies no anomalous absorption. In our simulations, we make some fairly extreme assumptions in order to force absorption ratios to exceed 1.4 or 1.5 across a range of liquid water path values. These assumptions are not AAC 1-1

2 AAC 1-2 SENGUPTA AND ACKERMAN: BRIEF REPORT Figure 1. Change in absorption as a result of (a) increase in external aerosol loading and change in composition, (b) change in surface albedo, (c) increase in absorption by cloud water droplets, and (d) incloud absorption from a gray absorber correlated to liquid water path. The change in absorption is depicted as a ratio when compared with a baseline. meant to represent real atmospheric behavior; they are meant to illustrate the expected behavior of solar transmission in the presence of anomalous absorption. For easier comparison of our results with the papers supporting anomalous absorption we also present our results as a ratio of cloud forcing at the top of the atmosphere and surface, where cloud forcing is defined as the difference in absorption between cloudy sky and clear by Pilewskie and Valero [1995]. [4] In all our simulations we use a lognormal distribution with effective radius of 7.5 mm and standard deviation of 1.46 mm to represent the cloud droplet distribution. While the effective radius is a representative value for the SGP [Sengupta, 2002] the standard deviation is an average from a number of continental stratus cases from Miles et al. [2000]. Clouds are taken to be 500 m thick with bases at 500 m, the liquid water path ranges from 0 to 0.4 mm and the constant solar zenith angle is 45. The midlatitude summer atmospheric profile from MODTRAN [Berk et al., 1989] for pressure, temperature, water vapor mixing ratio and ozone is used. The one-dimensional radiative transfer model (RAPRAD) used to compute surface and top-of-atmosphere broad-band fluxes model is based on the d2-stream numerical algorithm presented by Toon et al. [1989] and Kato et al. [1999]. [5] For externally mixed aerosol we use mineral dust with the Angstrom exponent value [Angstrom, 1929] from Kato et al. [1997] as a baseline. For comparisons we use soot, which has higher absorption than mineral dust over solar wavelengths, with mean radius of 0.12 mm and standard deviation of 2.0 mm [d Almeida et al., 1991]. The aerosol is taken to be evenly distributed in the lowest 1000 m of the atmospheric column. We consider three cases of soot with optical depths either same as the baseline, or two-fold, or ten-fold, increased. Our results show that we would require not only a highly absorbing aerosol but also a much higher optical depth to get a significant increase in the absorption (Figure 1a), which will then result in a comparably large decrease in surface flux (Figure 2a). It is interesting to note that an increase in absorption decreases the cloud forcing ratio to values below 1 (Figure 3a). [6] To investigate the surface albedo effects we compare a surface with spectrally constant albedo of 0.2 with a black surface and a surface with an albedo of 0.4. While change in surface albedo impacts downwelling surface flux (Figure 2b) there is no effect on the absorption (Figure 1b) and cloud forcing ratio (Figure 3b). [7] To simulate an internally mixed absorber with spectrally varying properties we increase the imaginary part of the index of refraction for water tenfold, 100-fold and fold with normal absorption being treated as the baseline. Large increases in the absorptive properties of the cloud, as expected, result in increased absorption (Figure 1c) and reduced surface flux (Figure 2c), while also increasing the cloud forcing ratio (Figure 3c).

3 SENGUPTA AND ACKERMAN: BRIEF REPORT AAC 1-3 Figure 2. Change in surface flux as a result of (a) increase in external aerosol loading and change in composition, (b) change in surface albedo, (c) increase in absorption by cloud water droplets, and (d) incloud absorption from a gray absorber correlated to liquid water path. The change in surface flux is depicted as a ratio when compared with a baseline. [8] Finally, we consider the impact of a gray absorber whose amount increases with increasing cloud liquid water path. Our baseline is the absence of this absorber. To create this scenario the number of gray absorbers is considered proportional to the cloud droplet concentration. As an example, for a liquid water path of 0.1 mm at 10% proportionality the gray absorber concentration is 18 cm 3. The absorbers are chosen to have an effective radius of 1 mm with a lognormal width of 0.20 and a single scattering albedo of It is seen that significant absorption (Figure 1d) corresponds to significant reduction in surface flux (Figure 2d) and significant increase in the cloud forcing ratio (Figure 3d). [9] In all three simulations, we find the expected result that any mechanism that increases absorption also reduces the solar transmission. Further, we find that the reduction in transmission is not a linear function of the liquid water path. It should also be noted that enhanced absorption is a result of a higher proportion of photons being absorbed on interaction with cloud droplets or other particles and aerosols and changes in the optical depth are negligible. 3. Comparison of Modeled and Observed Solar Fluxes [10] For our comparison study we select 18 days of single layer stratus at the ARM SGP site in the period between January 1997 and January 1998 using information from the Millimeter-wave Cloud Radar (MMCR) and Micropulse Lidar (MPL). To be labeled as a warm cloud the cloud base and top temperature have to be above 273 K and 253 K respectively. Atmospheric profile information is obtained from multiple sets of measurements which are used to construct the profiles at the highest possible temporal resolution (Gerald G. Mace, personal communication). Liquid water path information is obtained from retrievals using the dual-channel microwave radiometer (MWR) data [Liljegren et al., 2001]. As previously mentioned Min and Harrison [1996] use the 415 nm (a nonabsorbing band presumably) measurements from the Multifilter Rotating Shadowband Radiometer (MFRSR) to infer stratus cloud optical depths, which are subsequently used in conjunction with liquid water path measurements to calculate cloud droplet effective radii. Aerosol extinction optical depths obtained from the MFRSR [Michalsky et al., 2001] are used to retrieve the parameters of the Angstrom relationship [Angstrom, 1929]. The asymmetry parameter and single-scattering albedo of aerosol particles, treated as spheres, are computed using the Toon and Ackerman [1981] Mie codes for a mean radius of 0.58 mm and standard deviation of 1.35 mm[kato et al., 1997] with refractive indices of mineral dust particles from d Almeida et al. [1991]. The liquid water path, effective radius, cloud boundary, atmospheric profile and aerosol optical depth information are

4 AAC 1-4 SENGUPTA AND ACKERMAN: BRIEF REPORT Figure 3. Change in surface to top of atmosphere cloud forcing ratio as a result of (a) increase in external aerosol loading and change in composition, (b) change in surface albedo, (c) increase in absorption by cloud water droplets, and (d) in-cloud absorption from a gray absorber correlated to liquid water path. used as inputs to RAPRAD for computing downwelling broad-band fluxes at the surface. In our study we only consider cases with cloud fractions over 90%. Cloud fraction is calculated by a technique developed by Long et al. [1999] that uses a relationship between cloud cover fraction and the difference between the measured and estimated clear-sky downwelling shortwave diffuse irradiances. We compare our computed fluxes with measurements of surface flux provided by pyrheliometer observations of direct normal downwelling shortwave irradiance, and unshaded and shaded pyranometers observations of the total and diffuse downwelling shortwave irradiances, respectively. [11] Our data set consists of min, single layer, overcast, warm stratus clouds measurements spread over the 18 selected days. To eliminate solar zenith angle effects to first order we use calculated to observed ratios of surface fluxes. The mean flux ratio is seen to be 1.008, which points to an absence of bias in the comparison (Figure 4b). The correlation between calculations and observation is Using the T test the level of significance is computed to be 0.94 and the limits of the 95% confidence interval are and The standard deviation (s) of 0.26 is mostly from the scatter at low liquid water paths (Figure 4a) resulting from uncertainties in liquid water path measurements. To show that excess modeled absorption presented in the previous section would indeed translate to an effect which will be visible above the scatter in Figure 4, we computed surface fluxes using a spectrally varying absorber similar to the tenfold case in Figure 2c (Figure 5) and a gray absorber similar to the 10% case shown in Figure 2d (Figure 6). The mean flux ratio for spectrally varying enhanced absorption is with 95% confidence interval limits of and The running mean for 0.1 mm liquid water path bins in Figure 5a expectedly shows significant bias as opposed to the unbiased running mean in Figure 4a. The gray absorption shows similar bias with a mean flux ratio of and 95% confidence interval limits of and The running mean displayed in Figure 6a for 0.1 mm liquid water path bins also confirms our observation of bias for gray absorption. We therefore see that despite large scatter in the data, the large sample size leads to statistically significant differences in comparisons of the unbiased case with both the enhanced absorption examples. [12] The absence of bias in the mean flux ratio is what we would expect if the model adequately simulated atmospheric absorption. The only possible exception is the existence of a gray absorber that is correlated to the optical depth. This gray absorber would have to have near perfect correlation between absorption at 415 nm and absorption across the rest of the solar spectrum. Further, it would have to linearly increase with cloud optical depth (and cloud liquid water path). If the first condition is not met, we expect a bias in

5 SENGUPTA AND ACKERMAN: BRIEF REPORT AAC 1-5 [13] The results of our analysis show that there is no bias in our ratio of calculated to observed fluxes nor is there any observable dependence on liquid water path. The simulations show that we expect both bias and liquid water path dependence in the presence of significant anomalous absorption. Our analysis results do show a fairly large scatter that can be attributed primarily to cloud inhomogeneity and uncertainty in the measurement of liquid water path especially at small values. Because of this scatter, we cannot discount the presence of some small differences in absorption between model and actual atmosphere. Even so, the unbiased mean, the high correlation (0.87) between calculations and observations as well as the high level of significance of our T test results (0.94) argue strongly against the presence of anomalous absorption in our data set. Given the 18 days and 4850 points that we have included in our analysis, we think that our results are robust. While a very peculiar type of gray absorber cannot be detected by our comparison, the agreement between in situ data and narrow-band effective radius retrievals during ARESE II indicates the absence of such an in-cloud absorber. Furthermore other researchers have indicated that there is evidence to suggest the absence of any absorption at 500 nm [Valero et al., 1997], which ultimately negates the possibility of a gray absorber. Lastly we note that our results are perhaps not as rigorous as the results obtained from the Figure 4. Ratio of calculated to observed surface (a) as a function of liquid water and (b) as a histogram for min stratus data points. These data points are from SGP data and consist of 18 days. The solid line in Figure 4a represents a running mean with bin size 0.1 mm and is shifted by 0.01 mm for each calculation. The histogram has a bin size of the histogram (Figure 4b), while if the second condition is not met, we expect curvature in the scatterplot in Figure 4a. Additionally, if a gray absorber did exist, cloud droplet effective radius retrieval values would differ from in situ measurements. During the Atmospheric Enhanced Shortwave Radiation Experiment II (ARESE II), effective radii derived to match measured downwelling surface radiation agreed with aircraft measurements of particle size [Ackerman et al., 2003] thereby negating such a possibility. 4. Conclusion Figure 5. Ratio of calculated (including enhanced spectral absorption from increasing the imaginary part of the water refractive index tenfold) to observed surface flux (a) as a function of liquid water and (b) as a histogram for min stratus data points. These data points are from SGP data and consist of 18 days. The solid line in Figure 5a represents a running mean with bin size 0.1 mm and a shift of 0.01 mm. The histogram has a bin size of 0.05.

6 AAC 1-6 SENGUPTA AND ACKERMAN: BRIEF REPORT Figure 6. Ratio of calculated (including gray absorbers equal to 10% of the number of cloud droplets) to observed surface flux (a) as a function of liquid water and (b) as a histogram for min stratus data points. These data points are from SGP data and consist of 18 days. The solid line in Figure 6a represents a running mean with bin size 0.1 mm and a shift of 0.01 mm. The histogram has a bin size of ARESE II experiment [Ackerman et al., 2003] but do cover a broader range of conditions. [14] Acknowledgments. This research was supported by the Office of Biological and Environmental Research of the U.S. Department of Energy under contract number DE-AC06-76RL01830 as part of the Atmospheric Radiation Measurement Program. References Ackerman, T. P., D. M. Flynn, and R. T. Marchand, Quantifying the magnitude of anomalous solar absorption, J. Geophys. Res., 108(D9), 4273, doi: /2002jd002674, Angstrom, A., On the transmission of sun radiation and on dust in air, Geogr. Ann., 2, , Berk, A., L. S. Bernstein, and D. C. Robertson, MODTRAN: A moderate resolution model for LOWTRAN 7, Tech. Rep. 0122, Air Force Geophys. Lab., Hanscom Air Force Base, Mass., Cess, R. D., et al., Absorption of solar radiation by clouds: Observation versus models, Science, 267, , d Almeida, G. A., P. Koepke, and E. P. Shettle, Atmospheric Aerosols: Global Climatology and Radiative Characteristics, pp , A. Deepak, Hampton, Va., Dong, X., P. Minnis, T. P. Ackerman, E. E. Clothiaux, G. G. Mace, C. N. Long, and J. C. Liljegren, A 25-month database of stratus cloud properties generated from ground-based measurements at the Atmospheric Radiation Measurement Southern Great Plains site, J. Geophys. Res., 105, , Kato, S., T. P. Ackerman, E. E. Clothiaux, J. H. Mather, G. G. Mace, M. L. Wesely, F. Murcray, and J. Michalsky, Uncertainties in modeled and measured clear-sky surface shortwave irradiances, J. Geophys. Res., 102, 25,881 25,898, Kato, S., T. P. Ackerman, E. G. Dutton, N. Laulainen, and N. Larson, A comparison of modeled and measured shorwave irradiance for a molecular atmosphere, 61, , Liljegren, J. C., E. E. Clothiaux, G. G. Mace, S. Kato, and X. Dong, A new retrieval for cloud liquid water path using a ground-based microwave radiometer and measurements of cloud temperature, J. Geophys. Res., 106, 14,485 14,500, Long, C. N., T. P. Ackerman, and J. J. Deluisi, Estimation of fractional sky cover from broadband sw radiometer measurements, paper presented at the 10th AMS Conference on Atmospheric Radiation, Madison, Wis., Am. Meteorol. Soc., 28 June to 2 July, Michalsky, J. J., J. A. Schlemmer, W. E. Berkhaiser, J. L. Berndt, L. C. Harrison, N. S. Laulainen, N. R. Larson, and J. C. Barnard, Multilayer measurements of aerosol optical depth in the Atmospheric Radiation Measurement and Quantitative Links programs, J. Geophys. Res., 106, 12,099 12,107, Miles, N. L., J. Verlinde, and E. E. Clothiaux, Cloud droplet size distributions in low-level stratiform clouds, J. Atmos. Sci., 57, , Min, Q., and L. C. Harrison, Cloud properties derived from surface MFRSR measurements and comparison with GOES results at the ARM SGP site, Geophys. Res. Lett., 23, , Pilewskie, P., and F. P. J. Valero, Direct observations of excess solar absorption by clouds, Science, 267, , Ramanathan, V., B. Subasilar, G. J. Zhang, W. Conant, R. D. Cess, J. T. Kiehl, H. Grassl, and L. Shi, Warm pool heat budget and shortwave cloud forcing: A missing physics?, Science, 267, , Sengupta, M., Radiative impact of continental stratus at the Southern Great Plains, Ph.D. thesis, Pa. State Univ., Univ. Park, Toon, O. B., and T. P. Ackerman, Algorithms for the calculation of scattering by stratified spheres, Appl. Opt., 20, , Toon, O. B., C. P. Mckay, and T. P. Ackerman, Rapid calculation of radiative heating rates and photodissociation rates in inhomogeneous multiple scattering atmospheres, J. Geophys. Res., 94, 16,287 16,301, Valero, F. P. J., R. D. Cess, M. Zhang, S. K. Pope, A. Bucholtz, B. Bush, and J. Vitko Jr., Absorption of solar radiation by clouds: Interpretations of collocated aircraft measurements, J. Geophys. Res., 102, 29,917 29,928, Zender, C. S., B. Bush, S. K. Pope, A. Bucholtz, W. D. Collins, J. T. Kiehl, F. P. J. Valero, and J. Vitko Jr., Atmospheric absorption during the Atmospheric Radiation Measurement (ARM) Enhanced Shortwave Experiment (ARESE), J. Geophys. Res., 102, 29,901 29,915, T. P. Ackerman, Pacific Northwest National Laboratory, 3200-Q Avenue, MSIN K-9 24, Richland, WA 99352, USA. M. Sengupta, Cooperative Institute for Research in the Atmosphere (CIRA), Colorado State University, Foothills Campus, Fort Collins, CO 80523, USA. (sengupta@cira.colostate.edu)