Cooperative Institute for Research in Environmental Sciences

A winning proposal for the Innovative Research Program, 2006:

Highly resolved wavelength dependencies of aerosol optical properties in the shortwave spectrum

Investigators: Allison McComiskey (CIRES/NOAA), Thomas H. Painter (CIRES/NSIDC), Paul Ricchiazzi (ICESS/UCSB)

Objective. Our objective is to measure the wavelength dependence of aerosol optical properties at high spectral resolution across the solar spectrum. Quantification of these properties is necessary for modeling the impact of aerosols on climate, but direct, high spectral resolution measurements of these properties have never been made. These measurements require instrumentation that can determine the angular distribution of sky radiance at high angular resolution (< 0.1°) in certain angular domains with the capabilities of a field spectroradiometer. We propose here to modify an existing field-portable, hyperspectral goniometer system and its angular sampling protocols to produce the first data set of this kind.

Background and importance. Atmospheric aerosols directly scatter and absorb solar radiation and modify the development and characteristics of clouds, leading to a global average decrease in radiation and a cooling at the Earth’s surface. This cooling may partially offset the warming caused by an increase in greenhouse gases, but the uncertainty in observations of aerosol optical properties makes this effect difficult to quantify. Aerosol distributions vary greatly in space and time, and continuous measurements are sparse. Moreover, the nature of these measurements has been sufficiently costly and complex that measurements have only been made in fewer than 10 wavelengths in the visible and near-infrared portions of the shortwave spectrum. In order to compute broadband shortwave radiation budgets for climate modeling, the wavelength dependence of these properties must be known continuously across the spectrum. Currently, model inputs are extrapolated from a few measurements using wavelength relationships that are unverified. The first-order uncertainty in modeling the recent observed changes in climate derives from the uncertainties associated with aerosol optical properties.

The following aerosol properties are required to calculate radiation budgets for climate modeling: aerosol amount (typically an optical depth), the single scattering albedo (ratio of scattering to extinction), and the asymmetry parameter (a metric of the directionality of scattering). The single scattering albedo is derived from knowledge of both aerosol scattering and absorption. The asymmetry parameter is the cosine weighted integral of the aerosol scattering phase function. It is a single value that describes the angular scattering of aerosols that is computationally efficient for radiative transfer modeling. Routine methods for measuring these aerosol properties include in situ measurements from the surface or aircraft and ground radiometry, providing information for the entire atmospheric column. In situ measurements typically provide aerosol scattering and absorption at no more than three wavelengths in the visible spectrum. The asymmetry parameter can only be derived from an empirical relationship that requires large assumptions about the phase function of the aerosols. NASA’s AERONET Program operationally retrieves aerosol properties from Cimel Sunphotometers (CSPHOT) that make angular radiance measurements but are limited to the almucanter plane and measurements at 8 wavelengths from 0.340 to 1.022 µm.

A method for retrieving the single scattering albedo and phase function from angular measurements of sky radiation has been developed by Wang and Gordon (1994). This method was tested using CSPHOT data and extended by Ricchiazzi et al. (2006) to provide better accuracy by including a more detailed characterization of surface reflectance. We propose to apply this method to radiance data collected with the Automated Spectro-Goniometer (ASG), developed by Dr. Painter (Painter et al., 2003), a spherical robot that is coupled with a hyperspectral radiometer, an Analytical Spectral Devices (ASD) FieldSpec FR (Figure 1a). The ASD-FR has spectral resolution of 3 to 10 nm across the wavelength range 350 to 2500 nm. The ASG will measure sky radiance while moving though a comprehensive set of angles, similar to those accessed by the CSPHOT. Modification of the ASG to incorporate quadrant centering, reorientation to sky viewing, and introduction of sky-specific angular protocols with real-time solar tracking with updating at > 1 Hz will facilitate a quantum step forward in the field of aerosol characterization.

What makes this innovative? How might this be interdisciplinary? The use of this instrument is novel for atmospheric characterization and the configuration proposed here, to scan the atmosphere, was conceived by the proposors and is an outgrowth of our published experience using the ASD-FR to analyze the sky radiance data at a high-latitude site. Most instruments developed for observation of downwelling, atmospheric radiation are limited in either their spectral or angular resolution. This will be the first such instrument designed for combined, very high spectral and angular resolution, which provides the leverage to simultaneously retrieve the pertinent aerosol properties at high accuracy throughout the shortwave spectrum. This project also brings together researchers from different disciplines. The ASG was originally designed as a downlooking instrument for observations of surface reflectance properties. With the use of methods from the atmospheric sciences, the instrument can be applied to a very different and important problem in aerosol and climate science.

Research Plan. Radiometric retrievals of atmospheric aerosol properties require observations under cloudless conditions. In most cases, aerosol characteristics remain constant throughout a given day, though rapid changes are also possible when dictated by variations in meteorology and source strength. In order to incorporate as much of this variability as possible in our measurements, deployments will be made on clear sky days as often as possible throughout the year. Full sky-radiance scan will consist of measurements obtained at approximately 100 scattering angles, either in the principal plane, or at constant zenith viewing angle. In order to obtain these measurements at sufficient angular resolution, the instrument will be fitted with a 1o field of view optic. By retrofitting of the ASG with a solar quadrant centering unit, the goniometer will have solar tracking capabilities to ~0.1o required for accurate solar position location and accurate retrievals. This large number of sky-radiance samples will allow for the identification and removal of views contaminated by sub-visual cirrus clouds. In addition, having a large number of measurements at small scattering angles will be used to identify and correct for angular pointing errors.

The robot will be programmed to produce up to 10 sky-radiance scans per hour, which is about an order of magnitude greater than the typical sample rate from AERONET. The availability of several scans per hour will help differentiate the effects of natural aerosol variation from noise introduced by scene conditions (e.g., sub-visual cirrus). Standard data products available from the AERONET Cimel sunphotometers are produced from measurements with less angular resolution (Figure 1b). The ASD simultaneously reports radiance in 2151 spectral channels between 350 and 2500 nm for each angular position. For each radiance scan, a full set of aerosol properties in this wavelength range (size distribution, single scattering albedo and phase function) will be retrieved using the techniques outlined in Ricchiazzi et al. (2006). These observations will provide detailed information on aerosol properties at spectral and temporal scales that are not available with current instruments.

The retrieval of aerosol properties is sensitive to situational variables such as aerosol amount, the surface albedo, atmospheric pressure, and the presence of sub-visual cirrus cloud. In order to constrain these variables for our retrievals, we will make measurement near the Boulder Baseline Surface Radiation Network (BSRN) site, located at 40o 03’ N and 105o 00’ W. Continuous measurements at the site include solar downwelling global, direct, and diffuse radiation at the surface at the top a 300 m tower as well as solar upwelling global radiation at the top of the tower at 1 to 3 minute resolution. Having good estimates of aerosol optical depth from the Boulder BSRN observations will greatly reduce the need for precise absolute calibration of the ASD-FR. The main requirement for good retrievals of single scattering albedo and phase function will be an accurate relative calibration of the instrument as the gain is adjusted (either electronically or through the use of accurately characterized apertures) for viewing toward the sun versus viewing the sky. The BSRN direct and diffuse radiation observations will also be used to test closure by comparing modeled and observed radiation with the aerosol properties retrieved from the ASG. These tests will provide a good estimate of the accuracy of the retrievals.

Expected outcome and impact. These measurements will represent an unprecedented data set of spectrally resolved aerosol properties and a proven concept for dramatically more accurate and comprehensive measurements of these properties. In this set of data we will have the wavelength dependence of the asymmetry parameter and single scattering albedo across the entire shortwave spectrum for a variety of aerosol types, information that does not currently exist. Modeling studies that attempt to calculate the radiative impacts of aerosol on climate have had to use a ‘best guess’ at the dependence and much variation exists in how this is handled in different modeling studies. This information in itself will have a significant impact on the uncertainties associated with modeling the impact of aerosols on climate, and in turn, the resultant journal papers are expected to be widely cited. Finally, we will be able to assess the requirements for developing a robust and autonomous instrument using the ASD and goniometer configuration for continuous operation in the field.

References

Painter, T.H., B. Paden, and J. Dozier. Automated Spectro-Goniometer: a spherical robot for the measurement of the bi-directional reflectance of snow. Reviews of Scientific Instruments, 74(12), 5179-5188, 2003.

Ricchiazzi, P., C. Gautier, J. Ogren, and B. Schmid. 2006: A comparison of aerosol optical properties obtained from in-situ measurements and retrieved from sun and sky radiance observations during the May 2003 ARM aerosol intensive observation period. Journal of Geophysical Research, 111, D05S06, doi:10.1029/2005JD005863, 2006.

Wang, M. and H. Gordon. Estimating aerosol optical properties over the oceans with the multiangle imaging spectroradiometer: Some preliminary studies. Applied Optics, 33, 4042-4057,1994.

Figure 1 (a, above) The Automated Spectro-Goniometer (ASG) deployed for measurements of directional reflectance of snow. (b, below) Cimel almucantar (diamonds) and principal plane (squares) observations of radiance at 676 nm obtained at the ARM program Southern Great Plain facility on 29May2003. These data were used as part of a study to test for closure between in-situ and remotely sensed aerosol retrievals (Ricchiazzi et al., 2006).