Ocean Optics Spectrometers Measure Spectral Radiance Aerodyne Research, Inc. of Billerica, Mass., is a developer […]
Ocean Optics Spectrometers Measure Spectral Radiance
Aerodyne Research, Inc. of Billerica, Mass., is a developer and supplier of products and technology for environmental quality, remote sensing and optical appearance design and prediction.
Aerodyne is evaluating Ocean Optics spectrometers as part of a system that characterizes clouds. As the work continues, Herman Scott, the company’s executive vice president, described for us in this article the significance of cloud optical depth, new approaches to characterizing clouds, and the challenges of climate modelling:
Measuring Cloud Optical Depth
Cloud optical depth (COD) is one of several optical properties used to characterize clouds. The underlying cloud microphysical properties of primary importance are the effective radius of water droplets or ice crystals, the droplet concentration, and the thermodynamic phase (liquid or ice). As shown in Figures 1 and 2, COD is defined and pictured as the path-integrated extinction coefficient for solar photons along a vertical line of sight through the cloud. As the droplet concentration increases, a photon undergoes more scattering events and the cloud optical depth increases.
Figure 1. A simple one-dimensional model of clouds is assumed in the definition of cloud optical depth. A commonly used Atmospheric Transmission Model is MODTRAN, which incorporates DISORT (DIScrete Ordinate Radiative Transport). The Earth’s albedo is upwelling and must be taken into account since some fraction will be reflected (scattered) off the bottom side of the cloud back into the COD sensor’s field of view.
Figure 2. This notional plot shows how spectral radiance depends upon COD and illustrates the thick/thin cloud ambiguity.
In recent years Aerodyne Research has built and field tested several variations of a quantitative COD sensor or spectroradiometer, all designed around a modern compact grating spectrometer (CGS), specifically the Ocean Optics USB2000+ and STS-VIS models. This COD sensor is dubbed TWST for Three Waveband Spectrally-agile Technique. TWST is a calibrated spectroradiometer that stares at a narrow segment (0.5 degrees) of the sky directly overhead, recording the spectral radiance in the visible wavelength regime at 2-8 nm spectral resolution. Sample spectra for a range of COD values are plotted in Figure 3.
Figure 3. Sample spectra measured by TWST with the three spectral factors currently used in the COD retrieval algorithm identified. The solar zenith angle (SZA) was 65°.
A significant and well-known difficulty in the passive remote sensing of cloud optical depth, the so-called thick/thin cloud ambiguity, is clearly illustrated in Figure 2. The relationship between spectral radiance and COD is two-valued; one, marked OD1, is in the optically thin region where the brightness increases with increasing COD, and the other, marked OD2, is in the optically thick region where the brightness decreases with increasing COD. This ambiguity in COD is the principal complication inherent in the spectral radiance method of measuring COD. Using a spectroradiometer rather than a filter band radiometer enables TWST to measure the equivalent width of the 760 nm oxygen A-band in order to resolve this ambiguity (Niple 20141; Niple et al. 20162). The equivalent width is a monotonic function of the photon total pathlength (and thus also of the COD) and does not suffer from the ambiguity.
About the TWST Sensor
The TWST COD sensor is designed to be robust and transportable for field deployments while providing the user with accurate, real time COD values at a 1 Hz data logging rate. The heart of the sensor consists of a single fiber optic spectroradiometer with an entrance aperture that is well-baffled from the direct sun. Dark spectra collection is done automatically with a conventional shutter. This sensor is unique in providing high temporal resolution (up to 10 Hz sampling rate), high spatial resolution (0.5 degrees), spectral agility and high spectral resolution (2-8 nm), with a typical signal-to-noise-ratio (SNR) >1000:1 for 400 co-added spectra at a sampling rate of 1 Hz. Candid photographs of several successful TWST field deployments are shown in Figures 4, 5 and 6 with various versions of the weatherproof enclosure.
Figure 4. Two co-aligned COD sensors are deployed for reproducibility testing.
Figure 5. A COD sensor with reflective cover for thermal control is deployed in Australia.
Figure 6. The COD sensor is deployed at Hyytiälä, Finland, in 2014 as a guest on the Department of Energy Atmospheric Radiation Measurements campaign named Biogenic Aerosols Effects on Climate and Clouds.
The Challenge – Real 3D Clouds and Global Climate Models
Real 3D clouds and the diversity of cloud types present a much more difficult problem than represented by the 1D model shown in Figure 1. Consider the many scattering pathways for a photon to arrive in the field of view of the sensor (Figure 7). Forward modeling of the 3D radiative transfer occurring in real clouds becomes impractical.
Figure 7. This illustration shows the complexity imposed by 3D clouds in understanding the radiative transfer between the sun and the Earth.
One alternative is to resort to highly computational intensive numerical models. A good example is the NASA managed I3RC (Intercomparison of 3D Radiation Codes) Community Model. Such 3D RT models are so computationally intensive that they are run typically by only the leading experts in the world. Another alternative is to turn to one of the efforts to develop approximations that perform 3D RT computations. One recent example is SPARTACUS (Schafer et al. 20163; Hogan et al. 20164). Usually these approximations are designed to perform well for a limited regime or range of cases.
The point to be made here is this: In scientists’ efforts to understand and predict climate change, the results from these models, as well as global data collection networks (Holben et al. 19985; Chiu et al. 20106) covering time spans of decades, provide the inputs to our Global Climate Models (GCMs). We understand that clouds are a major player in regulating the temperature of the Earth, serving as a thermal blanket and preventing the long wavelength radiation from escaping into space (Figure 9). Clearly, the measurement and modeling of cloud effects are a major factor in our understanding of climate and our capability to predict changes. One would think that we need only to turn to the world’s impressive array of space-based sensors for the global data needed to validate models and predict climate changes. However, not all of the news is encouraging on that front. Let us refer to some authoritative sources.
Lauer and Hamilton have objectively compared cloud modeling results from two generations of GCMs in the Coupled Model Intercomparison Project (CMIP3 and CMIP5, each covering 20 years of model data) and also compared each of these to 20 years of satellite data7. In addition to finding no substantial improvements in the predictive performance of the later generation of models, the Lauer and Hamilton finding of most interest to us is that the grid size for the GCMs and the current satellite-based sensor spatial resolution is too large to correctly capture cloud microphysical properties on a global scale. This issue (among others) for space-based cloud sensors tells us there remains for the time being a real need for networks of ground-based cloud sensors such as TWST. Those of us still working to improve ground-based sensors have not yet been made irrelevant by our vast array of space-based sensors.
A confirmation of this continuing requirement for ground-based remote sensing is emerging indirectly from metrology standards committees, such as the NIST Committee on Radiation Measurements (CORM). CORM notes that the science underlying climate change is truly complex, involving many interrelated systems. To build a predictive model that really works will require several decades of data with less than 2% absolute uncertainty. That is a daunting Grand Challenge.
TWST COD Sensor — Encouraged by Success, New Developments and Real Requirements
Aerodyne has been very pleased with the performance of our Ocean Optics spectrometers. We have monitored the radiometric stability of specific units over periods of months to years and found the calibrations to be repeatable within a few percent, well within the stability of our calibration source.
Figure 8. This is a comparison of Cloud Optical Depth as measured by TWST and a Microwave Radiometer (MWR). Since the MWR data logging rate was one point per 5 minutes, the plot includes 5 minute rolling averages of the TWST data as well as the typical 1 Hz record.
Aerodyne purposely designed the TWST COD sensor around a spectrometer as opposed to a multiband radiometer. We have realized these advantages as a result:
- TWST uses a spectrometer in place of a multiband radiometer to record 2048 spectral channels simultaneously. This gives TWST the spectral agility to use any band as the primary band for determining the cloud optical depth. In addition, it allows TWST to measure the equivalent width of the 760 nm oxygen A-band and thereby to resolve the COD ambiguity. The equivalent width is a monotonic function of the photon total pathlength and of the COD. Thus it does not suffer from the COD ambiguity (Figure 2).
- TWST has high temporal resolution, 1 second or less versus 90 seconds for the comparable filter band radiometer. The TWST spectrometer, which sees all spectral resolution elements simultaneously, provides a valuable multiplex advantage. As a result TWST can collect samples at least one to two orders of magnitude faster without sacrificing signal-to-noise-ratio (SNR) and thereby better capture the fast evolution of cloud properties. To make the TWST temporal advantage specific, we point out here that for each TWST data point (at 1 Hz sample rate), the radiance values collected at 440 nm and 870 nm are truly simultaneous whereas there are several seconds of delay between the radiance values collected for the same two bands in the radiometer. This lack of simultaneity can introduce substantial unwanted errors into the COD measurements for rapidly evolving clouds.
- TWST has high spatial resolution of 0.5 degrees. This reduces clear sky background contamination and promotes the study of cloud edges where cloud-aerosol interactions are an important effect.
- TWST has high spectral resolution, approximately 300 spectral resolution elements versus 6 spectral bands in a typical radiometer. Again, the spectrometer puts more spectral information into the hands of the analyst, leading to higher sensitivity and enabling the extraction of more precise COD information.
Researchers are finding more value in using calibrated spectrometers in place of filter band radiometers and are now taking advantage of the entire spectral content. LeBlanc et al. 20158 have demonstrated the capability to retrieve effective droplet radius and thermodynamic phase in addition to COD by extending the spectral coverage beyond the visible into the NIR regime. Aerodyne is now moving forward to add this same NIR extension to TWST in order to retrieve additional cloud microphysical properties and thereby enhance the value of TWST.
Merlin et al. 20169 have demonstrated the retrieval of cloud geometrical thickness and cloud top altitude using multi-angular oxygen A-band measurements from above. These many applications of spectrometers to remote sensing of cloud properties are exciting developments that will most certainly encourage researchers to take advantage of affordable and field-worthy COTS compact grating spectrometers such as those offered by Ocean Optics.
As to the future of our TWST sensor, in addition to continuing applications in field campaigns, we envision the possibility of a growing network of users in the area of education and training tools. Given its affordability and ease of use, TWST would be an excellent choice as a training tool in spectroscopy for any environmental sciences department in high school, college and university. With such a network of users, TWST data could be uploaded to cloud storage sites such as a Dropbox or Google Drive on a daily basis. From there the data would be processed to create a global picture of cloud properties.
Please see reference Niple et al. 20162 for more TWST details and visit www.aerodyne.com for more information about the extensive projects in aerosol and cloud chemistry at Aerodyne Research. Aerodyne’s business model is based strongly upon collaboration with our instrument customers. That collaboration enables Aerodyne scientists to publish as lead author or participating author upwards of 40 refereed journal articles per year.
Figure 9. Cloud effects on Earth’s radiation.
1 Niple, E.R., 2014. Application of Oxygen A-band Equivalent Width for Cloud Optical Depth Measurement, 2014 American Meteorological Society Annual Meeting Sixth Symposium on Aerosol-Cloud-Climate Interactions, February 3, 2014 Atlanta, Georgia.
2 Niple, E. R., Scott, H. E., Conant, J. A., Jones, S. H., Iannarilli, F. J., and Pereira, W. E. Application of oxygen A-band equivalent width to disambiguate downwelling radiances for cloud optical depth measurement, Atmos. Meas. Tech., 9, 4167-4179, doi:10.5194/amt-9-4167-2016, 2016.
3 Schäfer, S.A.K., R.J.Hogan, C. Klinger, J. C. Chiu, and B. Mayer (2016), Representing 3-D cloud radiation effects in two-stream schemes: 1. Longwave considerations and effective cloud edge length, J. Geophys. Res. Atmos., 121, 8567–8582, doi:10.1002/2016JD024876.
4 Hogan, R.J., Schäfer, S.A., Klinger, C., Chiu, J.C. and Mayer, B., 2016. Representing 3D cloud-radiation effects in two-stream schemes: 2. Matrix formulation and broadband evaluation, J. Geophys. Res. Atmos., 121, 8583–8599, doi:10.1002/2016JD024875.
5 Holben, B.N., T.F. Eck, I. Slutsker, D. Tanré, J.P. Buis, A. Setzer, E. Vermote, J.A. Reagan, Y.J. Kaufman, T. Nakajima, F. Lavenu, I. Janowiak, and A. Smirnov, 1998. AERONET—A federated instrument network and data archive for aerosol characterization, Remote Sensing of the Environment, 66, 1-16.z
6 Chiu, J.C., C-H. Huang, A. Marshak, I. Slutsker, D.M. Giles, B.N. Holben, Y. Knyazikhin, and W.J. Wiscombe, 2010. Cloud optical depth retrievals from the Aerosol Robotic Network (AERONET) cloud mode observations, J. Geophys. Res. Atmos. 115, (D14202), doi:10.1029 /2009JD013121.
7 Lauer, A. and Hamilton, K., 2013. Simulating clouds with global climate models: A comparison of CMIP5 results with CMIP3 and satellite data. Journal of Climate, 26(11), pp.3823-3845.
8 LeBlanc, S.E., Pilewskie, P., Schmidt, K.S. and Coddington, O., 2015. A spectral method for discriminating thermodynamic phase and retrieving cloud optical thickness and effective radius using transmitted solar radiance spectra. Atmospheric Measurement Techniques, 8(3), pp. 1361-1383.
9 Merlin, G., Riedi, J., Labonnote, L.C., Cornet, C., Davis, A.B., Dubuisson, P., Desmons, M., Ferlay, N. and Parol, F., 2016. Cloud information content analysis of multi-angular measurements in the oxygen A-band: application to 3MI and MSPI. Atmospheric Measurement Techniques, 9(10), pp.4977-4995.