Happy New Year Everyone. I look forward to ‘99 particularly as we have a launch date of July 15 scheduled for EOS AM-1. This edition I would like to acknowledge my co-workers on the project who have put in an enormous amount of time to make AERONET work. I’ll discuss some of the operational issues including deployment schedules and a statement on how to join AERONET. We have an update on calibration issues including filter degradation, calibration on the GSFC 6 ft. sphere and our polarization cal. capability. I’ll briefly discuss our relationships to other networks, field programs such as INDOEX, LBA, and other collaborations. We have several relatively new issues that we are pursuing including support for a receive site in Hawaii for acquisition of data through GMS and an initiative to develop an internet access to AERONET and satellite data and other GIS data bases that will follow a Mapped Objects format. Finally we’ll wrap this up with some research activities from within our group and immediate collaborators.

Laura East is AERONET's part time office assistant and property manager. She is an important link to the outside world by organizing and e-mailing the instrument checks to the site managers, provides information to the many inquiries we receive particularly with respect to international agreements, shipping equipment and purchases.

Tom Eck and I have worked closely together in sun photometry since the mid 80's. Tom provides careful Langley and intercomparison calibration for all instruments in the network and, along with Sasha Smirnov, weekly checks and summary of every instrument operating in the network. Tom is developing a bootstrap (PAR) solar flux network as an add on to AERONET and is a CO-I on the LBA program. Tom is very successfully researching and publishing the data collected including studies on single scattering albedo and spectral dependence of AOD.

Alexander (Sasha) Smirnov is also playing the dual role of operations and research. Sasha is trained and experienced in sun photometry and provides a very critical evaluation of the instruments health and data quality through the weekly checks. Sasha is responsible decides whether an instrument meets our standards for deployment and if field data collected can be raised to our quality assured level 2 data base. Sasha developed the automatic QA algorithms and has written a paper on it which is currently under review and available on our homepage. Sasha is researching the data leading the effort on AOD climatology, optical properties of air masses and has participated in the ACE-2 and TARFOX experiments.

Wayne Newcomb handles all of the instruments providing roof top intercalibrations, sphere calibrations, fixes broken instruments, offers telephonic and email solutions to field instrument problems, daily checks the troubleshooting reports, supplies the Lab, and ships the instruments back to the field. I should note that our intercalibration procedure requires three calibrations, an ‘as is’ post field calibration, window cleaning calibration and deployment calibration. Considering we have a six month rotation on instruments and approximately 100 instruments in the network this represents a huge complicated effort.

Nader Abuhassen has so far delayed his research career to provide upgrades and electronic, software, and optical repairs to the sun sky radiometer system. His experience developing the cimel and engineering background has been fundamental to the operation of the network. Over the last year and a half Nader has made numerous design changes to the system that are transparent to the rest of us but has resulted in improved accuracy and much less down time to the field instruments. Nader has developed measurement protocols for the Cimel, status checks for the home page and has developed an interface for the solar flux network.

Ilya Slutsker has provided an increasingly complex series of programs to ingest, process, archive and retrieve the AERONET data base. This has evolved from original Mac type mouse driven menu of DEMONSTRAT, literally weekly upgrades and improvements since 1993 to the functionality of the current homepage. What is not apparent are the numerous programs, to retrieve data from GOES, EUMETSAT, individual downloads, special calibration programs and histories, tracking instrument and platform ID's, and development of special programs for our research team. He also says we won't have an YK2 bug.

Oleg Dubovik is our inversion specialist. We've felt from the inception of this project that the sky radiance measurements would be the most significant observational feature of AERONET and indeed the use of the Nakajima inversion code has provided significant steps forward in analysis of aerosol optical properties derived from the network measurements. Oleg has spent the last 1.5 years developing a comprehensive inversion code and sensitivity analysis that provides a physically meaningful extension of the Nakajima size distributions from 0.03 to 15 micron radii and including retrievals of the complex index of refraction for very high aerosol loading. These products will be evaluated and validated by Oleg, the research team and in collaboration with the research community.

Joel Schafer has worked with AERONET data for several years in association with the BOREAS program. Joel stepped in last year to help on many operational phases of the program while he waits for LBA funding to begin. Joel has provided a burst of energy to organized the lab, deploy instruments in all types of locations and weather and his broad experience with instruments and scientific expertise have provide links for improved communication between the scientists and lab folks. Joel has has been available for emergency travel, most recently a two week trip to the Maldives and is scheduled for a trip to the Amazon in mid January.

Norm O’Neill is on a one year sabbatical from Sherbrooke University in Quebec. Norm’s interests are radiative transfer and their remote sensing implications in the atmosphere, in coastal waters and over forested terrain. He is a professor in the Geography Dept. and is the head of AEROCAN, the Canadian branch of AERONET. He is currently working on regional climate models to estimate spatial and temporal distributions of aerosol optical properties which he hopes to tie to ground based measurements and satellite observations. Norm has made significant contributions to our research and operational program.

Brian Montgomery will work with the AERONET team over the next two years to develop a GIS mapped object data base that we expect will allow fast a comprehensive interface of AERONET data to satellite and modeled data bases over the internet. I feel this will be particularly important for Satellite validation efforts.

Tony Caporale works part time for AERONET developing international agreements which includes initiating contacts with the proper officials, providing explanations and suggestions for modifications as necessary. Tony reviews their comments and when he's satisfied the legal eagles at HQ will approve the agreement he recommends that HQ will send an official agreement to the host officials.

Although not apart of the AERONET paid staff, the following folks are instrumental to the continued development of the program.

DeVon Carroll at NASA HQ International Relations is responsible for getting the official documents through NASA and out to the host country. AERONET is only a small part of both Tony and DeVon's work load.

Gina Baldessari as Resource analyst for Code 923 tracks AERONET's budget, writes PRs and grants, computes manpower taxes, and handles the system's financial forecast requirements and budget reviews.

Darrel Williams, Branch head for Code 923, has given a great deal of Branch support for the AERONET program. He's provided visiting scientist support, found space for the staff of a growing program and provided advice and direction for the sustained growth of the program.

Michalel King and Robert Curran have recognized the need for such measurements and have provided the funds for AERONET based on peer reviewed proposals. They have provided scientific and programmatic guidance.

Finally the MODIS discipline team members Yoram Kaufman (atmospheres), Chris Justice (Land) and Wayne Esaias (Ocean) have continued their support for the program and have worked to integrate AERONET into their research and operational activities.

We've received the polarizer and it is operational. We will now calibrate and track the polarization characteristics of each polarization instrument. Since this is new (as of Jan 7), we are not sure of the time required nor the accuracy of the measurements required and we'll have to establish the tracking procedure and appropriate flags.

The GSFC calibration facility is managed by Peter Able who has recently had the 6 foot integrating sphere painted and recalibrated. The brightness has been increased by ~30%.

We regress the cimel spectral voltages (440, 675, 870 and 1020 nm) against lamp levels of 0, 1, 2, 4, 8, 12 and 16 lamps. Our experience is an absolutely linear relationship for all instruments. All calibrations are evaluated for linearity prior to acceptance and application.

The reference instruments, using Mauna Loa calibrations, are continuing to be cycled on a ~triweekly basis. Tom offers the latest info:

The percentage changes in Vo calibration coefficients for Cimel #37 from Mauna Loa Observatory (MLO) langleys made Sep. 1998 to Dec. 1998 (~3 months) are as follows:

Vo % change: ((Sep-Dec)/Sep))*100

Negative denotes higher Vo in December. Most channels showed insignificant changes in Vo (smaller than Vo uncertainty as defined by MLO repeatability), except the 380 and 940 nm filters (the 340 filter was changed). Since the 380 nm filter is adjacent to the 340 nm filter in the filter wheel, perhaps this filter was perturbed when the 340 filter was replaced on October 27, 1998. Note that the Langley Vo value variability at MLO for the 940 nm filter is much greater than other filters, so the change of -4.91% is probably not all due to filter changes given the Vo coefficient of variation at MLO for the 940 filter is ~3.5% to 4.2%.

In general, the Vo stability of most filters (Ion Assisted Deposition (IAD) interference filters) in the network has been very good over the last year (within ~1 % or better over a year). However there are 2 exceptions, the 340 and 670 nm filters. These both have shown degradation with time, with the Vo increasing with time for the 340 nm filter (becoming more transmitting with time) and the 670 nm Vo value decreasing with time. A typical example follows:

Cimel #99 - This instrument was at the CART site in Oklahoma with pre - deployment calibration done on Oct 6, 1997 and post deployment calibration on November 22, 1998. The Vo values were very close between the two calibrations: ~0.5-1.5% except for 675 nm (~6% lower) and 340 nm (~ 14.5% higher). Thus the average rate of filter degradation for 670 nm was ~0.5% per month and was ~1.1% per month for 340 nm filter.

We have recently installed 340 nm and 670 nm filters from the newest shipments of Barr filters in some instruments and we will be monitoring their stability in time.

NO2 absorption is not considered in our processing and can affect the 340, 380 and 440 nm channels. It is most likely to affect instruments in polluted urban environments, however, since the amount of NO2 is not known we've chosen so far to ignore it. Norm O'Neill provides the following discussion and suggestions.

Glen Shaw pointed out in 1976 that NO₂ optical depths at the wavelength of maximum absorption (390 nm) could vary between 0.008 for clean Arctic air to 0.087 for a polluted urban atmosphere. He clearly warned at the time that NO2 induced absorption in the ultraviolet, blue and green regions of the spectrum could significantly influence the calibration and application of multi-wavelength sun photometry.

The NO2 broadband continuum extends from 325 to 480 nm (half peak absorption coefficient magnitude) with a typical cross section ~ 5e-19 cm^2. This yields an (STP) volume absorption coefficient of 5e-19 * NSTP ~ 13 cm^-1 (NSTP = 2.55e19 cm^3). The range of NO2 abundance values corresponding to Shaw's extrema above was 0.4 to 5.0 milli-cm (total NO2 number density abundance compressed to an equivalent height at STP). The corresponding optical depth extrema for any channel between 325 and 480 nm is accordingly 0.005 and 0.065.

One of us (Norm O'Neill) verified Shaw's observations for a sampling of summer days in a moderately polluted urban atmosphere (York University, 20 km north of downtown Toronto, Canada*). Equivalent heights of NO₂ abundance acquired during the sampling period varied between 0.5 and 2.5 milli-cm. In terms of Langley calibration it was found that NO2 varied as the classical (Shaw) parabola to yield significant regression line shifts (between 1 and 5 % for a variety of days and wavelengths between 400 and 500 nm) which were all but undectable by sunphotometry analysis alone. In terms of the application of sunphotometry to aerosol analysis it was found, for example, that significant (aerosol turbidity dependent) increases in the Angstrom coefficient could occur (0.1 to 0.4) for a wavelength span of 400 to 1000 nm.

The ideal solution to these problems is to measure NO₂ optical depth independently and correct sunphotometer derived optical depths for the NO₂ contribution. The simplest solution is to subtract the NO₂ optical depth from all UV-blue-green channels for a realistic background value (~ 1 milli-cm) and hence to roughly divide errors in half.

NO₂ detection incorporated into AERONET

Three-band Brewer spectrophotometers have successfully been used to measure NO₂ absorption (0437.7, 444.8 and 450.0 nm channels of 0.3 nm bandwidth, Brewer et al., 1973, Nature, Vol. 246, pp. 129-133). These signals can be intelligently processed to minimize the effects of common channel radiometric and spectral shifts (in more recent versions, 5 channels have been employed to perform such processing). While it is not clear that a network sun photometer could operationally include an NO₂ detection capability the possibility certainly warrants consideration since this gas is a highly variable and important indicator of atmospheric pollution chemistry. The added value to AERONET would be significant both in terms of the capability of monitoring NO₂ as well providing a means of discriminating and correcting for NO₂ influences on aerosol optical depth.

One must however keep in mind that the objectives of monitoring NO₂ for its own sake as opposed to monitoring NO₂ as a correction to aerosol sun photometry are different and that this difference can have a significant design impact. In the first case, one would aim for precisions and accuracies better than say 1/20 of the natural variation of NO₂ (0.25 milli-cm or ~ 0.0033 OD) while in the latter case accuracies ~ 1/5 the natural variation might be satisfactory given the typical sun photometer target of 0.01 in OD (i.e. 1 milli-cm or ~ 0.013 OD).

We have a new server called AERONET which replaced SPAMER on November 30. The homepage address is http://AERONET.gsfc.nasa.gov/

We are now using 20 watt solar panels for the transmitter greatly reducing power related problems.

For the folks handling instruments please inform Ilya when an instrument is moved and please don’t change the integration times.

The Cimel software is being modified to make access to the integration times more difficult. Cimel is also changing the filter wheel positioning during triplets to protect certain filters from excess exposure. More details will follow on this upgrade when the final version is complete.

We have begun washing the outer opical windows of all instruments prior to deployment. We ask that site managers in areas where deposition on optics is likely to wash the optical windows weekly during field monitoring and to keep a log of this maintenance. Thus our post calibration procedure is the following: cal as is, clean windows and recal, maintain/repair instruments, wash windows again and make deployment calibration. It's a great deal of work and evaluation but we think it is necessary to achieve our accuracy goals.

The AERONET group will move to the new ESSB at Goddard the week of March 23. No data will be lost but data access will be interrupted for several days and all activities will be disrupted during and around that week.

The six foot integrating sphere at GSFC will move to the new Earth System Science Building (ESSB) scheduled for February 11 which will impact our operation. Sometime after that it will again be down for a final repainting.

International agreements have been completed for sites in Brazil, Mexico, Canada, Surinam and France!. Pending agreements are for Chile, Bolivia, Argentina, Belarus, Moldova, Maldives Islands, Botswana, Mozambique, Poland, Barbados, and Guadeloupe.

Periodically I’m asked how to join AERONET. The process is simple. Buy a CIMEL polarized or non polarized sun and sky scanning spectral radiometer and Vitel DCP (transmitter). Provide for it’s onsite maintenance, agree that all data collected will be available in the AERONET public domain data base (instrument PI’s have authorship rights on all data published). AERONET will provide instrument ID’s, some maintenance, calibration and characterization, processing, cloud screening and QA, and homepage access of the data. Aeronet will also provide, when possible, a swapout instrument during the semiannual recalibration of instruments.

It’s a good deal for everyone but as our resources are taken, the unlimited growth will have to stop. I feel have not reached that point but we are getting close.

AERONET has many collaborative partners including AEROCAN (Canada), AEROBRAS (Brazil), PHOTON (France), and strong affiliations with SIMBIOS, EOS validation, NRL, CSIRO as well as many other scientists and institutions. Other networks with similar types of measurements include, SKYNET (Japanese Prede instrument), USDA UV network (MFRSR), Forgan network (Australia), and Michaelsky network (shadowband). Links to these may be viewed from the AERONET homepage.

INDOEX- radiative forcing Indian ocean- two instruments AERONET, 3 inst. PHOTON

LBA-Amazon biosphere atmosphere interactions, LBA funded 2 inst. 4 inst. AERONET

EOPACE-coastal marine aerosols LBA two instruments

SAFARI2000-EOS validation, regional aerosol studies, 5 inst. AERONET, 1 Photon

Design: Modular

Spectral: 340/2, 380/4, 440/10, 500/10, 610/10, 675/10, 778/10, 870/10 , 940/10, 1030/10, 1640/25, 2130/50, linear polarization @870

Detectors: UV: AS-N; Vis & NIR: Si; SWIR: InGaS. Filters integrated with detectors on large thermal inertia plate

Measurements: Spectrally Instantaneous. sun, aureole, sky and dark current protocol. BRDF protocol option.

Ancillary data: Temperature, air pressure, and GPS

Communication: TBD: Integrated DCS (current system), cell phone, 2 way satellite phone.

Environmental Range: All environments including marine and arctic

Power: Solar panels & batteries with AC option

Commercialized Cost: $20 to 25 K

Much of the focus of the AERONET data is for satellite validation. We have taken the initiative to explore matching up data bases through internet GIS systems. As such our georeferenced data base as well as selected satellite data bases will be accessed through a ‘mapped objects’ format for what we hope will be a seamless extraction of coincident data bases for analysis.

Oleg has completed drafts of his inversion and sensitivity papers which appear on our home page. A beta version is currently being installed on demonstrat and the AERONET homepage which should be available in 1 or 2 months. The inversion products will include, size distribution (volume or number option), phase function, asymmetry factor, single scattering albedo, scattering AOD, real and imaginary index of refraction, total spectral downwelling flux, spectral up-welling flux, spectral direct normal downwelling flux. Uncertainty estimates will appear on all points.

The Nakajima inversion code products will continue to be presented and upgraded as new versions are received.

The abstract from his sensitivity analysis is in the research section.

The following data sets were quality assured using the automatic Smirnov cloud screening and manual assessment:

Several intercomparisons between Cimel, Prede and MFRSRs are underway with various collaborators however no results are ready to report.

O. Dubovik (1,2), A. Smirnov (1,2), B.N. Holben (1), M.D. King (3), T.F. Eck (1,4), I .Slutsker (1,2)

1. Biospheric Sciences Branch, code 923, NASA/GSFC, Greenbelt, MD 20771, USA

2. Also at Science Systems and Applications, Inc., Lanham, MD 29706, USA

3. Earth Science Directorate, code 900, NASA/GSFC , MD 20771, USA

4. Also at Raytheon STX, Lanham, MD 20771, USA

Abstract. Sensitivity studies regarding aerosol optical property retrieval from radiances measured by ground based Sun - sky scanning radiometers of the AErosol RObotic NETwork (AERONET) are conducted. The studies are based on the inversion algorithm for retrieving the aerosol size distribution and complex refractive index together with aerosol phase function and single scattering albedo from the spectral measurements of direct and diffuse radiation. The perturbations of the inversion resulting from random errors, possible instrumental offsets and known uncertainties in the atmospheric radiation model employed are analyzed. The Sun or sky channel miscalibration, inaccurate azimuth angle pointing during sky radiance measurements and inaccuracy in accounting for ground reflectance are considered as errors sources. The effects of these errors on the characterization of three typical and optically distinct with aerosols bi-modal size distributions (weakly absorbing water soluble aerosol, absorbing bio mass burning aerosol and desert dust) are considered. The aerosol particles are assumed in the retrieval to be polydisperse homogeneous spheres that have the same complex refractive index. Therefore, it is examined how the inversions with such an assumption may mislead us in the case of non-spherical dust aerosols and in the case of externally or internally mixed spherical particles with different refractive indices.

The analysis shows that the retrievals give useful information about aerosol optical properties even in presence of rather strong systematic or random uncertainties in the measured data. Quantitative estimates of the retrieval accuracy are given. The major concern relates to the characterization of low optical depth situations, where the high relative errors may occur in the direct radiation measurements of aerosol optical depth. Significant problems may also appear with the characterization of dust, where accurate azimuth angle pointing is critical. Scattering by non-spherical dust particles requires special analysis, whereby approximation of the aerosol by spheres allows us derive accurate size distributions by inverting sky-radiance measured in first forty degrees scattering angle only, where non-spherical effects are minor.

The inversion is designed as a search for the best fit of measured radiances by a theoretical model. It is shown that the value of the smallest obtained residual is sensitive to both the presence of experimental errors and failure of the radiative model. Therefore, this residual value can be adopted as an indicator of the quality of the retrieval.