the 2010 aop workshop summary report - nasa › archive › nasa › casi.ntrs.nasa... · the 2010...

March 2011

NASA/TM–2011–215871

The 2010 AOP Workshop Summary Report

Stanford B. Hooker, John H. Morrow, James W. Brown, and Elaine R. Firestone

https://ntrs.nasa.gov/search.jsp?R=20110008270 2020-07-16T09:15:28+00:00Z

��

�� !��"�� #��$�#��

�� #%�� #�&��'��(�� )�� *�� (�� #��&��*��*�� #��+

, �-'.�'�%�/0%'�� &��1�� 2 ��3��#�� (��2��*�� 2�� 3��

, �-'.�'�%�4-4�&��)/4 �� #��5�� 6��$��*��$�� )��3��#��

, '��&�'��&�&-�&� ��2�� #��2��

, '��7-&-�'-�/0%'�� '��#�� #��

, �-'�%�/0%'�� 1��*��21��

, �-'.�'�%��&��%�� -�� 2�� 2��(��

��5��2 ��(�� 5�� 5��

7�� *�� +

, �� .�� +88*** �� 8��2��

, -2��#��6��9��

, 7�3�#��6��.��)��$��!443) 757-5803

, ��.��)��$��!443"�757-5802

, :��+��.��)��$��'��

7115��)��.��4)�;<=>?

National Aeronautics and

Space Administration

Goddard Space Flight Center

Greenbelt, Maryland 20771

March 2011

Stanford B. HookerNASA Goddard Space Flight Center, Greenbelt, Maryland

John H. MorrowBiospherical Instruments Inc., San Diego, California

James W. BrownCSTARS University of Miami, Miami, Florida

Elaine R. FirestoneScience Systems and Applications, Inc., Lanham, Maryland

NASA/TM–2011–215871


Available from:

NASA Center for AeroSpace Information National Technical Information Service

7115 Standard Drive 5285 Port Royal Road

��

S.B. Hooker, J.H. Morrow, J.W. Brown, and E.R. Firestone

Abstract

The rationale behind the current workshop, which was hosted by Biospherical Instruments Inc. (BSI), was toupdate the community and get community input with respect to the following: topics not addressed duringthe first workshop, specifically the processing of above-water apparent optical property (AOP) data within theProcessing of Radiometric Observations of Seawater using Information Technologies (PROSIT) architecture;PROSIT data processing issues that have developed or tasks that have been completed, since the first workshop;and NASA instrumentation developments, both above- and in-water, that are relevant to both workshopsand next-generation mission planning. The workshop emphasized presentations on new AOP instrumentation,desired and required features for processing above-water measurements of the AOPs of seawater, working groupdiscussions, and a community update for the in-water data processing already present in PROSIT. The sixworking groups were organized as follows: a) data ingest and data products; b) required and desired features foroptically shallow and optically deep waters; c) contamination rejection (clouds), corrections, and data filtering;d) sun photometry and polarimetry; e) instrumentation networks; and f) hyperspectral versus fixed-wavelengthsensors. The instrumentation networks working group was intended to provide more detailed information aboutdesired and required features of autonomous sampling systems. Plenary discussions produced a number ofrecommendations for evolving and documenting PROSIT.

1. INTRODUCTIONThe processing of data collected to measure the appar-

ent optical properties (AOPs) of seawater is a fundamentalpart of the calibration and validation of ocean color satel-lite missions, because the data products are directly com-parable to the spaceborne measurement: the spectral radi-ance emerging from the sea surface or the so-called water-leaving radiance, LW (λ), where λ denotes wavelength. Theartificial variance imparted by processor-to-processor dif-ferences in estimating LW (λ) frequently equals or exceedsthe total uncertainty permitted in calibration and valida-tion field activities (Hooker et al. 2001), which is no morethan a few percent (currently about 3.5%). The princi-pal objective of the 2009 workshop (Wright and Hooker2009†) was to specify the desired and required featuresof a community-maintained, open-source, Web-based in-terface for the Processing of Radiometric Observations ofSeawater using Information Technologies (PROSIT‡).

Much of the recommended capabilities for a large vari-ety of fixed-wavelength instrument systems were incorpo-rated into PROSIT (Hooker and Brown 2011). The initialbeta version of PROSIT was produced in 2009, and ini-tial testing was completed in early 2010. The first opera-tional version was tested in the middle of 2010 and up untilthe second workshop was convened 8–10 December in SanDiego (California). The agenda was initially established

† The cited meeting summary plus the workshop agenda, listof attendees, and all of the talks that were presented by theparticipants are available on the Web at the following Website: http://oceancolor.gsfc.nasa.gov/DOCS/ by going tothe Meetings and Workshops heading and selecting “CVOAOP Workshop—January 2009.”

‡ Literally translated from Latin as “May it benefit.”

to accomplish the same thing for above-water AOP datathat the first workshop did for in-water AOP data (Fig.1). Because the in-water processing capability was rathermature, time was also reserved for demonstrating how in-water AOP processing is accomplished using PROSIT. Thelatter component included time to train alongside the prin-cipal PROSIT programmer. The training opportunity in-cluded extra time before the meeting for several scientistswho requested it.

Much of AOP data processing is inexorably tied to thedata acquisition activity, so time was also reserved for newtechnology presentations associated with the above-waterOptical Sensors for Planetary Radiant Energy (OSPREy),and in-water Compact-Optical Profiling System (C-OPS)instruments. The presentations showed how these twotechnologies are positioned to deal with the advanced plan-ning that is occurring for the Aerosol-Cloud-Ecosystems(ACE) mission and the Pre-Aerosol, Clouds, and OceanEcosystem (PACE) mission that was recently announced.Many of the scientists in attendance (Table 1) have alreadystarted using some of the new instrumentation in their re-search work, so time before, during, and after the workshopwas made available for discussions specifically tailored tothe science objectives being pursued.

The rationale behind the current workshop, which washosted by Biospherical Instruments Inc. (BSI), was to up-date the community and get community input with respectto the following:

1. Topics not addressed during the first workshop, spe-cifically the processing of above-water AOP datawithin the PROSIT architecture;

2. PROSIT data processing issues that have developedor tasks that have been completed since the firstworkshop; and

1


Fig. 1. The agenda for the 2010 AOP workshop, convened on 8–10 December in San Diego (California),showing informal meeting times (light blue), principal investigator (PI) presentations (orange), break-outsessions for working group discussions (green), posters and instrument displays (yellow), and alternativescheduling (gray blue).

2


Table 1. Attendees to the 2010 AOP Workshop are listed along with their affiliation, country, and e-mail address.

Investigator Affiliation Country E-mail Address

Samir Ahmed City College of New York USA [email protected] Antoine Laboratoire d’Oceanographie de Villefranche France [email protected] Ayoub University of South Florida USA [email protected] Bailey FutureTech/NASA/GSFC USA [email protected] Becu Universite Laval Canada [email protected] Bernhard Biospherical Instruments Inc. USA [email protected] Booth Biospherical Instruments Inc. USA [email protected] Brown CSTARS University of Miami USA [email protected] Court University of California USA [email protected] Dana HOBI Labs USA [email protected] Darecki Institute of Oceanology Poland [email protected] Dungan NASA/ARC USA [email protected] Frouin Scripps Institution of Oceanography USA [email protected] Garcia Fundacao Universidade do Rio Grande Brazil [email protected] Hargreaves Lehigh University USA [email protected] Hooker† NASA/GSFC USA [email protected] Kozlowski Scripps Institution of Oceanography USA [email protected] Kudela University of California Santa Cruz USA [email protected] Lind Biospherical Instruments Inc. USA [email protected] Morrow† Biospherical Instruments Inc. USA [email protected] Nelson University of California Santa Barbara USA [email protected] Ruddick Royal Belgian Institute of Natural Sciences Belgium [email protected] Schieber Scripps Institution of Oceanography USA [email protected] Sosik Woods Hole Oceanographic Institution USA [email protected] Stramski Scripps Institution of Oceanography USA [email protected] Subramaniam Lamont-Doherty Earth Observatory USA [email protected] Thomalla Council for Scientific and Industrial Research S. Africa [email protected] Toro-Farmer CMS/University of South Florida USA [email protected] Werdell SSAI/NASA/GSFC USA [email protected]

† Workshop organizer and co-chairman.

3. NASA instrumentation developments, both above-and in-water, which are relevant to both workshopsand next-generation mission planning.

The workshop emphasized presentations on new AOP in-strumentation (OSPREy and C-OPS), the desired and re-quired features for above-water AOP data processing,working group discussions, and a community update forthe in-water data processing already present in PROSIT.

The latter focus on new hardware, which is being de-veloped by BSI, was considered a fortuitous opportunityfor the community to be briefed directly by the manufac-turer. For the above-water sensors, this briefing is at acritical time in the development cycle, i.e., it allows forcommunity input regarding data processing requirementsbefore the designs of the instruments are finalized. For thein-water instruments, the community can provide input re-garding how data acquisition parameters feed directly intodata processing options.

The workshop themes were established by the hard-ware and software components for the AOP instrumenta-tion. The hardware theme is derived from two new sets

of instruments: the in-water C-OPS and the above-waterOSPREy, which are described by Morrow et al. (2010) andHooker et al. (2011), respectively. The software theme isdirectly connected to PROSIT, as well as the software as-sociated with data acquisition. The agenda was organizedto provide more details about the hardware and softwarecomponents:

• The development of microradiometers, which arethe building blocks for the new above- and in-waterinstruments;

• OSPREy data acquisition and anticipated data pro-cessing capabilities;

• C-OPS data acquisition capabilities directly influ-ence data processing capabilities;

• Existing (in-water) and anticipated (above-water)Web-based data processing capabilities, includingtraining for the former; and

• Documentation of the Web-based processor as a liv-ing document with the protocols and algorithmsthat are used included in the document.

3


The agenda was structured to provide the newest infor-mation first to maximize the amount of time available forcomments, questions, and discussion. Approximately 30scientists from the worldwide ocean color community par-ticipated in the workshop.

2. WORKING GROUPSThe six working groups were organized as follows:A Data ingest and data products;B Required and desired features for optically shallow

and optically deep waters;C Contamination rejection (clouds), corrections, and

data filtering;D Sun photometry and polarimetry;E Instrumentation networks; andF Hyperspectral versus fixed-wavelength sensors.

The instrumentation networks working group was intendedto provide more detailed information about desired andrequired features of autonomous sampling systems (e.g.,OSPREy sites).

2.1 Working Group AWorking Group A discussed data ingest and data prod-

ucts for above-water sensor systems. The initial discussionconcerned platforms, and it was decided that the data neednot be categorized by platform (drifting, fixed, etc.) aslong as the metadata are properly specified (i.e., a fixedplatform simply has invariant tilt and location). It wasnoted that the most widely-used above-water platform isprobably an aircraft. The group did not discuss aircraft-specific requirements, but again the only difference may bethe values of some metadata (high velocity, varying alti-tude and attitude, etc.).

The discussion on the basic data format establishedthat the American Standard Code for Information Inter-change (ASCII) format is preferred and tab-delimited val-ues are preferred, but not required. It was agreed that in-strument manufacturers would provide the necessary infor-mation to translate proprietary data formats into ASCII.This could be a simple text description, documentation ex-plaining algorithms, code fragments, or stand-alone soft-ware. Ingestion of raw data (volts or counts) is preferred,and every effort should be made to ingest data that is asraw as possible, so all processing steps can be controlledand properly executed.

The discussion on metadata centered on making surePROSIT users are aware they must satisfy the require-ments to support the current protocols (time and location,pressure tare, dark values, chlorophyll a concentration, skyconditions, etc.). It was also noted that certain other pa-rameters that might enable better processing in future,e.g., filter functions and field of view (FOV), should alsobe identified even though they may not be used in the near

future. If supplying these variables is not an impracticalburden, then users should be encouraged to provide them.To make submission easier, it was agreed that metadatamay be included in file headers or in separate files. Theprocessor has a defined vocabulary for tagging data types,so flexibility is already built in.

Protocols were discussed and it was noted that somedata products have multiple protocols, and only one shouldbe recommended. The so-called Modified Fresnel Reflect-ance Glint Correction method appears to have the widestuse and was recommended as the one protocol to use. TheOSPREy shadowband has several options for how data arecollected and processed, and it was suggested that thesedata can be processed like a profile.

The need for defined levels of processing and qualitywas discussed. The current version of PROSIT uses acoarse quality control (QC) rating, but mostly it is useroriented, because the user determines the necessary andappropriate QC criteria. It was recommended that theprocessor support more than just the calibration and val-idation perspective, but also research needs. It was ac-knowledged that the ocean color community does not re-quire from everybody a data quality level that is in keepingwith calibration and validation activities.

The data products were imagined to be produced by anOSPREy sensor suite—which has a separate solar referencewith shadowband attachment—and included the following:

Total radiance at the sea surface (LT );Sky radiance (Li);Water-leaving radiance (LW = LT − ρLi, where ρ isthe surface reflectance);Polarization (Stokes parameters) of LT and Li withspectrograph;Downward global-, diffuse-, and direct irradiance;andRatio of direct-to-global irradiance.

In addition, a suite of atmospheric data products are rec-ommended, which appear in Sect. 2.4, along with the max-imum diversity of QC products.

2.2 Working Group B

Working Group B discussed required and desired fea-tures for optically shallow and optically deep waters. Theinitial discussion attempted to come to a consensus on thedefinition of “optically shallow” and “optically deep.” Itwas generally agreed that actual bathymetry, although animportant variable that should be a required measurement,was not the defining parameter optically. As a workingdefinition, optically shallow waters exist when bottom fea-tures affect the light field; optically deep waters are notaffected by the bottom. Furthermore, optically deep oroptically shallow waters share an obvious spectral depen-dence with respect to bathymetry, and the discussion did

4


not disclose a specific boundary separating the two. Opti-cally shallow or deep waters were also not necessarily re-lated to optical complexity. For the purposes of acquiringdata, there was no difference in the measurement require-ments for the two cases, per se.

The use of multiple wavelengths as part of more so-phisticated atmospheric correction algorithms was recom-mended. One cited reason was the recurring failure to pro-duce data products when the assumption in calculations ofaerosol path radiance, i.e., the water-leaving radiances forthe 765 and 865 nm channels are zero, is invalid (turbidwaters or strong bubble field). Included in this recommen-dation was the use of higher spectral resolution in regionsof the spectrum that are changing rapidly, especially above1,000 nm.

In the discussion, it was observed that bubble cloudswithin the near-surface ocean layer can be prominent con-tributors to light scattering and water-leaving radiance re-gardless of whether the marine environment is opticallyshallow or deep. The time scales associated with the pro-duction and fate of bubbles within the near-surface oceanare short (on the order of a fraction of a second to min-utes). The intermittent nature of bubble clouds is a sig-nificant challenge that must be taken into account in ex-periments designed to measure radiometric quantities nearthe surface and for subsequent AOP determinations. Thisis especially true when wind speed exceeds about 5m s−1

and breaking waves inject bubbles into the water.In addition to the issue of whitecaps, the presence of

bubbles in the water column is also important within thecontext of atmospheric correction of ocean color measure-ments. The traditional black pixel assumption in the near-infrared portion of the spectrum is violated by the presenceof highly scattering bubbles near the surface. The inclu-sion of a meteorological package, accurately recorded time,and digital camera images for detection of whitecaps andbottom features were also highly recommended. An addi-tional requirement was added for a precise determinationof the FOV (spot size) with respect to each deployment ofinstruments above the water.

During the discussion, it was suggested that polariza-tion measurements should be strongly recommended. Skyradiance, Li, is polarized, and the reflectance from the wa-ter surface depends on the polarization state of the incom-ing radiation. The total radiance measured at the sea sur-face, LT , and the ratio of LT /Li, will depend on the degreeof polarization of the sky. Consequently, the most accurateretrievals of LW , require measurements of the Stokes pa-rameters of Li and LT . In addition, it was pointed out thatmeasurements of sky polarization may help in refining at-mospheric correction values. The following were suggestedas desirable, although not necessarily required: a) lidar fordetermining the concentration of colored dissolved organicmatter (CDOM), b) current meter, c) bottom reflectancemeter, and d) surface gravity wave sensor.

2.3 Working Group CWorking Group C discussed recommendations for the

processing of above-water radiometric data and protocols,with attention to contamination rejection (clouds), correc-tion, and data filtering. The group reaffirmed what PIsneed to be able to ingest into PROSIT, which are mostlyorganized as time series of data with geolocation(s) for thefollowing:

• Solar zenith angle (SZA) calculation;• Sky radiance, Li(λ, t);• Total radiance at the sea surface, LT (λ, t);• Global solar irradiance, Ed(0+, λ, t) or an estimate

from sun photometer data;• Three-angle geometry of the sensors, including az-

imuth;• Wind speed, W (t); and• Instrument metadata, e.g., calibration data, sensor

FOV, and sampling frequency, integration time.It was also noted that deployment information (photo-graphs) is inevitably valuable and that flexibility for manysystems and diverse protocols would be needed, e.g., three-sensor Trios, OSPREy, SIMBADA, SeaPRISM, etc. Inreturn, PROSIT should provide satellite ocean color re-flectances and matchup times, wind speed, cloud informa-tion, and real-time estimates of Ed(0+, λ, t).

Protocol issues were discussed and it was noted thatthere is a high diversity of protocols. The diversity is afunction of the relative azimuth angle used during datacollection; the sky glint correction methodology, e.g., windspeed dependence, near-infrared (NIR) correction, clearand turbid options, minimal LT filtering, and Cox-Munkto name a few; calibration information; and instrumentcorrections (cosine response, stray light, etc.).

Corrections and data filtering discussions centeredaround being able to identify anomalous geometry, point-ing errors (vertical tilts greater than 5◦, relative azimuthoff specification, and zenith angle off specification). Cloudflagging using the Li/Ed(0+) ratio in the NIR (750 nm,870 nm, or similar) should be less than 5%. The detectionand avoidance of precipitation was recommended, as wasensuring clean apertures, especially on moving platforms,which are subjected to more wind-blown spray.

Longer-term ideas were discussed and included calcula-tion of platform perturbations (which was recognized as asignificant effort) and the use of a gimballed solar referencesensor.

2.4 Working Group DWorking Group D discussed sun photometry and po-

larimetry. The sun photometer data products included to-tal optical depth, aerosol optical depth, Angstrom parame-ters; aerosol single scattering albedo; aerosol size distribu-tion; and aerosol scattering phase function. Different types

5


of sensor systems for the required measurements involvedwere discussed and included handheld-, shadowband-, andsun tracking sensors. These discussions also included fixedwavelength versus spectrograph sensors. The implicationsfor PROSIT centered around the acknowledgment that thealgorithms to calculate sun photometer data products arewell established, but the atmospheric and oceanographiccommunities use different symbols for data inputs and dataproducts. Another problem that was discussed was howbest to derive the global solar irradiance from instrumentslike SeaPRISM. An unresolved question was whether ornot PROSIT should be able to process a sun photometeronly data set.

The motivation for polarimetry was improved deriva-tion of the water-leaving radiance, LW = LT − ρLi, whereρ is the surface reflectance (which is a function of W );enhanced atmospheric correction; characterization of thein-water light field and scattering; and measurement offluorescence. A principal data product was the Stokes pa-rameters, I, Q, U , and V , which can be combined to de-scribe the coherent radiation (Q2 +U2 +V 2 = I2), as wellas the incoherent radiation (Q2 + U2 + V 2 < I2). Theimplication for PROSIT is the processor must be able toapply the Mueller matrix.

The sun photometry data products were imagined to beproduced by an OSPREy sensor suite, which has a separatesolar reference with shadowband attachment, and includedthe following:

Polarization (Stokes parameters) of LT and Li withspectrograph;

Downward global, diffuse, and direct irradiance;

Ratio of direct-to-global irradiance;

Direct-normal Sun irradiance (sun photometer appli-cation);

Total optical depth, aerosol optical depth, and Ang-strom parameters;

Aerosol single scattering albedo;

Aerosol size distribution and aerosol scattering phasefunction;

Total ozone column (from spectrograph);

Precipitable water column (from spectrograph); and

Cloud optical depth.

QC data products were also anticipated.

2.5 Working Group E

Working Group E discussed instrumentation networkswith specific application to the OSPREy sensor suite. Thedesired characteristics in a network, acknowledging therewill be different tiers of implementation, included the fol-lowing: the sensors should be part of a centralized activity;at least some of the data should be accessible in real time;

calibration of the sensors should be done at a single stan-dardized instrument calibration facility; coordinated logis-tical support should be available; the data processing anddistribution of data products should be centralized; a poolof instruments with a same-day swap capability should beestablished; and the instrumentation should be flexible,from both a scientific and technological perspective with aminimum set of criteria that should be satisfied.

The discussion of logical tiers established three types ofsensor suites: a) systems with maximum redundancy andcost would be deployed at a minimum number of sites; b)more numerous fixed validation sites with less capabilityand lower cost; and c) potentially very numerous mini-mal capability sites, which might include buoys, floats,etc. Other considerations discussed in connection with thetiers were the use of portable calibration sources at a siteto confirm calibration at the time of installation, as wellas data availability and timeliness of QC and data pro-cessing. It was acknowledged that participants might haveto buy in with more than one instrument to participate.Planning, in the context of NASA decadel survey missionsand associated preparatory activity funding, was also dis-cussed.

Near term recommendations to move forward includedthe following:

1. Populate one or two existing Aerosol RoboticNetwork-Ocean Color (AERONET-OC) sites withnew technology (e.g., OSPREy) to provide a proofof concept, a time series continuity and instrumentintercomparability, and a demonstration to overcomeresistance to change (within the existing network).

2. Open a dialog across users, agencies, and networkmanagers for the timely and efficient reporting ofprogress.

2.6 Working Group FWorking Group F discussed Web-based processing of

hyperspectral versus fixed-wavelength above-water sensors.The group advocated the processor should make no funda-mental distinction between hyperspectral and fixed wave-length sensors—hyperspectral sensors just have a lot ofchannels. The advantages of hyperspectral sensors aretheir optical bandwidth may be narrow and the band spac-ing is usually smaller than the actual bandwidth. Thedisadvantages are they may need spectral averaging, thedynamic range is almost always less, and the temporalresolution is less, both in terms of integration and datatransfer times.

The higher-order data products for hyperspectral sen-sors might require protocols and may not be standardized.Suggested products included red-edge position (common interrestrial applications), absorption feature depth, deriva-tive analysis, and spectral shape transforms. For the latter,it was noted that some applications (especially airborne)use transforms that reduce the data size of spectra, but

6


preserve their key features (analogous to compression al-gorithms for photographs).

It was noted that hybrid instruments, that is instru-ments having both hyperspectral and discrete components(e.g., OSPREy) could be treated as separate instruments.Such an approach might possibly remove a significant ad-vantage of such instruments, however, because the moreaccurate fixed wavelength channels can be used to keepthe spectrograph continuously calibrated.

The role of the processor was discussed at length. Thegroup advocated that the processor should not make re-search decisions. The PI is responsible for collecting mean-ingful data that adheres to the agreed upon sampling pro-tocols. The processor should support—but not require—highly sophisticated users and should offer options for min-imal processing. Standardization of processor functionswas considered important, because it allows researchers toduplicate each other’s processing on their own data sets,or try different processing on other data sets.

Specific functions that the processor should follow areas follows:

• All ingested sensor data and required metadatashould be in ASCII format; optional metadata (e.g.,photographs) can be in binary.

• Processing should start with the rawest data possi-ble, preferably counts, but partially processed (e.g.,dark-corrected) data should be accepted if there areno alternatives.

• The principal data product is the spectral water-leaving radiance, LW (λ), the global solar irradiance,Ed(0+, λ), plus the normalized variables created fromthese two observations, e.g., the remote sensing re-flectance, Rrs(λ), the normalized water-leaving radi-ance,

[LW (λ)

]N, etc.

• Output from the processing step should include a de-scription of the configurations, corrections, and op-tions applied in sufficient detail to unequivocally re-produce the processing that was done.

• For hyperspectral data, the output options shouldpermit binning to a uniform wavelength spacing forintercomparison purposes.

• There should be an option to convolve the instrumentdata with standard satellite sensor wavelength sets,including filter response functions.

• There was a (controversial) discussion about pro-viding derived atmospheric parameters from high-spectral-resolution data, e.g., the direct-to-diffuse ra-tio calculated by looking at oxygen absorption bandalgorithms.

• Comparison between hyperspectral and fixed wave-length instruments will require band-averaged datafrom the hyperspectral data to match filter wave-bands, and possibly the application of corrections tothe hyperspectral data based on the comparison (this

is inspired by the hybrid OSPREy sensors, but couldapply to other instrument designs as well).

• Data visualization can be complicated, because forhyperspectral sensors every sample is a spectrum.Sometimes it is sufficient to look at plots of specificwavebands (as is done with a filter radiometer), butsometimes a separate display is needed for looking atvariations in the full spectrum.

For the latter function, harmful algal bloom (HAB) appli-cations may require tracking the location of the red peak,which might shift slightly (selecting individual bands willnot suffice for this). It was also considered desirable tobe able to look at a movie or other ways of slicing thedata. A three-dimensional perspective plot was consideredappealing, but not always effective.

3. THE PROSIT MANUALThe basic architecture for PROSIT is to correctly ex-

ecute the existing protocols for processing AOP data toagreed upon data products. The utility of a set of proto-cols that are endorsed and maintained by a broader com-munity far exceeds the simple accomplishment of providingthe procedures for accomplishing certain tasks or measure-ments. As long as the protocols are a work in progress, pe-riodic updates provide a timely review of the state of theart and gives new ideas or procedures a forum for evalua-tion. This opportunity to discuss and document how thebasic tools for meeting PROSIT processing requirementsare being satisfied is a critical element for maintaining thesoftware.

Because the underlying protocols establish much of thesoftware architecture for PROSIT, it is appealing to in-clude the protocols in the PROSIT documentation, so thereis no confusion regarding specific PROSIT features. Partof the challenge of incorporating protocols documentationis facilitating the accessibility of the information, i.e., be-ing able to obtain and cite prior revisions, as well as thecurrent version. There is also a strong desire to learn fromprior efforts to publish protocols and make the informationeasy to access.

To prevent possible problems with scientists and re-searchers being able to access both current and legacy ver-sions of the protocols, meetings were held with personnelat the Center for AeroSpace Information (CASI). The jointwork developed a numbering and archival scheme that willbe sustainable over time for a new concept in documenta-tion for NASA—a so-called living document. The archi-tecture of this approach is encyclopedic in nature, but un-like most encyclopedias, it will be added to and updated,as needed, i.e., it “lives.” One of the first tasks was todetermine the clearest and most efficient method for num-bering and categorizing not only the initial document, butalso its future updates—without losing access to the legacymaterial—a pivotal concept in scientific literature.

7


A living document has not been attempted in the NASAScientific and Technical Information (STI) report series,so a number of challenges needed to be considered regard-ing many aspects of producing the document. Its struc-ture, how it would be printed, subsequent updating, thereport numbering scheme, and above all, the ease for re-searchers to access the archived data, have all been ofparamount importance. Agreements were also obtainedon how community members can provide material and re-ceive co-authorship. This new document type will still beclassified in the STI system as a Technical Memorandum(TM), however, which is familiar to most of the researchcommunity.

The current plan is for the PROSIT living document toinclude all the material needed to understand and use thesoftware: the protocols being implemented, user manualinformation, important coding routines, workshop sum-maries, etc. Updates and revisions to the document willbe done section by section as they are completed. Eachsection will have a unique pagination scheme for ease ofupdating and indexing. Sections will be distributed onthree-hole punched paper, so researchers can easily replaceone revision for another and store the TM in a binder ifdesired (Fig. 2).

Three-Ring Looseleaf Binder Tabbed Dividers

A

DC

IH

Sym

bols

Cum.Refs.

Index

Glossary

TM-2011-123456/Sect-A/Rev-1

Prologue

Ocean Color Calibration and Validation Protocols

of-view Sensor (SeaWiFS) Science Team and accepted byNASA. Of particular significance is that the SeaWiFSProject Office immediately started to implement a Sea-WiFS Validation Plan designed to ensure a best effort toachieve the product uncertainty goals, above (McClain etal. 1992). A critical aspect of the validation plan was thatin situ radiometric, optical, and bio-optical measurementsof uniformly high quality and accuracy be obtained forverifying SeaWiFS system performance and product un-certainties.

In 1991, the SeaWiFS Project Office sponsored a work-shop to recommend appropriate measurements, instrumentspecifications, and protocols specifying methods of cali-bration, field measurements, and data analysis necessaryto support SeaWiFS validation, which led to the publi-cation of Ocean Optics Protocols for SeaWiFS Validation,(Mueller and Austin 1992). Continued discourse within theocean color research community led to Revisions 1 (Muellerand Austin 1995), 2 (Fargion and Mueller 2000), and 3(Mueller and Fargion 2002) of these protocols.

1.2 Theoretical Basis

The Ocean Optics Protocols for Satellite Ocean ColorSensor Validation (Revision 4.0) was intended to providestandards, which if followed carefully and documented ap-propriately, would ensure that any particular set of opticalmeasurements would be acceptable for ocean color sensorvalidation and algorithm development. These protocolswere guidelines and may have been somewhat conserva-tive. In the case of ship shadow avoidance, for example,there were some circumstances in which acceptable radio-metric profiles could be acquired considerably closer to aship than was previously specified. When the protocolswere not followed in such cases, however, it was incum-bent upon the investigator to explicitly demonstrate thatthe actual error levels were within tolerance. Close adher-ence to these protocols was the most straightforward wayfor an investigator to establish a measurement was uncon-taminated by artifacts, such as ship shadow, and was ac-curate enough to meet the requirements of satellite oceancolor product validation.

Finally, having a standard set of measurement proto-cols is indispensable in developing consistency across thevariety of international satellite ocean color missions eitherrecently launched or scheduled for launch in the next fewyears. While each mission has its own validation effort, themission validation teams should not need to define sepa-rate validation measurement requirements. In the U.S.,for instance, ocean color validation support was derivedfrom four separate funding programs, i.e., the SeaWiFSProject, Moderate Resolution Imaging Spectroradiometer(MODIS) validation program, the Earth Observing Sys-tem (EOS) calibration and validation program, and theSensor Intercomparison for Marine Biology and Interdisci-plinary Oceanic Studies (SIMBIOS) Project (McClain and

Fargion, 1999a, 1999b). Continued development and re-finement of these protocols help ensure coordination, col-laboration, and communication between those involved.

Immediate concerns focused the early versions of theOcean Optics Protocols (Mueller and Austin 1992, 1995)on specific preparations for the SeaWiFS mission. In theinterim, not only SeaWiFS, but the Japanese Ocean ColorTemperature Sensor (OCTS), the Polarization DetectionEnvironmental Radiometer (POLDER), and the MODISglobal coverage ocean color systems have been successfullylaunched and brought into operation, and the near-termlaunch of several other such systems is anticipated (Ap-pendix A). The SIMBIOS Program goal is to assist theinternational ocean color community in developing a multi-year time-series of calibrated radiances that transcendsthe spatial and temporal boundaries of individual missions(Barnes et al. 2001). Specific objectives are to: (1) quan-tify the relative accuracies of the products from each mis-sion, (2) work with each project to improve the level ofconfidence and compatibility among the products, and (3)develop methodologies for generating merged level-3 prod-ucts. SIMBIOS has identified the primary instruments tobe used for developing global data sets. These instru-ments are SeaWiFS, OCTS, POLDER [Advanced EarthObserving Satellite (ADEOS)-I and II], MODIS (Terra andAqua), Multi-angle Imaging SpectroRadiometer (MISR),Medium Resolution Imaging Spectrometer (MERIS), andGlobal Line Imager (GLI). The products from other mis-sions [e.g., Ocean Color Imager (OCI), Ocean ScanningMultisprectral Imager (OSMI), and Modular OptoelectronicScanner (MOS)] will be tracked and evaluated, but are notconsidered as key data sources for a combined global dataset.

The scope of the protocols was, therefore, broadenedto support development of bio-optical databases that meetthe expanded requirements of the SIMBIOS goals and ob-jectives (Fargion and Mueller 2000). The key objectiveaddressed by the original working group was to recom-mend protocols and standards for supporting in situ opti-cal measurements. The original objectives remain valid to-day, albeit with broader requirements for detailed measure-ments and sensor characteristics (e.g. wavelength charac-teristics). The generalized protocol objectives address thefollowing subject areas:

1. The required and useful optical parameters to beused for validation of satellite ocean color sensornormalized water-leaving radiances and atmosphericcorrection algorithms, and for monitoring each satel-lite sensor’s calibration and stability, will be de-fined.

2. The instrumentation requirements, and standardsfor measuring the parameters in item 1, includ-ing definitions of measured quantities, wavelengths,field-of-view (FOV) and band specifications, sensi-tivity, uncertainty and stability, will be delineated.

2 TM-2011-123456/Sect-A/Rev-1

Hooker et al.

In-Water Apparent Optical Properties (AOPs)

Stanford B. HookerNASA/Goddard Space Flight Center

Greenbelt, Maryland

John Q. ResearcherState University CollegeSomewhere, America

Jane Z. InvestigatorForeign Country Institute

Anywhere, Worldwide

Abstract

The accurate determination of upper ocean apparent optical properties (AOPs) is essential for the vicariouscalibration of an ocean color sensor like SeaWiFS and the validation of the derived data products, becausethe sea-truth measurements are the reference data to which the satellite observations are compared. The onlyeconomically feasible approach for minimizing spatial biases for a global data set is to maximize the acquisitionof in situ measurements by soliciting data from the oceanographic community at large. To ensure uniformity ofthe requisite data quality requirements, minimization of uncertainties is critical. The uncertainties associatedwith in situ AOP measurements have various sources, such as, the deployment and measurement protocols usedin the field, the absolute calibration of the radiometers, the environmental conditions encountered during datacollection, the conversion of the light signals to geophysical units in a data processing scheme, and the stabilityof the radiometers in the harsh environment they are subjected to during transport and use. The protocolspresented here establish the agreed upon principles for reducing uncertainties from most of these sources.

1. Introduction

1.1 Importance to Ocean ColorDuring the period of 1985–1991, the National Aeronau-

tics and Space Administration (NASA) charged a series ofsuccessive science working groups with the task of recom-mending guidelines, goals, and mission design criteria forfuture satellite ocean color remote sensors. The deliber-ations of these working groups were based on the oceancolor science community’s experiences with the Nimbus-7Coastal Zone Color Scanner (CZCS). The highly success-ful CZCS mission firmly established ocean color remotesensing as a powerful tool for monitoring and studying thebio-optical properties of the global ocean. The caveat tothis, however, is that the radiometric responsivities of theCZCS channels declined progressively with time through-out its 8-year operating life. This decline firmly establishedthe need to independently verify the performance of a satel-lite’s sensor using in situ measurements of the ocean andatmosphere. From this perspective, the principal recom-mendations of these NASA Ocean Color Science WorkingGroups (collectively) included:

1. Baseline satellite ocean color products should in-cludea. Normalized water-leaving radiances LWN (λ)

(Gordon and Clark 1981),b. Aerosol radiances La(λ),c. Chlorophyll a concentration—chl [mgm−3],d. The diffuse attenuation coefficient K(490) at a

wavelength of 490 nm, ande. Calibrated radiances Lt(λ), observed at the satel-

lite.2. Principal goals for product uncertainties should be

a. Less than 5% uncertainty in LWN (λ) andb. Less than 35% uncertainty in Chl.

3. An ongoing satellite ocean color sensor system val-idation program is necessary, using in situ mea-surements of oceanic radiometric and bio-opticalproperties, and of atmospheric optical properties, toverify system performance—including algorithms—immediately after launch and throughout a satelliteocean color sensor’s operating lifetime.

These and other recommendations of the earlier work-ing groups were endorsed by the Sea-viewing Wide Field-

TM-2011-123456/Sect-A/Rev-1 1

Fig. 2. A representation of a living documentplaced in a looseleaf binder with tabbed dividersfor the glossary and symbols sections, in addition tospecialized sections for a high-level index and a cu-mulative references section. The “magnified” footerhighlights a sample CASI TM number, which willappear on each page.

4. DISCUSSIONThe workshop concluded with a plenary discussion of

problems and desired solutions already encountered withusing PROSIT. A summary of some of the topics discussedand the ensuing recommendations to be undertaken are asfollows:

1. Data ingested into PROSIT need to be processedwith the objective of a unified quality, so criteria needto be established to ensure all data can be processedto the same level of completeness. This means datamust be as raw as possible, metadata must be pro-vided, and the protocols must be followed as closelyas possible (because PROSIT simply follows the pro-tocols).

2. It was recommended that PROSIT does not have todeal with significant errors made during data acqui-sition, but does need to be able to flag such data,which might only be possible if the PI identifies theproblem. The latter assumes the PI knows aboutthe problem, which might not always be the case.For example, some PIs used irradiance sensors witha responsivity that was only valid in water as above-water solar references.

3. Given limited resources, there are only two optionswith compromised data quality: either work on theindividual data sets to improve them, or work onbetter programming routines to trap and flag baddata. The group consensus was the latter option isthe best. It should be the PI’s responsibility to cor-rect compromised data or investigate how the datacan be used for other objectives.

4. Some problems during acquisition can be solved withbetter PROSIT tools. For example, some PI’s do notapply a pressure tare, but the data acquisition mightinclude data when the profiler was on the deck of theship before or after the cast(s). The PROSIT CastEditor page should be modified to identify not onlydata rejection intervals (e.g., data collected after thebottom of a cast), but also to identify data to be usedfor the pressure tare or dark counts.

5. The automated submission of data and products tothe SeaWiFS Bio-Optical Archive and Storage Sys-tem (SeaBASS) is a required feature, but should in-clude processing headers that explicitly identify miss-ing criteria, configuration parameters, and qualityflags.

6. If all the established criteria are met for a particulardata ingestion, then the original data and metadataplus the resulting data products could be stored inSeaBASS, otherwise the data would be for the PI’sresearch purposes and not distributed.

7. The SeaBASS discussion ultimately resulted in a rec-ommendation to establish a submission (data qual-ity) threshold. If an ingested data set satisfies thethreshold, the data are accepted for full processingand submitted to SeaBASS; if not, the data can beprocessed, but the problems preventing acceptancewill likely not be mitigated and poorer quality re-sults will be obtained.

8


8. The SeaBASS threshold was seen as a way to re-lease PROSIT from the unbounded burden of tryingto create submission quality data, but does not pre-vent anyone from using PROSIT. The threshold wasalso seen as a motivator to improve data quality overtime.

9. The nascent training module in PROSIT is impor-tant and should be expanded. Training should beused to show how to process data, as well as what isrequired to collect data for correct processing. A fieldchecklist in the living document would be beneficialto novice practitioners.

10. A discussion on the above-water analogs to the abovein-water topics included similar recommendations.A separate recommendation, however, was to adoptonly one above-water method for PROSIT to im-plement, and the consensus was to use the Mod-ified Fresnel Reflectance Glint Correction method,because it is used the most. If another protocolis proved superior, then PROSIT will evolve andchange to that protocol.

11. A discussion on whether or not alternative protocolswill be considered inferior to the PROSIT protocolestablished the need for performance metrics, so al-ternatives can be evaluated objectively and the stateof the art can be advanced quantitatively.

12. Shadowband, sun photometry, and polarimetry pro-cessing were discussed within the context of port-ing whatever the OSPREy activity accomplishes intoPROSIT.

13. A future workshop, approximately in late 2011 orearly 2012, is recommended to discuss a) the above-water capabilities implemented in PROSIT, b) re-finement of the in-water capabilities, c) drafting ofperformance metrics for both above- and in-waterprocessing, and d) revision of the protocols involved(e.g., self-shading correction).

14. A discussion on making point measurements closeto the sea surface emphasized the difficulty of do-ing that using an extrapolation interval and led todiscussions on alternatives, for example, using mea-surements of the inherent optical properties (IOPs),and how to evaluate and include them. The uncer-tainty part of the discussion included calibration un-certainties and how this problem also applies to thealternatives. The recommendation was to keep anopen mind.

15. How best to improve early participation of PIs andnew practitioners with PROSIT established that theburden of working with problematic data falls on thePI and the community in general. A substantial ad-vantage will be PIs can sensibly compare results forthe first time.

Acknowledgments

The success of the 2010 AOP Workshop was in large measuredetermined by the excellent logistical support provided by Bio-spherical Instruments Inc. Although many people contributedto the success, the excellent work done by Ms. Anna Reed isparticularly acknowledged.

Glossary

ACE Aerosol-Cloud-EcosystemsAERONET Aerosol Robotic Network

AOP Apparent Optical PropertyASCII American Standard Code for Information In-

terchange

BSI Biospherical Instruments Inc.

CASI Center for AeroSpace Information (NASA)CDOM Colored Dissolved Organic MatterC-OPS Compact-Optical Profiling SystemCOTS Commercial Off-The-Shelf

CSTARS Center for Southeast Tropical Advanced Re-mote Sensing

FOV Field of View

GSFC Goddard Space Flight Center

HAB Harmful Algal BloomHOBI Hydro-Optics, Biology, and Instrumentation

IOP Inherent Optical Property

NASA National Aeronautics and Space Administra-tion

NIR Near-Infrared

OC Ocean ColorOSPREy Optical Sensors for Planetary Radiant Energy

PACE Pre-Aerosol, Clouds, and Ocean EcosystemPI Principal Investigator

PROSIT Processing of Radiometric Observations of Sea-water using Information Technologies

QC Quality Control

SeaBASS SeaWiFS Bio-Optical Archive and Storage Sys-tem

SeaPRISM SeaWiFS Photometer Revision for Incident Sur-face Measurement

SeaWiFS Sea-viewing Wide Field-of-view SensorSIMBADA Satellite Validation for Marine Biology and-

Aerosol Determination, AdvancedSSAI Science Systems and Applications, Inc.STI Scientific and Technical InformationSZA Solar Zenith Angle

TM Technical Memorandum

Symbols

Ed(0+, λ, t) The global solar irradiance, or an estimate fromsun photometer data.

I One of the Stokes parameters.

Li(λ, t) The sky radiance.LT (λ, t) The total radiance at the sea surface.LW (λ) The water-leaving radiance (defined as LT −

ρLi).[LW (λ)]

NThe normalized water-leaving radiance.

9


Q One of the Stokes parameters.

Rrs(λ) The remote sensing reflectance.

U One of the Stokes parameters.

V One of the Stokes parameters.

W (t) The wind speed.

λ Wavelength.

ρ The surface reflectance.

References

Hooker, S.B., and S. Maritorena, 2000: An evaluation ofoceanographic radiometers and deployment method-ologies. J. Atmos. Oceanic Technol., 17, 811–830.

, G. Zibordi, J-F. Berthon, D. D’Alimonte, S. Mari-torena, S. McLean, and J. Sildam, 2001: Results of theSecond SeaWiFS Data Analysis Round Robin, March2000 (DARR-00). NASA Tech. Memo. 2001–206892,Vol. 15, S.B. Hooker and E.R. Firestone, Eds., NASAGoddard Space Flight Center, Greenbelt, Maryland,71 pp.

, C.R. McClain, and A. Mannino, 2007: NASA Strate-gic Planning Document: A Comprehensive Plan forthe Long-Term Calibration and Validation of Oce-anic Biogeochemical Satellite Data. NASA SpecialPub. 2007–214152, NASA Goddard Space Flight Cen-ter, Greenbelt, Maryland, 31 pp.

, and J.W. Brown, 2011: Processing of RadiometricObservations of Seawater using Information Technolo-gies (PROSIT): In-Water User Manual. NASA Tech.Memo., NASA Goddard Space Flight Center, Green-belt, Maryland, (in prep.).

, G. Bernhard, J.H. Morrow, C.R. Booth, T. Cromer,R.N. Lind, and V. Quang, 2011: Optical Sensors forPlanetary Radiant Energy (OSPREy): Calibrationand Validation of Current and Next-Generation NASAMissions. NASA Tech. Memo. 2011–215872, NASAGoddard Space Flight Center, Greenbelt, Maryland,(in press).

Morrow, J.H., S.B. Hooker, C.R. Booth, G. Bernhard,R.N. Lind, and J.W. Brown, 2010: Advances in Mea-suring the Apparent Optical Properties (AOPs) ofOptically Complex Waters. NASA Tech. Memo. 2010–215856, NASA Goddard Space Flight Center, Green-belt, Maryland, 80 pp.

Wright, V.M., and S.B. Hooker, 2009: “AOP Workshop:Development of a Web-based Community AOPProcessor,” hosted by University of California SantaBarbara, Santa Barbara, California, 13–15 January2009, http://oceancolor.gsfc.nasa.gov/MEETINGS/CVO/AOP_2009/SummaryReport/AOP_Workshop_Report.pdf.

10

REPORT DOCUMENTATION PAGE Form Approved

OMB No. 0704-0188 The public repor ing burden for his collection of informa ion is estimated to average 1 hour per response, including the ime for reviewing instruc ions, searching existingdata sources, gathering and maintaining the data needed, and completing and reviewing the collec ion of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to Department of Defense, Washington Headquarters Services, Directorate for Information Operations and Reports (0704-0188), 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to any penalty for failing to comply with a collection of information if it does not display a currently valid OMB control number. PLEASE DO NOT RETURN YOUR FORM TO THE ABOVE ADDRESS.

1. REPORT DATE (DD-MM-YYYY) 2. REPORT TYPE 3. DATES COVERED (From - To)

4. TITLE AND SUBTITLE 5a. CONTRACT NUMBER

5b. GRANT NUMBER

5c. PROGRAM ELEMENT NUMBER

6. AUTHOR(S) 5d. PROJECT NUMBER

5e. TASK NUMBER

5f. WORK UNIT NUMBER

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION REPORT NUMBER

9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSORING/MONITOR'S ACRONYM(S)

11. SPONSORING/MONITORINGREPORT NUMBER

12. DISTRIBUTION/AVAILABILITY STATEMENT

13. SUPPLEMENTARY NOTES

14. ABSTRACT

15. SUBJECT TERMS

16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF

ABSTRACT 18. NUMBER

OF PAGES

19b. NAME OF RESPONSIBLE PERSON

a. REPORT b. ABSTRACT c. THIS PAGE 19b. TELEPHONE NUMBER (Include area code)

Standard Form 298 (Rev. 8-98)Prescribed by ANSI Std. Z39-18

05-03-2011 Technical Memorandum


NNX09AN94G

Stanford B. Hooker, John H. Morrow, James W. Brown, andElaine R. Firestone

NASA Calibration and Validation Office (CVO)BWtech, 1450 S. Rolling Rd., Suite 4.001Halethorpe, MD 21227

National Aeronautics and Space AdministrationWashington, DC 20546-0001

TM-2011-215871

Unclassified-Unlimited, Subject Category: 25, 48Report available from the NASA Center for Aerospace Information, 7115 Standard Drive, Hanover, MD 21076. (443)757-5802

John H. Morrow: Biospherical Instruments Inc., San Diego, CA; James W. Brown: CSTARS University of Miami, Miami, FL;and Elaine R. Firestone: Science Systems and Applications, Inc., Lanham, MD

The rationale behind the current workshop, which was hosted by Biospherical Instruments Inc. (BSI), was to update the community and get communityinput with respect to the following: topics not addressed during the first workshop, specifically the processing of above-water apparent optical property(AOP data) within the Processing of Radiometric Observations of Seawater using Information Technologies (PROSIT) architecture; PROSIT dataprocessing issues that have developed or tasks that have been completed, since the first workshop; and NASA instrumentation developments, both above-and in-water, that are relevant to both workshops and next generation mission planning. The workshop emphasized presentations on new AOPinstrumentation, desired and required features for processing above-water measurements of the AOPs of seawater, working group discussions, and acommunity update for the in-water data processing already present in PROSIT. The six working groups were organized as follows: a) data ingest and dataproducts; b) required and desired features for optically shallow and optically deep waters; c) contamination rejection (clouds), corrections, and datafilltering; d) sun photometry and polarimetry; e) instrumentation networks; and f) hyperspectral versus fixed-wavelength sensors. The instrumentationnetworks working group was intended to provide more detailed information about desired and required features of autonomous sampling systems. Plenarydiscussions produced a number of recommendations for evolving and documenting PROSIT.

SeaWiFS, apparent optical properties, AOP, optically complex waters, PROSIT

Unclassified Unclassified UnclassifiedUnclassified

10

Stanford B. Hooker

(410) 533-6451

the 2010 aop workshop summary report - nasa › archive › nasa › casi.ntrs.nasa... · the 2010...

Documents