UNIVERSITY OF CALIFORNIA, SAN DIEGO
UNIVERSITY OF CALIFORNIA, SAN DIEGO
3 1822 04101 6478
3 1822 04101 6478
Version 1.1
May 2007
Offsite MOSPHERIC OPTICS GROUP
(Annex-Johnical Note No. 271
rnals)
QC
974.5
.T43
Enhancement of Near-Real-Time
no. 271
Cloud Analysis
And Related Analytic Support
For Whole Sky Imagers
Published as Final Report for ONR Contract
N00014-01-D-0043 DO #4
UNIVERSITY
OF
CALIFORNIA
SAN DIEGO
Janet E. Shields
Monette E. Karr
Art R. Burden
Richard W. Johnson
William S. Hodgkiss
The material contained in this note is to be considered
proprietary in nature and is not authorized for distribution
without the prior consent of the Marine Physical Laboratory.
SCRIPPS
INSTITUTION
OF
OCEANOGRAPHY
MARINE PHYSICAL LAB San Diego, CA 92152-6400
.
- - -
- -
-
-
-
-
-
UNTERRITY OF CALIFORNIA, SAN DIEGO

3 1822 04101 6478
T-
10
+ ir
ATMOSPHERIC OPTICS GROUP
Technical Note No. 271
Version 1.1
May 2007
Enhancement of Near-Real-Time
Cloud Analysis
And Related Analytic Support
For Whole Sky Imagers
Published as Final Report for ONR Contract
N00014-01-D-0043 DO #4

UNIVERSITY
OF
CALIFORNIA
SAN DIEGO
Janet E. Shields
Monette E. Karr
Art R. Burden
Richard W. Johnson
William S. Hodgkiss
tersebut
The material contained in this note is to be considered
proprietary in nature and is not authorized for distribution
SCRIPPS
without the prior consent of the Marine Physical Laboratory.
INSTITUTION
OF
OCEANOGRAPHY I MARINE PHYSICAL LAB San Diego, CA 92152-6400
Enhancement of Near-Real-Time
Cloud Analysis
And Related Analytic Support
For Whole Sky Imagers
Table of Contents
1. Introduction.
2. Background .......
3. Statement of Work .....
4. Funding Increments and Optional Tasks....
5. Major Deliveries and Documentation.......
6. Discussion of Hardware Developments ....
7. Day Algorithm Developments and Analysis .........
8. Night Algorithm Developments and Analysis ..........................
9. Discussion of Wavelength Options for Optical Cloud Imaging ........
9.1 The Visible and SWIR Intercomparison Experiment ...........21
9.2 Thermal IR Systems ...........
..........
9.3 Comments Regarding Mid Wave IR Systems
9.4 Theoretical Study of LWIR Characteristics ..........
9.4.1 Discussion of the Approach...
9.4.2 Cloud Signatures, part 1.......
..............27
9.4.3 Background Signatures .............
...........29
9.4.4 Cloud Signatures, part 2..
............31
9.4.5 Resulting Cloud and Background Effective Temps....32
9.5 Summary of the Wavelength Analysis ..
...........34
...........
....................................
10. Summary ..........
........................35
11. Acknowledgements .......
12. References.
List of Illustrations
III
Fig. 1 Profile of Standard WSI Occultor......
Fig. 2 Profile of Improved Occultor.
............9
Fig. 3 Initial Design for computer control to replace ACP Panel .........11
Raw Red and Processed Cloud Decision for 20 Feb 02 1420....12
Raw Red and Processed Cloud Decision for 21 Feb 02 1700....12
Fig. 6 Raw Red and Processed Cloud Decision for 12 Aug 02 1706 ...13
Fig. 7 Raw Red and Processed Cloud Decision for 16 Aug 02 2200 ...13
Fig. 8 Forward Bias Haze Case ........
Fig. 9 Reverse Bias Haze Case ..........
Fig. 10 Red/Blue Ratio Through the Sun on 3 Typical Days ................
Fig. 11 Red/Blue Ratio Through the Sun on Forward Bias Days ........
Fig. 12 Red/Blue Ratio Through the Sun on Reverse Bias Days ...........
Fig. 13 No Moon Case in Oklahoma ...........
............16
Fig. 14 No Moon Case at SOR enhanced to show clds away from city 16
Fig. 15 No Moon Case at SOR enhanced to show clds near city ...........17
Fig. 16 Moonlight Example
...............17
Fig. 17 Night Raw and Processed Data for a clear case.......
Fig. 18 Night Raw and Processed Data for a broken cloud case .........
Fig. 19 Night Raw and Processed Data for an overcast case.
Fig. 20 Star Irradiance Comparison ..............
Fig. 21 Sample Cloud Transmittance Extraction Results .....................
Fig. 22 Overcast case, imagery at 1.6 um and 650 nm .............
Fig. 23 Thin clouds, imagery at 1.6 um and 650 nm .......
...........
Fig. 24 Thin and pre-emergent clouds at 1.6 um and 650 nm ................23
Fig. 25 Sunset, imagery at 1.6 um and 650 nm........ ...........................23
Fig. 26 Sample Cloud Image taken in the 3 - 5 um region .....................25
Fig. 27 Sample Cloud Image taken in the 8 - 12 um region ...................25
Fig. 28 Vertical Profile of Scattering Coefficients......... ...............28
Tables
Table 1 Cloud Altitudes used in Calculations........
....
Table 2 Estimated Vertical Beam Transmittance.
. ......27
.........................29
Table 3 Effective Radiating Temperatures used for Sky Backgrounds ......31
Table 4 Computed Cloud and Background Results for the Zenith .............32
Table 5 Computed Cloud and Background Results for 60° Zenith Angle..33
Table 6 Computed Cloud and Background Results for 85° Zenith Angle..33
Enhancement of Near-Real-Time
Cloud Analysis
And Related Analytic Support
For Whole Sky Imagers
Janet E. Shields, Monette E. Karr,
Art R. Burden, Richard W. Johnson, and William S. Hodgkiss
1. Introduction
This report describes the work done for the Starfire Optical Range, Kirtland Air Force
Base under Contract N00014-01-D-043 DO #4, between 25 May 01 and 31 September
06. This work relates to the Air Force's need to characterize the cloud distribution during
day and night, for a variety of applications, including support of satellite tracking, and
support of research into impact of clouds on laser communication. This contract
followed Contract N00014-97-D-0350 DO #6, which will be discussed in Section 2, and
is documented in Shields et al 2004b, Technical Note 265. The primary goals of
Delivery Order #4 discussed in this current report included further development of day
and night cloud algorithms and support of the fielded Whole Sky Imager instruments.
Much of the work done under DO #4 was completed by the end of 2004. Some
additional work was done in 2005 and 2006 under the DO #4 funding, but most of the
SOR work during this interval was done under a follow-on contract, ONR N00014-01-D-
0043 DO #11, funded September 04. The work under DO #11 will be reported under a
separate report.
We would also like to note that we were funded by several other sponsors during this
period. Although we have no desire to provide an overview of the other work, in some
cases it was related to the SOR goals, and may be discussed in this context.
2. Background
A series of digital, automated Whole Sky Imagers (WSI) have been developed by MPL
over many years, beginning in the early 1980's (Johnson et al. 1989 and 1991, and
Shields et al. 1993, 1994, 1997a and b). These systems are designed to acquire accurate
imagery of the full upper hemisphere in several spectral filters, in order to be able to
assess the presence of clouds at each pixel in the image. A system capable of 24-hour
operation, the Day/Night WSI, was developed by MPL under funding from the Air Force,
Navy, and Army in the early 1990's (Shields et al. 1998, 2003b and e, 2004 a and b, and
2005b and c). One of the first two units was fielded at the Air Force's Starfire Optical
Range in October 1992.
Related systems have been designed and fielded over the years. These include a new
Daytime Visible/NIR WSI (Feister et al. 2000, and Shields et al. 2003d), and visible and
Short-wave NIR systems for airborne use (Shields et al. 2003c). An imaging system for
measuring visibility was developed and successfully tested (Shields et al. 2005a and
2006). Also, a field calibration device for use with the Day/Night WSI was developed
(Shields et al. 2003a).
Partly as a result of SOR's experience with the Day/Night WSI fielded in 1992, they have
funded fabrication of additional instruments, as well as algorithm development and data
analysis in recent years. Under Contract N00014-97-D-0350 DO #2 (Sep 97 – June 01),
MPL was funded to develop and provide a new Day/Night WSI, which was designated
Unit 12 (Shields et al. 2003b). This unit included several design upgrades developed for
other sponsors. The funding also included analysis and processing of existing daytime
cloud decision images to provide statistical estimates of Cloud Free Line of Sight
(CFLOS) and related properties. Under the optional funding, the WSI was also upgraded
to run under Windows. This was a major upgrade, involving a new camera model, a new
camera software library, and new WSI instrument control software. The instrument was
delivered in Jan '99, and has been running well for much of the time since. The data
analysis results were also delivered in '98 and 99, and my understanding is that these
data have proven to be quite useful.
Under Contract N00014-97-D-0350 DO #6 (May 99 - May 03), we were funded to
provide two additional instruments, Units 13 and 14 and do analysis work (Shields et al.
2004b). These instruments included significant design upgrades, including integration of
the control computer into the outdoor environmental housing, new control hardware and
software, and new software to provide near-real-time cloud processing on an additional
display computer. One of the instruments was fielded at a site in California for an
experiment, and ran well prior to the completion of the experiment, at which time it was
returned to MPL. The other system was kept at MPL pending sponsor readiness for
deployment. It was allowed to run continuously, and ran well.
In addition, under the options, we were funded to develop a new Starlight Cloud
Algorithm based on detection of the contrast between the signal from stars and their
background. The concepts had been developed under funding from another sponsor
(Shields et al. 2002), and under Delivery Order 6, this concept was expanded to handle
moonlight, and converted to a fieldable C-code. The appropriate geometric calibrations
and background were extracted for the Unit 12 running at the SOR site, and the algorithm
was installed and provided reasonable results.
Toward the end of this prior contract, a funding increment was received that partially
funded the new contract, and partially funded the existing contract. The Statement of
Work (SOW) for the new contract, DO #4, reflected the priorities at the time of the
funding increment. However, we were asked to use the options under the existing
contract, DO #6, to begin the work. The part of this work that was completed under the
older contract was reported in Shields et al. 2004b, but it will also be reviewed in Section
3. Statement of Work
The Statement of Work for the Delivery Order #4 is given in italics below.
Primary Task:
The contractor shall, unless otherwise specified herein, supply the necessary personnel,
facilities, services, and materials to accomplish the following tasks within a one-year
period following receipt of funding.
1. Modify the SOR unit, Unit 12, to provide the current version of the night algorithm,
Near Infra-Red data acquisition, and cloud algorithm processing on a sponsor-
supplied display computer.
2. Provide those miscellaneous software upgrades, such as a modification of the Display
Computer interface, which are mutually agreed upon by the sponsor and by MPL to
be appropriate.
3. Continue development of the day and/or night cloud algorithm.
4. Provide a stand-alone image processing program, which may be used to process
archived data through the cloud algorithms.
5. Provide a final written report and interim verbal reports to the sponsor regarding the
results of the above work.
6. Other work such as minor repairs to the instrument are permissible under this task, if
they are mutually agreed upon by the sponsor and by MPL to be appropriate. In this
case, the work toward the other requirements in this statement of work would be
lessened accordingly.
Optional Task
The contractor shall, unless otherwise specified herein, supply the necessary personnel,
facilities, services, and materials to accomplish the following tasks within a two-year
period following receipt of funding.
1. Coordinate with the sponsor regarding the most appropriate tasks and estimated costs
for development
2. Provide personnel trained in the Whole Sky Imager and its capabilities to address
these tasks (at MPL) to the limit of funding provided under the contract. These tasks
may include analysis, software development, documentation, minor hardware
development, and other tasks related to the WSI that are mutually agreed upon by the
sponsor and by MPL to be appropriate.
.
3. Provide a final written report and interim verbal reports to the sponsor regarding the
results of the above work.
4. Funding Increments and Optional Tasks
The funding was sent from the Air Force to ONR in a number of funding increments, and
the optional tasks were defined as these funding increments were received. The funding
increments and associated tasks are listed below.
An initial funding increment was sent to ONR in Nov 00. Some of these funds were
allocated into the previous Contract N00014-97-D-0350 DO #6, and the remaining funds
were put into the new Contract N00014-01-D-043 DO #4. The new contract was
received about May 01. The priorities were to work on the tasks listed in the Statement
of Work Primary Tasks section. Because we were asked to work on the new priorities
under the existing contract, we were able to complete portions of the first two items in the
SOW under the earlier contract with this first split funding increment.
The second funding increment was received about Aug 03. This completed funding for
the Primary Tasks. In addition, we were given the following priorities under the options.
1. Unit 12 at SOR had been hit by lightning, and we were asked to assess and if possible
repair this unit.
2. Deploy Unit 14 to a site in Virginia.
In January 04, we received two additional small funding increments, with the associated
tasks given below. At that time, we were asked to keep track of these increments
separately because they came from separate funding sources, and so separate budgets
were set up. However, later we were told that this was no longer necessary. Thus in this
report, we do not distinguish between these different budgets.
3. Continue night transmittance map development and analysis work.
4. Begin to evaluate optimal WSI designs for the future, including hardware, software,
and algorithms, and evaluate costs technical and cost tradeoffs associated with
providing such instruments.
A fifth funding increment was received Mar 04. The priorities for this increment were:
5. Repair the occultor trolley that had failed on Unit 14, which was fielded in Virginia.
6. For Unit 14, radiometrically calibrated the camera system from Unit 13, and swap it
into Unit 14.
7. Provide a glass dome and heater for Unit 14 to minimize dew on the outside of the
dome.
8. Evaluate night imagery from Unit 14, which was fielded at a very bright location, and
adjust the flux control algorithm to acquire on-scale data.
A sixth funding increment was received on May 04. Part of the funding increment was
applied to this contract, and the remaining funds were applied to a new contract
o a new contract, Contract
N00014-01-D-043 DO #11, which was received Sep 04 and will be discussed in a
separate report. The priorities were for the full funding increment, and thus could be
addressed either under this contract or under the follow-on contract. Of these new
priorities, the one addressed under this contract is as follows:
9. Evaluate the pros and cons of eventually developing an IR WSI for detecting clouds
during the day and at night.
5.0. Major Deliveries and Documentation
All funded tasks were completed. Because there were so many primary and optional
tasks, we would like to document the major deliveries and documentation for these tasks.
In the later sections, we will provide an overview of the more significant advances related
to general capabilities. Additional documentation is provided in the technical
memoranda listed below and in the reference section. These technical memos can be
provided to the sponsor upon request.
Primary Task 1: In January 01, we took a trip to SOR to upgrade the WSI Unit 12. The
system was upgraded to include a Near Infrared (NIR) filter, to better handle haze and
thin clouds. A processing computer was added to the system, to enable real-time
processing of the cloud algorithm. Changes were made to the data acquisition software
to enable both of these upgrades, and a new processing program, to enable running both
the day and night algorithms, was written and installed. (This program, like the earlier
version that ran on the control computer, was a Windows program. The trip is
documented in Memo AV01-031t, and the software is documented in Memo AV01-029t.
In February 01, we delivered several further upgrades to the algorithms and software.
The day algorithm requires a background library of the clear sky red/blue ratios. Up until
this time, we had been using the background from another site and instrument. As part of
this contract, we wrote programs to enable extracting and processing this background
data for these instruments, and we extracted the background and set up the day algorithm
in the field to use these inputs. The day algorithm was also updated to include a horizon
mask and occultor mask, to mask out regions that do not represent sky data. The new
night algorithm was installed; this was an algorithm based on the contrast between the
star and its background, and adapted specifically for the lighting distribution at the SOR
site. The day algorithm processing is documented in Memos AV01-032t and 035t, and
the software upgrades are documented in Memos AV01-033t and 34. This work is
documented in Tech Note 265, and it completed Primary Task 1 in the SOW.
Primary Task 2: In January and February 01, we updated the Unit 13 and 14 software
interface code, as documented in Memo AV01-030t, and integrated in the algorithm
improvements made for the Unit 12 system, as documented in Memo AV01-034t. The
control and processing software for these units was different from that for Unit 12, partly
because of differing sponsor requirements (connections to another system), and partly
due to impacts of running different generations of operating systems. Unit 12 runs under
Windows 95, and Units 13 and 14 run under Windows NT. The control code for Units
13 and 14 was developed as a script running under the V++ image processing package,
because this was felt to be more efficient that converting to the new camera library
required for Windows NT. This work is documented in Tech Note 265, and it completed
Primary Task 2 in the SOW.
Primary Task 3 and 4: Continuing development of algorithms was required under Task 3
in the SOW, as well as in Optional Task 3 and Optional Task 9 in Section 4. In addition,
Task 4 in the SOW requires the development of a stand-alone image processing program.
Most of the work on the day algorithm in the early part of the contract was partially
funded under the earlier contract, and is already documented in Technical Note 265. This
includes adapting the algorithm to run in real time, extracting the constants for the SOR
site, and extracting the clear sky background for the SOR site. In June 03, we prepared a
detailed talk for SOR providing an overview of WSI systems, day and night algorithms.
To further develop the Day algorithm, we developed a stand-alone processing code (Task
4), documented in Memos AV04-038t and 039t, and several new utility programs
developed to support the effort, as documented in Memo AV04-043t. We processed a
fairly extensive data set from Oklahoma, partly with funding from another sponsor. This
gave us the opportunity to evaluate an extensive data set, and determine the strengths and
weaknesses of the algorithm. These results are documented in Memos AV04-047t, 049t,
and 050t. The results were presented to the sponsor in talks in June 04 and November 04,
and will be discussed in Section 7.
Some of the work on the night algorithm in the early part of the contract was partially
funded under the earlier contract, and is already documented in Technical Note 265. This
includes adapting the contrast-based algorithm for running in real time, updating it to
work under moonlight, extracting the geometric calibrations, adding the ability to mask
the horizon and occultor, and extracting the appropriate inputs for use at the SOR site.
This work is documented in Memo AV03-030t. We also began work toward extracting
beam transmittance for the night sky, and tested initial concepts for a full resolution night
algorithm. This work will be discussed in Section 8.
Although both the day and night algorithm have continued to evolve, we believe that this
work described above and further documented in Sections 7 and 8 completed the
requirements of Primary Tasks 3 and 4.
Primary Task 5 is to provide a final written report, which is this report. In addition,
interim verbal reports were provided via Power Point talks in June and December 2003;
January, July and December 2004; April, June, August and December 2005; and
February 2006. Much of the work in 2005 and 2006 was funded under the follow-on
contract. Copies of these presentations are available to the sponsor on request. This
report completes Primary Task 5.
Primary Task 6 in the SOW is to provide minor repairs as needed. This was for Unit 12,
the WSI at SOR. As it turns out, Unit 12 was hit by lightning, and more major repairs
were funded under the options, and are discussed below, as well as in Section 6. The
work documented in Section 6 completes Primary Task 6.
Optional Tasks 1 and 2 given in Section 3 essentially allow work on options coordinated
by MPL and SOR. These options are discussed in detail in Section 4, and become
Optional Tasks 1 - 11 in Section 4. They will be discussed below. Optional Task 3 in
Section 3 is to provide a final written report and interim verbal reports. The work is
discussed in this final report, and was reported in the verbal interim reports documented
two paragraphs above.
Optional Task 1 in Section 4 was to repair Unit 12, which had been hit by lightning.
Problems were diagnosed and mostly repaired during a trip in October 03, and in
cooperation with the sponsor during November 03, as documented in Memos AV03-049t
and AV03-052t. Repair to the unit included replacing the photodiodes in the filter
changer, replacing the DIO card, replacing the computer power supply, replacing a
camera card, and repairing the flow system. This work completed Primary Task 6 as well
as Optional Task 1 in Section 4.
Optional Task 2 in Section 4 was to deploy Unit 14 to Virginia. Earlier in `03 (and to
some extent '02), we had supported the test of Unit 14, and its preparation for
deployment. In November 03, we deployed the unit to Virginia. Memo AV03-047
documents the new control system software, and Memo AV03-048t documents the
processing computer software. Memo AV03-054t is the trip report, and the components
are given in Memo AV04-007t. This completed Optional Task 2 documented in Section
4. We also had some problems with this unit, and support memos AV03-055t and 560
and AV04-001t were provided to help support the field team. This work completed
Optional Task 2 in Section 4.
Optional Task 3 in Section 4 was to continue the night transmittancc map development
work; this was discussed earlier in conjunction with the algorithm work, and will be
further detailed in Section 8. This work completed Optional Task 3 in Section 4.
Optional Task 4 in Section 4 was to begin to evaluate optimal WSI designs for the future.
As part of this work, we evaluated occultor redesigns. We had designed and built an
occultor using an encoder under funding from another sponsor, but had not been happy
with the encoder performance. We purchased a different encoder, and tests with this onc
were successful. We also evaluated current cameras available. We found that the
Photometrics cameras are no longer available, but have been replaced by Princeton
Instruments cameras. However, the latter company is not experienced with bonding of a
fiber optic taper on the chip, so it may be necessary to redesign in the future to avoid this
issue. However, the new cameras are air cooled, and that will simplify the environmental
housing requirements. We did a pre-design for a “virtual ACP" that nearly eliminates the
Accessory Control Panel, and presented these concepts at the December 2004 talk. Some
of this work was funded under the follow-on contract that started in September 04. This
work completed Optional Task 4 in Section 4.
Optional Tasks 5 - 8 in Section 4 were all related to the Unit 14 support. A trolley repair
was completed in April 04, and documented in Memos AV04-032t, 33t, and 34t. This
completed Optional Task 5 in Section 4.
For Optional Task 6 in section 4, we calibrated both the fielded sensor and the repaired
sensor. We first documented in Memo AV04-002t (January 2004) how to apply the
calibrations that had been acquired earlier, in July 2000. The field calibration program
originally developed for another sponsor was updated to work with the SOR systems, as
documented in Memo AV04-030t. We evaluated some questions related to the impact of
shutter timing, as documented in Memo AV04-024t, and then calibrated the Unit 13
camera system in May 2004, as documented in Memo AV04-025t and 27t. The camera
housing was exchanged with that in the field in June 04, and documented in Memo
AV04-035t. As part of this trip, we also repaired an environmental housing temperature
sensor, and changed the GPS update interval to be compatible with the site constraints.
The returned camera was also calibrated on return, and documented in Memo AV04-045t
and 046t. This completed Optional Task 6 in Section 4.
For Optional Task 7 in Section 4, we built a glass dome with heater into the camera
housing. This is designed to evaporate dew and snow on the dome in the winter. It was
installed in the camera housing prior to the calibration discussed above, and installed in
the field in June 04, as documented in Memo AV04-035t. During fielding we found it
was better not to run it in the summer. This completed Optional Task 7 in Section 4.
For Optional Task 8 in section 4, we also updated the flux control to work for this very
bright location. We analyzed the open hole data, as documented in Memo AV04-054t,
and decided that we would like to evaluate data taken spectrally at this site. The flux
control was updated to acquire spectral data at night, as documented in Memo AV04-
035t. This completed Optional Task 8 in Section 4.
As noted earlier, Optional Task 9 was intended to be addressed primarily under the next
contract, DO #11, but could be addressed under this contract also. A theoretical and
practical analysis of the use of IR imagers in meeting the project needs was done, and the
results presented in July 04. These results are discussed in Section 9. This work
completed Optional Task 9 in Section 4.
As documented above we have, with the completion of this report, completed all contract
requirements for Contract N00014-01-D-043 DO #4 to the best of our knowledge.
6. Discussion of Hardware Developments
As discussed in Section 5, during this interval we continued maintenance of the WSI Unit
12 at SOR, and also deployed WSI Unit 14 to Virginia. The Unit 12 system generally
operated fairly well, although it required extensive repair following the lightning strike,

and somewhat more repairs than normal during the period after that. However, the in-
house records show a high number of repairs required for Unit 14. The unit ran
essentially flawlessly during a several month deployment to Northern California, and it
ran essentially flawlessly during extensive test periods at MPL. We had several problems
at the Virginia site. Some were, we believe, due to the inexperience of the personnel at
the site, and some may have been because this was a new version of the instrument. The
truth is that we cannot account for the relatively poor behavior of the instrument at the
site.
Some additional memos related to general instrument repair that have not already been
listed in Section 5 are memos AV01-068t, AV01-070t, AV01-080t, AV01-092t, AV03-
048t, AV04-009t, AV04-01lt, AV04-013t, AV04-026t, AV04-029t, AV04-031t, and
AV04-044t. These memos, listed in the Reference section, document the instrument part
numbers, return of the unit from Northern California, software upgrades, procedures for
making repairs, check lists, and so on.
Work under Optional Task 4 of Section 4 to evaluate optimal WSI designs for the future
was discussed in section 5. We would like to enlarge on two areas, the occultor
modernization and the ACP modernization.
The normal occultor design with the trolley, shown in Fig. 1, obscures about 2.1% of the
sky. However, we make the mask slightly larger than the actual occultor, because there is
some uncertainty in the position. We find that a 1% uncertainty covers most occurrences.
Because the rails and chains are close together, if there is a 1.0° uncertainty in occultor
position, the resulting occultor mask covers about 9.1% of the sky. A new occultor was
designed to address this, under funding from another program. It is not only smaller, but
also has a more accurate readout on position, so that a 0.5 to 1° tolerance can be used.
This is shown in Figure 2. This results in a 4.6 – 5.5% loss over the sky, as opposed to
the earlier 9.1% loss.
Fig. 1. Profile of standard WSI occultor
Fig. 2. Profile of Improved Occultor
In order to improve the position accuracy, this occultor is less mechanically flexible, and
it also uses absolute encoders to provide a more accurate readout. Also, it will be less
costly to build, as it does not require as much control electronics fabricated at MPL.
Unfortunately, at the time the occultor was designed, only one company provided an
encoder that met the necessary specifications, and this encoder turned out to have
reliability issues. Since that time, a similar encoder became available that was
manufactured by a more well-known company, and that also met the required
specifications. Under the SOR contract DO #4, we purchased one of these encoders and
its software and tested it, and found that it was both reliable and compatible with the WSI
system. In the future, if more WSI systems are built, we would recommend using this
new encoder with the new occultor design, unless a better occultor is designed in the
interim.
Another significant upgrade we evaluated is simplification of the Accessory Control
Panels. These ACP's control the occultor shown in Fig 1 and the filter changer, and they
provide readout of the various environmental housing protection sensors, such as the flow
sensor and the temperature sensors. They enable the system to be run either manually
using switches, or with computer control. At this time, we feel that the manual
interaction is not required, and a much simpler ACP can be designed. Preliminary
concepts were developed under this contract, and tested under the next contract. A
preliminary design of the interface, which would now be a computer interface rather than
a panel with meters, is shown in Fig. 3. This upgrade should both lower cost and increase
reliability.
7. Day Algorithm Developments and Analysis
As discussed in Section 5, during the early part of this contract, much of the work on the
day algorithm involved upgrading the SOR site. We had a day cloud algorithm that could
be run in the field, however it was running on the same computer as the data acquisition,
and did not allow the system to be run at 1-minute intervals. We added a processing
computer, and adapted both the control and the processing logic for this setup. We also
installed a NIR filter, which should do a better job with thin clouds under hazy
conditions, and adapted the programs to handle this data acquisition.
The clear sky background is a necessary part of the thin cloud part of the cloud algorithm,
and fairly time consuming to extract from the field data. Up until this time, we had used
the background extracted for another site, and told the sponsor that the thin cloud results
should not be considered valid. As required under Primary Task 1, we extracted the clear
sky background for SOR, along with the other parameters that are required to run the
algorithm, thus providing SOR with thin cloud results.

Virtual ACP
ARC DRIVE
TROLLEY DRIVE
CURRENT POSITION
CURRENT POSITION
90.00
90.00
BRAKE
SET
REVERSE
STOP
FORWARD
REVERSE
STOP
FORWARD
RELEASE
PULSE
REVERSE
PULSE
FORWARD
PULSE
REVERSE
PULSE
FORWARD
DRIVE ARC TO:
DRIVE TROLLEY TO:
90.00
90.00
GO
GO
SENSORS
FILTER CHANGER
CCD TEMP
-40.00
SPECTRAL WHEEL
STEP
FORWARD
POSITION
CAM HSG
TEMP
30.00
STEP
REVERSE
ENY HSG
TEMP
25.00
POSITION
FLOW
RATE
0.400
ND WHEEL
STEP
FORWARD
N2
PRESSURE
5.000
STEP
BEVERSE
Fig. 3. Initial design for computer control to replace Accessory Control Panels
Under this contract, we then developed a stand-alone day algorithm program for use at
MPL in testing the strengths and weaknesses of the algorithm, as we had not had the
opportunity to evaluate an extensive data set to see how well the algorithm performed.
Also, support programs, documented in Memo AV04-043t, were developed to make the
processing somewhat more efficient. We were funded by another sponsor to process data
from an Oklahoma site for a full two months and portions of several other months were
processed, in order to extract Cloud Free Line of Sight (CFLOS) (Shields et al. 2005b).
This provided the most efficient mechanism for us to evaluate the general strengths and
weaknesses of the algorithm for the SOR program, since the data was already processed,
and the site was midway in haze level between the SOR and the Virginia sites.
In general, we were quite pleased with the results. Sample results are shown for February
in Figures 4 and 5, and for August in Figures 6 and 7. In these figures, dark blue
represents regions identified by the algorithm as clear, light blue represents thin cloud,
and white represents opaque cloud. Black pixels represent “no data”. Note that the
horizon mask is higher in the August data, because the prairie grass had grown up into
the field of view.
Normally the two most difficult regions for a cloud algorithm are the horizon and the
solar aureole. The algorithm generally handled these regions quite well. Especially in
Figure 5, it is easy to see that the cloud discrimination near the horizon is excellent.
Similarly, the bias with solar scattering angle is very small, as can be seen particularly in

Figure 7. That is, the cirrus are identified quite well near the sun as well as at angles well
away from the sun.
Fig. 4. Raw Red and Processed Cloud Decision for 20 Feb 02,1420 GMT.
Fig. 5. Raw Red and Processed Cloud Decision for 21 Feb 02, 1700 GMT.
In our evaluation, we found that there were two primary problem areas with the
algorithm. First, we found that about a third of the time, the results were not valid for
sunrise and sunset. This is a difficult time regime that will require some work. Second,
the algorithm does not yet handle haze optimally. Examples are shown in Figures 8 and
9. In the February data, we did not lose any data due to haze, but we lost nearly a third of
the August data due to haze.
12

Fig. 6. Raw red and Processed Cloud Decision for 12 Aug 02, 1706 GMT.
Fig. 7. Raw red and Processed Cloud Decision for 16 Aug 02, 2200 GMT.
Fig. 8. Forward bias haze case
Fig. 9. Reverse bias haze case
13
During this period, we were funded to develop a better haze algorithm for the Day WSI
(Shields et al. 2003d). As documented in Memo AV04-014t, we were able to
significantly improve results for the Day WSI by making two changes. The new Day
WSI algorithm uses NIR/blue ratios in place of red/blue ratios, and it also adjusts each
image for the current haze amount. Similar upgrades have been partially completed for
the Day/Night WSI under the follow-on funding, and are anticipated to be the next major
improvement in the day algorithm
Although we feel that the developments made on the Day WSI should address the issue
with haze for most conditions, there was another really interesting thing about the haze
cases from August. Although generally the haze caused a false call of thin cloud near the
aureole, as shown in Fig. 8, some resulted in a false call of thin cloud everywhere but the
aureole, as shown in Fig. 9. We have not yet seen this occur elsewhere. The
identification of a thin cloud depends on the current red/blue ratio, in comparison with
the nominal clear sky ratio. When we were first developing the cloud algorithm, in the
mid 1980's, we found that in general, the shape of the clear sky ratio distribution is
reasonably fixed, as a function of scattering angle and look angle, even in varying haze
wever, in Figures 8 and 9, the ratio near the aureole is varying in
comparison with the rest of the sky, as shown in Figures 10 through 12. We would
anticipate that this is due to a change in the drop size distribution, which affects the
aureole most strongly. If this turns out to be a common occurrence, then we will need to
make the algorithm more sophisticated to handle this.
2002 SGP R/B Ratio Image Cross-Section, Row 205:
Solid - Aug. 14, Dash - Aug 18, Dotted - Aug 19
300

TTTTTTTTTT
TTT
TTTTTTT
-
-
-
1
Ratio
דדדדדדדדדדדדדדדדייוידיויזדי
~-unde
6
.
monadure mand og har ma dovrh traghuni
Tximin
hann var
500
100
200
300
400
Column Pixel
Fig. 10. Red/blue ratio in a column through the sun on 3 typical days.
14
2002 SGP R/B Ratio Image Cross-Section, Row 205:
Solid - Aug. 2, Dash - Aug 3
300

דר זוזו ז ח ו ז חרזדן זיך דחד
דוח ממד
- - - -
250E
====-------
-- -----
-
--
-
.
_
LL
ג
.
1
Ratio
*
או באוז
*
דדדדדדדדדדדדדדדיווייוויד
wanaumatislava on
m
.jo
100
400
500
200
300
Column Pixel
Fig. 11. Red/blue ratio in a column through the sun on 3 days with enhanced aureole ratios or forward bias.
Solid - Aug. 5, Dash - Aug 6, Dotted - Aug 8

דדדדדדדדדדדדדדדדדדדדדדדדדדדדדדדדדדדדדדדדדדדד ד י ד ד 700
read
ד ד
;
,לי - t
ד י
Ratio
ידידיוייוידדדדייוידייידיידי]
'
י
www.w,
- . */:
.
4
ע •
י
דדדדדדד
200
300
400
100
500
Column Pixel
Fig. 12. Red/blue ratio in a column through the sun on 3 days with low aureole ratios, resulting in reverse
cloud bias.

In addition, we hope in the future to extract aerosol information from the WSI. We see
the red/blue and NIR/blue ratios vary with the amount of haze. Variations in the radiance
across the sky, particularly near the aureole, may also become useful in extracting other
information related to the aerosol.
In short, the day algorithm work done under this contract enabled us to get a real time day
algorithm optimized for SOR with the correct clear sky background, and then we
developed a stand-alone version of the code for analysis of data at MPL. As a result, we
were able to evaluate an extensive data set, and determine the strengths and weaknesses
of the algorithm. Much of this work was presented in a talk at SOR in December 2004.
8. Night Algorithm Developments and Analysis
To identify the presence of clouds at night, it is first necessary to detect them, i.e. to have
raw data in which the clouds may be distinguished. Since often the assumption is made
that this is impossible with a visible system, we would first like to note that we have not
yet seen a situation in which the clouds are not clearly visible in the imagery - even when
they are not visible to the human eye. As an illustration, Figure 13 shows a moonless
night from a site with minimal lights in the surround. Figures 14 and 15 show a moonless
night from a site near a city. The system has a grey scale of 0 to 65,535, with a readout
noise near 1. As a result, it is difficult to show the full range in an illustration. Figure 14
has been windowed to show the clouds in the darker part of the sky, and Figure 15 has
been windowed to show the clouds in the brighter part of the sky. Figure 16 shows a
moonlight image. In all cases, the data is clearly present for detecting the clouds.
Fig. 13. No moon case in Oklahoma, File 73520120
Fig. 14. No moon case, SOR, enhanced to show
clouds away from city
Our first version of the night cloud algorithm was programmed for the SOR site just
before this contract started, and reported in Shields et al. 2004b. This is a moderate
16

resolution algorithm, based on contrast between the sky and background. It is still in use
at some WSI sites, and will be referred to as the contrast-based algorithm. Examples of
results are shown in Figures 17 - 19. This version of the algorithm required quite a bit of
development, as we had not previously handled moonlight, nor did we previously have an
automated algorithm. We were reasonably pleased with the results.
ANAL
Fig. 16. Moonlight example
Fig. 15. No moon case, SOR, enhanced to show
clouds near city
en onsmaturuwonneruduntugtune
Fig. 17. Night raw and processed data for a clear case (N/S reversed in these images)
This contrast-based night algorithm is based in part on the detection of approximately
2000 stars per image. An angular calibration using star position was developed, to
provide accuracy to within about 72 pixel or 1/6º. Decisions are made within each
roughly 5º zenith by 15° azimuthal cell. For more details, see Shields et al. 2004b.
17.

Although the algorithm worked reasonably well, it did not handle the city lights well, and
we wanted to develop, in the long run, a full resolution algorithm, i.e. one that makes a
decision at each pixel, as the day algorithm does.
DALIUWUTOTULUITOR UMUMUM TUTOR
Fig. 18. Night Raw and processed data for a broken cloud case (N/S reversed in these images, yellow is
thin cloud)
IUMINIUUUUUUUUUUUUUTHROாய
Fig. 19. Night Raw and processed data for an overcast case (N/S reversed in these images)
Under Delivery Order #4 (the contract being discussed in this report), we developed the
first concepts for a full resolution algorithm, as documented in Memo AV01-069t. The
first step required in order to develop the full resolution algorithm is to calibrate the
images for absolute radiance, and then extract the earth-to-space beam transmittance for
selected stars. All 3 of the SOR units were calibrated, as documented in Memos AV01-
041t through AV01-067t and AV01-094t. Software was written to apply these
calibration results to field images, as documented in Memo AV03-037t.
18

As noted in Section 5, later in the contract we were given the field calibration device we
had developed for another sponsor, and we adapted the software to work with the SOR
Windows systems, as documented in Memo AV04-030t. This made the recalibration of
Units 13 and 14 much more efficient (Memos AV04-024t, 025t, 027t, 045t, and 046t).
The next step required development of techniques to extract earth-to-space beam
transmittance. We already had techniques to detect the stars, documented earlier. This
transmittance work was heavily leveraged by funding from another sponsor, as reported
in Shields et al. 2004a, Memo AV01-097t, and AV04-004t. As part of this work, we
developed the equations for computing the inherent star irradiance in the WSI passband,
based on the star library star magnitude and color temperature. We also developed
techniques for determining the measured star irradiance from the calibrated radiance
image. Figure 20 shows the results for the theoretical star irradiance, in comparison with
measured star radiance, for approximately 1500 bright stars down to an irradiance of 10-9
watt/m²-um (down to a star magnitude between 3 and 4, depending on color
temperature). When stars to magnitude 6 are included, the results are reasonable,
although more scattered, with a R' correlation value of .85 for 32,000 stars. In later work,
to be reported in the next report, we were able to improve these results significantly, to
yield a R' correlation value of .935 from the same site.
Spectral Irradiances For WSI Passband, Zenith < 60 Degrees
Spectral Irradiance > 10- (W/m-um), Method: Pixel Addition
10- TITT TTTTTTTTTTTTTTT
TTTTTT
Unit 12, February 14, 1999
Measured Spectral Irradiance (w/ma-um)
5
Linear Coefficients:
A = 6.11882e-011
B = 1.01323
Correlation Coeff =
0.987889
10-94
1
10-9
10-6
10-6
10-7
Theoretical Spectral Irradiance (W/m-um)
Fig. 20. Comparison of measured star irradiance, as determined from the calibrated radiance image, with
the theoretical star irradiance, as determined for the WSI passband from the star library magnitude and
color temperature.
19

An example of the resulting cloud transmittance map is shown in Figure 21. This has
been generated by correcting the raw transmittance map with an aerosol transmittance of
0.8. The resulting ranges, less than .6 for cloud, and greater than .6 for the sky, seem
very reasonable, although there is more scatter in the values in the clear regions than we
would like.
20
Fig. 21. Sample cloud transmittance extraction result. Color key: Red: no star detected; orange, yellow,
and green correspond to transmittance ranges of 0 - .2, 2 - .4, and .4 - .6 respectively; purple, blue and
turquoise correspond to 6.-.8,.8 -1.0, and > 1.0 respectively
Additional work toward improved transmittances, a transmittance-based algorithm, a
high resolution algorithm, and methods for ground-truthing the results, were continued
under the next contract, and will be discussed in the next report.
9. Discussion of Wavelength Options for Optical Cloud Imaging
The sponsors would like to have the ability to detect clouds over the whole sky that will
be of importance in attenuating transmission at or near 1.6 um in the SWIR. Based on
scattering theory, we would anticipate that a visible sensor like the WSI should do very
well at this task. Clouds are typified by large droplets, in the regime of about 2 - 20 um
(Hobbs and Deepak 1981, see article by Cox, Fig. 1). Because the droplets are larger
than the wavelengths of interest, the scattering losses should be similar for the visible and
the SWIR. We also expect the scattering to be similar for the visible and the SWIR for
light fog and heavy fog (Holst, 2000 Figure 5.7). Based on models H, L and M
(McCartney 1976 Figure 3.8), haze tends to have drop sizes shorter than the 1 um. (Also
see Hobbs and Deepak 1981, article by Jiusto Fig. 9, for an example of haze developing
into fog.) Thus we expect that the scattering losses in the SWIR will be less than those in
the visible.
As a result, the WSI should detect all scattering losses of importance in the SWIR. The
WSI will also detect the haze, and it will be important that the haze not be identified as
cloud. However, it will be useful if the WSI algorithms identify cases with heavy haze
that might affect the SWIR. We should also note that although we have not extensively
evaluated the absorption, we would expect that absorption losses should not be
significant in the SWIR, except for ice near 1.61 um.
Although we feel the WSI should do an outstanding job detecting the scattering losses
that impact the SWIR, we and our sponsors felt that this question was worth further
evaluation. Under this contract, we made a preliminary, but fairly extensive, evaluation
of the pros and cons of using sensors in the visible, in the short wave infrared (SWIR) at
1.6 um, and in thermal infrared (IR) wavelengths for this purpose. The analysis was
presented in a talk on July 04. The Power Point file can be provided to the sponsors upon
request. In Section 9.1 below we discuss an experiment to evaluate how well SWIR
sensors can perform for this task. This experiment also evaluated how well the visible
sensor detects the clouds seen in the SWIR. In Section 9.2 we provide a general
evaluation of available imagery from IR cloud detection systems. In Section 9.3, we
briefly evaluate Mid Wave IR. In Section 9.4, we provide a theoretical analysis of the
Long Wave IR potential for our applications.
The Power Point talk also included a general overview of the theory, including plots of
the drop sizes associated with haze and with clouds, and the impacts of scattering and
extinction in the different bands. This is not included in this report, as it is already
covered in numerous theoretical texts.
Before we proceed further with the wavelength discussion, I would like to include one
comment made during the talk, regarding the bias of our group. I have to admit that we
love the WSI. We have found it to be a very capable and flexible instrument, and we
have found that very good algorithms can be developed if the analyst is familiar with
both atmospheric physics and the instrument performance and calibration characteristics,
Some desired aspects of the algorithms are not yet developed, and also nothing is ever
perfect. We have some experience with the IR and would love to gain more experience.
It is in the University's interest for us to present a fair and unbiased evaluation, and that is
what we will try to do.
9.1. The Visible and SWIR Intercomparison Experiment
If a whole sky imager could be developed that worked day and night, down to the
horizon, directly in the SWIR wavelengths, this could have an advantage, because we
would be measuring the clouds directly in the wavelengths of interest. During 2002 -
2004, under funding from another sponsor, we developed a SWIR calibrated fisheye
imager, for use in an aircraft or UAV, looking down at cloud tops (Shields et al., 2003c).
While the system was still at MPL, we ran an experiment to compare the visible and
SWIR systems looking up from the ground. One of the purposes of this experiment was
to determine whether the WSI system is able to detect all of the clouds of significance in
the SWIR wavelengths. A second purpose was to evaluate the effectiveness of the SWIR
imager in measuring the clouds from the ground.
21

Several examples of image sets are shown in Figures 22 through 25. The images were
not exactly simultaneous, because we had only one solar occultor. We acquired an image
with one camera, moved the occultor, and then acquired the image with the other camera,
obtaining images within about a minute. In each image, the SWIR image is shown on the
left, and the visible image is shown on the right.
Fig. 22. Overcast case, imagery at 1.6 um and at 650 nm 7 May 04 near 1040 Local
Fig. 23. Thin clouds, imagery at 1.6 um and at 650 nm 7 May 04 near 1230 Local
In Figure 22, we had an overcast of fairly heavy clouds, and the images are quite similar
in terms of the clouds detected. The quality of the raw data is superior in the visible, due
to the camera technology (note the vertical lines and poorer resolution in the SWIR
imagery on the left side).

Fig. 24. Thin and pre-emergent clouds, at 1.6 um and at 650 nm 7 May 04 near 1221 L
Fig. 25. Sunset, images at 1.6 um and at 650 nm 11 May 04 near 1942 L
Relatively thin clouds are shown in Figures 23 and 24, and Figure 24 also included pre-
emergent clouds. In this situation, the haze is beginning to acquire texture and become
slightly more white, and clouds are about to form. This is less obvious when seen
visually (with a human Photopic peak response near 555 nm), because the WSI image is
at 650 nm, where the spectral contrast between the background haze and the larger
droplets is increased over the visible. We do not yet know if it will be important to
identify these as clouds for SWIR applications, but at least in this situation they were not
seen in the SWIR image. From this imagery, it appears that the visible system detected
all of the clouds seen in the SWIR.
There are two primary reasons that we feel a visible system will be superior to a SWIR
system for the application of detecting significant losses in the SWIR transmittance.
23
system h both sensory superio,
y the SWT
First, the visible sensors are significantly superior, in terms of noise, uniformity, and
linearity. We have calibrated both sensors very carefully, and for example at a signal of
about 200, the SWIR system has a non-linearity error close to a factor of 1.8 (80% error).
Secondly, the SWIR system has much less dynamic range than the visible camera used in
the WSI, due to higher noise and limited grey scale range. Figure 25 shows an image
near sunset, where the SWIR image is not very useful. The SWIR system was using a
filter with a 50 nm passband filter. If this filter were removed, the sensor would see 2
logs (a factor of 100) darker. However, night imaging requires at least sensitivity 5 logs
darker than the example shown in Figure 25. The WSI has a sensitivity that goes 7 logs
darker than the setting shown in Figure 25. For these reasons, we do not recommend
using a system at 1.6 um at this time.
9.2. Thermal IR Systems
By making contacts with colleagues, and exploring the Internet, we were able to obtain
information and/or sample imagery from three IR ground-based sky imaging systems
with reasonably wide fields of view. These systems all are based on the Long Wave IR
(LWIR) between 8 and 12 um. These systems included: a) a system using an 18” mirror
and a Raytheon 300A Uncooled Ferroelectric camera at 10 um; b) a system using a 150°
lens and a New-Technology Uncooled Microbolometer operating at 8 – 13 um; c) a
system with a 180° lens and an uncooled microbolometer. As part of this study, we also
researched the available cameras from the IR manufacturers and evaluated their
characteristics for the purpose of detecting the clouds; however this is beyond the scope
we wish to include in this report.
Although we found several images on the web, we are in an awkward position, because
we are no longer able to find the web sites in order to determine whether the images may
be used in a formal document. In fact, one of the web sites disappeared shortly after we
extracted the images. We know that at least one of these instruments is still in use at the
current time, but we are not aware of images that are available for use in a report. At the
time of the July 04 presentation, we tried to contact the source of the images we
evaluated, and did not receive any response. As a result, we feel it is best if we do not
include the IR images in this report, but simply report on our findings.
The IR images we observed have much lower resolution and appear to have poorer focus.
than the WSI. Also, the dynamic range of the cameras appears to be less than the
dynamic range of the sky. That is, either the clouds or the sky are offscale, which would
make algorithm development (if required) difficult. On the positive side, we found that
the IR systems did not require a solar occultor, due to the relatively low brightness of the
sun at these wavelengths. This makes sense, since from Holst 2000 Table 6-1; the solar
radiation at ground level is only about 1.5 W/m², whereas the self-emission of a 290° k
target is 127 W/m² in the LWIR.
Regarding the dynamic range, the low opaque clouds appeared in the IR imagery to be
not only brighter than the background, but also offscale bright. We did not find any
examples of cirrus clouds detected by wide field of view IR imagers. In later searching,
24

we have found one example with cirrus clouds in the simultaneous visible image, and the
IR detector does not appear to have detected them.
We did not find any imagers with a cloud algorithm, beyond a simple threshold that is
tweaked from day to day depending on the haze level. If the some of the clouds are
offscale bright, as they appear to be, and some of the sky is offscale dark, as it appears to
be from the images, it may not be possible or easy to develop an algorithm to detect
cirrus, or to distinguish other thin clouds from haze. We also noted that the IR systems
appeared to have more noise. At least the IR imagers available at the time were also
fairly low resolution, except possibly for imagers that are far more costly than the visible
imagers. In Section 9.4, we further explore these impressions from a theoretical point of
view. First, however, we would like to comment briefly on the Mid Wave IR (MWIR).
9.3. Comments Regarding Mid Wave IR Systems
We are not aware of any groups attempting to provide routine whole sky cloud
monitoring using a Mid-wave IR (MWIR) system operating in the 3 – 5 um range.
However one of the IR camera manufacturers, Electrophysics, sent us sample narrow-
angle images taken in the 3 – 5 um region and the 8 - 12 um region, as shown in Figures
26 and 27. Both images were sent courtesy of C. Alicandro at Electrophysics.
Fig. 26. Sample cloud image taken with a cooled
camera operating in the 3 - 5 um region.
Fig. 27. Sample cloud image taken with an uncooled
Camera operating in the 8 - 12 um region.
From an evaluation of the energy levels, scattering and absorption mechanisms, we feel
that development of algorithms for the MWIR might be significantly more difficult than
development of algorithms for either the visible or the LWIR. Whereas the visible cloud
detection is driven primarily by scattering, and the LWIR cloud detection is driven
primarily by absorption and emission, the MWIR is impacted significantly by both
mechanisms (Zissis 1993, see Fig. 3.19). This is partly because the solar radiance is
fairly high in the MWIR. For example, Holst 2000 indicates that the solar ground
irradiance is about 24 W/m”, while the self-emission of a 290k blackbody is 4.1 W/m² in
the 3 – 5 um range. This may also imply that a solar occultor would be necessary in the
MWIR, whereas it is not in the LWIR. We decided not to evaluate the MWIR further at
this time for these reasons.
25
9.4. Theoretical Study of LWIR Characteristics
For the theoretical study, our goal was to evaluate how much the background sky
signature on a LWIR whole sky image might be expected to vary, and how the cloud
signature should compare with the background sky. More specifically, we wanted to
know whether there was sufficient difference in the signals to detect the clouds, and if so,
whether a cloud algorithm beyond a simple threshold would be required. In order to have
a better understanding, we wanted to derive this from the basic physics, although we
would like to make a more extensive evaluation in the future using standard atmospheric
models.
9.4.1. Discussion of the Approach
Although the source of the cloud signal is the thermal emission of the clouds, the signal is
not just that due to the temperature of the clouds. The radiance that reaches the sensor is
also impacted by the transmittance of the path of sight between the ground and the cloud,
and the path radiance, or radiance scattered into the path of sight from the surround. The
apparent radiance of any target t at range r may be given as
,L,(0,0)=,L,(0,0)*T,(0)+,L,(0,0)
where
L,(0,6) is the target radiance from range r at zenith angle 8 and azimuth angle o,
L. (0,0) is the target radiance from range 0 at zenith angle 8 and azimuth angle q,
T. (O) is the beam transmittance over the path from r to 0 at zenith angle o, and
L,(0,0) is the path radiance between the observer and the target at zenith angle 8 and
azimuth angle o.
The transmittance losses are very significant in the LWIR. For example, see Holst 2000
Figure 5-4, which illustrates that the transmittance at 10 um is on the order of .5 over a 2
km path length, and less than .l over a 10 km path length. Also, the path radiance may
be very significant in the LWIR. For example, see Holst 2000 Figure 5-8, which shows
an example of an image taken near the St. Louis Gateway Arch of path radiance due to
aerosol layers.
The LWIR sensors often are calibrated in Blackbody temperature in degrees (as in Fig.
27). After exploring a variety of theoretical approaches to our problem, we took the
following approach to calculating the cloud signature:
a) Determine typical cloud altitudes of interest, and the associated temperatures
b) Calculate the relative radiance for these temperatures
c) Correct for beam transmittance
d) Correct for path radiance
26
e) Convert back to effective temperature
f) Compare with the effective temperature of the backgrounds.
This will be discussed in more detail in the following sections.
9.4.2. Cloud Signatures, Part 1
Zissis 1993 (page 226) lists typical cloud altitudes in temperate regions as shown in Table
1, Column 2. For our sample calculations, we used the altitudes shown in Column 3.
The standard atmosphere temperatures (McCartney 1976 Table 2.6) were rounded to the
nearest 5°, and are shown in Column 4. Temperature extremes for 60° N Latitude in
winter and 30° N Latitude in summer were extracted from McCartney 1976 Appendix F,
and are shown in Columns 5 and 6.
Table 1
Cloud Altitudes used in Calculations

Altitude used
Cloud
Layer
Zissis
Altitude
Range
For
calculations
Temperatures Evaluated, °K.
Standard Cold
Warm
Atmosphere Extreme Extreme
Approximate 60° North 30° North
Temp
Winter
Summer
270
257
301
260
241
272
225
217
238
I
l
Low
Mid
0 - 2 km
2-7 km
5 - 13 km
1 km
5 km
10 km
L
L
To calculate the relative radiance for these clouds, we used the blackbody equation for
radiative exitance shown in Eq.2, as stated by Holst 2000 and Zissis 1993 Chapter 1.
M4,7)-> (2_) "
where:
M=Blackbody radiant exitance in W m2 um?
C1 = 3.7411 108 W um4 / m² (radiation constant)
C2 = 1.4388 104 um K (radiation constant)
a = wavelength in micrometers
T = temperature in degrees Kelvin
(Be careful not to confuse T, which is used for Temperature in Eq. 2, with Tr which is
used for transmittance in Eq. 1.) For our calculations, we used a wavelength of 10 um,
and used the exitance for all calculations, for consistency. Although this is not equal to
radiance, it should be proportional to radiance, and can be used if we are careful in the
handling not to mix units.
27

We also wanted to take into account the emissivity of the clouds, which describes the
efficiency with which objects radiate in comparison with a blackbody. From equations
given in Jacobs 1996, we estimate that the emissivity should be very close to 1 for the
opaque clouds. (Jacobs also reports the emissivity of the sky with values ranging from
about .2 to about .77, if we interpret him correctly.) We decided to make our calculations
using values of 1 for the low and the middle clouds. For high clouds, we used 1 for
opaque clouds, and .5 for thin clouds. By using a value of 1, we are reporting the best
possible behavior, in the sense that this would provide the best contrast for the IR imager
that is trying to detect the clouds. That is, we are being somewhat conservative in
making the IR system look as good as possible.
The next step is to determine the transmittance for use in Eq. 1. To do this, we first
extracted typical transmittances at 10 um over horizontal paths of 1 km, 2 km, and 5 km
from Holst 2000. We verified that the values yielded a consistent scattering coefficient at
10 um near ground level of about .35 km', and also verified in other references that this
value is reasonable. For the 1 km horizontal path, the resulting transmittance is about
.70.
The relationship between a horizontal transmittance and a vertical transmittance is
complex and variable, as it depends on the structure of the haze layers. To estimate
reasonable bounds on the vertical transmittance in relation to the horizontal
transmittance, we used measurements of vertical profiles of scattering coefficient taken
by our group from aircraft (Duntley et al. 1978). A sample scattering coefficient profile
is shown in Figure 28.
00
ALTITUDE ADLA
TOTAL VOLUME SCATTERIN
COEFFICIENT PEH ME
Fig. 28. Vertical profile of scattering coefficients
from Duntley et al. 1978.
The Duntley report includes the ground-based scattering coefficients, as well as the beam
transmittance computed from ground to altitude using the measured scattering coefficient
profiles. From this flight data, we found that a vertical path from the ground to 1 km
sometimes has a higher, and sometimes lower, transmittance than the horizontal path of 1
km. For a vertical path from the ground to 5 km and 10 km, we found that the
transmittance was about an average factor of .65 lower (for 5 km) and .52 lower (for 10
km). Thus we estimated a reasonable vertical transmittance to 1, 5, and 10 km at .70,.46,
and .36 as shown in Table 2. This is clearly an area where it would be useful to evaluate
the results with model scattering coefficient lapse rates, but we felt that for now this
approach would provide a reasonable assessment.
Table 2
Estimated Vertical Beam Transmittance
Using a Horizontal Transmittance over 1 km of.70

Cloud Vertical Tr. Estimated
Altitude Divided by Vertical
Horizontal | Transmittance
1 km Tr
1 km
.70
5 km
10 km
52
.65
.46
.36
To evaluate the cloud signatures at non-vertical look angles, we used the equation
T,(0)=T, (0)|secol
(3)
The next step is correcting for path radiance. However, we first would like to discuss the
clear sky background signatures, because we based the path radiance calculation on these
signatures.
9.4.3. Background Signatures
In evaluating quite a range of references, we considered either using measured
background signatures directly, or computing the signatures from other values such as the
effective emissivity of the atmosphere. Eventually, we decided to use a very simple
approach. Holst 2000 reports that a good rule of thumb for typical atmospheres is that
the effective temperature of the sky in the 8 – 12 um region is near 260° K overhead, and
near 293° K at the horizon, with a net variation of 33° K from the horizon to the zenith.
Furthermore, these temperatures vary with the ambient temperature.
We decided to use this information, and supplement it with information on how much the
ambient temperature might be expected to vary. Clearly, it would be good to do a more
extensive modeling study using Modtran in the future, but again we felt that this physical
approach would yield a better understanding of the variations in IR imagery and their
causes. We evaluated this rule of thumb in comparison with the sky measurements of
Jacobs 1996, and found them to be reasonably consistent, although Jacobs shows a larger
range of 49° K from horizon to zenith.
29
We looked at quite a few sources to evaluate how much these temperatures might be
expected to vary. First we looked at seasonal variations. From McCartney 1976, the
und temperature monthly means from winter at 60° N to summer at 30° N vary from
257° K to 301° K, for a net range of +22° K. Moran and Morgan 1991 list typical
CONUS (Continental US) ground temperatures as 265° K to 292° K in winter, and 292°
K to 312° K in summer, yielding a net variation in CONUS of+ 24° K.
In addition, there are diurnal variations. Moran and Morgan 1991 list typical diurnal
variations from the normal daily maximum to the normal daily minimum as about # 8º in
the winter, and +7° in the winter.
In addition, there is significant variation as a result of changes in aerosol and air mass.
For example, in Zissis 1993, Figures 3.22 and 3.23 show the spectral irradiance of a clear
nighttime sky for several angles of elevation above the horizon for stations in Colorado
and in Florida. Extracting the approximate spectral radiance in uW cm srum near 10
um from the plots, we find the following. The spectral radiance at the zenith is about 60
for Colorado and about 260 for Florida, and the spectral radiance near the horizon is
about 807 for Colorado and about 920 for Florida. In this example, the aerosol and air
mass made a very significant difference, especially at the zenith.
Some of the implications of these background signatures are as follows. Within a single
image, the effective background temperatures will be significantly higher at the horizon
than the zenith, with deltas of about 33 to 49° K, and potentially different variations
depending on air mass and aerosol considerations. From one image to another, we can
expect diurnal variations of roughly + 22 to 24° K, and diurnal variations of roughly + 7
to 8° K. Overall, taking the zenith to horizon range of 260 to 293° K from Holst 2000,
and adding these seasonal and diurnal variations, that means the sensor can expect to see
effective background temperatures ranging from about 228 to 325° K, if the extremes and
the differences in aerosol are not considered. We got detailed specifications on 3
cameras that would be reasonable to consider for this application, and their minimum
temperature readings, at the time of this study, were 257, 257, and 237° K. We will use
an estimated minimum of 247° K for this study. This implies that for some
environments, the background sky signature may be offscale dark, and thus very thin
clouds may also be offscale dark.
For our calculations, we wished to calculate the cloud and the background signatures for
low, middle, and high clouds, as defined in Table 1. We included computations for the
standard atmosphere, 60° N winter, and 30° N summer. At the time of this study, we
were told that we would like to sense to 0 to 5° above the horizon (85 - 90° zenith angle),
and must sense down to at least 20° above the horizon (70° zenith angle). Knowing that
requirements are sometimes compromised, we decided to calculate the IR signatures for
zenith angles of 0°, 60°, and 85° zenith angles. We based the ground temperature
variations on McCartney 1976, and the zenith angle effective temperature variations on
Holst 2000. The result was the effective background temperatures shown in Table 3.
Table 3
Effective Radiating Temperatures used for Sky Backgrounds

| Zenith Angle Standard Atm
Summer 30 N
0°
268
Winter 60 N
224°
247°
255°
260°
283°
291°
60°
291°
299°
850
9.4.4. Cloud Signatures, Part 2
The remaining step in the cloud signature calculation was applying the correction for the
path radiance. The effective background temperatures discussed in Section 9.4.3 are
essentially the temperatures that correspond to the path radiance from earth to space. The
path radiance from ground to altitude should be less than the path radiance from ground
to space, although they should be quite close for high altitudes. The largest contribution
to the path radiance will be from the lower warmer layers, so even for the lower cloud
altitudes, the path radiance may not be that much less than the earth to space path
radiances.
We decided to use the ground-to-space path radiances derived from the background
temperatures in calculating the cloud signatures, for two reasons. First, the ground-to-
space path radiance bounds the problem, and we do not expect large differences between
the ground-to-space and most ground-to-altitude path radiances, as discussed above.
Second, if we have an error in our background radiances, it will affect both the cloud and
background radiances equally, and thus have minimal effect on the conclusions. By
doing this, we are once again taking a very conservative approach, i.e. calculating
performance of the IR sensors that is better than the real performance, because by using a
path radiance higher than the true path radiance, the calculations will yield higher
contrast with the background than would actually be obtained.
To illustrate how this information is put together, we provide a sample calculation. For
an opaque cloud at 10 km at the zenith in the standard atmosphere, we use a cloud
temperature of 225 K (Table 1). Using Eq. 2 yields a relative radiance of 6.26. The
emittance of the cloud is estimated to be near 1. The transmittance for the zenith
direction and a cloud altitude of 10 km is estimated to be .36 (Table 2). The product of
the relative radiance, emittance, and transmittance yields a transmitted relative radiance
of 2.25. The background effective temperature, from Table 3, is estimated to be 260 K.
Using Eq. 2 to convert this to a relative radiance yields a relative path radiance of 14.8.
Using Eq. 1 to combine the transmitted relative radiance and the relative path radiance
yields a net effective radiance for the cloud of 17.1. Using the inverse of Eq. 2 to derive
the effective temperature corresponding to this relative radiance yields an effective
temperature for the cloud, as detected from the ground, of 266.8 K.
31

9.4.5. Resulting Cloud and Background Effective Temperatures
The results of the calculations are shown in Tables 4 – 6. In each of these tables, the
cloud altitude and zenith angle are shown in Columns 1 and 2. Columns 3, 5, and 7 show
the effective temperatures for the clouds, as sensed from the ground, and columns 4, 6,
and 8 show the temperature difference between the cloud and the background. In the
bottom row, we show the background effective temperature, as sensed from the ground,
and extracted from Table 3. The colors used in the table will be explained below.
Table 4
Computed Cloud and Background Results for the Zenith
(Colors explained in Text)
Standard Atm Winter 60N Summer 30N
Alt
(km)
1
Zen Eff | AT
Temp
0 298.2 38.2
Eff AT Eff | AT
Temp Temp
263.0 39.0 315.8 47.8
279.0 29.0 244.7 20.7 | 289.7 | 21.7
266.8
233.3
9.3 |276.6
8.6
10 Thin
263.5
3.5
228.9
4.9
272.4
4.4
Aerosol
260.0
224.0
268.0
There are four colors used in these tables. The light blue cells are cases where the cloud
signal is lower than the anticipated minimum sensitivity range of 247K for the IR sensor.
That is, in these cases, we expect the signal to be offscale dark, and it should not be
possible to sense the clouds or background. These cases occur in the Winter 60° N case
for the zenith. At 60° zenith angle, the clouds have come onscale, but the background is
still offscale dark, so that algorithm development would be difficult.
In Table 4, there are 3 cells that are colored yellow. Yellow indicates those cases for
which the difference between the cloud and the background is greater than the variation
in background from the zenith to 85° zenith angle. For example, in the standard
atmosphere, the difference in the background between 0 and 85° zenith angle was
estimated to be 31° K (Table 3). The low altitude clouds are about 38° K warmer than the
background. Thus, they should be detectable with a simple threshold cloud algorithm.
However, the clouds at other altitudes, as well as other zenith angles, have less
temperature difference from their background. For all the cases not marked with yellow,
we do not expect a simple threshold to work well, because it will either miss most of the
clouds, or call the horizon a cloud.
32

Table 5
Computed Cloud and Background Results for 60° Zenith Angle
(Colors explained in Text)
Standard Atm Winter 60N Summer 30N
Alt |Zen | Εff | ΔΤΙ Εff | ΔΤ | Εff | ΔΤ
(km)
Temp
Temp
1 | 60 | 305.7 | 22.7 | 269.1 22.0 | 320.3 29.3
Temp
290.1
7.1 254.3
7.3 299.3
8.3
10
284.9
1.9
| 249.4
2.4 | 293.5
2.5
10 Thin
284.0
1.0
248.2
1.2 |292.2
1.2
Aerosol
283.0
247.0
291.0
Table 6
Computed Cloud and Background Results for 85° Zenith Angle
(Colors explained in Text)
Standard Atm Winter 60N
Summer 30N
Zen
Eff
Temp
AT
Alt
(km)
1
Eff AT Eff | AT
Temp
Temp
255.8 0.78 300.1 1.06
85 291.8
0.8
291.0 .004
255.0.005
299.0 .005
5
10
10.
291.0.00007|255.01.0001 299.0 .001
0 Thin
| 291.0.00002|255.0.00007|299.0.00008
Aerosol
291.0
255.0
299.0
In Table 4, there are 4 cells marked with a light pink. These are cases where the
difference between the cloud and the background is less than 14 of the 31° K range in the
background. These are cases where the algorithm would have to be reasonably
sophisticated. The algorithm would need to detect a delta that is somewhat small in
comparison with the variation within an image. The difference between the cloud and the
33
background is even smaller in comparison with the background variation from image to
image. The need to take into account seasonal, diurnal, and other variations in the
background signal would make the algorithm more complicated.
In Table 5, and 60° zenith angle, all of the middle and high clouds except for one summer
30° latitude case are colored light pink, indicating that we would expect a reasonably
sophisticated algorithm to be needed to automatically recognize the clouds.
In Table 6, there is a dark pink category. This represents the cases where the temperature
differences between the cloud and background are less than the NEDT (Noise Equivalent
Delta Temperature) for the sensors we evaluated. That is, the clouds are buried in the
system noise, and would not be detected, according to these calculations.
Thus these calculations imply that for the conditions we studied, the IR sensor would not
sense middle or high clouds near the horizon. The sensor would require a reasonably
sophisticated algorithm to identify the presence of middle and high clouds at other angles.
At the zenith, identifying the presence of low clouds would be simple, and even without a
sophisticated algorithm. Middle and high clouds at the zenith would require a
sophisticated algorithm, and may be offscale dark at the zenith in winter in cold
environments.
Clearly, this is not a definitive study, and ideally we would like to expand the study both
with measurements and with modeling studies. However, we feel that these results give
us some understanding of what to expect from IR systems, and why. We were pleased
that our sponsors indicated that these results fit very well with experimental results they
have seen but that have not been published.
9.5. Summary of the Wavelength Analysis
As discussed in Section 9.1, we do not believe that sensors at 1.6 um will do a better job
for our application, because these cameras tend to provide a poorer image, and we do not
believe their sensitivity will be adequate for night-time applications. We did not evaluate
MWIR sensors in detail, because they appear not to provide any significant benefits, and
the algorithms would be more complex.
There are some LWIR systems in development, and at this point, their imagery does not
appear to be as good as that obtained in the visible. A theoretical analysis of the cloud
and background signals indicates that although low clouds at the zenith are easy to detect,
sophisticated algorithms will probably be required to detect low clouds at angles away
from the zenith. Also, a reasonably sophisticated algorithm will probably be required to
detect middle and high clouds at most angles. Close to the horizon, middle and high
clouds are expected to be buried in the noise. In cold environments, the middle and high
clouds at the zenith may be offscale dark.
In order to successfully identify the presence of clouds, two things are necessary. First,
the raw imagery must have sufficient difference between the cloud and background
signals. Second, algorithms must be able to sort out these differences and identify the
presence of the clouds. Our initial analysis leads us to believe that for an IR system, the
first constraint may not occur for many needed angles and cloud heights, and reasonably
sophisticated algorithms will be required for most angles and cloud heights. By contrast,
as illustrated in earlier sections, the visible sensors provide very high quality imagery,
and clouds are well detected down to the horizon, and at all cloud altitudes.
If cost constraints limit the application to a single sensor, we feel that the visible sensors
will be very much superior, because they have already demonstrated that they can detect
high and low clouds night and day, overhead and near the horizon, some of which may be
offscale or in the noise for IR sensors at the present time. Furthermore, the algorithms
are much more developed than the IR system algorithms we are aware of at the present
time, although the algorithms are continuing to be advanced even in the visible.
Also, because the transmission losses at 1.6 um are primarily driven by scattering
mechanisms, like the visible, whereas the thermal IR cloud detection is based on thermal
emission, we feel we have a better chance of detecting the clouds of importance to the
SWIR. We make these statements with the caveat that it may at some point become a
priority to further evaluate these beliefs, using measurements and modeling studies. In
particular, we note that it may turn out that a hybrid system, using both a visible and an
IR sensor, though much more costly, could provide enough advantage to be worth
evaluating.
10. Summary
During this contract period, significant progress was made in providing day and night
algorithms, in providing WSI hardware and maintenance, and in evaluation of the best
wavelength regimes to use for the application. We believe that we have completed all
contract requirements.
11. Acknowledgements
We would like to express our appreciation to the personnel of Starfire Optical Range and
their contractors from Boeing. Dr. Earl Spillar, the head of this program, Ann Slavin, our
acting contract monitor, and Dr. Robert Fugate, the director of SOR at the time of this
contract, were all helpful, and have been a pleasure to work with. We feel this work is
valuable, and very much appreciate having had the chance to advance the state of the art,
as well as meet our sponsor's specific needs. We also would like to express our
appreciation to ONR, which provided the funding vehicle for this work.
35
12. References
In-house Technical Memoranda, in order by memo number; available to sponsor on
request
Karr, M., “Unit 12 Software Updates”, Atmospheric Optics Group Technical
Memorandum AV01-029t, 29 January 2001
Karr, M., “Update to ProcWSI for Unit 13/14”, Atmospheric Optics Group Technical
Memorandum AV01-030t, 29 January 2001
Shields, J., “SOR Trip Report, Jan '01”, Atmospheric Optics Group Technical
Memorandum AV01-031t, 28 February 2001
Shields, J., “Procedure for Processing SOR Cloud Library Data”, Atmospheric Optics
Group Technical Memorandum AV01-032t, 28 February 2001
Karr, M., “SOR RunWSI Updates”, Atmospheric Optics Group Technical Memorandum
AV01-033t, 28 February 2001
Karr, M., “SOR ProcWSI Updates”, Atmospheric Optics Group Technical Memorandum
AV01-034t, 28 February 2001
Shields, J., “SOR Clear Sky Background Library File Generation for Unit 12”,
Atmospheric Optics Group Technical Memorandum AV01-035t, 6 April 2001
Burden, A. R., “Basic Geometric Calibration of WSI Unit 12”, Atmospheric Optics
Group Technical Memorandum AV01-041t, 10 June 2001
Burden, A. R., “Unit 12 Rolloff Calibration", Atmospheric Optics Group Technical
Memorandum AV01-042t, 18 June 2001
Burden, A. R., “Unit 12 Flat Field Calibration", Atmospheric Optics Group Technical
Memorandum AV01-043t, 23 June 2001
Burden, A. R., “Unit 12 Linearity and Exposure Calibration Results”, Atmospheric
Optics Group Technical Memorandum AV01-044t, 23 June 2001
Burden, A. R., “Unit 12 Effective Lamp Irradiance Computation", Atmospheric Optics
Group Technical Memorandum AV01-045t, 20 June 2001
Burden, A. R., “Unit 12 Aperture Calibrations”, Atmospheric Optics Group Technical
Memorandum AV01-046t, 22 June 2001
Burden, A. R., “Absolute Calibration Results for Unit 12”, Atmospheric Optics Group
Technical Memorandum AV01-047t, 20 June 2001
36
Burden, A. R., “Unit 12 Calibration Summary”, Atmospheric Optics Group Technical
Memorandum AV01-048t, 27 June 2001
Burden, A. R., “Basic Geometric Calibration of WSI Unit 13”, Atmospheric Optics
Group Technical Memorandum AV01-049t, 10 June 2001
Burden, A. R., “Unit 13 Rolloff Calibration", Atmospheric Optics Group Technical
Memorandum AV01-050t, 28 June 2001
Burden, A. R., “Unit 13 Flat Field Calibration", Atmospheric Optics Group Technical
Memorandum AV01-051t, 27 June 2001
Burden, A. R., “Unit 13 Linearity and Exposure Calibration Results", Atmospheric
Optics Group Technical Memorandum AV01-052t, 26 June 2001
Burden, A. R., “Unit 13 Effective Lamp Irradiance Computation", Atmospheric Optics
Group Technical Memorandum AV01-053t, 22 June 2001
Burden, A. R., “Unit 13 Aperture Calibrations", Atmospheric Optics Group Technical
Memorandum AV01-054t, 26 June 2001
Burden, A. R., “Unit 13 Acrylic Dome Calibration", Atmospheric Optics Group
Technical Memorandum AV01-055t, 26 June 2001
Burden, A. R., "Absolute Calibration Results for Unit 13”, Atmospheric Optics Group
Technical Memorandum AV01-056t, 26 June 2001
Burden, A. R., “Unit 13 Calibration Summary”, Atmospheric Optics Group Technical
Memorandum AV01-057t, 29 June 2001
Burden, A. R., “Basic Geometric Calibration of WSI Unit 14”, Atmospheric Optics
Group Technical Memorandum AV01-058t, 10 June 2001
Burden, A. R., “Unit 14 Rolloff Calibration", Atmospheric Optics Group Technical
Memorandum AV01-059t, 29 June 2001
Burden, A. R., “Unit 14 Flat Field Calibration", Atmospheric Optics Group Technical
Memorandum AV01-060t, 27 June 2001
Burden, A. R., “Unit 14 Linearity and Exposure Calibration Results”, Atmospheric
Optics Group Technical Memorandum AV01-061t, 27 June 2001
Burden, A. R., “Unit 14 Effective Lamp Irradiance Computation”, Atmospheric Optics
Group Technical Memorandum AV01-062t, 27 June 2001
37
Burden, A. R., “Unit 14 Aperture Calibrations”, Atmospheric Optics Group Technical
Memorandum AV01-063t, 27 June 2001
Burden, A. R., “Unit 14 Acrylic Dome Calibration", Atmospheric Optics Group
Technical Memorandum AV01-064t, 27 June 2001
Burden, A. R., “Absolute Calibration Results for Unit 14”, Atmospheric Optics Group
Technical Memorandum AV01-065t, 27 June 2001
Burden, A. R., “Unit 14 Calibration Summary”, Atmospheric Optics Group Technical
Memorandum AV01-066t, 29 June 2001
Burden, A. R., “CalWSI FlatField - IDL Software for Flat Field and Rolloff
Calibrations”, Atmospheric Optics Group Technical Memorandum AV01-067t, 29 June
2001
Baker, J., “Unit 13 Trolley Failure", Atmospheric Optics Group Technical Memorandum
AV01-068t, 27 June 2001
Shields, J., “Research Approach for High Resolution Night Cloud Algorithms”,
Atmospheric Optics Group Technical Memorandum AV01-069t, 3 July 2001
Shields, J., “Return of SOR Unit 14 from Northern California, 3 July 01”, Atmospheric
Optics Group Technical Memorandum AV01-070t, 10 July 2001
Shields, J., “Unit 13 Parts List”, Atmospheric Optics Group Technical Memorandum
AV01-080t, 6 September 2001
Karr, M., "SOR Unit 13 and 14 run time history”, Atmospheric Optics Group Technical
Memorandum AV01-092t, 28 September 2001
Karm, M. CSOR Voi:103 and
Shifeles urne r Shuter
Shields, J., “Camera Shutter Delays under Windows and SOR Software”, Atmospheric
Optics Group Technical Memorandum AV01-094t, 2 October 2001
Shields, J. and A. Burden, “Derivation of Star Irradiance for the Beam Transmittance
Project”, Atmospheric Optics Group Technical Memorandum AV01-097t, 9 October
2001
Burden, A., “Improvements Made to Starlight-Based WSI Cloud Algorithm”,
Atmospheric Optics Group Technical Memorandum AV03-030t, 19 May 2003
Burden, A., “ImgCal: New Software for the Radiometric Calibration of D/N WSI
Images”, Atmospheric Optics Group Technical Memorandum AV03-037t, 9 September
2003
Karr, M., “Control Computer operations overview for WSI units 13 and 14, Version 4.2”,
Atmospheric Optics Group Technical Memorandum AV03-047t, 06 November 2003
Karr, M., “Display/Processing Computer operations overview for WSI Units 13 and 14,
Version 2.1.", Atmospheric Optics Group Technical Memorandum AV03-048t, 06
November 2003
Baker, J., “Filter Changing Alignment Procedure for SOR Unit 12”, Atmospheric Optics
Group Technical Memorandum AV03-049t, 07 November 2003
Baker, J., “SOR WSI Unit 12 Field Diagnostics and Repair Trip - October 2003”,
Atmospheric Optics Group Technical Memorandum AV03-052t, 24 November 2003
Baker, J., and J. Shields, "SOR-TASC Virginia Unit 14 Install Trip - November 2003”,
Atmospheric Optics Group Technical Memorandum AV03-054t, 04 December 2003
Baker, J., “Acrylic Dome Installation Procedure”, Atmospheric Optics Group Technical
Memorandum AV03-055t, 08 December 2003
Baker, J., “Unit 13 and 14 Trolley Calibration and Alignment Procedure (Revised)”,
Atmospheric Optics Group Technical Memorandum AV03-056t, 12 December 2003
Baker, J., “Unit 13 & 14 Trolley Cart Removal and Fixed Shade Installation Procedure”,
Atmospheric Optics Group Technical Memorandum AV04-001t, 08 January 2004
Shields, J., “Unit 14 Radiometric Calibrations”, Atmospheric Optics Group Technical
Memorandum AV04-002t, 19 January 2004
Burden, A., “Earth-To-Space Beam Transmittance from Nighttime WSI Imagery:
Preliminary Results”, Atmospheric Optics Group Technical Memorandum AV04-004t,
20 January 2004
Shields, J., “Unit 14 Parts list and Overview”, Atmospheric Optics Group Technical
Memorandum AV04-007t, 25 March 2003
Baker, J., “Upgrading the WSI flow package from Series 100 to Series 800”,
Atmospheric Optics Group Technical Memorandum AV04-009t, 08 April 2004
Baker, J., “Unit 12 Arc Calibration and Alignment Procedure Revisited”, Atmospheric
Optics Group Technical Memorandum AV04-010t, 08 April 2004
Baker, J., “Unit 12 Trolley Calibration and Alignment Procedure Revisited”,
Atmospheric Optics Group Technical Memorandum AV04-011t, 08 April 2004
Shields, J., “Summary of the NIR Day Cloud Algorithm for the German Day WSI”,
Atmospheric Optics Group Technical Memorandum AV04-014t, 9 Apr 2004
39
Baker, J., “Unit 13 & 14 Environmental Housing Temp Sensor Replacement",
Atmospheric Optics Group Technical Memorandum AV04-019t, 07 May 2004
Shields, J., “Shutter Tests with Windows-based D/N WSI Systems using Units 13 and 2”,
Atmospheric Optics Group Technical Memorandum AV04-024t, 21 May 2004
Burden, A., “Unit 13 Effective Lamp Radiances for Field Calibration", Atmospheric
Optics Group Technical Memorandum AV04-025, 28 May 2004
Baker, J., “WSI System Checklist for Unit 13&14”, Atmospheric Optics Group Technical
Memorandum AV04-026t, 01 June 2004
Burden, A., “Unit 13 Field Calibration Results”, Atmospheric Optics Group Technical
Memorandum AV04-027t, 02 June 2004
Baker, J., “Filter Changer Alignment procedure for SOR Unit 13&14”, Atmospheric
Optics Group Technical Memorandum AV04-029t, 09 July 2004
Karr, M., “Revised Field calibration program, Fieldcal for Windows Version 1.0”,
Atmospheric Optics Group Technical Memorandum AV04-030t, 16 July 2004
Baker, J., “Trolley Drive Gear and Chain Adjustment Procedures – Unit 13&14”,
Atmospheric Optics Group Technical Memorandum AV04-031t, 20 July 2004
Baker, J., “Unit 14 SOR/TASC Virginia Trip – Trolley Drive Analysis and Repair
4/12/04”, Atmospheric Optics Group Technical Memorandum AV04-032t, 20 July 2004
Ora
Baker, J., “Unit 13 Updated parts List and Overview”, Atmospheric Optics Group
Technical Memorandum AV04-033t, 20 July 2004
Baker, J., “Unit 14 Updated Parts List and Overview”, Atmospheric Optics Group
Technical Memorandum AV04-034t, 20 July 2004
Baker, J., “Unit 14 SOR/TASC Virginia Trip – Camera Install and FC Repair 6/7/04”, .
Atmospheric Optics Group Technical Memorandum AV04-035t, 20 July 2004
Shields, J., "Processing of SGP Feb 02 Data for Cloud Decision", Atmospheric Optics
Group Technical Memorandum AV04-038t, 31 August 2004
Shields, J., “Processing of SGP Mar 02 Data for Cloud Decision”, Atmospheric Optics
Group Technical Memorandum AV04-039t, 31 August 2004
Karr, M., “Image Utility Programs”, Atmospheric Optics Group Technical Memorandum
AV04-043t, 22 September 2004
40
Baker, J., “Procedure for Changing WSI Shutter in Units 13&14”, Atmospheric Optics
Group Technical Memorandum AV04-044t, 28 September 2004
Burden, A., “Unit 14v1 and 13v2 Effective Lamp Radiances for Field Calibration”,
Atmospheric Optics Group Technical Memorandum AV04-045t, 12 August 2004
Burden, A., “Unit 14v1 and 13v2 Field Calibration Results”, Atmospheric Optics Group
Technical Memorandum AV04-046t, 12 August 2004
Shields, J., “Abnormal Red/blue ratios in Aug 02 SGP D/N WS Data Set”, Atmospheric
Optics Group Technical Memorandum AV04-047t, 14 October 2004
Shields, J., “Processing of SGP August 02 Data for Cloud Decision”, Atmospheric Optics
Group Technical Memorandum AV04-049t, 23 October 2004
Shields, J., “Reprocessing of SGP Feb and Mar 02 Data for Cloud Decision",
Atmospheric Optics Group Technical Memorandum AV04-050t, 03 November 2004
Shields, J., “Initial Virginia night data analysis”, Atmospheric Optics Group Technical
Memorandum AV04-054t, 31 December 2004
Published References and Technical Notes in order by date
C
McCartney, Earl J., (1976), “Optics of the Atmosphere – Scattering by Molecules and
Particles”, Wiley Series in Pure and Applied Optics, Stanley S. Ballard, Advisory Editor,
John Wiley and Sons, New Your, London, Sydney, Toronto
Duntley, Seibert Q., R. W. Johnson, and J. I. Gordon (1978), “Airborne Measurements of
Atmospheric Volume Scattering Coefficients in Northern Europe, Summer, 1977”,
Visibility Laboratory, Scripps Institution of Oceanography, University of California San
Diego, SIO 78-28, AFGL-TR-78-0168
Hobbs, Peter V., and Adarsh Deepak, (1981), “Clouds, Their Formation, Optical
Properties, and Effects”, Academic Press, A Subsidiary of Harcourt Brace Jovanovich,
Publishers, New York, London, Toronto, Sydney, San Francisco. We referenced articles
in this book by James Jiusto and by Stephen K. Cox.
Johnson, R. W., W. S. Hering and J. E. Shields (1989), “Automated Visibility and Cloud
Cover Measurements with a Solid-State Imaging System", Marine Physical Laboratory,
Scripps Institution of Oceanography, University of California San Diego, SIO 89-7, GL-
TR-89-0061, NTIS No. ADA216906
Johnson, R. W., J. E. Shields, and T. L. Koehler (1991), “Analysis and Interpretation of
Simultaneous Multi-Station Whole Sky Imagery”, Marine Physical Laboratory, Scripps
Institution of Oceanography, University of California San Diego, SIO 91-3, PL-TR-91-
2214
Physical Laboratory, Scrippe
41
Moran, Joseph M., and Michael D. Morgan, (1991), "Meteorology – The Atmosphere
and The Science of Weather", Third Edition, Macmillan Publishing Company, New
York, and Collier Macmillan Canada, Toronto
Shields, J. E., R. W. Johnson, and T. L. Koehler, (1993), “Automated Whole Sky
Imaging Systems for Cloud Field Assessment”, Fourth Symposium on Global Change
Studies, 17 – 22 January 1993, American Meteorological Society, Boston, MA
Zissis, George J., Editor (1993), "Volume 1: Sources of Radiation” of “The Infr
Electro-Optical Systems Handbook, Joseph S. Accetta and David L. Shumaker Executive
Editors, Copublished by Infrared Information Analysis Center, Environmental Research
Institute of Michigan, Ann Arbor, Michigan, and by SOIE Optical Engineering Press,
Bellingham, Washington, USA. We referenced Chapter 1 by William Wolfe, and
Chapter 3 by David Kryskowski and Gwynn Suits.
Shields, J. E., R. W. Johnson, and M. E. Karr, (1994), “Upgrading the Day/Night Whole
Sky Imager from Manual/Interactive to Full Automatic Control, Marine Physical
Laboratory, Scripps Institution of Oceanography, University of California San Diego,
Report MPL-U-140/94
Jacobs, Pieter A., (1996), “Thermal Infrared Characterization of Ground Targets and
Backgrounds”, Tutorial Texts in Optical Engineering volume TT26, Donald O'Shea
Series Editor, SPIE Optical Engineering Press, A Publication of SPIE – The International
Society for Optical Engineering, Bellingham, Washington
Shields, J. E., R. W. Johnson, M. E. Karr, R. A. Weymouth, and D. S. Sauer, (1997a),
“Delivery and Development of a Day/Night Whole Sky Imager with Enhanced Angular
Alignment for Full 24 Hour Cloud Distribution Assessment”, Marine Physical
Laboratory, Scripps Institution of Oceanography, University of California San Diego,
Report MPL-U-8/97
Shields, J. E., M. E. Karr, and R. W. Johnson, (1997b), “Service Support for the Phillips
Laboratory Whole Sky Imager", Marine Physical Laboratory, Scripps Institution of
Oceanography, University of California San Diego, Report MPL-U-10/97
Shields, J. E., R. W. Johnson, M. E. Karr, and J. L. Wertz, (1998), “Automated
Day/Night Whole Sky Imagers for Field Assessment of Cloud Cover Distributions and
Radiance Distributions”, Tenth Symposium on Meteorological Observations and
Instrumentation, 11 – 16 January 1998, American Meteorological Society, Boston, MA
Feister, U., Shields, J., Karr, M., Johnson, R., Dehne, K. and Woldt, M. (2000), “Ground-
Based Cloud Images and Sky Radiances in the Visible and Near Infrared Region from
Whole Sky Imager Measurements”, Proceedings of Climate Monitoring - Satellite
Application Facility Training Workshop sponsored by DWD, EUMETSAT and WMO,
Dresden 2000.
Holst, Gerald C., (2000), “Common Sense Approach to Thermal Imaging” Copublished
by JCD Publishing, Winter Park, Florida, and SPIE Optical Engineering Press, A
Publication of SPIE – The International Society for Optical Engineering, Bellingham,
Washington
Shields, J. E., M. E. Karr, A.R. Burden, R.W. Johnson, and J. G. Baker, (2002), “Analytic
Support for the Phillips Laboratory Whole Sky Imager, 1997 - 2001”, Marine Physical
Laboratory, Scripps Institution of Oceanography, University of California San Diego.
Shields, J. E., R.W. Johnson, M. E. Karr, A.R. Burden, and J. G. Baker, (2003a), “WSI
Field Calibration System Operations Manual”, Marine Physical Laboratory, Scripps
Institution of Oceanography, University of California San Diego, Technical Note 252,
February 2003.
Shields, J. E., M. E. Karr, A.R. Burden, R.W. Johnson, and J. G. Baker, (2003b),
“Analysis and Measurement of Cloud Free Line of Sight and Related Cloud Statistical
Behavior – Published as Final Report for ONR Contract N00014-97-D-0350 DO #2”,
Marine Physical Laboratory, Scripps Institution of Oceanography, University of
California San Diego, Technical Note 262, June 2003.
Shields, J. E., R. W. Johnson, M. E. Karr, A. R. Burden, and J. G. Baker (2003c),
Calibrated Fisheye Imaging Systems for Determination of Cloud Top Radiances from a
UAV, International Symposium on Optical Science and Technology, SPIE the
International Society for Optical Engineering, 2003.
Shields J. E., R. W. Johnson, M. E. Karr, A. R. Burden, and J. G. Baker, (2003d),
Daylight Visible/NIR Whole Sky Imagers for Cloud and Radiance Monitoring in Support
of UV Research Programs, International Symposium on Optical Science and Technology,
SPIE the International Society for Optical Engineering, 2003.
Shields, J. E., R. W. Johnson, M. E. Karr, A. R. Burden, and J. G. Baker (2003e), Whole
Sky Imagers for Real-time Cloud Assessment, Cloud Free Line of Sight Determinations
and Potential Tactical Applications, The Battlespace Atmospheric and Cloud Impacts on
Military Operations (BACIMO) Conference, Monterey, CA.
http://www.nrlmry.navy.mil/bacimo.html, 2003.
Shields, J. E., A.R. Burden, M. E. Karr, R.W. Johnson, and J. G. Baker, (2004a),
“Development of Techniques for Determination of Nighttime Atmospheric Transmittance
and Related Analytic Support for the Whole Sky Imager - Published as Final Report for
ONR Contract N00014-01-D-0043 DO #5”, Marine Physical Laboratory, Scripps
Institution of Oceanography, University of California San Diego, Technical Note 263,
April 2004.
Shields, J. E., M. E. Karr, A.R. Burden, R.W. Johnson, and J. G. Baker, (2004b), “Project
Report for Providing Two Day/Night Whole Sky Imagers and Related Development
43.
Work for Starfire Optical Range – Published as Final Report for ONR Contract N00014-
97-D-0350 DO #6”, Marine Physical Laboratory, Scripps Institution of Oceanography,
University of California San Diego, Technical Note 265, May 2004.
Shields, J. E., J. G. Baker, M. E. Karr, R. W. Johnson, and A. R. Burden, (2005a),
Visibility measurements along extended paths over the ocean surface, International
Symposium on Optical Science and Technology, SPIE the International Society for
Optical Engineering, August 2005.
Shields, J. E., A.R. Burden, R.W. Johnson, M. E. Karr, and J. G. Baker, (2005b), “Cloud
Free Line of Sight Probabilities and persistence Probabilities from Whole Sky Imager
Data”, Marine Physical Laboratory, Scripps Institution of Oceanography, University of
California San Diego, Technical Note 266, August 2005.
Shields, J. E., A. R. Burden, R. W. Johnson, M. E. Karr, and J. G. Baker (2005c),
Measurement and Evaluation of Cloud Free Line of Sight with Digital Whole Sky
Imagers, The Battlespace Atmospheric and Cloud Impacts on Military Operations
(BACIMO) Conference, Monterey, CA. http://www.nrlmry.navy.mil/bacimo.html, 2005.
Shields, J. E., R. W. Johnson, J. G. Baker, M. E. Karr, and A. R. Burden, (2006),
Multispectral scattering measurements along extended paths using an imaging system,
International Symposium on Optical Science and Technology, SPIE the International
Society for Optical Engineering, August 2006.