UNIVERSITY OF CALIFORNIA, SAN DIEGO UC SAN DIEGO LIBRARY 3 1822 04429 7463 OPTICAL SYSTEMS GROUP TECHNICAL NOTE NO, 220 June 1990 Offsite (Annex-jo rnals) 974.5 • T43 no. 220 THE APPLICATION OF WSI CLOUD STATISTICS TO ASAT ENGAGEMENT SCENARIOS TECHNICAL OUTLINE AS OF DRAFT, 30 JUNE 1990 R. W. Johnson UNIVERSITY OF CALIFORNIA SAN DIEGO SITY TERS UNO THE. ORNIA 1868 ST SCRIPPS INSTITUTION OF OCEANOGRAPHY MARINE PHYSICAL LAB San Diego, CA 92152-6400 UNIVERSITY OF CALIFORNIA, SAN DIEGO . WRISKE w TAASISW ww . ww.www. . WASAP WASIWGRESSwwwwwxx . SASS.. WW. twist W 3 1822 04429 7463 TECHNICAL OUTLINE AS OF DRAFT, 30 JUNE 1990 I. Front Cover & UCSD Forms Research Plan II. A. Discussion 1. Introduction 2. Basic Concepts 3. Implementation a. Data Acquisition b. Supporting Instrumentation c. Application to ASAT Scenarios d. Data Processing, Interpretation and Analysis e. Alternative Products f. Hardware Options B. Budget Review & Summary .1- PROPOSAL NO.: UCSD-90 IRS #95-6006144-W Congressional District #41 THE REGENTS OF THE UNIVERSITY OF CALIFORNIA UNIVERSITY OF CALIFORNIA, SAN DIEGO Scripps Institution of Oceanography, Mail Code A-010 La Jolla, California 92093 Telephone (619) 534-4570 PROPOSAL FOR RESEARCH TO BE CONDUCTED UNDER THE SPONSORSHIP OF Title of Proposal: Project Period: Amount Requested: Agency Contract or Grant No.: Principal Investigator: Richard W. Johnson Principal Development Engineer Marine Physical Laboratory of the Scripps Institution of Oceanography University of California, San Diego San Diego, California 92152 (619) 534-1772 Make checks payable to: The Regents of the University of California Mail to: Accounting (Fiscal) Officer, UCSD OFFICER TO WHOM AWARD DOCUMENTS SHOULD BE MAILED: N. J. Sattler, Manager SIO Contract and Grants Administration, A010 Scripps Institution of Oceanography Date: Signature Principal Investigator(s): Date: . . .. ... . . ... . .. Signature Typed Name & Title: Kenneth M. Watson, Director OFFICIALS AUTHORIZED TO SIGN FOR INSTITUTION: Date: Signature Typed Name & Title: Date: wignature Typed Name & Title: II. RESEARCH PLAN A. DISCUSSION 1. Introduction It is a well accepted truism that as more and more vital elements of our national defense become increasingly sophisticated, they also become more sensitive to the influences imposed upon them by their surrounding operational environment. This fact is nowhere more clearly illustrated than in the arena of electro-optical system performances in turbid, cloudy atmospheres where cloud attenuation of the sensor/target path of sight can result in the complete and near instantaneous negation of system performance. It is therefore only prudent that intense and innovative research into the development of techniques for the assessment and prediction of appropriate cloud population characteristics and cloud field dynamics, particularly at small spatial and temporal scales, be actively pursued. This technical brief is directed toward that end. The WSI (Whole Sky Imagery) analysis effort is adaptable to a variety of studies designed to help assess GBL (Ground Based Laser) performance as achieved through multiple site operations as well as multiple lines of sight at one GBL location. The basic statistical summaries can be developed simultaneously with the WSI operation at each designated site. As the data base grows beyond its existing content, CFLOS and CFARC probability distributions are readily calculatable with respect to total sky cover, arc length, arc position and orientation, season, time of day, etc. The statistical summaries tend to follow a general format to ensure the widest possible application to analytical model development and validation. Of unique interest and importance is the high resolution of the WSI data base in both time and space. The data acquisition schedule has provided simultaneous imagery at multiple sites on a time scale as short as one minute at reduced grid resolutions throughout the day. Thus, it is possible to determine joint frequency distributions of initial and final CFLOS and CFARC conditions at a single station or between multiple sites over very short time intervals. Appropriate conditional probability summaries may be constructed so as to be directly applicable to both the evaluation and refinement of simulation models related to the definition of cloud impacts upon electro-optical system performances, and the near real- time support of operational case studies. - 2- The format flexibility inherent in the output product of the WSI data processing stream enables convenient interfacing with task specific software; a feature appropriate to the insertion of meteorological impacts as a relatively separable element into many more complex operational engagement scenarios. 2. Basic Concepts The Whole Sky Imager (WSI) data network appropriate to this task is a sub-set of the seven station array originally deployed. The subset consists of the three camera sites located in New Mexico, specifically the C-STATION and HELSTF sites at WSMR and the optional station at Kirtland AFB in Albuquerque. As is illustrated in Table 1, the emerging data base has been accumulating since 30 Mar 88, and thus currently represents about two years worth of data for each location. If the sites remain operational an average of about 80% of the time which is their typical historical rate, then an additional 123 data days will be accumulated by 30 Sep 90, yielding an initial working base of approximately 730 data days per site. It is a reasonable estimate that the 1460 data days describing the WSMR cloud coverage will reside on approximately 260 data tapes. If data collection and archival at these sites continues into FY 91 as anticipated, the data base will grow at a rate of approximately one tape per site per week. As noted in the earlier references, each seven day data tape contains approximately 1200 megabytes of image data plus around 200 megabytes of housekeeping space. The data quantities thus are substantial, and will require carefully conceived and efficiently executed processing procedures to insure downstream utilization. The general concept guiding the post-acquisition processing of the WSI data base is illustrated in Fig. 1, "WSI BASIC IMAGE PROCESSING FLOW CHART". This conceptual game plan is implemented through a series of proprietary software routines designed to impose strict quality and calibration controls upon the raw data, while simultaneously reducing the number of images required for insertion into the optimized cloud/no cloud decision processes. Carried to its nominal conclusion, this processing plan reduces the basic imagery (that created by each camera in the system) to about one fourth of its original volume at the composite ratio stage and then, if desired, by an additional - 3 - BASIC IMAGERY CORRECTED IMAGERY COMPOSITE RATIO DELIVERABLE DATABASE CALIBRATION FUNCTIONS : - RADIANCE CONVERSIONS CALIB. BLUE IMAGE 512 X 512 x 8 CLOUD/NO-CLOUD DECISION ALGORITHMS BLUE IMAGE RADIOMETRIC LINEARITY I BLUE/RED RATIOS RADIOMETRIC SENSITIVITY UP-SUN RED IMAGE 512 X 512 x 8 CALIB. RED IMAGE DERIVED PRODUCTS DOWN-SUN IMAGE RATIO COMPUTATIONS | DOWN-SUN OPTICAL DISTORTIONS NEAR HORIZON SPATIAL I DISTRIBUTIONS - COMPOSITE BLUE/RED RATIO IMAGE OPTIMUM CLOUD/NO-CLOUD IMAGE 4 FIELD OF VIEW DEFINITION TWILIGHT TEMPORAL | DISTRIBUTIONS - BLUE +N.D. IMAGE 512 X 512 x 8 CALIB. (BLUE) IMAGE DAWN SENSOR CHIP UNIFORMITY STATISTICAL PARAMETERS PIXEL SELECTIONS FOR ONS 7 AS REQ'D I OPTIMUMIZED COMPOSITE (BLUE)(RED) RATIOS REGISTRATION ADJUSTMENTS RED + N.D. IMAGE 512 X 512 x 8 CALIB. (RED) IMAGE | FLUX CONTROL | THRESHOLDS WSI BASIC IMAGE PROCESSING FLOW CHART Figure 1 increment depending upon the output format specified by the user of the derived products. Currently (Jun 1990) the optimum cloud/no cloud images are being created at the same eight bit resolution as are the composite ratio images. Whereas further data compaction might prove attractive for later data transfers, the additional software development deemed necessary has not seemed justified at this time. Table 1 WSI DATA TAPE STATUS FOR WSMR AREA (AS OF 30 APRIL 1990) C-STA 30 MAR 88 HELSTF 29 MAR 88 KIRTLAND 17 MAY 88 TOTAL DAYS IN FIELD 762 1763 714 TOTAL DAYS OF DATA 608 595 604 FULL DATA DAYS (12 HR) 507 445 498 PARTIAL DATA DAYS 101 150 106 E DAYS DOWN - NO DATA 147 154 BACKLOG - NOT PROCESSED -0- LATEST DATE RECEIVED 23 Apr | 16 Apr 7 Apr DAYS NOT REC'D AS OF 30 April 1990 | 7 14 23 A = B + E + F + H Supplementing the image oriented data base, are copies of the Form 10a and 10b hourly station observations for these same sites. These hard copy documents are used at MPL primarily for trouble shooting and diagnostic purposes, but they are typical of the weather observation data that forms the long term historical record available through ETAC, etc. .5. 3. Implementation In order to fully define the task of preparing an existing image oriented data base for an orderly transfer to a new applications environment, it is appropriate to breakdown the existing archival methodology into its component tasks and evaluate the impact of each upon the transferral process. This stepwise evaluation will minimize the probability of overlooking a critical procedural sequence and provide a reasonable set of smaller sub-tasks upon which the necessary time/cost estimates can subsequently be based. a. Data Acquisition The WSI cloud data base, as illustrated in Table 1, presently resides in the custody of the Marine Physical Laboratory (MPL), Scripps Institution of Oceanography, San Diego, CA, 92093. The preparation of the 260 raw data tapes plus their appropriate output products with the procedural documentation deemed necessary for subsequent usages is outlined in this note. b. Supporting Instrumentation The WSI cloud data base is stored in digital format on 8mm video tape cassettes (Sony P6-120MP, NTSC). These small cassettes, when used with the EXABYTE 8200 streaming tape sub-system, can hold up to 2.2 Gigabytes of digital information each. All WSI data manipulations conducted at MPL are built around this small compact tape system under the control of an IBM/AT class microcomputer. All software is MS-DOS compatible, and is generally written in either FORTRAN or C programming language. The basic hardware configuration required for WSI image processing is essentially the same as that required for the initial acquisition. That is, a system able to read the 8mm data tapes, an image processing system able to handle the 512 X 480 pixel image format, and a host computer suitable for overall control. At MPL, the data acquisition system shown in Fig. 2 is also appropriate for all subsequent image processing tasks. For bulk processing, however, the basic computer unit is supplemented with an .6. Image Acquisition & Analysis System Hardware Block Diagram E/O System 5 Heath con SONY PVM 1271Q MONITOR . GE 2710 SOLID STATE VIDEO CAMERA TMI COMPUTER (IBMAT CLONE) . AUTOMATIC EQUATORIAL SOLAR OCCULTOR ASSY. VIDEO IMAGE PROCESSING SUB-SYSTEM (II FG 100) ARCHIVAL 110 SUB-SYSTEM (SEAGATE 65 M byte H.D.) 340- REMOTE CONTROLLED IRIS ASSY. EXABYTE EXB 8200 2.2 G byte 8mm CARTRIDGE TAPE SYSTEM ANALOG EO CAMERA ACCESSORY CONTROL PANEL ACCESSORY CONTROL PANEL REMOTE CONTROLLED OPTICAL FILTER ASSY. . STOWED KEYBOARD EXTERIOR SENSOR INSTALLATION INTERIOR CONTROLLER INSTALLATION IT Figure 2 Eighteen Eight Labs PL 1250 floating point array processor to accelerate the system throughput, and multiple EXABYTE drives are attached to each control computer. c. Application to ASAT Scenarios The WSI image oriented data base, by virtue of its relatively high spatial and temporal resolution, is particularly appropriate for applications associated with the acquisition and tracking of earth orbiting devices. The statistical characterization of local cloud effects upon earth to space optical communication links is, of course, the reason the existing multi-station data collection network was originally established and continues to operate. The generalized output product from this network, the temporally varying pixel by pixel definition of cloud thickness throughout the observable sky dome, is directly applicable to all cloud free line of sight (CFLOS) and cloud free arc (CFARC) determinations either as real time operational support products, or as statistically predictive model inputs. Ground based laser operations which are particularly impacted by cloud field dynamics can employ the appropriate variations of CFLOS/CFARC persistence and recurrence statistics directly into mission planning scenarios with increasingly short response times. With regard to determining the impact of cloud effects upon definition of a specific ASAT systems engagement space, it is clear that there are several contributing elements. The characteristics of the satellites orbit, the characteristics of the intervening atmosphere specifically including the nature of the prevailing cloud field, and the nature of the ground based laser itself must all be considered. Whereas it is appropriate to be fully cognizant of the methodology intended for the specification of these various characteristics, we feel that the optimum analytic scheme would employ as complete a decoupling between the major sub-elements as is practical. This is particularly true with regard to the atmospherics. To the degree that this decoupling of the cloud field statistics and related characterizations from the other performance characterizations within the general ASAT system performance model can be achieved, the more readily a truly definitive set of CFARC statistics can be generated. In the long term view, we feel that atmospheric decoupling is an extremely important issue and should be addressed early and creatively in defining the analytic approach. For the application at hand, i.e., meteorological support products for application within the ASAT Scenario discussed in Technical Note 219 and Technical . -8- Memorandum AV90-074t attached, we will assume meteorological decoupling, and the production of statistical products from the WSI data base for White Sands Missile Range. In accordance with our current understanding of the task, our initial recommendation would be to proceed in the following sequence. A. Process the existing WSMR data base to provide full resolution "10 minute" ratio imagery. B. Process full resolution imagery to cld/nocld format using new B-based cloud decision algorithm. C. Collate cld/no cld data base for hierarchal sorting by site, total cloud cover, user defined interval characterizations, temporal intervals, and model specific sub-sets. D. Evaluate all sub-sets of cloud imagery to specify adequacy for insuring statistical confidence in output products. E. Define sub-sets of orbital tracks appropriate for specific evaluation with selected cloud sub-set.. F. Process cloud/orbit sub-sets to provide CFARC conditional probabilities as outlined in TN 219. G. Coordinate insertion of derived CFARC probabilities into sponsor's ASAT System Performance model. Estimating the cost of implementing the general sequence indicated above can be done at several levels of detail ranging from an absolute "bare bones" approach defined by a minimal data base applied to a single orbit using existing software/hardware resources only, to an extensive array of multi-site/multi-orbit scenarios exercised on a stand alone dedicated workstation. The following budgetary estimates are intended to identify an implementation adequate for entry level operations yielding reliable data for a limited number or orbital selections. Costs associated with generating the working data base of full resolution imagery are outlined in section d, as are the costs associated with re-processing, if required, using an upgraded decision algorithm. Modification of program TAPRAT output files to enable more sophisticated hierarchal sorting techniques will require only a moderate reprogramming effort. First estimates for this task indicate the requirement for one analyst-mo to code, debug and implement modified header structures amenable to the desired sorting flexibilities. The completion of the currently emerging B-based decision algorithm should be accelerated to enable its use on the full resolution imagery recommended for use in the ASAT task. This is essentially a substantially upgraded version of program CLDDEC, which will compensate for the directional effects of aerosol scattering within the lower atmosphere. It is not unreasonable to anticipate an accelerated effort of three to four analyst-mos to complete this task in time to enable an orderly execution of step B in the general sequence. Evaluation of the overall data base in its various sub-setted configurations to insure adequacy for statistical confidence will be an ongoing operation absorbed for the most part by the interpretation and analysis (I/A) functions included under section 3d. Some validation and oversight by our senior meteorologist will be required, as will some modest software development. An additional analyst-mo should be adequate for this evaluation. The prototype software used to create the sample statistics illustrated in TN 219 was intentionally simplistic in nature as a matter of expediency. To fairly address the task at hand, this software package will require expansion to enable the processing of full resolution imagery, and to retain the resultant pixel by pixel library of cloud specifications for statistical accumulation and display. Redesign of the computational scheme to permit the processing of optimum sized groups of cloud/orbit combinations may involve several iterations, but as the number of sets to be processed increases beyond just a few, the additional redesign effort will more than pay for itself. A reasonable estimate for this package upgrade is two or three analyst-mos. Procedural coordination with GBL/TRW software developments is an on-going necessity, but probably a relatively low level effort. One analyst-mo per year for most www AD . 10 - circumstances is probably adequate. However, expanding the scope of the software or analytic interfacing task could require a re-definition of this task element. In light of the estimates outlined above, it seems reasonable that a first pass through the existing WSMR data base, with adequately upgraded code, and sufficient technical oversight to insure statistical confidence would require on the order of 8 to 10 analyst-months of effort. At existing University rates this level of salary and salary related expenses is summarized in Table 7, Task 2. d. Data Processing, Interpretation and Analysis Data Processing With the basic data tapes in hand, the processing hardware and software on-line, and the data base management system implemented, one can address the main thrust of this proposed data process; the stepwise manipulation of all component elements from raw tape to computed arc statistics, and the application of these end product statistics to defining cloud impacts upon engagement space boundaries. In the following paragraphs this stepwise process will be described on the basis of a single data tapes sequential manipulation, and from this unit tape estimate, the overall program time/cost estimates will be produced. In assessing the following estimates of the processing times associated with each procedural sequence, one should be aware that throughput times are often heavily influenced by different machine/software interfaces even when determined on systems containing nominally "identical" hardware and code versions. With this caveat in mind, the estimates included herein and indicated on Fig. 3, were derived from actual full length processing episodes running in the hardware/software configuration listed in Table 2. Processing the complete stepwise procedure required to convert a field tape containing seven days of raw unedited cloud imagery into a fully processed tape containing the same seven days of data in the form of color coded cloud-no cloud . 11. FIG. 3 Raw Imagery to Decision Imagery Conversions one site - one week RAW TAPE TAPEQC TAPEQC OUTPUT FILES T.O.F. BACK-UPS 7 RAW TAPE EXCOPY RAW DATA BACK-UP (ARCHIVE) e deling - Comerciante T.O.F. (BACK-UPS CHECLAN TAPRAT INPUT FILES prunenwes (INTERACTIVE) RAW TAPE TAPRAT INPUT FILES TAPRAT RATIO TAPE 1/10 LOG FILE DIAG FILE A 5.* EXCOPY RATIO TAPE 1/10 RATIO BACK-UP 1/10 (OPTIONAL ARCHIVE) ca 2 LOG FILE DATA BASE SUMMARY D.B. SUMMARY RATIOS DIAG. FILE CLDDEC RATIO TAPE 1/10 CLDDEC TAPE 1/10 CLDDEC OUTPUT FILES EXCOPY (ARCHIVE) CLDDEC TAPE 1/10 CLDDEC BACK-UP 1/10 D.B. 9. CLDDEC OUTPUT FILES DATA BASE SUMMARY .. .. SUMMARY DECISION *Separate runs required for each one minute or ten minute image set. - 12 - imagery is illustrated in Fig. 3, and the run-time summary for both machine time and supporting analyst time is shown in Table 3. Table 2 WSI Standard Data Processing Configurations HARDWARE SOFTWARE TMI Model 2001A Computer Proprietary Software Programs 1. 80386 CPU 2. 80387 Co Processor 3. PL 1250 Array Processor 4. ASC-88 SCSI Interface 5. EXABYTE 8200, 2 ea w/ 4 $25 PROM 6. Seagate 60 MB Hard Disc 1. TAPEQC, Version xxx 2. EXCOPY, Version xxx 3. TAPRAT, Version xxx 4. CLDDEC, Version xxx 5. DBSUM, Version (In development) Table 3 Sequence No. Raw to Decision Conversions Run Time Summary Machine Hours Analyst Hours File Copy | Total Housekeeping &Oversight w (From Fig. 3)| Primary www.... .wwwwwwwwwwww 5.5 w ww .5+.5 25 w..www .2 0.5 2.5 2.5 12+12 0.5+0.5 2+2 .2+.2 4.4 0.5+0.5 1 1.2 0.5 6+6 .5+.5 13 0.5+0.5 1+1 .2+.2 2.4 0.5+0.5 .2 0.5 Totals 1+10 58 64.2 1 or 10 37 4.8 41.8 Separate runs required for each one minute or ten minute image set. Times indicated are based upon "clean" runs without appreciable software or hardware glitches. www 1.2 6.2 10 wp - 13- In estimating the overall processing task, as outlined in Fig. 3 & Table 3, it is important to note that the totals shown in Table 3 are really for best effort data "processing" only. They do not include the very substantial effort which is heavily related to "housekeeping and oversight", but which is more specifically oriented toward on-line data "interpretation and analysis". Nor do they include the more stylized but highly important implementation and maintenance of the Data Base Management functions. However, having noted these caveats, one may proceed with estimating the processing task associated with an assumed ASAT support scenario. It is reasonable to assume the quantities established in Sec. 2, i.e., an initial data base as of 1 Oct 90 which contains WSMR imagery from two sites residing on approximately 260 data tapes. As reflected in Fig. 4, these tapes will have completed their initial QC runs, and a substantial portion of them will have been run through the entire processing sequence illustrated in Fig. 3. Whereas these preliminary data will be useful in establishing initial techniques for sorting by cloud amount and related distributions, they represent the one-minute data arrays and do not contain the full resolution 10 minutes imagery deemed most suitable for application to ASAT engagement scenarios. Neither have they been processed using the solar scattering angle (B) based algorithm which is currently under development. For producing thin cirrus detection algorithms suitable for use in analyses related to the ASAT task, the improvements inherent in the B based are essential. The basis for initial estimates is now established. A set of 260 data tapes must be processed for full resolution imagery using the B-based algorithm. These CLDDEC output data must then be processed for statistical application as outlined in the low resolution, more simplistic example contained in Technical Note No. 219. The estimated processing task of pushing 260 tapes through steps 3 through 9 of Fig. 3 one time to generate full resolution, i.e., 10 minute data, is 7098 machine-hours and 1430 man-hours. These figures are more readily addressed as 8.2 man-mo, and - 14 - MAR APR MAY JUN JUL AUG JAN SEP FEB OCT NOV DEC. JAN. FEB MAR FEB. MAR. APR MAY JUN JUL AUG SEP OCT. NOV. DEC. e CSTA CSTA CSTA CSTA HELSTF HELSTF HELST HELSTF KIRTLAND KIRTLAND KIRTLAND - 15 WSI DATA STATUS & PROCESSING SUMMARY * GOOD DATA - - SPECIAL HANDLING REQUIRED J= SYSTEM DOWN 1988 1989 1990 = INSTALLATIONMAINTENANCE = RATIO - CLOUD DECISION / SINGLE THRESHOLD Figure 4 for the machines, assuming a 16 hour day for 80 hours per week with a four machine processing system, 22.2 system-wks or 5.2 system-mo. Data Interpretation & Analysis · The Table 3 comment related to "clean" runs is a comment ignored at extreme peril. One must keep in mind that all of WSI field systems were monitored by on-site host personnel with widely varying degrees of technical background and program related interests. Not all received the imposition of the WSI system into their daily routines with equal exuberance. The vagaries of local weather extremes, power fluctuations, hardware faults and conflicting on-site priorities all contributed to sometimes substantial variations in the normally full automatic mode that the WSI attempted to maintain. With this "real-world" context in mind, it is easy to visualize the nature of the on-line data interpretation and analysis that must accompany the more formalized data "processing" outlined in Fig. 3. The primary tools employed by the analyst to diagnose and sanitize those raw data tapes which fail either their initial TAPEQC runs, or their subsequent TAPRAT runs, are the regularly produced TAPEQC output files, the Local Apparent Noon (LAN) image files, and the raw data tapes themselves, all in concert with the analysts highly honed understanding of the WSI hardware and software operating regimes. These diagnostic skills are not readily transferable and they are absolutely essential to the orderly and efficient data flow shown in Fig. 3. The incidence with which one must expect to contend with non-ideal data is illustrated to a first approximation by the non-green entries in Fig. 4. This general purpose status chart may be used to select data intervals containing minimum service interruptions, stations containing simultaneous data, etc. However, it does not contain information at a fine enough level of detail to preclude the need for some on-line interpretive assessments even when processing "green line" data which in contain system time line variances and recording format anomalies. Experience to date in processing "green line" data through TAPRAT and CLDDEC has indicated a need for an analysts on-line interpretation time at least equivalent to that needed for "Housekeeping and Oversight" tasks. It is clear from Fig. 4 that as the backlog processing moyés into the 1988 data inventory, the required interpretation and analysis - 16- effort will surely increase. It is not unrealistic to assume that the increase would be in the order of 50 to 80 percent beyond the estimates in Table 3. Some estimate of the additional machine time required to support the "interpretation and analysis" (I/A) function plus the attendant requirement for data re-runs should be provided as part of the overall task definition. As a first approximation, one can use the yellow, i.e. special handling fraction of the Fig. 4 data summary as the beginning estimate, and add the green line increment in accordance with current experience. Following this approach, one finds about 7% of the data designated as special handling. During the first quarter of calendar 1990, an estimated 10% of the accumulated data runs were repeated due to specific I/A decisions and another 10% were re-run due to miscellaneous hardware peculiarities, primarily EXABYTE incompatibilities. Thus one might assume a worst case requirement for up to 27% of miscellaneous re-runs, however as "production line" processing settles down, a 10 to 15% re-run estimate seems more reasonable. The overall task estimates reviewed in the preceding paragraphs are summarized in Table 4. Table 4 Data Processing, Interpretation, and Analysis Time Estimates Task System - mos Machine 5.2 | (@.15) 0.8 Man - mos Analyst 8.2 Data Processing Interp. & Anal. 1.2 TOTALS 6.0 9.4 e. Alternative Products All preceding comments have been directed toward the "standard" data processing mode as illustrated in Fig. 1. The output product for this standard sequence is a database containing derived cloud/no cloud images based upon a single threshold decision algorithm (APR '90). There are other output products which have been - 17 - discussed at different times during the data acquisition interval that should be commented upon as viable options. The first alternative involves the use of improved decision algorithms to up-grade the cloud/no cloud data base. In general this up-grade can be exercised at anytime and as often as desired, once the composite ratio image library has been established. Re- running TAPRAT output tapes through a revision of CLDDEC is straight forward and relatively inexpensive. An estimated cost for implementing this alternative is illustrated in Table 5. Upgrade CLDDEC Executing an up-grade to the cloud/no cloud imagery via an improved cloud decision algorithm implies re-running steps 7, 8 & 9 as shown in Fig. 3. The associated machine and analyst hours necessary for this task are listed in Table 3, and represent 16.6 machine-hours and 2.5 analyst hours per ratio tape. Assuming ratio tapes remain at 1:1 with raw data tapes, then about 260 tapes would require this partial re-processing. Some tape compaction is possible through a judicious application of EXCOPY, but it is not currently an automatic option. Thus, the re-run of 260 tapes through CLDDEC, etc., represents a task of 2.2 analyst-mo and 1.7 system-mo. Table 5 Estimated Cost to Accomplish CLDDEC Reruns 15,730 2,800 Salaries, related costs, w/o Univ load (2.2 m-mos) Computer operating costs: 10 machines, maintenance only $2000/yr/machine @ 0.14 sys-yr Working tapes, backup & scratch @ 4 per ratio tape, 1040 @ $6 Supplies & Materials @ 5% (Sal & EB) Equipment up-grades Travel SUBTOTAL UCOH @ .27 TOTAL 6,240 786 ܚ ܀ ܝ ܀ 25,556 6,900 32,456 AN - 18 - The second alternative involves the modification of the Fig. 3 sequence to generate calibrated radiance imagery, either in addition to or in lieu of the cloud/no cloud imagery. This sequence modification is not trivial since the calibrated radiance imagery is not normally held intact in image memory, but for maximum computational throughput is used piecemeal and not retained once the ratio computations are complete. An estimated cost for implementing this alternative is illustrated in Fig. 5 and Table 6. Create Calibrated Radiance Library Modifying the processing sequence shown in Fig. 3 to produce output tapes of d radiance imagery has not yet been implemented, but does not represent a major programming effort provided one makes a clean interrupt at the point of insertion within the software program TAPRAT. There will be however, changes to the logical flow within the program which may result in substantial increases in processing times. In any case, the creation of a calibrated radiance library would procedurally involve the re-running of modified steps 4, 5, & 6 shown in Fig. 3. These modifications would be of the form shown in Fig. 5, with programs ABSRAD and SELECT representing the necessary new code. Since the absolute radiance values which are desired as output reside temporarily within the existing sequence, and their accumulation should be reasonably simple, the machine time necessary to execute the ABSRAD task should represent a relatively short interval. A reasonable first approximation for the ABSRAD execution time is its being equivalent to or less than the CLDDEC runs. To the degree with which this approximation is true, then the incremental cost of executing the sequence shown in Fig. 5 can be estimated. Assuming a clean insertion, it is reasonable to assume approximately 3 man-mo to code, test and debug the ABSRAD & SELECT software. Housekeeping and Oversight (H/O) analyst hours required for processing should not exceed the TAPRAT rate, nor should the I/A time. Thus an estimate of around 4 analyst-hours per radiance tape, about half the TAPRAT rate, seems reasonable. Rerunning the nominal 260 raw data tapes through the SELECT/ABSRAD sequence represents an incremental increase of approximately 1.1 system-mo. and 9 . 19. Fig. 5 Raw Imagery to Calibrated Radiance Data ABS RAD ABS RADIANCE TAPE 1/10 SELECT (4.*)a RAW TAPE TAPRAT INPUT FILES TAPRAT RATIO TAPE 1/10 LOG FILE DIAG FILE (5.*)a (ABS RAD EXCOPY TAPE 1/10 RADIANCE BACK-UP 1/10 (OPTIONAL ARCHIVE) LOG FILE (6.)a DATA BASE SUMMARY 11 D.B. SUMMARY RADIANCE DIAG. FILE * Separate runs required for each one minute or ten minute image set. - 20 - analyst-mo. An estimated incremental cost for implementing this alternative processing sequence is illustrated in Table 6. Table 6 Estimated Incremental Cost to Create Absolute Radiance Library 64,350 7,200 4,680 1. Salary related costs, w/o Univ. load (9m-mos) 2. Computer operating costs: 10 machines, maintenance only $2000/yr/machine @ 0.9 sys-yr Working tapes, backup & scratch @ 3 per raw tape, 780 @ $6 3. Supplies & Materials @ 5% (Item 1) Equipment upgrade Travel SUBTOTAL 6. UCOH @ 0.27 TOTAL 3,217 79,447 21,451 100,898 f. Hardware Options The various cost estimates outlined in the preceding sections of this note assume that the processing and analysis tasks will be accomplished using, at least initially, the suite of desk top computers currently on-hand and in use at MPL. There are significant programmatic advantages in not relying on time-share prioritizations, but in having some minimum core of dedicated hardware available to ones task. In a similar manner, the basic data base that is accumulating at the WSMR is acquired by camera systems which were built under the auspices of the Office of Naval Research. Whereas we do not anticipate the withdrawal of these systems from service, it is only prudent to consider one's options should a withdrawal occur. In order to apprise you of the relative costs involved in backing up critical hardware components, we have included Table A, E/O Camera System 5A, As-built parts list, reflecting our costs as of 16 Jun 89. Two options based upon Table A are summarized for your review in the following tables. . 21 TABLE A E/O CAMERA, SYSTEM 5A As-Built Parts List 16 Jun 89 Item # Model # Manufacturer/Vendor Function Approx cost $ AWN 2001A B386S/16-2MEG 80387-16 KBIAT BVMC11 Texas Microsystems Inc Texas Microsystems Inc Texas Microsystems Inc Texas Microsystems Inc Texas Microsystems Inc Computer Base Unit CPU Math Co-Processor Keyboard Color Video Card 1500 2121 1000 210 100 ST-277R HCRA2 Seagate Western Digital 65MB Hard disc Disc Controller 490 140 EXB-8200 ASC-88 Exabyte Advanced Storage Concepts 2.2 GB Tape Streamer SCSI Host 3900 200 DIO-96 ICS Computer Products TTL I/O Ports 660 VS-100-1024 -3-U-AT Imaging Technology Inc Video FG & IP 4700 Zoom HC2400 Zoom Telephonics Smart Modem 250 Sony PVM1342Q Voice & Video Color Monitor 900 GE 2710 DRF Associates Video Camera 3450 GC-1000 Heath/Zenith NBS Radio Clock 300 EO3-1200-01 MPL/UCSD Opt. Filter Changer 8100 EO3-3000-01 MPL/UCSD Acc. Control Panel 5000 EO3-3500-01 MPL/UCSD WXP Camera Housing 4000 EO3-3600-01 MPL/UCSD WXP Lens Housing 2500 6000 MPL/UCSD MPL/UCSD EO3-2101-01 EO3-3700-01 & 3800-01 Cable Set Solar Attenuator Stand & Shroud 2500 MPL/UCSD CAM/ACP, etc. 500 Enclosure MPL/UCSD Environmental Housing 2500 BUD 300 91F1697 SRI TMI 19" Rack Rack Slides 70 AFC-3 Aquanetics Systems Camera Chiller 600 CBEC 8G6 Wallow Recirculation Heater 275 4MD-SC Little Giant Coolant Pump 120 1200 VA SPS Clear Signal Stand by Power Source 1400 NOTE: Items 1 thru 15 and 24 thru 29 are retail unit costs only, i.e., no associated labor costs. Items 16 thru 22 are approximate materials and labor costs for in-house MPL fabrication. Final assembly and check-out labor costs are currently estimated at $15000 per unit and are reflected in our normal salaries budget. - 22 - Hardware Option 1: Data Processing Computer System Items 1 thru 13 in Table A: PL1250 Array Processor $16,171 2.700 $18,871 Hardware Option 2: Complete Field Deployable Camera System Items 1 thru 28 in Table A: $53,786 The selection and use of the components discussed herein does not imply endorsement or recommendation of the tested products by the Marine Physical Laboratory or its sponsors to the exclusion of other products which may be suitable. B. BUDGET REVIEW & SUMMARY Whereas labor costs are relatively straight forward to estimate, machine costs are somewhat variable depending upon the amortization schedule imposed by ones accounting system. One approach is to amortize the cost of the machine over the lifetime of the warranty. In this instance a TMI computer as described in Sec. 3b costs about $24,000 with all peripherals. If one runs it 10hrs/day for 250 days/yr it accumulates about 2500 machine-hrs/yr, whence $10/hr for machine time will pay for it at the end of its one year warranty, or $5/hr will pay for it at the end of a two year extended warranty agreement if available. At an estimated $5/hr for machine time, the task outlined in Table 4 represents approximately $35,490 in computer operating costs. Out of warranty returns for either the EXABYTE or FG-100 frame grabber sub- system currently cost approximately $1,000 per unit, per return. To date, there have been 18 EXABYTE returns for repair from a total of 23 available systems, and slightly fewer returns for the frame grabbers. Thus, once the hardware systems are acquired, it might be more cost effective to amortize only the extra cost of estimated repairs over the life of the program, and then throw the machines away. It really depends upon the choice of long term or short term policy and projected system upgrade schedules. - 23 - Estimates of analysts costs are premised on the basis of an average salary of $60,000 per annum with a University load factor of 1.7 yielding a total cost of $102,000 per analyst yr. Using the estimate from Table 4, one arrives at a total estimated labor related cost of $79,900. Miscellaneous costs related to equipment maintenance supplies and materials, office supplies, travel and new equipment are in general straightforward to estimate, either as fixed percentages or as mutually agreed upon amounts as requested by the sponsor. The cost estimates discussed in the preceding paragraphs are summarized in Tables 7, 8 and 9. Table 7 TASK 1: Process data amounts estimated in Sec. 3a in accordance with schedule illustrated in Fig. 3 & Table 3, and with interpretation and analysis as estimated in Table 4. Estimated Cost to Accomplish Task 1 1. 67,210 35,490 12,000 Salaries, Employee Benefits, w/o Univ load (9.4 m-mo) Computer operating costs: 4 machine Processing System @ $5/hr 6 machine analyst support, maintenance only assume $2000/yr/machine @ 1 yrs Working tapes, back-up + scratch @ 8 per datatape, 2080 @ $6 ea Supplies & Materials @ 5% (Sal. & EB) Equipment upgrade Travel, 20 md @ $100 day 2,000 4 Airfare @ 1000 4,000 10 d auto @ $60 day 600 6,600 SUBTOTAL UCOH @.27 12,480 3,360 ܚ ܕ ܝ 6,600 137,140 37,028 6. TOTAL 174,168 - 24 - Table 8 TASK 2: Provide software upgrades, technical analysis and oversight in accordance with the general sequence for application to ASAT scenarios as defined in Sec. 3c. Estimated Cost to Accomplish Task 2 71,500 3,575 Salaries, Employee Benefits, w/o Univ load (10 m-mo) omputer operating costs: included under Task 1 Supplies & Materials @ 5% (Sal. & EB) Equipment Travel SUBTOTAL UCOH @ .27 6. 75,075 20,270 95,345 TOTAL OPTION 1: Same as Task 1, except amortization of computer maintenance only. Estimated Cost to Accomplish Option 1, Task 1 67,210 Salaries, related costs, w/o Univ load (9.4 m-yr) Computer operating costs: 10 machines, maintenance only @ $2K/yr/machine for 1 yr Working tapes Supplies & Materials @ 5% (Item 1) Equipment up-grades Travel SUBTOTAL UCOH @ .27 20,000 12,480 3,360 6,600 6. 109,650 29,605 | 139,255 TOTAL . 25. Summary Several tasks related to the processing and evaluation of the Whole Sky Imager (WSI) data base, and interfacing an analysis of its cloud/no cloud statistical characterization to an appropriate set of ASAT engagement scenarios, or for modifying its output to provide alternative products, have been reviewed. Each of the tasks has had its incremental cost estimated assuming the as-built data base status on or about 1 Oct 90. Table 9 lists the estimated costs of the several procedural options. Table 9 Summary of Estimated Costs Task Ident Total A. Task 1, Fig. 3, Tables 3 & 4 Task 1, Option 1 B. Task 2, Sec 3c C. Up-graded Algorithm Re-runs* D. Absolute Radiance Re-runs* E. Hardware Options 1. Dual Data Process Sys. 2. Field Camera Sys. Salary (Item 1) 67,210 67,210 71,500 15,730 64,350 Estimated Costs ($) Machine Support (Item 2) (Items 3-6) 59,970 46,988 32,480 39,565 (See Task 1) 23,845 9,040 7,686 11,880 24,668 174,168 139,255 95,345 32,456 100,898 37,741 53,786 37,741 53,786 *Incremental costs assuming Task 1 or Task 1, Option 1 is completed. - 26.