UNIVERSITY OF CALIFORNIA, SAN DIEGO 3 1822 04429 7489 OPTICAL SYSTEMS GROUP TECHNICAL NOTE NO, 222 September 1990 Offsite (Annex-Jo finals) QC 974.5 . T43 no. 222 PRELIMINARY CLOUD FREE ARC STATISTICS FROM WSI IMAGERY T. L. Koehler UNIVERSITY OF CALIFORNIA SAN DIEGO 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 and the Geophysics Laboratory ITV 31 UNO. Contract Monitor, B. Kunkel Atmospheric Sciences Division TRE RNIA KOOL 1868 est Prepared for The Geophysics Laboratory, Air Force Systems Command United States Air Force, Hanscom AFB, Massachusetts 01731 under contract No. F19628-88-K-0005 SCRIPPS INSTITUTION OF OCEANOGRAPHY MARINE PHYSICAL LAB San Diego, CA 92152-6400 UC SAN DIEGO LIBRARY UNIVERSITY OF CALIFORNIA. SAN DIEGO w Pued .www.www. www. wwww........... wo IS wiper www ... . dewasa ww exx-WPS www. weweselne .. w.w. e wwwwwww Medien ww. w... werd-Wwwwwwwwwwwww. .. w ..- Swe www 3 1822 04429 7489 . Preliminary Cloud Free Arc Statistics from WSI Imagery Table of Contents 1.0. Introduction 2.0 White Sands Cloud Cover Climatology 3.0 CFARC Data Reduction NWAY WHYW w wwwwwww 4.0 wwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwww. ..www Results 4.1 General CFARC Distributions 4.2 Cumulative CFARC Length Distributions 4.3 Conditional Probability Distributions 4.4 Concluding Remarks . 5.0 References List of Figures CS Figures Tables 1.0 INTRODUCTION 1 The preliminary processing of the Whole Sky Imager (WSI) data base for four sites for the period from February 1989 through March 1990 has recently been completed. With these data, a wide range of analytic studies are possible. Of particular interest to our group are studies that illustrate the applicability of the WSI data base and identify data peculiarities that might be corrected in the final data processing. Several limitations of the data have already been noted in previous reports (eg Shields, et al., 1990). For example, the single threshold cloud decision algorithm used in generating the preliminary data overestimates the presence of clouds near the horizons at low solar elevation angles (around sunrise and sunset), and near the aureole, while it underestimates the presence of thin cloud in the downsun direction. Generally, preliminary comparisons of WSI total sky cover to human observations agree reasonably well. The character of the cloud cover over the White Sands Missile Range (WSMR) in New Mexico is of vital interest to the SDIO program. This note describes a brief study of the cloud free arc (CFARC) statistics derived from the WSI data collected at the C-Station WSMR site. 2.0 WHITE SANDS CLOUD COVER CLIMATOLOGY Considering our desire to limit the resources allocated to this task, and the preliminary nature of the cloud/no cloud data available, we decided to focus on two monthly samples that illustrate the extremes in the cloud cover distributions at WSMR. Since the degree of cloud cover along an arc in the sky is strongly dependent on the total cloud cover, the selection of which monthly periods to examine was based on the climatological sky cover data from Stallion Site at WSMR illustrated in Fig. 1. The strong summer diurnal cycle with clearer mornings (blue) and cloudier afternoons (red) is clearly evident. The extreme periods (black circles) occur in consecutive months, with June AM being the clearest on average, and June PM being cloudiest on average. The green (+) symbols show the corresponding C-Station WSI average cloud covers from 1989. The WSI one year average matches the Stallion 13-year site climatology in the July PM cloudy period. In contrast, the June AM averages differ significantly. Fig. 2 illustrates the cloud cover distributions from WSI and climatology for the periods of interest. The completely clear and completely cloudy categories are in close agreement for the June AM results. However, the sky cover amounts in the 3-9 tenths categories are smaller in the WSI results compared to climatology, appearing instead in the 1 tenth category, leading to the lower WSI average cloud cover. Similarly, the WSI results from July PM are smaller in the same 3-9 tenths categories than those from climatology, but both the 1 tenth and 10 tenths cases have increased, resulting in no difference in average total cloud cover. The differences noted can arise from several sources, including natural year-to-year variability, different sites (C-Station is approximately 100 miles south of Stallion Site) and surrounding terrain, and known biases in the human observations vs. WSI (see, for example, McGuffie and Henderson-Sellers, 1989). . 3.0 CFARC DATA REDUCTION The basic data employed in the study is a subset of the preliminary one-minute cloud/no cloud data release to TASC. Software was written to extract a 337 pixel cloud/no cloud sample from image column 255 starting at row 72 at 5 minute intervals. This column runs from south to north through zenith, and the sample is bounded by the 60° zenith circle (Fig. 3). For each subset extracted, the maximum arc free of either opaque (white) or thin (yellow) cloud pixels was determined, and the corresponding starting pixel and arc length, values saved. The maximum CFARC in Fig. 3 is highlighted in dark blue. All available data from the C-Station site, 27 days for June 1989 and 28 days from July 1989, were, processed in this manner. The results that follow are based only on the morning data from June and the afternoon data from July. Any period with an arc containing occulter mask pixels was ignored. 4.0 RESULTS 4.1 General CFARC Distributions 1 The first step in the data analysis was to classify the maximum CFARC starting and ending pixel positions and the resulting arc lengths into 24 categories, each category representing an arc of 5° zenith angle. The June AM results are shown in Fig. 4. All three distributions do not include the completely cloudy arc (4.6%) or completely clear arc (75.4%) cases. As expected, the start pixel distribution has a maximum at the south edge, and the end pixel distribution has a maximum at the north edge. The start pixel distribution also has an unusual secondary maximum near the center of the arc, due primarily to spurious cloud identifications made near the sun, but outside the occulter mask in the 20 to 30 minute periods just before and after the mask intersects the arc of interest. A similar peak occurs in the center of the arc length distribution, implying that these are probably completely clear arc cases that have been misidentified as partial arc cases. Thus, the completely clear case frequency would be increased by almost 2%. Note also the the end pixel maximum exceeds the start pixel maximum by over 6%, indicating an azimuthal preference for clouds to the south compared to the north during the June mornings. :: As might be expected, the July PM results (Fig. 5) have greater frequency values than those from June AM, due to fewer completely clear cases. Maximum CFARCs starting at the southern edge now exceed those ending on the northern edge, suggesting a reversal in the cloud preference compared to June AM. The strong peak in the first arc length category (0° - 59) is probably caused by anomalous clear pixels appearing in completely cloudy CFARC situations. 4.2 Cumulative CFARC Length Distributions Users of CFARC information often ask for the probability of having an arc length of at least some specified value, or alternatively, the arc length that is met or exceeded some specified percentage of the time. Fig. 6 provides just such information, with the vertical axis depicting the probability of exceeding the maximum CFARC specified along the horizontal axis. For example, the maximum CFARC along the specified track is greater than 40° 90% of the time in the June AM sample,but only 48% in the July PM sample. Note also that the completely clear arc frequency appears at the right side of the diagram, and the completely cloudy frequency can be obtained by subtracting the value on the extreme left from 100%. Another interesting feature of these results is that fewer than 80% of the July PM cases have clear arcs of any length. the completely chopelely clear are for sample,but onlu Sone specified track: 1 The following four figures provide additional insight needed for interpreting the results in Fig. 6. Fig. 7 combines the WSI cloud determination image from Fig. 3 with the June AM curve from Fig. 6. The maximum CFARC for this case was 74°, as indicated by the green vertical line in the inset graph. This length was exceeded roughly 80% of the time in the June AM sample. In other words, 80% of the cases had maximum CFARCs clearer than this case. COM In the worst case scenario (July PM), the 80% level must include completely cloudy arcs, as shown illustrated in Figs 8 and 9. While it is easy to see that a selected arc will be completely cloudy in an overcast situation (Fig. 8), one must also realize that completely cloudy arcs also arise in cases with partial cloud cover (Fig. 9). In further contrast with the best case example in Fig. 7, Fig. 10 illustrates a case with a 35° maximum CFARC. Only 50% of the July PM cases exceeded this value. 4.3 Conditional Probability Distributions In many laser scenarios, the position of the maximum CFARC can be almost as important as its length. Tables 1 and 2 give the percentage occurrences of maximum CFARC length, given that the CFARC begins at certain positions along the arc. The data have been organized in twelve 10° categories. For example, 0.4% of the 1572 July PM maximum CFARCs started in the southern 30° - 40° zenith angle interval and extended into the 30° - 40° arc length interval, as did the case shown in Fig. 10. Fig. 11 gives a graphical illustration of the conditional probability tables. The blue and gray maxima in start pixel categories 6 and 7 (the region 10° on either side of zenith) in the June AM panel correspond to the secondary maximum noted previously in the start and end pixel distributions. The July PM blue maximum in arc length category 1 (0° - 5º) suggests that the peak in this category noted earlier in the bottom panel of Fig. 5 is localized in one part of the arc, and may be produced by nonuniformities in the imaging chip. Except for the problems just noted, the conditional probability values are quite reasonable, and are consistent in structure to those from other periods (not shown). 4.4 Concluding Remarks wiem Potential users of the WSI data base should find the results from this study encouraging. Previous comparisons reported in Shields, et al, (1990) showed reasonable agreement between overall sky cover estimates made by the WSI and the human observer. The CFARC results presented here imply that useful information regarding finer scale features, such as single arcs, can also be extracted from the WSI data. 5.0 REFERENCES McGuffie, K. and A. Henderson-Sellers (1989). Almost a Century of "Imaging" Clouds Over the Whole Sky Dome, Bulletin of the American Meteorological Society, Vol. 100, No. 10, 1243-1253. Shields, J. E., T. L. Koehler, M. E. Karr and R. W. Johnson (1990). Automated Cloud Cover & Visibility Systems for Real Time Applications. Tech. Note 217, Optical Systems Group, Marine Physical Laboratory, SIO, UCSD. AMA Texwww. 1.**.* www .***.. .. .... ........ ... . ....... AWW , , ww .. min .... ............................................. . . .................. .. w ww . w ww . wik . ..... ......... ... ............. ... .... ....... ...... . . wiwin' s how. ... ... ...www . Winsle mwil. com ............. . .. ................ . .. ... .. . . . W i th V n W i wid .. .' .H. n ..... . . .. . . . ... . mail ini .WALAY W Euriyyetl" in ' w w w .. ..... .... ... ..XXX .... *****...... . ... ..... w w w i . w .w' Value TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT - July (PM) 50 40- 30 - 20 - . 10 MPC 20 40 60 80 100 Completely Cloudy 120 Completely Clear | CFARC Value (degrees) Fig. 6 WHITE SANOS MISSILE RANGE C-STATION 21 JUNE 1989, 13:20 Z 9 TEREDE SaaS Prob. (%) Ττττ 0 40 80 120 CFARC Volue (degrees) Fig. 7 MPL Prob. (8) ht 40 80 120 WHITE SANDS MISSILE RANGE C-STATION 7 JULY 1989, 21:00 2 $88 CFARC Value (degrees) Fig. 8 WHITE SANDS MISSILE RANGE C-STATION 7 JULY 1989, 22:40 Z so $3.* es NE * SYAR88 382 6 2833 FATALURI OF. 2 MPE XXX SELE ed WXXN SES 22:48 23 2 :22 Prob. (8) 40 80 120 CFARC Value (degrees) Fig. 9 WHITE SANOS MISSILE RANGE C-STATION 23 JULY 1989, 23:40 Z WE ve 23 *** . . AK . МРО SA 1972 13 X ht 22 PARSON 30 SXXX SALES WA * ws ** 23 3 XXX:.. BER SA BW EXO XOXO Wii Prob. (8) CE 40 80 120 CFARC Value (degrees) Fig. 10 Conditional Cloud Free Arc Probability C-Station, June 1989 (AM) Probability (%) RAUEN VIBE VET Arc Length Start Pixel are o 9 Conditional Cloud Free Arc Probability C-Station, July 1989 (PM) ELE HEHE Probability (%) RUTE Arc Length 701938 Start Pixel Fig. 11 Table 1. Conditional maximum CFARC probability (%) given arc start position for the June AM sample totaling 1365 cases. 110° Arc Length Category 200 300 400 500 300 400 500 600 00 100 100 200 Zenith Angle Category to toto 600 700 800 toto.to 700 800 900 900 1000 toto 1000 110° 0.9 0.3 0.3 0.0 0.7 S: 0.4 0.0 0.1 0.5 0.4 0.0 0.0 0.0 0.1 0.7 0.0 0.0 0.1 0.6 -60° to -50° -50° to -40° -40° to -30° -30° to -20° -20° to -10° -10° to 0° 0° to 10° 10° to 20° 20° to 30° 30° to 40° 40° to 50° 50° to 60° 0.1 0.1 0.0 0.1 0.1 0.0 0.1 0.0 0.0 0.3 0.4 0.3 0.4 0.2 0.1 0.0 0.1 0.1 0.1 0.1 0.5 0.3 0.0 0.0 0.2 0.1 0.0 0.1 0.8 0.1 0.1 0.0 0.0 0.1 0.0 0.0 0.0 0.1 0.0 0.5 0.3 0.0 0.2 0.0 0.0 0.0 0.0 0.1 0.0 0.4 0.3 0.1 0.1 0.0 0.0 1.4 2.8 Start Pixel N Table 2 - Conditional maximum CFARC probability (%) given arc start position for the July PM sample totaling 1572 cases. 00 to 100 Zenith Angle Category Arc Length Category 200 300 400 500 to t o to to 300 400 500 600 100 to 200 600 to 700 700 to 800 800 to 900 900 to 1000 100° to 1100 -60° to -50° 4.3 1.4 . : 1.3 0.3 0.4 0.4 0.8 1.8 0.6 0.2 0.6 2.2 0.1 0.9 S. 1.3 0.6 0.1 0.3 0.4 1.8 0.3 0.3 0.3 0.3 2.6 0.1 0.5 0.1 0.8 1.4 0.3 0.4 0.9 0.3 0.4 0.3 0.2 0.4 -50° to 40° 40° to -30° -30° to -20° -20° to -10° -10° to 00 0° to 10° 10° to 20° 20° to 30° 30° to 40° 40° to 50° 50° to 60° Start 1.8 0.5 0.6 0.5 0.8 1.8 0.5 0.1 0.6 1.0 1.3 2.3 1.4 0.8 0.3 0.3 0.4 0.4 0.1 0.3 0.8 0.6 1.2 1.8 - 0.4 0.3 0.5 0.4 0.4 0.1 0.2 0.3 0.7 Pixel 0.1 0.4 Wap . Me 翻院 ​